ROLE OF TUBULIN GLYCYLATION IN MICROTUBULAR ORGANELLE BIOGENESIS

IN THE CILIATE TETRAHYMENA THERMOPHILA.

by

RUPAL MOHANDAS THAZHATH

(Under the Direction of Jacek Gaertig)

ABSTRACT

Tubulin glycylation is an evolutionarily conserved posttranslational modification, which occurs

on specific glutamic acid residues on the carboxy terminal tail domains of α− and β− tubulin.

Tubulin glycylation is highly evolutionarly conserved from protists to mammals, and is mostly

found in cells with cilia and flagella. In ciliates, the modification also occurs on non-ciliary

microtubules. Previous studies showed that the glycylation domain of β−tubulin has an essential

role in cell survival in Tetrahymena. However, not much is known about the cellular function of

this modification. The studies presented here explore the effects that a deficiency in the levels of

glycylation has on ciliary and cortical structures in Tetrahymena thermophila. A lethal β−tubulin

glycylation domain mutation, βDDDE440, was induced as a result of conjugation of genetic

heterokaryons. The study revealed completely opposite effects of the same mutation on cilia and

cortical organelles. The axoneme showed a lack of structural integrity, with an absence of the B-

tubule of peripheral doublets and the central pair. The mutants underwent consecutive and

incomplete rounds of cell cycle with defective cytokinesis. Cortical microtubular structures such

as longitudinal microtubules failed to undergo proper severing and appeared to impede the

cleavage furrow. Cilia in the glycylation mutants were mostly affected at the early stages of their

assembly with an initial delay in assembly of new cilia and complete failure in assembly of the

central pair. The modification was also required for long-term maintenance of proper structure

and length of cilia. Cortical microtubules such as longitudinal microtubules and basal bodies

showed signs of hypertrophy. A polarized effect was seen with the anterior subcell most affected

by the mutation. An arrest in nuclear and cortical growth and dissociation of a cortical

microtubule associated protein occurred selectively in the anterior half of the mutant cells. The

results presented here show multiple role of the tubulin glycylation domain in assembly and

maintenance of microtubules. Tubulin glycylation appears to act in a context-specific manner

and its function is dependent on both subcellular localization and antero-posterior cell polarity.

INDEX WORDS: Glycylation, Tetrahymena, Cytokinesis, Ciliary assembly, IFT,

Glutamylation, subunit exchange

STUDIES ON THE ROLE OF TUBULIN GLYCYLATION IN MICROTUBULAR

ORGANELLE BIOGENESIS IN THE CILIATE TETRAHYMENA THERMOPHILA.

By

RUPAL MOHANDAS THAZHATH

B.S, University of Mumbai, India, 1996

M.S, University of Mumbai, India, 1998

A Dissertation Submitted to the Graduate Faculty of the University of Georgia in Partial

Fulfillment of the Requirements for the Degree

DOCTOR OF PHILOSOPHY

ATHENS, GEORGIA

2003

2003

Rupal Mohandas Thazhath

All Rights Reserved

ROLE OF TUBULIN GLYCYLATION IN MICROTUBULAR ORGANELLE BIOGENESIS

IN THE CILIATE TETRAHYMENA THERMOPHILA.

BY

RUPAL MOHANDAS THAZHATH

Major Professor: Jacek Gaertig

Committee: Marcus Fechheimer

Edward T Kipreos

Michael McEachern

Boris Striepen

Electronic Version Approved:

Maureen Grasso

Dean of the Graduate School

The University of Georgia

December 2003

iv

DEDICATION

The work presented here is dedicated to my parents Prof. T. A. Mohandas and Prof. Shanthi

Mohandas, and other members of my family, specially Cannizzarro and my brother

Unnikrishnan who supported me through this extremely long and trying period. Mom and Dad,

your trust, love and patience with me have made me a better person. I would also like to dedicate

this to my family here in the United States: Abby, Kajoli and Atulya Tankha, Shipra Vaishnava,

Raghunandan Kandala and Dr. Vikas Dhingra. Abby, you are what kept me sane through the last

two years. Kajoli and Atul, you gave Abby and me unconditional love. Shipra, Raghu and Vikas,

you always listened and made sure I never once gave up on anything. A special thanks to Shane

Peters for all the help you have given me in the last two months, and to Dhruv Maniktala for

always caring. It has been moments with you, my family that has held everything together for me

through the past 5 years. Finally, I would also wish to dedicate this work to my grandmother who

was always proud of me, and unfortunately passed away before she could see me graduate.

v

ACKNOWLEDGEMENTS

The work presented here would not have been possible without the involvement of a

number of people. First, my mentor Dr. Jacek Gaertig has been an outstanding role model for me

to follow. He is honest, caring and very supportive. I could not imagine a better person to guide

me through the past five years. He is a source of encouragement and his scientific curiosity and

his excitement following the smallest discoveries are infectious to everyone in the lab. He has

helped me make the right decisions at all times during my graduate career here at UGA. My

Committee Members: Dr. Marcus Fechheimer, Dr. Michael McEachern, Dr. Edward Kipreos and

Dr. Boris Striepen have been patient and supportive. I would like to thank all of them for the

advice and exciting ideas they have put forward. Dr. Marcus Fechheimer has always gone above

and beyond the call of duty and has been tremendously supportive and helpful.

Dr. Mark Farmer, Dr. John Shields (CAUR) and Dr. Patrick J P Brown are my friends

and teachers. The microscopy presented here is a direct result of being trained by them. I would

like to thank them for their patience toward me, and my incessant questioning. They have

tolerated me for hours on end without complaint and were always there to help if needed. Dr.

Joseph Frankel at the University of Iowa has been instrumental in helping me with the intricate

details surrounding Tetrahymena cell biology and ultrastructure. The 12G10 and 13C4

monoclonal antibodies were raised by E. Marlo Nelsen and Dr. Joseph Frankel, Dr. Maria Jerka-

Dziadosz and her students at Warsaw University helped with some EM studies presented here. It

is thanks to Dr. Jerka-Dziadosz that certain aspects of the phenotype were revealed. Dr. Jeffery

Salisbury provided the 20H5 anti-centrin antibodies used extensively in this project. I would also

vi

like to thank Dr. Marty Gorovsky and Jianming Duan for providing the metallothionein construct

that was modified and used in some studies. Dr. Gianni Piperno kindly provided the 6-11 B-1

and Dr. Marie-Helene Bré and Dr. Nicolette Levilliers kindly provided the AXO49 and TAP952

antibodies. Chengbao Liu provided an important immunofluorescence image for the Nature-Cell

Biology paper, highlighting the presence of tubulin polyglycylation in the cortex of

Tetrahymena.

This work would not have been possible without the people in the lab who are not only

collegues, but also my friends. Dr. Dorota Wloga has been generous with sharing her data about

the NIMA kinases. She is also the reason why everything went smoothly in the lab as she made

sure equipment was in working order and reagents were in stock. Dorota and Kris Rogowski

helped me at a number of times both in and out of the lab. Dr. Jason Brown, a former graduate

student was always supportive and willing to discuss problems and results. Thanks to Marta

Gaertig for doing the dishes.

A note of thanks to everyone in the Cellular Biology Office, especially Kathy Vinson,

Genia King, Marian Thomas and Brett Rudolph for constant help and support. Finally, I would

like to thank Dr. Joe Crim for always supporting me with his enthusiasm, interest and excitement

for my work.

vii

TABLE OF CONTENTS

PAGE

ACKNOWLEDGEMENTS ........................................................................................................ v

CHAPTER

1. INTRODUCTION ......................................................................................... 1

2. REVIEW OF LITERATURE......................................................................... 5

3. POLYGLYCYLATION DOMAIN OF β-TUBULIN CONTROLS

AXONEMAL ARCHITECTURE AND AFFECTS CYTOKINESIS

IN TETRAHYMENA............................................................................ 79

4. FUNCTION OF THE β-TUBULIN GLYCYLATION DOMAIN IS

DEPENDENT ON MICROTUBULE TYPE AND CELL

POLARITY IN TETRAHYMENA ..................................................... 100

5. CONCLUSIONS ....................................................................................... 145

1

CHAPTER 1

INTRODUCTION

Microtubules are linear polymers of heterodimers of α and β tubulin, which play

many essential roles in eukaryotic cells. For example, the microtubule-based 9+2

axoneme, which comprises the structure of cilia and flagella, is responsible for the ciliary

beating, which generates cell motility. The mitotic spindle is assembled from

microtubules, and is responsible for chromosome segregation into daughter cells.

Microtubules are known to assemble complex and diverse structures, often within a

single cell. How cells are able to generate the sophisticated MT structures and achieve

their diverse functions remains largely unknown.

One way by which the properties of MTs in individual types of organelles can be

varied is through the posttranslational modification of tubulin, which is the building

blocks of MTs. MT properties are extensively altered, with the addition of various

modifications, most of which, have been found to localize to the COOH terminal tail

domains of tubulins located on the surface of MTs. Although the precise biochemical

function of these modifications remains to be uncovered, the evolutionary conservation

across species makes suggests that they are important.

Two types of so-called polymodifications, glycylation and glutamylation, are

evolutionarly conserved and localize to the COOH terminus of both α- and β−tubulin.

Glycylation was first identified in another ciliate, Paramecium by Andre Adoutte and

colleagues (Redeker et al., 1994). Anti-tubulin glycylation antibodies revealed the

2

presence of this type of modified tubulin in cilia and flagella of diverse organisms,

ranging from ciliated protozoa to mammals (Bré et al., 1996). Mass spectrometry studies

identified the possible sites on the COOH terminus and also uncovered the lengths of the

side chains that could be added on (Vinh et al., 1999). Glutamylation was first identified

in pig brain tubulin (Edde et al., 1990). Early studies carried out by injecting antibodies

against glutamylation caused disassembly of centrioles in mammalian cells, indicating a

role in assembly or maintenance of centrioles and structurally related basal bodies

(Bobinnec et al., 1998). Injection of antibodies against polyglycylation caused a loss of

motility in reactivated spermatozoa (Bré et al., 1996), leading investigators to believe a

possible role in locomotory functions. However, the exact mode of action of these

modifications in the function of MT based organelles remained to be elucidated.

This study uses the ciliate Tetrahymena thermophila, as a model to study the role

of tubulin glycylation in the formation and function of microtubular organelles.

Tetrahymena is an ideal system to study gene products with essential functions, with

routine knockouts and mutagenesis studies being carried out owing to incorporation of

exogenous DNA into the genome exclusively through homologous recombination.

Furthermore, Tetrahymena assembles diverse types of microtubular organelles, including

many types shared with mammals such as the cilia, basal bodies, and the mitotic spindle.

Also, Tetrahymena uses most if not all types of post-translational modifications of

tubulins, including glycylation.

Previous studies had shown that this modification is essential to survival of the

cell in Tetrahymena thermophila, although the exact mode of its action was unknown

(Xia et al., 2000). My research has led to characterization of the consequences of deletion

3

of sites of the modifications from tubulin proteins in vivo using novel genetic approaches.

This dissertation is organized into five chapters. Chapter 2 is a review of the literature

including an extensive review on studies on glycylation in Tetrahymena. Chapters 3 and

4 describe my own studies on the role of this modification in vivo in Tetrahymena

thermophila. Chapter 3 describes the study of a lethal glycylation domain mutant using a

novel strategy of germline mutant heterokaryons, which aids in the visualization of the

gradual change in phenotype from wildtype to mutant. The phenotype described in

Chapter 3 is investigated further in Chapter 4, where we were able to use an inducible

system and epitope tagging to achieve higher spatial and temporal resolution of events

involved in the manifestation of the phenotype.

Chapter 5 will summarize all the findings so far and the conclusions drawn from

these studies and propose further experiments to enhance our understanding of the

function of these polymodifications. In conclusion, this work has demonstrated an

important role for the tubulin glycylation domain in vivo, using the ciliate Tetrahymena

thermophila as a model.

References:

Bobinnec, Y., A. Khodjakov, L.M. Mir, C.L. Rieder, E. B., and M. Bornens. 1998.

Centriole disassembly in vivo and its effect on centrosome structure and function

in vertebrate cells. J. Cell Biol. 143:1575-1589.

Bré, M.-H., V. Redeker, M. Quibell, J. Darmanaden-Delome, C. Bressac, J. Cosson, P.

Huitore, J.-M. Schmitte, J. Rossier, T. Johnson, A. Adoutte, and N. Levilliers.

1996. Axonemal tubulin polyglycylation probed with two monoclonal antibodies:

4

widespread evolutionary distribution, appearance during spermatozoan maturation

and possible function in motility. J. Cell Sci. 109:727-738.

Edde, B., J. Rossier, J.P. Le Caer, E. Desbruyeres, F. Gros, and P. Denoulet. 1990.

Posttranslational glutamylation of alpha tubulin. Science. 247:83-85.

Redeker, V., N. Levilliers, J.-M. Schmitter, J.-P. Le Caer, J. Rossier, A. Adoutte, and M.-

H. Bré. 1994. Polyglycylation of tubulin: a post-translational modification in

axonemal microtubules. Science. 266:1688-1691.

Vinh, J., J.I. Langridge, M.-H. Bre, N. Levilliers, V. Redeker, D. Loyaux, and J. Rossier.

1999. Structural characterization by tandem spectroscopy of the posttranslational

modifications of tubulin. Biochemistry. 38:3133-3139.

Xia, L., B. Hai, Y. Gao, D. Burnette, R. Thazhath, J. Duan, M.-H. Bré, N. Levilliers,

M.A. Gorovsky, and J. Gaertig. 2000. Polyglycylation of tubulin is essential and

affects cell motility and division in Tetrahymena thermophila. J.Cell Biol.

149:1097-1106.

5

CHAPTER 2

REVIEW OF LITERATURE

MTs: Structure and Function

Microtubules (MTs) are ubiquitously present elements of the eukaryotic cytoskeleton.

MT based organelles are involved in essential cellular processes such as maintaining cell shape,

intracellular transport, cell motility, signal transduction, establishment of cell polarity and cell

division. The major structural component of MTs is a heterodimer, composed of two related

polypeptides α- and β-tubulin. The tubulin heterodimers assemble end to end to form linear

structures known as protofilaments, which then associate laterally to form tubules with a hollow

lumen of 25 nm diameter. MTs form a variety of morphologically and functionally distinct

structures such as cilia, flagella, centrioles, basal bodies (BBs) and the mitotic and meiotic

spindle. Little is known about the mechanisms used by cells to assemble distinct types of MT

structures, which often co-exist in the same cell.

The determination of the 3D structure of the α/β tubulin dimer based on cryoelectron

microscopy (Nogales et al., 1998) and the subsequent 20Å resolution model of the MT (Luduena

et al., 1992; Nogales et al., 1999) have helped increase our understanding of MT assembly and

function. The predicted structures of both α− and β-tubulin are nearly identical with each formed

by a core of 2 β sheets surrounded by 12 α helices. Each monomer can be divided into 3

functional domains: an amino terminal domain which contains the nucleotide (GTP) binding

domain, an intermediate domain which contains the paclitaxel binding site (for β-tubulin) and the

COOH-terminal domain which overlays the previous domains and sits on the outside surface of

6

the dimer. The COOH-terminal domain is probably involved in the binding of Microtubule

Associated Proteins (MAPs) and motor proteins. The atomic model lacks the last 18 and 10

COOH-terminal residues of α-tubulin and β-tubulin, respectively (Nogales et al., 1998).

Although not well resolved, the COOH-terminal tail regions are predicted to be located on the

outside surface of the MT (Nogales et al., 1998), as indicated by the accessibility of this region to

proteases, antibodies and influence on the binding of MAPs (Andreu and De Pereda, 1993;

Luduena et al., 1992; Sackett, 1995). The COOH-terminal tail domains are subjected to

extensive post-translational modifications (PTMs), whose roles are poorly understood at the

molecular level and are the central subject of this thesis.

MTs are polarized structures as the direct result of head-tail association of α/β-tubulin

dimers within protofilaments. The minus ends are located near the sites where MTs initiate their

growth, so called Microtubule Organizing Centers (MTOCs), such as the BBs and centrosomes

(Brinkley, 1985; Gould and Borisy, 1977; Kirschner, 1978; Pickett-heaps, 1969). The plus end is

distal to an MTOC and often extends to the plasma membrane. The high-resolution model of the

MT has also definitively established that the minus end has an outside ring of α-tubulin of the

proximal most dimer, while the plus end terminates with a crown of β-tubulin (Nogales et al.,

1999). Both α and β-tubulin are GTP -binding proteins. There are two GTP binding sites on the

dimer called E for exchangeable and N for non-exchangeable (Hesse et al., 1987). The E site is

located on β− tubulin while the N-site is located on α-tubulin. Microtubule polymerization is

accompanied by hydrolysis of the GTP bound to the E site on β-tubulin and the formed GDP

remains tightly bound to the MT lattice (Davis et al., 1994). The GTP in the N site on α-tubulin

is non-hydrolyzable probably because it is trapped at the interface between the two subunits

inside the dimer. The role of GTP binding to α-tubulin is unknown, although the binding of a

7

single magnesium ion to the N site plays a role in MT stability after assembly (Menendez et al.,

1998). The status of the nucleotide at the E site in the most distal dimer that faces the solvent at

the plus end of a MT is crucial for control of polymer properties. As long as the plus end has

GTP on the E site at the distal end, the MT is stable, can polymerize and protofilament ends have

a straight conformation (Nogales et al., 1999). When the rate of addition of new dimers to the

MT end, is exceeded by the rate of GTP hydrolysis within the polymer, GDP can be present at

the plus end terminal E site of the MT. The loss of the GTP cap with a conversion to GDP

tubulin leads to a dramatic increase in the curvature of the MT tip, leading to rapid

depolymerization (Muller-Reichert et al., 1998).

Stochastic changes in the size of the GTP cap led to the phenomenon of dynamic

instability (Mitchison and Kirschner, 1984), which can be observed in vivo and in vitro. During

dynamic instability, MTs undergo rapid transitions between elongation and shortening. Rapid

loss of subunits at the plus end leads to a shortening of the MT called a catastrophe, whereas

addition of subunits at the plus end leads to growth, a phase known as rescue (Walker et al.,

1988). This behavior is very important to function of MTs in vivo, as the dynamic state of MTs is

essential during mitosis and cell migration, and is therefore highly regulated (Kirschner and

Mitchison, 1986). Another important parameter in MT kinetics is the critical concentration (Cc),

which is the concentration of free tubulin dimers when the rate of polymerization of MTs, equals

that of disassembly leading to a steady state. At dimer concentrations above the Cc, dimers

polymerize into MTs with preferential addition at the plus end. At concentrations below Cc, MTs

depolymerize. In vivo, the concentration of free dimers is always lower than the Cc, but the

MTOCs can accelerate the nucleation of MTs at lower dimer concentrations, providing cells with

a tool for a tight control of polymerization state of MTs.

8

The process of biogenesis of MT organelles includes a series of events; starting from

transcription of tubulin genes to assembly of MTs. Furthermore, already assembled organelles

are subjected to maintenance mechanisms, which are required for renewal of used subunits and

organelle size regulation. There is a tight regulation of tubulin gene transcription and translation,

based on feedback mechanisms that sense the existing levels of tubulins in the cell, a process

referred to as autoregulation of tubulin synthesis (Ben-Ze'ev et al., 1979; Caron et al., 1985a;

Caron et al., 1985b; Cleveland et al., 1981; Cleveland et al., 1983b). Tubulin production can be

regulated at the transcriptional level, differentially in different tissues (for review see MacRae

and Langdon, 1989). However, autoregulation through the modulation of mRNA stability is the

best-known mechanism in vertebrate cells (Cleveland, 1988; Cleveland, 1989). For example,

mammalian cells regulate the levels of tubulin mRNAs by a posttranscriptional mechanism,

sensing the concentrations of unpolymerized dimers in the cell. Treatment of mammalian cells

with either a MT depolymerizing drug or a MT polymerizing drug, leads to a decrease and

increase in the tubulin mRNA levels respectively (Ben-Ze'ev et al., 1979; Cleveland, 1989;

Cleveland et al., 1981; Cleveland et al., 1983a).

However, the regulation of respective α and β-tubulin subunit levels appears to be

mediated by different pathways. The sequence necessary and sufficient for the selective

degradation of β-tubulin mRNA in response to changes in the level of free tubulin subunits

resides within the first 13 nucleotides of mRNA that encode the amino-terminal sequence of β -

tubulin, Met-Arg-Glu-Ile (MREI) (Yen et al., 1988). The MREI sequence is involved in co-

translational recognition of the nascent amino-terminal β-tubulin tetrapeptide as it emerges from

the ribosome. This leads to a co-translational degradation of β-tubulin mRNA, which could be

mediated through binding of one or more cellular factors to the β-tubulin nascent peptide.

9

Interestingly, the mechanism of autoregulation of tubulin synthesis appears to be absent in

invertebrates and protists, which use the initiation of transcription as the main level of regulation.

Gu et al demonstrated that the ciliate, Tetrahymena thermophila responds to MT

depolymerization (and resulting increase in the unpolymerized dimer concentration) by

increasing α-tubulin gene transcription, thereby increasing α-tubulin synthesis and α-tubulin

mRNA levels, without affecting rates of tubulin message turnover (Gu et al., 1995). Surprisingly,

a MT polymerizing agent, paclitaxel (which increases the mass of the polymer versus

unpolymerized tubulin) had the same effect, which indicated that the response to any

perturbation in the cytoskeleton leads to an increase in α-tubulin gene transcription. In addition,

there appears to be a differential response to events affecting the ciliary components versus those

that affect the cytoplasmic MT network. The response to removal of cilia (deciliation) involved

an increase in transcription of all 3 major tubulin genes, the single major α-tubulin ATU1 and the

two major β-tubulin genes, BTU1 and BTU2. However, a disturbance in the cytoplasmic MT

network using agents that do not affect cilia, led to an increase in the α-tubulin transcription with

an increase in mRNA levels of only one of the two major β-tubulin genes, BTU1 (Gu et al.,

1995). Thus, Tetrahymena and possibly other eukaryotes have complex feedback type and gene-

specific mechanisms that regulate tubulin gene expression.

After translation, both α and β-tubulins, are captured (separately) by chaperonin

complexes, called CCT (Kubota et al., 1994; Rommelaere et al., 1993). The CCT complexes are

barrel shaped conglomerates, which take unfolded tubulins into their lumen and partially fold

while hydrolyzing ATP. CCT is also involved in folding of other cytoskeletal proteins, including

actin and γ-tubulin. In order to produce assembly competent α/β tubulin heterodimers, an

additional set of proteins called cofactors (A, B, C, D, E) are also required (Cleveland et al.,

10

1978; Gao et al., 1993; Melki and Cowan, 1994; Tian et al., 1996; Tian et al., 1997). After

release from the CCT chaperonins, partially folded α and β tubulins follow different pathways.

α-tubulin is captured by cofactor B and β-tubulin is captured by cofactor A, which are

subsequently replaced by cofactors E and D, respectively. The two pathways then converge with

the formation of a quaternary complex (α-tubulin/E + β-tubulin/D). In the final step, co-factor C

binds the complex and upon GTP hydrolysis, assembly-competent α/β tubulin heterodimers are

released (Lewis et al., 1997).

In vivo, MTs are organized into various structures that are often anchored at the MTOCs,

for example, BBs in ciliated/flagellated cells and centrosomes in mammalian cells (Brinkley,

1985; Gould and Borisy, 1977; Kirschner, 1978; Pickett-heaps, 1969). The study of the

components of the MTOCs, have revealed the ubiquitous presence of γ-tubulin, which has been

found to be essential for MT nucleation (Horio et al., 1991; Joshi et al., 1992; Marshall et al.,

1996; Oakley et al., 1990; Sobel and Snyder, 1995; Stearns et al., 1991; Sunkel et al., 1995).

Overexpression of γ-tubulin leads to abnormal ectopic sites of MT nucleation (Shu et al., 1995).

γ-tubulin functions in a 2.2 MDa complex with a number of associated proteins. Isolated

complexes form capped ring structures with the diameter close to the diameter of the MT (25

nm) (Zheng et al., 1995). It is likely that the γ-tubulin ring provides a direct template for the MT

polymerization by associating with the ring of α-tubulin at the minus ends (Li and Joshi, 1995;

Moritz et al., 1995; Spang et al., 1996; Zheng et al., 1995). Consistently, in vitro studies showed

that γ-tubulin is a minus-end MT binding protein (Li and Joshi, 1995). Thus, tubulin

heterodimers are captured at the MT nucleating sites in vivo, and are assembled into various

functional arrays, depending on structure and location in the cell. In the case of cilia and flagella,

11

the BB serves as the template for the assembly of the axonemal structure (to be described

below).

MTs are organized into diverse organelles ranging from simple cytoplasmic networks, to

more complex structures such as, the mitotic spindle, BBs, centrioles and axonemes. It is

important to note that MT structures differ not only at the level of organization, but also differ in

their stability. Some organelles are transient such as the mitotic spindle, as they appear only at

specific stages of the cell cycle. Other organelles are more stable, such as centrioles and the BBs,

which persist throughout the cell cycle. MTs also show varying levels of polymer growth

dynamics and tubulin subunit turnover. The spindle MTs turnover more rapidly whereas those in

the BBs /centrioles and cilia (see my results in Chapter 4) show extremely slow turnover.

Most MTs including ubiquitous cytoplasmic and nuclear spindle MTs, have 13

protofilaments. The number of protofilaments may be derived from the diameter of the γ-tubulin

nucleating ring because MT assembled from purified dimers without γ-tubulin in vitro have a

highly variable number of protofilaments. There are specialized types of MTs that show a

different tubular arrangement, such as those in centrioles, BBs, cilia and flagella. BBs and related

centrioles have 9 triplet structures arranged radially, whereas the axoneme of cilia and flagella is

usually comprised of 9 outer doublet MTs and 2 central singlets. The MT triplets are made of a

complete 13-protofilament tubule A and incomplete 10-protofilament tubules B and C, fused to

each other. The doublets have only the A and B-tubule. The structure of both the axoneme and

BB, which is pertinent to this study, will be described in detail later. To summarize, MTs are

capable of forming very complex and diverse structures within the same cytoplasm without the

presence of any intervening membranes. How this is accomplished, remains largely unknown.

12

In order to maintain spatial and temporal control over the assembly processes, cells must

be able to generate biochemical differences between the distinct functional types of MTs. One

radical solution utilized by many organisms is expression of tubulin primary sequence variants

(isotypes), which are adapted for assembly of only certain types of organelles. Multiple isotypes

of α and β tubulins are widespread among eukaryotes, occurring in plants, animals, fungi and

protists (for review see Luduena, 1998). There is a wide variability in the degree of sequence

divergence among isotypes of an organism, with the trend leaning toward an increased number

of primary sequence variants in multicellular organisms and a low number in lower eukaryotes

(with a possible special case of ciliates – see below). In some cases the tubulin isotypes

(identified by standard cloning and amplification approaches) in the same organism may differ

from each other at only one position, as is the case of α-tubulin from the green algae Polytomella

and Volvox (Conner et al., 1989; Harper and Mages, 1988). However, it is possible that more

divergent sequences are also present in these organisms, which have not been detected so far,

and may only be apparent if their genomes are completely sequenced. It has been observed that

in some cases, expression of different sequence variants contributes to the adaptation of MTs to

specific functions in multicellular organisms. For example, in Drosophila, BBs contain only β1

tubulin and axonemes are assembled exclusively from β2 tubulin (Kemphues et al., 1982; Raff et

al., 1997). These two tubulins differ at only 25 of 447/446 amino acids respectively. It is

interesting to note that 8 of these differences lie near the COOH-terminal tail, the region

subjected to extensive post-translational modifications (see below) and involved in binding

MAPs. When the potential of β1 tubulin to function in axonemes was tested by placing its

coding region under male germ-line-specific promoter, it was found that β1 alone could not

generate axonemes (Nielsen et al., 2001). Thus relatively small sequence differences between the

13

two isoforms could determine differences in the MT biogenesis capacity, although the two

organelles (cilia and centrioles/BBs) posses a similar architecture (i.e. the BB serves as a

template for the assembly of the axoneme and both exhibit the characteristic nine-fold

symmetry). Importantly, when the COOH-terminal of the β1 isoform was replaced with that of

β2, 9+2 axonemes were assembled, although the resulting structures were shorter than normal

(Nielsen et al., 2001). Even more strikingly, replacement of just 2 amino acids of the COOH tail

of β1 to β2-specific amino acids was sufficient to allow the assembly of 9+2 axonemes by the

resulting protein chimera (Nielsen et al., 2001).

Specific isoforms are involved in generation of some very unusual MTs. In C. elegans, a

specific β-tubulin isotype is required for the assembly of 15 protofilament MTs in the touch

receptor neurons (Savage et al., 1989). In the moth, a specific β-tubulin isotype, which is

expressed only in the sperm, gives rise to unusual 16 protofilament MTs. Drosophila lacks such

structures, however, expression of this particular isotype in flies led to the assembly of 16-

protofilament structures (Raff et al., 1997). Thus, isotypes confer their unique assembly

properties in a heterologous system, suggesting that to some extent the sequence affects the

intrinsic properties of the resulting MTs. Vertebrates also have multiple isotypes differing mainly

in the tail domains. Interestingly, there are isotype orthologs across different vertebrate species,

which show specific and conserved patterns of tissue-specific expression. Less is known about

the function of isotypes in vertebrates, but recently a lineage-specific β- isotype was found to be

important in differentiation of mammalian blood platelets (Schwer et al., 2001). The biochemical

basis of isotype specificities remains unclear. However, in vitro experiments showed that

different isotype mixtures have different polymerization kinetics and stabilities (Banerjee and

Luduena, 1992). It is also possible that there are isotype differences in binding of MAPs. Finally,

14

there is a strong possibility that some isotype differences are related to the presence or absence of

post-translational modifications (see below).

However, many cell types assemble complex and diverse MT organelles while utilizing

only a single major isotype of the α/β tubulin dimer, For example, Drosophila male germ cells

assemble axonemes and several additional types of MT organelles in other cells, such as sensory

cilia from a single isotype of α- and β-tubulin (Hoyle and Raff, 1990; Hutchens et al., 1997). In

general, cells with axonemes tend to have a limited number of isotypes present, which appear

multifunctional. It appears that the presence of an axoneme constrains evolution of tubulin

isotype and selects against sequence divergence (Gaertig et al., 1993). More divergent tubulins

appear to participate in the assembly of non-axonemal MTs such as those forming the networks

and the spindle in fungi.

Less is known about to what extent isotype diversity contributes to generation of

differences in MT-based organelles within a single cell. In some cases, isotypes were found to be

sorted to distinct organelles. For example, an extremely divergent β-tubulin of Tetrahymena,

Btl1p, can assemble into the micronuclear mitotic spindle, but is excluded from cilia (M.

Gorovsky, personal communication). In cases of isotype exclusion, it is possible that, either they

lack organelle-targeting determinants, translation of their mRNAs is spatially restricted or they

lack ability to polymerize into certain types of organelles. The exact mechanism of Btl1p sorting

and its significance remain to be established. In the case of Btl1p, unfortunately the isotype

deletion did not produce an informative phenotype (Clark and Gorovsky, personal

communication). In the cases where different isotypes copolymerize, having a more

heterogeneous composition of MTs could affect the polymerization kinetics and stability as

shown in vitro by Luduena (Luduena et al., 1992). However, as already mentioned, there are

15

many examples of cell types, which assemble diverse types of MTs from a single type of α/β

primary sequence dimer.

On the other side, MT diversity can also be achieved through posttranslational tubulin

modifications (PTMs). It has been thought that PTMs play an important role in the

differentiation of MT based organelles, since they are very diverse biochemically and are

evolutionarily conserved. Furthermore, in many cases PTMs could work in concert with the

isotype diversity. For example, the primary sequence of a particular isotype could carry or lack a

particular site for a PTM. Thus, one purpose of expressing multiple isotypes in multicellular

organisms might be to enable the action a certain PTM in particular cell lineages and organelle

types. For instance, glutamylation is concentrated on centrioles, mitotic spindle and neuronal

bundle MTs, whereas glycylation is found in ciliated and flagellated cells. Thus, it is quite

conceivable that PTM enzymes might, in some way specify the structure, and/or function of the

MT based organelle, by recognizing and modifying a specific tubulin isotype.

There are a number of tubulin-related proteins (also known as alternative tubulins) that

are found in various cell types, the association or action of which, could affect the assembly,

structure or function of various MT types. These include and are not limited to γ-, δ- and ε-

tubulin, which participate in MT nucleation and control the protofilament structure (Chang and

Stearns, 2000). The alternative tubulins are quantitatively minor as compared to α and β-

tubulins. γ-tubulin is essential for nucleation of the cytoplasmic network MTs and duplication of

BBs in Paramecium andTetrahymena (Ruiz et al., 1999; Shang et al., 2002a). The mammalian δ-

tubulin was found associated with centrosomes of non-ciliated cells, and in the manchette and

axonemes of spermatid cells (Smrzka et al., 2000). ε-tubulin is a component of the centrosomes

with a non-overlapping localization with γ-tubulin (Chang and Stearns, 2000). δ-tubulin and ε-

16

tubulin have been implicated in the process of BB/centriolar assembly. Silencing of either δ- or

ε-tubulin gene led to a block in duplication of the BBs in Paramecium (Dupuis-Williams et al.,

2002; Ruiz et al., 2000). Mutants lacking functional versions of these tubulins show defects in

the ordered assembly process of the triplet structure of BBs/centrioles. Strikingly, δ-tubulin

mutants lack the C-tubule of the triplet structure and therefore assemble abnormal centrioles with

doublet MTs (Dutcher and Trabuco, 1998; Ruiz et al., 2000). ε-tubulin mutants cannot assemble

the B-tubule and this results in BBs with singlet MTs lacking both the B and C-tubules (Dupuis-

Williams et al., 2002). A ζ- tubulin encoding sequence was identified in Trypanosoma although

localization and function of this protein remains unknown (Vaughn et al., 2000). However, fungi,

and some invertebrates such as C. elegans and Drosophila lack apparent homologs for δ, ε and

ζ-tubulins. On the other side, C. elegans lacks typical centrioles or BBs (O'Connell, 2000),

Drosophila assembles singlet or doublet MTs of the BBs (Callaini et al., 1997) and fungi lack

centrioles altogether. Thus, these genes could have been lost from the genomes of these

organisms as they may be required for assembly of the typical triplet MTs of centrioles and BBs

but not their simplified versions.

Another mode of achieving MT diversity is through the binding and action of MAPs.

This group of factors includes structural MAPs, molecular motors and plus end binding proteins.

Structural MAPs are proteins, which lack a nucleotide-dependent activity and bind to MTs

reversibly with affinities in the low micromolar range. Structural MAPs such as MAP2 and Tau

promote tubulin polymerization and contribute to MT stability by suppressing dynamic

instability and promoting bundle formation (Gustke et al., 1994). Molecular motors are protein

machines, which utilize the energy from ATP hydrolysis to move directionally along cytoskeletal

filaments, and often transport specific cargo. The two known classes of MT motors, kinesins and

17

dyneins, each consists of an extended family of related proteins that function in organelle

transport, spindle formation, chromosome motility in dividing cells, and cell motility. Certain

MAPs could be specifically located in a given area of the cell, conferring a distinct biochemical

property in that location. Indeed, pursuant to this hypothesis, many MAPs show a restricted

subcellular localization. One way to achieve these spatial differences is to restrict mRNA

localization and subsequently restrict protein translation. In neurons, the axon specific MAP,

Tau, is localized in the cell body and the proximal segment of the axon, whereas MAP2 is

restricted to the cell body and the dendrites (Litman et al., 1994; Litman et al., 1993). It has also

been shown in Naegleria, that there is a specific relocalization of flagellar protein mRNAs

during the differentiation from amoeba to temporary flagellate (Han et al., 1997). These mRNAs

localize to the base of the growing flagella, adjacent to the BB and MTOC. However, restricted

mRNA localization alone cannot account for the targeting of MAPs to specific MTs within

organelles (such as cilia or flagella). These organelles lack the protein translation machinery and

all components are synthesized in the cell body and transported to the growing cilium or

flagellum. Thus, the proper attachment of the dynein arms to specifically the A tubule of the

outer doublets and the localization of Klp1 kinesin to only one specific MT of the central pair

(Bernstein et al., 1994) cannot be explained by differential localization of mRNA alone. Thus,

there might be some structural determinants in the MT itself, such as differences in the lattice of

protofilaments or surface modifications of MTs, which could aid in this process. For example,

the lattice differences between the A- and B- or C-tubules could be responsible for differential

binding of some MAPs. It is also possible that one of the protofilaments of the axonemal

doublets is made of a non-tubulin protein called tektins (Pirner and Linck, 1994). It is more

difficult to explain the longitudinal periodicity of binding of MAPs along certain MTs. For

18

example, the sets of radial spoke complexes and inner dynein arms bind to doublet MTs of

flagella with 96 nm periodicity, in some cases even in vitro. Furthermore, there is longitudinal

stratification of MAP isoforms within the flagellum, with different subtypes of inner dynein arms

present in different segments of the axoneme (Piperno et al., 1990). It is very tempting to

speculate that secondary modifications of tubulin proteins, which affect the tail domains on the

MT surface, could play a role in altering the structural and functional properties of MTs (to be

described in detail in a following section). In agreement with this hypothesis, some tubulin

modifications, such as tubulin glutamylation, affected the in vitro binding of certain MAPs to

tubulins (Bonnet et al., 2000; Boucher et al., 1994; Larcher et al., 1996).

After the tubulin dimers are assembled into polymers, they are subject to the action of

tubulin modifying enzymes that act on the assembled MT. Modifications have been speculated to

change either the intrinsic properties of MTs themselves, or to act via a secondary mechanism by

modulating the binding and/or action of MAPs and motor proteins. Most PTMs are generally

considered to be post-assembly events, as they tend to accumulate on MTs after organelle

assembly. The vast majority of literature regarding PTMs, represent morphological studies on

their differential distribution within cells, tissues and organs, which tell us little about the

function of these evolutionarily conserved mechanisms. Only recently, has the biological

significance of PTMs, come to light, in large part as a result of research from our laboratory

(Rosenbaum, 2000). The following section will attempt to address the properties and functions of

various PTMs identified so far.

Acetylation: Acetylation occurs on K40 at the N-terminal domain of α-tubulin (L'Hernault and

Rosenbaum, 1985; LeDizet and Piperno, 1987), and was first identified in Chlamydomonas

19

flagellar MTs (L'Hernault and Rosenbaum, 1983). The enzyme responsible for acetylation in

Chlamydomonas α-tubulin (acetylase) is localized to the flagellum and is highly specific for α-

tubulin (Greer et al., 1985; Maruta et al., 1986). The enzyme does not require polymerized

tubulin as a substrate although its activity on the dimer is weaker as compared to polymerized

MTs. A tubulin deacetylase activity, which catalyzes the reversal of this modification on

α−tubulin is located in the cytoplasm in Chlamydomonas which correspond with the lower

abundance of acetylation on cell body MTs (Maruta et al., 1986). This PTM was until very

recently, thought to be related in some manner to the stabilization of MTs, due to the

accumulation of the modification on a subset of long-lived MT networks. However, this

modification does not cause MT stability, rather it appears to be a cellular marker for longevity

of the MT structure. Expression of a non-acetylatable α-tubulin in Chlamydomonas by

substituting the K40 to A or R (up to 80% of total α-tubulin) had no detectable effect on the cell

(Kozminski et al., 1993). Mutating K40 of the major Tetrahymena α−tubulin (Atu1p) to an

unmodifiable residue had no detectable effect on the phenotype (Gaertig et al., 1995).

Importantly, although the nearly complete genome sequence revealed several additional and

highly divergent α−tubulin gene sequences, none of them has a lysine at the position

homologous to K40 of Atu1p and α-tubulins of other organisms (J.G personal communication).

Furthermore, acetylated tubulin could not be detected using epitope-specific antibodies in the

mutants with R40 codon in the ATU1 gene of Tetrahymena (Gaertig et al., 1995). Thus, it can be

safely concluded that Tetrahymena does not require acetylation of K40 for survival and this

process does not affect any major cellular activities, or Tetrahymena cells can adapt easily to

lack of this mechanism. K40 acetylation is also dispensable in C.elegans (Fukushige et al.,

1999). Mutation in a specific α-tubulin gene mec-12 (G69A) affected the formation of 15-

20

protofilament MTs in the axons and resulted in a touch-insensitive phenotype. mec-12 encodes

an α−tubulin (MEC-12), which is the only acetylatable α−tubulin in C. elegans. However, a

mec-12 gene with a point mutation leading to lack of K40 acetylation site was able to rescue

touch sensitivity phenotype of mec-12 mutants without any apparent consequences on the

phenotype of the organism, indicating that acetylation on α−tubulin is not essential in C.elegans.

However, there might be a subtle function of this PTM in higher eukaryotes. A recent

breakthrough was the identification of the gene, HDAC6, whose product encodes a tubulin

deacetylase, responsible for removal of acetylation from K40 in vivo and in vitro (Hubbert et al.,

2002). Hubbert et al. showed recently that decreasing the levels of K40 acetylation by

overexpressing the tubulin deacetylase HDAC6, leads to an increase in the chemotactic motility

of murine fibroblasts (Hubbert et al., 2002). Reciprocally, blocking the function of HDAC6

deacetylase with a small molecule inhibitor tubacine led to a decrease in cell motility and

disrupted the localization of p58, a MAP that mediates the binding of Golgi elements to MTs

(Haggarty et al., 2003). Lack of HDAC6 gene in ES cells was not lethal but slightly decreased

the rate of proliferation. The effects of HDAC6 on MTs appear not to be mediated by modulation

of polymer stability but rather by changing binding of MAPs (Palazzo et al., 2003). These results

could indicate a possible role for α−tubulin acetylation/ deacetylation in regulating the activity of

MAPs and/or motor proteins that play roles in cell motility.

Detyrosination: The terminal Y at the COOH terminus of α−tubulin undergoes cycles of

removal and addition. The removal is achieved through the action of an α-tubulin specific

carboxypeptidase (Arce and Barra, 1983; Kumar and Flavin, 1981) that acts preferentially on

polymeric tubulin as a substrate (Wehland and Weber, 1987b). The removal of the terminal Y or

21

detyrosination, is found like acetylation, on more stable MT networks, and occurs after MT

assembly (Webster et al., 1990). This was clearly demonstrated in a study using gold labeled

antibodies and TEM, that showed detyrosination on segments of Trypanosoma pellicular MTs

which had assembled early, with no detection on newly assembled plus ends of the MT (Sasse

and Gull, 1988). The enzyme tubulin tyrosine ligase (TTL) is responsible for the conversion of

detyrosinated tubulin back into its unmodified form (Ersfeld et al., 1993). This enzyme is ATP

dependant, and is highly specific for α-tubulin detyrosination and prefers the tubulin dimer as

substrate. Wehland et al showed that the binding site for the enzyme is on β-tubulin rather than

on α−tubulin (Wehland and Weber, 1987a). The differences in the requirement for soluble

versus polymeric tubulin between the reverse and forward enzyme may result in a cyclic process

of tyrosinated tubulin used for MT assembly and detyrosination post-assembly. After

depolymerization, detyrosinated tubulin could be re-tyrosinated and reused with newly

synthesized tyrosinated tubulin for MT assembly (Gundersen et al., 1987).

As K40 acetylation, detyrosination appears to be the result and not the cause of increased

MT stability. For instance, MTs stabilized by taxol are rapidly detyrosinated in cells, which

usually do not show a high level of this modification (Wehland and Weber, 1987b). Increased

detyrosination, achieved by injecting antibodies against TTL, did not change the stability of MTs

in vivo (Webster et al., 1990). In cancer cells, increased tubulin detyrosination, apparently

resulting from the suppression of TTL activity and the resulting unbalanced activity of tubulin

carboxypeptidase, appears to provide a strong growth advantage (Lafanechere et al., 1998).

Chang et al showed that inhibiting the detyrosination of MTs in myoblasts with an inhibitor 3-

nitrotyrosine blocked morphological differentiation and accumulation of muscle specific factors,

revealing a connection between posttranslational modifications of MTs and progression of the

22

myogenic program (Chang et al., 2002). Acetylated and detyrosinated α−tubulin often occur on

the same MTs but their rates or patterns of appearance are not the same (Bulinski et al., 1988;

Sasse and Gull, 1988), indicating differential regulation and function. In the already mentioned

study by Chang et al using 3-nitrotyrosine as a detyrosination inhibitor, myogenesis was

prevented when MT detyrosination was blocked without affecting acetylation, indicating that the

two modifications are independent functionally (Chang et al., 2002).

Detyrosination appears to strongly affect the interactions of MTs with the motor protein

kinesin-I and this interaction may be required for proper distribution of other cytoskeletal

components (Kreitzer et al., 1999). Vimentin intermediate filaments preferentially co-align with

detyrosinated MTs in cultured mammalian cells (Gurland and Gundersen, 1995), and this

interaction could be disrupted by injecting detyrosinated α−tubulin molecules, which were

chemically altered to prevent their polymerization (presumably acting as a competitor for factors

binding to modified tubulins). Also, these distribution of intermediate filaments (IFs) along

detyrosinated MTs appears to be mediated by kinesin-I motor as injection of anti-kinesin also led

to a collapse of the IF network in vivo. In vitro studies with a non-polymerizable detyrosinated

α-tubulin as a competitor showed a strong inhibition of the binding of kinesin to MTs. Thus, it

appears that kinesin moves IFs specifically along detyrosinated MTs.

Phosphorylation: This modification was first identified in the rat brain β-tubulin (Eipper, 1972).

It occurs on the COOH-terminal tail domain of β-tubulin (S444) (Alexander et al., 1991; Diaz-

Nido et al., 1990) and accompanies neuronal differentiation in vivo (Aletta, 1996; Gard and

Kirshner, 1985). Axonemal tubulin also appears to be phosphorylated. Sea urchin axonemes

show phosphorylation of both α and β-tubulin (Otsubo et al., 1975). In Chlamydomonas

23

axonemes, one study reported the apparent phosphorylation of α-tubulin, using radiolabeled 32P

(Piperno and Luck, 1976). The study however, did not separate the labeled proteins at high

resolution, thus leaving the possibility that another phosphorylated protein might be co-migrating

with α-tubulin. β-tubulin was shown to be hyperphosphorylated in neurons from Alzheimer’s

disease patients (Vijayan et al., 2001). There appears to be a role for this modification in

adaptation to cold. Most MTs are sensitive to low temperature. However, phosphorylated tubulin

was enriched in a small cold stable population of neuronal MTs of vertebrates (Alexander et al.,

1991; Diaz-Nido et al., 1990) and in the cold stable MTs of the marginal band of avian

erythrocytes (Rudiger and Weber, 1993). It was also shown that β-tubulin is phosphorylated in

the cold-adapted Antarctic ciliate Euplotes focardii (Pucciarelli et al., 1997). The precise

function of this modification however, has not been addressed experimentally.

Palmitoyation: Palmitoylation was identified in porcine brain and the site of this modification is

conserved in all known sequences of α -tubulin (Caron, 1997). Cys-377 was identified as the

primary site of this PTM on α -tubulin (Ozols and Caron, 1997). Palmitoylation may be the only

PTM that exists in yeast (Caron et al., 2001). The modification consists of the covalent linkage

of the long chain fatty acid, palmitate, to cysteine residues via thioester bonds. In general, this is

a reversible protein modification, which regulates membrane-associated activities (Mumby,

1997). For example, palmitoylation results in the sequestration of some proteins such as NO-

synthase (Garcia-Cardena et al., 1996) and proteins involved in T-cell receptor signaling within

membrane domains (Zhang et al., 1998). This modification also determines protein-protein

interactions. Palmitoyation of the β-adrenergic receptor prevents its phosphorylation by cAMP

24

dependent protein kinase C, which in turn reduces its ability to interact with G proteins (Moffet

et al., 1996). Thus, the modification could localize soluble tubulin or MTs to the membranes.

Both α and β−tubulins are palmitoylated in resting human platelets and the levels are

decreased during thrombin activation, in association with the disassembly of the peripheral band

of MTs that are juxtaposed to the cell cortex (Caron, 1997). A C377S mutation in S. cerevisiae

α-tubulin results in a 50-60% drop in the levels of tubulin palmitoyation (Caron et al., 2001).

Numerous MT organization defects were observed including excessive number, abnormal

lengths and/or misplaced MTs, which also affected positioning of the mitotic spindles through

the bud neck. Thus, palmitoylation may play a role in a subset of MTs during nuclear migration.

It appears that this modification is involved in anchoring the MT plus ends to the cell cortex,

possibly regulating the length of MTs during different stages of the cell cycle. C377 is conserved

in the Atu1p α -tubulin of Tetrahymena. However, mutating this amino acid to alanine did not

affect cell growth and motility (Xia and Gaertig, unpublished observations). It remains to be

established whether the mutation affects nuclear migration (during conjugation), as it is the case

in yeast. It is also possible that additional modification sites exist in Atu1p, which act

synergistically with Cys 377.

Glutamylation: Glutamylation is one of the 2 types of the so-called polymodifications that

localizes to the COOH-terminal tail domain of both α and β -tubulin. Glutamylation occurs by

addition of either a single glutamic acid or a chain of polyglutamic acid of variable length to the

γ -carboxyl group of specific Es in the primary sequence (Audebert et al., 1993; Bre et al., 1994a;

Edde et al., 1990; Gagnon et al., 1996; Schneider et al., 1998; Schneider et al., 1997). This PTM

occurs in many organisms including mammals and unicellular eukaryotes and is particularly

25

abundant in nerve cells and cells with axonemes. However, glutamylation also labels spindle

MTs and increases in abundance during mitosis (Bobinnec et al., 1998b). Glutamylated tubulins

localize to the BBs, nuclear spindle MTs and cilia in Tetrahymena and Paramecium (Bre et al.,

1994b). In mammalian fibroblasts, this PTM is concentrated on centrioles and is less abundant

on cytoplasmic MTs. Injection of monoclonal antibodies specific to the glutamylated tubulin

caused disappearance of the centrioles (Bobinnec et al., 1998a). Thus, glutamylation may be

important in maintaining already assembled centrioles and could have a role in centriole

assembly. It was shown earlier, that the same antibody interfered with motility of in vitro ATP-

reactivated axonemes, probably by affecting the binding of dynein to the B-tubule of the outer

doublets (Gagnon et al., 1996). Glutamylation has also been implicated in the differential

regulation of the binding of structural and motor MAPs to tubulin in vitro. This regulation might

facilitate the selective recruitment of these MAPs into distinct MT populations, hence

modulating their functional properties (Boucher et al., 1994; Larcher et al., 1996).

Interestingly, this modification is not restricted to tubulin. A monoclonal antibody

directed against glutamylated tubulin detected additional glutamylated proteins present in HeLa

cells (Regnard et al., 1999), which were identified as NAP-1 and NAP-2, members of the

nucleosome assembly protein family (Regnard et al., 2000). NAPs play a role in deposition of

core histone complexes onto chromatin. Recently, Westermann and Weber purified the tubulin

glutamylase activity from the flagellate Crithidia and found a single protein, which was

identified as the member of the NIMA kinase family, CfNek (Westermann and Weber, 2002). It

is not known yet whether CfNek is the catalytic subunit of tubulin glutamylase. Westermann

could not obtain any glutamylase or kinase activity for CfNek in a heterologous system. It is

possible that CfNek is a kinase, which is an upstream regulator of tubulin glutamylase. In

26

agreement with this hypothesis, Edde and collegues identified additional peptides in the tubulin

glutamylase from murine brain and found no evidence of a NIMA kinase (Regnard et al., 2003)

(B. Edde, personal communication).

Glycylation: This modification is structurally related to glutamylation. It was first discovered in

Paramecium by Andre Adoutte and collegues. These authors noticed that antibodies against

ciliary tubulin of Paramecium decorated MTs in other species but only in cells with cilia and

flagella (Adoutte et al., 1991). A subsequent study showed that the epitopes of these antibodies

appeared on newly assembled ciliary MTs with a significant delay after assembly, suggesting

that the antibodies recognized a secondary modification which slowly accumulated on already

assembled MTs (Fleury et al., 1995; Levilliers et al., 1995). Mass spectrometry identified the

modification as glycylation (Redeker et al., 1994). This polymodification occurs on the COOH

terminus of both α and β-tubulin where glycine chains of varying lengths are added on to

specific conserved glutamic acid residues in the primary sequence (Vinh et al., 1999). These

glutamic acid sites tend to be evolutionarily conserved especially in cells with axonemes (Bré et

al., 1996). Glycylation was detected directly by mass spectrometry analysis of tubulins from

diverse sources, including protists such as Giardia (Weber et al., 1997) and mammals (Rudiger

et al., 1995). This modification is found mainly in ciliated and flagellated cells but very recently

was also found in mammalian brain tubulin (Banerjee, 2002). It is possible that glycylated

tubulin in the brain originated from axonemes of cilia on the surface of neurons and ependymal

cells. Monoclonal antibodies specific to glycylated tubulins inhibited the motility of reactivated

sea urchin spermatozoa (Bré et al., 1996), suggesting a role in regulation of ciliary dynein. The

length of the glycine side chains varies depending on subcellular location. Monoglycylation is

27

the addition just one glycine, and is found on all MT types in Tetrahymena and Paramecium.

(Bré et al.,1998). Polyglycylation is the addition of polyglycine side chains of 2 or more residues

and mass spectrometry studies have shown that up to 34 glycine residues can be added onto a

single tubulin molecule (Redeker et al., 1994). Recent studies revealed that glycylation also

occur on a 150kD non -tubulin protein in the cell body of Tetrahymena (Gorovsky, unpublished

results). The function of tubulin glycylation is the main focus of this thesis and experiments on

the role of this PTM will be described below.

Ubiquitination: Ubiquitin is an abundant 76 amino acid polypeptide, which can be covalently

conjugated to specific proteins by the formation of an isopeptide bond between its COOH

terminus and the amino group of a lysine residue of the target proteins. This process requires a

number of enzymes that ensure its specificity. The ubiquitin-activating enzyme E1 forms a thiol

ester linkage between a cysteine residue and the COOH terminal glycine in ubiquitin in an ATP

dependent manner (Hochstrrasser, 1996). This activated moiety is then transferred to one of

several possible carrier proteins (E2) via the formation of another thiol ester linkage. Finally,

ubiquitin is transferred to the substrate with the assistance of an E3 ligase or ligase complex.

Several internal residues within ubiquitin can be used as substrates to create polyubiquitinated

chains, which can act as signals for proteosome-mediated degradation.

Given the mechanisms regulating tubulin polypeptide synthesis and the subsequent

complex pathways of folding into the heterodimer, it is inevitable that at some point along the

way, misfolded tubulin monomers are produced. Furthermore, depolymerization of MTs also

leads to an increase in tubulin degradation (Cleveland, 1989). The misfolded and depolymerized

tubulins, which are toxic to the cell, are degraded by an unknown mechanism. There is evidence

28

indicating a possible role of misfolded tubulin in the Parkinson’s Disease (PD). A gene known as

parkin, which is a protein ubiquitin E3 ligase, has been linked to the Autosomal Recessive

juvenile Parkinson’s Disease (ARPD) (Kitada et al., 1998). Mutations in this gene are found in

patients with ARPD. Ren et al showed a strong co-localization and affinity of parkin for MTs

(Ren et al., 2003). This strong binding led to an increased tubulin ubiquitination and an

accelerated degradation of α- and β-tubulins in human embryonic kidney cells. Ubiquitinated

tubulins were also detected in rat brain lysates (Ren et al., 2003). The mutated versions of parkin

found in PD patients however, failed to either ubiquitinate or degrade tubulin suggesting

degradation of tubulin plays an important role in the brain.

Tetrahymena thermophila as a model to study microtubule organelles and the role of

tubulin posttranslational modifications

To study the functions of PTMs in vivo, one can take three approaches: 1) mutate the

sites of modification on tubulins so that the modification cannot be added, 2) inhibit or remove

the forward modifying enzyme or 3) hyperactivate the reverse enzyme. Most of the enzymes

responsible for PTMs remain unknown, including those responsible for tubulin glycylation and

glutamylation. Only two reverse PTM enzymes have been identified: K40 deacetylase HDAC6

(Hubbert et al., 2002) and TTL (Raybin and Flavin, 1977a; Raybin and Flavin, 1977b). Little is

known about enzymes depositing or removing polymodifications. In the absence of any

information about forward PTM enzymes, the best approach is to target the modification sites on

tubulins. Our laboratory’s choice of the model organism for these studies is Tetrahymena

thermophila and we believe that this is the most optimal organism long term, for use of all of the

3 approaches listed above. There are several drawbacks to using other model systems for these

29

studies, even though some may have better developed genetic and biochemical methodologies.

Yeast apparently lacks most PTMs (Alfa and Hyams, 1991) and the flagellate Chlamydomonas

does not show a high level of homologous recombination, making it difficult to introduce

mutations or knockouts. Tetrahymena is a rapidly growing ciliate which shows tremendous

potential for use as a genetic tool. Also, the recently sequenced 100 MB genome of Tetrahymena

with 8X redundancy provides an important tool for protein gene identification and subsequent

localization and knockout studies. Integration of exogenous DNA into the genome occurs

exclusively by homologous recombination, and thus, mutated genes can be targeted easily to a

given locus (Turkewitz et al., 2002).

The main genetic feature of this organism is the nuclear dualism, which allows for some

unusual approaches in bringing phenotypes to expression (see below). Tetrahymena has a

somatic macronucleus (MAC), which determines the phenotype during the vegetative life cycle,

and a transcriptionally silent micronucleus (MIC). During cell division, the MIC divides

mitotically while the MAC divides amitotically with random segregation of its chromosomes.

This leads to phenotypic assortment where heterozygous MACs randomly segregate alleles at

each division, eventually becoming homozygous (Doerder et al., 1992). Tetrahymena can also

undergo a developmental process of conjugation where under starvation; two cells of different

mating types undergo a phase of sexual reproduction (Frankel, 1999). Soon after cell pairing,

meiotic prophase occurs during which the 10 bivalent chromosomes form the so called crescent

structure. After meiosis only 1 of 4 postmeiotic haploid nuclei undergoes an additional mitotic

division. The remaining 3 postmeiotc nuclei undergo degradation. The mitotic division of the

retained haploid nucleus results in a migratory and stationary pronucleus. Each member of the

mating pair exchanges a migratory pronucleus and the fusion of the migratory and stationary

30

pronuclei result in diploid zygotic micronuclei present in each exconjugant. The zygotic

micronucleus undergoes 2 postzygotic divisions yielding 4 nuclei positioned as pairs at the two

opposite poles. The posterior two nuclei develop into MICs while the anterior two develop into

MACs. One of the MICs degenerates. The old MAC degenerates by an apoptotic mechanism.

During the first cell division after conjugation, the remaining MIC divides mitotically, while the

MACs are segregated into daughter cells without division. Thus, the process of conjugation leads

to replacement of the parental MAC by a new MAC, which is a product of zygotic MICs.

Consequently, genes that were once transcriptionally silent in the MIC of the parent strains are

now brought to expression in the MAC of the progenies. The activation of expression in the

silent genome during conjugation can be used in creative way to obtain inducible phenotypes

(see below).

We can routinely manipulate genes in both the MIC and MAC, using biolistic gun

transformation protocols. One could create knockout heterokaryons with disrupted or otherwise

mutated copies of the gene to be studied in the MIC retaining WT copies in the MAC (Hai and

Gorovsky, 1997). Strains with a lethal or deleterious mutation in the MIC can be maintained

indefinitely in the vegetative life cycle because the mutated gene is silent. Two such strains can

be mated and the effects of mutating the gene on the phenotype can be observed in the

conjugation progeny, even if the gene is essential and the phenotype is terminally lethal. On

conjugation, the lethal mutations are expressed and there is a transition from the wild type to the

mutant phenotype owing to the gradual replacement of pre-existing wild type gene products with

mutant ones, in the progeny. This offers a window of opportunity to sample and study the effects

of the mutation, and progression of the phenotype over a period of time. If the phenotype is

lethal or selectable, versions of the mutated gene can also be introduced into the conjugation

31

progeny of heterokaryons without any flanking selectable markers, by phenotypic rescue. This

strategy has been used with considerable success in studying the effects of eliminating and/or

mutating many genes in Tetrahymena and was effectively used to study the effects of mutating

the known glycylation sites on both α and β tubulin (see below). I used the heterokaryon method

to analyze the phenotype of a lethal mutant of the glycylation domain (described in detail in the

appendix) and the results are described in Chapter 3.

Another recent advancement in the use of Tetrahymena as a genetic tool was the

development of an inducible gene expression system by Shang et al (Shang et al., 2002b).

Tetrahymena has a non-essential metallothionein gene, MTT1, whose product is used to

neutralize and accumulate cadmium, and whose promoter is induced by the presence of

cadmium. Shang et al have engineered this locus to be the site of an inducible expression system.

One can easily create a construct with the coding region of the gene of interest flanked by the 5’

and 3’ UTRs of the MTT1 gene, under control of the MTT1 promoter. Such a fragment can then

be targeted to the MTT1 locus by biolistic transformation and homologous recombination. Thus,

one could switch on expression of a given gene by adding cadmium to the medium. I used this

strategy was used to create a novel conditional tubulin glycylation mutant and the results of that

study are discussed in Chapter 4.

Ciliates represent a striking case of MT diversity, comparable to what is usually found in

an entire multicellular organism. Tetrahymena thermophila can assemble up to 17 different types

of MTs (Gaertig, 2000). These microtubule-based structures assemble into complex and diverse

organelles such as the cilia, cell cortex, cytoplasmic networks and nuclear MTs in both the

mitotic micronuclear spindle and macronuclear assemblages during amitosis. In ciliates there

appeared to be little sequence variation among tubulin genes (Gaertig et al., 1993). However, this

32

view needs to be reevaluated in the light of a huge amount of information from the recently

sequenced genome. Tetrahymena has a single “conventional” α-tubulin gene ATU1, which is

essential. Tetrahymena also has two redundant and multifunctional β-tubulin genes, BTU1 and

BTU2 (encoding the same protein), and the presence of either one is essential for survival of the

cell (Xia et al., 2000). Surprisingly, recent sequencing of the Tetrahymena genome has revealed

5-7 additional α and β-tubulin genes, which show 45-65% of sequence identity to the

conventional tubulins Atu1p, Btu1p and Btu2p (Jacek Gaertig, personal communication). The

function of divergent tubulins remains to be explored, with the exception of Btl1p, which was

found to be non-essential for growth and mating but may have a subtle role in the micronucleus

during mitosis, where it was found to localize (M. Gorovsky, personal communication).

Tetrahymena also has most if not all types of alternative tubulins including γ, η, and ε tubulins.

The Tetrahymena “conventional” (non-divergent) α and β-tubulins, Atu1p and Btu1/2p occur in

a very large number of isoforms, which are created by diverse PTMs. These include acetylation,

detyrosination/tyrosination, and two polymodifications, glycylation and glutamylation.

Subcellular localization of modifications in Tetrahymena

Here I will discuss the localization of acetylation, glutamylation and glycylation in

Tetrahymena, as probed by the modification-specific antibodies. The localization of these three

modifications in Tetrahymena is similar to that in Paramecium. α-tubulin acetylation as probed

by the monoclonal anti-acetylated tubulin antibody 611B1, localizes to cilia, the cell cortex and

intra-micronuclear MTs. This modification is excluded from the more labile intracytoplasmic

MT network (Gaertig et al., 1995). Glutamylated tubulins were found in cilia, BBs, the

micronuclear spindle and intracytoplasmic MTs using a monoclonal antibody GT335 (Bre et al.,

33

1994b). Tubulin monoglycylation is found on all types of MTs using the TAP952 antibody,

which recognizes a single glycine attached to the primary sequence of tubulin (Callen et al.,

1994). However, in cilia, this modification is found to be highly concentrated on tips.

Polyglycylation, as probed by the monoclonal antibody AXO49, which detects 3 or more

glycines attached to the primary sequence of tubulin (Levilliers et al., 1995), was found on

ciliary and all types of cortical MTs (Chapter 3). Interestingly, polyglycylation appears to be

excluded from ciliary tips, where the PTM is limited to monoglycylation.

It appears that some of the modifications have a differential localization within the

axoneme. Pechart et al used immunogold TEM analysis with antibodies directed against various

tubulin isoforms (Pechart et al., 1999). In Paramecium, the distribution of the acetylated form of

α-tubulin was uniform along the whole length of the cilia, the central pair and the MT doublets.

With glutamylation, only the BB and the proximal part of the axoneme were strongly labeled.

This modification showed a proximo-distal decreasing gradient along the length of the axoneme,

with dramatic decrease in the transitional region, which was six times less compared to the

adjacent axoneme. Conversely, mono and polyglycylated tubulin isoforms were more

concentrated in the transitional region than in the BB MTs. The axonemes of Paramecium also

contain more of mono and polyglycylated tubulin isoforms than the BB. In mammalian

spermatozoa, a differential labeling was detected between the peripheral outer doublets (Kann et

al., 1998). There was a predominant labeling of glutamylated isoforms in peripheral doublets 1-

5-6 and a predominant labeling of monoglycylated epitopes in doublets 3-8. In sea urchin

axonemes, the B tubules have more glycylated isoforms than the A tubules of the outer doublets

(Multigner et al., 1996). Care must be used when comparing the localization data obtained using

antibodies specific to modified tubulin, in different organisms. This is because, all these

34

antibodies are directed against specific tubulin sequences along with the presence of the

modification. As a result, subtle differences in the primary sequence may alter the binding

properties of the antibody, in turn affecting the degree and extent of labeling. Thus, the

differential localization observed across a range of organisms may be attributed to the

differences in the primary sequences of various modified isotypes recognized by the antibody. In

fact, recent studies by Kann et al have shown a differential expression and distribution of mono-

and polyglutamylated tubulin isoforms in various cellular models using two antibodies directed

against the glutamylated form of tubulin, GT335 and B3 (Kann et al., 2003). Finally, the

accessibility of the modified epitopes to antibodies could be compromised by MAPs.

Unfortunately, PTM-specific antibodies are so far the only approach that can be used to detect

the isoforms in intact organelles.

Studies on the role of Tubulin Glycylation in Tetrahymena thermophila

Glycylation occurs on several clustered glutamic acid residues on the COOH terminus of

both α and β-tubulin in Paramecium. (Vinh et al., 1999). These sites are also conserved in

Tetrahymena tubulins (Table 2.1). Two monoclonal antibodies are available that are specific for

the two forms of the modification: TAP952 detects a single glycine attached to the glutamic acid

at the COOH terminal tail of tubulin. The epitope includes some primary sequence but

unmodified COOH-terminal peptide is not recognized. AXO49 detects glycine chains of 3 or

more glycines. This antibody does not appear to recognize the primary sequence (Callen et al.,

1994; Levilliers et al., 1995). The antibodies revealed polyglycylation in the cilia of

Tetrahymena and showed that the modification is excluded from all types of cytoplasmic and

nuclear MTs. A very faint signal was detected with the anti-polyglycylation antibody on the

35

cortical MTs (see Chapter 4). Thus, it appears that the numerous cilia on the cell surface, conceal

the presence of the modification on the MTs that make up the cell cortex. This was confirmed by

deciliating the cells to reveal that the modification was present on all types of cortical MTs

including longitudinal MTs (LMs), Transverse MTs (TMs), the contractile vacuole pores (CVPs)

and post-ciliary MT bundles (PCs). Monoglycylation is found on all types of MTs in

Tetrahymena including the nuclear spindle, macronuclear MTs, cytoplasmic MTs, cortical MTs

and cilia, with the strongest labeling at the tips of the cilia (Xia et al., 2000). It appears therefore

that the length of lateral chains is under strict regulation and longer chains are produced in a

smaller subset of organelles. The enzymatic machinery responsible for glycylation is unknown

but it is possible that two enzymatic activities are involved, a monoglycylase which adds to first

glycine onto the γ-carboxyl group of a glutamic acid in the primary sequence, and a

polyglycylase which extends the chain by adding more glycines. However, it is also possible that

a single enzyme catalyzes both steps and its activity is either inhibited more strongly, or that it is

more strongly counteracted by deglycylases in the cell body.

In order to determine the sites of modification in Tetrahymena, site directed mutagenesis

was used to make charge-conserving substitutions at the potential glycylation sites to aspartic

acids. Tubulin knockout heterokaryons were used for rescue transformation using either mutated

α−tubulin or β−tubulin genes (Xia et al., 2000). For α-tubulin mutational studies, α-tubulin

knockout heterokaryons were used, that had the major essential α-tubulin gene, ATU1 disrupted

by the neomycin resistance cassette in the MIC and had wildtype ATU1 alleles in the MAC.

After mating of two such heterokaryons, the progeny lack functional α-tubulin, and die unless

rescued with a functional copy of the ATU1 gene. This method can be used to quickly assess

functionality of mutated versions of the gene. Failure to rescue would indicate a possible lethal

36

effect of the mutated gene or sensitivity to the drug selection method (a selective drug,

paromomycin, is used to eliminate non-conjugating cells which do not express the neo marker

gene). Initial studies indicated that the modification was dispensable on α-tubulin as mutating 3

potential sites of modification at positions E445-447 (based on homology to the sites of

modification identified by MS in Paramecium, Table 2.1) had no effect on the cell (Table 2.2).

However, the triple site mutants showed a small amount of glycylation when tubulin where

analyzed by high resolution 2D western blotting (Gao and Gaertig, unpublished data) or by mass

spectrometry (Redeker and Bre unpublished data). It was hypothesized that the 3 remaining

glutamic acids at positions 438, 441 and 443 serve as additional sites, possibly only when the

primary sites at positions 445-447 are missing. To address this possibility, ATU1 heterokaryons

were rescued with all six sites mutated (α6D) (Duan and Gorovsky, unpublished data). Viable

transformants were obtained. The α6D mutant cells showed slow growth and slow pair

formation during conjugation. Western blots using probes against the mono- and polyglycylated

forms showed a complete elimination of the modification on a-tubulin (Rogowski and Gaertig,

unpublished results). Because glutamylation sites are also located somewhere on the 6 Es of the

COOH-terminal tail domain, we can safely conclude that neither glycylation nor glutamylation

are essential on α-tubulin but could play a role in pair formation during mating. Interestingly,

MTs were observed forming around the tips of cells during early stage of the membrane fusion

during conjugation in the ciliate Blepharisma (Bedini et al., 1978). It is possible that α−tubulin

polymodification(s) play a role in organization or function of MTs that are associated with the

development of conjugal junction, for example, by directing membranous organelles to the

plasma membrane.

37

When I first started working in the lab, a former graduate student Lu Xia was

investigating the role of the putative sites of glycylation on β-tubulin. Knockout heterokaryons

were constructed and used for the rescue transformation that had both major β-tubulin genes

BTU1 and BTU2 disrupted by drug resistance cassettes. It was found that E437 was a major site

of modification in Paramecium (Redeker et al., 1994) with the three additional E’s downstream

of this site also polyglycylated (E438-440)(Vinh et al., 1999). There is additional potential site at

E442 in Tetrahymena. A series of mutations was made at the E’s at positions 437-442 (Xia et al.,

2000) (Table 2.3). It was found that no single site was essential for survival of the cell. Also,

double substitutions in varying combinations at these sites had no detectable effect on the

phenotype of the cell. However, most gene fragments containing triple and quadruple mutations

failed to rescue.

Interestingly, one of the triple mutations, which had the E at position 437 unchanged,

with the other 3-4 sites downstream changed to D’s, was a viable mutant. The phenotype of these

mutants showed slow growth and motility, occasionally defective cytokinesis and small size.

What appeared to be a defect in the positioning of the cleavage furrow was also observed, but

later (in my own studies –see the following Results chapters) was identified as a combination of

cytokinesis failure and lack of growth in the anterior member of the subdivided cell.

When the levels of polyglycylation were tested on β-tubulin in the βEDDD440 mutant,

there was a dramatic decrease (only 24% of that present in WT). No signal was detected when

western blots were probed with anti-monoglycylated antibodies (this could be the result of lack

of epitope recognition by the antibody and some monoglycylated tubulin could still be present).

Surprisingly, there was a reduction in polyglycylated α-tubulin detected with the AXO49

antibody though not as dramatic as that on β-tubulin. Even more surprising was more than two

38

fold increase of the levels on monoglycylation on α-tubulin in this mutant. It was the first

indication that there might be a functional interaction between the modification sites on the

COOH-terminal tails of α and β tubulin, which might be modulating the levels of the

modification in vivo. Usually, more of glycylation is observed on β−tubulin compared to α-

using western blots. It may just be a quantitative difference between the two tubulins with

overlapping functions, and thus the removal of the modification from α-tubulin could be

tolerated. If the total amount of polyglycylation on MTs is important, it may be possible to

rescue a lethal β-tubulin mutation if one could shift some polyglycylation from β-tubulin to α-

tubulin by increasing the number of sites available. However, if mutated β-tubulin was unable to

assemble into functional structures, transferring the modification onto α-tubulin should not be

able to rescue the lethal β-tubulin mutations.

To distinguish between the two possibilities, a chimeric α-tubulin was created (Table 2.3)

which had the COOH-terminal tail of β-tubulin was used to replace the corresponding sequence

of α-tubulin. Thus, the construct had most of the sequence of α-tubulin fused to the COOH-

terminal tail domain of β-tubulin. When this construct was used in conjunction with a lethal

triple β-tubulin mutation, transformants were obtained that were viable and showed a normal

phenotype. Western blot analysis showed that there was an increase in the level of glycylation on

α-tubulin, with a reduced signal on β-tubulin. Thus, the modification appears to migrate with the

tail sequence from β-tubulin to α-tubulin. We can conclude therefore that the subunit location of

the modification sites is not important. The same strategy was subsequently used to assess

transferability of the function of all 9 putative polymodifications sites on β-tubulin. When all 9

glutamic acids on β-tubulin were mutated, not surprisingly this set of mutations by itself failed to

yield any viable mutants and was presumably lethal (Duan and Gorovsky, 2002). However, when

39

the chimeric α/β-tubulin was co-transformed with this 9-D mutation of β-tubulin, viable

transformants were obtained with an apparent normal phenotype. Western blot analyses showed

a complete elimination of mono- and polyglycylation and glutamylation on β-tubulin (Gao and

Gaertig, unpublished results) and an increase in polymodifications on the α/β tubulin chimera. It

is worthwhile mentioning that in addition to glycylation being added to these COOH-terminal

glutamic acid residues, glutamylation may also act at the same or nearby sites. The exact sites of

modification in the case of glutamylation are unknown. However, in the βDDDDGD442 mutant

co transformed with the α/β chimera, the levels of polyglutamylation detected by antibodies

were unaffected on β-tubulin, suggesting that the possible sites are upstream of these residues

(Xia et al., 2000). In the β9D mutant co-transformed with the chimera however, both

polyglutamylation and glycylation were removed from β-tubulin. This meant the subunit location

is not important for both glycylation and glutamylation. This result also argues that there is

nothing wrong with folding of α- and β-tubulin mutants lacking the sites of polymodifications,

as long as the modifications are present on the partner subunit in the dimer.

The COOH termini of both tubulins may also have other essential functions in addition to

PTMs because they cannot be deleted. Duan and Gorovsky demonstrated that the tail is required

for the essential function of both α and β-tubulin (Duan and Gorovsky, 2002). However, the tails

are interchangeable and cells grow normally with either tail type on both tubulins. The highly

acidic COOH-terminal tails were not resolved in the high resolution model of the MT, making it

difficult to propose a structural basis for the essential function of the tail. However, this function

could be related to proper folding of the dimer, or motor/MAP binding. For example, the tail

domains are known to be required for efficient folding of tubulins in vitro using CCT chaperonin

and tubulin cofactors (Fontalba et al., 1995). PTMs are probably not essential for proper folding

40

of tubulins as most if not all are not present at that time and are mostly added on after the

polymer is assembled. Also, polymodifications do not occur in yeast and tubulin folding is

essential in this organism. Thus, the function of the COOH-terminal domain in folding, likely

represents the primary sequence involvement rather then PTMs.

Until this point, it was known that the presence of glycylation in sufficient amounts was

required for survival in Tetrahymena and that the subunit location of the modification sites was

not important. However, the exact functional role of the modification remained elusive. The

studies described in the following chapter on the lethal βDDDE440 mutant using mutant

heterokaryon strains yielded very surprising results, which placed the mechanism of action of the

modification in cilia and cell cortex, (see Results Chapters). Next, I will describe in detail, what

is known about the cortical units and associated structures such as the cilium, in Tetrahymena.

Cell architecture and surface pattern formation in Tetrahymena thermophila

Tetrahymena cells have a very obvious polarity: the oral apparatus, the feeding organelle,

marks the anterior part of the cell whereas the contractile vacuole pores mark the posterior end of

the cell (Frankel, 1999). The entire cell cortex is covered with BBs, and associated cilia. The

polarity is also manifested by orientation of several cytoskeletal appendages associated with the

BBs. Longitudinal MTs (LMs) run from the anterior to the posterior ends of the cells. In cross

section, a WT array of these MTs are usually comprise of 7-9 protofilaments. However,

individual MTs of the LM bundles do not span the entire length of the cell. Rather an LM bundle

is a system of partly overlapping single MTs (Frankel, 1999). Perpendicular to these MTs are the

transverse MTs, short bundles of 8 protofilaments. Transverse MTs originate from the anterior

end of the BB structure. The post ciliary MTs (PCs) originate from the posterior ends of the BBs

41

whereas the kinetodesmal fiber, the only non-MT based structure associated with the BB runs

anterior to the BB. The oral apparatus also has a very precise arrangement of structures both in

terms of intrinsic organization and spatial arrangement with respect to the whole cell. Cilia

surrounding the oral cavity are connected into clusters called membranelles or membranes whose

coordinated beating funnels and filtrates food particles. An undulating membrane (UM) and 3

associated membranelles (which comprise the Adoral Zone of Membranelles, AZM) form the

oral apparatus along with about 120 associated cilia. The UM is made up of a double file of BBs

of which only one is ciliated. The BBs of the oral apparatus also show specific organization

where lengths of the post ciliary fibers vary from row to row within the AZM. Only the posterior

most row within a membranelle has a well developed postciliary fiber whereas the two anterior

ones assemble only a single post ciliary fiber (Jerka-Dziadosz, 1981).

Cell division in Tetrahymena thermophila

Prior to cell division in Tetrahymena, a number of events must take place such as

duplication or assembly of certain structures to be transmitted to the daughter cell. A new oral

apparatus (OA) forms in the mid body region of the cell, just below the presumptive position of

the future cleavage furrow, to be inherited by the posterior daughter cell. The old CVPs are

inherited by the posterior daughter cell while new CVPs are assembled in the posterior part of

the future anterior daughter cell. Breaks occur in LMs and BB rows in the cleavage furrow

region. Contraction of the fission furrow is associated with the formation of a cortical

subdivision free of cortical elements. Constriction of the cleavage furrow leads to separation of

the two daughter cells. The final stage of division is accompanied by a phenomenon called

42

rotokinesis, which involves a rotational movement of the anterior daughter in a counterclockwise

direction to break the narrow bridge remaining between the two cells (Brown et al., 1999a).

Basal bodies: Structure and Function

Centrioles and BBs are structurally similar. Centrioles are found in pairs near the nucleus

and serve to organize the cytoplasmic MT arrays and mitotic spindles during cell division. BBs

anchored below the cell membrane give rise to cilia or flagella. In mammalian epithelial cells,

BBs originate from modified centrioles. There is also interconversion of centrioles and BBs in

Chlamydomonas (for review see Beisson and Wright, 2003).

BBs and centrioles show a complete 13 protofilament A-tubule, to which are attached the

incomplete (10 protofilament) B and C-tubules. Nine such triplets are then arranged in a nine

fold symmetry to form the typical “centriolar structure”. This geometry is strikingly conserved

with rare exceptions. For example, C. elegans centrioles have 9 singlets (O'Connell, 2000) and 9

doublets have been observed in the BBs of Mastigamoeba and Drosophila (Callaini et al., 1997;

Simpson et al., 1997). The assembly and duplication of the BB triplet structure, still remains

under investigation although recent work has now identified the possible sequence of assembly

and some key players in the process. It appears that some unknown factors act to initiate the

assembly process at key points in the cell cycle, be it BBs or centrioles. It appears that the nine-

fold symmetry is laid down before assembly begins, starting with the appearance of a generative

disc (Dippell, 1968). Mutational analyses and other studies in various systems have revealed

that γ-tubulin and centrin are recruited to the assembly site before the formation of a MT scaffold

with α and β-tubulin (Suh et al., 2002). There is a sequential polymerization of the A, B and C

tubules in that order (Dippell, 1968). η-tubulin appears to be involved in this initial stage

43

whereas ε-tubulin and δ-tubulin appear essential for subsequent addition of the B and C tubules,

as described earlier. It is not clear whether ε and δ- tubulins play a role in shaping the B and C

tubule respectively or whether they stabilize the organelle after assembly. Studies done in

Chlamydomonas and Paramecium indicate that ε-tubulin is necessary for the cohesion of the

centriolar structure (Dupuis-Williams et al., 2002; Dutcher et al., 2002). These tubulin isoforms

may thus be involved in stabilization of the structure. Thus, although not entirely elucidated, this

process is an ordered assembly of multiple proteins at defined sites in the cell, depending on the

cell type.

Many studies have led to the belief that the assembly of a new centriolar structure can

proceed via two partially distinct mechanisms. These are commonly referred to as the templated

and de novo pathways. In the templated assembly process a new BB/centriole is formed at right

angle to the pre existing structure, thus it is assumed that the mother structure serves as a

template for the assembly of the daughter structure. The best known example of this mode of

duplication is the duplication of the centrosome in animal cells. The centrioles of the centrosome

are always found in pairs and at right angles to each other. Prior to duplication, the two centrioles

separate. Each acts as a template for the formation of a new daughter at right angles to the

preexisting one. Thus, the centrosome duplication is semi-conservative. Another example of the

apparent templated mechanism is the duplication of BBs within the somatic rows of cilia in

Tetrahymena. The preexisting BB serves as a template for the formation of a new daughter BB.

The new BB forms just anterior to the old one within the cortical rows, presumably using the

information from the older organelle as a template. However, it has been well documented in

various systems that BBs and centrioles can also assemble in the absence of a preexisting

structure. In CHO cells, after centrosome ablation by laser microsurgery, de novo formation of

44

pericentriolar material, followed by reappearance of centrioles was observed (Khodjakov et al.,

2002). De novo BB formation has also been observed in the ciliate Oxytricha during the

emergence of cells from the cyst stage devoid of BBs into the vegetative stage, which has a

precise patterning of BBs (Hasimoto, 1963). In Chlamydomonas, cells with a vfl1 mutation are

aflagellate, but can reform BBs de novo after a couple of generations (Marshall et al., 2001). In

addition, as will be described in a later section, it appears that Tetrahymena is also capable of

assembling BBs associated with cortical rows de novo after γ-tubulin depletion and subsequent

reintroduction (Shang et al., 2002b). Thus, it has been postulated that cells can utilize either the

de novo or the templated pathway, but cells prefer the kinetically dominant templated pathway,

repressing the de novo pathway (Marshall et al., 2001).

The cortex of Tetrahymena is very well organized in that the orientation of ciliary units

and the insertion of newer ones occur in a very precise manner. The detail involved in the

assembly of these structures is very well coordinated with the cell cycle and the positioning and

orientation of the newer units is based on the older pre-existing units. i.e that the older units may

act as so-called template structures for the formation of newer ones (Allen, 1969). In 1975, D. L.

Nanney studied precisely this process by using protargol stained dividing Tetrahymena cells in

order to visualize newly assembled BBs. In this case, the naked BBs were the assumed to be the

new ones, which had not yet developed cilia. These studies suggested that the proliferation of

new BBs was more prominent in the mid and posterior parts of the cell. Additional studies also

showed that the anterior 1/3rd of the cell had a higher concentration of old (ciliated) cortical units

and insertion of new units occurs primarily in the posterior 2/3rd of the cell. Thus, one can safely

assume that the posterior daughter inherits not only more newly assembled BBs but also their

associated structures. Thus, the daughter cells produced at division are unequal in their

45

inheritance of old and new structures. The posterior daughter inherits more newly assembled

structures whereas the anterior daughter retains a greater proportion of the older pre existing

structures.

Cilia and Flagella

Most cilia and flagella have a characteristic structure, commonly referred to as the 9+2.

The axoneme (the microtubular part) consists of 9 outer doublets surrounding a central pair of

MT singlets, forming the axoneme. The outer doublets comprise of a complete 13 protofilament

MT called the A-tubule to which is attached an incomplete 10 protofilament B tubule. The

axoneme represents a very striking example of how these MT based structures are assembled, in

a conserved manner across species. About 250 different components have to come together in a

precise manner to form the eukaryotic cilium or flagellum. Cilia and flagella are extensions of

the triplet BB structure and the doublets of the cilium are continuous with the A- and B- tubules

of triplet MTs of the BB. The eukaryotic axoneme has a polarity, like most other MT based

organelles. The plus end is distal whereas the minus end is anchored at the BB. Cells also

regulate the lengths of the cilium or flagellum, after assembly (Marshall and Rosenbaum, 2001).

How assembly occurs and how selected proteins are transported to the growing cilium is still

under intense investigation. One phenomenon, which has recently been implicated in the process

of ciliary assembly and maintenance, is the process of Intraflagellar Transport (IFT), which will

be described in the following section.

Cilia and flagella have very diverse roles that depend on the cell type. For example, the

flagellum of the sperm cell is responsible for cell motility and consequently is critical for

fertilization of the egg. There are non-motile cilia found in the rod and cone cells of the

46

vertebrate eye, where they may perform the function of intracellular transport (Marszalek et al.,

2000; Pazour et al., 2002). In ciliated epithelial cells such as in the respiratory tract and oviduct,

these organelles are responsible for moving fluids and suspended particles across the tissue

surface. Importantly, motile cilia on the embryonic node have been implicated in the

determination of left-right axis determination (Afzelius, 1985). Mutagenesis of Chlamydomonas

revealed a number of mutants with anomalies in ultrastructure and beating, Many of the cloned

mutated genes of Chlamydomonas turned out to be relevant to the study of two human

conditions, Primary Ciliary Dyskinesia (PCD) and lateralization defects (LD). In PCD, the cilia

of the respiratory epithelial cells that line the lower and upper respiratory tracts do not beat

normally. This results in reduced mucociliary clearance leading to recurrent infections of the

respiratory tract, sinus and middle ear. About half of PCD patients have situs inversus, which is a

complete left right reversal of internal organs. These patients suffer from the Kartagener

syndrome, which is PCD with situs inversus. LD involves mispositioning of lateral organs of the

thorax (eg: heart) and/or the abdomen (eg: stomach), which might be complete (situs inversus) or

incomplete (situs ambiguous). Extensive research has established that the situs invertus

phenotype is a direct consequence of lack of function of cilia present on the surface of the node

during embryonic development (for review see Brueckner, 2001).

The early embryo begins as a bilaterally symmetrical organism until ciliary beating

apparently promotes leftward movement of the fluid surrounding the part of the embryo

composed of nodal cells. This ‘nodal flow’ is driven by specialized 9+0 cilia that project from

the surface of the node. A knockout of essential murine kinesin-II genes leads to a complete

block in the assembly of monocilia on embryonic nodal cells, leading to situs inversus (Nonaka

et al., 1998). It has been experimentally shown that nodal flow is sufficient to initiate left-right

47

orientation in the mouse embryo. Shigenori Nonaka and collegues demonstrated that fluid flow

can control left right development, by placing embryos in a chamber and forcing fluid flow

across the node (Nonaka et al., 1998). The question remains as to what are the downstream

consequences of the nodal flow during the generation of left-right asymmetry.

One hypothesis is that this fluid movement creates a gradient of some extracellular

morphogen, which then induces a gradient of gene expression. This could promote the folding of

the organ primordial in the proper direction, orienting the heart on the left and so on. However,

there are several lines of evidence against this model. Asymmetric protein distribution has not

been detected and the strong flow used experimentally in vitro would be expected to wash away

or dilute signaling factors. A new hypothesis has emerged as an alternative mechanosensory

model (McGrath et al., 2003). This model is based on the presence of two distinct populations of

cilia, motile and non-motile, on the nodal cells. The L/R determination process requires the

axonemal dynein, left-right dynein (lr-dynein) and polycystin-2. In the absence of lr-dynein,

nodal cilia do not move and the left right determination becomes random. Also, it has been

shown that the cation channel polycystin-2-the product of the polycystic kidney disease type 2

gene Pkd-2- plays a role in left right determination, as mice with mutations in this protein have

right isomerism (Pennekamp et al., 2002).

Recently, McGrath et al showed that there are two distinct populations of nodal cilia, one

subset being centrally located and motile and the other located around the periphery of the node

and nonmotile (McGrath et al., 2003). The motile cilia contain lr-dyenin whereas the non-motile

cilia lack lr-dyenin. Nodal monocilia in mice lacking the lr-dyenin protein as well as mice with a

targeted disruption of lr-dyenin, are completely immotile. Polycystin-2 is present in all

populations of nodal monocilia. They proposed that left right asymmetry is laid down in mice,

48

initially by a ciliary mechanism. The motile lr-dyenin containing cilia generate a leftward nodal

flow whereas the immotile cilia play a mechanosensory role. The leftward flow generated by the

centrally located lr-dyenin containing cilia induces a bending of the peripherally located sensory

cilia, generating a release of calcium at the left border of the node. Calcium signals at the left of

the node could trigger enhanced expression of a left side specific factor, thus providing a link

between leftward flow and asymmetric organogenesis. Thus, cilia play an important role in left

right determination in early embryogenesis.

Intraflagellar Transport

During ciliary/ flagellar assembly, new subunits are added onto the distal tip of the

structure, and the process of turnover requires the removal of old subunits and addition of newer

ones in the axonemal structure. The machinery responsible for the synthesis of new ciliary

proteins is present in the cell body, far removed from the cilium. Thus, the cell faces a problem

of efficient and precise cargo delivery to the site of assembly and subunit turnover. The

discovery of the process of IFT has helped us in our understanding of how the cell deals with this

problem. IFT is the bi-directional transport occurring in both motile and non-motile cilia, as well

as in flagella (Rosenbaum et al., 1999; Sloboda, 2002). The main function of IFT is the delivery

of axonemal building blocks from the BB region to the tip of the flagellum where the axoneme

assembles. During this process, protein complexes (IFT particles) move just below the plasma

membrane along the B sub-fiber of the outer doublet MTs. In Chlamydomonas, IFT particles are

made of 2 complexes, Complex A containing 4 proteins and complex B which contains 11

proteins (Cole et al., 1998; Piperno and Mead, 1997). IFT particles move in both the anterograde

49

and retrograde directions at different rates. Presumably, the retrograde movement is required for

the recycling of subunits and motor proteins involved in the process.

The anterograde IFT direction is mediated by kinesin-II. Chlamydomonas and

Tetrahymena cells with mutations in the kinesin-II gene are unable to assemble cilia or flagella

(Brown et al., 1999b; Walther et al., 1994). In Chlamydomonas, studies on the flagellar assembly

mutants first identified the FLA10 gene, which encodes a subunit of the kinesin-II motor,

showed that cells with temperature sensitive alleles are unable to form flagella at the restrictive

temperature. Shifting cells with fully formed flagella to the restrictive temperature caused a

shortening of flagella. Concomitant with this shortening, a cessation of IFT and a disappearance

of IFT particles from the flagella were observed (Kozminski et al., 1995). FLA10 is a member of

the family of heterotrimeric plus end directed motors. Studies in sea urchin embryos showed that

injection of antibodies against a motor subunit of Kinesin-II, caused formation of short flagella

lacking the central pair (Morris and Scholey, 1997). Furthermore, nodal cilia in mouse embryos

failed to assemble when either motor subunit genes were knocked out (Okada et al., 1999;

Takeda et al., 1999). Recently published data from our own lab showed that disruption of the

Tetrahymena homologue of IFT52 (a component of complex B) led to loss of cilia and

cytokinesis defects (Brown et al., 2003).

Cytoplasmic dynein has been shown to play a role in the recycling of IFT particles and

motors back to the cell body, via retrograde IFT. Deleting the gene that encodes for the dynein

heavy chain isoform DHC1b in Chlamydomonas resulted in a mutant that had very short flagella

(Porter et al., 1999). These flagella had accumulated IFT particles at the tip of the structure

indicating that anterograde transport was unaffected while the return of IFT particles was

blocked (Pazour et al., 1999; Pazour et al., 1998). When a temperature sensitive mutant of

50

DHC1b was allowed to form flagella at the permissive temperature and then shifted to the

restrictive temperature, flagella shortened. Therefore, retrograde transport is needed for the

maintenance of fully formed flagella as well. It is quite possible that the role of retrograde

transport could be simply to recycle IFT components.

There are instances however, where IFT does not appear to be involved in axonemal

assembly and maintenance. Kinesin-II is not needed to assemble sperm flagella but is required

for ciliary assembly in sensory neurons of Drosophila (Sarpal et al., 2003). This could be a

special case though, because Drosophila spermatogenesis occurs via a unique pathway.

Spermatids develop from a syncitial cyst, not fully enclosed by a plasma membrane until the

axonemes have fully elongated. This means that the growing tip of the axoneme is fully

accessible to the cell body cytoplasm. In addition, the Drosophila sperm tails are extremely long,

which might indicate a possible alternate pathway of assembly of these long axoneme tails, not

mediated by IFT.

Subunit exchange in preformed organelles

Already assembled microtubular organelles, such as cilia and flagella are dynamic, as

they exchange their structural components including tubulins. Not much was known about the

mechanism of tubulin exchange in preformed organelles. It is possible that IFT play some role in

exchange in the case of cilia and flagella. Traditionally, cilia and flagella were considered very

static structures with little or no turnover occurring at all. Some recent studies show this not to be

the case. Song and Dentler used the technique of labeling Chlamydomonas flagellar polypeptides

with 35S and found that under conditions that promote flagellar stability, at least 80 polypeptides

in the “stable” flagella become labeled in a 4hr period (Song and Dentler, 2001). Using a novel

51

epitope tagging method, Marshall and Rosenbaum showed that tubulin subunit exchange does

occur in axonemes of Chlamydomonas flagella (Marshall and Rosenbaum, 2001). However, this

turnover was limited to the tip of the flagellum and loss was not detected at the base of the

flagellum, indicating that simple treadmilling (preferential addition of new subunits at the tip

combined with loss at the transition zone/BB region) might not occur. However, Trypanosoma

flagella show exchange of paraflagellar rod proteins (a structure unique to this group of proteins

made of non-tubulin proteins and located along the entire axoneme), occurring not only at the tip

of the structure, but along the entire length of the axoneme (Bastin et al., 1999). One could infer

from this study that, tubulin in the adjacent flagellum of Trypanosoma might also be subjected to

a similar pattern of exchange. Song and Dentler also showed that inhibiting tubulin

polymerization by addition of colchicine did not inhibit the turnover of other flagellar

polypeptides (Song and Dentler, 2001). Thus, it appears that a more complex mechanism may be

in play during the assembly/maintenance process in cilia.

It is also not known exactly whether tubulin is transported in the form of subunits or

stable polymers. The same issue is important in the maintenance of axonal extension of neurons,

which requires constant transport of tubulin. It had been speculated earlier that tubulin is

transported out along axons in the form of assembled polymers (Ahmad et al., 1998; Bass and

Brown, 1997). In contrast, observations on the bulk movement of fluorescent tubulin in squid

axons have led others to conclude that axonal MTs are stationary and that tubulin moves in the

form of free subunits or small oligomeric complexes along axonemal MTs (Chang et al., 1999;

Funakoshi et al., 1996; Miller and Joshi, 1996). However, these studies are not conclusive as to

the precise nature of the tubulin cargo transported and utilized for axonal assembly. Terada et al.

investigated the process of slow axonal transport of tubulin in squid giant axons using

52

fluorescent-labeled tubulin (Terada et al., 2000). They showed that the directional movement of

the fluorescent profile was dependent on kinesin motor function as injection of antibodies

directed against kinesin blocked tubulin transport significantly. Recently, our research in our lab

has led to discovery of tubulin cofactor B in the soluble/membrane compartment of Tetrahymena

cilia (G. Pandiyan. D. Dave and J. Gaertig, personal communication). This observation raises a

possibility that folding of tubulin occurs inside cilia and that tubulin may be transported in an

unfolded and non-dimerized state. Another group discovered the presence of the CCT

chaperonin complexes in cilia of Terahymena, which led to a suggestion that CCT components

act as MAPs (Seixas et al., 2003). However, together with the observations from our lab it is

should be considered that the entire folding pathway occurs inside cilia, and could be coupled to

IFT. One way or another, the delivered tubulins are used to replace the existing tubulins without

disassembling the structure. This turnover process occurs not only in cilia but also in all types of

other stable MT organelles including BBs. One of the main goals of my own research was to

determine whether tubulin glycylation acts during or after MT assembly and to what extent

modification participates in the turnover (see Chapter 4).

References:

Adoutte, A., P. Delgado, A. Fleury, N. Levilliers, M.-C. Lainé, M.-C. Marty, E. Boisvieux-

Ulrich, and D. Sandoz. 1991. Microtubule diversity in ciliated cells: Evidence for its

generation by post-translational modification in the axonemes of Paramecium and quail

oviduct cells. Biol.Cell. 71:227-245.

Afzelius, B.A. 1985. The immotile-cilia syndrome: a microtubule-associated defect. CRC

Crit.Rev.Biochem. 19:63-87.

53

Ahmad, F.J., C.J. Echeverri, R.B. Vallee, and P.W. Bass. 1998. Cytoplasmic dyenin and

dynactin are required for the transport of microtubules into the axon. J Cell Biol.

140:391-401.

Aletta, J.M. 1996. Phosphorylation of type III b-tubulin in PC12 cell neutrites during NGF-

induced process outgrowth. J. Neurobiol. 31:461-475.

Alexander, J.E., D.F. Hunt, M.K. Lee, J. Shabanowitz, H. Michel, S.C. Berlin, T.L. Macdonald,

R.J. Sundberg, L.I. Rebhun, and A. Frankfurter. 1991. Characterization of post-

translational modifications in neuron-specific class III b-tubulin by mass spectrometry.

Proc.Natl.Acad.Sci.USA. 88:4685-4689.

Alfa, C.E., and J.S. Hyams. 1991. Microtubules in the fission yeast Schizosaccharomyces pombe

contain only the tyrosinated form of alpha-tubulin. Cell Motility and the Cytoskeleton.

18:86-93.

Allen, R.D. 1969. The morphogenesis of basal bodies and accessory structures of the cortex of

the ciliate protozoan Tetrahymena pyriformis. J.Cell Biol. 40:716-733.

Andreu, J.M., and J.M. De Pereda. 1993. Site-directed antibodies to tubulin. Cell

Motil.Cytoskeleton. 26:1-6.

Arce, C.A., and H.S. Barra. 1983. Association of tubulinyl-tyrosine carboxypeptidase with

microtubules. FEBS Lett. 157:75-78.

Audebert, S., E. Desbruyéres, C. Gruszczynski, A. Koulakoff, F. Gros, P. Denoulet, and B. Eddé.

1993. Reversible polyglutamylation of α- and β-tubulin and microtubule dynamics in

mouse brain neurons. Mol.Biol.Cell. 4:615-626.

Banerjee, A. 2002. Coordination of posttranslational modifications of bovine brain alpha-tubulin.

Polyglycylation of delta2 tubulin. J Biol Chem. 277:46140-4.

54

Banerjee, A., and R.F. Luduena. 1992. Kinetics of colchicine binding to purified β-tubulin

isotypes from bovine brain. J.Biol.Chem. 267:13335-13339.

Bass, P.W., and A. Brown. 1997. Slow axonal transport: the polymer transport model. Trends

Cell Biol. 7:380-384.

Bastin, P., T.H. McRae, S.B. Francis, K.R. Matthews, and K. Gull. 1999. Flagellar

morphogenesis: Protein targeting and assembly in the paraflagellar rod of Trypanosomes.

Mol. Cell. Biol. 19:8191-8200.

Bedini, C., A. Lanfranchi, R. Nobili, and A. Miyake. 1978. Ultrastructure of meiosis inducing

(heterotypic) and non-inducing (homotypic) cell unions in conjugation of Blepharisma. J

Cell Sci. 32:31-43.

Beisson, J., and M. Wright. 2003. Basal body/centriole assembly and continuity. Curr Opin Cell

Biol. 15:96-104.

Ben-Ze'ev, A., S.R. Farmer, and S. Penman. 1979. Mechanisms of regulating tubulin synthesis in

cultured mammalian cells. Cell. 17:319-325.

Bernstein, M., P.L. Beech, S.G. Katz, and J.L. Rosenbaum. 1994. A new kinesin-like protein

(Klp1) localized to a single microtubule of the Chlamydomonas flagellum. J.Cell Biol.

125:1313-1326.

Bobinnec, Y., A. Khodjakov, L.M. Mir, C.L. Rieder, E. B., and M. Bornens. 1998a. Centriole

disassembly in vivo and its effect on centrosome structure and function in vertebrate cells.

J. Cell Biol. 143:1575-1589.

Bobinnec, Y., M. Moudjou, J.P. Fouquet, E. Desbruyeres, B. Eddé, and M. Bornens. 1998b.

Glutamylation of centriole and cytoplasmic tubulin in proliferating non-neuronal cells.

Cell Motil. Cytoskeleton. 39:223-232.

55

Bonnet, C., D. Boucher, S. Lazereg, B. Pedrotti, K. Islam, P. Denoulet, and J.C. Larcher. 2000.

Differential binding regulation of microtubule-associated proteins MAP1A, MAP1B, and

MAP2 by tubulin polyglutamylation. J. Biol. Chem. 276:12839-12848.

Boucher, D., J.C. Larcher, F. Gros, and P. Denoulet. 1994. Polyglutamylation of tubulin as a

progressive regulator of in vitro interactions between the microtubule-associated protein

Tau and tubulin. Biochemistry. 33:12471-12477.

Bré, M.-H., V. Redeker, M. Quibell, J. Darmanaden-Delome, C. Bressac, J. Cosson, P. Huitore,

J.-M. Schmitte, J. Rossier, T. Johnson, A. Adoutte, and N. Levilliers. 1996. Axonemal

tubulin polyglycylation probed with two monoclonal antibodies: widespread evolutionary

distribution, appearance during spermatozoan maturation and possible function in

motility. J. Cell Sci. 109:727-738.

Bré, M.-H., V. Redeker, J. Vinh, J. Rossier, and N. Levilliers. 1998. Tubulin polyglycylation:

differential posttranslational modification of dynamic cytoplasmic and stable axonemal

microtubules in Paramecium. Mol. Biol. Cell. 9:2655-2665.

Bre, M.H., B. De Nechaud, A. Wolff, and A. Fleury. 1994a. Glutamylated tubulin probed in

ciliates with the monoclonal antibody GT335. Cell Motil Cytoskeleton. 27:337-49.

Bre, M.H., B. De Nechaud, A. Wolff, and A. Fleury. 1994b. Glutamylated tubulin probed in

ciliates with the monoclonal antibody GT335. Cell Motil Cytoskeleton. 27:337-349.

Brinkley, B.R. 1985. Microtubule Organizing Centers. Annual Review of Cell Biology. 1:145-

172.

Brown, J.M., N.A. Fine, G. Pandiyan, R. Thazhath, and J. Gaertig. 2003. Hypoxia regulates

assembly of cilia in suppressors of Tetrahymena lacking an intraflagellar transport

subunit gene. Mol. Biol. Cell. 14:3192-3207.

56

Brown, J.M., C. Hardin, and J. Gaertig. 1999a. Rotokinesis, a novel phenomenon of cell

locomotion-assisted cytokinesis in the ciliate Tetrahymena thermophila. Int. Cell Biol.

Rep. 23:841-848.

Brown, J.M., C. Marsala, R. Kosoy, and J. Gaertig. 1999b. Kinesin-II is preferentially targeted to

assembling cilia and is required for ciliogenesis and normal cytokinesis in Tetrahymena.

Mol. Biol. Cell. 10:3081-3096.

Brueckner, M. 2001. Cilia propel the embryo in the right direction. Am J Med Genet. 101:339-

344.

Bulinski, J.C., J.E. Richards, and G. Piperno. 1988. Posttranlational modifications of alpha-

tubulin: detyrosination and acetylation differentiate populations of interphase

microtubules in cultured cells. J Cell Biol. 106:1213-1220.

Callaini, G., W. Whitfield, and M. Riparbelli. 1997. Centriole and centrosome dynamics during

the embryonic cell cycles that follow the formation of the cellular blastoderm in

Drosophila. Exp Cell Res. 234:183-190.

Callen, A.-M., A. Adoutte, J.M. Andrew, A. Baroin-Tourancheau, M.-H. Bré, P.C. Ruiz, J.-C.

Clérot, P. Delgado, A. Fleury, R. Jeanmaire-Wolf, V. Viklicky, E. Villalobo, and N.

Levilliers. 1994. Isolation and characterization of libraries of monoclonal antibodies

directed against various forms of tubulin in Paramecium. Biol.Cell. 81:95-119.

Caron, J.M. 1997. Posttranslational modification of tubulin by Palmitoylation: I. in vivo and cell-

free studies. Mol. Biol. Cell. 8:621-636.

Caron, J.M., A.L. Jones, and M.W. Kirschner. 1985a. Autoregulation of tubulin synthesis in

hepatocytes and fibroblasts. J.Cell Biol. 101:1763-1772.

57

Caron, J.M., A.L. Jones, L.B. Rall, and M.W. Kirschner. 1985b. Autoregulation of tubulin

synthesis in enucleated cells. Nature. 317:648-651.

Caron, J.M., L.R. Vega, J. Fleming, R. Bishop, and F. Solomon. 2001. Single site α-tubulin

mutation affects astral microtubules and nuclear positioning during anaphase in

Sacharomyces cerevisiae: possible role for palmitoylation of α-tubulin. Mol. Biol. Cell.

12:2672-2687.

Chang, P., and T. Stearns. 2000. Delta-tubulin and epsilon-tubulin: two new human centrosomal

tubulins reveal new aspects of centrosome structure and function. Nat. Cell Biol. 2:30-35.

Chang, S., T.M. Svitkina, G.G. Borisy, and S.V. Popov. 1999. Speckle microscopic evaluation of

microtubule transport in growing nerve processes. Nat. Cell Biol. 1:399-403.

Chang, W., D.R. Webster, A.A. Salam, D. Gruber, A. Prasad, J. Eiserich, and J.C. Bulinski.

2002. Alteration of the C-terminal amino acid of tubulin specifically inhibits myogenic

differentiation. J Biol Chem. 277:30690-30698.

Cleveland, D.W. 1988. Autoregulated instability of tubulin mRNAs: A novel eukaryotic

regulatory mechanism. TIBS. 13:339-343.

Cleveland, D.W. 1989. Autoregulated control of tubulin synthesis in animal cells.

Curr.Opin.Cell Biol. 1:10-14.

Cleveland, D.W., M.W. Kirschner, and N.J. Cowan. 1978. Isolation of separate mRNA for alpha

and beta tubulin and characterization of the corresponding in vitro translation products.

Cell. 15:1021-1031.

Cleveland, D.W., M.A. Lopata, P. Sherline, and M.W. Kirschner. 1981. Unpolymerized tubulin

modulates the level of tubulin mRNAs. Cell. 25:537-546.

58

Cleveland, D.W., M.F. Pittenger, and J.R. Feramisco. 1983a. Elevation of tubulin levels by

microinjection suppresses new tubulin synthesis. Nature. 305:738-740.

Cleveland, D.W., M.F. Pittenger, and M.A. Lopata. 1983b. Autoregulatory control of expression

of * and b tubulin. J.Submicrosc, Cytol. 15:353-358.

Cole, D.G., D.R. Diener, A.L. Himelblau, P.L. Beech, J.C. Fuster, and J.L. Rosenbaum. 1998.

Chlamydomonas kinesin-II-dependent intraflagellar transport (IFT): IFT particles contain

proteins required for ciliary assembly in Caenorhabditis elegans sensory neurons. J. Cell

Biol. 18:993-1008.

Conner, T.W., M.D. Thompson, and C.D. Silflow. 1989. Structure of the three β-tubulin-

encoding genes of the unicellular alga, Polytomella agilis. Gene. 84:345-358.

Davis, A., C.R. Sage, C.A. Dougherty, and K.W. Farrell. 1994. Microtubule dynamics

modulated by guanosine triphosphate hydrolysis activity of β-tubulin. Science. 264:839-

842.

Diaz-Nido, J., L. Serrano, C. Lopez-Otin, J. Vandekerckhove, and J. Avila. 1990.

Phosphorylation of a neuronal-specific β-tubulin isotype. J.Biol.Chem. 265:13949-13954.

Dippell, R.V. 1968. The development of basal bodies in Paramecium. Proc.Natl.Acad.Sci.U.S.A.

61:461-468.

Doerder, F.P., J.C. Deak, and J.H. Lief. 1992. Rate of phenotypic assortment in Tetrahymena

thermophila. Dev.Genet. 13:126-132.

Duan, J., and M.A. Gorovsky. 2002. Both carboxy terminal tails of alpha and beta tubulin are

essential, but either one will suffice. Curr Biol. 12:313-316.

59

Dupuis-Williams, P., A. Fleury, N. Garreau De Loubresse, H. Geoffery, L. Vayssie, A. Galvani,

A. Espigat, and J. Rossier. 2002. Functional role of epsilon tubulin in the assembly of the

centriolar microtubule scaffold. J Cell Biol. 158:1183-1193.

Dutcher, S.K., N.S. Morrissette, A.M. Preble, C. Rackley, and J. Stanga. 2002. Epsilon tubulin is

an essential component of the centriole. Mol Biol Cell. 13:3859-3869.

Dutcher, S.K., and E.C. Trabuco. 1998. The UNI3 gene is required for assembly of basal bodies

of Chlamydomonas and encodes δ-tubulin, a new member of the tubulin superfamily.

Mol. Biol. Cell. 9:1293-1308.

Edde, B., J. Rossier, J.P. Le Caer, E. Desbruyeres, F. Gros, and P. Denoulet. 1990.

Posttranslational glutamylation of alpha tubulin. Science. 247:83-85.

Eipper, B.A. 1972. Rat brain microtubule protein: purification and determination of covanlently

bound phosphate and carbohydrate. Proc. Natl Acad. Sci. USA. 69:2283-2287.

Ersfeld, K., J. Wehland, U. Plessmann, H. Dodemont, V. Gerke, and K. Weber. 1993.

Characterization of the tubulin-tyrosine ligase. J.Cell Biol. 120:725-732.

Fleury, A., A.-M. Callen, M.-H. Bré, F. Iftode, R. Jeanmaire-Wolf, N. Levilliers, and J.-C.

Clérot. 1995. Where and when is microtubule diversity generated in Paramecium? II.

Deployment of microtubular networks in the interphase cell and during division. J.Cell

Sci. in press.

Fontalba, A., J. Avila, and J.C. AZabala. 1995. Beta-tubulin folding is modulated by the isotype-

specific carboxy-terminal domain. J Mol Biol. 246:628-636.

Frankel, J. 1999. Cell Biology of Tetrahymena. Methods Cell Biology. 62:27-125.

60

Fukushige, T., Z.K. Siddiqui, M. Chou, J.G. Culotti, C.B. Gogonea, S.S. Siddiqui, and M.

Hamelin. 1999. MEC-12, an alpha-tubulin required for touch sensitivity in C.elegans. J.

Cell Sci. 112:395-403.

Funakoshi, T., S. Takeda, and N. Hirokawa. 1996. Active transport of photoactivated tubulin

molecules in growing axons revealed by a new electron microscopic analysis. J Cell Biol.

133:1347-1353.

Gaertig, J. 2000. Molecular Mechanisms of microtubular organelle assembly in Tetrahymena. J.

Euk. Microbiol. 47:185-190.

Gaertig, J., M.A. Cruz, J. Bowen, L. Gu, D.G. Pennock, and M.A. Gorovsky. 1995. Acetylation

of lysine 40 in alpha-tubulin is not essential in Tetrahymena thermophila. J.Cell Biol.

129:1301-1310.

Gaertig, J., T.H. Thatcher, K.E. McGrath, R.C. Callahan, and M.A. Gorovsky. 1993.

Perspectives on tubulin isotype function and evolution based on the observations that

Tetrahymena thermophila microtubules contain a single α- and β-tubulin. Cell Motility

and the Cytoskeleton. 25:243-253.

Gagnon, C., D. White, J. Cosson, P. Huitorel, B. Edde, E. Desbruyeres, L. Paturle-Lafanechere,

L. Multigner, D. Job, and C. Cibert. 1996. The polyglutamylated lateral chain of alpha-

tubulin plays a key role in flagellar motility. J. Cell Sci. 109:1545-1553.

Gao, Y., I.E. Vainberg, R.L. Chow, and N.J. Cowan. 1993. Two cofactors and cytoplasmic

chaperonin are required for the folding of α - and β-tubulin. Mol.Cell.Biol. 13:2478-

2485.

61

Garcia-Cardena, G., P. Oh, J. Liu, J.E. Schnitzer, and W.C. Sessa. 1996. targeting of nitric oxide

synthase to endothelial cell caveolae via plamitoylation: implications for nitric oxide

signalling. Proc. Natl. Acad. Sci. U.S.A. 93:6448-6453.

Gard, D.L., and M.W. Kirshner. 1985. Polymer-dependent increase in phosphorylation of b-

tubulin accompanies differentiation of a mouse neuroblastoma cell line. J.Cell Biol.

100:764-774.

Gould, R.R., and G.G. Borisy. 1977. The pericentriolar material in chinese hamster ovary cells

nucleates microtubule formation. J Cell Biol. 73:601-615.

Greer, K., H. Maruta, S.W. L'Hernault, and J.L. Rosenbaum. 1985. Alpha-tubulin acetylase

activity in isolated Chlamydomonas flagella. J Cell Biol. 10:2081-2084.

Gu, L., J. Gaertig, L.A. Stargell, and M.A. Gorovsky. 1995. Gene specific signal transduction

between microtubules and tubulin genes in Tetrahymena thermophila. Mol.Cell Biol.

15:5173-5179.

Gundersen, G.G., S. Khawaja, and J.C. Bulinski. 1987. Postpolymerization detyrosination of α-

tubulin: A mechanism for subcellular differentiation of microtubules. . J.Cell Biol.

105:251-264.

Gurland, G., and G.G. Gundersen. 1995. Stable, detyrosinated microtubules function to localize

vimentin intermediate filaments in fibroblasts. J.Cell Biol. 131:1275-1290.

Gustke, N., B. Trinczek, J. Biernat, E.-M. Mandelkow, and E. Mandelkow. 1994. Domains of tau

protein and interactions with microtubules. Biochemistry. 33:9511-9522.

Haggarty, S.J., K.M. Koeller, J. Wong, C.M. Grozinger, and S.L. Schreiber. 2003. Domain

selective small-molecule inhibitor of histone deacetylase 6 (HDAC6)-mediated tubulin

deacetylation. Proc. Natl. Acad. Sci. U.S.A. 100:4389-4394.

62

Hai, B., and M.A. Gorovsky. 1997. Germ-line knockout heterokaryons of an essential alpha-

tubulin gene enable high-frequency gene replacement and a test of gene transfer from

somatic to germ-line nuclei in Tetrahymena thermophila. Proc. Natl. Acad. Sci. U.S.A.

94:1310-1315.

Han, J.W., J.H. Park, M. Kim, and J. Lee. 1997. mRNAs for microtubule proteins are

specifically colocalized during the sequential formation of basal body, flagella and

cytoskeletal microtubules in the differentiation of Naegleria grubei. J. Cell Biol.

137:871-879.

Harper, J.F., and W. Mages. 1988. Organization and structure of Volvox beta-tubulin genes.

Mol.Gen.Genet. 213:315-325.

Hasimoto, K. 1963. Formation of ciliature in excystment and induced re-encystment in

Oxytricha fallax stein. J. Protozool. 10:156-166.

Hesse, J., M. Theirauf, and H. Ponstingl. 1987. Tubulin sequence region beta 155-174 is

involved in binding exchangeable guanosine triphosphatte. HNO. 35:451-454.

Hochstrrasser, M. 1996. Ubiquitin-dependent protein degradation. Ann. Rev. Genet. 30:405-439.

Horio, T., S. Uzawa, M.K. Jung, B.R. Oakley, K. Tanaka, and M. Yanagida. 1991. The fission

yeast gamma-tubulin is essential for mitosis and is localized at microtubule organizing

centers. J.Cell Sci. 99:693-700.

Hoyle, H.D., and E.C. Raff. 1990. Two Drosophila beta tubulin isoforms are not functionally

equivalent. J.Cell Biol. 111:109-1026.

Hubbert, C., A. Guardiola , R. Shao, Y. Kawaguchi, A. Ito, A. Nixon, M. Yoshida, X.F. Wang,

and T.P. Yao. 2002. HDAC6 is a microtubule-associated deacetylase. Nature. 417:455-

458.

63

Hutchens, J.A., H.D. Hoyle, R. Turner, and E.C. Raff. 1997. Structurally similar Drosophila α-

tubulins are functionally distinct in vivo. Mol. Biol. Cell. 8:481-500.

Jerka-Dziadosz, M. 1981. Patterning of ciliary structures in janus mutant of Tetrahymena with

mirror image cortical duplications. An ultrastructural study. Acta Protozool. 20:337-356.

Joshi, H.C., M.J. Palacios, L. McNamara, and D.W. Cleveland. 1992. gamma-Tubulin is a

centrosomal protein required for cell cycle-dependent microtubule nucleation. Nature.

356:80-83.

Kann, M.L., Y. Prigent, N. Levilliers, M.H. Bre, and J.P. Fouquet. 1998. Expression of

glycylated tubulin during the differentiation of spermatozoa in mammals. Cell Motil

Cytoskeleton. 41:341-352.

Kann, M.L., S. Soues, N. Levilliers, and J.P. Fouquet. 2003. Glutamylated tubulin: Diversity if

expression and distribution of isoforms. Cell Motil Cytoskeleton. 55:14-25.

Kemphues, K.J., T.C. Kaufman, R.A. Raff, and E.C. Raff. 1982. The testis-specific beta-tubulin

subunit in Drosophila melanogaster has multiple functions in spermatogenesis. Cell.

31:655-670.

Khodjakov, A., C. Rieder, G. Sluder, G. Cassels, O. Sibon, and C. Wang. 2002. De novo

formation of centrosomes in vertebrate cells arrested during S phase. J Cell Biol. 158.

Kirschner, M.W. 1978. Microtubule assembly and nucleation. Int Rev Cyt. 54:1-71.

Kirschner, M.W., and T. Mitchison. 1986. Microtubule dynamics. Nature. 324:621-625.

Kitada, T., S. Asakawa, N. Hattori, H. Matsumine, Y. Yamamura, S. Minoshima, M. Yokochi,

Y. Mizuno, and N. Shimizu. 1998. Mutations in the parkin gene cause autosomal

recessive juvenile parkinsonism. Nature. 392:605-608.

64

Kozminski, K.G., P.L. Beech, and J.L. Rosenbaum. 1995. The Chlamydomonas kinesin-like

protein FLA10 is involved in motility associated with the flagellar membrane. J.Cell

Biol. 131:1517-1527.

Kozminski, K.G., D.R. Diener, and J.L. Rosenbaum. 1993. High level expression of

nonacetylatable α-tubulin in Chlamydomonas reinhardtii. Cell Motility and the

Cytoskeleton. 25:158-170.

Kreitzer, G., G. Liao, and G.G. Gundersen. 1999. Detyrosination of tubulin regulates the

interaction of intermediate filaments with microtubules in vivo via a kinesin-dependent

mechanism. Mol. Biol. Cell. 10:1105-1118.

Kubota, H., G. Hynes, A. Carne, A. Ashworth, and K. Willison. 1994. Identification of six

related Tcp-1 gees encoding divergent subunits of the TCP-1 containing chaperonin. Curr

Biol. 4:89-99.

Kumar, N., and M. Flavin. 1981. Preferential action of a brain detyrosinolating carboxypeptidase

on polymerized tubulin. J.Biol.Chem. 256:7678-7680.

L'Hernault, S.W., and J.L. Rosenbaum. 1983. Chlamydomonas α -tubulin is posttranslationally

modified in the flagella during flagellar assembly. J.Cell Biol. 97:258-263.

L'Hernault, S.W., and J.L. Rosenbaum. 1985. Chlamydomonas α -tubulin is posttranslationally

modified by acetylation on the e-amino group of a lysine. Biochem. 24:473-478.

Lafanechere, L., C. Courtay-Cahen, T. Kawakami, M. Jacrot, M. Rudiger, J. Wehland, D. Job,

and R.L. Margolis. 1998. Suppression of tubulin tyrosine ligase during tumor growth. J

Cell Sci. 111:171-181.

65

Larcher, J.C., D. Boucher, S. Lazereg, F. Gros, and P. Denoulet. 1996. Interaction of kinesin

motor domains with alpha- and beta-tubulin subunits at a tau-independent binding site.

Regulation by polyglutamylation. J Biol Chem. 271:22174-22124.

LeDizet, M., and G. Piperno. 1987. Identification of an acetylation site of Chlamydomonas α -

tubulin. Proc.Natl.Acad.Sci.USA. 84:5720-5724.

Levilliers, N., A. Fleury, and A.M. Hill. 1995. Monoclonal and polyclonal antibodies detect a

new type of post-translational modification of axonemal tubulin. J.Cell Sci. 108:3013-

3028.

Lewis, S.A., G. Tian, and N.J. Cowan. 1997. The alpha and beta-tubulin folding pathways.

Trends Cell Biol. 7:479-484.

Li, Q., and H.C. Joshi. 1995. gamma-tubulin is a minus end-specific microtubule binding

protein. J.Cell Biol. 131:207-214.

Litman, P., J. Barg, and I. Ginzburg. 1994. Microtubules are involved in the localization of tau

mRNA in primary neuronal cell cultures. Neuron. 13:1463-1474.

Litman, P., J. Barg, L. Rindzooski, and I. Ginzburg. 1993. Subcellular localzation of Tau mRNA

in differentiating neuronal cell culture: implications for neuronal polarity. Neuron.

10:627-638.

Luduena, R.F. 1998. Multiple forms of tubulin: different gene products and covalent

modifications. Int. Review Cytol. 178:207-274.

Luduena, R.F., A. Banerjee, and I.A. Khan. 1992. Tubulin structure and biochemistry. Curr Opin

Cell Biol. 4:53-57.

MacRae, T.H., and C.M. Langdon. 1989. Tubulin synthesis, structure, and function: what are the

relationships? Biochem.Cell Biol. 67:770-790.

66

Marshall, L.G., R.L. Jeng, L. Mulholland, and T. Stearns. 1996. Analysis of Tub4p, a yeast γ-

tubulin-like protein: Implications for microtubule-organizing center function. J. Cell Biol.

134:443-454.

Marshall, W.F., and J.L. Rosenbaum. 2001. Intraflagellar transport balances continuous turnover

of outer doublet microtubules: implications for flagellar length control. J Cell Biol.

155:In Press.

Marshall, W.F., Y. Vucica, and J.L. Rosenbaum. 2001. Kinetics and regulation of de novo

centriole assembly. Implications for the mechanism of centriole duplication. Curr Biol.

11:308-317.

Marszalek, J.R., X. Liu, E.A. Roberts, D. Chui, J.D. Marth, D.S. Williams, and L.S. Goldstein.

2000. Genetic evidence for selective transport of opsin and arrestin by kinesin-II in

mammalian photoreceptors. Cell. 102:175-187.

Maruta, H., K. Greer, and J.L. Rosenbaum. 1986. The acetylation of alpha-tubulin and its

relationship to the assembly and disassembly of microtubules. J.Cell Biol. 103:571-579.

McGrath, J., S. Somio, S. Makova, X. Tian, and M. Brueckner. 2003. Two populations of node

monocili initiate left-right asymmetry in the mouse. Cell. 114:61-73.

Melki, R., and N.J. Cowan. 1994. Facilitated folding of actins and tubulins occurs via a

nucleotide-dependent interaction between cytoplasmic chaperonin and distinctive folding

intermediates. Mol.Cell.Biol. 14:2895-2904.

Menendez, M., G. Rivas, J.F. Diaz, and J.M. Andreu. 1998. Control of the structural stability of

the tubulin dimer by one high affility bound magnesium ion at nucleotide N-site. J Biol

Chem. 273.

Miller, K.E., and H.C. Joshi. 1996. Tubulin transport in neurons. J Cell Biol. 133:1355-1366.

67

Mitchison, T., and M.W. Kirschner. 1984. Dynamic instability of microtubule growth. Nature.

312:237-242.

Moffet, S., L. Adams, H. Bonin, T.P. Loisel, M. Bouvier, and B. Mouillac. 1996. Palmitoylated

cysteine 341 modulates phosphorylation of the beta2 adrenergic receptor by the cAMP

dependant protein kinase. J Biol Chem. 271:21490-21497.

Moritz, M., M.B. Braunfeld, J.W. Sedat, B. Alberts, and D.A. Agard. 1995. Microtubule

nucleation by gamma-tubulin-containing rings in the centrosome. Nature. 378:638-640.

Morris, R.L., and J.M. Scholey. 1997. Heterotrimeric kinesin-II is required for the assembly of

motile 9+2 ciliary axonemes on sea urchin embryos. J. Cell Biol. 138:1009-1022.

Muller-Reichert, T., D. Chretien, F. Severin, and A.A. Hyman. 1998. Structural changes at

microtubule ends accompanying GTP hydrolysis in microtubules: information from a

slowly hydrolyzable analogue of GTP, guanylyl (alpha,beta)methylenediphosphonate.

Proc. Natl. Acad. Sci. U.S.A. 95:3661-3666.

Multigner, L., I. Pignot-Paintrand, Y. Saoudi, D. Job, U. Plessmann, M. Rüdiger, and K. Weber.

1996. The A and B tubules of the outer doublets of sea urchin sperm axonemes are

composed of different tubulin variants. Biochemistry. 35:10862-10871.

Mumby, S.M. 1997. Reversible palmitoylation of signaling proteins. Curr Opin Cell Biol. 9:148-

154.

Nielsen, M.G., F.R. Turner, J.A. Hutchens, and E.C. Raff. 2001. Axoneme-specific beta-tubulin

specialization: a conserved C-terminal motif specifies the central pair. Curr Biol. 11:529-

33.

Nogales, E., M. Whittaker, R.A. Milligan, and K.H. Downing. 1999. High-resolution model of

the microtubule. Cell. 96:79-88.

68

Nogales, E., S.G. Wolff, and K.H. Downing. 1998. Structure of the α/β tubulin dimer by electron

crytallography. Nature. 391:199-202.

Nonaka, S., Y. tanaka, Y. Okada, S. Takeda, A. harada, Y. Kanai, M. Kido, and N. Hirokawa.

1998. Randomization of left-right asymmetry due to loss of nodal cilia generating

leftward flow of extraembryonic fluid in mice lacking KIF3B motor protein. Cell.

95:829-837.

O'Connell, K. 2000. The centrosome of the early C.elegans embryo: inheritance, assemly,

replication and developmental roles. Curr Opin Dev Biol. 49:365-384.

Oakley, B.R., C.E. Oakley, Y. Yoon, and M.K. Jung. 1990. γ-Tubulin is a component of the

spindle pole body that is essential for microtubule function in Aspergillus nidulans. Cell.

61:1289-1301.

Okada, Y., S. Nonaka, Y. Tanaka, Y. Saijoh, H. Hamada, and N. Hirokawa. 1999. Abnormal

nodal flow preceds situs inversus in iv and inv mice. Mol Cell. 4:459-468.

Otsubo, K., H. Sakai, H. Murofushi, and R. Kuriyama. 1975. Electrophoretic separation of

tubulin alhpa and beta subunits after S-sulfonation. J Biochem (Tokyo). 77:17-21.

Ozols, J., and J.M. Caron. 1997. Posttranslational modification of tubulin by palmitoylation: II.

Identification of sites of palmitoylation. Mol. Biol. Cell. 8:637-645.

Palazzo, A., B. Ackerman, and G.G. Gundersen. 2003. Cell Biology: Tubulin acetylation and cell

motility. Nature. 417:455-458.

Pazour, G.J., S.A. Baker, J.A. Deane, D.G. Cole, B.L. Dickert, J.L. Rosenbaum, G.B. Witman,

and J.C. Besharse. 2002. The intraflagellar transport protein, IFT88, is essential for

vertebrate photoreceptor assembly and maintenance. J.Cell.Biol. 157:103-113.

69

Pazour, G.J., B.L. Dickert, and G.B. Witman. 1999. The DHC1b (DHC2) isoform of cytoplasmic

dynein is required for flagelar assembly. J.Cell.Biol. 144:473-481.

Pazour, G.J., C.G. Wilkerson, and G.B. Witman. 1998. A dynein light chain is essential for the

retrograde particle movement of intraflagellar transport (IFT). J. Cell Biol. 141:979-992.

Pechart, I., M.L. Kann, N. Levilliers, M.H. Bre, and J.P. Fouquet. 1999. Composition and

organization of tubulin isoforms reveals a variety of axonemal models. Biol. Cell.

91:685-97.

Pennekamp, P., C. Karcher, A. Fischer, A. Schweickert, B. Skyrabin, J. Horst, M. Blum, and B.

Dworniczak. 2002. The ion channel polycystin-2 is required for left right axis

determination in mice. Curr Biol. 12:938-943.

Pickett-heaps, J.D. 1969. The evolution of the mitotic apparatus: An attempt at comparative

ultrastructural cytology in dividing plant cells. Cytobios. 3:257-280.

Piperno, G., and D.J. Luck. 1976. Phosphorylation of axonemal proteins in Chlamydomonas

reinhardtii. J Biol Chem. 251:2161-2167.

Piperno, G., and K. Mead. 1997. Transport of a novel complex in the cytoplasmic matrix of

Chlamydomonas flagella. Proc. Natl. Acad. Sci. USA. 94:4457-4462.

Piperno, G., Z. Ramanis, E.F. Smith, and W.S. Sale. 1990. Three distinct inner dynein arms in

Chlamydomonas flagella: molecular composition and location in the axoneme. J.Cell

Biol. 110:379-389.

Pirner, M.A., and R.W. Linck. 1994. Tektins are heterodimeric polymers in flagellar

microtubules with axial periodicities matching the tubulin lattice. J.Biol.Chem.

269:31800-31806.

70

Porter, M.E., R. Bower, J.A. Knott, P. Byrd, and W. Dentler. 1999. Cytoplasmic dynein heavy

chain 1b is required for flagellar assembly in Chlamydomonas. Mol. Biol. Cell. 10:693-

712.

Pucciarelli, S., P. Ballarini, and C. Miceli. 1997. Cold-adapted microtubules: characterization of

tubulin posttranslational modifications in teh antarctic ciliate Euplotes focardii. Cell

Motil Cytoskeleton. 38:329-340.

Raff, E.C., J.D. Fackenthal, J.A. Hutchens, H.D. Hoyle, and H.F. Turner. 1997. Microtubule

architecture specified by a β-tubulin isoform. Science. 275:70-73.

Raybin, D., and M. Flavin. 1977a. Enzyme which adds tyrosine to the α−chain of tubulin.

Biochemistry. 16:2189-2194.

Raybin, D., and M. Flavin. 1977b. Modification of tubulin by tyrosylation in cells and extracts

and its effect on assembly in vitro. J. Cell Biol. 73:498-504.

Redeker, V., N. Levilliers, J.-M. Schmitter, J.-P. Le Caer, J. Rossier, A. Adoutte, and M.-H. Bré.

1994. Polyglycylation of tubulin: a post-translational modification in axonemal

microtubules. Science. 266:1688-1691.

Regnard, C., E. Desbruyeres, P. Denoulet, and B. Edde. 1999. Tubulin polyglutamylase:

isozymic variants and regulation during the cell cycle in HeLa cells. J. Cell Sci.

112:4281-4289.

Regnard, C., E. Desbruyeres, J.C. Huet, C. Beauvallet, J.C. Pernollet, and B. Edde. 2000.

Polyglutamylation of nucleosome assembly proteins. J. Biol. Chem. 275:15969-15976.

Regnard, C., D. Fesquet, C. Janke, D. Boucher, E. Desbruyeres, A. Koulakoff, C. Insina, P.

Travo, and B. Edde. 2003. Characterization of PGs1, a subunit of a protein complex co-

purifying with tubulin polyglutamylase. J Cell Sci. 116:4181-4190.

71

Ren, Y., J. Zhao, and J. Feng. 2003. Parkin binds to α/β tubulinand increases their ubiquitination

and degradation. J Neurosci. 23:3316-3324.

Rommelaere, H., M. Van Troys, Y. Gao, R. Melki, and N.J. Cowan. 1993. Eukaryotic

cytoplasmic chapronin contains t-complex polypeptide 1 and seven related subunits.

Proc. Natl. Acad. Sci. U.S.A. 90.

Rosenbaum, J.L. 2000. Cytoskeleton: functions for tubulin modifications at last. Current Biol.

10:R801-R803.

Rosenbaum, J.L., D.G. Cole, and D.R. Diener. 1999. Intraflagellar transport: the eyes have it. J.

Cell Biol. 144:385-388.

Rudiger, M., U. Plessmann, A.H. Rudiger, and K. Weber. 1995. Beta tubulin of bull sperm is

polyglycylated. FEBS Lett. 364:147-151.

Rudiger, M., and K. Weber. 1993. Characterization of the post-translational modifications in

tubulin from the marginal band of avian erythrocytes. Eur J Biochem. 218:107-116.

Ruiz, F., J. Beisson, J. Rossier, and P. Dupuis-Williams. 1999. Basal body duplication in

Paramecium requires gamma-tubulin. Curr. Biol. 14:43-46.

Ruiz, F., A. Krzywicka, C. Klotz, A. Keller, J. Cohen, F. Koll, G. Balavoine, and J. Beisson.

2000. The SM19 gene, required for duplication of basal bodies in Paramecium, encodes a

novel tubulin, eta-tubulin. Curr. Biol. 10:1451-1454.

Sackett, D.L. 1995. Structure and function in the tubulin dimer and the role of the acidic carboxy

terminus. Proteins: Structure, Function and Engineering:255-302.

Sarpal, R., S.V. Todi, E. Sivan-Loukianova, S. Shirolikar, N. Subramanian, E.C. Raff, J.W.

Erickson, K. Ray, and D.F. Ebrel. 2003. Drosophila KAP interacts with the Kinesisn II

72

motor subunit KLP64D to assemble chordotonal sensory cilia, but not sperm tails. Curr

Biol. 13:1687-1696.

Sasse, R., and K. Gull. 1988. Tubulin post-translational modifications and the construction of

microtubular organelles in Trypanosoma brucei. J. Cell Sci. 90:577-589.

Savage, C., M. Hamelin, J.G. Culotti, A. Coulson, D.G. Albertson, and M. Chalfie. 1989. mec-7

is a b-tubulin gene required for the production of 15-protofilament microtubules in

Caenorhabditis elegans. Genes Dev. 3:870-881.

Schneider, A., U. Plessmann, R. Felleisen, and K. Weber. 1998. Posttranslational modification of

trichomonad tubulins: identification of multiple glutamylation sites. FEBB Lett. 429:399-

402.

Schneider, A., U. Plessmann, and K. Weber. 1997. Subpellicular and flagellar microtubules of

Trypanosoma brucei are extensively glutamylated. J. Cell Sci. 110:431-437.

Schwer, H.D., P. Lecine, S. Tiwari, J.E. Italiano, J.H. Hartwig, and R.A. Shivdasani. 2001. A

lineage-restricted and divergent beta-tubulin isoform is essential for the biogenesis,

structure and function of blood platelets. Curr. Biol. 11:579-586.

Seixas, C., C. Casalou, L.V. Melo, S. Nolasco, P. Brogueira, and H. Soares. 2003. Subunits of

the chaperonin CCT are associated with Tetrahymena microtubule structures and are

involved in cilia biogenesis. Exp Cell Res. 290:303-321.

Shang, Y., B. Li, and M.A. Gorovsky. 2002a. Tetrahymena thermophila contains a conventional

gamma tubulin that is differentially required for the maintenance of different micotubule

organizing centers. J Cell Biol. 158:1195-1206.

Shang, Y., X. Song, J. Bowen, R. Corstanje, Y. Gao, J. Gaertig, and M.A. Gorovsky. 2002b. A

robust inducible-repressible promoter greatly facilitates gene knockouts, conditional

73

expression, and overexpression of homologous and heterologous genes in Tetrahymena

thermophila. Proc Natl Acad Sci. 99:3734-3739.

Shu, H.B., Z. Li, M.J. Palacios, Q. Li, and H.C. Joshi. 1995. A transient association of gamma-

tubulin at the midbody is required for the completion of cytokinesis during the

mammalian cell division. J.Cell Sci. 108:2955-2962.

Simpson, A.G.B., C. Bernard, T. Fenchel, and D.J. Patterson. 1997. The organization of

Mastigamoeba schizzophrenia n. sp: More evidence of ultrastructural idiosyncrasy and

simplicity in pelobionts protists. Eur J Protozool. 33:87-98.

Sloboda, R.D. 2002. A healthy understanding of intraflagellar transport. Cell Motil

Cytoskeleton. 52:1-8.

Smrzka, O.W., Delgehyr.N, and Bornens.M. 2000. Tissue-specific expression and subcellular

localisation of mammalian delta-tubulin. Curr. Biol. 10:413-416.

Sobel, S.G., and M. Snyder. 1995. A highly divergent g-tubulin gene is essential for cell growth

and proper microtubule organization in Saccharomyces cerevisiae. J. Cell Sci. 131:1775-

1788.

Song, L., and W. Dentler. 2001. Flagellar protein dynamics in Chlamydomonas. J. Biol. Chem.

276:29754-29763.

Spang, A., S. Geissler, K. Grein, and E. Schiebel. 1996. g-tubulin-like tub4p of Saccharomyces

cerevisiae is associated with the spindle pole body substructures that organize

microtubules and is required for mitotic spindle formation. J. Cell Biol. 134:429-441.

Stearns, T., L. Evans, and M. Kirschner. 1991. Gamma-tubulin is a highly conserved component

of the centrosome. Cell. 65:825-836.

74

Suh, M., J. Han, Y. No, and J. Lee. 2002. Transient concentration of a gamma tubulin related

protein with a pericentrin related protein in the formation of basal bodies and flagella

during the differentiation of Naegleria gruberi. Cell Motil Cytoskeleton. 52:66-81.

Sunkel, C.E., R. Gomes, P. Sampaio, J. Perdigao, and C. Gonzalez. 1995. Gamma-tubulin is

required for the structure and function of the microtubule organizing centre in Drosophila

neuroblasts. EMBO J. 14:28-36.

Takeda, S., Y. Yonekawa, Y. Tanaka, Y. Okada, S. Nonaka, and N. Hirokawa. 1999. Left-right

asymmetry and kinesin superfamily protein KIF3A: new insights in determination of

laterality and mesoderm induction by kif3A-/- mice analysis. J. Cell Biol. 145:826-835.

Terada, S., M. Kinjo, and N. Hirokawa. 2000. Oligomeric tubulin in large transporting complex

is transported via kinesin in squid giant axons. Cell. 103:141-155.

Tian, G., Y. Huang, H. Rommelaere, J. Vandekerckhove, and C. Ampe. 1996. Pathway leading

to correctly folded beta-tubulin. Cell. 86:287-296.

Tian, G., S.A. Lewis, B. Feierbach, T. Stearns, and H. Rommelaere. 1997. Tubulin subunits exist

in an activated conformational state generated and maintained by protein cofactors. J Cell

Biol. 138.

Turkewitz, A.P., E. Orias, and G. Kapler. 2002. Functional genomics: the coming of age for

Tetrahymena thermophila. Trends in Genetics. 18:35-40.

Vaughn, S., T. Attwood, M. Navarro, V. Scott, P.G. McKean, and K. Gull. 2000. New tubulins

in protozoal parasites. Curr Biol. 10:R258-R259.

Vijayan, S., E. El-Akkad, I. Grundke-Iqbal, and K. Iqbal. 2001. A pool of beta-tubulin is

hyperphosphorylated at serine residues in Alzheimer disease brain. FEBS Lett. 509:375-

381.

75

Vinh, J., J.I. Langridge, M.-H. Bre, N. Levilliers, V. Redeker, D. Loyaux, and J. Rossier. 1999.

Structural characterization by tandem spectroscopy of the posttranslational modifications

of tubulin. Biochemistry. 38:3133-3139.

Walker, R.A., E.T. O'Brian, N.K. Pryer, M.F. Soboerio, W.A. Voter, H.P. Erickson, and E.D.

Salmon. 1988. Dynamic instability of individual microtubules analyzed by video light

microscopy: rate constants and transition frequencies. J Cell Biol. 107.

Walther, Z., M. Vashistha, and J.L. Hall. 1994. The Chlamydomonas FLA10 gene encodes a

novel kinesin-homologous protein. J.Cell Biol. 126:175-188.

Weber, K., A. Schneider, S. Westermann, N. Muller, and U. Plessmann. 1997. Posttranslational

modifications of alpha- and beta-tubulin in Giardia lamblia, an ancient eukaryote. FEBS

Lett. 419:87-91.

Webster, D.R., J. Wehland, K. Weber, and G.G. Borisy. 1990. Detyrosination of alpha tubulin

does not stabilize microtubules in vivo. J.Cell Biol. 111:113-122.

Wehland, J., and K. Weber. 1987a. Tubulin -tyrosine ligase has a binding site on beta-tubulin: a

two domain structure of the enzyme. J Cell Biol. 104:1059-1067.

Wehland, J., and K. Weber. 1987b. Turnover of the carboxy-terminal tyrosine of α -tubulin and

means of reaching elevated levels of detyrosination in living cells. J. Cell Biol. 88:185-

203.

Westermann, S., and K. Weber. 2002. Identification of CfNek, a novel member of the NIMA

family of cell cycle regulators, as a polypeptide copurifying with ubulin. J Cell Sci.

115:5003-5012.

76

Xia, L., B. Hai, Y. Gao, D. Burnette, R. Thazhath, J. Duan, M.-H. Bré, N. Levilliers, M.A.

Gorovsky, and J. Gaertig. 2000. Polyglycylation of tubulin is essential and affects cell

motility and division in Tetrahymena thermophila. J.Cell Biol. 149:1097-1106.

Yen, T.J., D.A. Gay, J.S. Pachter, and D.W. Cleveland. 1988. Autoregulated changes in stability

of polyribosome-bound b-tubulin mRNAs are specified by the first 13 translated

nucleotides. Mol.Cell.Biol. 8:1224-1235.

Zhang, W., R.P. Trible, and L.E. Samelson. 1998. LAT palmitoylation: its essential role in

membrane microdomain targeting and tyrosine phosphorylation during T cell activation.

Immunity. 9:239-246.

Zheng, Y., M.L. Wong, B. Alberts, and T. Mitchison. 1995. Nucleation of microtubule assembly

by a gamma-tubulin-containing ring complex. Nature. 378:578-583.

77

Table 2.1: COOH-terminal amino acid sequence of major α-tubulin and major β-tubulin from

Tetrahymena and Paramecium

Organism Tubulin isotype COOH-terminal sequence

Tetrahymena α IETAEGEGEEEGY

Paramecium α1 IETAEGEGEE.GEG

Tetrahymena β TAEEEGEFEEEEGEN

Paramecium β TAEEEGEFEEE .GEQ

Table 2.2: Site directed mutagenesis of putative polymodification sites of Tetrahymena α-tubulin

Type of Substitution Sequence of ATU1 fragment usedfor rescue

Rescue Phenotype

None (WT) IETAEGEGEEEGY449 YES Normal

Single IETAEGEGDEEGY449 YES Normal

Single IETAEGEGEDEGY449 YES Normal

Single IETAEGEGEEDGY449 YES Normal

Triple (α3D) IETAEGEGDDDGY449 YES Normal

All 6 sites (α6D) IDTADGDGDDDGY449 YES Abnormal pairformation, slower

growth

78

Table 2.3: Site directed mutagenesis of putative polymodification sites of Tetrahymena β-tubulin

based on Xia et al. (Xia et al., 2000) and, Duan and Gorovsky (Duan and Gorovsky, 2002)

Sequence of BTU1 fragment used

for rescue

Rescue Phenotype Lethality rescue by

tail replacement on

α-tubulin&

TAEEEGEFEEEEGEN443 YES Normal NA

TAEEEGEFDEEEGEN443 YES Normal NA

TAEEEGEFEDEEGEN443 YES Normal NA

TAEEEGEFEEDEGEN443 YES Normal NA

TAEEEGEFEEEDGEN443 YES Normal NA

TAEEEGEFDDEEGEN443 YES Normal NA

TAEEEGEFEEDDGEN443 YES Normal NA

TAEEEGEFDDDEGEN443 NO Lethal Yes

TAEEEGEFDDEDGEN443 NO Lethal ND

TAEEEGEFDEDDGEN443 NO Lethal ND

TAEEEGEFEDDDGEN443 YES Abnormal* NA

TAEEEGEFDDDDGEN443 NO Lethal YES

TAEEEGEFEDDDGDN443 YES Abnormal* NA

TAEEEGEFDDDDGDN443 NO Lethal YES

TADDDGDFDDDDGDN443 NO Lethal YES

*slow growth, slow motility, occasional cytokinesis defects &COOH-terminal sequence of α/β chimeric gene:

IETAEGEFEEEEGEN (Sequence in italics indicates β-tubulin COOH-terminal sequence)

79

CHAPTER 3

POLYGLYCYLATION DOMAIN OF β-TUBULIN MAINTAINS AXONEMAL

ARCHITECTURE AND AFFECTS CYTOKINESIS IN TETRAHYMENA.1

1Rupal Thazhath, Chengbao Liu and Jacek Gaertig. This work has been published in Nature Cellbiology Vol. 4, 256-259, the only definitive repository of the content that has been certified andaccepted after peer review. Copyright and all rights therein are retained by Macmillan MagazinesLtd. Copyright 2002 Macmillan Magazines Ltd. This chapter reprinted with permission fromthe publisher.

80

Abstract

Polyglycylation occurs by post-translational addition of a polyglycine peptide to the γ-

carboxyl group of glutamic acids near the COOH terminus of α- and β-tubulin (Redeker et al.,

1994) and has been found only in cells with axonemes, from protists to humans (Adoutte et al.,

1985; Bré et al., 1996). In Tetrahymena, multiple sites of polyglycylation on α-tubulin are

dispensable. In contrast, mutating similar sites on β-tubulin gives site-specific effects, and either

affects cell motility and cytokinesis or results in cell death (Xia et al., 2000). Here, we address

the cause of lethality of polyglycylation deficiency in vivo using a powerful approach of

Tetrahymena heterokaryons (Hai et al., 1999). Cells with a lethal mutation of the polyglycylation

domain of β-tubulin assembled axonemes lacking the central pair, B-subfibers and the

transitional zone of outer microtubules (MTs). Furthermore, an arrest in cytokinesis occurred,

and was associated with lack of proper severing of cortical MTs positioned near the cleavage

furrow. Thus, tubulin polyglycylation is required for maintenance of some stable microtubular

organelles that are all known to be polyglycylated in vivo, but its effects on MTs appear to be

organelle type-specific.

Materials and Methods

Construction of mutant heterokaryons

A SmaI site was created by site-directed mutagenesis, downstream of the transcription

termination site of BTU1, on the plasmid pBTU1-DDDE (Xia et al., 2000), and used to insert a

neo2 gene cassette as an EcorV-SmaI fragment (Gaertig et al., 1994). The resulting btu1-ddde440-

neo fragment was targeted to the MIC by biolistic bombardment (Cassidy-Hanley et al., 1997). A

transformant heterozygous for the btu1-ddde440-neo/BTU1 was crossed to a strain homozygous

81

for btu2:bsr/btu2::bsr (Xia et al., 2000). Strains were made homozygous for both mutant alleles

and drug resistant alleles were excluded from the MAC by phenotypic assortment as described

(Xia et al., 2000).

Phenotypic analyses

T. thermophila cells were grown and mated as described (Brown et al., 1999b; Xia et al.,

2000). To bring the phenotypes conferred by the MIC genomes of heterokaryons to expression,

two strains homozygous for either btu1-ddde440-neo/btu1-ddde440-neo; btu2:bsr/btu2::bsr (UG15,

UG16) or BTU1-neo/BTU1-neo; btu2::bsr/btu2::bsr (UG17, UG18) in the MIC, were mated.

Eight hr after mixing, single pairs were isolated into drops of MEPP (Rasmussen and Orias,

1975) medium. Petri dishes with drops were incubated in a moist chamber at 30oC.

Immunofluorescence and Electron Microscopy

The method A is modified after Brown et al. (Brown et al., 1999b). About 100 cells were

isolated into 30 µl of 10 mM Tris, pH 7.5, on a coverslip. An equal volume of 0.5% Triton X-

100 in the PHEM buffer, pH 6.9 (60 mM PIPES, 25 mM HEPES, 10 mM EGTA, 2 mM MgCl2,

1 µM paclitaxel, 0.5 µg/ml leupeptin, 10 µg/ml E-64, 10 µg/ml chymostatin, 12.5 µg/ml of

antipain) was added, and followed after 1 min by 30 µl of 2% paraformaldehyde in PHEM buffer

(without paclitaxel and protease inhibitors), and samples were air dried at 30οC. In method B,

modified after Williams et al. (Williams et al., 1990), to 30 µl of cells, an equal volume of 50%

ethyl alcohol, 0.2% Triton X-100 was added, and cells allowed to dry. Coverslips were processed

for immunofluorescence (Gaertig et al., 1995). The following mouse monoclonal antibodies were

used: 20H5 against centrin (Errabolu et al., 1994) at 1:25 dilution, 12G10 which recognize α-

82

tubulin (Jerka Dziadosz et al., 1995) at 1:25, AXO49 specific for polyglycylated tubulin

(Levilliers et al., 1995) at 1:100, and 6-11 B1 specific for acetylated α-tubulin (Ledizet and

Piperno, 1991) at 1:10 dilution. The method A was used with 6-11 B-1 and AXO 49 antibodies.

Nuclei were stained with propidium iodide. Secondary antibodies were goat anti-mouse-FITC

(Zymed) conjugates at 1:100 dilution. Cells were viewed using a Bio-Rad MRC 600 confocal

microscope.

For electron microscopy, cells were washed with 10 mM Tris, pH 7.5 and fixed in 4%

glutaraldehyde in 10 mM Tris buffer at 4oC for 1 hr, washed 3 times with 10 mM Tris and

postfixed in 4% osmium tetroxide for 1 hr at 4oC. Cells were embedded in Epon after

dehydration in graded steps from 30-100% ethanol. Sections were stained with uranyl acetate

and lead citrate and were visualized on a JEOL 100CXII transmission electron microscope.

Results and Discussion

Ciliates have two nuclei, the micronucleus (MIC) and the macronucleus (MAC). Genes

present in the MIC are transcriptionally repressed, which allows for introduction of potentially

lethal mutations. Furthermore, the mutant alleles of the MIC genome can be brought to

expression by inducing conjugation during which a new MAC forms from the MIC (Hai and

Gorovsky, 1997). T. thermophila has two redundant genes which encode the major β-tubulin:

BTU1 and BTU2 (Gaertig et al., 1993; Xia et al., 2000). The wildtype sequence of the COOH

terminal tail peptide of β-tubulin of T. thermophila is TAEEEGEFEEEEGEN443 (with the major

polyglycylation sites underlined). We constructed heterokaryons which have a lethal triple,

mutation of the polyglycylation domain (with the tail sequence TAEEEGEFDDDEGEN443)

named βDDDE440 (Xia et al., 2000) in BTU1 and lack BTU2 in the MIC, but have wildtype (WT)

83

alleles in the MAC. As a positive control, we constructed similar heterokaryons except that they

had a WT coding sequence of BTU1. The phenotype conferred by the MIC sequences of

heterokaryons was revealed by inducing conjugation. In WT cells, after mating, progeny cells

with new MACs separate, and enter the first vegetative life cycle. Nuclear divisions are initiated

by mitosis of the MIC, followed by an amitotic division of the MAC, which coincides with

cytokinesis (Fig. 3.1A-D). In contrast to WT, after mating, the βDDDE440 progeny cells divided

completely only 2-3 times and subsequently continued to grow in size, but repeatedly failed to

complete cytokinesis, forming chains of "subcells" with arrested cleavage furrows (Fig. 3.1E-G).

Subcells had sets of nuclei, indicating that the nuclear divisions had occurred normally (Fig.

3.1E-G). Although the mutants continued to assemble cilia (Fig. 3.1G), they rapidly lost ciliary

motility and were nearly paralyzed within 24 hr. In sharp contrast to the βDDDE440 cells,

progeny cells entirely lacking a zygotically expressed β-tubulin, died prior to the first vegetative

division (results not shown). Thus, the βDDDE440 phenotype is not the same as a total loss of

function of β-tubulin.

While control cells had a normal 9 + 2 ultrastructure of axonemes (Fig. 3.2A,H), at 48 hr

after mating, nearly all mutant axonemes lacked the central pair of MTs (Table I, Fig. 3.2 B, C).

Thus, the ciliary paralysis can be explained by the loss of the central pair, which is known to be

required for motility of flagella (Smith and Lefebvre, 1997). Many axonemes showed incomplete

outer MTs, which lacked their B-tubules (Fig. 3.2C, D). Furthermore, at 72 hr 27% of axonemes

(n=22) contained a single MT positioned near the center (Fig. 3.2 E, F). Some axonemes also had

a reduced number of all peripheral tubules (Fig. 3.2G). In axonemes with a single central MT,

the number of peripheral MTs was reduced to 8 (Fig. 3.2E, F). Thus, it appears that following the

loss of central pair, a peripheral singlet translocated into the central position. Some cross-

84

sections show likely intermediate stages of such transpositions (Fig. 3.2D). Longitudinal sections

showed that mutants (63% of sections, n=23) had large discontinuities in the transitional zone

between the axoneme and the basal body (Fig. 3.2H, I). The absence of the central pair and the

lack of attachment of outer MTs to the basal body therefore could facilitate transposition of

peripheral singlets. Only singlets were found to be transposed. Transpositions of peripheral

doublet MTs have been described in cases of immobile cilia syndrome (Sturgess et al., 1980) and

at the distal segments of primary cilia, which naturally lack the central pair (Wheatley, 1982). In

contrast to dramatic defects in the axonemes, the structure of the basal bodies was normal, with 9

triplet MTs and associated cytoskeletal fibers (results not shown).

Tetrahymena cells are polarized, with the oral apparatus located at the anterior and

contractile vacuole at the posterior end, respectively (Fig. 3.3A, B). The cell cortex comprises

rows of ciliated basal bodies. Each basal body is associated with transverse and postciliary bands

of MTs and each row is accompanied by the longitudinal microtubule (LM) bundle (Fig. 3.3B).

Although polyglycylated tubulin isoforms are predominantly present in cilia (Fig. 3.3C), they are

also detectable at lower level on all cortical MTs (Fig. 3.3D). The 20H5 antibody specific to

centrin (Errabolu et al., 1994) labels basal bodies of Tetrahymena (Jerka Dziadosz et al., 1995)

and can be used to reveal the overall cortical pattern. In WT cells, one of the first signs of cell

division is the formation of a new oral apparatus in the equatorial region (Fig. 3.1A). Prior to

constriction, an interruption appears in the rows of basal bodies, just above the new oral

apparatus (Fig. 3.1A). The cleavage furrow is formed within the region of this so called "cortical

subdivision" (Frankel, 2001) (Fig. 3.1B). This region is free of transverse and post-ciliary MTs

(Fig. 3.1C). The LMs are initially present in the subdivision zone, but develop small breaks as

the cleavage furrow progresses (Fig. 3.1C, D).

85

Within 48 hr, the majority of mutant cells consisted of 2-4 subcells (Fig. 3.1E, G; Fig.

3.4A-E). Generally, the cortical rows of mutant cells showed numerous defects including

abnormal torsion, breaks and fragmentation (Fig. 3.1E, F; asterisks). Strikingly, in the region of

constriction of mutants, the cortical subdivision was either incomplete or absent (Fig. 3.1E, F;

Fig. 3.4A, C, E). It was difficult to determine whether LMs were properly severed within the

highly constricted fission regions of mutants (e.g. Fig. 3.1E, 3.4E), even using confocal

sectioning. However, some cleavage furrows had retracted, allowing for examination of the

cortical details in the region of fission. Two types of defects were detected in cells with retracted

furrows: 1) cortical rows were interrupted, but their ends continued to grow and invaded the

fission area (Fig. 3.1F and 3.4A) and 2) individual LMs failed to be severed (Fig. 3.4C). Thus,

the cleavage furrow arrest could result from mechanical resistance exerted by uninterrupted or

overgrown LMs. At 72 hr ciliary rows underwent further extensive breakage, the cleavage

furrows completely retracted and subcells fused into giant polynucleated monsters (Fig. 3.4F)

and died within a few days.

Previously we described a viable mutation of the polyglycylation region, βEDDD440,

which led to slow growth, slow motility and an infrequent block in cytokinesis (Xia et al., 2000).

Using anti-centrin, we now find that the βEDDD440 cells also have highly defective ciliary rows,

with break and fragmentation similar to those seen in βDDDE440 (results not shown). Although

only a few percent of βEDDD440 mutants fail to complete cytokinesis (Xia et al., 2000), those

with an arrested cleavage furrow generally lacked a proper cortical subdivision and often had one

or more of LMs that were either overgrown or not severed in the fission area (Fig. 3.4D). We did

not find any ultrastructural defects in the axonemes in the βEDDD440 (results not shown).

However, this mutant has a reduced rate of ciliary motility, suggesting that structural defects may

86

occur, but only in a small subset of axonemes. Alternatively, a mild reduction in polyglycylation

may not alter the axonemal organization, but may affect the activity of axonemal dynein, as we

previously suggested (Xia et al., 2000). However, at least in the cell cortex, two different

mutations of the polyglycylation domain lead to a similar set of defects and differ mainly in the

level of penetrance.

The axonemal defects of βDDDE440 resemble the consequences of swapping of tubulin

isotypes in Drosophila. When the axoneme-specific β2 was replaced by the β1 isotype, the

resulting sperm axonemes lacked the central pair (Raff et al., 2000). Furthermore, a double

substitution immediately upstream of the polyglycylation region allowed β1 to support the

assembly of 9 +2 axonemes (Raff et al., 2000). The axonemes of Drosophila are subjected to

polyglycylation (Bressac et al., 1995), and therefore the difference between β1 and β2 could

involve the ability of β2 to be polyglycylated.

It appears that the deficiency in tubulin polyglycylation affected the stability of the

axoneme. Unlike the highly labile cytoplasmic network MTs, the axonemal MTs are known to be

stable, as manifested by their high resistance to the action of microtubule-depolymerizing drugs

and cold (Behnke and Forer, 1967; Stargell et al., 1992; Tilney and Gibbins, 1968). Strikingly,

the βDDDE440 mutation affected mainly parts of the axoneme, which are already known to be

less resistant to various treatments. For example, the central pair but not the doublet MTs, can be

depolymerized using a low-salt buffer with EDTA (Gibbons, 1965). Furthermore, the gaps in the

transition zone of βDDDE440 axonemes resemble the interruptions in the same region of

Chlamydomonas flagella subjected to deflagellation using a calcium shock (Quarmby, 2000).

Thus, lack of proper polyglycylation could lead to an overall decrease in the structural integrity

of axonemes. The cause of the cytokinesis defects is less clear. Although Tetrahymena cells

87

utilize ciliary motility at the final stage of cytokinesis to rupture the cytoplasmic bridge (Brown

et al., 1999a), cytokinesis still occurs in 50% of cells completely lacking cilia. Also, unlike the

βDDDE440 mutants, even those cilia-less mutants, which fail to complete cytokinesis still develop

a normal cortical subdivision (Brown et al., 1999b). Clearly, the cytokinesis arrest in βDDDE440

occurs at an earlier stage compared to the cilia-less mutants and is likely unrelated to ciliary

paralysis. It is striking that the arrest in cytokinesis correlates with defects in the organization of

cortical MTs. In particular, LMs often are not properly severed and basal bodies not excluded

from the cleavage area. It appears that in the βDDDE440 the arrest in constriction could simply

be caused by the mechanical resistance imposed by uninterrupted cortical MTs. Thus, tubulin

polyglycylation may be required for the cell-cycle-regulated equatorial breakage of LMs,

inhibition of elongation of ends of LMs after their breakage and exclusion of replication of new

basal body units in the cleavage area.

It was quite unexpected that the same mutation destabilized the ciliary MTs but led to an

apparent increased stability or hypertrophy of some longitudinal cortical MTs. It possible

therefore that polyglycylation does not alter the intrinsic properties of MTs, but instead provides

signals for binding of certain Microtubule Associated Proteins (MAPs). Such polyglycylation-

regulated MAPs could mediate diverse effects on MTs (stabilizing, severing, capping, etc.) and

could be restricted to specific cellular compartments.

88

Acknowledgement

This work was supported by NIH grant GM 54017 and ACS grant RPG-99-245-01-CSM. We are

grateful to the staff of the Center for Advanced Ultrastructural Research of the University of

Georgia for training the first author and extensive technical assistance. We thank Joseph Frankel

and Joel Rosenbaum for numerous comments and stimulating ideas. We also thank Marty

Gorovsky, Maria Jerka-Dziadosz, Norman Williams and Jim Lauderdale for critical reading of

the manuscript, Jeffrey Salisbury for 20H5, Gianni Piperno for the 6-11 B-1 and Marie-Helene

Bré and Nicolette Levilliers for the AXO49 antibodies. The 12G10 monoclonal antibody was

raised by E. Marlo Nelsen and Joseph Frankel, and will be available from the Developmental

Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by the

University of Iowa Dept. of Biological Sciences.

References:

Adoutte, A., M. Claisse, R. Maunoury, and J. Beisson. 1985. Tubulin evolution: Ciliate-specific

epitopes are conserved in the ciliary tubulin in metazoa. J.Mol.Evol. 22:220-229.

Behnke, O., and A. Forer. 1967. Evidence for four classes of microtubules in individual cells.

J.Cell Sci. 2:169-172.

Bré, M.-H., V. Redeker, M. Quibell, J. Darmanaden-Delome, C. Bressac, J. Cosson, P. Huitore,

J.-M. Schmitte, J. Rossier, T. Johnson, A. Adoutte, and N. Levilliers. 1996. Axonemal

tubulin polyglycylation probed with two monoclonal antibodies: widespread evolutionary

distribution, appearance during spermatozoan maturation and possible function in

motility. J. Cell Sci. 109:727-738.

89

Bressac, C., M.H. Bré, J. Darmanaden Delorme, M. Laurent, N. Levilliers, and A. Fleury. 1995.

A massive new posttranslational modification occurs on axonemal tubulin at the final

step of spermatogenesis in Drosophila. Eur.J.Cell Biol. 67:346-355.

Brown, J.M., C. Hardin, and J. Gaertig. 1999a. Rotokinesis, a novel phenomenon of cell

locomotion-assisted cytokinesis in the ciliate Tetrahymena thermophila. Int. Cell Biol.

Rep. 23:841-848.

Brown, J.M., C. Marsala, R. Kosoy, and J. Gaertig. 1999b. Kinesin-II is preferentially targeted to

assembling cilia and is required for ciliogenesis and normal cytokinesis in Tetrahymena.

Mol. Biol. Cell. 10:3081-3096.

Calzone, F.J., and M.A. Gorovsky. 1982. Cilia regeneration in Tetrahymena. Exp.Cell Res.

140:474-476.

Cassidy-Hanley, D., J. Bowen, J. Lee, E.S. Cole, L.A. VerPlank, J. Gaertig, M.A. Gorovsky, and

P.J. Bruns. 1997. Germline and Somatic Transformation of Mating Tetrahymena

thermophila by Particle Bombardment. Genetics. 146:135-147.

Errabolu, R., M.A. Sanders, and J.L. Salisbury. 1994. Cloning of a cDNA encoding human

centrin, an EF-hand protein of centrosomes and mitotic spindle poles. J. Cell Sci. 107:9-

16.

Frankel, J. 2001. Cell polarity in ciliates. In Cell Polarity. D.G. Drubin, editor. Oxford University

Press.

Gaertig, J., M.A. Cruz, J. Bowen, L. Gu, D.G. Pennock, and M.A. Gorovsky. 1995. Acetylation

of lysine 40 in alpha-tubulin is not essential in Tetrahymena thermophila. J.Cell Biol.

129:1301-1310.

90

Gaertig, J., L. Gu, B. Hai, and M.A. Gorovsky. 1994. High frequency vector-mediated

transformation and gene replacement in Tetrahymena. Nucleic Acids Res. 22:5391-5398.

Gaertig, J., T.H. Thatcher, K.E. McGrath, R.C. Callahan, and M.A. Gorovsky. 1993.

Perspectives on tubulin isotype function and evolution based on the observations that

Tetrahymena thermophila microtubules contain a single α- and β-tubulin. Cell Motility

and the Cytoskeleton. 25:243-253.

Gibbons, I.R. 1965. Chemical dissection of cilia. Arch.Biol. 76:317-352.

Hai, B., J. Gaertig, and M.A. Gorovsky. 1999. Knockout heterokaryons enable facile mutagenic

analysis of essential genes in Tetrahymena. Methods Cell Biol. 62:513-531.

Hai, B., and M.A. Gorovsky. 1997. Germ-line knockout heterokaryons of an essential alpha-

tubulin gene enable high-frequency gene replacement and a test of gene transfer from

somatic to germ-line nuclei in Tetrahymena thermophila. Proc. Natl. Acad. Sci. U.S.A.

94:1310-1315.

Jerka Dziadosz, M., L.M. Jenkins, E.M. Nelsen, N.E. Williams, R. Jaeckel Williams, and J.

Frankel. 1995. Cellular polarity in ciliates: persistence of global polarity in a disorganized

mutant of Tetrahymena thermophila that disrupts cytoskeletal organization. Dev.Biol.

169:644-661.

Ledizet, M., and G. Piperno. 1991. Detection of acetylated α-tubulin by specific antibodies.

Methods of Enzymology. 196:264-274.

Levilliers, N., A. Fleury, and A.M. Hill. 1995. Monoclonal and polyclonal antibodies detect a

new type of post-translational modification of axonemal tubulin. J.Cell Sci. 108:3013-

3028.

91

Quarmby, L. 2000. Cellular Samurai: katanin and the severing of microtubules. J. Cell Sci.

113:2821-2827.

Raff, E.C., J.A. Hutchens, H.D. Hoyle, M.G. Nielsen, and F.R. Turner. 2000. Conserved

axoneme symmetry altered by a component beta-tubulin. Current Biol. 10:1391-1394.

Rasmussen, L., and E. Orias. 1975. Tetrahymena: growth without phagocytosis. Science.

190:464-465.

Redeker, V., N. Levilliers, J.-M. Schmitter, J.-P. Le Caer, J. Rossier, A. Adoutte, and M.-H. Bré.

1994. Polyglycylation of tubulin: a post-translational modification in axonemal

microtubules. Science. 266:1688-1691.

Smith, E.F., and P.A. Lefebvre. 1997. The role of central apparatus conponents in flagella

motility and microtubule assembly. Cell Motil. Cytoskeleton. 38:1-8.

Stargell, L.A., D.P. Heruth, J. Gaertig, and M.A. Gorovsky. 1992. Drugs affecting microtubule

dynamics increase *-tubulin mRNA accumulation via transcription in Tetrahymena

thermophila. Mol.Cell.Biol. 12:1443-1450.

Sturgess, J.M., J. Chao, and J.A.P. Turner. 1980. Transposition of ciliary microtubules - another

case of impaired ciliary motility. New England J. Med. 303:318-322.

Tilney, L.G., and J.R. Gibbins. 1968. Differential effects of antimitotic agents on the stability

and bahavior of cytoplasmic and ciliary microtubules. Protoplasma. 65:167-179.

Wheatley, D.N. 1982. The centriole: a central enigma of cell biology. Elsevier Biomedical Press,

Amsterdam.

Williams, N.E., J.E. Honts, and J. Kaczanowska. 1990. The formation of basal body domains in

the membrane skeleton of Tetrahymena. Development. 109:935-942.

92

Xia, L., B. Hai, Y. Gao, D. Burnette, R. Thazhath, J. Duan, M.-H. Bré, N. Levilliers, M.A.

Gorovsky, and J. Gaertig. 2000. Polyglycylation of tubulin is essential and affects cell motility

and division in Tetrahymena thermophila. J.Cell Biol. 149:1097-1106.

93

Table 3.1. Morphology of axonemal microtubules in WT and ββββDDE440 mutants

Cross-section morphology WT

(n=32)

Mutant 48 hr

(n=37)

Mutant 72 (n=22)

Normal axonemes (9+2) 32 2 0

Lack of central MTs (9+0) 0 30 3

Axonemes lacking the

central pair, with

peripheral singlets

0 1 6

Axonemes lacking the

central pair with

transposed peripheral

singlets

0 2 4

Axonemes lacking the

central pair and with

reduced total number of

peripheral tubules (<9)

(with transposed tubules)

0 (0) 2 (0) 9 (2)

94

Figure 3.1. Course of cytokinesis in WT and mutant cells labeled with propidium iodide for

DNA (red) and one of the following antibodies (green): 20H5 anti-centrin, 12G10 anti-

α−tubulin, 6-11 B1 anti-acetylated α-tubulin. A-D: Cytokinesis in WT cells. Note that the cell

shown in B. has two dividing MICs, which a frequent variation in wildtype strains. E-F. Mutant

cells after 48 hr. Abbreviations: oa, oral apparatus; noa, new oral apparatus; cs, cortical

subdivision; ma, macronucleus; mi, micronucleus; cs, cortical subdivision; ci, cilium. Bar = 25

µm

Figure 3.2. Ultrastructure of WT (A,H) and βDDDE440 cilia at 48 (B-C) and 72 hrs (D-G,I). A.

Cross-sections of WT cilia. B-D. Mutant cilia lacking the central pair and having singlet

peripheral MTs. In C arrows show singlet MTs with dynein arms. In D. one of the singlets

(asterisk) is shifted toward the center of the axoneme. E,F. Axonemes with a single central and 8

peripheral MTs, after apparent transposition. G. An axoneme missing entirely one of the

peripheral microtubules. H. A longitudinal section of WT cilium and its basal body. I. A

longitudinal section of a mutant cilium. In the transitional region, the axonemal outer

microtubules are separated from the basal body by a large gap. The surrounding plasma

membrane is collapsed (arrow). Bar = 0.1 µm.

Figure 3.3. Immunofluorescence images of WT, nondividing cells A. A cell labeled for

acetylated α-tubulin B. A cell labeled by 12G10 antibody, which recognizes α-tubulin primary

sequence. In cilia, the antibody epitope is only accessible at the tip. The inset shows higher

magnification of cell cortex. C. A cell labeled with AXO49, an antibody, which recognizes

polyglycylated tubulin isoforms. D. A deciliated cell (Calzone and Gorovsky, 1982) labeled by

95

AXO49. Abbreviations: oa, oral apparatus; cvp, contractile vacuole pore; pf, postoral fiber; bb,

basal body; p, postciliary MT band; t, transverse MT; lm, longitudinal MT bundle. Bar = 25 µm.

Figure 3.4. βDDDE440 mutant cells 24 (A-C), 48 (E) and 72 hrs (F) after phenotype induction

and βEDDD440 cells arrested in cytokinesis (D), labeled with either anti-centrin or anti-tubulin

antibodies. A-B. Mutant cells with 2 subcells showing either an absence (A) or incomplete (B)

cortical subdivision. In B. each of the subcells had already entered a new round of cell cycle as

manifested by the formation of new oral primordia. C- A mutant with two subcells in which the

cleavage furrow had retracted. The arrows on the left show an LM, which is uninterrupted. D.

Cells from the viable mutant strain βEDDD440 arrested in cytokinesis. Arrows show LMs, which

appear uninterrupted. E. A four subcell chain labeled by anti-tubulin. Note lack of cortical gaps

in the regions of constriction. F. A monster comprising multiple subcells showing extensive

breakage and disorganization of ciliary rows. Abbreviations: oa, oral apparatus; noa, new oral

apparatus Bar = 25 µm.

100

CHAPTER 4

FUNCTION OF THE β-TUBULIN GLYCYLATION DOMAIN IS DEPENDENT ON

MICROTUBULE TYPE AND CELL POLARITY IN TETRAHYMENA.

1Rupal Thazhath, Maria-Jerka Dziadosz, Jianming Duan, Dorota Wloga, Martin A. Gorovsky,

Joseph Frankel, and Jacek Gaertig. To be submitted to Molecular Biology of the Cell.

101

Abstract

Tubulin glycylation is a post-translational modification commonly found in cells with

cilia or flagella. The ciliate Tetrahymena has glycylation on ciliary and cortical microtubules.

Mutating three glycylatable amino acids on β-tubulin produced immotile 9+0 ciliary axonemes

and inhibited cytokinesis. The mutant tubulin was targeted to newly assembled and preformed

organelles at the same rate as wild type. In axonemes, the defects, including lack of the central

pair, occurred during assembly. The glycylation domain was also required for the long term

maintenance of the proper length of preformed cilia. In contrast to the aberrant assembly and

instability in cilia, mutant cortical organelles showed abnormally high number of microtubules.

A cell cycle arrest and dissociation of cortical MAPs occurred selectively in the anterior member

of incompletely divided mutant cells. Thus, the effects of tubulin glycylation sites are specific for

microtubule organelle type and dependent on cell polarity.

Introduction

Microtubules (MTs), polymers made of α/β-tubulin, are subject to a set of conserved

post-translational tubulin modifications (PTMs) (reviewed in Rosenbaum, 2000). Tubulin

glycylation (TuG) found mainly in cells with cilia or flagella (Adoutte et al., 1985; Bressac et al.,

1995), and in mammalian brain (Banerjee, 2002), occurs by addition of a glycine or a

polyglycine chain to the γ-carboxyl group of glutamic acids near the COOH-termini of α- and β-

tubulin (Redeker et al., 1994).

The ciliate Tetrahymena has polyglycylated ciliary and cortical MTs, while

monoglycylation occurs on nuclear and intracytoplasmic MTs (Thazhath et al., 2002; Xia et al.,

102

2000). Three TuG sites could be eliminated from Tetrahymena α-tubulin without detectable

consequences (Xia et al., 2000), but a similar mutation on β-tubulin led to ciliary paralysis and

arrest in cytokinesis (Thazhath et al., 2002). The mutant cilia lacked the central pairs (CPs) and

had incomplete peripheral doublets. The cytokinesis arrest was associated with incomplete

severing of longitudinal cortical MTs (LMs) (Thazhath et al., 2002).

PTMs appear to be post-assembly events that accumulate on relatively stable organelles,

including the basal bodies (BBs) cilia and flagella (Adoutte et al., 1991; Bobinnec et al., 1998;

Bré et al., 1996; Johnson, 1998; Sasse and Gull, 1988). However, even these “stable”

microtubular organelles undergo tubulin subunit exchange (Marshall and Rosenbaum, 2001;

Rosenbaum and Child, 1967; Song and Dentler, 2001; Stephens, 1997). Here, we attempt to

differentiate between the roles of TuG in assembly and post-assembly maintenance of organelles

using epitope dilution experiments to follow the fate of TuG domain mutant β-tubulin for many

generations. We found that the TuG domain regulates the length of ciliary peripheral doublets, is

required for the assembly of the CP and for the long term maintenance of axonemes. In contrast

to assembly/stability defects in cilia, organelles in the mutant cortex had an excessive number of

MTs. Furthermore, cell cycle progression was arrested in an antero-posterior gradient in

subdivided mutant cell chains. Thus, TuG has multiple roles that are organelle dependent and

strongly influenced by cell polarity.

Materials and Methods

Construction of strains with inducible expression of tagged β-tubulin

The pTTMN plasmid (Shang et al., 2002) was used to construct the pBHM plasmid, by

replacing the MTT1 gene coding region with that of BTU1 and inserting the HA epitope tag

103

sequence in front of the stop codon. The βDDDE440 heterokaryons (Thazhath et al., 2002) were

mated and after 24 hrs transformed biolistically with the MTT1-BTU1-HA-MTT1 fragment

obtained by SacI digestion of pBHM. Transformants were selected on SPP medium with

paramomycin (pm, 120 µg/ml) and cadmium chloride (Cd, 2 µg/ml). Transformants (βDDDE440-

CD) had a wildtype (WT) phenotype developed the βDDDE440 mutant phenotype on the medium

lacking Cd. To obtain a WT strain suitable for studies on the flux of tubulin, the β-tubulin

knockout heterokaryons (Xia et al., 2000) were mated and rescued by transformation with the

MTT1-BTU1-HA-MTT1 fragment. Transformants that integrated the fragment into the MTT1

locus were selected in SPP with 120 µg/ml pm and 5 µg/ml Cd. A transformant strain (Cd-

dependent for growth) was subjected to biolistic transformation with the WT Hind III-Bgl II

BTU1 fragment, to replace the disrupted region of BTU1. The final transformants (named WT-

CD) were selected which were able to grow without Cd.

Analyses of the βDDDE440 phenotype

Tetrahymena cells were grown and mated and the constitutive βDDDE440 phenotype was

induced by mating mutant heterokaryon strains (Thazhath et al., 2002). For phenotype induction

in the βDDDE440-CD strain, cells were grown in medium containing Cd for several generations

and then starved for 8 h in 10mM Tris pH 7.5. The cell concentration was adjusted to 1 x 104

cells/ml and cells were transferred to MEPP medium (Orias and Rasmussen, 1976) lacking Cd

and copper salts. Cell concentration was maintained between 1 x 104-1 x 105 by periodic

dilutions. For immunofluorescence (Thazhath et al., 2002), the following primary antibodies

were used: monoclonal acetylated α-tubulin 6-11 B1 (1:10 dilution, provided by Gianni Piperno,

Mount Sinai Medical School, NY), monoclonal anti 13C4p (1:1 dilution); monoclonal anti-α-

104

tubulin 12G10 (1:25); polyclonal or monoclonal anti-HA (Clontech, 1:100) and polyclonal anti-

total Tetrahymena tubulin (SG) antibodies (1:200). Secondary antibodies were goat anti-mouse

FITC, and goat anti-rabbit-Cy3 (Zymed) conjugates (1:100). Cells were viewed using a Leica

TCS SP2 Spectral Confocal Microscope with Coherent Ti: sapphire multiphoton laser (Mira

Optima 900-F). TEM was done as described (Jerka-Dziadosz et al., 2001). The deciliation was

carried out as described (Calzone and Gorovsky, 1982).

Results

Construction of a conditional, cadmium-dependent TuG domain mutant

We constructed a conditional β-tubulin TuG domain mutant, whose phenotype can be

induced in vegetative cells, and in which the WT and mutant β-tubulin can be distinguished

using epitope tagging. The βDDDE440-CD strain has two β-tubulin genes: a βDDDE440 TuG

domain mutant gene located at the β−tubulin encoding BTU1 locus and a second gene Cd-

inducible, WT BTU1 coding region with a COOH terminal hemagglutinin (HA) tag in the MTT1

locus controlled by the MTT1 promoter (Shang et al., 2002). In the βDDDE440 mutation 3

adjacent glutamic acids homologous to glycylation sites in Paramecium (Vinh et al., 1999),

(positions 437-439) are replaced with aspartic acids (Thazhath et al., 2002). The βDDDE440-CD

cells exhibit a WT phenotype in the presence of Cd, which induces the WT-HA β-tubulin (the

βDDDE440 mutation is recessive). After removal of Cd, the βDDDE440-CD cells express mainly

mutant tubulin and develop the βDDDE440 phenotype (see below). We also constructed a control

strain (WT-CD), WT in the BTU1 locus with a WT–HA-tagged BTU1 coding region in the

MTT1 locus. These cells showed a normal phenotype + Cd indicating that overexpression of

tagged β-tubulin was not detrimental. Heterokaryon βDDDE440 mutants that express the

105

phenotype after conjugation undergo 2-3 normal cell cycles (corresponding to a period of

dilution of the preexisting WT tubulin), before becoming immotile chains of subcells with 3 or

more cleavage furrows and multiple nuclei (Thazhath et al., 2002). The βDDDE440-CD mutant

cells also divided normally up to 3 times, after Cd removal, and before losing motility and

arresting in cytokinesis forming characteristic subdivided chains (FIG. 4.1B, Fig. 4.2, Fig. 4.3).

The tubulin glycylation domain affects the cilia length and is required for formation of the

central pair

Since the βDDDE440-CD mutants express both HA-tagged WT tubulin and the nontagged

βDDDE440 mutant tubulin, we could trace the localizations of both tubulins in the same cells. We

initially grew the βDDDE440-CD mutants in the presence of Cd for many generations to achieve

uniform labeling of all MTs with WT-HA tubulin (Fig. 4.2A). To induce the phenotype, these

cells were transferred to a medium lacking Cd and labeled with anti-HA and anti-total tubulin

antibodies.

At 13-15 hrs after Cd removal, we observed two types of normal length cilia,

distinguishable by the intensity of the HA signal (green in Fig. 4.2, Fig.4.3) relative to the total

tubulin (red in Fig. 4.2, 4.3). Some cilia showed strong (arrowheads, Fig. 4.2B, C) while other

cilia showed weaker HA labeling (asterisks, Fig. 4.2B, C), in both cases uniform along the entire

length. In Tetrahymena, new basal bodies (BBs) within rows of somatic cilia form by a so-called

templated mechanism in which they appear to form in association with old BBs (Allen, 1969).

New BBs are then inserted anterior to old ones within the ciliary row and undergoes ciliation.

Neither the old BB nor its cilium is resorbed. Thus, the simplest interpretation of this staining

pattern is, that the strongly HA-positive cilia are old cilia formed prior to Cd removal, while the

106

weakly labeled cilia were formed after Cd removal, and contain higher amounts of mutant

tubulin and less residual WT tubulin. Our analysis of the control WT-CD strain showed a similar

pattern (Fig. 4.4A).

After 15h, large areas devoid of cilia appeared in the mutant, mainly in the equatorial

region (Fig. 4.2C arrow). Thus, either BB duplication or ciliogenesis was affected. However, the

formation of BBs appeared normal. A longitudinal section of the 2 subcell stage showed cortical

rows with densely packed BBs and normal associated structures (Fig. 4.5A). A longitudinal

section of a cell in a similar stage revealed multiple adjacent BBs lacking axonemes. These BBs

had a transitional plate, the site from which the CP normally assembles (Fig. 4.6A). Thus, BBs in

the mutant experience a delay in ciliation (see below), which was not observed in the WT-CD

strain (Figs. 4.4A-D).

Starting at 22 hrs, the mutant cortex was covered with new cilia made primarily of mutant

tubulin, (Fig. 4.2D, Fig. 4.3A arrows), which, however, were much shorter than WT (Fig. 4.2D,

Fig. 4.3A-C). TEM of the 2 subcell mutants revealed very short cilia lacking CP (Fig. 4.6A-C).

Particularly striking are images of newly developing ciliary oral membranelles composed of

many exclusively 9+0 axonemes (Fig. 4.6B, E, F). Finally, some new cilia had swollen tips with

doublet MTs terminating prematurely (Fig. 4.6C). Thus, the mutant tubulin can be utilized to

assemble new axonemes but the resulting structures are short and incomplete.

Two “invariant zones” in the cell cortex contain old ciliary units in which tubulin subunits

slowly exchange

In mutant cells, a number of cilia retained a high content of HA tubulin and normal

length for many generations after Cd removal. Large clusters of old cilia were particularly

107

prominent at the anterior end (Fig. 4.3A, labeled “a”). Electron micrographs at the 2 subcell

stage showed ultrastructurally normal cilia concentrated in the apical anterior region of a mutant

(Fig. 4.6D). Thus, in the “old” anterior cilia, the presence of a normal 9+2 structure correlates

with retention of WT tubulin. At 22.5 hr, some cells also had a smaller posterior region enriched

in old cilia (Fig. 4.2D, labeled “p”). At about the same time we also observed WT-CD cells with

either an anterior or posterior group of old cilia, (Fig. 4.4B, C). The existence of these so called

“invariant zones” (Iftode et al., 1989) can be explained by lack of insertion of new basal bodies

in the areas close to the cell extremities as the cell doubles its size prior to division (Fig. 4.1A).

The sizes of these invariant zones diminished with increased number of generations in WT-CD

and mutant cells (Fig. 4.4B and D, 4.3A-C), suggesting that insertion of ciliary units near the cell

poles probably occurs every cell cycle.

In the mutants with an anterior invariant region, the membranelles of the oral apparatus

(OA) labeled strongly with anti-HA antibodies, indicating little exchange (Fig. 4.3A, oa). In

contrast, cells with a posterior invariant zone invariably had a weak HA signal in the OA (Fig.

4.2D). A similar correlation was seen for WT-CD cells (Fig. 4.4B, C). The two types of cells are

most likely derived from the anterior or posterior daughter cell of the previous cell cycle

respectively (Fig. 4.1A). A WT example of such a dividing cell (prior to separation of the

anterior and posterior invariant zones by division) is shown in Fig. 4.4A. Prior to cell division, a

new OA is formed in the equatorial region (Fig. 4.1A). At the same time, the old OA undergoes

only limited remodeling (Frankel, 1999). Thus, the cell with an anterior invariant region and

conserved OA is likely to be the anterior daughter while the cell with the posterior invariant

region and new OA is likely derived from the posterior daughter (Fig. 4.1A). Importantly, the

108

pattern of insertion of new ciliary units and conservation of the content of old oral structures

appears unchanged in the mutant.

Interestingly, old cilia in both the βDDDE440-CD and WT-CD strain showed

incorporation of newly made tubulin. At 22-36 hrs, many old cilia of mutant cells had short

segments at their tips that were devoid of WT-HA tubulin but were labeled by the general anti-

tubulin antibody, indicating that they were made predominately of βDDDE440 tubulin

(arrowheads in Fig. 4.2D and 4.3A, B). Thus, mutant tubulin can be added to the distal ends of a

preexisting largely WT axoneme. With similar kinetics, exchange at the tips of old cilia was seen

in WT-CD cells (Fig. 4.4D arrowheads). Thus, there is no defect in the ability of mutant tubulin

to be incorporated into preformed cilia.

At later stages, when mutant cells were composed of 3-4 subcells, the HA-positive old

cilia were shorter, and the HA signal was barely detectable at 80 hrs (Fig. 4.3C insets,

arrowheads). This reduction in HA signal occurred along the length of the mutant axoneme (Fig.

4.3C inset). Thus, cilia made primarily of mutant tubulin shorten whether they were assembled

“de novo” from mutant tubulin or obtained a high content of mutant tubulin by post-assembly

exchange. We were unable to verify whether the shortened old cilia retained their CP, but that

appears unlikely because at 85 hrs the mutant chains were completely paralyzed and DIC showed

no movement of cilia at the cell ends (data not shown).

In the WT-CD strain, with increased time (62-72 hrs), we observed 2 types of cells with

old cilia. Some cells had only a few or sometimes just one old cilium in the central region (Fig.

4.4E). Such cells are most likely derived from the posterior daughter of a cell with an anterior

invariant zone or from an anterior daughter of a cell with a posterior invariant zone. At 72 hrs

(corresponding to 13 generations), very few cells showed weakly HA-positive cilia within the

109

anterior invariant zone, and had low content of HA-tagged tubulin uniformly distributed along

the axoneme, was clearly above background, (Fig. 4.4F inset). Thus, both mutant and WT-

tubulin are capable of replacing pre-existing WT tubulin in preformed cilia, in what appears to be

along the length of the cilium. TuG does not appear to affect the rate and pattern of exchange;

however, in addition to its role in proper assembly, TuG is also required for the long-term

maintenance of the proper size and structure of cilia.

Slow exchange of tubulin occurs at different rates in different cortical organelles

To study cortical tubulin exchange, we partially deciliated and examined the mutants.

The pattern of tubulin flux in the mutant cortex was similar to that in cilia, with a clear anterior

and a less prominent posterior invariant zone (Fig. 4.7B,C). Similar results were obtained with

the WT-CD strain (Fig. 4.7A). Furthermore, old cortical MTs also undergo slow exchange. Many

old transverse MT bundles (TMs) showed tips with higher contents of mutant tubulin compared

to the more proximal region of the same bundle (Fig. 4.7A, 4.7C, inset), demonstrating that

subunit exchange in TMs is initiated distal to the basal body, and proceeds to the basal body

proximal end, as in cilia. The subunit exchange in LMs was much more rapid compared to TMs

in both types of cells, as LMs labeled only weakly with the HA antibody by 12hrs (Fig. 4.7A,B).

The contractile vacuole pore (CVP) also underwent a relatively rapid subunit exchange as no

significant HA labeling was seen at 12 hr (Fig. 4.4A, note that the cell shown has both an

anterior and a posterior invariant zone indicating that the posterior CVPs in this cell were formed

prior to Cd removal). Thus, the rates of exchange in the cell cortex are dependent on the

organelle type. Importantly, the subunit targeting, exchange pattern and rates appear unaffected

for the glycylation-deficient β−tubulin.

110

Cortical microtubular organelles in the βDDDE440 mutants are hypertrophic

At late times after Cd removal, the density of cilia of the βDDDE440-CD mutants

increased dramatically compared to WT-CD cells (Fig. 4.3B, C). This correlated with an

increased density of BBs within cortical rows (Fig. 4.5A). Although the majority of BBs were

ultrastructurally normal, some showed additional MTs associated with the triplets (arrows, Fig.

4.5C, D). The additional basal body and cortical MTs were not seen in WT cells.

The organization of LMs in the mutants was of special interest because we found

previously that these MTs fail to sever during cytokinesis and therefore could obstruct the

contractile ring. Indeed, we now found that LM bundles in the mutants at the 2 subcell stage

contained almost twice as many MTs compared to WT. LMs in the mutant had an average of

11.7 MTs (±3.7, n= 37) as compared to 6.3 in WT (±1.8, n= 12). Fig. 4.5B shows an LM of the

βDDDE440 mutant with a profile of 14 MTs (arrows). This hypertrophy of LMs could be directly

related to their impaired severing during cytokinesis.

Tetrahymena has a complex OA comprised of 4 compound ciliary structures with about

120 cilia. The undulating membrane (UM) is made of a double file of BBs, of which only one is

ciliated. Three membranelles, each consisting of three files of BBs (Adoral Zone of

Membranelles, AZM), are located to the cell’s left of the UM (Frankel, 1999). In a WT AZM,

among the 3 rows of BBs, only the most posterior one has well developed postciliary (PC) MTs,

while in the 2 more anterior rows PC is limited to a single MT (Jerka-Dziadosz, 1981).

Strikingly, PCs in some mutant OAs the anterior and middle AZM rows had PCs with multiple

MTs (Fig. 4.5E). Thus, many microtubular organelle types in the mutant cortex show

hypertrophic features.

111

Polarity- dependent dissociation of a MAP in βDDDR440 mutants.

To further examine changes in the cell cortex of the βDDDE440 mutant, we studied the

distribution of a MAP recognized by the 13C4 antibody (Jerka-Dziadosz et al., 1995), which

recognizes a single Triton X100-insoluble protein band of 80 kD (M. Nelsen and JF, unpublished

observations). In WT, the 13C4 epitope was found on all cortical MTs (Fig. 4.8A). As the WT

cell entered division (Fig. 4.8A), newly formed and old organelles were uniformly labeled.

However, the βDDDE440 mutant chains show a loss of association of MTs with 13C4p (Fig.

4.8B-E). Furthermore, the extent of 13C4p loss was dependent on cell polarity. In the mutant

chains one can easily determine the polarity: the most anterior subcell has an OA close to the

chain’s end, while the posterior most subcell has its CVPs near the end of the chain (labeled “p”

in Fig. 4.8). Strikingly, 13C4p preferentially disassociated from the more anterior subcells (Fig.

4.8B, D). Concomitant with this loss of association with the cortex, an increase of granular

labeling in the cell body was observed (Fig. 4.8D, E) while cortical MTs were still present (Fig.

4.8F-I). When the intensity of 13C4p label is compared among the subcells, there is a clear

gradient of intensity – with the most anterior subcell having the least amount of cortical 13C4p

(Fig. 4.8C).

After a long period of mutant phenotype induction, subcells underwent cortical

integration by retraction of the cleavage furrow and sliding of cortices of subcells. This

integration led to extensive twisting and breaking of ciliary rows. Without exception, we

observed that the process of integration started at the anterior most ends of the chains. Chains

were commonly observed with 4 sets of nuclei in which the 3 anterior subcells were integrated

while the most posterior subcell remained unintegrated (Fig. 4.8D, I). In these cases, the

integrated subcells had markedly less cortex-bound 13C4 while the posterior subcell showed

112

normal cortical 13C4 localization. Thus, in addition to hypertrophic features, the mutant cortex

also shows degenerative features, including dissociation of the 13C4p MAP, that are influenced

by cell polarity.

The anterior subcell of βDDDE440 mutant exhibits selective arrest of the cell cycle

The pattern of dissociation of 13C4p indicated that there might be an unusual mode of

growth and division in the mutant chains. To investigate the division pattern, we induced the

constitutive mutant phenotype (by mating mutant heterokaryons) and examined living cells by

video microscopy. Observations on living mutants revealed a "posterior dominance" phenotype,

observed in 76% of cells (n=68) (Fig. 4.10A) in which, after the first cleavage furrow arrest (Fig.

4.9A), only the posterior subcell grew and divided. This led to the formation of a three-subcell

chain (Fig. 4.9B). In most cases, the next round of division also resumed in the posterior-most

subcell of the same chain leading to the formation of a four-subcell chain (Fig. 4.9C) in which, 3

subcells formed as anterior daughters (each formed at a different time) and only one subcell was

formed as a posterior daughter (summarized in Fig. 4.10B). Importantly, the arrest in cortical

growth in the anterior subcells was also associated with an apparent arrest in the nuclear cell

cycle. When the nuclei were undergoing division in the most posterior subcell, the nuclei of

anterior subcells appeared not to divide (Fig. 4.9E). Furthermore, the size of the macronucleus

(MAC) in the anterior subcells was considerably smaller than the size of the predivision

(presumably G2) MAC in the posterior subcell, suggesting that they had not replicated their

DNA (Fig. 4.9E). Division of the MIC is also restricted to the posterior-most subcell (Fig. 4.9E).

Therefore, the βDDDE440 mutants also display a novel phenotype based on selective cortical and

nuclear growth arrest in the anterior half of the dividing cell (Fig. 4.10B).

113

We noticed that βDDDE440 mutant cells periodically undergo oral replacement. This

phenomenon is often observed in Tetrahymena cells subjected to stressful conditions (Frankel

and Williams, 1973). During oral replacement, a new OA is formed immediately posterior to the

old OA. It is likely that oral replacement in mutants occurred as a response to abnormal assembly

of MTs around the oral BBs as described above. However in the mutant chains, there was a

gradient of replacement – the posterior subcell was usually in the most advanced stage (Fig.

4.9D). The simplest explanation is that the process of oral replacement was initiated first within

the posterior subcell, and began later in more anterior subcells. This suggests that the posterior

subcell produces a signal that triggers the morphogenetic process in the anterior subcells. These

results argue that while the mutant anterior subcells are less active in initiating growth and

cortical morphogenesis, they can still respond to some types of morphogenetic signals

transmitted from the posterior subcell.

Discussion

This study was undertaken to address the ability of TuG-deficient tubulin to participate in

assembly and post-assembly maintenance of organelles and to shed light on differences between

the cortical and ciliary functions of TuG in Tetrahymena.

We observed an effect of the TuG domain on ciliary assembly. As the levels of

preexisting WT tubulin were depleted in the βDDDE440-CD strain, assembly of the entire

axoneme was inhibited. With a delay, however, new cilia formed from mutant tubulin but they

remained very short and lacked the CP. The fact that both the CP and the outer doublets were

affected is consistent with the presence of TuG in both structures (Pechart et al., 1999). The

shortness of the axonemes deficient in TuG is probably a consequence of low TuG rather than

114

the absence of CP, since Chlamydomonas mutants that lack CP, have normal length flagella

(Adams et al., 1981).

We can propose three explanations for why deficiency in TuG leads to a failure to form

CP. One possibility is that TuG affects intraflagellar transport (IFT), the motility system

operating inside cilia and flagella that delivers components from the cell body (Rosenbaum and

Witman, 2002). Short doublet MTs (but not CP) can be assembled in Tetrahymena IFT mutants

(Brown et al., 2003; Brown et al., 1999b) and in sea urchin eggs in which IFT was inhibited by

antibodies (Morris and Scholey, 1997). Thus, CP assembly could be more dependent on IFT than

the doublets.

A second explanation for the stronger effect on CP reflects the mode of CP nucleation.

The CP MTs are nucleated from the transitional plate while the peripheral MTs are extensions of

the triplets of the BB. A deficiency in γ-tubulin led to assembly of 9+0 axonemes in

Trypanosoma (McKean et al., 2003) indicating CP MTs are more sensitive to lack of γ-tubulin

than doublets. TuG could be involved in the nucleation of CP MTs within the transitional region

in conjunction with γ-tubulin.

Finally, it is possible that lack of TuG affects the ends of growing axonemal MTs,

making the CP less stable. CP formation involves attachment of the plus end of growing MTs to

the ciliary membrane via capping complexes (Sale and Satir, 1977). TuG could participate in

association of the capping complexes with the ends of CP MTs. Capping complexes also

connect the doublets with the ciliary membrane (Dentler, 1980). However, doublets may still

polymerize without a proper attachment due to their higher intrinsic stability. In agreement with

this hypothesis, some mutant cilia had highly disorganized doublets near the tips (see Fig. 4.6C).

115

While the TuG domain mutation led to reduced polymerization or stability of ciliary

MTs, it had a seemingly opposite effect on cortical MTs. We observed an increased density of

BBs and supernumerary MTs near BB triplets. Furthermore, the LMs as well as PC MTs of oral

membranelles in mutant cells contained an abnormally high number of MTs. The abnormal LMs

could be too thick to be properly severed at the time of contractile ring constriction leading to

failure of cytokinesis (Thazhath et al., 2002). Alternatively, the presence of TuG could provide a

signal for recognition of LMs by severing factors at the time of cytokinesis, and hypertrophy of

LMs could be a result of lack of severing. Interestingly, another polymodification, tubulin

glutamylation was recently linked to axonemal MT severing. The Chlamydomonas FA2 mutant

has a MT severing deficiency phenotype, leading to lack of deflagellation (Mahjoub et al., 2002).

The FA2 gene encodes a NIMA-type kinase, and another NIMA-related kinase of Crithidia

copurified with tubulin glutamylase activity (Westermann et al., 1999). Thus, both tubulin

glutamylation and glycylation could be involved in marking MTs for severing.

Two explanations for the origin of the hypertrophic cortex can be considered. First, it

might be secondary to ciliary defects. Lack of proper assembly of axonemal MTs could lead to

an increased concentration of unpolymerized tubulin in the cell body, which in turn could drive

hyperpolymerization of MTs. However, at least in the case of LMs, this appears unlikely.

Mutants that fail to assemble cilia due to the absence of key IFT proteins (kinesin-II and Ift52p),

are able to assemble and disassemble LMs and divide normally, as long as external physical

force is provided to substitute for the motile process (rotokinesis) used by Tetrahymena for

abscission of the cytoplasmic bridge (Brown et al., 2003; Brown et al., 1999a). Thus, it is more

likely that TuG has a direct function in the morphogenesis of cortical MTs and that this function

is required for limiting MT polymerization.

116

The fundamentally different influence of the same mutation on cortical and ciliary MTs

suggests that either TuG changes the intrinsic properties of different MTs in distinct ways, or that

the modification is a signal which can be interpreted differently by localized trans-acting

mechanisms. The first hypothesis would require that the TuG content, site utilization or lateral

chain length be different between the cortical and ciliary MTs. It is possible that among the 5

major sites of TuG on β-tubulin, nonidentical sites are used in the cell cortex and in cilia.

However, this is extremely unlikely because cells can have a nearly normal phenotype when 4 of

these sites are inactivated, and none of the 5 sites is essential, either individually or in pairs (Xia

et al., 2000). Also, virtually all sites of TuG (and of tubulin glutamylation) can be inactivated on

β-tubulin by simultaneously replacing the COOH-terminal tail of α-tubulin with the

corresponding sequence of β-tubulin (Duan and Gorovsky, 2002; Xia et al., 2000). It is therefore

more likely that it is not the site but rather the glycine side chain length that differs between the

cortex and cilia. Consistent with this, mass spectrometry revealed that the cortical and cell body

tubulins of Paramecium had much shorter chains compared to axonemal tubulins (Bre et al.,

1998). The second hypothesis (not mutually exclusive with the first one) postulates that TuG acts

as a signal for binding distinct MAPs. In cilia, such organelle-targeted MAPs could have a

polymerization –promoting and stabilizing effect, while in the cell cortex different MAPs may be

required for limiting polymerization.

An important byproduct of this study was increased understanding of how the dynamics

of tubulin flow in preexisting cilia and cortical organelles relates to the growth of the cortical

cytoskeleton. In both mutant and WT cilia, nearly complete replacement of tubulins required

about 80 hrs, which is equivalent to ~13 generations in WT cells under the conditions used here.

Thus, assuming that the exchange rate was constant, we observed an exchange of about 1% per

117

hr. Earlier pulse chase studies in Tetrahymena pyriformis, showed a similar low turnover rate of

2.25 %/hour (Nelsen, 1975). Studies on flagella also indicated a slow rate of exchange in

Chlamydomonas (Gorovsky et al., 1970). A more recent study in Chlamydomonas (Song and

Dentler, 2001) showed a higher turnover rate of a minimum of 20% of total flagellar proteins

within 6h. Thus, our study confirms previous reports documenting subunit exchange in

preformed organelles that is relatively slow and requires many generations for completion.

We also obtained new information about the pattern of exchange of subunits inside cilia

and cortical MTs. A study of tubulin exchange in Chlamydomonas flagella using tagged tubulin

in dikaryons showed that preformed flagella exchange tubulin subunits within the distal 1/3 of

the structure (Marshall and Rosenbaum, 2001). However, in that, tubulin subunit exchange

beyond the distal 1/3 could not be studied, owing to complete resorption of flagella in

Chlamydomonas dikaryons in preparation for meiosis. Although we observed exchange at the

distal tips of axonemes, longer times demonstrated a labeling with the HA epitope tagged tubulin

and untagged tubulin along the entire length of the axoneme in a uniform manner. The labeling

patterns observed, indicated a possible uniform exchange occurring along the length of the

cilium. A similar pattern of exchange along the entire axoneme was observed for a component of

the paraflagellar rod in preformed flagella of Trypanosoma (Bastin et al., 1999). Further studies

should address this possible insertion of new subunits along the length of the axoneme.

In Tetrahymena, stable ciliary units rich in relatively old tubulin were present in two

clusters – a larger one at the anterior end and a smaller zone at the posterior end of the cell.

These areas can be described as “invariant zones” – territories of the cortex with a low

probability of insertion of new cortical units during the growth phase of the cell cycle. Our

results agree with the previously observed preferential insertion of new BBs in the equatorial

118

region of Tetrahymena (Frankel et al., 1981; Kaczanowski, 1978; Nanney, 1975). The posterior

and anterior invariant zones have been extensively documented in another ciliate Paramecium

tetraurelia, based on the mapping of newly assembled BBs (Iftode et al., 1989).

In Paramecium, two waves of basal body proliferation lead to insertion of new units

within the variable territories, while the invariant zones seemed to remain unaffected (Iftode et

al., 1989). In Tetrahymena, we kept track of the shape and size of the invariant zones for many

generations, and we observed that with time the size of these zones decreased. We can account

for this by proposing that a wave triggering insertion of BBs (comparable to the first wave in

Paramecium) propagates decrementally from the equatorial region toward both the anterior and

posterior ends. In the invariant zones this signal becomes very weak, yet with time even the low

probability of insertion of a new unit into the invariant zones leads to translocation of some

preexisting units toward the equator and reduction in the size of invariant zones in subsequent

generations. Apparently in Tetrahymena, the waves of signaling that control formation of new

BBs may affect the posterior end more strongly, as this invariant zone is much smaller, and

disappears earlier.

The invariant zones are not only structurally static but also are distinct at the molecular

level, as evident from relatively slow exchange of preexisting tubulins. This is in agreement with

studies on Oxytricha fallax, which showed remarkable conservation of the AZM at both the

molecular and morphological level over several generations (Grimes and Gavin, 1987).

However, we also revealed major differences between the rates of subunit exchange among

subtypes of “stable” ciliary and cortical MTs. More rapid subunit exchange was found in the

LMs. Interestingly, in Paramecium this structure (Sundaraman and Hanson, 1976), subsequently

named the “cytospindle” (Cohen et al., 1982) appears transiently during cell division and

119

disassembles when division is completed (Iftode et al., 1989). Although in Tetrahymena LMs

persist during the entire cell cycle, they exchange subunits more rapidly than other cortical

structures, which is in agreement with the dynamic character of cytospindle in Paramecium.

Furthermore, CVP MTs also exchange tubulin rapidly. The differences in the exchange rates of

specific types of cortical structures likely reflect different functions of these organelles.

An unexpected observation was the apparent polarized effect that lack of glycylation had

on cortical growth and nuclear cell cycle. Starting from a two-subcell stage, cortical growth and

the nuclear divisions are arrested selectively in anterior subcells (Fig. 4.10B). Two conclusions

can be drawn. First, it appears that there is a difference between the anterior and posterior subcell

(see below). Second, it is evident that Tetrahymena cells have a coupling mechanism, that

coordinates cortical and nuclear development (Frankel et al., 1976), and can operate

differentially in space. It has been observed in the ciliate Stentor that when cells of different

stages of the cell cycle are fused together, they tend to undergo cell cycle events synchronously

(Tartar, 1966). This is not the case with βDDDE440 mutant chains. One possibility is that there is

no true cytoplasmic exchange between the subcells of βDDDE440 chains. However, confocal

sections show relatively large cytoplasmic channels between the subcells, making this very

unlikely. Also, subcells of chains created by electrofusion have narrower cytoplasmic connection

and yet they undergo both the nuclear cell cycle and cortical development in a synchronous

manner (Gaertig and Iftode, 1989). Furthermore, most subcells of βDDDE440 chains eventually

integrate by retraction of cleavage furrows and cortical sliding, confirming that the cytoplasmic

connections between the individual subcells exists. Finally, subcells periodically undergo oral

replacement, with the most posterior subcell most advanced, suggesting this process is initiated

in the most posterior subcell and some factor is transmitted from the posterior to anterior

120

subcells. All these data argue that, cytoplasmic exchange or cortical transmission of signals

exists between the subcells, but that once a given territory is established as anterior following

partial cytokinesis, it is subjected to arrest in the cell cycle, accompanied by highly localized

communication between the local cortical domain and the nuclei within a subcell, so that nuclei

are also arrested.

Why would the anterior subcell be selectively affected by the TuG mutation? A simple

possibility is that as the production of WT tubulin ceased, the mutant tubulin is incorporated

preferentially into the anterior half of the cell. Our tagging experiments lead us to reject this

hypothesis. We did not observe a preferential accumulation of mutant tubulin in the anterior

subcell, and the anterior subcell actually inherits a larger supply of old WT tubulin due to the

presence of a larger anterior invariant zone. An opposite hypothesis could be considered – the

anterior subcell’s arrest could be the consequence of increased content of “old” tubulin even

though it is WT. There could be a requirement for rejuvenation of old tubulin, based on subunit

exchange, which may not function properly in the mutant. This hypothesis is contradicted by our

observations showing that targeting and rate of exchange of pre-existing tubulin is unaffected in

the mutant. It should therefore be considered that there are properties of the anterior subcell that

are uniquely dependent on TuG. It should be noted that we could not detect any difference in the

levels of mono- and polyglycylation between the anterior and posterior subcell of dividing WT

cells using antibody staining (data not shown). However, we proposed earlier that the action of

TuG is dependent on the presence of MAPs, which have organelle specific functions. In a similar

way, there might be MAPs yet to be discovered, whose action is confined to the anterior half of

the dividing cell.

121

Acknowledgements

We thank a UGA undergraduate, Henry Heffler for assistance in the construction of WT-

CD strains and Iza Strzyzewska-Jowko (Nencki Institute, Warsaw) for 12G9

immunofluorescence. We are grateful to the staff of the Center for Ultrastructural Research of

the UGA for continuing technical support. We thank Keith Gull (University of Oxford, UK),

Peter Satir (Albert Einstein School of Medicine, NY) and Norman Wiliams (University of Iowa)

for helpful comments. The 12G9 and 12G10 monoclonal antibodies were raised by E. Marlo

Nelsen (University of Iowa). The 12G10 antibodies are available from the Developmental

Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by the

University of Iowa. This work was supported by the NSF grant number 235826 to JG and an

NIH award GM026873 to MAG.

122

References:

Adams, G.M.W., B. Huang, G. Piperno, and D.J. Luck. 1981. Central pair microtubular complex

of Chlamydomonas flagella: polypeptide composition as revealed by analysis of mutants.

J Cell Biol. 91:69-76.

Adoutte, A., M. Claisse, R. Maunoury, and J. Beisson. 1985. Tubulin evolution: Ciliate-specific

epitopes are conserved in the ciliary tubulin in metazoa. J.Mol.Evol. 22:220-229.

Adoutte, A., P. Delgado, A. Fleury, N. Levilliers, M.-C. Lainé, M.-C. Marty, E. Boisvieux-

Ulrich, and D. Sandoz. 1991. Microtubule diversity in ciliated cells: Evidence for its

generation by post-translational modification in the axonemes of Paramecium and quail

oviduct cells. Biol.Cell. 71:227-245.

Allen, R.D. 1969. The morphogenesis of basal bodies and accessory structures of the cortex of

the ciliate protozoan Tetrahymena pyriformis. J.Cell Biol. 40:716-733.

Banerjee, A. 2002. Coordination of posttranslational modifications of bovine brain alpha-tubulin.

Polyglycylation of delta2 tubulin. J Biol Chem. 277:46140-4.

Bastin, P., T.H. McRae, S.B. Francis, K.R. Matthews, and K. Gull. 1999. Flagellar

morphogenesis: Protein targeting and assembly in the paraflagellar rod of Trypanosomes.

Mol. Cell. Biol. 19:8191-8200.

Bobinnec, Y., A. Khodjakov, L.M. Mir, C.L. Rieder, E. B., and M. Bornens. 1998. Centriole

disassembly in vivo and its effect on centrosome structure and function in vertebrate cells.

J. Cell Biol. 143:1575-1589.

Bré, M.-H., V. Redeker, M. Quibell, J. Darmanaden-Delome, C. Bressac, J. Cosson, P. Huitore,

J.-M. Schmitte, J. Rossier, T. Johnson, A. Adoutte, and N. Levilliers. 1996. Axonemal

tubulin polyglycylation probed with two monoclonal antibodies: widespread evolutionary

123

distribution, appearance during spermatozoan maturation and possible function in

motility. J. Cell Sci. 109:727-738.

Bre, M.H., V. Redeker, J. Vinh, J. Rossier, and N. Levilliers. 1998. Tubulin polyglycylation:

differential posttranslational modification of dynamic cytoplasmic and stable axonemal

microtubules in Paramecium. Mol. Biol. Cell. 9:2655-2665.

Bressac, C., M.H. Bré, J. Darmanaden Delorme, M. Laurent, N. Levilliers, and A. Fleury. 1995.

A massive new posttranslational modification occurs on axonemal tubulin at the final

step of spermatogenesis in Drosophila. Eur.J.Cell Biol. 67:346-355.

Brown, J.M., N.A. Fine, G. Pandiyan, R. Thazhath, and J. Gaertig. 2003. Hypoxia regulates

assembly of cilia in suppressors of Tetrahymena lacking an intraflagellar transport

subunit gene. Mol. Biol. Cell. 14:3192-3207.

Brown, J.M., C. Hardin, and J. Gaertig. 1999a. Rotokinesis, a novel phenomenon of cell

locomotion-assisted cytokinesis in the ciliate Tetrahymena thermophila. Int. Cell Biol.

Rep. 23:841-848.

Brown, J.M., C. Marsala, R. Kosoy, and J. Gaertig. 1999b. Kinesin-II is preferentially targeted to

assembling cilia and is required for ciliogenesis and normal cytokinesis in Tetrahymena.

Mol. Biol. Cell. 10:3081-3096.

Calzone, F.J., and M.A. Gorovsky. 1982. Cilia regeneration in Tetrahymena. Exp.Cell Res.

140:474-476.

Cohen, J., A. Adoutte, S. Grandchamp, L.M. Houdebine, and J. Beisson. 1982. Immunochemical

study of microtubular structures throughtout the cell cycle of Paramecium. Biol. Cell.

44:35-44.

124

Dentler, W.L. 1980. Structures linking the tips of ciliary and flagellar microtubules to the

membrane. J Cell Sci. 42:207-20.

Duan, J., and M.A. Gorovsky. 2002. Both carboxy terminal tails of alpha and beta tubulin are

essential, but either one will suffice. Curr Biol. 12:313-316.

Frankel, J. 1999. Cell Biology of Tetrahymena. Methods Cell Biology. 62:27-125.

Frankel, J., L.M. Jenkins, and L.E. DeBault. 1976. Casual relations among cell cycle processes in

Tetrahymena pyriformis. An analysis employing temperature sensitive mutations. J. Cell

Biol. 71:242-260.

Frankel, J., E.M. Nelsen, and L.M. Jenkins. 1981. Development of the ciliature of Tetrahymena

thermophila. II. Spatial subdivision prior to cytokinesis. Dev. Biol. 88:39-54.

Frankel, J., and N.E. Williams. 1973. Cortical development in Tetrahymena. In The biology of

Tetrahymena. A.M. Elliott, editor. Dowden, Hutchinson, and Ross, Stroutsburg, PA. 375-

409.

Gaertig, J., and F. Iftode. 1989. Rearrangement of the cytoskeleton and nuclear transfer in

Tetrahymena thermophila cells fused by electric field. J.Cell Sci. 93:691-703.

Gorovsky, M.A., K. Carlson, and J.L. Rosenbaum. 1970. Simple method for quantitative

densitometry of polyacrylamide gels using fast green. Anal. Biochem. 35:359-370.

Grimes, G.W., and R.H. Gavin. 1987. Ciliary protein conservation during devleopment in the

ciliated protozoan, Oxytricha. J.Cell Biol. 105:2855-2859.

Iftode, F., J. Cohen, F. Ruiz, A.T. Rueda, L. Chen-Shan, A. Adoutte, and J. Beisson. 1989.

Development of surface pattern during division in Paramecium.

1. Mapping of duplication and reorganization of cortical cytoskeletal structures in the wild type.

Development. 105:191-211.

125

Jerka-Dziadosz, M. 1981. Patterning of ciliary structures in janus mutant of Tetrahymena with

mirror image cortical duplications. An ultrastructural study. Acta Protozool. 20:337-356.

Jerka-Dziadosz, M., L.M. Jenkins, E.M. Nelsen, N.E. Williams, R. Jaeckel-Williams, and J.

Frankel. 1995. Cellular polarity in ciliates: persistence of global polarity in a disorganized

mutant of Tetrahymena thermophila that disrupts cytoskeletal organization. Dev. Biol.

169:644-61.

Jerka-Dziadosz, M., I. Strzyewska-Jowko, U. Wojsa-Lugowska, W. Krawczynska, and A.

Krzywicka. 2001. The dynamics of filamentous structures in the apical band, oral

crescent, fission line and the postoral meridional filament in Tetrahymena thermophila

revealed by the monoclonal antibody 12G9. Protist. 152:53-67.

Johnson, K.A. 1998. The axonemal microtubules of the Chlamydomonas flagellum differ in

tubulin isoform content. J. Cell Sci. 111:313-320.

Kaczanowski, A. 1978. Gradients of proliferation of ciliary basal bodies and the determination of

the position of the oral primordium in Tetrahymena. J. Exp. Zool. 204:417-430.

Mahjoub, M.R., B. Montpetit, L. Zhao, R.J. Finst, B. Goh, A.C. Kim, and L.M. Quarmby. 2002.

The FA2 gene of Chlamydomonas encodes a NIMA family kinase with roles in cell cycle

progression and microtubule severing during deflagellation. J Cell Sci. 115:1759-68.

Marshall, W.F., and J.L. Rosenbaum. 2001. Intraflagellar transport balances continuous turnover

of outer doublet microtubules: implications for flagellar length control. J Cell Biol.

155:In Press.

McKean, P.G., A. Baines, S. Vaughan, and K. Gull. 2003. Gamma-tubulin functions in the

nucleation of a discrete subset of microtubules in the eukaryotic flagellum. Curr Biol.

13:598-602.

126

Morris, R.L., and J.M. Scholey. 1997. Heterotrimeric kinesin-II is required for the assembly of

motile 9+2 ciliary axonemes on sea urchin embryos. J. Cell Biol. 138:1009-1022.

Nanney, D.L. 1975. Patterns of basal body addition in ciliary rows in Tetrahymena. J. Cell Biol.

65:503-512.

Nelsen, E.M. 1975. regulation of tubulin during ciliary regeneration of nongrowing

Tetrahymena. Exp. Cell Res. 94:152-158.

Orias, E., and L. Rasmussen. 1976. Dual capacity for nutrient uptake in Tetrahymena. IV. growth

without food vacuoles. Exp. Cell Res. 102:127-137.

Pechart, I., M.L. Kann, N. Levilliers, M.H. Bre, and J.P. Fouquet. 1999. Composition and

organization of tubulin isoforms reveals a variety of axonemal models. Biol. Cell.

91:685-97.

Redeker, V., N. Levilliers, J.-M. Schmitter, J.-P. Le Caer, J. Rossier, A. Adoutte, and M.-H. Bré.

1994. Polyglycylation of tubulin: a post-translational modification in axonemal

microtubules. Science. 266:1688-1691.

Rosenbaum, J.L. 2000. Cytoskeleton: functions for tubulin modifications at last. Current Biol.

10:R801-R803.

Rosenbaum, J.L., and F.M. Child. 1967. Flagellar regeneration in protozoan flagellates. J. Cell

Biol. 34:345-364.

Rosenbaum, J.L., and G.B. Witman. 2002. Intraflagellar transport. Nat Rev Mol Cell Biol. 3:813-

25.

Sale, W.S., and P. Satir. 1977. The termination of the central microtubules from the cilia of

Tetrahymena pyriformis. Cell Biol Inter. 1:45-49.

127

Sasse, R., and K. Gull. 1988. Tubulin post-translational modifications and the construction of

microtubular organelles in Trypanosoma brucei. J. Cell Sci. 90:577-589.

Shang, Y., X. Song, J. Bowen, R. Corstanje, Y. Gao, J. Gaertig, and M.A. Gorovsky. 2002. A

robust inducible-repressible promoter greatly facilitates gene knockouts, conditional

expression, and overexpression of homologous and heterologous genes in Tetrahymena

thermophila. Proc Natl Acad Sci. 99:3734-3739.

Song, L., and W. Dentler. 2001. Flagellar protein dynamics in Chlamydomonas. J. Biol. Chem.

276:29754-29763.

Stephens, R.E. 1997. Synthesis and turnover of embryonic sea urchin ciliary proteins during

selective inhibition of tubulin synthesis and assembly. Mol. Biol. Cell. 8:2187-2198.

Sundaraman, V., and E.D. Hanson. 1976. Longitudinal microtubules and their functions during

asexual reproduction in Paramecium tetraurelia. Genet.Res.Cambr. 27:205-211.

Tartar, V. 1966. Synchronization of oral primordia in Stentor coeruleus. J Exp Zool. 161:53-62.

Thazhath, R., C. Liu, and J. Gaertig. 2002. Polyglycylation domain of beta-tubulin maintains

axonemal architecture and affects cytokinesis in Tetrahymena. Nature Cell Biol. 4:256-

259.

Vinh, J., J.I. Langridge, M.-H. Bre, N. Levilliers, V. Redeker, D. Loyaux, and J. Rossier. 1999.

Structural characterization by tandem spectroscopy of the posttranslational modifications

of tubulin. Biochemistry. 38:3133-3139.

Westermann, S., A. Schneider, E.K. Horn, and K. Weber. 1999. Isolation of tubulin

polyglutamylase from Crithidia; binding to microtubules and tubulin, and glutamylation

of mammalian brain alpha- and beta-tubulins. J. Cell Sci. 112:2185-2193.

128

Xia, L., B. Hai, Y. Gao, D. Burnette, R. Thazhath, J. Duan, M.-H. Bré, N. Levilliers, M.A.

Gorovsky, and J. Gaertig. 2000. Polyglycylation of tubulin is essential and affects cell

motility and division in Tetrahymena thermophila. J.Cell Biol. 149:1097-1106.

129

Figure Legends:

Fig. 4.1 (A) A schematic representation of MT-based organelles in Tetrahymena. The

intracytoplasmic, macronuclear, oral fiber and cytoproct MTs are not shown. An interphase (left)

and dividing (right) cell are shown. The inset shows a magnification of a region of the cortex

surrounding 2 adjacent BBs. In the dividing cell, we distinguish between preformed (black) and

newly made organelles (grey). We also use grey color to depict preformed organelles with a

relatively high rate of subunit exchange even though some of them are maintaining structural

integrity (LMs and CVPs). Note the presence of an anterior and posterior invariant zone in the

dividing cell. Abbreviations: ss, separation spindle (anaphase B) of the micronucleus; mic,

micronucleus; mac, macronucleus; cvps, contractile vacuole pores; ncvps, new contractile

vacuole pores, m, oral membranelles; um, undulating membrane; oa, oral apparatus; noa, new

oral apparatus; lm, longitudinal MT; tm, transverse MT, kd, kinetodesmal fiber

(nonmicrotubular); pc, postciliary MT; bb, basal body. (B) Growth of βDDDE440-CD cells in the

presence (open squares) and absence of Cd (solid squares).

Fig. 4.2: Tubulin flow in βDDDE440-CD cells after Cd removal. Cells were stained with a

combination of anti-HA (green) and anti-tubulin antibodies, SG (red) at 0 (A), 13 (B), 15 (C) and

22 (D) hrs after Cd removal. Structures with higher green signal represent “older” organelles. A:

A cell grown for about 2 weeks in Cd, showing uniform distribution of WT-HA and total tubulin.

B and C: Note older cilia showing uniform HA label (arrowhead) and cilia with lower level of

HA tubulin that were formed sometime after Cd removal (asterisk). Also, in C note regions

devoid of cilia (arrow). D: Note the presence of a posterior invariant zone in D (p). Arrowheads

130

show old cilia with a distal tip made up mainly of mutant tubulin. Arrows indicate new and short

cilia that are made primarily of mutant tubulin. Scale Bar = 8µm

Fig 4.3: Tubulin flow in βDDDE440-CD cells after Cd removal (part of the same series as in Fig.

2). Cells were stained at 22 (A), 44 (B), and 80 (C) hrs after Cd removal. Note the presence of an

anterior invariant zone in A (a) associated with conserved cilia of membranelles of the anterior

oral apparatus (oa). Arrowheads show old cilia with a distal tip made up mainly of mutant

tubulin. Arrows indicate new and short cilia that are made primarily of mutant tubulin. B: A cell

composed of 3 subcells. An anterior end of another cell is seen on the left side. Note

conservation of old cilia in the anterior and posterior invariant zones of the chain. Cilia made of

primarily mutant tubulin are much shorter compared to older cilia retaining WT-HA tubulin.

Tubulin subunit exchange occurs at tips of older cilia (arrowhead). C: A 4 subunit cell. Note

nearly complete absence of HA label indicating nearly complete exchange with mutant tubulin.

A few old cilia are present at the anterior end and the HA label of old tubulin is distributed

uniformly (inset in both anti-HA and merged panels). Scale Bar = 8µm

Fig. 4.4: Tubulin exchange in WT-CD cells. Cells in panels A-E were labeled by

immunofluorescence using 12G10 monoclonal antibodies (green) against total α-tubulin and

anti-HA tag antibodies (red). After Cd washout, structures with stronger red signal represent

“old” organelles. The cells were analyzed at 12,5 (A), 20.5 (B-C), 25.5 (D), 62.5 (E), and 71.5

(F) hrs after Cd removal. A: A dividing cell. Note the presence of old cilia (arrowhead)

assembled prior to Cd removal. The asterisk indicates a cilium formed after Cd removal, which

has a lower HA tagged tubulin content. The newly formed OA (noa) and the more dynamic

131

structures such as the mitotic micronuclear separation spindle (ss) are less strongly labeled with

the anti-HA antibodies (see middle panel). The contractile vacuole pores (cvp) already show a

lower intensity of HA labeling. B and C: Note the presence of an anterior invariant zone (a) in B

and the posterior invariant zone in C (p). D: Arrowheads indicate cilia with reduced HA labeling

at their tips indicating tubulin exchange. E: Almost complete dilution of old units due to clonal

growth has occurred with the exception of one old cilium. F: Cells labeled at 71.5h with

monoclonal anti-HA (green) and polyclonal anti-total tubulin (red) antibodies. This image was

chosen to emphasize that in addition to the dilution effect, older static cilia do undergo subunit

exchange along their entire length. Arrowhead in inset points to a cilium that still labels with the

anti-HA antibody along the length of the cilium, although weakly. For comparison, the arrow in

the inset points to a cilium that does not label with the anti-HA antibody. Scale bar = 8µm.

Fig. 4.5: TEM analysis of the βDDDE440 mutant cortex. A: A longitudinal section showing

dense packing of BBs. B: An LM bundle showing profiles of 14 MTs. C and D: Cross-sections

through the BBs. Arrows point to abnormal MTs associated with the triplets. E: Cross-section of

an OA. The membranelles of the AZM show overgrown PC ribbons (arrows) in the first and

second row. The AZM rows are labeled 1, 2 and 3 respectively.

Fig. 4.6: TEM analysis of cilia in the βDDDE440 mutants. A: A longitudinal section of the cortex

showing a delay in the ciliation of new BBs. B: A section of the newly formed OA of the

posterior mutant subcell at a 2-subcell stage. Virtually all newly formed cilia lack CP. C: A

section through a short mutant cilium. Note the bulbous tip and a lack of attachment of the outer

doublets to the ciliary membrane. D: A longitudinal section through the apical region of an

132

anterior mutant subcell at a 2-cell stage. The cilia shown belong to the anterior invariant zone

and appear normal in every respect. E and F: Cross-sections through mutant cilia showing

defective axonemal structures lacking the CP and conversion of outer doublets into singlets

(asterisks).

Fig. 4.7: Tubulin exchange in the cortex of βDDDE440-CD and WT-CD cells.

Cells were partly deciliated to enable visualization of cortical MTs. The WT-CD cell in panel A

was labeled with anti-HA (red) and 12G10 anti-α-tubulin antibodies (green). The βDDDE440-CD

cells in panels B and C were labeled with anti-HA (red) and 6-11-B-1 anti acetylated α-tubulin

(green). A: A WT-CD cell 12.5h after Cd removal. Note scattered old cortical units indicated by

stronger anti-HA labeling. Inset- Tubulin exchange in TMs occurs at the distal tips. B and C:

βDDDE440-CD cells labeled at 12 h and 22 h respectively. Note the cluster of older cortical units

in the anterior (in B) and in the posterior region (in C). The inset in C shows TMs exchanging

subunits at distal tips as in A.

Fig. 4.8: Dissociation of the 13C4p MAP from cortical MTs in the βDDDE440 mutants.

Cells labeled with 13C4 antibody (A-E) and 12G10 anti α-tubulin antibodies (F-I).

A: A WT dividing cell showing a uniform localization of 13C4p along all cortical MTs.

B-E: βDDDE440 mutants showing a gradient of 13C4p MAP localization between anterior and

posterior subcell (labeled p). Mutants are at 24 h (B) and 48h (C) after induction of phenotype. D

and E show an increase in granular labeling in the cell body of the anterior subcell at 48h and

72h. Panels F-I: βDDDE440 mutants labeled at similar stages of phenotype development as in B-E

with 12G10 (anti α-tubulin). Note that despite dissociation of the 13C4p MAP, cortical MTs

133

appear to be intact. In panels D, E, and I, originally distinct subcells have undergone cortical

integration.

Fig. 4.9: A-C: A video imaging analysis of a single βDDDE440 mutant as it develops the

phenotype; p, posterior most subcell.

A: A mutant cell with one arrested cleavage furrow giving rise to two subcells.

B. The same cell at 37 hrs; the posterior subcell has undergone cytokinesis, giving rise to a chain

of 3 subcells. Note that together, the two posterior cells which are derived from the posterior cell

in A are larger than the anterior cell and the most posterior subcell is bigger than the middle cell,

indicating that the posterior cell consistently grows faster. C. At later time the division of the

most posterior subcell leads to a chain made of 4 subcells and 3 cleavage furrows. The numbers

in A-C indicate successive cleavage furrows. D: Oral replacement in the βDDDE440 mutants. The

βDDDE440 mutant cell is labeled with the 12G9 antibody, which recognizes an unknown cortical

protein. The posterior most subcell is toward the bottom right of the panel. op, oral replacement

primordium; oa, old oral apparatus. Note that the op in the posterior most subcell is more

advanced while the op in the anterior most subcell is still in the initial stages of formation. E: A 3

subcell chain showing a larger posterior subcell in which the micronuclei (mi) are in the

configuration characteristic of late anaphase of the mitotic division. The macronucleus (ma) of

the most posterior subcell is also much larger than those in the 2 more anterior subcells.

134

Fig. 4.10: A: A composite of the cell division pathways observed in a number of isolated mutant

cells. The majority of mutant cells show the posterior dominance phenotype, in which at any

given stage the most posterior subcell grew and divided. B: A summary of the most prevalent

mode (posterior dominance) of cell division in the mutants.

145

CHAPTER 5

CONCLUSIONS

The focus of my research was to identify the function(s) of tubulin glycylation (TuG) in

vivo. The fact that the sites of modification on both α and β tubulin proteins are evolutionarily

conserved, compounded by the remarkable conservation of the modification across diverse

species indicated its important role. Before I initiated my studies, it was known that the

modification was essential for survival of Tetrahymena, and there were indications of a possible

role in ciliary motility, growth and cell division based on the phenotype of a hypomorphic

glycylation-deficient mutant (βEDDD440, (Xia et al., 2000). The presence of TuG almost

exclusively in ciliated and flagellated cells of diverse species initially led investigators to believe

that the modification had a role in the axoneme function or assembly. However, findings by Xia

et al., suggested that ciliates could be using the modification for non-ciliary functions as well.

The exact mode of action through which the modification affects the ciliary or non-ciliary MTs

was unknown. My studies have attempted to shed light on the mechanism of action of tubulin

glycylation in vivo. Specifically, my initial goal was to explain why a triple mutation of the

glycylation sites, βDDDE440, was lethal.

The knockout heterokaryon approach (Hai et al., 1999) has been used with great success

in the study of various gene products in Tetrahymena thermophila. The heterokaryon method can

be used in two ways. One can study the phenotype of cells lacking a particular gene by direct

analysis of the progenies of mating heterokaryons. Furthermore, it is possible to establish

whether a particular mutated version of the gene of interest is functional by attempting a rescue

of mating heterokaryons having a disruption at that locus. The rescue approach is quick but not

146

very informative in the case of lethal mutations. Furthermore, the rescue transformants require

drug selection (usually paramomycin) to eliminate nonconjugants. The use of selective drugs

could therefore modify the phenotypic outcome. It should be mentioned that mutations in other

genes are known to modify the sensitivity to paramomycin. For example, strains lacking genes

required for assembly of cilia (such as IFT52) are paramomycin-sensitive despite the presence of

the neo marker gene, which confers resistance in normal cells (J. Gaertig, unpublished

observations). Thus, lack of transformants using the triple glycylation site mutant gene in the

rescue approach (Xia et al., 2000) could be explained by either lethality conferred by the

transformed gene or by lethality of the selection method, or a synergistic effect of both. To

develop a better tool for analysis of potential lethal phenotypes, I modified the knockout

heterokaryon strategy to create mutant (non-knockout) heterokaryons with a linked marker gene

(Chapter 3). The mutant heterokaryon strains carried a potentially lethal mutation (βDDDE440) in

the coding region of β-tubulin gene, BTU1, in the MIC and WT alleles in the MAC. The

glutamic acids at positions 437-439 were mutated to aspartates, which from previous studies,

were known to be unmodifiable by the tubulin glycylases in vivo. This approach allowed us to

maintain cells with a WT phenotype as long as they were grown vegetatively. To reveal the

mutant phenotype, mutant heterokaryon strains were allowed to conjugate, and progenies

analyzed cytologically. This method allowed us to eliminate the effects of drug selection used in

knockout rescue transformations.

The initial phenotypic changes that were observed in the progenies of βDDDE440 mutant

heterokaryons included a loss of motility, and a partial block in cytokinesis giving rise to chains

of subcells. Eventually, mutant cells died within 7 days. The direct cause of death is not clear.

The first possibility that I investigated was that mutant cells died simply due to inability to take

147

up food via phagocytosis. This principal route of feeding is restricted to the oral apparatus in

Tetrahymena, and requires motility of oral cilia. Since cilia are strongly affected in the

βDDDE440 mutant, cell death could result from lack of beating of oral membranelles, and

therefore could be indirect. However, even cilia-less mutants obtained previously in our lab

could be grown on an enriched medium, MEPP (Brown et al., 1999b). MEPP is believed to

stimulate micropinocytosis as an alternative food uptake pathway and allows for growth of cells

lacking entirely the oral apparatus (Rasmussen and Orias, 1975). To investigate further the

possibility that βDDDE440 mutant cells died of starvation, I isolated and grew the progenies of

βDDDE440 mutant heterokaryons in MEPP and compared the growth to that of cilia-less kinesin-

II mutants, which were known from previous studies to survive in enriched MEPP media (Fig.

5.1). However, the βDDDE440TuG domain mutants were unable to grow in the MEPP medium,

suggesting a reason other than starvation for lethality. It was also a possibility that the βDDDE440

mutants died due to lack of proper folding of tubulin caused by the alteration of the primary

sequence of the COOH terminal tail of β-tubulin, that could result in lack of proper dimerization

and incorporation into MTs. To investigate this possibility, we compared the growth of the

βDDDE440 mutants to progeny of knockout heterokaryons, which lacked the major β-tubulin

entirely (Xia et al., 2000). I found that the progenies of these β-tubulin knockout heterokaryons

died without dividing, within 1-2 days after conjugation. In contrast, the βDDDE440 mutants were

able to complete a few rounds of normal cell divisions before arresting in cytokinesis and losing

motility (Fig. 5.2). Thus, we can safely conclude that the lethality of βDDDE440 was not a result

of a lack of functional β-tubulin or a α/β dimer. It is more likely that cell death resulted from the

lack of proper cell organization and intracellular transport in extremely disorganized

multinucleated cells of βDDDE440.

148

Immunofluorescence analyses in Tetrahymena showed the presence of polyglycylation in

the ciliary and cortical MTs, while monoglycylation was localized to virtually all types of MTs

(ciliary, cortical, cytoplasmic, macro- and micronuclear) (Chapter 3). Remarkably, we saw a

very selective effect of the βDDDE440 mutation on specific types of MT organelles. The mutation

affected MT organelles that are known to be polyglycylated, more strongly. On the other side,

the organelle types known to have exclusively monoglycylated tubulins, like micro- and macro-

nuclear MTs, appeared unaffected. Furthermore, the ciliary MTs were more affected compared

to the cortical MTs, which correlates with longer length of polyglycine chain in cilia compared to

cell body of Paramecium, shown by mass spectrometry (Vinh et al., 1999).

The lack of effect of the βDDDE440 mutation on MT systems, which are only

monoglycylated, can be explained in two ways. It is possible that monoglycylation is not

important per se and represents only a step in maturation of MTs required for future

polyglycylation. In other words, in Tetrahymena, all tubulin could be subjected to

monoglycylation shortly after dimerization, so that a sub-pool of tubulin could be further

modified in selected organelles to achieve polyglycylation. From a biochemical standpoint, it

would make sense if two different enzymes were involved, one of them catalyzing the addition

of the first glycine, and another one responsible for subsequent elongation and resulting in

formation of polyglycine. However, mass spectrometric analysis of ciliary tubulins showed a

significant pool of tubulins which lack any modifications of the COOH terminal tail (Redeker,

Bre, and Gaertig, unpublished data). Thus, not all tubulin undergoes monoglycylation, even in

organelles, which have the highest level of polyglycylation. The presence of unmodified tubulin

in cilia indicates that monoglycylation is not a mere preparation of all tubulin for potential

polyglycylation and that monoglycylation may have its own function. Furthermore, it is possible

149

that functionally significant monoglycylation occurs on additional E residues, which were not

mutated in the βDDDE440 mutants (there are 6 non-mutated Es in the tail domain of βDDDE440).

It should be mentioned that recently Krzysztof Rogowski, a graduate student in our laboratory,

has made a β-tubulin heterokaryon strain in which all 9 Es of the C-terminal tail domain were

replaced by non-modifiable Ds. In addition to a somewhat stronger version of the βDDDE440

phenotype, these β9D mutants are also defective in nuclear division and positioning. These

results indicate that either monoglycylation or glutamylation on nuclear MTs has an important

function and that one or both of these PTMs occur on the 6 glutamic acid residues of the tail

domain of β-tubulin that are not mutated in βDDDE440.

The axoneme was most severely affected by the βDDDE440 mutation. Specifically, the

CP was not present on many cilia and the B-tubule was absent from the outer doublets. The A-

tubules and associated dynein arms seemed normal. The question arises as to what is the basis of

the selective effect of the βDDDE440 mutation on the CP and B-tubules. I should first explain that

the effects of lack of proper glycylation sites on β-tubulin in cilia are both MT subtype-specific

and organelle-wide. On one side, the CP and B-tubules were more affected compared to the A-

tubules of outer doublets. On the other side, the entire cilium made of mutant tubulin was

shorter.

The important question is to what extent the observed MT type specific phenotypes (such

as lack of CP) correlate with the intra-organellar distribution of the modification. Previous

studies on the localization of polyglycylation using immunogold labeling and TEM showed that

in sea urchin axonemes, polyglycylation preferentially localizes to the B-tubule of the outer

doublets and the CP MTs (Multigner et al., 1996). However, it appears that there is a great

variation in the labeling pattern (with the same antibodies) across different species. For example,

150

a similar study in the related ciliate Paramecium showed that polyglycylation was equally

present in both tubules of the outer doublets and the CP while monoglycylation was excluded

from the CP with no variation between the A and B-tubules of the outer doublets (Pechart et al.,

1999). Although similar studies have not been reported for Tetrahymena, it is likely that the

distribution of modifications is similar to that seen in the closely related ciliate Paramecium.

Thus, it appears that there is no correlation between the amounts of glycylation on individual

MTs and their dependence on the modification sites.

If we assume that both CP and doublets are equally strongly polyglycylated, we need

another explanation for the stronger effect of the same mutation of the glycylation domain on CP

compared to the doublets. First, we need to consider that glycylation affects IFT, the motility

process, which is critical for delivery of ciliary subunits during the process of assembly. The rate

of IFT could be affected, and the CP could be more dependent on this process than the outer

doublets. IFT occurs on the surface of doublet MTs. We also know that the COOH terminal tail

domains strongly interact with kinesin during their motility in vitro on isolated MTs (Wang and

Sheetz, 2000). Thus, it is possible that glycylation of doublet MTs affects the interactions with

kinesin-II (the principal IFT anterograde motor) during IFT. However, the IFT hypothesis would

also need to assume that CP is more sensitive to the deficiency in IFT compared to doublets.

Importantly, the IFT mutants in Tetrahymena (lacking either kinesin-II or its cargo particles)

occasionally assemble cilia, which are always very short and lack CP (Brown et al., 2003). Thus,

the ciliary phenotype of IFT mutants resembles the phenotype of the tubulin glycylation site

mutant studied here. Thus, the experimental evidence indicating increased dependence of CP

assembly on IFT is already available. It is possible that the doublet MTs can assemble slowly

without IFT, based on diffusion of tubulin dimers into the ciliary space and that doublets require

151

lower critical concentration of tubulin dimers for assembly compared to CP due to their

increased internal stability.

The second possible explanation of the differential effect of the mutation on CP lies in

the different mode of CP nucleation. CP assembles out of the plaque structure in the transition

zone, while doublets grow by extension of the A and B-tubule of the triplets of the adjacent BB.

Thus, the assembly of CP requires formation of new MTs (nucleation), while the doublets grow

simply by extension of preexisting MTs of the BB. Furthermore, it is already known that

deficiency in γ-tubulin affects selectively CP in axonemes (McKean et al., 2003). Thus, we can

hypothesize that tubulin glycylation is specifically involved in nucleation of the CP MTs and that

it may interact with γ-tubulin. At any rate, this potential interaction with γ-tubulin during

nucleation would need to be specific to the cilium, because the same mutation of the glycylation

sites did not affect the assembly of nuclear MTs, which are dependent on the same γ-tubulin in

Tetrahymena (Shang et al., 2002a).

The third possibility is that TuG is involved in the stabilization of the tips of the

axonemal MTs through the capping structures, which connect them to the plasma membrane of

the cilium (Sale and Satir, 1977). It is possible that the modification enables acquisition of the

cap forming proteins at the tips of cilia during an early stage of assembly. An absence of the

stabilizing caps could lead to a lack of CP assembly or instability in a short time after assembly.

Furthermore, it is possible that the CP and its associated cap play a role in determining the length

of the axoneme. The outer doublets may still polymerize without proper capping owing to their

potentially higher intrinsic stability. Recently Pedersen and Rosenbaum showed that a conserved

plus-end MT binding protein, EB1, is present at the tips of flagella in Chlamydomonas (Pedersen

et al., 2003). It is therefore possible that EB1 is a part of the capping complexes. However, EB1

152

does not appear to play a role in stabilizing the plus ends of MTs. Studies in cultured mammalian

cells shows the presence of this protein at the plus ends of growing cytoplasmic MTs, and is lost

on catastrophe. Cytoplasmic MTs are however, the more labile population of MTs as compared

to ciliary MTs, thus EB1 could be performing different functions in the two subsets of MTs. An

EB1 homolog was located in the Tetrahymena genome (D. Wloga, personal communication). It

will be important to determine whether EB1 is present at ends of ciliary microtubules in the

βDDDE440 mutants. If the capping hypothesis is correct, we would also need to explain why the

modification is present along the entire axoneme and not just at the tips of MTs. We will

postulate later (see below) that the modification could have multiple functions, depending on its

location, even within the same organelle. For example, it is intriguing that polyglycylation and

polyglutamylation occur in gradients along the distal-proximal length of cilia, with the highest

level near the base (and possibly with longer side chains). For example, in Paramecium, a

proximo-distal decreasing gradient of labeling (using modification-specific antibodies) was

observed from the transition zone out toward the tip of the axoneme for polyglycylation (Pechart

et al., 1999). Monoglycylation showed a reverse proximo-distal gradient as compared to

polyglycylation (Pechart et al., 1999). The existence of such gradients can be simply explained

by rapid deposition of monoglycylation on newer segments of MTs (near the plus ends of the

distal tips) and gradual replacement by polyglycylation on older segments of the axoneme

(toward the minus ends near the BB). It is possible that some parameters of the modification,

such as the length of the side chain, change gradually along the structure, providing different

properties in highly restricted localizations (such as only at the tip). Furthermore, I will argue

(see below) that the polymodifications act primarily by regulating the binding of MAPs. Thus,

153

the modification could have a unique action at the tip, as long as appropriate MAP(s) (such as the

capping complexes) are also restricted to the tip.

Another prominent phenotype observed in the βDDDE440 mutants was a block in

cytokinesis, with mutants undergoing several rounds of arrested divisions giving rise to chains of

mutant subcells. There were several possible reasons why the mutants were unable to divide

completely. First, we again need to consider that the cytokinesis phenotype is an indirect

consequence of lack of proper ciliary motility. This is because at the end of cytokinesis,

Tetrahymena undergoes the process of rotokinesis, which occurs at the final stage of cell division

where the two daughter cells are connected only by a narrow cytoplasmic bridge (Brown et al.,

1999a). In order to break this connection, the cells use a cilia motility-based mechanism where,

the posterior daughter undergoes a series of unidirectional rotations. Due to loss of ciliary

motility, rotokinesis was likely impeded in the βDDDE440 mutants. However, it appears that the

block in cytokinesis in the βDDDE440 mutant chains occurs at a much earlier stage, when a

relatively wide cytoplasmic connection is still present and the daughter cells normally do not

rotate. The behavior of subcells in the chains indicates that there is substantial cytoplasmic and

cortical continuity between the adjacent cell units in the βDDDE440 mutant. Jason Brown, a

former graduate student in our lab, has extensively studied the multinucleated chains of subcells

formed solely due to lack of rotokinesis in ciliary assembly (IFT) mutants. He found that

following a block in rotokinesis, the subcells of IFT mutants quickly integrated into

multinucleated “monsters”, due to rapid sliding of individual units in the arrested chains.

However, in the βDDDE440 mutants the individual subcells remain distinct (nonintegrated) for a

relatively long time. Furthermore, the reason why subcells of the βDDDE440 mutants integrate

slowly appears to be that some ciliary rows are continuous between the adjacent subcells, due to

154

lack of severing and absence of cortical subdivision. In contrast, the ciliary rows of subcells of

IFT mutants are well separated by the cortical subdivision, which develops at the time of

cytokinesis (Brown et al., 2003; Brown et al., 1999b).

Ultrastructural analysis revealed that certain microtubular organelles in the mutant cortex

were hypertrophic. It appeared that the density of BBs in the ciliary rows was increased

compared to wildtype. Interestingly, BBs and LM MTs contained an increased number of

individual microtubules than are normally present. As shown in Chapter 4, BBs showed the

presence additional MTs attached to the triplets. Furthermore, BBs present in the membranelles

of the oral apparatus, showed associated hypertrophic PC MT bundles. Thus, it is entirely

possible that the hypertrophic features of the cortex, impede the cleavage furrow progression,

leading to the block in cytokinesis in the βDDDE440 mutants. The other possibility is that the lack

of a proper cortical subdivision was an effect and not the cause of the arrested cytokinesis

phenotype. However, as argued in Chapter 4, in the cilia-less IFT mutants arrested in cytokinesis,

LMs undergo proper severing, suggesting that they are not hypertrophic. Thus, at least the

severing phenotype appears to be an effect of the TuG mutation and not an indirect consequence

of lack of cytokinesis. It remains to be determined whether the increased number of MTs in the

LM bundles is the cause or the result of lack of severing. It is possible that TuG determines the

site of severing and the increased number of MTs is the consequence of lack of severing at a

proper time during the cell cycle. Alternatively, TuG could play a role in limiting polymerization

of cortical MTs and lack of severing could be caused by excessive number of MTs.

An important part of this project was to study the temporal aspects of expression of the

glycylation deficiency phenotype. Specifically, I examined the timing of appearance of axonemal

defects in cilia of mutant cells. Using the original mutant βDDDE440 heterokaryons, I discovered

155

that cilia in mutant cells often lacked CP. However, using my initial approach I could not

determine whether the CP was initially assembled and then lost or whether there was a lack of

assembly altogether. This is an important distinction. In one scenario, glycylation would be

involved in the assembly machinery, including transport of tubulin, its ordered addition to ends

of axonemal MTs and association with dynein arms. In the second scenario, the assembly

mechanism could be functional but the structure may be unstable after assembly due to lack of

proper capping, association with stabilizing MAPs etc. Another question related to timing of

expression is about the origin of the ciliary length defect. Is the short size of cilia a result of slow

assembly or lack of proper size regulation during the post-assembly maintenance? It was

apparent to us that we needed to develop a more advanced experimental system to achieve better

spatial and temporal resolution of the sequence of events that occurred during the progression of

the phenotype.

For this purpose, I constructed a novel inducible system, wherein the phenotype of the

cell could be switched from wildtype to mutant and vice versa, depending on the presence or

absence of cadmium ions in the growth media. The MTT1 gene is a non- essential

metallothionein gene, whose product binds heavy metals. The MTT1 promoter is strongly

induced in response to cadmium uptake and this feature was used in constructing the first

inducible gene expression system in Tetrahymena (Shang et al., 2002b). As described in Chapter

4, I created a strain that had the βDDDE440 mutant gene in the endogenous BTU1 locus, and an

additional WT BTU1 coding region in the MTT1 locus under control of the MTT1 promoter. The

WT BTU1 coding region in the MTT1 locus encoded an HA epitope tag at the COOH terminal

end, which we used to distinguish between the WT and mutant tubulin in the same cells.

Addition of the tag did not change the phenotype of WT cells, indicating that the tag itself does

156

not have significant side effects. When cadmium was added to the medium containing mutant

cells, expression of WT BTU1-HA was induced, and the cells maintained a WT phenotype due

to the recessive nature of the βDDDE440 mutation. When cells were washed free of cadmium, the

expression of the WT BTU1-HA was downregulated, and mainly the mutant βDDDE440 tubulin

was produced, and consequently the cells started developing the βDDDE440 phenotype. There

was a lag period before the phenotype developed owing to the process of dilution of already

present WT-BTU1-HA gene products in the mutant cell.

The phenotype of the conditional βDDDE440 mutant was not as severe in the

manifestation of the cytokinesis defects as compared to mating heterokaryons described in

Chapter 2. A large population only underwent a single arrested round giving rise to two subcell

mutants as apposed to the multiple subcell chains formed using mating heterokaryons. Only

some conditional mutants did undergo 2-3 arrested rounds of division giving rise to 3-4 subcell

chains. This could be attributed to the apparent leakiness of the MTT1 promoter, observed in the

previous studies (Brown et al., 2003). There appears to be a basal level of expression from the

MTT1 locus even in the absence of cadmium. However, the expression of the ciliary phenotype

in the conditional mutants was almost identical to what we observed earlier, with short CP-less

cilia present all over the surface of the cell. There could be different phenotypic thresholds for

the contribution of mutant tubulin in a mixture with WT tubulin between the two subsets of MTs.

A lower content of mutant tubulin in the mixture could be required to achieve the phenotypic

effect in the cilia as compared to the cortex. Raff et al tested the ability of BB specific β1 tubulin

to function in axonemes in place of the β2-tubulin, which is axoneme specific in Drosophila

(Raff et al., 2000). They found that β1 by itself could not assemble axonemes. When they tested

the ability of a mixture of the two tubulin isotypes to generate functional axonemes, they found

157

that at equimolar ratios, the mixture of β1 and β2 could generate fully functional sperm

axonemes. However, when the ratio of β1 to β2 was increased to 2:1, the axonemes were a

mixture of normal 9+2, and abnormal 9+0 and 10 doublet axonemes. Thus, it appears that there

is a certain ratio of inappropriate β1 to proper β1 that can be tolerated during spermatogenesis,

beyond which the small isotype specific differences in the COOH terminal sequence affect the

overall structure of the organelle. Our study gave us a tremendous amount of detail concerning

the actual spatial and temporal incorporation of tubulin in vivo, and allowed us to visualize the

process using epitope tagging.

Using this inducible system, we were able to determine in a definitive manner that mutant

tubulin was incorporated into newly assembled structures of several types and was capable of

normal targeting and exchange in the preexisting structures. This study was the first attempt to

reconstruct the pattern of tubulin flow in assembled and maintained organelles in Tetrahymena.

Although the main focus of my work was to evaluate the properties of the glycylation site

deficient tubulin, as a control, I did tubulin flow studies in the WT background. Therefore, as a

byproduct, I also provided the first description of tubulin flow pattern for normal Tetrahymena

cells. For WT, as expected, the intracytoplasmic network MTs were found to be most labile.

However, this result is not surprising because the intracytoplasmic MTs of Tetrahymena

resemble the cytosolic network of singlet MTs known from animal cells and like the former may

undergo cycles of dynamic instability. More surprising were my observations on the cortical

MTs. (BBs, PC, TM, LM, and CVP). These are considered as “stable” MTs because they do not

undergo any obvious changes in their shape and size following assembly. However, as explained

in Chapter 4, even highly morphostatic cilia are known to undergo slow tubulin subunit

158

exchange, which occurs without organelle disassembly. Thus, it was of high interest to determine

whether cortical MTs also undergo exchange (see below).

Another aspect of the tubulin flow study is related to the cell pattern and its influence on

the assembly and maintenance of organelles. Indeed, I found that the sites of assembly of new

cortical units, was strongly dependent on cell polarity. The insertion of new BBs and associated

cortical units was more prominent in the mid body and the posterior region of the cell. This

confirmed earlier studies on the patterns of BB duplication done by counting non-ciliated BBs

(Frankel et al., 1981; Kaczanowski, 1978; Nanney, 1975). However, in our study we could also

follow the old units over many generations and determine to what extent they exchange tubulin

subunits. We found that the non-uniform insertion of new BBs units throughout the cortex leads

to the formation of territories near the ends of the cell with largely old units, named invariant

zones. The presence of similar invariant zones has been well documented in another ciliate,

Paramecium tetraurelia in which the newly assembled BBs can be morphologically

distinguished from the old units more easily compared to Tetrahymena (Iftode et al., 1989). In

the invariant zones of Tetrahymena, detectable old BBs and their cilia persisted for about 13

generations. However, with time the old organelles underwent slow subunit exchange. We

obtained two important pieces of information regarding the process of tubulin exchange in

preformed organelles. First, the subunit exchange occurs in a polarized manner inside the

organelle. Specifically, both in cilia and cortical TMs, the exchange was initiated at the distal end

of MTs. This observation is consistent with the fact that plus ends of MTs are located at tips of

cilia. In the flagella of Chlamydomonas, subunit exchange was also seen at the plus ends (Song

and Dentler, 2001). However, my results shed light on the organization and polarity of the much

shorter non-ciliary MTs associated with the BB. Specifically, based on the pattern of subunit

159

exchange, I can postulate that the plus ends of the TM and PC bundles are distal from the BB.

With time, I saw clearly that the exchange expanded from the distal end and eventually involved

the entire structure, in the case of cilia. In the previous similar study in Chlamydomonas, the

subunit exchange could be visualized only for the distal 1/3 of the flagellum, due to experimental

system used (dikaryons) in which flagella underwent natural resorbtion (Marshall and

Rosenbaum, 2001). Here I was able to establish that although the subunit exchange starts at the

distal end, with time it involves the entire structure. The process is slow, and requires about 13

generations or over 80 hrs for nearly complete replacement of preexisting tubulins in cilia. In the

future, it will be important to determine, whether the exchange occurs by insertion along the

entire length of the axoneme or due to end activity (dynamic instability). Insertion of new

subunits into the already built MT lattice laterally, would appear to be unfavorable with respect

to the kinetics of the reaction. We would need to carry out further investigation into the process

of exchange of ciliary tubulins using higher resolution approaches such as TEM, and approaches

that can be used with living cells (GFP-tubulin)

Another observation we made in this study was that the rates of turnover were not the

same for all cortical organelles. Specifically, the turnover rate for TMs and PCs appeared to be

slow, with the rate resembling that in cilia. However, the MTs of LMs and CVP exchanged

subunits much more rapidly. Thus, the rates of subunit exchange are organelle-specific. The

question arises as to why the exchange rates are different in preformed cortical organelles of

different type. One possibility is that even though the MT organelles appears morphostatic after

assembly, in reality they could be undergoing rapid reorganization of microtubules by end

polymerization and depolymerization that are below the limit of detection. This could well be the

case of LM bundles, which consist of a number of partly overlapping MTs spanning the length of

160

the cell (Frankel, 1999). For example, in Paramecium the homologous structure known as

cytospindle, assembles only transiently during cell division (Iftode et al., 1989; Sundaraman and

hanson, 1976). Another possibility is that the observed higher turnover rate does not involve

depolymerization and repolymerization at the ends but rather, as it could be the case of cilia,

insertional removal and substitution of subunits along the entire length. This aspect could not be

discerned due to the small size of cortical organelles. Such more rapid exchange could be

important to the function of the CVP, which is involved in periodic evacuation of fluids from the

contractile vacuole system.

It is also possible that in some cases the slow exchange is related to structural

rejuvenation of organelles. The longer modification side chains could be an indication of “age”

of that region of the organelle may be subjected to tubulin subunit exchange. It is important to

note that our experiments only addressed the question of whether a mutated tubulin can replace a

preexisting wildtype tubulin in preformed organelle. However, in the future one would need to

address a reverse question, of whether a wildtype tubulin could replace a preexisting mutant

tubulin already built into the polymer. Such an experiment would directly assess the role of TuG

as a marker of aged tubulin for rejuvenation. Thus, one could use the conditional mutants to

produce axonemes made primarily of the mutant tubulin and by adding cadmium, reactivate

wildtype tubulin gene and follow the population of structures made of mutant tubulin to

determine whether they can be rejuvenated.

By comparing the pattern of flow between the cadmium-dependent βDDDE440 mutant and

an appropriate control strain, I was able to determine that the mutant tubulin deficient in

glycylation sites does not affect the pattern of insertion of new BB units. Furthermore, the

mutant tubulin was capable of targeting to all types of new organelles as well as to preexisting

161

organelles, and supported slow subunit exchange in preformed organelles at apparently normal

rates. Thus, the mutant tubulin does not appear to be defective in its ability to assemble and

rejuvenate existing organelles made of normal tubulin. However, when mainly mutant tubulin

was used to assemble cilia, the newly formed structures were incomplete and short. Furthermore,

shortening of preexisting cilia occurred with time as the mutant tubulin replaced the preexisting

WT tubulin. Thus, as discussed above, we determined that glycylation plays an important role

during assembly of cilia, but is also required for proper long-term maintenance of cilia. Thus,

even if the structure was initially assembled properly, gradual replacement of its tubulins with

mutant versions leads to a corresponding change in the structure. This observation suggests that

the overall size of the organelle may be a sum of highly localized interactions between individual

tubulin subunits along the length of the organelle.

My study has provided an abundant documentation of the consequences of deficiency in

the sites of β-tubulin glycylation on various types of MTs. Before I discuss how the modification

could be acting at the molecular level (see below) I have to address two important issues. First, I

need to have confidence that the observed phenotypes resulted mainly from the alteration of the

post-translational modifications and not from the changes in the primary sequence of tubulins.

Second, I need to address the potential interference of tubulin glutamylation, which is related in

many ways to glycylation and remains poorly characterized in ciliates.

The βDDDE440 mutation involved making a conserved change from glutamates to

aspartates at positions 437-439 in the primary sequence. There is a possibility that some or all of

the phenotypes that we observed were not directly related to the lack of the modification but are

a result of mutating the primary sequence. The mutations could, in some way affect tubulin

function at the level of folding, or dimerization, before the acquisition of PTMs. There are

162

several indications from our studies that this is not the case. As discussed earlier, the phenotype

of βDDDE440 was far less severe from the phenotype of a β-tubulin null mutant, showing that the

mutant tubulin was at least partly functional. In our tubulin exchange studies we were able to

confirm that the mutant βDDDE440 tubulin was capable of normal targeting, assembly and

exchange as the rates measured were comparable to those in the WT control strain (discussed in

Chapter 4). As a result, it is likely that the folding and consequently all of the downstream events

in which the nascent β-tubulin protein is involved prior to dimerization, are unaffected.

Furthermore, Xia et al showed that they could rescue the same lethal TuG domain mutation on β-

tubulin (βDDDE440 ) by introducing a chimeric α-tubulin that had its COOH terminal tail

substituted by that of WT-β-tubulin (Xia et al., 2000). Thus, the βDDDE440 mutation is not lethal

when a WT tail of β-tubulin in place of the corresponding domain on α−tubulin. These data

provide another argument that there is no defect in the properties of β-tubulin between

translation and dimerization.

Studies done by our collaborators showed that a COOH terminal tail is essential for

survival of the cell, and the tails are interchangeable with cells growing normally with either a

α− or β-tubulin tail on both tubulins (Duan and Gorovsky, 2002). In addition, they found that a

lethal mutation on β-tubulin that lacks all possible sites of modification (β9D) can be rescued by

an α-tubulin containing a duplicated α-tubulin sequence of the COOH tail. However, this

chimeric α-tubulin cannot rescue a β-tubulin mutant lacking entirely the COOH terminal tail,

indicating an essential function of the COOH tail, different from that of tubulin modifications.

Hoyle et al investigated the capacity of a COOH terminal truncated form of β-tubulin to function

in axonemes in Drosophila (Hoyle et al., 2001). The truncated tail-less version of β-tubulin

could be incorporated into the dimer with subsequent assembly into axonemes depending on the

163

ratio of mutant tubulin in the dimer pool. Axonemes assembled from the tail-less tubulin

however, were non-motile owing to a disruption of the periodicity of the interactions with non-

tubulin components such as radial spokes (Hoyle et al., 2001). Taking these results together with

those by Duan and Gorovsky, we can speculate that the COOH terminal region of tubulins could

be interacting with motor proteins and structural MAPs in ways that do not require

modifications. There is already evidence for interactions between molecular motors and the

primary sequence of the COOH terminal tail of tubulin. It was shown that a lysine rich positively

charged loop of the kinesin-I motor domain (so called K-loop) interacts with the acidic glutamate

rich COOH terminal tails of tubulins (called the E-hook) during in vitro motility assays (Okada

and Hirokawa, 1999). Subtilisin digestion identified the E-hook as the binding partner of the K-

loop and it was demonstrated that this interaction was essential for the one-dimensional

Brownian-like movement of kinesin along MTs in the weak binding state of the motor (Okada

and Hirokawa, 1999). In relation to this finding, one could imagine that PTMs might be strongly

interfering with the binding of MAPs/ motor proteins to the COOH termini, due to their bulky

nature of these polymodifications localized there. Thus, it is possible that the PTMs may be

acting as a negative regulators of interactions of motor proteins with the primary sequence. To

summarize, it is unlikely that the mutation affected tubulin prior to assembly. However, some

interference with the function of the primary sequence after assembly cannot be excluded at this

time. If this is the case, based on the tail switch experiments, the function of the primary

sequence of the tail is transferable between the two subunits of the dimer.

Another important issue that must be addressed is whether the mutation that were introduced in

this study, affected only glycylation or possibly other PTMs as well. As was mentioned in

Chapter 2, the exact site location in the primary sequence of another related polymodification,

164

glutamylation, is not known. Glutamylation can be easily detected in both Paramecium and

Tetrahymena using modification-specific antibodies. However, mass spectrometry failed to

detect the glutamylated peptides in tubulins isolated from axonemes of Paramecium (Redeker et

al., 1994). It is likely that this modification is quantitatively minor in Paramecium. A similar

mass spectrometry analysis showed that presence of glutamylated tubulins in axonemes of

Tetrahymena (Redeker, Bre and Gaertig unpublished data). Both α- and β-tubulins were found to

be glutamylated. Often both glutamylation and glycylation were found on the same tail domains.

Unfortunately, the exact sites of glutamylation are unknown for Tetrahymena. In mouse tubulin,

the addition of this modification occurs on E445 of mouseα1/2 isotype (Edde et al., 1990), E 443

and E445 of mα4 (Redeker et al., 1998), E 438 of mβ6 (Alexander et al., 1991), E 435 of mβ2 and

mβ3 (Fouquet et al., 1994; Redeker et al., 1992; Rudiger et al., 1992), E 434 of mβ4 and E 441 of

mβ5 (Mary et al., 1994). When the sequences of β-tubulin tails of isotypes with known sites of

glutamylation are aligned with the Paramecium and Tetrahymena tail sequences, in the majority

of cases of other species, glutamylation was found on E residues upstream of the sites mutated in

the βDDDE440 mutants. Furthermore, the phenotype of βDDDE440 corresponds largely to the

types of MTs, which from immunological studies are known to have the highest level of

glycylation and not glutamylation. For example, in the βDDDE440 mutant, the strongest

phenotypic effect is observed in cilia, while BBs are affected infrequently and in a more subtle

way. These phenotypes correspond with the highest level of polyglycylation in cilia. On the

other side, glutamylation is highly concentrated in the BB (Fouquet et al., 1997; Pechart et al.,

1999). Thus, we certainly mutated the major sites of glycylation. However, the proximity of sites

of glutamylation raises a possibility that we affected this modification as well. Future studies

should use mass spectrometry and establish whether only glycylation, or also glutamylation,

165

decrease in the βDDDE440 mutant. Furthermore, one would also need to analyze the α-tubulin in

the same mutant because compensating changes in the PTM distributions on the partner subunit

of the dimer could occur. The cadmium-inducible mutant will provide a large quantity of

genetically homogenous material for biochemical and structural studies.

Ultimately, even better distinction between the contributions of glutamylation and

glycylation will be achieved when we will identify the responsible enzymes and target them in

vivo. Intense work in this direction is under way in several laboratories, including our own (see

below). Finally, it should be mentioned that the distinction between glutamylation and

glycylation might not be as important as it now appears. This is because evolutionary evidence

suggests that the functions of these two structurally related modifications are similar. For

example, the flagellate Trypanosoma assembles a variety of MT organelles, including a motile

flagellum, while the tubulins are only glutamylated (Schneider et al., 1997). Furthermore, while

glutamylation is abundant in Chlamydomonas (Kann et al., 2003), glycylation appears to be

limited to monoglycylation (Pechart et al., 1999), based on immunological evidence. Thus, the

two modifications may have similar functions. Certainly, in some cases glutamylation can fulfill

the function of both types of polymodifications. Whether a reverse is also true is not known.

In order to directly address the question of how tubulin modifications function in vivo, it

would be very informative to clone and sequence the modifying enzymes and carry out

subsequent deletion/overexpression studies to observe the effects on the MT subsets. Given the

recent sequencing of the Tetrahymena genome and the use of the genome in identification of

genes of interest based on data from other organisms, this seems quite feasible. Efforts have been

made in our lab to clone and purify the glycylase enzyme by a graduate student, Krzysztof

Rogowski. He was able to achieve 7000- fold purification of enzyme activity using a series of

166

chromatographic columns. However, so far, due to extremely low quantities of the enzyme, and

lack of Tetrahymena genome annotation (required for MS fingerprinting analysis), it was not

possible to identify the catalytic subunit of the enzyme. Although identification of the

glycylase(s) is an important goal, the functional analysis of the role of this enzyme may also be

complex, because at least one additional protein besides tubulins is also subjected to glycylation

(K. Clark and M. Gorovsky, personal communication). Thus, one enzyme could have several

targets and deleting the enzyme gene could lead to a complex phenotype. For this reason, the role

of the modification will be best understood from combined studies on the enzymes as well as

their tubulin targets.

Recently, NIMA kinases have been implicated as potentially important regulators of

tubulin glutamylation/glycylation. NIMA (Never in Mitosis) kinases form a family of cell-cycle

related, serine-threonine kinases that have been implicated in the G2/M transition, mitotic

spindle formation and regulation of the centrosome replication cycle (Faragher and Fry, 2003;

Mahjoub et al., 2002; Mayor et al., 1999; O'Connell et al., 1994). These protein kinases are

structurally related to the NIMA kinase of Aspergillus nidulans (NIMA-related kinases, Nrks,

are also known as Nek kinases) and are evolutionarily conserved from protists to man. Recently,

Westermann and Weber showed that a NIMA kinase consistently co-purified with the tubulin

glutamylation activity in the flagellate Crithidia (Westermann et al., 1999b). Westermann et al

initially isolated a nearly homogenous preparation of tubulin polyglutamylase from Crithidia.

The enzyme is bound to the stable MT cytoskeleton isolated by Triton X-100 extraction, which is

in agreement with the idea that the enzyme prefers polymeric tubulin as a substrate. In addition,

the tubulin glutamylase preparation from Crithidia was able to utilize mammalian brain tubulin

as a substrate and could modify synthetic peptides, which contained the COOH terminus of α

167

and β-tubulin (Westermann et al., 1999a). Westermann et al. identified a 54kD polypeptide that

was found in the final fraction of tubulin glutamylase (Westermann and Weber, 2002) as a novel

member of the NIMA family of mitotic kinases (CfNek). However, it would be too early to

conclude that CfNek is the tubulin glutamylase for several reasons. For example, Westermann

and Weber were not able to express a functional enzyme with glutamylating activity in a

heterologous system. A possibility remains that CFNek is not a bi-functional enzyme (a kinase

and a tubulin glutamylase), but rather an upstream regulator, which activates the tubulin

glutamylase by phosphorylation. It is furthermore possible that in the Crithidia purification

scheme, the enriched CfNek activated endogenous tubulin glutamylase that could be associated

with brain MTs that were used as an in vitro substrate. Regnard et al., previously reported the

partial purification of mouse brain tubulin polyglutamylase as part of a multimeric complex, and

confirmed the view that the enzyme prefers MTs as a substrate (Regnard et al., 1998). In a follow

up study, the same group reported evidence that the mammalian brain tubulin polyglutamylase

exists as a complex of at least 3 major protein species (Regnard et al., 2003). However, their

results indicate that brain tubulin polyglutamylase does not contain a NIMA/Nek-related subunit.

(B. Edde, personal communication). It could be that the two enzymes (in Crithidia and in a

mouse) follow a different catalytic mechanism. However, this appears to be highly unlikely –

why would the glutamylase evolve twice using different enzymatic strategies? It appears rather

that the catalytic subunit of the tubulin glutamylase is still unknown, and that the NIMA kinase is

its upstream regulator. Once, identified, the sequence of the glutamylase may also be useful in

identification of the glycylase, which could belong to the same family of enzymes (both proteins

bind to MTs, modify the γ-carboxyl group of the glutamic acid and function as amino acid

ligases).

168

Initially, two NIMA kinases were identified in Tetrahymena (D. Wloga, personal

communication), called Nrk1p and Nrk2p. Recently, up to 30 putative NIMA related kinase

sequences have been identified in Tetrahymena using the published genome (D. Wloga, personal

communication). Dr. Dorota Wloga in the lab has recently carried out overexpression studies of

the Nrks in Tetrahymena. Nrks had overlapping but nonidentical sites of localization with the

polymodifications, glutamylation and glycylation. Furthermore, overproduction of Nrk1p and

Nrk2p led to an increase in the levels of glutamylation, glycylation and acetylation in the areas

where these kinases localized. These interesting data indicate a role for the Nrks in regulating of

multiple PTMs, either directly by modulating related enzymes activities, or through an indirect

role by influencing other processes, which might in turn affect the levels of modifications (such

as MT stability).

Importantly, genetic studies have already linked a NIMA kinase to the function of

axonemal MTs. Mahjoub et al reported the identification of a new member of the NIMA kinase

family in Chlamydomonas reinhardtii, FA2 (Mahjoub et al., 2002). The FA2 gene was

discovered during a genetic screen for mutants defective in the deflagellation pathway in

Chlamydomonas (Finst et al., 1998). Deflagellation occurs upon increase in the Ca2+

concentration in the area of the transitional zone between the BB and the flagellum. As a result,

the axonemal microtubules are ruptured and the entire flagellum detaches from the cell. The

microtubule severing protein, ATPase katanin, has been implicated in deflagellation by its

localization to the transitional zone and based on the observation that anti-katanin antibodies

inhibited severing of flagellar MTs in vitro in a permeabilized cell model of Chlamydomonas

(Lohret et al., 1998). Katanins are heterodimeric MT severing protein complexes that promote

MT disassembly (Hartmann et al., 1998; McNally and Vale, 1993; McNally et al., 2000).

169

Accumulated data on katanin suggests a model for MT severing (Hartmann and Vale, 1999).

MTs act as a scaffold upon which katanin oligomerizes after it has exchanged its ADP for ATP.

Once a complete ring is assembled, the ATPase activity of katanin is stimulated leading to a

conformational change in katanin. This results in a destabilization of tubulin-tubulin contacts. It

has been shown that mutations in the FA2 gene lead to a defect in the calcium induced axonemal

MT severing (Finst et al., 1998; Quarmby, 1996). The discovery that FA2 is a NIMA kinase

indicates a possible role of this family of kinases in the regulation of MT severing. In

conjunction with the work from Weber’s lab, showing that a NIMA kinase could be a regulator

of glutamylation, there is a possibility that NIMA kinases regulates katanin action by interacting

with its substrate – tubulins. More specifically, it is entirely possible that glutamylation or

glycylation provide a mark on MTs, which stimulates the binding of katanin and promote MT

severing. Although glutamylation is excluded from the transitional zone between the BB and the

flagellum (Pechart et al., 1999), a zone of the highest level of glutamylation (and glycylation) is

immediately adjacent. Thus it is possible that katanin severs at the zone of the highest contrast

between modified and unmodified segments of the same MTs.

Remarkably, recent circumstantial evidence brings a possibility that PTM –dependent

MT severing also plays a role in the regulation of mitosis in C. elegans. Out of the two β-tubulin

isotypes expressed during early development in this organism, katanin prefers the TBB-2

isotype, over TBB-1 (Lu et al., 2003) as a substrate. For example, a mutation in the TBB-2

isotype rescues the embryonic phenotype caused by overexpression of katanin. What is quite

amazing is the location of the mutation in TBB2 that rescues the katanin phenotype – a

substitution of a glutamic acid in the COOH terminal tail to lysine. The specific glutamic acid

that is mutated is homologous to the strongest site of glycylation mapped for Tetrahymena and

170

Paramecium. The compositions of polymodifications in C.elegans is unknown but glutamylation

and not glycylation was detected in the related worm (Mansir and Justine, 1998) suggesting that

the glutamic acid mutated in TBB2 in C. elegans is subjected to glutamylation. Thus, the

mutated TBB2 isotype may rescue the katanin overexpression phenotype simply because the

mutant tubulin is glutamylated to a lesser extent. Remarkably, the βDDDE440 mutant also

displays lack of severing of LM MTs in the cleavage furrow region. Thus, one of the functions

of glycylation and related glutamylation may be to mark a subset of MTs for severing during a

proper stage of the cell cycle. It will be important to see if the βDDDE440 cytokinesis defects

could be rescued by overexpressing either one or both of the Tetrahymena Nrks. An upregulation

of the NRK1 or NRK2 gene could lead to an upregulation of tubulin modifications on LMs might

result in rescue of the cytokinesis defect in the mutant. Furthermore, the severing phenotype in

the βDDDE440 mutants could also be rescuable by overexpression of katanin. Two katanin

homologs have been identified in the Tetrahymena genome (D. Wloga, personal

communication), which are yet to be investigated further.

The potential involvement of polymodifications in marking MTs for severing may be a

reflection of a general mode of how PTMs work. It is possible that the primary mode of action of

PTMs is by enabling binding or regulating properties of various MAPs attached to MTs. In vitro,

the MAP (Tau, Map2 and kinesin) binding to unpolymerized brain tubulins was dependent on

the polyglutamyl chain length (Bonnet et al., 2000; Boucher et al., 1994; Larcher et al., 1996).

The affinity of these proteins for tubulins increased for isoforms with 1-3 glutamate residues and

then progressively decreases when the chain length increased further, up to 6 units. Thus, as the

MT polymer grows, the varying lengths of the polymodification side chains could be providing

spatial and temporal cues to various MAPs, in terms of regulating their binding activity. As

171

mentioned earlier, gradients have been observed in the case of monoglycylation, polyglycylation

and polyglutamylation along the lengths of the axoneme. The number and lengths of side chains

added on to tubulins by the modification enzymes could simply be a function of how long the

assembled polymer has been exposed to the enzyme, assuming that they are distributed

uniformly along the length of the axoneme. Thus for polyglycylation and polyglutamylation, the

parts of the polymer that have persisted for longer periods of time (i.e at the minus end of the

axoneme) would have longer chains of polymodifications. In effect, the polymeric nature of

polymodifications provides a unique mechanism for fine tuning of binding of a variety of MAPs

along the length of the microtubule, as a function of the distance of a given segment of a MT

from the plus end. It needs to be mentioned that the length of the polymodification chain could

provide not only spatial but also temporal information, as the side chain length is dependent on

both the age of the polymer as well as the distance from the plus end.

Furthermore, our results suggest that a single modification type could affect multiple

types of MTs in different ways. For example, the same mutation has a destabilizing effect on

cilia and cause hypertrophy and lack of proper depolymerization in the cortex. Thus, the effects

of polyglycylation appear to be context-specific. We postulate that the main function of PTMs is

in regulation of binding of MAPs to MTs. The simplest explanation of the context-dependent

action of glycylation is that different types of MAPs bind to modified MTs in different

organelles. Thus, the targeting of MAPs would need to be independent of the presence of PTMs.

We showed here that the 13C4 MAP is restricted to the cortex, and that its association with

cortical MTs was compromised by deficiency in the tubulin glycylation domain. Future studies

should establish which MAPs are affected by the modification in other locations including cilia.

172

Probably the most surprising result of our study is the demonstration that the phenotypic

consequences of the tubulin glycylation mutation are dependent on cell polarity. The

βDDDE440 mutant displays a selective arrest of growth and nuclear division specifically in the

anterior subcell of the mutant chains. Related to this observation, the localization of the cortical

MAP, 13C4p, is dependent on the TuG domain specifically in the anterior cells, indicating

another strong polarized effect of the mutation. Thus, it appears that the TuG domain is involved

in regulating the binding of this MAP in a polarity-dependent manner. The polarity effects

observed could indicate a possible role of the TuG domain in setting up the anterior/posterior

differences prior to cell division. As is the case with 13C4p, one could possibly expect other

MAPs to also be regulated by the TuG domain. This regulation of binding could be affecting

MAPs that might be involved in the process of cell division in Tetrahymena, leading up to the

cytokinesis defects we observe in the βDDDE440 mutant. As discussed in preceding sections, the

polymodifications might serve as a marker for binding of MAPs. Overexpression studies of the

NIMA kinase (Nrk2p) in Tetrahymena, showed an increased localization of this kinase in BBs

just posterior to the cleavage furrow and a corresponding increase in polyglycylation and

polyglutamylation in the same region (D. Wloga, personal communication). Thus, these cell-

cycle regulated kinases might serve as morphogenetic markers for binding/dissociation of

necessary MAPs along the antero-posterior axis during the process of cell division by

upregulating the levels of tubulin modification.

Asymmetric partitioning of cell fate determinants in higher organisms is employed to

generate cell type diversity (Horvitz and Herskowitz, 1992). The direction of division and the

ability of a cell to divide asymmetrically or symmetrically is strongly related to the functions of

the cytoskeleton. Six polarity related genes, PAR genes have been identified in C. elegans,

173

whose products appear responsible for establishing asymmetries that define the antero-posterior

body axis in C. elegans (reviewed in (Golden, 2000)). The functions of the P AR genes are

conserved in Drosophila and human cells (Wodarz, 2002). Recent evidence showed that some

PAR proteins regulate the dynamics of MTs in the early embryo of C.elegans (Labbe et al 2003).

Specifically, MTs near the posterior cortex of the embryo were found to be more dynamic.

Tetrahymena cell also has a very strong polarity. However, there appears to be no homologs of

PAR genes in the Tetrahymena genome (J. Gaertig, K. Rogowski, personal communication). On

the other side, we could speculate that one of more of PAR gene products modulates the

properties of MTs by affecting tubulin modifications.

References:

Alexander, J.E., D.F. Hunt, M.K. Lee, J. Shabanowitz, H. Michel, S.C. Berlin, T.L. Macdonald,

R.J. Sundberg, L.I. Rebhun, and A. Frankfurter. 1991. Characterization of post-

translational modifications in neuron-specific class III β-tubulin by mass spectrometry.

Proc.Natl.Acad.Sci.USA. 88:4685-4689.

Bonnet, C., D. Boucher, S. Lazereg, B. Pedrotti, K. Islam, P. Denoulet, and J.C. Larcher. 2000.

Differential binding regulation of microtubule-associated proteins MAP1A, MAP1B, and

MAP2 by tubulin polyglutamylation. J. Biol. Chem. 276:12839-12848.

Boucher, D., J.C. Larcher, F. Gros, and P. Denoulet. 1994. Polyglutamylation of tubulin as a

progressive regulator of in vitro interactions between the microtubule-associated protein

Tau and tubulin. Biochemistry. 33:12471-12477.

174

Brown, J.M., N.A. Fine, G. Pandiyan, R. Thazhath, and J. Gaertig. 2003. Hypoxia regulates

assembly of cilia in suppressors of Tetrahymena lacking an intraflagellar transport

subunit gene. Mol. Biol. Cell. 14:3192-3207.

Brown, J.M., C. Hardin, and J. Gaertig. 1999a. Rotokinesis, a novel phenomenon of cell

locomotion-assisted cytokinesis in the ciliate Tetrahymena thermophila. Int. Cell Biol.

Rep. 23:841-848.

Brown, J.M., C. Marsala, R. Kosoy, and J. Gaertig. 1999b. Kinesin-II is preferentially targeted to

assembling cilia and is required for ciliogenesis and normal cytokinesis in Tetrahymena.

Mol. Biol. Cell. 10:3081-3096.

Duan, J., and M.A. Gorovsky. 2002. Both carboxy terminal tails of alpha and beta tubulin are

essential, but either one will suffice. Curr Biol. 12:313-316.

Edde, B., J. Rossier, J.P. Le Caer, E. Desbruyeres, F. Gros, and P. Denoulet. 1990.

Posttranslational glutamylation of alpha tubulin. Science. 247:83-85.

Faragher, A.J., and A.M. Fry. 2003. Nek2A kinase stimulates centrosome dysjunction and is

required for formation of bipolar mitotic spindles. Mol Biol Cell. 14:2876-2889.

Finst, R.J., P.J. Kim, and L. Quarmby. 1998. Genetics of the deflagellation pathway in

Chlamydomonas. Genetics. 149:927-936.

Fouquet, J.P., B. Edde, M.L. Kann, A. Wolff, E. Desbruyeres, and P. Denoulet. 1994.

Differential distribution of glutamylated tubulin during spermatogenesis in mammalian

testis. Cell Motil.Cytoskeleton. 27:49-58.

Fouquet, J.P., M.L. Kann, I. Pechart, and Y. Prigent. 1997. Expression of tubulin isoforms during

the differentiation of mammalian spermatozoa. Tissue Cell. 29:573-583.

Frankel, J. 1999. Cell Biology of Tetrahymena. Methods Cell Biology. 62:27-125.

175

Frankel, J., E.M. Nelsen, and L.M. Jenkins. 1981. Development of the ciliature of Tetrahymena

thermophila. II. Spatial subdivision prior to cytokinesis. Dev. Biol. 88:39-54.

Golden, A. 2000. Cytoplasmic flow and the establishment of polarity in C.elegans 1 cell

embryos. Curr Opin Genet Dev. 10:414-420.

Hai, B., J. Gaertig, and M.A. Gorovsky. 1999. Knockout heterokaryons enable facile mutagenic

analysis of essential genes in Tetrahymena. Methods Cell Biol. 62:513-531.

Hartmann, J.J., J. Mahr, K.P. McNally, K. Okawa, A. Iwamatsu, S. Thomas, S. Cheesman, J.

Heuser, R.D. Vale, and F.J. McNally. 1998. Katanin, a microtubule-severing protein, is a

novel AAA ATPase that targets to the centrosome using a WD40 domain. Cell. 93:277-

287.

Hartmann, J.J., and R.D. Vale. 1999. Microtubule disasembly by ATP dependent

oligomerization of the AAA enzyme katanin. Science. 286:782-784.

Horvitz, H.R., and I. Herskowitz. 1992. Mechanism in asymmetric cell division: two B's or not

to B's, that is the question. Cell. 68:237-255.

Hoyle, H.D., F.R. Turner, L. Brunick, and E.C. Raff. 2001. Tubulin sorting during dimerization

in vivo. Mol. Biol. Cell. 12:2185-2194.

Iftode, F., J. Cohen, F. Ruiz, A.T. Rueda, L. Chen-Shan, A. Adoutte, and J. Beisson. 1989.

Development of surface pattern during division in Paramecium. 1. Mapping of

duplication and reorganization of cortical cytoskeletal structures in the wild type.

Development. 105:191-211.

Kaczanowski, A. 1978. Gradients of proliferation of ciliary basal bodies and the determination of

the position of the oral primordium in Tetrahymena. J. Exp. Zool. 204:417-430.

176

Kann, M.L., S. Soues, N. Levilliers, and J.P. Fouquet. 2003. Glutamylated tubulin: Diversity of

expression and distribution of isoforms. Cell Motil Cytoskeleton. 55:14-25.

Larcher, J.C., D. Boucher, S. Lazereg, F. Gros, and P. Denoulet. 1996. Interaction of kinesin

motor domains with alpha- and beta-tubulin subunits at a tau-independent binding site.

Regulation by polyglutamylation. J Biol Chem. 271:22174-22124.

Lohret, T.A., F.J. McNally, and L. Quarmby. 1998. A role for katanin-mediated axonemal

severing during Chlamydomonas defalgellation. Mol. Biol. Cell. 9:1195-1207.

Lu, C., M. Srayko, and P.E. Mains. 2003. The C. elegans microtubule severing complex

MEI/MEI-2 katanin interacts differentially with two superficially redundant beta tubulin

isotypes. Mol. Biol. Cell. In Press.

Mahjoub, M.R., B. Montpetit, L. Zhao, R.J. Finst, B. Goh, A.C. Kim, and L.M. Quarmby. 2002.

The FA2 gene of Chlamydomonas encodes a NIMA family kinase with roles in cell cycle

progression and microtubule severing during deflagellation. J Cell Sci. 115:1759-68.

Mansir, A., and J.L. Justine. 1998. The microtubular system and posttranslationally modified

tubulin during spermatogenesis in a parasitic nematode with amoeboid and aflagellate

spermatozoa. Mol Reprod Dev. 49:150-167.

Marshall, W.F., and J.L. Rosenbaum. 2001. Intraflagellar transport balances continuous turnover

of outer doublet microtubules: implications for flagellar length control. J Cell Biol.

155:In Press.

Mary, J., V. Redeker, J.P. Le Caer, J.C. Prome, and J. Rossier. 1994. Class I and IVa beta-

tubulin isotypes expressed in adult mouse brain are glutamylated. FEBS Lett. 353:89-94.

Mayor, T., P. Meraldi, Y.D. Steirhof, E.A. Nigg, and A.M. Fry. 1999. Protein kinases in control

of the centrosome cycle. FEBS Lett. 452:92-95.

177

McKean, P.G., A. Baines, S. Vaughan, and K. Gull. 2003. Gamma-tubulin functions in the

nucleation of a discrete subset of microtubules in the eukaryotic flagellum. Curr Biol.

13:598-602.

McNally, F.J., and R.D. Vale. 1993. Identification of katanin, an ATPase that severs and

disassembles stable microtubules. Cell. 75:419-429.

McNally, K.P., O.A. Bazirgan, and F.J. McNally. 2000. Two domains of p80 katanin regulate

microtubule severing and spindle pole targeting by p60 katanin. J Cell Sci. 113:1623-

1633.

Multigner, L., I. Pignot-Paintrand, Y. Saoudi, D. Job, U. Plessmann, M. Rüdiger, and K. Weber.

1996. The A and B tubules of the outer doublets of sea urchin sperm axonemes are

composed of different tubulin variants. Biochemistry. 35:10862-10871.

Nanney, D.L. 1975. Patterns of basal body addition in ciliary rows in Tetrahymena. J. Cell Biol.

65:503-512.

O'Connell, M.J., C. Norbury, and P. Nurse. 1994. Premature chromatin condensation upon

accumulation of NIMA. EMBO J. 13:4926-4937.

Okada, Y., and N. Hirokawa. 1999. Mechanism of the single-headed processivity: diffusional

anchoring between the K-loop of kinesin and the C-terminus of tubulin. Proc. Natl. Acad.

Sci.U.S.A. 18:640-645.

Pechart, I., M.L. Kann, N. Levilliers, M.H. Bre, and J.P. Fouquet. 1999. Composition and

organization of tubulin isoforms reveals a variety of axonemal models. Biol. Cell.

91:685-97.

178

Pedersen, L.B., S. Geimer, R.D. Sloboda, and J.L. Rosenbaum. 2003. The microtubule plus end

tracking protein EB1 is localized to the flagellar tip and basal bodies in Chlamydomonas

reinhardtii. Curr Biol. 13:1969-1974.

Quarmby, L. 1996. Ca2+ influx activated by low pH in Chlamydomonas. J Gen Physiol.

108:351-361.

Raff, E.C., J.A. Hutchens, H.D. Hoyle, M.G. Nielsen, and F.R. Turner. 2000. Conserved

axoneme symmetry altered by a component beta-tubulin. Current Biol. 10:1391-1394.

Rasmussen, L., and E. Orias. 1975. Tetrahymena: growth without phagocytosis. Science.

190:464-465.

Redeker, V., N. Levilliers, J.-M. Schmitter, J.-P. Le Caer, J. Rossier, A. Adoutte, and M.-H. Bré.

1994. Polyglycylation of tubulin: a post-translational modification in axonemal

microtubules. Science. 266:1688-1691.

Redeker, V., R. Melki, D. Promé, J.-P. Le Caer, and J. Rossier. 1992. Structure of tubulin C-

terminal domain obtained by subtilisin treatment. The major α- and β- tubulin isotypes

from pig brain are glutamylated. FEBS Lett. 313:185-192.

Redeker, V., J. Rossier, and A. Frankfurter. 1998. Posttranslational modifications of the C-

terminus of α-tubulin in adult rat brain: a4 is glutamylated at two residues. Biochemistry.

37:14838-14844.

Regnard, C., S. Audebert, Desbruyeres, P. Denoulet, and B. Edde. 1998. Tubulin

polyglutamylase: partial purification and enzymatic properties. Biochemistry. 37:8395-

8404.

179

Regnard, C., D. Fesquet, C. Janke, D. Boucher, E. Desbruyeres, A. Koulakoff, C. Insina, P.

Travo, and B. Edde. 2003. Characterization of PGs1, a subunit of a protein complex co-

purifying with tubulin polyglutamylase. J Cell Sci. 116:4181-4190.

Rudiger, M., U. Plessmann, K.D. Kloppel, J. Wehland, and K. Weber. 1992. Class II tubulin, the

major brain β-tubulin isotype is polyglutamylated on glutamic acid residue 435. FEBS

Lett. 308:101-105.

Sale, W.S., and P. Satir. 1977. The termination of the central microtubules from the cilia of

Tetrahymena pyriformis. Cell Biol Inter. 1:45-49.

Schneider, A., U. Plessmann, and K. Weber. 1997. Subpellicular and flagellar microtubules of

Trypanosoma brucei are extensively glutamylated. J. Cell Sci. 110:431-437.

Shang, Y., B. Li, and M.A. Gorovsky. 2002a. Tetrahymena thermophila contains a conventional

gamma tubulin that is differentially required for the maintenance of different micotubule

organizing centers. J Cell Biol. 158:1195-1206.

Shang, Y., X. Song, J. Bowen, R. Corstanje, Y. Gao, J. Gaertig, and M.A. Gorovsky. 2002b. A

robust inducible-repressible promoter greatly facilitates gene knockouts, conditional

expression, and overexpression of homologous and heterologous genes in Tetrahymena

thermophila. Proc Natl Acad Sci. 99:3734-3739.

Song, L., and W. Dentler. 2001. Flagellar protein dynamics in Chlamydomonas. J. Biol. Chem.

276:29754-29763.

Sundaraman, V., and E.D. hanson. 1976. Longitudinal microtubules and their functions during

asexual reproduction in Paramecium tetraurelia. Genet.Res.Cambr. 27:205-211.

180

Vinh, J., J.I. Langridge, M.-H. Bre, N. Levilliers, V. Redeker, D. Loyaux, and J. Rossier. 1999.

Structural characterization by tandem spectroscopy of the posttranslational modifications

of tubulin. Biochemistry. 38:3133-3139.

Wang, Z., and M.P. Sheetz. 2000. The C-terminus of tubulin increases cytoplasmic dynein and

kinesin processivity. Biophys. J. 78:1955-1964.

Westermann, S., U. Plessmann, and K. Weber. 1999a. Synthetic peptides identify the minimal

substrate requirements of tubulin polyglutamylase in side chain elongation. FEBS Lett.

459:90-94.

Westermann, S., A. Schneider, E.K. Horn, and K. Weber. 1999b. Isolation of tubulin

polyglutamylase from Crithidia; binding to microtubules and tubulin, and glutamylation

of mammalian brain alpha- and beta-tubulins. J. Cell Sci. 112:2185-2193.

Westermann, S., and K. Weber. 2002. Identification of CfNek, a novel member of the NIMA

family of cell cycle regulators, as a polypeptide copurifying with ubulin. J Cell Sci.

115:5003-5012.

Wodarz, A. 2002. Establishing cell polarity in development. Nat. Cell Biol. 4:E39-E44.

Xia, L., B. Hai, Y. Gao, D. Burnette, R. Thazhath, J. Duan, M.-H. Bré, N. Levilliers, M.A.

Gorovsky, and J. Gaertig. 2000. Polyglycylation of tubulin is essential and affects cell motility

and division in Tetrahymena thermophila. J.Cell Biol. 149:1097-1106.

181

Figure legends:

Figure 5.1: Growth curve of βDDDE440 mutants (open symbols) compared to that of kinesin

knockout mutants (closed symbols) in MEPP medium. Note that Kinesin knockouts can be

maintained almost indefinitely and appear to divide normally in MEPP whereas the βDDDE440

mutants fail to divide more than 10-12 cells per drop.

Figure 5.2: Growth curve of βDDDE440 mutants (open symbols) plotted against that of β-tubulin

knockout mutants (closed symbols) in MEPP medium. The β-tubulin knockout mutants fail to

divide completely and die within 48h. The βDDDE440 mutants are able to complete at least a few

normal divisions. After a prolonged period of time, they arrest in cytokinesis and eventually die

in 7-8 days (Data not shown).