1

γγγγ-Phosphate Protonation and pH-Dependent Unfolding of the

ras·GTP·Mg2+ Complex: A Vibrational Spectroscopy Study‡

Hu Cheng, Sean Sukal, Robert Callender�* and Thomas S. Leyh§*

From The Department of Biochemistry, Albert Einstein College of Medicine, 1300 Morris

Park Ave, Bronx, New York 10461-1926

Supported by the National Institutes of Health Grants GM35183† and GM54469§,

and the National Science Foundation MCB-9727439†

*To whom reprint requests should be sent.

§To whom editorial correspondence should be sent.

Address: The Department of Biochemistry

Albert Einstein College of Medicine

1300 Morris Park Ave.

Bronx, New York 10461-1926

Phone: 718-430-2857

Fax: 718-430-8565

E-mail: leyh@aecom.yu.edu

Running Title: The Mechanism of Ras-Catalysed GTP Hydrolysis

Copyright 2000 by The American Society for Biochemistry and Molecular Biology, Inc.

JBC Papers in Press. Published on December 21, 2000 as Manuscript M009295200 by guest on A

pril 11, 2018http://w

ww

.jbc.org/D

ownloaded from

2

Summary

The interdependence of GTP hydrolysis and the second messenger functions of

virtually all GTPases has stimulated intensive study of the chemical mechanism of the

hydrolysis. Despite numerous mutagenesis studies, the presumed general base, whose

role is to activate hydrolysis by abstracting a proton from the nucleophilic water, has not

been identified. Recent theoretical and experimental work suggest that the γ-phosphate

of GTP could be the general base. The current study investigates this possibility by

studying the pH dependence of the vibrational spectrum of the ras·GTP·Mg2+ and

ras·GDP·Mg2+ complexes. Isotope edited Raman studies of the ras·GTP·Mg2+ complex

show that GTP remains bound to ras at pH as low as 2.0, and that the γ-phosphate is not

protonated at pH ≥ 3.3, indicating that the active-site decreases the γ-phosphate pKa by at

least 1.1 pKa units compared to solution. Amide I studies show that the ras·GTP·Mg2+

and ras·GDP·Mg2+ complexes partially unfold in what appear to be two transitions. The

first occurs in the pH range 5.4-2.6, and is readily reversible. Differences in the

pH-unfolding midpoints for the ras·GTP·Mg2+ and ras·GDP·Mg2+ complexes, 3.7 and 4.8

respectively, reveal that the enzyme·γ-phosphoryl interactions stabilize the structure. The

second transition, pH 2.6-1.7, is not readily reversed. The pH-dependent unfolding of

the ras·GTP·Mg2+ complex provides an alternative interpretation of the data that had been

used to support the γ-phosphate mechanism, thereby raising the issue of whether this

mechanim is operative in GTPase-catalyzed GTP hydrolysis reactions.

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

3

Introduction

The cellular activities of the ras protein are causally linked to tumorogenesis in a

high percentage and wide variety of human tissues (1). Binding of this cellular

messenger to Raf-1 protein kinase initiates a metabolic cascade that leads to cell growth

and differentiation (2, 3). Ras plays essential roles in regulating the cell cycle (4, 5),

apoptosis (6), mitogen-activated protein kinase pathways (7, 8) and retroviral activation

(9). The allosteric interactions that mediate signaling between GTPases and their targets

are typically switched from on to off by the hydrolysis of GTP (10, 11); however, there

are exceptions (12). The interdependence of hydrolysis and signaling functions of

GTPases has fostered intensive studies of the chemical mechanism of the ras-catalyzed

hydrolysis of GTP.

Classical general-base theory suggests that ras might catalyze hydrolysis by

abstracting a proton from the nucleophilic water, thereby activating it to attack the

γ-phosphate. Careful inspection of the GTP-bound active-site of ras suggests five

possible general base candidates (asp38, asp57, gln61, glu62, and glu63). The GTPase

activity of mutants of each of these amino acids have been studied (13-16), and it is

generally agreed that the presumed base has not been identified. The lack of a conclusive

identification of the base has spawned alternative suggestions for the chemical

mechanism. A recent theory that has attracted considerable attention is the γ-Phosphate

General Base theory, which hypothesizes that the γ-phosphate of GTP, rather than an

amino acid R-group, is the general base.

The idea that the γ-phosphate might be the general base in GTPase catalyzed

hydrolysis reactions was described first in a computational study that calculated that the

hydrolysis ∆G� for this mechanism is 24 ± 5 kcal/mole, which is close to the

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

4

experimentally observed value, ~ 22 kcal/mole (17). Support for this theory was

provided in a subsequent paper in which the 31P-NMR of ras-bound GTP was studied as a

function of pH. The α, β, and γ-resonances of GTP shift concomitantly at a pH midpoint

of 2.9, which agrees, within error, with the midpoint obtained from pH-rate studies of

ras-catalyzed GTP hydrolysis. pH-rate studies of ras kcat-mutants were used to construct

a free-energy plot (∆G� vs pKa); the plot was linear, with a slope of 2.1 (18). These

observations are consistent with, but do not prove that the observed midpoint is the pKa

of the γ-phosphate of the bound GTP. Vibrational spectroscopy provides an

unambiguous means of assigning the protonation state of phosphate, and can be used to

accurately define the structural and bonding changes that occur upon protonation (19,

20). To definitively assess whether the midpoint is, in fact, the pKa of the γ-phosphate,

and to define the bonding changes associated with the protonation, the pH-dependence of

the vibrational spectrum of the ras·GTP·Mg2+ complex was investigated.

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

5

MATERIALS AND METHODS

The materials and suppliers used in this study are as follows: restriction enzymes

(New England Biolabs), Pfu polymerase (Strategene), PEI F-cellulose TLC1 plates (EM

Science), KCNO (Acros Organics), radionucleotides (NEN Life Science Products), 18O-water (96.2 atom % 18O) (Isotec), Q Sepharose Fast Flow resin (Pharmacia), P6 Spin

Columns (Biorad), DNA primers (Oligonucleotide Synthesis Facility, Albert Einstein

College of Medicine), BL21(DE3) E. coli and pET-23a(+) plasmid (Novagen), Centricon

10 (Amicon). Buffers, media, enzymes, nucleotides, PCl5, and solvents were the highest

grades available from Aldrich, Sigma or Fisher Scientific.

Construction of the ras expression vector. Oligonucleotide directed PCR was used to

amplify the DNA coding sequence for amino acids 1-166 of human Ras from a yeast

expression plasmid using the following oligonucleotides 5�-ATTCCATATGACCGAAT

ACAAACTGG-3� and 5�-GCGGATCCGTTAGTGCTGACGGATTTCACGAAC-3�.

The amplified product contained Nde I and BamH I restriction sites at the 3� and 5� ends,

respectively, that were used to insert the ras coding region into pET-23a(+). Each strand

of the ras coding region in the pET-23a(+)/ras expression vector was sequenced. The

primary sequence predicted by the coding region exactly matched that of the published

human ras sequence (GenBank accession #J00277). The expression system produced ras

at ≥ 10% of the soluble cellular fraction (SDS-PAGE).

The ras purification. Ras was purified using standard procedures (21, 22).

Approximately 10 mg of > 95% pure protein ras was obtained per liter of cell culture.

The pure protein was stored in buffer containing 0.10 mM GDP at -70 oC.

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

6

The ras activity assay. GDP was removed from ras stocks by incubating ras (52 µM,

final) for 5 min in exchange buffer (23, 24) (50 mM Tris (pH 7.5), 20 mM EDTA, 200

mM (NH4)2SO4, 10 mM DTT, T = 25oC). The enzyme/exchange solution was then

loaded onto a P6 spin column (equilibrated in 50 mM Tris, pH 7.5, 10 mM DTT) and

eluted by centrifugation directly into 1.2 active-site equivalents of GTP/γ-32P-GTP tracer.

Ras was nucleotide free for less than 1 minute. The GTP hydrolysis reaction was

initiated by the addition of MgCl2 (10 mM, final) and quenched by the addition of

Na4EDTA, pH 9.5 (200 mM, final). Ten-minute time points were taken, spotted on PEI

F-Cellulose TLC plates, and developed in a 0.90 M LiCl mobile phase. The radioactive

reactants were quantitated using an AMBIS 2-dimensional radioactivity scanner. The

resulting time course fit a single exponential to give a calculated turnover number of

0.013 ± 0.005 min-1.

The synthesis of γ-18O3-GTP. 460 µmoles of fresh PCl5 was added quickly to a 30-fold

molar excess (250 µl) of rapidly stirred 18O-water. The solution was brought to pH 7.0

by addition of 10 M KOH. 460 µmoles sodium acetate was then added from a 1.0 M

stock. The mixture was diluted two-fold with unlabelled water, and 7.5 molar

equivalents of KCNO were added from a freshly prepared 6M stock. The reaction

solution was maintained at 30 oC, and kept at pH 6.5 by the addition of glacial acetic acid

in five minute intervals over 30 minutes. Two additional molar equivalents of KCNO

were added, and pH monitoring was continued for an additional 30 minutes. 4.0 ml of a

solution containing ADP (115 mM (460 µmoles), MgSO4 ( 116 mM) and Tris (0.58 M,

pH 7.3) was added, followed by 17 U (µmoles/minute) carbamate kinase, and the mixture

incubated at 38 oC for 30 minutes to convert the ADP and carbamyl phosphate to

γ-18O3-ATP. The nucleotide was diluted to 5.5 mM using an 11 mM Tris (pH 7.5), 11

mM MgSO4 buffer. Five mole equivalents GDP and 200 U nucleoside-5�-diphosphate

kinase were added, and the mixture was incubated at 37 oC for 1 hour.

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

7

The nucleotide was purified on a Q Sepharose Fast Flow anion exchange column

using a 10 mM to 1.2 M triethylamine-bicarbonate (pH 7.6) linear gradient. GTP eluted

at 1.0 M triethylamine-bicarbonate. Fractions containing GTP were collected and rerun

under similar conditions to further remove nucleotide contaminants. Triethylamine was

removed by rotary evaporation to dryness; the GTP was suspended and rotovaped three

times in methanol and three times in water. The nucleotide was then dissolved in water,

and the pH was adjusted to 8.5 by addition of 10 M KOH. The purified product was

stored at -70 oC. TLC on PEI F-Cellulose plates showed that the γ-18O3-GTP was greater

than 95% pure.

Electrospray mass spectrometry. The nucleotides were analyzed by mass spectrometry at

the Hunter College Mass Spectrometry Facility, NY, NY, on an Agilent Technologies

(formerly Hewlett Packard) HP1100 LC/MSD instrument. Negative ion electrospray

spectra were analyzed to give 89.6% and 88.0% enrichment in 18O for γ-18O3-GTP.

Nucleotide Exchange. Ras·GDP storage buffer was replaced with nucleotide exchange

buffer (NEB) (200 mM Tris-HCl, 200 mM (NH4)2SO4, 0.55 mM DTT, 0.5 mM NaN3, 0.6

mM EDTA, pH 7.5) by repeated concentration and dilution at 4°C using Centricon 10

concentrators. After buffer exchange, ras was adjusted to ~0.50 mM and a 10-fold molar

excess of γ-18O3-GTP was added. Nucleotide exchange was complete after 30 minutes at

25 ± 2 ºC.

Sample preparation for γ-phosphate experiments. Following nucleotide exchange, NEB

buffer and excess nucleotide were removed, as described above, and replaced by 20 mM

Tris-HCl, 10 mM MgCl2, 0.5 mM NaN3, pH 7.5. The pH was adjusted by addition of

HCl or NaOH. The final concentration of ras used in the experiments was 2 ~ 4 mM.

Sample preparation for amide I experiments. Ras·GDP was exchanged into deuterized

buffer (20 mM Tris-HCl; 10 mM MgCl2; 0.5 mM NaN3; D2O solvent, pD 7.5). The pD

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

8

was adjusted with DCl or NaOD. The final ras concentration was 0.4 mM. Ras·GTP was

exchanged into 20 mM Tris-HCl, 0.5 mM NaN3, 0.6 mM EDTA, D2O solvent, pD 7.5.

The EDTA sequesters Mg2+ and prevents GTP hydrolysis. Concentrated protein was

then added to different buffers (chloroacetic acid, formic acid, and acetic acid) containing

MgCl2 just before the IR experiments.

Spectroscopy. The IR studies of the phosphate stretch modes involve isotope editing

procedures that have been discussed previously (19). All FTIR spectra were measured on

either an IFS-66 Fourier transform spectrometer (Bruker Instruments Inc., Billerica, MA)

or a Magna 760 Fourier transform spectrometer (Nicolet Instrument Corp., WI). A dual

cell shuttle accessory was used to perform the 18O � 16O difference spectroscopy. This

procedure makes sure the environments of two subtracted samples are essentially

identical. The Teflon spacer between two BaF2 cells was set at 25 µm. Typically, the

sample and reference spectra are not exactly identical. In this case, the parent spectra are

factored such that the amide II bands are nulled after subtraction. The protocol for buffer

subtraction is identical except that the baseline between 1750 cm-1 and 2000 cm-1 is

nulled. The contribution from vapor was also removed by subtraction.

Amide I measurements. The IR spectra of the protein complexes in the amide I region

were obtained at various pH values. The spectra were enhanced by Fourier Spectral

Deconvolution (FSD) (enhancement factor = 1.5, intrinsic bandwidth = 17 cm-1). The

FSD method helps to resolve the peak position of each component in the original spectra.

From the FSD spectra, we could determine the frequencies that characterize the high and

low pH forms. To find the pKa of the transition from low to high pH, the difference in

the intensity pairs that changed the most during the titration were plotted as a function of

pH (1640-1635 cm-1 and 1641-1638 cm-1 for the ras·GTP·Mg2+ and ras·GDP·Mg2+

complexes, respectively). Due to slightly varying protein concentrations at the different

pHs, the intensity differences were normalized by the area of the corresponding amide I

band. A linear transformation of the data was performed to make the differences lie

between 0 and 1, and the resulting Normalized Intensity vs pH plots were fit using the

Henderson-Hasselbalch equation to obtain apparent pKa values (Figure 3).

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

9

Results and Discussion

The validity of the γ-phosphate general-base hypothesis rests on the proposition that

the γ-phosphate of ras-bound GTP partially abstracts a proton from the nucleophilic

water, making it more hydroxide-like, thereby activating it to attack the γ-phosphate and

cleave the β,γ-bond of GTP. Experimental support for this hypothesis is provided by 31P-NMR pH-titration studies of ras·GTP·Mg2+ in which the α−, β− and γ-31P resonance

positions shifted concomitantly at a pH midpoint of 2.9. This midpoint was interpreted as

the pKa of the GTP γ-phosphate in the ras·GTP complex (16). In solution, the α−, β− and

γ-31P resonances shift independently as they become protonated (25). If the results of the

enzyme-bound studies are interpreted on the basis of the addition of a single proton to the

tripolyphosphate chain of GTP, the simultaneous shift of all three resonances creates an

unavoidable ambiguity regarding which of the three phosphoryl-groups receives the

proton � it is as if the proton �belongs� to each of them.

The vibrational spectrum of phosphate provides a sensitive, site-specific measure of

its protonation state. The frequencies of the symmetric and antisymmetric modes of the

P···O bonds of phosphate and phosphate monoesters undergo large (~ 75 cm-1) shifts as

phosphate switches between its mono- and dianionic forms (20). In favorable cases (i.e.,

when both the symmetric and antisymmetric frequencies of the oscillator can be

assigned) it is possible to quantitate changes in bond length and bond order with

accuracies of better than ± 0.004 Å, and 0.04 valence units (19), respectively. To obtain

a more complete description of the structural and bonding changes that the γ-phosphate

undergoes as the pH passes through the 2.9-midpoint, the pH dependence of the

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

10

vibrational spectrum of the γ-phosphate of GTP in the ras·GTP·Mg2+ complex was

studied using Raman spectroscopy.

The vibrational spectrum of the ras·GTP·Mg2+ complex is quite complex, and

identifying the stretch band of γ-phosphate of GTP is difficult (26). This problem can be

overcome using an isotope-editing scheme in which the natural-abundance spectrum of

ras·GTP·Mg2+ is subtracted from that of an identical complex in which the non-bridging

oxygen atoms of the GTP γ-phosphate have been enriched in 18O. Increasing the mass of

an oxygen nucleus shifts the vibrational frequencies of the modes in which it participates;

this frequency-offset is what is revealed in the difference spectrum. Properly subtracted

parent spectra produce a difference spectrum that contains only the differences in the

absorption bands associated with the bonds that involve the labeled nuclei (19, 26).

To aid in assigning the ras·GTP·Mg2+ difference spectra, and to provide

comparisons between the active-site and solution behavior of the γ-phosphate, the

difference spectra for the γ-phosphate-protonated and -deprotonated forms of GTP·Mg2+

in solution were obtained. To select the pH�s at which to acquire the protonated and

deprotonated spectra, the pKa of the GTP·Mg2+ complex (4.2 ± 0.2 ) was determined in

water at 25 ºC (data not shown). The unprotonated (pH 5.9), intermediate (pH 4.5) and

fully protonated (pH 2.8) difference spectra are presented in Panel A of Figure 1. The

protonated and deprotonated spectra do not change at higher and lower pH�s,

respectively. The intermediate spectrum, pH 4.5, appears to be a weighted composite of

the protonated and deprotonated spectra, suggesting that the system is well described by a

two-state (protonated/deprotonated) model.

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

11

The pH-dependence of the ras·GTP·Mg2+ difference spectra is shown in Panel B of

Figure 1. The position and 18O-isotopic shift of the ras-bound (pH 7.5) and

solution-phase (pH 5.9) γ-phosphate difference spectra of GTP·Mg2+ are extremely

similar. The ~ 12 cm-1 spectral shift that occurs when the nucleotide binds to ras results

from specific interactions between the γ-phosphate and the active-site (19, 27, 28). These

similarities strongly support that the γ-phosphate is dianionic at the active-site. The

spectrum begin to change at pH 3.3, indicating that the γ-phosphate remains largely

unprotonated at this pH. Thus, the active-site of ras has decreased the pKa of the

γ-phosphate by > 1.1 units compared to solution. As the pH decreases below 3.3, the

difference spectrum becomes distinctly different, indicating a considerable change in the

electronic structure of the γ-phosphate. However, the spectrum does not resemble that for

the protonated GTP·Mg2+ in solution. Therefore, the spectral change below pH 3.3

cannot be explained simply on the basis of a protonation of the γ-phosphate. Whether or

not γ-phosphate protonation occurs, there must be a substantial change at the active-site

to yield the very unusual observed bands.

The conclusions from the Raman data agree in many respects with previous NMR

results. The decreased pKa and change in the γ-phosphate environment are consistent

with the NMR results, as does the fact that most (greater than 90%) of the nucleotide

remains bound at the low pH � if this were not the case, the protonated solution spectrum

would be observed. However, despite our efforts to reverse the low pH transition, we

could only partially recover the high pH spectrum, suggesting that an irreversible event

occurs at low pH. Given the low pH of the event, it seemed likely to be caused by an

unfolding of the complex. To investigate further the unfolding issue, the pH-dependence

of the amide I spectrum of the nucleotide-bound complex was studied

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

12

The pH-dependent unfolding of ras. Protein amide I spectra (~ 1600 � 1700 cm-1) arise

from the C=O stretch of the polypeptide backbone. The position of a C=O band within

the amide I region is determined primarily by its hydrogen bonding, and transition dipole

coupling to other C=O groups, which is distance- and orientation-dependent (29). The

unique geometric and bonding characteristics of the different secondary structures result

in structure-specific amide I absorbance bands. Previous correlations between band

position and secondary structure (30, 31) have assigned absorbance centered at 1650-

1655 cm-1 to α-helix; absorbance at 1635 and 1675 cm-1 to the low (intense) and high

(weak) frequency components of β-sheet structure; peaks at 1668, 1675, and 1686 cm-1

are indicative of turns; the 1645 cm-1 region is associated with disordered polypeptide.

Despite ongoing efforts, accurate quantitation of the secondary structural composition

from amide I spectra is not yet generally feasible (32); however, changes in an amide I

spectrum are frequently used as a sensitive indicator of secondary-structural change (33-

36).

The amide I band of the ras·GTP·Mg2+ and ras·GDP·Mg2+ complexes at three

different pD values are shown in Panels A and B of Figure 2. The samples were prepared

in D2O to shift the intense IR absorption band of water out of the amide I region. To

better visualize the amide I regions that change as a function of pD, the spectra were

enhanced by Fourier Spectral Deconvolution (FSD) (37, 38). The FSD spectra are shown

in Panels C and D of Figure 2. The pD-dependent changes in the amide I spectra of these

complexes clearly indicate changes in their secondary structure. The GTP complex

unndergoes two denaturing transitions. The first, between pD* 5.4 and 2.6, is well

behaved and reversible. The second, between 2.6 and 1.7, is not well behaved. The

results were not reproducible at pD* values below 2.7. We suspect that this is due to a

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

13

tendency for ras to aggregate at this low pD* value - a common finding for acid-unfolded

proteins.

Ras contains large sections of β-sheet and α-helix, and, as expected, the amide I

region of the FSD spectrum of the ras·nucleotide complexes near neutrality show

appropriate amplitude in the α-helical (1656 cm-1) and β-sheet (1635 and 1674 cm-1)

regions. Low pH denatures both the α-helical and β-sheet structures, since the 1635 and

1651 cm-1 peaks in the high pD* are replaced by peaks at 1640 and 1657 cm-1. At lower

values of pD, below 2.6, the changes in the amide I absorbance indicate disruption of still

more secondary structure in the second transition

To better understand the pD-dependence of the first (reversible) transition,

titration curves of the amide I changes were constructed (Figure 3). The curves were fit

using the Henderson-Hasselbalch equation (the fits are shown as solid curves). The

ras·GDP·Mg2+ titration is fit well by a model in which the observed changes are due to

the addition of a single proton, with a pKa of 4.8. The single-proton scheme did not

describe the ras·GTP·Mg2+ titration well; however, a two-proton uptake model provided

an excellent fit. The apparent pKa predicted by the fit was 3.7. The 1.1 pKa difference in

the midpoints of the denaturation curves for the GTP and GDP complex indicates that

γ-phosphate·ras contacts stabilize protein toward acid-driven unfolding. The protein�s

structure is altered by these contact(s) such that the unfolding appears to be simultaneous

with the addition of approximately two protons, instead of one. The apparent identity of

the pKa�s of the groups protonated in the transition can be explained either on the basis of

synergism in the binding of these protons, or more simply, by, say, two carboxylates with

pKa�s within ~ 0.1 pKa units of one another. Ligand induced stabilization of the folded

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

14

state of proteins is well documented (39-41), and recent work has shown that GDP and

MgGDP stabilize the urea-driven unfolding of ras by -5.7 and -12.3 kcal/mole,

respectively (42, 43).

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

15

Conclusions

The difference spectrum for the deprotonated GTP γ-phosphate in solution resembles

that for the ras·GTP·Mg2+ complex at pH ≥ 3.3, and differs markedly from the protonated

γ-phosphate spectrum. It is clear from this and other vibrational studies (26-28) that the

GTP γ-phosphate is not protonated in the ras·GTP·Mg2+ complex. The pH-dependence of

the ras·GTP·Mg2+ difference spectrum indicates that the pKa of the γ-phosphate of bound

GTP is < 3.3, which is substantially lower than that for GTP·Mg2+ in solution, 4.2.

Below pH 3.3, GTP remains bound to the enzyme, and the γ-phosphate undergoes a

significant change in its environment. The spectrum does not resemble that for

protonated GTP and therefore the γ-phosphate clearly does not undergo a simple

protonation. The lack of detail in the low-pH 180-difference-spectrum is indicative of a

heterogeneous environment for the bound GTP. The protein unfolds at low pH � both its

α- and β- structures are altered. The pH dependence of the amide I spectrum of the

ras·GTP·Mg2+ complex reveals two unfolding transitions. One that is reversible, with an

apparent pKa of 3.7, and a second, that is not reversible, and has a lower apparent pKa,

2.6 � 1.7. The unfolding of the protein and the atypical difference-spectrum indicate that

the ras·GTP·Mg2+ complex, including its active site, is heterogeneous at low pH.

The current study affirms several of the interpretations made in previous work. The

α−, β− and γ-phosphate 31P-NMR resonances of ras·GTP·Mg2+ were shown to shift

concomitantly with a midpoint of 2.9 that was interpreted to be the pKa of the

γ-phosphate (16, 44). It is not possible, based solely on this correlation, to

unambiguously assign the physical basis of the chemical shift to protonation. For

example, the shift could be caused by a pH-dependent migration of the Mg2+ ion on the

tripolyphosphate chain, or protein structural change in the near-environment of the

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

16

γ-phosphate. In support of the previous interpretation, the current study provides direct,

site-specific confirmation of the fact that the γ-phosphate is not protonated in the

ras·GTP·Mg2+ complex, and that its pKa is shifted from 4.5 in solution to below 3.3 by

the active-site environment.

Consistent with our unfolding studies, the pH-dependence of the 1H-NMR spectrum of

ras·GTP·Mg2+shows that the complex partially unfolds between pH 6.8 and 3.3 (16). In

the pH range 3.3 � 2.2, which includes the 2.9 midpoint, the 1H-NMR changes are slight

suggesting that no major structural changes occur in the complex; however, our amide I

data clearly shows that the secondary structure of the complex changes in this pH range.

A pH-dependent change in the structure of the ras·GTP·Mg2+ complex can explain the

observed pH-dependence of the 31P chemical shift as well as the simultaneous shift in the

α, β, and γ-31P resonances. Structural change could also be the basis for the pH-rate

inactivation profiles used to construct the free energy plot (∆G� vs pKa) described in the

introduction. The slope of the free energy plot (i.e., 2.1) has been interpreted in several

ways, including a protein isomerization that precedes hydrolysis (44). This isomerization

could, in fact, be the unfolding of the protein. The pH-dependent unfolding of the

ras·GTP·Mg2+ complex provides a fully consistent, alternative interpretation of the

experimental support for the γ-Phosphate General Base Hypothesis, and thus raises the

issue of whether this mechanism is operative in GTPase catalyzed reactions.

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

17

Figure Legends

Figure 1. pH-Dependence of the GTP γ-Phosphate Raman Difference Spectra. Panel A.

The GTP·Mg2+ solution spectra. The sample composition was as follows: (γ-18O3)GTP

or GTP (4.0 mM), MgCl2 (10 mM), T = 20 ± 1°C. Panel B. The ras·GTP·Mg2+Spectra.

The solutions contained: ras·(γ-18O3)GTP·Mg2+ or ras·GTP·Mg2+ (2.0 mM), Mg2+

(10 mM), Tris/Cl (20 mM), pH 7.5, T = 4 ± 1°C. Samples were prepared and

measurements were made as described in Materials and Methods.

Figure 2. The pH-Dependence of the Amide I Spectra of ras·GTP·Mg2+ and

ras·GDP·Mg2+. Panels A and B. The pH dependence of the amide I spectra of

ras·GTP·Mg2+ and ras·GTP·Mg2+, respectively. Panels C and D. Fourier spectral

deconvolutions of the data shown in Panels A and B, respectively. Samples were

prepared, acquired and deconvoluted as described in Material and Methods.

Figure 3. The pH dependent unfolding of ras·GTP·Mg2+and ras·GDP·Mg2.

The pH titrations of the amide spectral changes that occur in the ras·GTP·Mg2+ (●) and

ras·GDP·Mg2 (▲) complexes are shown. The apparent pKa values were obtained by

fitting the data using the Henderson-Hasselbalch equation for one (▲) or two (●)

protons. pD* is the uncorrected pH meter reading, it does not account for D2O in the

samples. The best-fits are represented by the solid lines. The construction of the plots is

discussed in Materials and Methods.

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

18

References

1. Barbacid, M. (1987) Annu Rev Biochem 56, 779-827.

2. Zhang, X. F., Settleman, J., Kyriakis, J. M., Takeuchi-Suzuki, E., Elledge, S. J.,

Marshall, M. S., Bruder, J. T., Rapp, U. R. & Avruch, J. (1993) Nature 364, 308-

13.

3. Bourne, H. R. (1997) Science 278, 1898-9.

4. Peeper, D. S., Upton, T. M., Ladha, M. H., Neuman, E., Zalvide, J., Bernards, R.,

DeCaprio, J. A. & Ewen, M. E. (1997) Nature 386, 177-81.

5. Leone, G., DeGregori, J., Sears, R., Jakoi, L. & Nevins, J. R. (1997) Nature 387,

422-6.

6. Cardone, M. H., Roy, N., Stennicke, H. R., Salvesen, G. S., Franke, T. F.,

Stanbridge, E., Frisch, S. & Reed, J. C. (1998) Science 282, 1318-21.

7. Luttrell, L. M., Ferguson, S. S., Daaka, Y., Miller, W. E., Maudsley, S., Della

Rocca, G. J., Lin, F., Kawakatsu, H., Owada, K., Luttrell, D. K., Caron, M. G. &

Lefkowitz, R. J. (1999) Science 283, 655-61.

8. Marais, R., Light, Y., Mason, C., Paterson, H., Olson, M. F. & Marshall, C. J.

(1998) Science 280, 109-12.

9. Coffey, M. C., Strong, J. E., Forsyth, P. A. & Lee, P. W. (1998) Science 282,

1332-4.

10. Bourne, H. R., Sanders, D. A. & McCormick, F. (1990) Nature 348, 125-32.

11. Bourne, H. R., Sanders, D. A. & McCormick, F. (1991) Nature 349, 117-27.

12. Iiri, T., Farfel, Z. & Bourne, H. R. (1998) Nature 394, 35-8.

13. Cales, C., Hancock, J. F., Marshall, C. J. & Hall, A. (1988) Nature 332, 548-51.

14. John, J., Rensland, H., Schlichting, I., Vetter, I., Borasio, G. D., Goody, R. S. &

Wittinghofer, A. (1993) J Biol Chem 268, 923-9.

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

19

15. Krengel, U., Schlichting, L., Scherer, A., Schumann, R., Frech, M., John, J.,

Kabsch, W., Pai, E. F. & Wittinghofer, A. (1990) Cell 62, 539-48.

16. Schweins, T., Geyer, M., Scheffzek, K., Warshel, A., Kalbitzer, H. R. &

Wittinghofer, A. (1995) Nat Struct Biol 2, 36-44.

17. Neal, S. E., Eccleston, J. F., Hall, A. & Webb, M. R. (1988) J Biol Chem 263,

19718-22.

18. Schweins, T. & Warshel, A. (1996) Biochemistry 35, 14232-43.

19. Cheng, H., Sukal, S., Leyh, T. S. & Callender, R. (200) Biochemistry

(submitted).

20. Deng, H., Wang, J. H., Ray, W. J. & Callendar, R. (1998) J. Phys. Chem. B. 102,

3617-3623.

21. Tucker, J., Sczakiel, G., Feuerstein, J., John, J., Goody, R. S. & Wittinghofer, A.

(1986) Embo J 5, 1351-8.

22. Gibbs, J. B., Sigal, I. S., Poe, M. & Scolnick, E. M. (1984) Proc Natl Acad Sci U

S A 81, 5704-8.

23. Hoshino, M., Clanton, D. J., Shih, T. Y., Kawakita, M. & Hattori, S. (1987) J

Biochem (Tokyo) 102, 503-11.

24. Moore, K. J., Webb, M. R. & Eccleston, J. F. (1993) Biochemistry 32, 7451-9.

25. Cohn, M. & Hughes, T. (1960) J. Biol. Chem. 235, 3250-3253.

26. Wang, J. H., Xiao, D. G., Deng, H., Webb, M. R. & Callender, R. (1998)

Biochemistry 37, 11106-16.

27. Cepus, V., Scheidig, A. J., Goody, R. S. & Gerwert, K. (1998) Biochemistry 37,

10263-71.

28. Du, X., Frei, H. & Kim, S.-H. (2000) J. Biol. Chem. 275, 8492-8500.

29. Bandekar, J. (1992) Biochim Biophys Acta 1120, 123-43.

30. Prestrelski, S. J., Byler, D. M. & Thompson, M. P. (1991) Int J Pept Protein Res

37, 508-12.

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

20

31. Prestrelski, S. J., Byler, D. M. & Liebman, M. N. (1991) Biochemistry 30, 133-43.

32. Surewicz, W. K., Mantsch, H. H. & Chapman, D. (1993) Biochemistry 32, 389-

94.

33. Fraser, P. E., Nguyen, J. T., Surewicz, W. K. & Kirschner, D. A. (1991) Biophys J

60, 1190-201.

34. Kennedy, D. F., Slotboom, A. J., de Haas, G. H. & Chapman, D. (1990) Biochim

Biophys Acta 1040, 317-26.

35. Muga, A., Mantsch, H. H. & Surewicz, W. K. (1991) Biochemistry 30, 7219-24.

36. Trewhella, J., Liddle, W. K., Heidorn, D. B. & Strynadka, N. (1989) Biochemistry

28, 1294-301.

37. Wi, S., Pancoska, P. & Keiderling, T. A. (1998) Biospectroscopy 4, 93-106.

38. Kauppinen, J. K., Moffatt, D. J., Mansh, H. H. & Cameron, D. G. (1981) Appl.

Spectrosc. 35, 271-276.

39. Shortle, D. & Abeygunawardana, C. (1993) Structure 1, 121-34.

40. Shortle, D. (1996) Faseb J 10, 27-34.

41. Yutani, K., Ogasahara, K. & Kuwajima, K. (1992) J Mol Biol 228, 347-50.

42. Zhang, J. & Matthews, C. R. (1998) Biochemistry 37, 14881-90.

43. Zhang, J. & Matthews, C. R. (1998) Biochemistry 37, 14891-9.

44. Schweins, T., Geyer, M., Kalbitzer, H. R., Wittinghofer, A. & Warshel, A. (1996)

Biochemistry 35, 14225-31.

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

21

Abbreviations

EDTA, ethylenediamine-N,N,N’,N’-tetraacetic acid; PEI, polyethyleneimine; FTIR,

Fourier Transform Infrared Spectroscopy; Tris, tris-(hydroxymethyl)aminomethane; U

(unit), micromoles of substrate converted to product per minute at Vmax.

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

22

116011201080Wavenumber cm

-11160112010801040

Wavenumber cm-1

5.9

3.3

2.0

7.5

4.5

3.4

IR A

bsor

banc

eA. B.

pH

Figure 1

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

23

pH 5.4 pH 2.6 pH 1.7

(A) ras•GTP•Mg

16801660164016201600

Wavenumber cm-1

(C) FSD spectra of ras•GTP•Mg

pH 5.4 pH 2.6 pH 1.7

170016801660164016201600

Wavenumber cm-1

(D) FSD spectra of ras•GDP•Mg pH 7.7 pH 2.4

(B) ras•GDP•Mg pH 7.7 pH 2.4

20

40

60

0

20

40

IR A

bsor

banc

e (m

OD

)

0

80

1635

1640

1651

1657

1677

1674

1638

1641

1653

1675

1658 16

75

Figure 2

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

24

1.0

0.8

0.6

0.4

0.2

0.0

Nor

mal

ized

Int

ensi

ty

8.07.06.05.04.03.02.0

pD*

ras•GTP•Mg2+

ras•GDP•Mg2+

pKa = 3.7

pKa = 4.8

Figure 3

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Hu Cheng, Sean Sukal, Robert Callender and Thomas S. LeyhA vibrational spectroscopy study

+ complex:2-Phosphate protonation and pH-dependent unfolding of the ras.GTP.Mgγ

published online December 21, 2000J. Biol. Chem. 

  10.1074/jbc.M009295200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

by guest on April 11, 2018

http://ww

w.jbc.org/

Dow

nloaded from