Strategy for purification of aggregation prone β ...web.iitd.ernet.in/~tkchaudhuri/New...

10
New Biotechnology Volume 29, Number 3 February 2012 RESEARCH PAPER Strategy for purification of aggregation prone b-glucosidases from the cell wall of yeast: a preparative scale approach Mohammad Asif Shah 1,2 , Tapan Kumar Chaudhuri 1 and Saroj Mishra 2, 1 Structural Biology, Department of Biochemical Engineering and Biotechnology, Indian Institute of Technology Delhi, Hauz Khas, New Delhi 110016, India 2 Biochemical Research Laboratory, Department of Biochemical Engineering and Biotechnology, Indian Institute of Technology Delhi, Hauz Khas, New Delhi 110016, India Purification of biotechnologically important proteins is of vital interest to the biotech industry. b-Glucosidases, belonging to Family 1 and Family 3 of the glycosylhydrolases, have varied applications as carbohydrate hydrolyzing and synthesizing enzymes. Obtaining high quantities of these enzymes is important for exploring their biosynthetic potential, structural information and catalytic activities. Classical methods for their preparation fail to deliver high yields because of adoption of several/ hydroxyapatite chromatography steps. We report here a preparative method for purification of large quantities of two closely related cell bound b-glucosidases (BGL I and BGL II) from Pichia etchellsii that belong to Family 3 glycosylhydrolases. A combination of ion-exchange and gel filtration chromatography was used to process milligram quantities of protein with recoveries of up to 53%. A simple affinity based separation resulted in resolution of BGL I and BGL II with high recovery and high specific activities of 74 IU/mg and 32 IU/mg protein respectively. Peptide sequences of BGL II indicated it to be a novel member of Family 3. Methods reported here present a successful strategy for obtaining large quantities of these enzymes. Introduction b-Glucosidases constitute a major group among glycosylhydro- lases. They occur in all domains of living organisms (eubacteria, archaea, and eukarya). These enzymes specifically cleave glyco- sidic bonds (either O-linked b-glycosidic bonds, b-D-glucoside glucohydrolase, EC 3.2.1.21, or S-linked b-glycosidic bonds, myr- osinase, or b-D-thioglucoside glucohydrolase, EC 3.2.3.1) in oli- gosaccharides and their aryl- and alkyl conjugates [1,2]. The functional diversity exhibited by these enzymes has been utilized in a multitude of biotechnological applications that include cellulose to glucose conversion [3,4], cellulose to ethanol con- version [5], release of flavored aromatic compounds [6–8] from their flavorless glucosides. Under defined conditions, the enzymes can synthesize glucosyl bonds by reverse hydrolysis and transglycosylation approaches [9–11] leading to the synthesis of carbohydrates, considered as potential therapeutics [12]. Microarrays of synthetic carbohydrates have already been devised for probing carbohydrate–protein interactions [13,14] demand- ing the availability of sugars on large scale. While several Family 1 enzymes have been used in synthesis, very few studies have been conducted on Family 3 b-glucosidases. The structure of only a single Family 3 glycosylhydrolase, Hordaeum vulgare, has been solved by X-ray crystallography [15]. These pro- teins belong to a category of large proteins (approximately 850 aa) and are organized in multi-domain structures consisting of an (a/b) 8 fold TIM barrel and an (a/b) 6 fold b sandwich domain. They occur as closely related isozymes which are difficult to resolve. Many of these enzymes are cell surface bound with large hydrophobic patches which make them prone to aggregation. Pichia etchellsii, a thermo- tolerant yeast produces multiple b-glucosidases out of which two are cell wall bound. Both the enzymes show good synthetic potential. The synthesis of alkyl glucosides (and aryl glucosides or saccharides with degree of polymerization from 3 or more) can occur by (i) reverse hydrolysis or by (ii) transglycosylation. In the first method, Research Paper Corresponding author: Mishra, S. ([email protected]) 1871-6784/$ - see front matter ß 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.nbt.2011.06.010 www.elsevier.com/locate/nbt 311

Transcript of Strategy for purification of aggregation prone β ...web.iitd.ernet.in/~tkchaudhuri/New...

ResearchPap

er

New Biotechnology �Volume 29, Number 3 � February 2012 RESEARCH PAPER

Strategy for purification of aggregationprone b-glucosidases from the cell wall ofyeast: a preparative scale approach

Mohammad Asif Shah1,2, Tapan Kumar Chaudhuri1 and Saroj Mishra2,

1 Structural Biology, Department of Biochemical Engineering and Biotechnology, Indian Institute of Technology Delhi, Hauz Khas, New Delhi 110016, India2Biochemical Research Laboratory, Department of Biochemical Engineering and Biotechnology, Indian Institute of Technology Delhi, Hauz Khas, New Delhi

110016, India

Purification of biotechnologically important proteins is of vital interest to the biotech industry.

b-Glucosidases, belonging to Family 1 and Family 3 of the glycosylhydrolases, have varied applications

as carbohydrate hydrolyzing and synthesizing enzymes. Obtaining high quantities of these enzymes is

important for exploring their biosynthetic potential, structural information and catalytic activities.

Classical methods for their preparation fail to deliver high yields because of adoption of several/

hydroxyapatite chromatography steps. We report here a preparative method for purification of large

quantities of two closely related cell bound b-glucosidases (BGL I and BGL II) from Pichia etchellsii that

belong to Family 3 glycosylhydrolases. A combination of ion-exchange and gel filtration

chromatography was used to process milligram quantities of protein with recoveries of up to 53%. A

simple affinity based separation resulted in resolution of BGL I and BGL II with high recovery and high

specific activities of 74 IU/mg and 32 IU/mg protein respectively. Peptide sequences of BGL II indicated

it to be a novel member of Family 3. Methods reported here present a successful strategy for obtaining

large quantities of these enzymes.

Introductionb-Glucosidases constitute a major group among glycosylhydro-

lases. They occur in all domains of living organisms (eubacteria,

archaea, and eukarya). These enzymes specifically cleave glyco-

sidic bonds (either O-linked b-glycosidic bonds, b-D-glucoside

glucohydrolase, EC 3.2.1.21, or S-linked b-glycosidic bonds, myr-

osinase, or b-D-thioglucoside glucohydrolase, EC 3.2.3.1) in oli-

gosaccharides and their aryl- and alkyl conjugates [1,2]. The

functional diversity exhibited by these enzymes has been utilized

in a multitude of biotechnological applications that include

cellulose to glucose conversion [3,4], cellulose to ethanol con-

version [5], release of flavored aromatic compounds [6–8] from

their flavorless glucosides. Under defined conditions, the

enzymes can synthesize glucosyl bonds by reverse hydrolysis

and transglycosylation approaches [9–11] leading to the synthesis

of carbohydrates, considered as potential therapeutics [12].

Corresponding author: Mishra, S. ([email protected])

1871-6784/$ - see front matter � 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.nbt.2011.06.010

Microarrays of synthetic carbohydrates have already been devised

for probing carbohydrate–protein interactions [13,14] demand-

ing the availability of sugars on large scale.

While several Family 1 enzymes have been used in synthesis, very

few studies have been conducted on Family 3 b-glucosidases. The

structure of only a single Family 3 glycosylhydrolase, Hordaeum

vulgare, has been solved by X-ray crystallography [15]. These pro-

teins belong to a category of large proteins (approximately 850 aa)

and are organized in multi-domain structures consisting of an (a/b)8

fold TIM barrel and an (a/b)6 fold b sandwich domain. They occur as

closely related isozymes which are difficult to resolve. Many of these

enzymes are cell surface bound with large hydrophobic patches

which make them prone to aggregation. Pichia etchellsii, a thermo-

tolerant yeast produces multiple b-glucosidases out of which two are

cell wall bound. Both the enzymes show good synthetic potential.

The synthesis of alkyl glucosides (and aryl glucosides or saccharides

with degree of polymerization from 3 or more) can occur by (i)

reverse hydrolysis or by (ii) transglycosylation. In the first method,

www.elsevier.com/locate/nbt 311

RESEARCH PAPER New Biotechnology �Volume 29, Number 3 � February 2012

Research

Pap

er

enzyme-glycosyl intermediate is intercepted by another glucose (or

monosaccharide or an alcohol) resulting in the formation of sugars

of increasing chain length. This reaction is under thermodynamic

control and the equilibrium is shifted towards synthesis by reducing

the water activity and/or by increasing the substrate concentration.

In the second approach, an activated donor (e.g. lactose, cellobiose

or an aryl glucoside) is intercepted by an alcohol to generate a

glycosidic bond. This reaction is kinetically controlled. Both these

approaches have been used with the P. etchellsii b-glucosidases for

the synthesis of alkyl glucosides like hexyl-b-D-glucoside, heptyl b-D-

glucoside, octyl-b-D-glucoside, decyl-b-D-glucoside [16], oligosac-

charides like glucobiose, glucotriose and glucopentose [17]. Bio-

synthesis of monoterpenyl glucosides namely neryl, geranyl and

citronellyl-glucosides has also been reported [18]. Recently, the

whole cell suspensions of P. etchellsii, bearing the wall bound b-

glucosidases, have been reported to synthesize octyl-glucoside by

transglucosylation with conversions up to 60% [19]. Given the

importance of these enzymes, we report a preparative scale

approach for purification of these aggregation prone b-glucosidases.

The strategy presented here should be applicable for the purification

of other large molecular weight proteins of this category.

Materials and methodsStrain, maintenance medium and culture conditionsThe organism used in the study was an oenological yeast strain P.

etchellsii (DBEB-Y1015) JFG-2201 (Deutsche Sammlung von Mik-

roorganismen und Zellkulturen, DSMZ: Germany). The yeast was

maintained on YEPD slants or stabs containing (per liter); yeast

extract: 5 g, peptone: 2.5 g, glucose: 10 g, and agar: 20 g. Yeast

extract, peptone and agar were from HiMedia Laboratories, India

while glucose was from Qualigens Fine Chemicals, India. The pH

was adjusted to 5.5. The growth was carried out for 48 h on this

medium at 408C. The strain was maintained on YEPD and stored at

48C.

For enzyme production, a single colony was inoculated in 50 ml

of phosphate succinate (PS) minimal medium in 250 ml Erlen-

meyer flask, which contained (per liter); yeast extract: 2.5 g, pep-

tone: 5 g, succinic acid: 6 g, CaCl2: 0.3 g, K2HPO4: 8.7 g,

(NH4)2SO4: 4 g, MgSO4: 0.5 g, pH 4.7. All the components were

the highest grade available from Qualigens. The cells were allowed

to grow at 408C and 220 rpm in an orbital shaker (Scigenics, India)

for 8–10 h. This initial inoculum (2%, v/v) was transferred to

200 ml PS medium contained in 1000 ml Erlenmeyer flasks (12

in number) and allowed to grow for 14 h. Induction of b-gluco-

sidase was carried out by adding cellobiose (Sisco Research Labora-

tories, India) to a final concentration of 10 mM. The cells were

allowed to grow for another 6–8 h. Ten liters of cells was grown for

enzyme isolation.

Cell disruptionCells were harvested by centrifugation at 5000 � g for 10 min at

48C in a Sorvall RC 5B centrifuge. The pellet was washed with

50 mM sodium phosphate buffer, pH 7 (hereafter referred to as

buffer) three times by re-suspension and re-centrifugation. The

washed pellet was resuspended in phosphate buffer to make a cell

suspension of 15–25% (wcw/v) as described below. This suspen-

sion was used for the extraction of b-glucosidase enzymes using

French press or Dyno-Mill treatment. In the case of French press,

312 www.elsevier.com/locate/nbt

the cell pellet was resuspended in buffer to a final concentration of

15%, 20% or 25% (wcw/v). The cells were broken by passing the

suspension through the pressure cell at 1300 psi. A total of 5 passes

were carried out. The cell suspension was centrifuged at 10,000 � g

for 30 min at 48C and enzyme activity and protein was estimated

in the supernatant after every pass.

In the case of Dyno-Mill, three different modes of operation

(batch, continuous, and discontinuous batch) were evaluated for

maximum release of b-glucosidase activity from cell wall.

Batch operationFor batch mode, a 20% cell suspension of P. etchellsii was made in

buffer. Eighty milliliters of ice chilled glass beads was added to

20 ml of cell suspension. The cell suspension-glass bead mixture

was taken in a KD Dyno-Mill (Willy A Bachofen, Switzerland) and

agitated at 1500 rpm for various time periods ranging from 1 to

5 min. After each minute, samples were withdrawn and assayed for

enzyme and protein release. The operation was carried out at 48Cwhich was maintained by a circulating water bath (Pharmacia,

Sweden). The residual enzyme was recovered from glass beads by

washing the mill with same buffer. After disruption, the cell debris

was removed by centrifugation at 10,000 � g for 30 min at 48C in a

Sigma 3K30 centrifuge. The supernatant was used as a crude source

of enzyme. Protein concentration and enzyme activity were deter-

mined in the supernatant.

Continuous operationThe cell suspension was made as described above and the suspen-

sion was pumped continuously through the ice-chilled mill at

125–520 ml/min till 80–90% of cells were broken. The disruption

was monitored by examining the cells microscopically after dif-

ferent time intervals. After disruption, the mill was washed with

350–400 ml of buffer to remove the residual activity from the

beads. The cell debris was removed by centrifugation at

10,000 � g for 30 min at 48C. The cell-free extract was used as a

source of crude enzyme and enzyme activity and protein was

estimated in the cell-free extract.

Discontinuous batch operationThe cell suspension was prepared as described above. It was then

applied to Dyno-Mill at 125 ml/min for the first two passes. The

subsequent passes were performed at 525 ml/min. Samples were

collected after every pass and assayed for the release of protein and

b-glucosidase activity. A total of seven passes were carried out. The

mill was washed with l L of phosphate buffer to release the residual

enzyme from the glass beads. All the operations were carried out at

48C. The broken cell suspension was centrifuged at 10,000 � g for

30 min at 48C to remove the cell debris.

Column chromatographyThe cell-free extract was subjected to anion-exchange chromato-

graphy on Q-Sepharose Fast flow in XK16/30 or XK26/70 column

(Pharmacia, Sweden) (Amersham Biosciences, USA). The resin was

packed into the columns according to manufacturer’s instructions

(Pharmacia, Sweden) in Akta Explorer chromatographic system

(Amersham Biosciences, USA). The column was equilibrated with

buffer and cell-free extract applied at a flow rate of 0.5–1 ml/min

either with a peristaltic pump (Pharmacia, Sweden) or through

New Biotechnology �Volume 29, Number 3 � February 2012 RESEARCH PAPER

ResearchPap

er

P960 of Akta Explorer. The sample volume ranged from 0.5 to 3.0 L

containing a total protein of 0.25–3.3 g. The unbound proteins

were extensively washed with the same buffer. The washing was

monitored by observing the baseline stability. Elution was per-

formed with the same buffer containing KCl (130 or 200 mM).

Samples of 2–7 ml were collected and analyzed for the presence of

total protein and b-glucosidase activity.

Gel filtration was carried out by using Sephadex G 200 matrix or

Superdex 75 pre-packed column. In the case of Sephadex G 200

two columns were packed (XK 16/100 and XK 26/100, Pharmacia).

Relevant samples from anion exchange column were concentrated

by ammonium sulphate precipitation (80% saturation) and the

concentrate (in 2% of the bed volumes of the column) was applied

to Sephadex G 200 or Superdex 75 gel filtration columns. Fractions

of 2–5 ml size were collected at 0.2 ml/min flow rate, assayed for b-

glucosidase activity and checked on 10% SDS-PAGE. Those show-

ing BGL I and BGL II were pooled followed by the addition of solid

ammonium sulfate to a final concentration of 1 M. This was

applied to phenyl sepharose column (XK16/20, Pharmacia), pre-

equilibrated with buffer containing 1 M ammonium sulfate. The

unbound proteins were washed with the same buffer and elution

was carried out with a linear decreasing salt gradient (1–0 M

ammonium sulphate). Fractions of 1.5 ml size were collected at

0.25 ml/min. Those showing b-glucosidase activity were analyzed

on 10% SDS-PAGE. All steps were carried out at 208C.

Analytical methodsb-Glucosidase activity was measured on 4 mM pNPG (Sisco, India

and Aldrich) prepared in 50 mM phosphate citrate buffer, pH 6

according to the modified method [20]. A suitably diluted enzyme

(150 ml) was taken and added to 1.35 ml of pNPG solution pre-

incubated at 508C for 10 min. Reaction was allowed to occur for

10 min and stopped by adding 750 ml of 1 M Na2CO3. The absor-

bance of the resulting solution was measured at 420 nm against a

reagent blank. A p-nitrophenol (pNP, Sisco) standard curve (0.01–

0.1 mmol/ml) was made from a stock of 1 mM pNP prepared in

50 mM PC buffer, pH 6.0. Absorbance was taken at 420 nm. One

unit of b-glucosidase activity corresponded to the release of 1 mM

of pNP per min per ml of the reaction mixture. The specific activity

was expressed as IU/mg of protein.

Protein estimation was carried out by Bradford (Sigma–Aldrich)

dye binding method [21]. A suitably diluted 0.1 ml of protein

sample was taken and to this, 3.1 ml of Bradford reagent was

added. The amount of protein was estimated from a standard

curve of BSA (0.1–1.4 mg/ml) prepared in 50 mM sodium phos-

phate buffer, pH 7. Proteins from each purification step were

analyzed by SDS-PAGE on 10% gel containing 6% stacking gel,

and stained with Coomassie Brilliant Blue R-250.

Mass spectrometry and internal peptide sequencing of BGL IIThe purified proteins from phenyl sepharose chromatography

were dialyzed against Milli-Q water. The samples were concen-

trated by lyophilization. For analysis, the protein samples were

diluted in 1:1 ratio of acetonitrile and water containing 0.1%

acetic acid to a concentration of 10 ppm. The molecular weights

were determined by direct injection of 20 ml of the sample into

spectrometer (Applied Biosystems QSTAR XL fitted with Q/QTOF).

The analysis was done in the positive ion mode. Other MS settings

were as follows – flow rate: 10 ml/min; ion spray voltage: 5500 V;

nebulizer gas: 15 lb/in2; curtain gas: 20 lb/in2; declustering poten-

tial: 280 V; focusing potential: 265 V. The multiply charged pro-

tein peaks obtained in the mass spectrum were convoluted using

Bayesian Protein Reconstruct Software.

The internal peptide sequences of BGL II were determined by in-

gel digestion of the protein and sequencing of different peptides by

mass spectrometry. The product ion spectra were analyzed and

sequence was determined. The analysis was performed in the

Protein Chemistry Core Lab at Baylor College of Medicine, Hous-

ton, TX.

ResultsEnzyme isolationThe two enzymes BGL I and BGL II are cell wall bound proteins

[16]. The extraction of proteins from cell wall in an active form and

with high yield is a prerequisite for generating milligram quan-

tities of proteins. In general, for extraction of intracellular cell

components, both solid shear (high pressure homogenizers) and

liquid shear (bead mills) based methods have proven successful on

large scale [22,23]. Both these methods of cell disruption were

evaluated for the isolation of wall bound BGL I and BGL II and

enzyme released is shown in IU/l and IU/mg protein.

In the case of French press, the maximum release of enzyme was

observed with 20% cell suspension and highest specific activity of

1.10 IU/mg of protein was obtained. Because French press is sui-

table for volumes in the range of 5–50 ml, Dyno-Mill was used for

processing as this is generally a preferred method at large scale.

Different modes of operation were evaluated for enzyme release. In

the case of batch operation, a 20% (wcw/v) cell suspension was

made and maximum enzyme was released after agitating the cells

in the mill for 2 min. The extent of breakage was monitored

microscopically and estimated at 60–70%. In continuous opera-

tion, the cells were completely broken while in the case of dis-

continuous batch operation, the cells were found to be broken

partially. A comparison of these methods in terms of enzyme

release and specific activities is presented in Fig. 1. Highest recov-

ery was 840 IU/l and the specific activity was 6.4 IU/mg in the

discontinuous process, which was used in all subsequent opera-

tions. No colloidal glass from the beads was observed and thus the

material was suitable for the next stage.

Column chromatographiesThe cell-free extract obtained after centrifugation was directly

applied to Q-Sepharose FF anion-exchange column for both con-

centration (of protein of interest) and purification. This method is

also least expensive (compared to hydroxyapatite or affinity). A

pilot experiment was carried out on a small analytical (XK16/30,

bed height 30 cm) scale to determine optimum conditions for

elution. For this, a total protein of 250 mg (1180 units) was loaded

onto the column and elution was carried out at 200 mM and

130 mM KCl step gradients. The elution profile is shown in

Fig. 2a and b respectively. At 200 mM KCl, the eluted fractions

showed several contaminant proteins in addition to BGL I and

BGL II. Higher specific activity (27.47 IU/mg) was obtained with

130 mM KCl step gradient than with 200 mM KCl (15.02 IU/mg) at

a nearly equal (92%) recovery. The SDS-PAGE analysis from the

130 mM eluted fractions showed that a lot of contaminating

www.elsevier.com/locate/nbt 313

RESEARCH PAPER New Biotechnology �Volume 29, Number 3 � February 2012

0

200

400

600

800

DiscontinousContinousBatch

Enz

yme

activ

ity (I

U/l)

0

2

4

6

8

Spec

ific

activ

ity (I

U/m

g)

FIGURE 1

Total b-glucosidase yield (IU/l) and specific activity (-&-) in different celldisruption processes. For all processes, a 20% wet cell weight (wcw) was

taken for reporting activities.

FIGURE 2

Elution profile and 10% SDS-PAGE analysis of fractions from ion-exchange column X

right side is shown 10% SDS-PAGE analysis of fractions. Lane 1: Marker lane, Lanes 2–

carried out with a step gradient of 130 mM KCl. On the right side is shown 10% Sappearing after the application of 130 mM KCl step gradient.

314 www.elsevier.com/locate/nbt

Research

Pap

er

proteins were indeed removed (Fig. 2b). On the basis of this, a large

column (XK26/70, bed height 50 cm) was used and the effect of

protein loadings (0.5–3.0 g) was investigated on enzyme recovery.

A quick scaled down quantitative b-glucosidase assay was done in

an ELISA plate. The specific activity and enzyme recovery for each

case are shown in Fig. 3. As observed, with an initial increase in

enzyme loading, there was an increase in specific activity up to a

loading of 1.76 g. However, further increase led to a decrease in the

same. While enzyme recovery was higher at 2.2 g, more contam-

inating proteins led to lower specific activities. Analysis of relevant

fractions by SDS-PAGE (Fig. 4a–c) confirmed that the preparation

was cleaner at a loading of 1.76 g (Fig. 4b). It was also observed that

enzyme recovery was initially as low as 64% at this protein loading

and could be enhanced to nearly 95% by the addition of 10 mM

cellobiose in the feed.

The concentrated ion-exchange fractions (5 mg total protein),

showing b-glucosidase activity, when applied to Sephadex G 200

column (XK16/100) gave a superior purity (70 IU/mg) compared to

that obtained on Superdex 75 column (15 IU/mg) while maintain-

K16/30. (a) Elution was carried out with a step gradient of 200 mM KCl. On the

6: fractions 2–6 appearing after the application of 200 mM KCl. (b) Elution was

DS-PAGE analysis of fractions. Lane 1: Marker lane, Lanes 2–6: fractions 7–12

New Biotechnology �Volume 29, Number 3 � February 2012 RESEARCH PAPER

10

20

30

40

50

60

70

80

3.532.521.510.50Enzyme lo ading (g)

Spec

ific

activ

ity (I

U/m

g)

30

45

60

75

90

105

Enz

yme

reco

very

(%)

FIGURE 3

Specific activity (-&-) and enzyme recovery (-~-) at different protein

loadings in XK 26/70 preparative column. Elution was carried out with a step

gradient of 130 mM KCl.

ResearchPap

er

ing the same level of yield (50%). On the basis of these results, 30–

40 mg protein sample was resolved on a preparative XK 26/100

column and the results are shown in Fig. 5a and b. As seen in the

SDS-PAGE, the two b-glucosidases were the major proteins. The

final separation of BGL I and BGL II was achieved on phenyl

Sepharose column. Two peaks were obtained, one was during

the flow-through (peak I) while the other peak (peak II) was eluted

at 100% concentration of elution buffer (Fig. 6a). Both showed b-

glucosidase activity and when checked on SDS-PAGE, peak 1

corresponded to BGL II while peak 2 to BGL I (Fig. 6b). The final

specific activity of BGL I was 74 IU/mg and that of BGL II 32 IU/

mg. A complete summary of purification is given in Table 1.

Purity of eluted b-glucosidases and internal peptide sequencesof BGL IIThe purity and the exact molecular weight of the proteins were

determined from the Electrospray spectrum shown in Fig. 7a (BGL

I) and b (BGL II). Typically, a protein will carry one charge per

1000 Da, or less if there are very few basic residues. As seen in Fig. 7,

multiply charged ions were formed for both the proteins and the

relative intensities were higher for BGL I than for BGL II. This was

advantageous as it improves the sensitivity at the detector and

allows a good analysis of high molecular weight molecules.

The presence of successive peaks differing from each by a single

charge confirmed the purity of the proteins. The data from both

the figures were convoluted by transformation of multiply charged

peaks in the spectrum using the algorithm in the mass spectro-

meter and molecular weight was determined as 97,000 for BGL I

and 94,000 for BGL II. The internal peptide sequences were deter-

mined for BGL II and were LSLSWPFK. . ., . . .NSDNPAYLNFHTER. . ..,

. . .EIYLEPFR. . ., . . .NDI/LDVTR. . .., . . .DQAVYF/MQVTPTR. . .. Two

of these (shown in bold) matched with the internal peptide

sequences of many Family 3 b-glucosidases.

DiscussionThis is the first report on the preparative scale approach for

resolution and purification of two closely related b-glucosidases.

The proteins are the cell wall bound enzymes of P. etchellsii and

belong to Family 3 of the glycosylhydrolase families. Relatively

little is known about this family and only one crystal structure is

available which is of a 657 aa enzyme [15]. Because the Family 3 b-

glucosidases occur as closely related isozymes, particularly in yeast,

their purification with concomitant high yield still remains a

challenge. In this work, we have introduced new strategies based

on the prior knowledge of substrate specificities of these enzymes

to resolve and purify them with high yield.

Breakage of yeast cells has been tried with solid shear and liquid

shear. Continuous shear, through the use of Dyno-Mill, is com-

mon for the release of enzymes from yeast cells [24] particularly for

large scale work. This is a larger device in which beads are agitated

by rapidly rotating discs. Because heat generation is a problem

with this device, adequate care has to be taken to maintain the

temperature at 48C. In our study, discontinuous liquid shear was

most successful with high recovery and very high specific activity

of b-glucosidases. Through this method, it is anticipated that

during the first two passes, as the flow rate is small, the cells

experience high shear to release the proteins. High flow rates kept

in the subsequent passes do not impart very high shear leading to

cells only being partially broken, which was observed under these

conditions. An increase in the release of total proteins is likely to

be due to stripping of proteins from the cell surface which also

explains increase in specific activity of b-glucosidases through this

method. In contrast to our previous work [16] where ammonium

sulphate precipitation was used as the first step in purification,

larger volumes of cell-free extract from Dyno-Mill (2–3 L), pre-

sented the problem of adding kilograms of ammonium sulphate.

This was circumvented by directly loading cell-free extract onto

anion-exchange column. It served dual purpose of eliminating the

use of DNase I (which is used for reducing the viscosity of the

solution) and preventing aggregation of the active protein by

maintaining low protein concentration in the cell-free extract

(0.5–0.7 mg/ml).

Because this is a high capacity technique, it was successfully

used to concentrate and purify proteins during the early stage.

Very large volumes could be loaded on to the column. Loading of

proteins at low concentrations minimized localized pH and ionic

strength fluctuations through rapid counter-ion release. This is

likely to have contributed to high recovery (above 60%) at this

stage. Recovery of the b-glucosidases from the column was further

increased to >90% by the addition of cellobiose to the cell-free

extract before loading. It served the purpose of ‘locking’ the active

site and thus prevented activity loss that may occur on account of

unfolding. We also used the strategy of absorbing the protein on

ion-exchange first followed by desorption as it allows a greater

degree of protein fractionation during the elution step. The fact

that it is also less expensive (when compared to hydrophobic

interaction, hydroxyapatite, metal chelate, covalent binding chro-

matography) justified its use at the early stage of purification.

Elution of b-glucosidases from ion-exchange matrix was

achieved by step gradient. This was first evaluated on a small

analytical column wherein b-glucosidases eluted around 200

and 130 mM KCl [Fig. 2a and b]. Higher specific activity

(27.4 IU/mg) with a yield of 88% was obtained at 130 mM KCl

and analysis of gel data indicated a purer preparation. It is also

important to note that nearly 60% of the column capacity was

www.elsevier.com/locate/nbt 315

RESEARCH PAPER New Biotechnology �Volume 29, Number 3 � February 2012

FIGURE 4

Elution profile and 10% SDS-PAGE analysis from ion-exchange preparative column (XK26/70) at a protein loading of (a) 1.1 g, (b) 1.76 g and (c) 3.3 g. For all

protein loadings shown, elution was carried out with a step gradient of 130 mM KCl. Fractions loaded in the gel are numbered as they appear from the column(see the figure on the left in each case).

Research

Pap

er

taken up by the cell-free extract leaving about 40% for effective

separation. This worked well as step elution was being employed

which requires less volume. Two problems that normally occur

due to step elution viz. co-elution of other proteins resulting in the

loss of selective desorption and increase in partition coefficient,

316 www.elsevier.com/locate/nbt

which occurs as a protein is eluted so that a fraction of the protein

of interest is still bound, were largely avoided. Only a slight tailing

was observed. These results are better than our previous study [16]

wherein much lower yields were reported. In the scale-up opera-

tion on a preparative XK 26/70 column, the results could be

New Biotechnology �Volume 29, Number 3 � February 2012 RESEARCH PAPER

FIGURE 5

Elution profile and 10% SDS-PAGE analysis from Sephadex G 200 (XK26/100) column. On the left side are the data for OD280 (-*-) and b-glucosidase activity (-~-).Fractions testing positive for b-glucosidase activity (24–31) were analyzed on 10% SDS-PAGE (shown on the right side). M: Marker lane, IEX: pooled sample from

the previous ion-exchange chromatography step.

ResearchPap

er

reproduced. The loss in recovery of b-glucosidase (about 40% loss)

in this scale-up could be circumvented by the addition of 10 mM

cellobiose, a natural substrate of b-glucosidase. This inclusion led

to nearly 95% recovery of the enzyme. Thus efficient trapping of b-

glucosidases by the Q-Sepharose matrix was achieved. This might

be due to higher percentage of acidic residues on the surface of

these proteins [16]. The effect of protein loading on the recovery of

b-glucosidases was also studied. The results indicated that higher

amounts of b-glucosidases adsorbed to the matrix with an increase

in protein loading but nonspecific elution at higher loadings lead

to decrease in specific activity. The removal of the lower molecular

weight contaminants was achieved with Sephadex G 200 gel

filtration which gave better purification than Superdex 75.

Although the lower molecular weight proteins were successfully

removed in this step, there was about 50% loss in the recovery of

BGL I and BGL II. It is possible that the loss in b-glucosidase

activity was due to the removal of some contaminant proteins

FIGURE 6

(a) Elution profile from phenyl Sepharose column. (b) 10% SDS-PAGE analysis of

purified BGL II, appeared in fraction # 7. Peak II, corresponded to purified BGL I,

which might be providing non-specific force for maintaining the

native structure of the proteins. This was supported by the fact that

purified BGL I was stabilized against denaturation by the addition

of bovine serum albumin which might be providing the crowding

effect for stability. Multi-domain proteins are known to be stabi-

lized by molecular crowding effect [25]. Glycerol was found to be

more effective than BSA (data not shown). However, BGL II

appeared to be quite stable even in the absence of any crowding

agent. This may be explained in terms of its ability to tetramerize

which might be providing the ‘self crowding’ effect to the poly-

peptide chains. It is proposed that crowding provides a stabilizing

effect to the folded protein because of indirect compaction of the

denatured states [26–30].

The final resolution of BGL I and BGL II was attempted by

hydroxyapatite or phenyl sepharose chromatography. In the for-

mer method, low recovery (around 15–20%) was observed. The

major disadvantage with the use of this matrix is that it has a

fractions (7, 35–41) showing b-glucosidase activity. Peak I, corresponding towhich appeared in fractions # 35–41.

www.elsevier.com/locate/nbt 317

RESEARCH PAPER New Biotechnology �Volume 29, Number 3 � February 2012

FIGURE 7

Electrospray spectrum of purified BGL I (a) and BGL II (b). The m/z (in Th) and the number of charges are indicated on each peak. The molecular mass was

determined as 97,000 Da for BGL I and 94,000 Da for BGL II.

Research

Pap

er

tendency to absorb carbon dioxide which can make a hard crust as

a layer on the column top. This leads to an increase in the back

pressure reducing the flow rate. Hydrophobic interaction chroma-

tography using phenyl sepharose was thus tried. It served both as

an affinity matrix for adsorption of the two enzymes, as BGL I and

318 www.elsevier.com/locate/nbt

BGL II are bound to the cell wall and do have hydrophobic patches,

as well as selective separation based on their preferences for the

aryl groups. BGL I shows preferential activity on aryl substrates

[16] and thus should bind to this matrix strongly and under

conditions of elution, BGL II should elute first. This strategy

New Biotechnology �Volume 29, Number 3 � February 2012 RESEARCH PAPER

TABLE 1

Summary of purification of BGL I and BGL II from Pichia etchellsii

Fraction Volume(ml)

Activity(IU/ml)

Proteinconcentration(mg/ml)

Specificactivity(IU/mg)

Total units(IU/ml)

Total protein(mg)

Yield(%)

Cell free extract 2000 4.21 1.10 3.81 8400 2200 100

Ion exchange 150 58.85 2.10 28 7980 315 95

Gel filtration 132 32.03 0.57 22 4229 75 53

301* 10*

Phenyl Sepharose

BGL I 12 10.62 0.143 74 164 1.70 54

BGL II 3 12.24 0.380 32 1.15

* Sample applied to phenyl sepharose column.

ResearchPap

er

worked and led to good and efficient separation of the two

enzymes. BGL I was eluted with much higher specific activity of

74 IU/mg as opposed to 35 IU/mg obtained with hydroxyapatite

[16]. Similarly, BGL II was also recovered to a specific activity of

32 IU/mg as opposed to 13.6 IU/mg reported earlier. It is also

possible that the hydrophobic matrix favored retention of enzyme

conformation during the process of elution. Although affinity

based matrices have been used for the purification of b-glucosi-

dases, these are limited to heterologous expression in Escherichia

coli employing His-Tag [31,32]. We also propose that phenyl

sepharose can be used as an affinity matrix for the separation of

isozymes based on their differential specificity towards aryl sub-

strates.

The high purity of the enzymes was confirmed by electronspray

ionization spectrum wherein ‘clean’ successive multiply charged

ions were obtained for the two enzymes. In this study, we also

identified a novel Family 3 glucosidase, namely BGL II; several

peptide sequences of which were novel. Availability of these

enzymes on large scale is likely to lead to a greater understanding

of Family 3 enzymes.

AcknowledgementsDepartment of Science and Technology (Govt. of India) is thanked

for providing financial support to one of the authors (SM) for

carrying out this work. The doctoral fellowship to MAS by Indian

Institute of Technology Delhi is gratefully acknowledged.

References

1 Bhatia, Y. et al. (2002) Microbial b-glucosidases: cloning, properties and

applications. Crit. Rev. Biotech. 22, 375–407

2 Morant, A.V. et al. (2008) b-Glucosidases as detonators of plant chemical defense.

Phytochemistry 69, 1795–1813

3 Barnett, C.C. et al. (1991) Cloning and amplification of the gene encoding an

extracellular b-glucosidase from Trichoderma reesei: evidence for improved rates of

saccharification of cellulosic substrates. Bio/Technology 9, 562–567

4 Bothast, R.J. and Saha, B.C. (1997) Ethanol production from agricultural biomass

substrates. Adv. Appl. Microbiol. 44, 261–286

5 Choa, K.M. et al. (1999) Integration of endo/exoglucanase and b-glucosidase genes

into the yeast chromosomes for direct conversion of cellulose to ethanol. Enzyme

Microb. Technol. 25, 23–30

6 Mamma, D. et al. (2004) Biochemical and catalytic properties of two intracellular

b-glucosidases from the fungus Penicillium decumbens. J. Mol. Catal. B 27, 183–190

7 Vasserot, Y. et al. (1995) Monoterpenyl glucosides in plants and their

biotechnological transformation. Acta Biotechnol. 15, 77–95

8 Winterhalter, P. and Skouroumounis, G.K. (1997) Glycoconjugated aroma

compounds: occurrence, role and biotechnological transformation. Adv. Biochem.

Eng. Biotechnol. 55, 73–105

9 Crout, D.H. and Vic, G. (1998) Glycosidases and glycosyl transferases in glycoside

and oligosaccharide synthesis. Curr. Opin. Chem. Biol. 2, 98–111

10 Hancock, S.M. et al. (2006) Engineering of glycosidases and glycosyltransferases.

Curr. Opin. Chem. Biol. 10, 509–519

11 Wang, L.-X. and Huang, W. (2010) Enzymatic transglycosylation for

glycoconjugate synthesis. Curr. Opin. Chem. Biol. 13, 592–600

12 Sharon, N. (2006) Carbohydrates as future anti-adhesion drugs for infectious

diseases. Biochim. Biophys. Acta 1760, 527–537

13 Huang, C. et al. (2006) Carbohydrate microarrays for profiling the antibodies

interacting with Globo H tumor antigen. Proc. Nat. Acad. Sci. U. S. A. 103,

15–20

14 Ratner, D.M. et al. (2004) Probing protein carbohydrate interactions with

microarrays of synthetic oligosaccharides. Chem. Biol. Chem. 5, 379–383

15 Varghese, J.N. et al. (1997) Three dimensional structure of barley b-D-glucan

exohydrolase, a family 3 glycosylhydrolase. Structure Fold. Des. 7, 179–190

16 Wallecha, A. and Mishra, S. (2003) Purification and characterization of two b-

glucosidases from a thermo-tolerant yeast Pichia etchellsii. Biochim. Biophys. Acta

1649, 74–84

17 Bhatia, Y. et al. (2002) Biosynthetic activity of Escherichia coli expressed Pichia

etchellsi b-glucosidase II. Appl. Biochem. Biotech. 102–103, 367–380

18 Bachhawat, P. et al. (2004) Enzymatic synthesis of oligosaccharides, alkyl and

terpenyl glucosides by recombinant Escherichia coli expressed Pichia etchellsii b-

glucosidase. Appl. Biochem. Biotech. 118, 269–282

19 Rather, M.Y. et al. (2010) b-Glucosidase catalysed synthesis of octyl-b-D-

glucopyranoside using whole cells of Pichia etchellsii in microaqueous media. J.

Biotechnol. 150, 490–496

20 Pandey, M. and Mishra, S. (1995) Cloning and expression of b-glucosidase gene

from the yeast Pichia etchellsii. J. Ferment. Bioeng. 80, 446–453

21 Bradford, M.M. (1976) Rapid and sensitive method for quantification of

microgram quantities of proteins utilizing the principle of protein dye binding.

Anal. Biochem. 72, 248–254

22 Chisti, Y. and Young, M. (1986) Disruption of microbial cells for intracellular

products. Enzyme Microb. Technol. 8, 194–204

23 Pingoud, A. et al. (2002) Biochemical Methods: A Concise Guide for Students and

Researchers. Wiley-VCH

24 Melendres, A.V. et al. (1991) A kinetic analysis of cell disruption by bead mill.

Bioseparation 2, 231–236

25 Ellis, R.J. (2001) Macromolecular crowding: obvious but underappreciated. Trends

Biochem. Sci. 26, 597–604

26 Berg, B. et al. (1999) Effects of macromolecular crowding on protein folding and

aggregation. EMBO J. 18, 6927–6933

www.elsevier.com/locate/nbt 319

RESEARCH PAPER New Biotechnology �Volume 29, Number 3 � February 2012

Research

Pap

er

27 Rivas, G. et al. (2004) Life in a crowded world: workshop on the biological

implications of macromolecular crowding. EMBO J. 5, 23–27

28 Zimmerman, S.B. and Minton, A.P. (1993) Macromolecular crowding:

biochemical, biophysical and physiological consequences. Ann. Rev. Biophys.

Biomol. Struct. 22, 27–75

29 Minton, A.P. (2005) Influence of macromolecular crowding upon the stability and

state of association of proteins: predictions and observations. J. Pharm. Sci. 94,

1668–1675

320 www.elsevier.com/locate/nbt

30 Stagg, L. et al. (2007) Molecular crowding enhances native structure and

stability of a/b protein flavodoxin. Proc. Natl. Acad. Sci. U. S. A. 104,

18976–18981

31 Opassiri, R. et al. (2004) b-Glucosidase, exo-b-glucanase and pyridoxine

transglucosylase activities of rice B Glu1. Biochem. J. 379, 125–131

32 Seidle, H.F. et al. (2006) Trp-49 of the family 3 b-glucosidase from Aspergillus niger is

important for its transglucosidic activity: creation of novel b-glucosidases with low

transglucosidic efficiencies. Arch. Biochem. Biophys. 445, 110–118