Mutagenesis of key amino acids alters activity of a Saccharomycescerevisiae endo-polygalacturonase expressed in Pichia pastoris

Pilar Blanco a;1, Graham Thow a, Craig G. Simpson a, Tomas G. Villa b,Brian Williamson a;�

a Scottish Crop Research Institute, Invergowrie, Dundee DD2 5DA, UKb Departamento de Microbiolog|¤a, Facultad de Farmacia, Universidad de Santiago de Compostela, 15706 Santiago de Compostela, Spain

Received 11 December 2001; received in revised form 14 February 2002; accepted 7 March 2002

First published online 11 April 2002

Abstract

A polygalacturonase (PG)-encoding gene from Saccharomyces cerevisiae (PGU1) was successfully expressed in the methylotrophic yeastPichia pastoris. PG secretion was efficiently directed by the S. cerevisiae K-factor signal sequence, while the native (PGU1) leader peptidewas unable to direct protein export in P. pastoris. The level of PGU1 activity achieved in P. pastoris was significantly enhanced whencompared to activity using the same gene in S. cerevisiae. Expression of PG proteins, engineered by site-directed mutagenesis, in P.pastoris showed that aspartic acid residues at positions 179, 200, and 201, and histidine 222 were essential for enzyme activity. Mutationof the two potential glycosylation sites in PGU1 showed that the two residues individually (N318D, N330D) did not affect secretedenzyme activity, but the double mutant caused a 50% reduction in enzyme activity when compared to the wild-type PGU1transformant. : 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.

Keywords: Endo-polygalacturonase; Protein expression; Directed mutagenesis ; Saccharomyces cerevisiae ; Pichia pastoris

1. Introduction

Pectic substances are the major constituent of primarycell walls and the middle lamella of higher plant cells [1].Pectin is a complex heteropolysaccharide composedmainly of D-galacturonic acid residues joined by K-1,4-linkages that forms a homogalacturonan skeleton. Thisbackbone structure (‘smooth regions’) alternates withbranched regions (‘hairy regions’) that contain rhamnose.The rhamnogalacturonan chains consist of L-rhamnoseand D-galacturonic acid linked by K-1,2-bonds [2]. Sidechains of neutral sugar are usually attached to rhamnosevia K-1,4-linkages and the D-galacturonic acid residues canbe methylated. Because of this complexity, a wide range of

enzymes called pectinases are necessary for degradation ofpectic substances.

Pectinases are classi¢ed into esterases and depolymeris-ing enzymes: esterases remove the methoxyl residues. Thedepolymerising enzymes are divided into hydrolases andlyases. Endo-polygalacturonases (3.2.1.15) (endo-PGs) arepectic enzymes that catalyse the random hydrolysis ofK-1,4-D-glucosidic bonds between two non-methylated ga-lacturonic acid residues, whereas exo-PGs (3.2.1.67) re-move terminal residues [3]. Endo-PG production hasbeen widely reported in plants and microorganisms. PlantPGs are important in fruit ripening and senescence [4],whereas the microbial enzymes are deployed during infec-tion of plants [5] or substrate colonisation in saprophyticgrowth.

The presence of pectinolytic enzymes has been reportedin yeast species and several strains of Saccharomyces cere-visiae produce an endo-PG [6^8]. Recently, the genes en-coding endo-PGs have been cloned and characterised fromS. cerevisiae and Kluyveromyces marxianus [9^13]. Aminoacid sequences derived from these genes show high simi-larity with each other and also with endo-PGs from ¢la-mentous fungi. The similarity was particularly high in the

0378-1097 / 02 / $22.00 : 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.PII: S 0 3 7 8 - 1 0 9 7 ( 0 2 ) 0 0 6 3 4 - 1

* Corresponding author. Tel. : +44 (1382) 562731;Fax: +44 (1382) 562426.

E-mail address: b.williamson@scri.sari.ac.uk (B. Williamson).

1 Present address: Departamento de Microbiolog|¤a, Facultad deBiolog|¤a, Universidad de Extremadura, Avda. de Elvas s/n, 06071Badajoz, Spain.

FEMSLE 10454 23-5-02

FEMS Microbiology Letters 210 (2002) 187^191

www.fems-microbiology.org

region where the putative catalytic site is located and mu-tation of these highly conserved residues abolished PGactivity [14,15].

The ability of endo-PGs to degrade pectic substances isused in the industrial processing of fruits and vegetablesand speci¢cally in the beverage industry where pectinscause problems in ¢ltration and clari¢cation. The mainsource for commercial pectinases is Aspergillus niger, butthis fungus secretes a mixture of several enzymes and someof them have undesirable activities, such as arabinofura-nosidases that cause haze formation [16] or esterases thatrelease methanol. It has recently been reported [17] thatthe use of endo-PG over-producer S. cerevisiae strains inoenological fermentations resulted in an improvement ofwine ¢ltration and reduced levels of methanol. Yeasts likeS. cerevisiae that secrete few enzymes and produce mainlyendo-PG could provide an alternative source for com-mercial pectinases. However, their productivity is lowcompared to those reported for fungi. Pichia pastoris isa useful system for heterologous over-expression of glyco-sylated eukaryotic proteins. In addition, the Pichia systemis well developed for industrial-scale fermentation and itcan be easily grown to high cell density in relatively inex-pensive media. In this work we report the expression ofthe endo-PG gene (PGU1) from S. cerevisiae in P. pastoris.This system was used to study, by site-directed mutagen-esis, the importance of the aspartic acid and histidine res-idues at the putative S. cerevisiae endo-PG catalytic siteand also to investigate the role of two putative glycosyla-tion sites on enzyme activity.

2. Materials and methods

2.1. Strains and plasmids

Yeast strains P. pastoris X-33 and KM71 and the ex-pression vectors pPICZB and pPICZKB were obtainedfrom Invitrogen. These plasmids are identical except thatthe latter contains a S. cerevisiae K-factor signal sequencefor secretion of expressed product into the medium. Thebacterial strains used were Escherichia coli DH5K (LifeTechnologies) and XL1-Blue (Stratagene). S. cerevisiae1389-8b+pYES-PGU1 was as described previously [9].

2.2. Media composition and growth conditions

Bacterial and yeast cells were cultured using protocolslisted in the manual for the EasySelect Pichia Expressionkit (Invitrogen). Bacterial clones were grown at 37‡C over-night in low-salt LB medium with zeocin (25 Wg ml31).Parental P. pastoris strains were maintained on YPD me-dium. P. pastoris expression cultures (KM71 and X-33transformants) were grown for 2 days, with vigorous shak-ing, in BMGY or MGY medium supplemented with his-tidine (40 Wg ml31) when it was required. Cells were col-

lected by centrifugation and re-suspended with BMMY orMM medium respectively, containing methanol as the in-ducer. S. cerevisiae strain was grown in SD medium sup-plemented with the required amino acids [18]. After cen-trifugation the cells were re-suspended in SD with 2%galactose as carbon source and as inducer. All yeast cul-tures were grown at 28‡C.

2.3. DNA manipulation and cloning

Standard methods for performing restriction digests, li-gation reactions and plasmid isolations were used [19].DNA fragments were puri¢ed from agarose gels with thePrep-a-Gene kit (Bio-Rad) using the manufacturer’s pro-tocols. Constructs were sequenced using the Dye Termi-nator Cycle Sequencing kit (Perkin Elmer) and productsrun on an ABI 377 sequencer. Site-directed mutagenesiswas carried out on pPICS2 using the QuickChange Site-Directed Mutagenesis kit (Stratagene) following the man-ufacturer’s instructions. All oligonucleotides were obtainedfrom MWG-Biotech and a list of primers used is availableon request. All mutant constructs were con¢rmed by DNAsequencing.

2.4. Expression constructs pPICS1 and pPICS2

A KpnI/NotI fragment containing the entire S. cerevisiaePGU1 cDNA was isolated from plasmid pBSK-PGU1 [9]and cloned into pPICZB to generate the pPICS1 con-struct. The pPICS2 construct fused the mature PGU1coding sequence in-frame with the K-factor signal se-quence of pPICZKB. The PGU1 fragment was generatedby PCR with Pfu polymerase (Promega) using pBSK-PGU1 as template and the oligonucleotides Scpg1 (5P-ACACCTTTGTCAAAAAGAGATTC-3P) and T3a (5P-CCCTCACTAAAGGGAACAAAAGC-3P). This PCRproduct was cut with XbaI and ligated with PmlI/XbaI-digested vector.

2.5. Transformation of P. pastoris

Transformation of P. pastoris by electroporation wasperformed using the Invitrogen protocol. ConstructspPICS1 and pPICS2 (10 Wg) were linearised with SacIand electroporated using a Bio-Rad Gene Pulser. P. pas-toris X-33 (wild-type, Mutþ) and KM71 (Muts) weretransformed with each construct and selected on YPDSplates with zeocin (100 Wg ml31). Positive transformedcolonies were con¢rmed by direct PCR screening of cells.Transformants were cultured in liquid medium and in-duced with methanol. To maintain the concentration ofinducer, methanol (0.5%, v/v) was added every 24 h.

2.6. Assay of PG activity

Staining of polygalacturonic acid (PGA) breakdown

FEMSLE 10454 23-5-02

P. Blanco et al. / FEMS Microbiology Letters 210 (2002) 187^191188

products with ruthenium red [20] was used to detect PGactivity of cells grown on agarose plates. Cells were re-moved from the culture medium by centrifugation after24 h and the activity of PGs in solution was measuredby a modi¢ed Nelson^Somogyi assay [21] using PGA(0.4%, w/v) in 50 mM sodium acetate bu¡er (pH 5.2) assubstrate. A typical reaction contained 500 Wl of sampleplus 500 Wl of substrate and was incubated for 1 h at 37‡C.

3. Results and discussion

3.1. PG secretion

To identify clones expressing extracellular endo-PG ac-tivity, a series of ¢ve positive clones and two controlstrains were grown on agarose plates containing PGA todetect PG secretion (Fig. 1). There was no evidence ofextracellular pectolytic activity in the pPICS1 transform-ants (Fig. 1A). However, all of the pPICS2 transformantstested were able to hydrolyse the PGA in the medium asindicated by the presence of a pink halo after staining withruthenium red (Fig. 1B). These results demonstrate thatthe S. cerevisiae PGU1 gene can be expressed in P. pasto-ris, but only when the gene was fused to the S. cerevisiaeK-factor signal sequence. The native (PGU1) signal se-quence was unable to secrete PG in P. pastoris.

Parental strains and the PICS2 clones that showed pec-tolytic ability on plates were cultured in liquid media toquantify their activity over time. They were grown initiallyin BMGY medium for 2 days, the cells were then recov-ered by centrifugation and re-suspended in the inductionmedium BMMY with methanol (2%, v/v) and incubatedover 192 h. Samples were taken at 4, 8, 24, 48, 72 and192 h post-induction (Fig. 2).

All transformants showed PG activity after inductionfor 8 h and the highest activity was reached in 4 days inKM71 clones and between 24 and 48 h for X-33 clones.Both Pichia strains reached a similar maximum PG activ-

ity, but di¡ered in their overall activity pro¢le. Activity inthe X-33 clones decreased close to zero after 3 days post-induction, while PG activity was maintained at a high levelin the KM71 clones (Fig. 2). The di¡erence in expressionof the S. cerevisiae PGU1 between these two strains maybe due to their Mut phenotype. Although X-33 (Mutþ)cells grew better on methanol and consequently reachedexponential phase earlier than KM71 (Muts) cells, thesecells are susceptible to high levels of methanol which canlead, in the presence of oxygen, to toxic levels of form-aldehyde. The slower growth found in KM71 cells alloweda rapid accumulation of PG activity that was maintainedover time.

The initial methanol concentration used to induce pro-tein expression a¡ected the levels of PG activity. When theinitial concentration of methanol was 0.5% (v/v), the levelof PG activity in X-33 cells increased at a slower ratecompared to cells induced with 2% (v/v) methanol (Fig.3). However, at 48 h post-induction there was no di¡er-ence in the activity obtained from the di¡erent cultures.For KM71 clones, lower methanol concentration showed

Fig. 1. Detection of extracellular PG production by recombinant P. pastoris X-33 containing pPICS1 (A) and pPICS2 (B) plasmids. Five transformantsof each construct were grown on medium containing PGA, with subsequent ruthenium red staining of breakdown products. C3 (strain X-33 non-trans-formed) and Cþ (X-33 transformed with pPICS2) are the negative and positive control respectively.

Fig. 2. PG activity of pPICS2 transformants measured over a timecourse of 192 h. Enzyme activity was measured in culture supernatantsof X-33 (Mutþ) and KM71 (Muts) transformants grown in BMMY. PGactivity was expressed in Wmol ml31 h31.

FEMSLE 10454 23-5-02

P. Blanco et al. / FEMS Microbiology Letters 210 (2002) 187^191 189

a negative e¡ect on the level of PG activity after 32^48 h(Fig. 3). These results again are likely to be caused by theMut phenotype of the respective strains. Therefore, KM71transformants induced with 2% methanol gave the highestlevels of PG activity over a wider time scale (Figs. 2 and3).

The results con¢rmed that P. pastoris produced highlevels of S. cerevisiae PG activity. To compare the activitywith that obtained for the same gene over-expressed inS. cerevisiae, clones from both systems were analysed.The value reached with P. pastoris (261.82 Wmol ml31

h31) was about 100 times higher than that found for S.cerevisiae (2.68 Wmol ml31 h31). The signi¢cantly in-

creased production observed in P. pastoris clones indicatesthis yeast has potential application as an alternative sourceof PG in the food processing industry.

3.3. Site-directed mutagenesis of S. cerevisiae PGU1

A series of PGU1 site-directed mutants were producedthat were tested for PG activity in P. pastoris. A quartet ofhighly conserved amino acid residues at positions 179, 200,201, and 222, identi¢ed as having a role in endo-PG ac-tivity in ¢lamentous fungi [15], were mutated. Initially thewild-type residues were altered from aspartate to aspara-gine, and the histidine to alanine (D179N, D200N,D201N, H222A). In each case PG activity was undetect-able in either plate assay (Fig. 4, M1^M4) or PG assay ofculture supernatant (Table 1). These results support theimportance of these residues for catalytic activity in bud-ding yeast, as found for PGs from ¢lamentous fungi [14].Conservative mutation of the aspartate at position 201 to

Fig. 3. E¡ect of methanol concentration on PG activity. PICS2 trans-formants of X-33 or KM71 were induced in BMMY medium containingeither 0.5 or 2.0% methanol (v/v). Samples of the culture supernatantwere analysed for PG activity. PG activity was expressed in Wmol ml31

h31.

Table 1PG activity of wild-type and mutant S. cerevisiae PGU1

Strain (mutation) PG activity on plate PG activity (Wmol ml31 h31)

Mutations targeting active siteX-33-PICS2 (non-mutated) + 58.35X-33-PICS2-M1 (D179N) 3 ^X-33-PICS2-M2 (D200N) 3 ^X-33-PICS2-M3 (D201N) 3 ^X-33-PICS2-M4 (H222A) 3 ^X-33-PICS2-M9 (D201E) 3 ^X-33-PICS2-M10 (H222R) 3 ^Mutations targeting putative glycosylationX-33-PICS2-M5 (N318D) + 52.13X-33-PICS2-M6 (N330D) + 58.46X-33-PICS2-M7 (N318D/N330D) + 29.22KM71-PICS2 (non-mutated) + 48.93KM71-PICS2-M7 (N318D/N330D) + 6.18

Extracellular PG activity was detected on plates by ruthenium red staining (+/3), PG activity was measured in culture supernatants.

Fig. 4. E¡ect of mutating PGU1 on PG activity. PG secreted fromtransformants of X-33 was detected by ruthenium red staining of PGAbreakdown products. The mutants are M1 (D179N), M2 (D200N), M3(D201N), M4 (H222A), M5 (N318D), M6 (N330D), M7 (N318D/N330D), M9 (D201E), and M10 (H222R).

FEMSLE 10454 23-5-02

P. Blanco et al. / FEMS Microbiology Letters 210 (2002) 187^191190

a glutamate and histidine at position 222 to arginine(D201E, H222R) also did not show any PG activity inP. pastoris transformants (Fig. 4, M9 and M10, Table1). This ¢nding indicates that the charge of the residuemay not be so important as the disposition of the aminoacid side chain and its in£uence in the formation of thecatalytic domain.

There are two potential glycosylation sites in PGU1 atresidues 318 and 330, these residues were mutated to in-vestigate their role in PG activity. Mutating the individualsites separately from asparagine to aspartate (N318D,N330D) did not a¡ect the level of PG activity in culturesupernatants (Fig. 4, M5 and M6, Table 1). When bothsites were mutated together, the recombinants hydrolysedPGA on plates (Fig. 4, M7). However, PG activity in liq-uid media was approximately half that of the wild-typeenzyme and half that of the above single mutants ex-pressed in P. pastoris X-33, and eight-fold less in KM71transformants expressing the PG double mutant.

Mutation of the putative glycosylation sites singly hasno e¡ect on enzymatic activity, mutating both sites togeth-er, however, does have a major e¡ect. Although glycosyl-ation is a feature of PGs neither their number nor theirposition appears to be conserved. In the case of this dou-ble mutant it is possible that changing two amino acids inclose proximity had a more deleterious e¡ect on enzymestructure such that catalytic activity and possibly proteinstability were compromised. The production of Pichia-ex-pressing lines demonstrates the utility of this system to testfor the development of PG-derived sequences with higherPG activity pro¢les.

Acknowledgements

P.B. held a fellowship from Fundacio¤n Ramo¤n Areces(Madrid, Spain). This work has been partially supportedby Scottish Executive Environment and Rural A¡airs De-partment (SEERAD).

References

[1] Willats, W.G.T., McCartney, W., Mackie, W. and Knox, J.P. (2001)Pectin: cell biology and prospects for functional analysis. Plant Mol.Biol. 47, 9^27.

[2] Schols, H.A. and Voragen, A.G.J. (1996) Complex pectins: structureelucidation using enzymes. P. 3 19. In: Progress in Biotechnology,Vol. 14. Pectin and Pectinases (Visser, J. and Voragen, A.G.J., Eds.),Elsevier Science, Amsterdam.

[3] Fogarty, W.M. and Kelly, C.T. (1983) Pectic enzymes. In: Microbial

Enzymes and Biotechnology (Fogarty, W.M., Ed.), pp. 131^182. Ap-plied Science Publishers, London.

[4] Had¢eld, K.A. and Bennet, A.B. (1998) Polygalacturonases: manygenes in search of a function. Plant Physiol. 117, 337^343.

[5] Lang, C. and Do«rnenburg, H. (2000) Perspectives in the biologicalfunction and the technological application of polygalacturonases.Appl. Microbiol. Biotechnol. 53, 366^375.

[6] Blanco, P., Sieiro, C. and Villa, T.G. (1999) Production of pecticenzymes in yeasts. FEMS Microbiol. Lett. 175, 1^9.

[7] Blanco, P., Sieiro, C., Diaz, A. and Villa, T.G. (1994) Production andpartial characterization of an endopolygalacturonase from Saccharo-myces cerevisiae. Can. J. Microbiol. 40, 974^977.

[8] Gainvors, A., Frezier, V., Lemarsquier, H., Lequart, C., Aigle, M.and Belarbi, A. (1994) Detection of polygalacturonase, pectin-lyaseand pectin-esterase activities in a Saccharomyces cerevisiae strain.Yeast 10, 1311^1319.

[9] Blanco, P., Sieiro, C., Reboredo, N.M. and Villa, T.G. (1998) Clon-ing, molecular characterization, and expression of an endo-polygalac-turonase-encoding gene from Saccharomyces cerevisiae IM1-8b.FEMS Microbiol. Lett. 164, 249^255.

[10] Hirose, N., Kishida, M., Kawasaki, H. and Sakai, T. (1998) Molec-ular cloning of a polygalacturonase gene in Saccharomyces cerevisiae.J. Ferment. Bioeng. 86, 332^334.

[11] Gognies, S., Gainvors, A., Aigle, M. and Belarbi, A. (1999) Cloning,sequence analysis and overexpression of a Saccharomyces cerevisiaeendopolygalacturonase-encoding gene (PGL1). Yeast 15, 12^22.

[12] Siestele, R., Bartkeviciute, D. and Sasnauskas, K. (1999) Cloning,targeted disruption and heterologous expression of the Kluyveromycesmarxianus endopolygalacturonase gene (EPG1). Yeast 15, 311^322.

[13] Jia, J. and Wheals, A. (2000) Endopolygalacturonase genes and en-zymes from Saccharomyces cerevisiae and Kluyveromyces marxianus.Curr. Genet. 38, 264^270.

[14] van Santen, Y., Benen, J.A.E., Schro«ter, K.H., Kalk, K.H., Armand,S., Visser, J. and Dijkstra, W. (1999) 1.68-AX crystal structure ofendopolygalacturonase II from Aspergillus niger and identi¢cationof active site residues by site-directed mutagenesis. J. Biol. Chem.274, 30474^30480.

[15] Armand, S., Wagemaker, M.J.M., Sanchez-Torres, P., Kester,H.C.M., van Santen, Y., Dijkstra, B.W., Visser, J. and Benen, A.E.(2000) The active site topology of Aspergillus niger endopolygalactur-onase II as studied by site-directed mutagenesis. J. Biol. Chem. 275,691^696.

[16] Whitaker, J.R. (1984) Pectic substances, pectic enzymes and hazeformation in fruit juices. Enzyme Microbiol. Technol. 6, 341^349.

[17] Vilanova, M., Blanco, P., Corte¤s, S., Castro, M., Villa, T.G. andSieiro, C. (2000) Use of a PGU1 recombinant Saccharomyces cerevi-siae strain in oenological fermentations. J. Appl. Microbiol. 89, 876^883.

[18] Sherman, F., Fink, G.R. and Hicks, J.B. (1989) Methods in YeastGenetics : A Laboratory Manual, 2nd edn. Cold Spring Harbor Lab-oratory Press, Cold Spring Harbor, NY.

[19] Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Molecular Clon-ing: A Laboratory Manual. Cold Spring Harbour Laboratory Press,Cold Spring Harbour, NY.

[20] McKay, A.M. (1988) A plate assay method for the detection offungal polygalacturonase secretion. FEMS Microbiol. Lett. 56,355^358.

[21] Johnston, D.J. and Williamson, B. (1992) Puri¢cation and character-ization of four polygalacturonases from Botrytis cinerea. Mycol. Res.96, 343^349.

FEMSLE 10454 23-5-02

P. Blanco et al. / FEMS Microbiology Letters 210 (2002) 187^191 191