Roles of DegP in Prevention of Protein Misfolding in the Periplasm upon Overexpression of Penicillin Acylase in Escherichia coli

Enhancement of the production of soluble recombinant penicillinacylase in Escherichia coli via coexpression of a periplasmicprotease/chaperone, DegP, was demonstrated. Coexpression ofDegP resulted in a shift of in vivo penicillin acylase (PAC)synthesis flux from the nonproductive pathway to the productiveone when pac was overexpressed. The number of inclusion bodies,which consist primarily of protein aggregates of PAC precursorsin the periplasm, was highly reduced, and the specific PAC activitywas highly increased. DegP was a heat shock protein inducedin response to pac overexpression, suggesting that the proteincould possibly suppress the physiological toxicity caused bypac overexpression. Coexpression of DegP_S210A, a DegP mutantwithout protease activity but retaining chaperone activity,could not suppress the physiological toxicity, suggesting thatDegP protease activity was primarily responsible for the suppression,possibly by degradation of abnormal proteins when pac was overexpressed.However, a shortage of periplasmic protease activity was notthe only reason for the deterioration in culture performanceupon pac overexpression because coexpression of a DegP-homologousperiplasmic protease, DegQ or DegS, could not suppress the physiologicaltoxicity. The chaperone activity of DegP is proposed to be anotherpossible factor contributing to the suppression.

INTRODUCTION

Top

Introduction

The well-known genetic information of Escherichia coli and successfulapplications of recombinant DNA technology make it possiblefor a variety of attempts with genetic engineering techniquesto overproduce recombinant proteins. Upon performing the cultivationfor recombinant protein production, there are two primary goals:high-cell-density cultivation and high-level gene expression.Culture performance can be optimized when the two goals areachieved simultaneously. High-cell-density culture can be obtainedby fed-batch cultivation (48), in which concentrated mediumis fed gradually into the bioreactor. The primary concern ofthis operation is developing an optimum feeding strategy, basedon which cells can be maintained in a high-energy state forenhancing gene expression while cells are growing.

On the other hand, various genetic strategies have been developedfor high-level gene expression (24). The use of strong promoters(e.g., tac, trc, and T7) for regulation of gene expression hasbeen successfully applied to enhance recombinant protein productionby improving transcriptional efficiency and perhaps translationalefficiency as well (35). However, it is a common problem thatinsoluble protein aggregates, known as inclusion bodies, tendto accumulate inside the cells upon overproduction of gene products(14, 44). While the mechanism of inclusion body formation isnot completely understood, it is believed that the overexpressedgene products cannot be suitably processed by folding modulatorsto develop a proper protein structure (46). For proteins destinedto be exported, the protein formation mechanism would be morecomplicated, and the efficiency of translocation, posttranslocationalfolding, processing, and targeting becomes important. In principle,the precursors, intermediates, or final gene products can possiblyform inclusion bodies in the cytoplasm and/or periplasm upongene overexpression (4). This raises an important issue that,in addition to improving the efficiency of each gene expressionstep (i.e., transcription, translation, and posttranslationalsteps), balanced protein synthesis flux throughout these stepsshould be properly maintained to avoid the accumulation of anyprotein species upon the overproduction of recombinant proteins.

Within the past decade, the issues of protein misfolding inthe bacterial periplasm and heat shock responses to extracytoplasmicstresses began to gain attentions. A specific periplasmic heatshock regulon of ${sigma}$ ^E, which is similar to the ${sigma}$ ^H heat shock regulonin the cytoplasm, has been identified in E. coli (9, 29, 32).Just as occurs in the cytoplasm, the heat shock responses toextracytoplasmic stresses were driven by synthesis of a varietyof heat shock proteins expressing protease activity that degradesmisfolded proteins and/or chaperone activity that renaturesmisfolded proteins. From the viewpoint of application, periplasmicproteins with these protease and/or chaperone activities wouldbe proper candidates for relieving extracytoplasmic stresseswhen cells are overexpressing foreign gene products.

More than 10 periplasmic proteases have been identified in E.coli (25). On the other hand, several periplasmic proteins,including DegP (also known as HtrA or Do) (42), Skp (11), SurA(20), and FkpA (2), have been identified as having chaperoneactivity and could play a role in the folding or targeting ofextracytoplasmic proteins. Among these periplasmic proteasesand chaperones, DegP and FkpA are two gene products inducedin response to extracytoplasmic stresses, such as heat shockor the presence of misfolded proteins, via the ${sigma}$ ^E heat shockregulon and Cpx two-component systems (32).

DegP is perhaps the only periplasmic heat shock protein withboth protease and chaperone activities. It has an inducibleserine protease activity for breakdown of aberrant periplasmicproteins arising upon extracytoplasmic stresses (17, 45). Anotheruncommon feature of DegP is that this protein can exhibit eitherprotease or chaperone activity with temperature as the regulatoryswitch (42). With these multiple functions, DegP is known tobe involved in relieving extracytoplasmic stresses upon overexpressionof several gene products, including alkaline phosphatase (PhoA)(13), DsbA'-PhoA (10), MalS (42), and maltose-binding protein(MalE) variants (3, 28).

We previously demonstrated that the presence of exogenous DegPcould enhance the production of recombinant penicillin acylase(PAC), an important industrial enzyme for the production ofmany ß-lactam antibiotics (37), in E. coli (21). Theformation of mature PAC in the periplasm involves a series ofposttranslational steps, including translocation and periplasmicprocessing/folding steps, which are unusual for prokaryoticproteins (Fig. 1) (38). The periplasmic processing mechanismis known to consist of various proteolytic steps via intramolecularautoproteolysis (15, 38). The formation of inclusion bodies,which are composed primarily of PAC precursors in the periplasm,was recently identified as an important obstacle to the overproductionof PAC in E. coli (8, 36, 43). Hence, efforts have been directedto reducing the number of periplasmic PAC inclusion bodies.

View larger version (19K):
[in this window]
[in a new window]
FIG. 1. Synthesis and maturation of PAC in E. coli. The structural pac gene from E. coli ATCC 11105 encodes a 95-kDa polypeptide precursor (preproPAC) composed of, from the N terminus to the C terminus, a signal peptide (S), ${alpha}$ subunit ( ${alpha}$ ), connecting peptide (C), and ß subunit (ß). The signal peptide directs the export of preproPAC into the periplasm and is removed after translocation. Another type of PAC precursor, proPAC, 92 kDa, is formed in the periplasm. Periplasmic processing, which involves a series of proteolytic steps, follows to remove the connecting peptide, and the two subunits ( ${alpha}$ at 24 kDa and ß at 62 kDa) become available for assembly of mature PAC.

In this study, we provide evidence that DegP functions primarilyas a protease to improve cell physiology by preventing the misfoldingand aggregation of periplasmic PAC precursors when pac is overexpressed.Cultivation performance for the production of recombinant PACwas significantly enhanced due to improved cell physiology aswell as simultaneous increases in the pac gene expression leveland culture cell density.

MATERIALS AND METHODS

Top

Materials and Methods

Bacterial strains and plasmids. Strains and plasmids used in the study are summarized in Table1 and briefly described here. MD ${Delta}$ P7 was used as the host forthe production of recombinant PAC. Compared to the parent strain,E. coli HB101, MD ${Delta}$ P7 could potentially have higher pac translationaland periplasmic processing efficiency for pac expression (7).MC4100 ${lambda}$ ${phi}$ [htrA-lacZ] contains a single copy of the P_degP::lacZtranscriptional fusion at the ${lambda}$ phage attachment site of thechromosome (12). Molecular cloning was performed according tostandard protocols (34), and HB101 was the host for cloning.Restriction enzymes were purchased from New England Biolabs(Beverly, Mass.). PCR was conducted in an automated thermalcycler (Amplitron II; Thermolyne, Dubuque, Iowa). Purificationof plasmid DNA was performed with a spin column kit purchasedfrom Clontech (Palo Alto, Calif.) or Viagen (Taipei, Taiwan).Plasmid transformation was carried out with an electroporator(E. coli Pulser; Bio-Rad, Hercules, Calif.).

View this table:
[in this window]
[in a new window]
TABLE 1. Strains, plasmids, and oligonucleotides

E. coli genes degP, degQ, and degS were amplified by PCR withPFU DNA polymerase (Stratagene, La Jolla, Calif.) and appropriateprimers (Table 1). The PCR templates were pKS12 (45) for degPand HB101 genomic DNA for degQ and degS. The PCR products flankedwith NcoI and KpnI sites (1.55, 1.44, and 1.24 kb for degP,degQ, and degS, respectively) were purified and cloned intothe corresponding restriction sites of pTrcKn99A. DNA sequencingof the PCR-cloned genes was performed with pTrcKn99A-derivedvectors. The NcoI-KpnI DNA fragments containing degP, degQ,and degS were subcloned into the NcoI-KpnI backbone of pAR3,which had been treated by complete KpnI digestion and partialNcoI digestion, to form pARDegP, pARDegQ, and pARDegS, respectively.The design of the NcoI site in the sense primers resulted inchanges in the second amino acid of the signal peptides of Lys $->$ Alafor DegP and DegQ and Phe $->$ Val for DegS, and these mutations didnot appear to affect translocation.

In vitro site-directed mutagenesis was conducted with the QuickChangekit (Stratagene, La Jolla, Calif.) according to the manufacturer'sprotocol. pARDegP was the mutagenesis template, and PSA/APSAwere the mutagenesis primers; they contain a silent mutationwith an extra NarI site for screening purposes. pTrcKnPAC2902contains the pac operon, whose transcription is regulated bythe trc promoter (8). It has the pBR322 replication origin andis therefore compatible with various pAR3-derived plasmids carryingthe pACYC184 replication origin.

Cultivation. Cells were revived by streaking the stock culture stored at-80°C on a Luria-Bertani (LB) agar plate (5 g of NaCl, 5g of Bacto yeast extract, 10 g of Bacto tryptone, and 15 g ofBacto agar per liter). The plate was incubated at 37°C forapproximately 20 h. An isolated single colony was picked toinoculate 25 ml of LB medium, which was then incubated at 37°Con a rotary shaker at 220 rpm for approximately 15 h. The mediumwas supplemented with kanamycin at 25 µg/ml or chloramphenicolat 32 µg/ml when necessary. The seed culture at a volumeof 10 ml was used to inoculate a table-top bioreactor (Omni-Culture;VirTis, Gardiner, N.Y.) containing 1 liter of working volumeof LB medium. Unless otherwise specified, isopropyl-ß-D-thiogalactopyranoside(IPTG) and arabinose were added simultaneously to induce thesynthesis of PAC and protease (DegP, DegQ, or DegS), respectively.The culture was supplemented with 10 µl of antifoam 289per liter (Sigma, St. Louis, Mo.) to avoid excessive foaming.Filter-sterilized air at 1.5 liters/min was purged into theculture for aeration. The culture pH was held at 7.0 ±0.1 by adding 3 N NaOH or 3 N HCl with a combined pH electrode(Mettler-Toledo, Berne, Switzerland), a pH controller (PC310;Suntex, Taipei, Taiwan), and two peristaltic pumps (101U/R;Watson Marlow, Falmouth, United Kingdom). The bioreactor wasoperated at 28°C and 500 rpm for approximately 50 h.

Analytical methods. The culture sample was appropriately diluted with saline solutionfor measuring cell density as the optical density at 600 nm(OD₆₀₀) with a spectrophotometer (V-530; Jasco, Tokyo, Japan).For the preparation of cell extract, 40 OD₆₀₀ units of cellswere centrifuged at 2°C and 6,000 x g for 5 min. The supernatantwas assayed for the extracellular enzyme activity. The cellpellet was resuspended in 2 ml of sodium phosphate buffer (0.05M, pH 7.5). The cell suspension was sonicated for 2 min withan ultrasonic processor (Sonics & Materials, Danbury, Conn.)and then centrifuged at 2°C and 20,000 x g for 20 min. Thesupernatant containing soluble proteins was assayed for theintracellular enzyme activity. The pellet containing insolubleproteins and cell debris was washed with phosphate buffer, resuspendedin TE-SDS buffer (10 mM Tris-HCl [pH 8.0], 1 mM EDTA, 1% sodiumdodecyl sulfate [SDS]) and heated at 100°C for 5 min. Theprotein content of the pellet was analyzed as the insolublefraction. All the analytical methods in this study, includingmicrobiological screening of PAC-producing strains, PAC assay,SDS-polyacrylamide gel electrophoresis (SDS-PAGE), and immunologicalanalysis (Western blotting), were performed as described previously(8). ß-Galactosidase was assayed as described previously(26). All assays were conducted in duplicate.

RESULTS

Top

Results

Effect of DegP coexpression on pac overexpression. For the control experiments with MD ${Delta}$ P7 harboring pTrcKnPAC2902and MD ${Delta}$ P7 harboring pTrcKnPAC2902 and pAR3, the culture performancein pac overexpression is summarized in Fig. 2. In all cases,formation of inclusion bodies was observed (Fig. 3), thoughthe production of PAC was effective. The inclusion bodies wereprimarily composed of proPAC in the periplasm, as describedin previous reports (6, 36, 43). The presence of exogenous DegPresulted in a significant decrease in the number of PAC inclusionbodies and increase in the specific PAC activity when pac wasoverexpressed. For the culture of MD ${Delta}$ P7 harboring pTrcKnPAC2902and pARDegP with IPTG induction at 0.05 mM but without arabinoseinduction, formation of PAC inclusion bodies was still observed,but there were significantly fewer than in the control experimentswith the same induction conditions (Fig. 3). The reduction inthe number of PAC inclusion bodies resulted from leaky expressionof degP by pARDegP, suggesting that a relatively small amountof exogenous DegP would be enough to improve the misfoldingof proPAC. The inclusion bodies completely disappeared uponincreasing the intracellular DegP concentration for the cultureof MD ${Delta}$ P7 harboring pTrcKnPAC2902 and pARDegP with simultaneousIPTG induction at 0.05 mM and arabinose induction at 0.05 g/liter(Fig. 3).

View larger version (26K):
[in this window]
[in a new window]
FIG. 2. Effect of DegP on pac overexpression: Time profiles of cell density (panel A), specific PAC activity (panel B), and cell lysis level (panel C) for cultures of MD ${Delta}$ P7 harboring various plasmids are shown. The level of cell lysis was defined as the percentage of total PAC activity detected in the extracellular medium. All cultures were supplemented with 0.05 mM IPTG for induction of PAC synthesis when the first sample was taken (i.e., OD₆₀₀ of 1.0 to 1.3). Open symbols represent cultures without arabinose supplementation, whereas solid symbols represent those with arabinose supplementation (0.05 g/liter) for induction of DegP synthesis when the first sample was taken.

View larger version (81K):
[in this window]
[in a new window]
FIG. 3. SDS-PAGE (panels A and B) and immunological (panel C) analyses of protein contents of soluble (panel A) and insoluble (panels B and C) fractions of the last samples of various cultures shown in Fig. 2 and 4. Lanes: M, protein markers; C, pTrcKnPAC2902 without IPTG as the control experiment; 1, pTrcKnPAC2902 with 0.05 mM IPTG; 2, pTrcKnPAC2902 and pAR3 with 0.05 mM IPTG; 3, pTrcKnPAC2902 and pAR3 with 0.05 mM IPTG and 0.05 g of arabinose per liter; 4, pTrcKnPAC2902 and pARDegP with 0.05 mM IPTG only; 5, pTrcKnPAC2902 and pARDegP with 0.05 mM IPTG and 0.05 g of arabinose per liter; 6, pTrcKnPAC2902 and pARDegP with 0.1 mM IPTG only; 7, pTrcKnPAC2902 and pARDegP with 0.1 mM IPTG and 0.05 g of arabinose per liter.

Cell physiology was significantly improved when degP was coexpressed.Up to 16% of PAC activity was detected in the extracellularmedium for the control experiment, versus only 1% for the cultureof MD ${Delta}$ P7 harboring pTrcKnPAC2902 and pARDegP with simultaneousIPTG induction at 0.05 mM and arabinose induction at 0.05 g/liter,suggesting that a much smaller number of cells were lysed whendegP was coexpressed. As a result, culture performance was significantlyimproved (Fig. 2). Compared to the control culture of MD ${Delta}$ P7 harboringpTrcKnPAC2902, the final cell density and specific PAC activitywere increased by 58% (7.6 versus 4.8 OD₆₀₀ units) and 13% (717versus 634 U/liter/OD₆₀₀ units), respectively, resulting inan 80% increase in the volumetric PAC activity (5,424 versus3,011 U/liter). An increase in arabinose induction concentrationto 0.1 g/liter did not further improve culture performance (datanot shown).

More distinct DegP effect upon extremely high level pac expression. The positive effect of DegP on the production of PAC was evenmore distinct when pac was overexpressed with a higher IPTGconcentration for induction of 0.1 mM (Fig. 4). With these cultureconditions, the number of PAC inclusion bodies was extremelylarge for the control experiment with MD ${Delta}$ P7 harboring pTrcKnPAC2902(Fig. 3). The amount of leaky degP expression for MD ${Delta}$ P7 harboringpTrcKnPAC2902 and pARDegP was not enough for proper functionwhen arabinose was not added. Cell growth was significantlyinhibited due to the extracytoplasmic stress. Increasing theintracellular DegP concentration by adding 0.05 g of arabinoseper liter resulted in a significant decrease in the number ofPAC inclusion bodies (Fig. 3). Coexpression of degP could alsoimprove cell physiology by reducing the level of cell lysis,relieving the inhibition of cell growth. Compared to the controlculture of MD ${Delta}$ P7 harboring pTrcKnPAC2902, the final cell densityand specific PAC activity were increased by 114% (6.0 versus2.8 OD₆₀₀ units) and 39% (933 versus 673 U/liter/OD₆₀₀ unit),respectively, resulting in an approximately twofold increasein the volumetric PAC activity (5,593 versus 1,880 U/liter).Up to 50% of PAC activity was detected in the extracellularmedium for the control experiment, whereas only 2% was foundfor the culture of MD ${Delta}$ P7 harboring pTrcKnPAC2902 and pARDegPwith simultaneous IPTG induction at 0.1 mM and arabinose inductionat 0.05 g/liter.

View larger version (28K):
[in this window]
[in a new window]
FIG. 4. Effect of DegP on pac overexpression. Same as Fig. 2 except all the cultures were supplemented with 0.1 mM IPTG. Solid rhombus, triangle, and square represent cultures supplemented with 0.1 mM IPTG when the first sample was taken but with 0.05 g of arabinose per liter added 4 h, 7 h, and 12 h, respectively, after the first sample was taken.

The effect of degP induction timing was investigated by cultivationsin which degP coexpression was induced 4 h, 7 h, and 12 h afterpac induction (Fig. 4). It appears that the positive effects(i.e., enhanced cell growth, increased pac expression, and reducedcell lysis) on pac overexpression were not observed until degPcoexpression. In addition, these effects gradually diminishedwhen degP coexpression was delayed and could hardly be observedif degP coexpression was delayed by 12 h (Fig. 4). The resultssuggest that cell physiology improved immediately after degPcoexpression and that the timing of degP coexpression was rathercritical. The PAC inclusion bodies formed before degP coexpressiondisappeared gradually toward the end of the cultivations (datanot shown). The results suggest that DegP not only preventedfurther formation of PAC inclusion bodies but also degradedthe misfolded proPAC. However, the question of whether the misfoldedproPAC was simply degraded or was rescued and subsequently shiftedinto the productive pathway remains to be answered.

DegP is a heat shock protein induced in response to pac overexpression. MC4100 ${lambda}$ ${phi}$ [htrA-lacZ] contains a single copy of the P_degP::lacZtranscriptional fusion at the ${lambda}$ phage attachment site of thechromosome of MC4100 (12). The strain can be used to estimatethe level of the cell's heat shock response to extracytoplasmicstresses by assaying ß-galactosidase activity. WithMC4100 ${lambda}$ ${phi}$ [htrA-lacZ] harboring pTrcKnPAC2902, an increase in degPpromoter activity was observed as a physiological response topac overexpression, which resulted in slight growth inhibitionbut significant increases in specific PAC activity, the numberof PAC inclusion bodies, and the level of cell lysis (Fig. 5).The results suggest that DegP is a heat shock protein inducedin response to pac overexpression.

View larger version (25K):
[in this window]
[in a new window]
FIG.5. Heat shock response to pac overexpression. Time profiles of cell density (panel A), specific PAC activity (panel B), specific ß-galactosidase activity (panel C), and cell lysis level (panel D) for cultures of MC4100 ${lambda}$ ${phi}$ [htrA-lacZ] harboring pTrcKnPAC2902 induced with IPTG at 0, 0.05, and 0.1 mM are shown. Immunological analysis of the protein contents of the insoluble fractions of the last samples of various cultures is shown in panel E. Lane 1, no IPTG; 2, 0.05 mM IPTG; 3, 0.1 mM IPTG.

Effect of coexpression of DegP_S210A and other DegP-like periplasmic proteases on pac overexpression. The role that DegP plays in pac overexpression can be identifiedby comparing culture performance among MD ${Delta}$ P7 harboring pTrcKnPAC2902,MD ${Delta}$ P7 harboring pTrcKnPAC2902 and pARDegP, and MD ${Delta}$ P7 harboringpTrcKnPAC2902 and pARDegP_S210A. The results are summarized inFig. 6. Unlike wild-type DegP, DegP_S210A could not suppressthe physiological toxicity caused by pac overexpression. Cultureperformance (i.e., pac expression level, number of PAC inclusionbodies, cell growth, and cell lysis) of MD ${Delta}$ P7 harboring pTrcKnPAC2902and pARDegP_S210A was similar to that of the control experimentwith MD ${Delta}$ P7 harboring pTrcKnPAC2902. The results suggest thatthe protease activity of DegP is responsible for suppressionof the physiological toxicity caused by pac overexpression (seelater discussion).

View larger version (45K):
[in this window]
[in a new window]
FIG. 6. Effect of various periplasmic proteases on pac overexpression. Same as Fig. 2 except the solid symbols represent cultures further supplemented with 0.05 g of arabinose per liter for induction of protease synthesis when the first sample was taken. Immunological analysis of the protein contents of the insoluble fractions of the last samples of various cultures is shown in panel D. Lanes: 1, pTrcKnPAC2902; 2, pTrcKnPAC2902 and pARDegP; 3, pTrcKnPAC2902 and pARDegP_S210A; 4, pTrcKnPAC2902 and pARDegQ; 5, pTrcKnPAC2902 and pARDegS.

Since the protease activity of DegP was primarily responsiblefor suppression of the physiological toxicity caused by pacoverexpression, it would be interesting to see if any otherperiplasmic protease had the same phenotype. The effect of DegQand DegS, two periplasmic serine proteases homologous to DegP,on pac overexpression was investigated with MD ${Delta}$ P7 harboring pTrcKnPAC2902and pARDegQ or pARDegS, and the results are summarized in Fig.6. Culture performance of MD ${Delta}$ P7 harboring pTrcKnPAC2902 and pARDegQand of MD ${Delta}$ P7 harboring pTrcKnPAC2902 and pARDegS was similarto or even worse than that in the control experiment with MD ${Delta}$ P7harboring pTrcKnPAC2902. The results suggest that neither DegQnor DegS could suppress the physiological toxicity caused bypac overexpression.

DISCUSSION

Top

Discussion

Coexpression of DegP resulted in the following improvementsupon pac overexpression: (i) it prevented misfolding of proPACdue to a reduction in the number of PAC inclusion bodies; (ii)it shifted PAC synthesis flux from the nonproductive pathwayto the productive one due to an increase in pac expression level;and (iii) it improved cell physiology due to enhanced cell growthand less cell lysis (Fig. 2, 3, and 4). The result that DegPis a heat shock protein induced in response to pac overexpression(Fig. 5) suggests that DegP could suppress the physiologicaltoxicity caused by pac overexpression. DegP was also found tobe a heat shock protein induced in response to the misfoldingof many periplasmic proteins, such as PhoA (13), DsbA'-PhoA(10), MalS (42), OmpF mutants (27), and maltose-binding protein(MalE) variants (3, 28). Misfolding of these proteins is usuallythe consequence of overexpression of gene products or a foldingdefect caused by certain mutations. While the heat shock responsewas initiated primarily for degrading or renaturing misfoldedproteins in the periplasm, the increased DegP level might notbe high enough for complete function and the misfolded geneproducts still accumulated as periplasmic inclusion bodies (10,13, 28). Normally, heat shock proteins express the proteaseactivity for degrading misfolded proteins and/or chaperone activityfor renaturing misfolded proteins. We therefore tried to identifythe exact role(s) that DegP plays as a heat shock protein inresponse to pac overexpression.

There are at least two hypothetical conditions which are notcompletely mutually exclusive to explain the positive effectof DegP when pac was overexpressed. First, the periplasmic processingmachinery was saturated with PAC precursors of proPAC transientlyformed in an excess amount. The periplasmic processing of proPACis initiated by proteolysis on the Thr₂₆₃-Ser₂₆₄ bond via intramolecularautoproteolysis (15). Two protein chains (i.e., ${alpha}$ plus C andß in Fig. 1) are formed and fold with each other.Enzyme activity at approximately the same level as that of maturePAC could be detected for the folding complex (22, 23). Apparently,the production of PAC was limited by the initiation step, asproPAC was the major aggregation species upon pac overexpression.In other words, folding of proPAC in the periplasm was not properlycarried out for subsequent autoproteolysis, possibly due toan increased level of the local protein concentration. Thislimitation possibly can be resolved by periplasmic chaperones.Second, cell physiology was significantly affected as a resultof extracytoplasmic stresses, such as protein misfolding and/orfailure of degradation of misfolded proteins in the periplasmupon pac overexpression. This limitation possibly can be resolvedby periplasmic proteases and/or chaperones.

DegP has both protease and chaperone activities (42), by whichcell physiology can be improved. For example, DegP was identifiedas a periplasmic heat shock protein for degrading abnormal proteinsin the periplasm when phoA was overexpressed (13). Overexpressionof phoA in a degP mutant caused a severe defect in export ofseveral proteins, including PhoA itself, and such a defect wasnot observed (even in the degP mutant) in the absence of phoAoverexpression (13). The proteolytic activity of DegP couldalso suppress the physiological toxicity originating from overexpressionof a hybrid protein of DsbA'-PhoA in a similar way (10). Asa result, DegP was identified as a peripheral membrane proteasewhich degrades misfolded proteins accumulated near the periplasmicside upon extracytoplasmic stresses (13, 30, 40). On the otherhand, DegP was also a heat shock protein in response to themisfolding of ovexpressed MalS in the periplasm and the physiologicalrole was, however, proposed to be a chaperone (42). While MalSsolubility was improved, the number of MalS inclusion bodieswas not reduced, and the in vivo chaperone activity of DegPwas not clearly demonstrated (42). It would be interesting tosee which activity contributed the improvement documented inFig. 2 to 4 when pac was overexpressed.

DegP is a serine protease whose active center consists of thecatalytic triad residues His105, Asp135, and Ser210 (33). TheDegP variant harboring the S210A mutation completely lost theprotease activity but retained the chaperone activity (41, 42),as the mutation induced minor changes in the tertiary structureof the protein (39). Hence, the role of DegP in pac overexpressioncan be identified with this mutant, DegP_S210A, being coexpressed.Unlike wild-type DegP, DegP_S210A could not suppress the physiologicaltoxicity caused by pac overexpression (Fig. 6), suggesting thatDegP functioned as a protease to improve culture performanceupon pac overexpression. Similar approaches were conducted toidentify the protease activity of DegP for suppression of thephysiological toxicity when the hybrid protein DsbA'-PhoA wasoverexpressed (10) or the chaperone activity of DegP for suppressionof heat shock responses originating from misfolding of MalS(42) and OmpF mutants (27).

More than 10 periplasmic proteases have been identified in E.coli. Among them, DegQ and DegS are the two with DegP-like proteinfunction and structure. All three proteins are serine proteases(47). However, only DegP is heat inducible and only E. colidegP mutants have a temperature-sensitive phenotype (47). DegQcan functionally substitute for DegP under certain conditions(47). Though DegQ is not a heat shock protein, it can transientlydegrade denatured, unfolded proteins which accumulate in theperiplasm upon heat shock or extracytoplasmic stresses and/ornewly secreted proteins prior to folding and disulfide bondformation (17). Similar to many intracellular proteases, whichform large oligomeric complexes, both DegP and DegQ form hexamersor dodecamers in solution (17, 18). DegP and DegQ are similarin size and highly homologous (47). In addition, they both containtwo PDZ domains, which are responsible for substrate recognition,and possibly have a common substrate recognition mechanisms(30). On the other hand, DegS is also homologous to DegP, thoughthe homology is not as strong as that of DegQ and DegP (47).It also has the PDZ domain for substrate recognition, but onlyone (30). The physiological function of DegS was identifiedto be related to the regulation of ${sigma}$ ^E activity (1).

Unlike DegP, neither DegQ nor DegS could suppress the physiologicaltoxicity caused by pac overexpression (Fig. 6). The apparentdifferent effects of the two highly homologous proteases, DegPand DegQ, on pac overexpression suggest that the improvementin culture performance was not simply caused by DegP proteaseactivity. In that case, the chaperone activity could be anotherpossible improving factor. It is known that DegP can exhibitprotease or chaperone activity with temperature as an environmentalswitch; namely the protease and chaperone activities dominateunder high and low temperatures, respectively (42). Hence, thepossibility that the positive effect of DegP on pac overexpressionwas caused by the chaperone activity should not be excludedsince all the cultivations were conducted at a relatively lowtemperature of 28°C in this study. The chaperone activitywas responsible for shifting the PAC synthesis flux from thenonproductive pathway to the productive one possibly by increasingthe solubility of proPAC. In addition, increasing the efficiencyof the current limiting step (i.e., autoproteolysis of proPAC)would require a proper folding status of proPAC that could possiblybe achieved by DegP chaperone activity.

It should be noted that the chance that DegP protease activitydirectly assisted the periplasmic processing of proPAC, whichconsists of several proteolytic steps, would be slim due tothe following reasons. First, DegP targets its protein substratesby recognizing a state of protein denaturation rather than specificamino acid sequences in particular proteins (19). The PDZ domainsof DegP are responsible for substrate recognition (30). In addition,unfolding of the protein substrates would be essential for theiraccess into the inner chamber of the double ring-shaped DegP,where cleavage of peptide bonds may occur (16). Second, bothDegP and DegQ have two PDZ domains and an active center withthe catalytic triad residues of His, Asp, and Ser for exhibitingprotease activity. One would expect to see the same effect forDegQ as for DegP on pac overexpression if DegP protease activitydirectly assisted autoproteolysis of proPAC.

ACKNOWLEDGMENTS

This study was supported by the National Science Council ofTaiwan (grant no. NSC 91-2214-E-035-007).

We thank K. Makino, J. Gutierrez, and J. Beckwith for providingstrains for this study.

FOOTNOTES

* Corresponding author. Mailing address: Department of Chemical Engineering, Feng Chia University, 100 Wenhwa Road, Taichung, Taiwan 407, ROC. Phone: 886 4 2451 7250, ext. 3678. Fax: 886 4 2451 0890. E-mail: cpchou{at}fcu.edu.tw.

REFERENCES

Top