Detergent-independent in vitro activity of a

truncated Bacillus signal peptidase

Maarten L. van Roosmalen, Jan D. H. Jongbloed, Anne de Jong,Jaap van Eerden, Gerard Venema, Sierd Bron and Jan Maarten van Dijlt

503632348.Author for correspondence: jan Maarten van Dijl. Tel: +31503633079. Fax: +e-mail: j.M.VAN.DIjL@FARM.RUG.NL

Department of Genetics,Groningen BiomolecularSciencesandBiotechnology Institute,PO Box 14, 9750 AA Haren,The Netherlands

The Gram-positive eubacterium Bacillus subtilis contains five chromosomallyencoded type I signal peptidases (SPases) for the processing of secretory pre-proteins. Even though four of these SPases, denoted SipS, SipT, SipU and SipV,are homologous to the unique SPase lof Escherichia co/i, they are structurallydifferent from that enzyme, being almost half the size and containing onemembrane anchor instead of two. To investigate whether the uniquemembrane anchor of Bacillus SPases is required for in vitro activity, solubleforms of SipS of B. subtilis, SipS of Bacillus amyloliquefaciens and SipC of thethermophile Bacillus caldolyticus were constructed. Of these three proteins,only a hexa-histidine-tagged soluble form of SipS of B. amyloliquefaciens couldbe isolated in significant quantities. This protein displayed optimal activity atpH 10, which is remarkable considering the fact that the catalytic domain ofSPases is located in an acidic environment at the outer surface of themembrane of living cells. Strikingly, in contrast to what has been previouslyreported for the soluble form of the E. co/i SPase, soluble SipS was active in theabsence of added detergents. This observation can be explained by the factthat a highly hydrophobic surface domain of the E. co/i SPase, implicated indetergent-binding, is absent from SipS.

Keywords: Bacillus subtilis, protein secretion, SipS, Sip(

INTRODUCTION plasmid-encoded type I SPases, denoted SipP (Meijeret al., 1995; Tjalsma et al., 1999a). Even though thepresence of paralogous SPases is not unusual ineubacteria, archaea and eukaryotes, the number ofparalogous SPases in B. subtilis is unusually high,suggesting that some of these SPases may havespecialized functions. Indeed, it was recently shown thatSipS, SipT and SipP are of major importance for thesecretion of degradative enzymes and cell viability,unlike SipU, SipV and SipW (Tjalsma et al., 1998,1999a). Furthermore, SipW is of major importance forthe processing of precursors of certain proteins, such asTasA (St6ver & Driks, 1999a) and y qxM (St6ver &Driks, 1999b). Taken together, these observations in-dicate that there are significant differences in thesubstrate specificities of the various SPases of B. subtilis.Nevertheless, SipS, SipT, SipU and SipV wereshown toprocess the hybrid precursor pre-(A13i)-p-lactamase(pre-A13i-Bla; Tjalsma et al., 1997), and all five chromo-somally encoded SPases were involved in the processingof the Bacillus amyloliquefaciens IX-amylase AmyQ(Tjalsma et al., 1998). The latter observations showed

Secretory proteins are synthesized as pre-proteins withan N-terminal signal peptide. Type I signal peptidases(SPases) remove these signal peptides during, or shortlyafter, pre-protein translocation across the cytoplasmicmembrane, thereby releasing the mature proteins fromthe trans side of the membrane (for reviews, see Pugsley,1993; Dalbey et al., 1997). The largest number of type ISPases has, thus far, been found in the Gram-positiveeubacterium Bacillus subtilis, which contains five para-logous, chromosomally encoded enzymes of this type(SipS, SipT, SipU, Sip V and Sip W) (van Dijl et al.,1992; Bolhuis et al., 1996; Tjalsma et al., 1997, 1998) .Inaddition, certain strains of B. subtilis (natto) contain

t Present address: Department of Pharma(euti(al Biology, University ofGroningen, Antonius Deusinglaan 1,9713 AV Groningen, The Netherlands.

Abbreviations: Bam, Bacillus amyloliquefaciens; Bsu, Bacillus subtilis;CBB, Coomassie brilliant blue R; pre-A13i-Bla, pre(A13i)-p-Ia(tamase; pre-A2-AmyL, pre(A2)-()(-amylase; SPase, signal peptidase.

0002-4401 @ 2001 SGM

.J, #SipC (Bca) 1 MTKQKEKRGRRWPWFVAVCVVATLRLFVFSNYVVEGKSMMPTLESGNLLIVNKLSYDIGPIRRFDI

SipV (Bsu) 1 MKKRFWFLAGVVSVVLAIQVKNAVFIDYKVEGVSMNPTFQEGNELLVNKFSHRFKTIHRFDI

SpSB (Sau) 1 MKKELLEWIISIAVAFVILFIVGKFIVTPYTIKGESMDPTLKDGERVAVNIIGYKTGGLEKGNV

SipS (BSU) 1 MKSENVSKKKSILEWAKAIVIAVVLALLIRNFIFAPYVVDGDSMYPTLHNRERVFVNMTVKYIGEFDRGDI

SipS (Bam) 1 MKSEKEKTSKKSAVLDWAKAIIIAVVLAVLIRNFLFAPYVVDGESMEPTLHDRERIFVNMTVKYISDFKRGQI

SipP (pTAl015) 1 MTKEKVFKKKSSILEWGKAIVIAVILALLIRNFLFEPYVVEGKSMDPTLVDSERLFVNKTVKYTGNFKRGDI

SipT (Bsu) 1 -MTEEKNTNTEKTAKKKTNTYLEWGKAIVIAVLLALLIRHFLFEPYLVEGSSMYPTLHDGERLFVNKTVNYIGELKRGDI

SipU (BSU) 1 MNAKTITLKKKRKIKTIVVLSIIMIAALIFTIRLVFYKPFLIEGSSMAPTLKDSERILVDKAVKWTGGFHRGDI

Lep (Eco) 1 MAARDSLDKATLKKVAPKPGWLETGASVFPVLAIVLIVRSFIYEPFQIPSGSMMPTLLIGDFILVEKFAYENGHPKRGDI/\ ** ** * /\

2-39 109-120

# # #SipC (Bca) 67 IVFHANKKE--DYVKRVIGLPGDRIAYKNDILYVNGKKVDEPYLRPYKQKLLDGRLTGDFTLEEVTGKTR VPPGCI

SipV (Bsu) 63 VLFKGPDHK--VLIKRVIGLPGETIKYKDDQLYVNGKQVAEPFLKHLKSVSAGSHVTGDFSLKDVTGTSK VPKGKY

SpsB (Sau) 65 VVFHANKND- -DYVKRVIGVPGDKVEYKNDTLYVNGKKQDEPYLNYNLKHKQGDYITGTFQVKDLPNANPKSNVIPKGKY

SipS (Bsu) 72 VVLNGDDV- --HYVKRIIGLPGDTVEMKNDQLYINGKKVDEPYLAANKKRAKQDGFDH- LTDDFG- PVKVPDNKY

SipS (Bam) 74 VVLNGENE- --HYVKRIIGLPGDTVQMKNDQLYINGKKVSEPYLAANKKKAKQDGYT- LTDDFG-PVKVPDDKY

SipP (pTAl015) 73 IILNGKEKS-THYVKRLIGLPGDTVEMKNDHLFINGNEVKEPYLSYNKENAKKVGIN LTGDFG-PIKVPKDKY

SipT (Bsu) 80 VIINGETSK-IHYVKRLIGKPGETVQMKDDTLYINGKKVAEPYLSKNKKEAEKLGVS LTGDFG-PVKVPKGKY

SipU (BSU) 75 IVIHDKKSG-RSFVKRLIGLPGDSIKMKNDQLYINDKKVEEPYLKEYKQEVKESGVT- LTGDF- -EVEVPSGKY

Lep ( Eco ) 131 VVFKYPEDPKLDYI KRA VGLPGDKVTYDQTFSRRNGGEATSGFFEVPKNETKENG IRLSERKETLGDVTHATWIVPPGQY** * ** /\ * /\ *

159-193 236-258

# #

SipC (Bca) 141 FVLGDNRLSSWDSR-HFGFVKINQIVGKVDFRYWPFKQFAFQF SipV (Bsu) 137 FVVGDNRIYSFDSR-HFGPIREKNIVGVISDAE SpsB (Sau) 143 LVLGDNREVSKDSR-AFGLIDEDQIVGKVSFRFWPFSEFKHNFNPENTKN SipS (BSU) 142 FVMGDNRRNSMDSRNGLGLFTKKQIAGTSKFVFYPFNEMRKTN SipS (Bam) 143 FVMGDNRRNSMDSRNGLGLFTKKQIAGTSKFVFFPFNEIRKTK SipP (pTA1O15) 144 FVMGDNRQESMDSRNGLGLFTKDDIQGTEEFVFFPFSNMRKAK SipT (Bsu) 151 FVMGDNRLNSMDSRNGLGLlAEDRIVGTSKFVFFPFNEMRQTK SipU (Bsu) 145 FVMGDNRLNSLDSRNGMGMPSEDDIIGTESLVFYPFGEMRQAK Lep (Eco) 269 FMMGDNRDNSADSR-YWGFVPEANLVGRATAIWMSFDKQEGEWPTGLRLSRIGGIH

**** * *** * *

F;g. ,. Alignment of deduced amino acid sequences of signal peptidases of Gram-positive eubacteria and E. coli. Thealignment includes the following type I SPases: SipC (Bca) from 8. caldolyticus, SipV (Bsu) from 8. subtilis (Tjalsma et al.,1997), SpsB (Sau) from Staphylococcus aureus (Cregg et al., 1996), SipT (Bsu) from 8. subtilis (Tjalsma et al., 1997), SipP(pTA 1015) from 8. subtilis (natto) (Meijer et al., 1995), SipS (Bsu) from 8. subtilis (van Dijl et al., 1992), SipS (Bam) from 8.amyloliquefaciens (Meijer et al., 1995), SipU (Bsu) from 8. subtilis (Tjalsma et al., 1997) and the E. coli SPase I [lep (Eco)](Wolfe et al., 1983). Residues that are critical for activity are labelled with a #. The conserved arginine residue, which isthe second residue in the soluble forms (sf) of SipC, SipS (Bsu) and SipS-His (Bam), is indicated by an arrow. Regions oflep (Eco) that are absent from the aligned SPases are not shown, but their relative positions and the correspondingresidue numbers in the lep (Eco) sequence are indicated below the alignment. Note that residues 106-115 of lep (Eco),which represent a very hydrophobic p-hairpin, are not conserved in SipS (Bam) or other 8acillus SPases. Identical aminoacids (*) or conservative replacements (.) are marked. Residues printed in bold are predicted to be part of the membraneanchors of the SPases.

that the B. subtilis SPases have at least partly overlappingsubstrate specificities (Tjalsma et al., 1997).

Recent studies have provided compelling evidence thatthe type I SPase of Escherichia coli, also known as leaderpeptidase, makes use of a serine-lysine catalytic dyad(Tschantz et al., 1993; Paetzel et al., 1998). This is alsoprobably true for SipS, SipT, SipU and Sip V of B. subtilisand for the SPases from other eubacteria and organellesthat are related to the SPase lof E. coli (van Dijl et al.,1995; Dalbey et al., 1997). Nevertheless, apart from theresidues involved in catalysis (Paetzel et al., 1998 ;Bolhuis et al., 1999c), very little is known of thosefactors that determine the activity of SPases fromeubacteria other than E. coli. Considering the fact thatthe type I SPases of B. subtilis display different substratespecificities in vivo, these enzymes appeared to beattractive models for the identification of importantdeterminants of SPase activity. Furthermore, these

enzymes are structurally very different from the SPase ofE. coli as they are much smaller (e.g. SipS of B. subtiliscomprises 184 residues but the E. coli SPase has 323residues; Fig. 1) and contain only one membrane anchorinstead of two (for review, see Dalbey et al., 1997}.Notably, like the SPases from B. subtilis, the majority ofknown eubacterial SPases contain only one N-terminalmembrane anchor (Dalbey et al., 1997).

Interestingly, the two membrane anchors of the E. coliSPase I are dispensable for the catalytic activity of theextracytoplasmic domain of this enzyme (Kuo et al.,1993; Tschantz et al., 1995). Therefore, to initiate the invitro characterization of Bacillus type I SPases with theaim of identifying important conserved and/or non-conserved determinants for SPase activity, we wanted tomake use of soluble forms of these enzymes. Suchsoluble forms are, in principle, easier to characterizethan intact proteins with a membrane anchor that have

910

to be solubilized with detergents. In the present studies,we constructed soluble forms of SipS of B. subtilis, SipSof B. amyloliquefaciens and, with the aim of obtaining apresumably more stabIe enzyme, SipC of the thermo-phile Bacillus caldolyticus. This last organism can growand secrete proteins at a temper at ure of 70 aC. Theresults showed that only a truncated, hexa-histidine-tagged soluble form of SipS of B. amyloliquefacienscould be isolated in significant quantities for in vitrocharacterization. Strikingly, as demonstrated with pre-{A2)-cx-amylase {pre-A2-AmyL), a precursor that isefficiently processed in vivo and synthesized in vitro{Smith et at., 1987, 1988; Bolhuis et at., 1999b), thissoluble form of SipS was active in the absence of addeddetergents or phospholipids, unlike the soluble form ofthe E. coli SPase I {Tschantz et at., 1995). This ob-servation is of general importance, because the highlyhydrophobic surface domain of the E. coli SPase thatwas recently implicated in detergent-binding {Paetzel etat., 1998) is not conserved either in SipS of B. amylo-tiquefaciens, or in the large majority of knowneubacterial SPases. It thus seems that, at least withrespect to its in vitro requirement for added detergentsor phospholipids, the E. coli enzyme could represent arelatively small subgroup of the known SPases.

pT713. The sipC-specific fragment was amplified with theprimers SipCO02 (5'-GGAA TTCGTCGACGGAGGACAA-TT A TGCGGTTGTTCGTGTTCAGCAA TT ACG-3') andSipCO04 (5'-GGAA TTCGGA TCCA T AGAAGCGGAAGA-CTCC-3'), using pGDL46.36 DNA as a template. PlasmidpT7dCH, specifying a hexa-histidine-tagged soluble form ofSipC, was constructed by ligating an EcoRI- and SaII-cleaved,PCR-amplified fragment of sipC into the corresponding sitesof pT712. The sipC-specific fragment was amplified with theprimers SipCO02 and SipCHisO04 (5'-GCTCT AGAA TTCT -

TAGTGATGGTGATGGTGATGAAACTGAAAGGCGA-ACTGTTT AAACGG-3'), using plasmid pGDL46.36 DNAas a template. Plasmid pT7dS, specifying a soluble form ofSipS from B. subtilis (Bsu), was constructed by ligating a Sa/1-and BamHI-cleaved, PCR-amplified fragment of sipS (Bsu)into the corresponding sites of pT712. The sipS (Bsu)-specificfragment was amplified by PCR with the primers SipSO01 (5'-GACTAGTCGACGGAGGACAATTATGCGCAACTTT-A TTTTTGC-3') and Lbs91 (5'-CGGGA TCCCGGGACT -AA TTTGTTTTGCGC-3') using B. subtilis 168 chromo-somal DNA as a template. Plasmid pT7dAH, specifying ahexa-histidine-tagged soluble form of SipS from B. amyl-oliquefaciens (Barn), was constructed by ligating an EcoRI-and Sa/1-cleaved PCR-amplified fragment of sipS (Barn) intothe corresponding sites of plasmid pT712. The sipS (Bam)-specific fragment was amplified by PCR with the primersSipAO01 (5'-GGAA TTCGTCGACGGAGGACAA TT A TG-CGCAACTTTTTATTTGCTCC-3') and SipAHis02 (5'-GGAATTCTTAGTGATGGTGATGGTGATGGATCGA-TTTCGTCTTGCGAA-3') using pGDL46.21 as a template.T o prevent the selection of non-overexpressing variants ofpT7dC, pT7dCH, pT7dS and pT7dAH, these plasmids werefirst constructed using E. coli MC1061, which does not containthe gene for the T7 RNA polymerase. Subsequently, E. coliBL21(de3) was used for the production of soluble SPasesspecified by pT7dC, pT7dCH, pT7dS and pT7dAH.

Protein overproduction and purification. For overproductionand subsequent purification of the hexa-histidine-tagged.?:oluble form of SipS (Barn), denoted sf-SipS-His (Barn), E. coliBL21(de3) was transformed with plasmid pT7dAH. Trans-formants were grown overnight in TY medium at 37 °C. Next,11itre of fresh TY medium was inoculated with 10 ml of thisovernight culture and incubated at 37 °C. When the culturereached an OD6oo of 0.6--{).9, the production of sf-SipS-His(Barn) was induced by adding IPTG to a final concentration of0.5 mM. Cells were collected by centrifugation about 3 h afterinduction. The cell pellet was resuspended in 10 mllysis buffer(50 mM Tris/HCI, 300 mM NaCI, 1 mM phosphoramidon,1 mM PMSF, pH 8.0) and disrupted by three passages througha chilled French pressure cell at 10000 p.s.i. (69 MPa). Cellsand debris were removed from the extract by centrifugation at5000 9 (15 min, 4 °C). To separate the soluble sf-SipS-His(Barn) from the membrane-bound E. coli SPase I, membraneswere removed from the supernatant by two subsequentultracentrifugation steps (100000 g, 30 min, 4 °C). Finally, sf-SipS-His (Barn) was isolated by metal-affinity chromato-graphy, using a column containing 5 ml TALON resin(CLONTECH Laboratories) that was pre-equilibrated withlysis buffer. The column was washed with 25 mllysis bufferand sf-SipS-His (Barn) was eluted with elution buffer (i.e. lysisbuffer with 300 mM imidazole). To verify the levelofpurification, samples from 0.5 ml fractions were analysed bySDS-P AGE and subsequent staining with Coomassie brilliantblue R (CBB). Fractions containing the purified sf-SipS-His(Barn) were pooled and transferred to a buffer containing50 mM HEPES, 50 mM NaCI, 0.1 mM EDT A, 1 mM DTT

METHODS

Plasmids. bacterial strains and media. Table 1 lists theplasmids and bacterial strains used. TY medium (tryptone/yeast extract) contained Bacto-tryptone (1% , w /v), Bactoyeast extract (0.5% , w /v) and NaCI (1% , w /v). If required,the medium for E. coli was supplemented with ampicillin(50 J.lg mi-l) and kanamycin (40 J.lg mi-l).

DNA techniques. Procedures for DNA purification, restric-tion, ligation, agarose gel electrophoresis and transformationof E. coli were carried out as described by Sambrook et al.(1989). Enzymes were from Roche Molecular Biochemicals.PCR was carried out with Vent DNA polymerase (NewEngland Biolabs) as described by van Dijl et al. (1995). DNAand protein sequences were analysed with the PCGeneAnalysis Program (version 6.7; IntelliGenetic) and CLUSTAL wversion 1.74 (Thompson et al., 1994).

The sipC gene of B. caldolyticus was cloned as previouslydescribed for the cloning of the sipS gene of B. subtilis (vanDijl et al., 1992). First, chromosomal DNA of B. caldolyticuswas cleaved with Sau3A and ligated into the unique Bc/1 site ofplasmid pGDL46. Subsequently, E. coli MC1061 was trans-formed with the ligation mixture and kanamycin- andampicillin-resistant transformants were selected. Finally,transformants that were able to process the hybrid precursorpre-A13i-Bla, and release mature A13i-Bla into the surround-ing growth medium, were selected using a plate (halo) assayfor p-Iactamase activity (van Dijl et al., 1992) .Notably, the E.coli SPase I is unable to process pre-A13i-Bla in vivo.Consequently, this hybrid precursor remains attached to themembrane and no mature A13i-Bla is released into thesurrounding medium, unless a heterologous SPase is expressedthat is able to process pre-A13i-Bla. As shown by DNAsequencing, the plasmid pGDL46.36 thus selected containedthe sipC gene of B. caldolyticus (SWISS-PROT # P41027).

Plasmid pT7dC, specifying a soluble form of SipC, wasconstructed by ligating an EcoRI- and BamHI-cleaved PCR-amplified fragment of sipC into the corresponding sites of

==~~==

tM CS

MRLFVFSNYVVEGKQM.

~ACGGAGGACAA1'rATG" (,be)

-T7. IRBSI

tM CS

sf-SipC

sf-SipS (Bsu)

sf-SipC-His

sf-SipS-His (Bam)

Fig. 2. Schematic presentation of constructs for overproduction of truncated soluble signal peptidases. AII constructsused for the overproduction of sf-SipC, sf-SipC-His, sf-SipS (Bsu) and sf-SipS-His (Barn) are based on the plasmids pT712 orpT713 (filled parts of the bar). PCR-amplified sequences specifying the sf-Sip proteins we re cloned downstrearn of the T7promoter of pT712/pT713, using restriction sites (indicated by M CS) in the multiple cloning sites of these plasmids. AIItruncated genes contained the ribosome-binding site and start codon of the B. subtilis obg gene as indicated. Residuesprinted in bold are remnants of the mernbrane anchors of the corresponding wild-type SPases. The catalytic serineresidue of these enzymes is underlined. The sf-SipC-His and sf-SipS-His (Barn) contained a C-terminal hexa-histidine tag.

Fig. 3. Overproduction and purification of atruncated soluble Bacillus signal peptidase.(a) Cells of E. coli BL21(de3), transformedwith plasmid pT7dAH, were grown in TYmedium and the overproduction of sf-SipS-His (Bam) was induced with IPTG asdescribed in Methods. Cells containing the'empty vector' (pT712) were used as anegative control. Samples, collected 3 hafter induction, were separated by SDS-PAGE and, subsequently, the gel was stainedwith CBB. The positions of sf-SipS-His (Bam)(arrow) and molecular mass referencemarkers are indicated. (b) The sf-SipS-His(Bam) was purified by metal-affinitychromatography from a cytosolic fraction ofIPTG-induced cells of E. coli BL21(de3)containing plasmid pT7dAH as described inMethods. Samples obtained from thedifferent fractions, before and after metal-affinity chromatography, were analysed bySDS-PAGE and subsequent CBB staining.Cytosol, the cytosolic fraction of IPTG-induced cells; wash, the wash fraction;elution, the fraction obtained by elutionwith buffer containing imidazole. Thepositions of sf-SipS-His (Bam), the specificdegradation products d1 and d2 of sf-SipS-His (Bam) and molecular mass referencemarkers are indicated.

",,"

,,~"8!

~'ó

&',,"0

$""

(a) (b)

kDa

45

21

14.4

by rnetal-affinity chromatography. To overproduce thesoluble Bacillus SPases, the corresponding S'-truncatedsip genes were provided with the efficient ribosome-binding site and start codon of the obg gene of B. subtilis(Welsh et al., 1994), cloned downstrearn of the T7promoter on plasmids pT712 or pT713 (Studier et al.,1990) and introduced into an E. co li strain that producesthe T7 RNA polymerase upon induction with IPTG.The IPTG-induced expression of these truncated genesresulted in the production of sf-SipC, sf-SipS (Bsu) andsf-SipS-His (Barn) to levels allowing their detection inCBB-stained SDS-PAGE gels. In contrast, sf-SipC-Hiswas barely visible in CBB-stained SDS-PAGE gels (Fig.3a; only the results for sf-SipS-His [Barn ] are shown) .Notably, compared to the production of intact SipS

(Bsu) ar SipS (Barn) with thethe saluble farms af thesesignificantly lawer levelsvatians).

Characterization of sf-SipS

As our attempts to purify sf-SipC or sf-SipS (Bsu) tohomogeneity by conventional chromatography (e.g. ionexchange, gel filtration) met with little success, and sf-SipC-His produc~vels were too low (data notshown), our efforts to purify a soluble form of a BacillusSPase were focused on sf-SipS-His (Barn). Due to its C-terminal hexa-histidine tag, sf-SipS-His (Barn) could beisolated in one step by metal-affinity chromatography

913

same T7 expression system,enzymes were produced at(our unpublished obser-

(a)

TX-100SPase

(b)

TX-100SPase

Lysate

QJ

~...10EQJCl10...c:QJ

~QJ

0.

Fig. 4. Detergent-independent precursor processing by purifiedsf-5ip5-His (Bam). (a) 3s5-labelled pre-A2-AmyL, synthesized invitro with a B. subtilis 5-135 extract, was incubated (16 h at37 aC) with 7.5 JlM purified sf-5ip5-His (Bam) in 50 mM HEPE5(pH 7.5), 5 mM Mg504, in the presence or absence of 0.5 %Triton X-100 (TX-100). In parallel, labelled pre-A2-AmyL wasincubated without sf-5ip5-His (Bam). Samples were analysed by5D5-PAGE and fluorography. p, pre-A2-AmyL; m, mature A2-AmyL. (b) To verify the absence of 5Pase-Iike activities in thecytoplasm of E. coli, 3s5-labelled pre-A2-AmyL was synthesizedin vitro, using an E. coli 5-30 extract. Next, this precursor wasincubated (60 min at 37 aC) in a Iysate of E. coli as previouslydescribed (van Dijl et al., 1991 a). As a control for the processingcompetence of pre-A2-AmyL under these conditions, thisprecursor was also incubated in the presence of purified 5Pase Iof E. coli (Lep).

al

~+"'10Eal0\10+"'Cal~al

Q.

20 40 80 100 12060

Time (min)

(Fig. 3b). Unfortunately, sf-SipS-His (Barn) appeared tobe prone to proteolytic degradation, as evidenced bythe co-purification of two SipS-specific degradationproducts (denoted dl and dl).

To characterize the enzymic activity of sf-SipS-His(Barn), a hybrid precursor, denoted pre-Al-AmyL(Smith et at., 1987, 1988) was used. This precursorcontains the first 60 residues of the YvcE protein of B.subtilis, including a typical signal peptide (Bolhuis et at.,1999b). The 35S-labelled pre-Al-AmyL was generatedby in vitro transcription-translation in a B. subtilis S-135 extract and incubated for 16 h with purified sf-SipS-His (Barn), in the absence of added phospholipids and inthe absence or presence of 0.5% Triton X-100. Asshown in Fig. 4(a), a significant fraction of the labelledpre-Al-AmyL was processed to the mature form,irrespective of the presence of Triton X-100. Notably,no mature Al-AmyL was produced upon synthesis andincubation of the corresponding precursor (which wascompetent for processing by the SPase lof E. coli) inextra cts of E. coli (Fig. 4b). This was true even when pre-Al-AmyL was incubated for 16 h in an E. coli extract(data not shown), showing that SPase-like activities areabsent from the cytoplasm of E. coli. Similarly, the wild-type p-lactamase (Bla) precursor and the hybrid pre-cursor pre-Al-Bla (Smith et at., 1987, 1988) remainedunprocessed in extra cts of E. coli (van Dijl et at., 1991b),confirming the absence of cytoplasmic SPase-like ac-tivities in E. coli. Taken together, these observations

Fig. 5. In vitro processing of pre-A2-AmyL by sf-SipS-His (Barn).(a) The pH-dependent activity of sf-SipS-His (Barn). In vitro-synthesized, 3sS-labelled pre-A2-AmyL was incubated (30 minat 37 aC} with 7.5 I!M purified sf-SipS-His (Barn) in 50 mM aceticacid/NaOH (pH 4.0 and pH 5.0), 50 mM Tris/HCI (pH 6.0, pH 7.0and pH 8.0), or 50 mM glycine/NaOH (pH 9.0, pH 9.5 andpH 10.0). Samples we re analysed by SDS-PAGE, fluorographyand scanning of films. The percentage of pre-A2-AmyLprocessing was calculated as the amount of labelled matureAmyL, divided by the total amount of labelled A2-AmyL(precursor + mature A2-AmyL). The results of two independentexperiments are indicated with filled or open bullets. About10% of the labelled A2-AmyL was mature after in vitrosynthesis and no additional processing was observed in mock-treated samples. (b) Temperature dependence of the adivity ofsf-SipS-His (Barn). In vitro-synthesized, 35S-labelled pre-A2-amyLwas incubated for 30 min with 7.5 I!M purified sf-SipS-His (Barn)in 50 mM glycine/NaOH buffer (pH 10.0) at the temperaturesindicated. Processing of pre-A2-AmyL was analysed as in (a). (c)Processing of pre-A2-AmyL as a fundion of time. In vitrosynthesized 35S-labelled pre-A2-amyL was incubated at 30 acwith 7.5 I!M purified sf-SipS-His (Barn) in 50 mM glycine/NaOHbuffer (pH 10.0) for different periods of time. Processing of pre-A2-AmyL was analysed as in (a).

dernonstrate that the purified sf-SipS-His (Barn) wasactive, even in the absence of added phospholipids ordetergents.

In vitro processing of pre-A2-AmyL was used todetermine the pH and temperature optima of sf-SipS-His (Barn). As shown in Fig. 5(a, b), optimal activity ofsf-SipS-His (Barn) was observed at pH 10 and atincubation temperatures between 30 cC and 37 cC. AtpH values above 10, the activity of sf-SipS-His (Barn)was significantly reduced (data not shown). No SPaseactivity was detected at pH values lower than 5 ortemperatures below 4 cC (data not shown) or above50 cC. Finally, as pre-A2-AmyL processing by sf-SipS-His (Barn) was frequently incomplete (irrespective of thepresence or absence of Triton X-100), varying betweenabout 40 and 80% of the labelled AmyL, the processingof pre-A2-AmyL was determined as a function of thetime of incubation at optimal pH and temperature. Asshown in Fig. 5(c), maximal levels of pre-A2-AmyLprocessing were observed af ter approximately 20 min ofincubation. As can be inferred from the comparison ofFigs 4(a) and 5(a), the presence of MgSO4 did not affectthe levelof pre-A2-AmyL processing. Notably, theremaining pre-A2-AmyL was not even processed uponthe addition of fresh sf-SipS-His (Barn) or the intactSPase lof E. coli (data not shown), suggesting that thisfraction of the labelled pre-protein was incompetent for

processlOg.

frorn the remaining 53 known SPases of this type (ourunpublished observation) suggests that the E. coli SPaseI could be atypical with respect to its detergent re-quirement in vitro. We consider it unlikely that theobserved detergent-independent in vitro activity of sf-SipS-His (Barn) relates to the particular pre-protein thatwas used in our studies (i.e. pre-A2-ArnyL), rather thanto the structural features of sf-SipS-His (Barn). Dalbeyand co-workers were recently able to show that even theintact SipS (Bsu), which was produced frorn a fusionwith the maltose-binding protein (Carlos et at., 2000),was capable of efficient processing of a pro-OmpAnuclease A hybrid precursor in the absence of detergent(R. E. Dalbey, personal communication). This is thesame precursor that was used to demonstrate thedetergent-dependent activity of the soluble derivative ofthe E. coli SPase I (Tschantz et at., 1995). Moreover, SipS(Bsu) and SipS (Barn) are highly similar proteins (91 %identical residues and conservative replacements ; Meiieret at., 1995). Finally, pre-A2-ArnyL is a typical pre-protein that is efficiently processed, both in B. subtilisand in E. coli (Smith et at., 1987, 1988; van Dijl et at.,1991b; Bolhuis et at., 1999b).

The soluble derivative of SipS (Barn) was shown to haveoptimal activity at alkaline pH and a similar result wasobtained with the wild-type form of this enzyme (ourunpublished observations). These observations are con-sistent with the proposed catalytic mechanism of type ISPases, involving a serine-lysine catalytic dyad (Dalbeyet at., 1997; Paetzel & Dalbey, 1997; Paetzel et at.,1998), as an alkaline pH would allow the active-sitelysine residue to act as a general base in catalysis.Compared to the E. coli SPase, which has optimalactivity between pH 8.5 and 9 (Zwizinski et at., 1981),sf-SipS-His (Barn) displayed a slightly higher pH op-timum (pH 10). Similarly, the lipoprotein-specific (typeII) SPase of E. coli, which appears to belong to a novelfamily of aspartic proteases (Tjalsma et at., 1999b), wasshown to have optimal activity at alkaline pH(Tokunaga et at., 1985). Even though the high pHoptima of these SPases are consistent with their proposedcatalytic mechanisms, they are in apparent conflict withthe fact that the catalytic sites of type I and type II SPasesare probably located at the outer surface of thecytoplasmic mernbrane (Dalbey et at., 1997; Tjalsma etat., 1999b), which has an acidic pH. In fact, the pH at theouter surface of the mernbrane is likely to be lower than6, due to the transmernbrane H+ gradient that is presentin living cells. At such low pH values, the type I SPasesat least are barely active, suggesting that their activitycould be regulated by the pH. Thus, the potentiallydeleterious proteolysis of mernbrane proteins by SPaseswould be minimized in the acidic environment at theouter mernbrane surface, until such enzymes are acti-vated to perform their specific task by interactions withpre-proteins or other, as yet unidentified, components ofthe protein export machinery. Alternatively, the pKa ofthe active-site lysine residue of type I SPases could belowered by the hydrophobic interactions with mell\;:.brane components, such as phospholipids. Consistent

DISCUSSION

In the present studies, we have shown that a solublevariant of the B. amyloliquefaciens SPase, SipS, is activeeven in the absence of added detergents or phospho-lipids. In fact, the activity of this so-calIed sf-SipS-His(Barn) protein was not stimulated by the addition ofTriton X-IOO. The latter observation is remarkable, inview of the previously documented observation that theactivity of a soluble derivative of the E. coli SPasestrongly depended on the addition of Triton X-IOO(Tschantz et al., 1995; van Klompenburg et al., 1998).On the basis of the crystal structure of the solublederivative of the E. coli SPase, Paetzel et al. (1998)postulated that the stimulation of SPase activity byTriton X-IOO is due to the presence of a large, highlyhydrophobic surface domain of this enzyme. Asignificant part of this exposed hydrophobic surface isprovided by a p-hairpin between residues 106 and 124 ofthe E. coli SPase (Paetzel et al., 1998). Strikingly, themost hydrophobic part of this p-hairpin (residues106-115) is not conserved in SipS (Barn) or other BacillusSPases (Fig. 1). Thus, the lack of a stimulating effect ofTriton X-IOO on sf-SipS-His (Barn) activity supports theview (Paetzel et al., 1998) that this hydrophobic p-hairpin is responsible for the detergent requirement ofthe E. coli SPase. Interestingly, the region betweenresidues 106 and 124 of the E. coli SPase is only conservedin nine known type I SPases from Bordetella pertussis,

Bradyrhizobium japanicum, Haemophilus influenzae,Helicobacter pylori, Pseudomonas fluorescens, Rhado-bacter capsulatus, Salmonella typhimurium andRickettsia prowazekii. The fact that this region is absent

915

{Associatie Biologische Onderzoeksscholen Nederland) ;].D.H.]. was supported by a grant {805-33.605) from S LW{Stichting Levenswetenschappen) ; and A. d. ] ., S. B. and].M. v.D. we re supported by EU Biotechnology Grants Bio2-CT93-0254, Bio4-CT96-0097, QLK3-CT -1999-00415 andQLK3-CT-1999-00917.

with the latter hypothesis, the crystal structure of the E.coli SPase indicates that hydrophobic interactions be-tween the side-chain of the active-site lysine residue andvarious other residues, such as phenylalanine 133 andtyrosine 143, are important for catalysis (Paetzel et al.,1998) .This view is supported by the recent observationthat the equivalents of the latter two residues in the B.subtilis SPase SipS (i.e. leucine 74 and tyrosine 81) arerequired for the catalytic activity of SipS (Bolhuis et al.,1999a). Whether the hydrophobic environment of themernbrane contributes to the lowering of the pKa of theactive-site lysine residue of type I SPases is presentlyunknown.

Zwizinski et al. (1981) reported that Mg2+ can have aninhibitory effect on the activity of the E. co li SPase I.Interestingly, 5 mM Mg2+ does not seem to inhibit sf-SipS (Barn) .Thus, the presence of MgSO 4 cannot be thereason for the observed incomplete processing of pre-A2-ArnyL shown in Fig. 4(a). Similar levels of processingwere detected in the same buffer (HEPES, pH 7.0 orpH 8.0) without MgSO 4' as shown in Fig. 5 (a) .The mostlikely explanation of why pre-A2-ArnyL preparationsare not completely processed is that some molecules ofthis precursor become incompetent for processing. Invitro this can be due to (micro)aggregation of theprecursor, which leads to translocation and processingincompetence (van Wely et al., 1998) .In fact, this wouldexplain why the use of different precursor preparationsresulted in different maximum levels of processing.Moreover, even in the living cell, pre-proteins canbecome incompetent for processing due to foldingof the mature protein before processing (van Dijlet al., 1991b; Bolhuis et al., 1999c). Notably, the pH-,temperature- and time-dependence of processing wasnot influenced when different pre-protein preparationswere used (data not shown), indicating that the in-complete processing of pre-A2-ArnyL, as observed in thepresent studies, has no bearing on our conclusions.

Finally, our present results show that sf-SipS-His (Barn)is prone to degradation and that this seems to be evenworse for sf-SipC, sf-SipC-His and sf-SipS (Bsu). Thissuggests that the latter three proteins are more sensitiveto degradation than sf-SipS-His (Barn). However, wecannot exclude the possibility that the different pro-duction levels reflect different expression levels. Mostlikely, the degradation of the soluble SPase mutantproteins used in these studies is due both to self-cleavageand to cleavage by cytosolic proteases of E. co li (vanRoosmalen et al., 2000). In fact, as exernplified by sf-SipS-His (Barn), proteolysis is a major bottleneck for theoverproduction, purification and structural analysis ofsoluble derivatives of Bacillus SPases. At present, we aretrying to develop novel strategies to overcome thisproblem.

ACKNOWLEDGEMENTS

We thank Drs A. Bolhuis, H. Tjalsma, A. J. M. Driessen andother members of the European Bacillus Secretion Group forstimulating discussions. M. L. v. R. was supported by theDutch Ministry of Economic Affairs through ABON

Bilgin, N., Lee, J.I., Zhu, H.Y., Dalbey, R. E. & von Heijne, G.(1990). Mapping of catalytically important domains inEscherichia coli leader peptidase. EMBO J 9, 2717-2722.

Bolhuis, A., Sorokin, A., Azevedo, V., Ehrlich, S. D., Braun, P. G.,de Jong, A., Venema, G., Bron, S. & van Dijl, J. M. (1996). Bacillussubtilis can modulate its capacity and specificity for proteinsecretion through temporally controlled expression of the sipSgene for signal peptidase I. Mol Microbiol 22, 605-618.

Bolhuis, A., Tjalsma, H., Stephenson, K., Harwood, C. R., Venema,G., Bron, S. & van Dijl, J. M. (1999a). Different mechanisms forthermal inactivation of Bacillus subtilis signal peptidase mutants.J Biol Chem 274, 15865-15868.

Bolhuis, A., Tjalsma, H., Smith, H., de Jong, A., Meima, R.,Venema, G., Bron, S. & van Dijl, J. M. (1999b). Evaluation ofbottlenecks in the late stages of protein secretion in Bacillussubtilis. Appl Environ Microbiol65, 2934-2941.

Bolhuis, A., Venema, G., Quax, W. J., Bron, S. & van Dijl, J. M.(1999c). Functional analysis of paralogous thiol-disulfide oxido-reductases in Bacillus subtilis. J Biol Chem 274,24531-24538.

Carlos, J. L. , Paetzel, M., Brubaker, G., Karla, A., Ashwell, C. M.,Lively, M. 0., Cao, G., Bullinger, P. & Dalbey, R. E. (2000). The roleof the membra!1e spanning domain of type I signal peptidase insubstrate cleavage site selection. J Biol Chem 275, 38813-38822.

Cregg, K. M., Wilding, E.I. & Black, M. T. (1996). Molecularcloning and expression of the spsB gene encoding an essential typeI signal peptidase from Staphylococcus aureus. J Bacteriol178,

5712-5718.

Dalbey, R. E., Lively, M. 0., Bron, S. & van Dijl, J. M. (1997). Thechemistry and enzymology of the type I signal peptidases. ProteinSci 6, 1129-1138.

van Dijl, J. M., de Jong, A., Smith, H., Bron, S. & Venema, G.(1991a). Non-functional expression of Escherichia coli signalpeptidase I in Bacillus subtilis. J Gen Microbiol137, 2073-2084.

van Dijl, J. M., de Jong, A., Smith, H., Bron, S. & Venema, G.

(1991b). Signal peptidase loverproduction results in increasedefficiencies of export and maturation of hybrid secretory proteinsin Escherichia coli. Mol Gen Genet 227,40-48.

van Dijl, J. M., de Jong, A., Vehmaanpera, J., Venema, G. & Bron,S. (1992). Signal peptidase lof Bacillus subtilis: patterns ofconserved amino acids in prokaryotic and eukaryotic type I signalpeptidases. EMBO J 11,2819-2828.

van Dijl, J. M., de Jong, A., Venema, G. & Bron, S. (1995).Identification of the potential active site of the signal peptidaseSipS of Bacillus subtilis: structural and functional similaritieswith LexA-Iike proteases. J Biol Chem 270,3611-3618.

van Klompenburg, W., Paetzel, M., de Jong, J. M., Dalbey, R. E.,Demel, R. A., von Heijne, G. & de Kruijff, B. (1998).

Phosphatidylethanolamine mediates insertion of the catalyticdomain of leader peptidase in membranes. FEBS Lett 431,75-79.

Kunst. F., Ogasawara, N., Moszer, I. & 148 other authors (1997).The complete genome sequence of the Gram-positive bacteriumBacillus subtilis. Nature 390,249-256.

Kuo, D. W., Chan, H. K., Wilson, C. J., Griffin, P. R., Williams, H. &Knight. W. B. (1993). Escherichia coli leader peptidase: pro-

916

identification of a eubacterial homolog of archaeal and eukaryoticsignal peptidases. Genes Dev 12,2318-2331.

Tjalsma, H., van den Dolder, J., Meijer, W. J. J., Venerna, G., Bron,S. & van Dijl, J. M. (1999a). The plasmid-encoded signal peptidaseSipP can functionally replace the major signal peptidases SipS andSipT of Bacillus subtilis. J Bacteriol181, 2448-2454.

Tjalsma, H., Zanen, G., Venerna, G., Bron, S. & van Dijl, J. M.(1999b). The potential active site of the lipoprotein-specific (type

II) signal peptidase of Bacillus subtilis. J Biol Chem 274,28191-28197.

Tokunaga. M., Loranger. J. M., Chang, S. Y., Regue, M.. Chang, S.& Wu, H. C. (1985). Identification of prolipoprotein signalpeptidase and genomic organization of the lsp gene in Escherichiacoli. J Biol Chem 260,5610-5615.

Tschantz, W. R., Sung, M., Delgado-Partin, V. M. & Dalbey, R. E.(1993). A serine and a lysine residue implicated in the catalyticmechanism of the Escherichia coli leader peptidase. J Biol Chem268, 27349-27354.

Tschantz, W. R., Paetzel, M., Cao, G., Suciu, D., Inouye, M. &Dalbey, R. E. (1995). Characterization of a soluble, catalyticallyactive form of Escherichia coli leader peptidase: requirement ofdetergent or phospholipid for optimal activity. Biochemistry 34,3935-3941.

Vehmaanpera. J., G()rner, A., Venema, G., Bron, S. & van Dijl,J. M. (1993).ln vitro assay for the Bacillus subtilis signal peptidase

SipS: systems for efficient in vitro transcription-translation andprocessing of precursors of secreted proteins. FEMS MicrobiolLett 114,207-214.

Welsh, K. M., Trach, K. A., Folger, C. & Hoch, J. A. (1994).Biochemical characterization of the essential GTP-binding pro-tein Obg of Bacillus subtilis. J Bacteriol176, 7161-7168.

van Wely, K. H. M., Swaving, J. & Driessen, A. J. M. (1998).Translocation of the precursor of alpha-amylase into Bacillussubtilis membrane vesicles. Eur J Biochem 255, 69Q-697.

Wertman, K. F., Wyman, A. R. & Botstein, D. (1986). Host/vectorinteractions which affect the viability of recombinant phagelambda clones. Gene 49, 253-262.

Wolfe, P. B., Wickner, W. & Goodman. J. M. (1983). Sequence ofthe leader peptidase gene of Escherichia coli and the orientationof leader peptidase in the bacterial envelope. J Biol Chem 258,12073-12080.

Wu, x. C.. Lee, W.. Tran, L. & Wong, S.-L. (1991). Engineering aBacillus subtilis expression-secretion system with a straindeficient in six extracellular proteases. J Bacteriol173, 4952-4958.

Zwizinski, C., Date, T. & Wickner, W. (1981). Leader peptidase isfound in both the inner and outer membranes of Escherichia coli.J Biol Chem 256,3593-3597.

duction of an active form lacking a requirement for detergent and

development of peptide substrates. Arch Biochem Biophys 303,274-280.

Meijer, W. J. J., de Jong, A., Wisman, G. B. A., Tjalsma, H.,

Venema, G., Bron, S. & van Dijl, J. M. (1995). The endogenousBacillus subtilis (natto) plasmids pT A1015 and pT A1040 contain

signal peptidase-encoding genes : identification of a new structuralmodule on cryptic plasmids. Mol Microbiol17, 621-631.

Paetzel, M. & Dalbey, R. E. (1997). Catalytic hydroxyl/aminedyads within serine proteases. Trends Biochem Sci 22, 28-31.

Paetzel, M., Dalbey, R. E. & Strynadka, N.-C. J. (1998). Crystalstructure of a bacterial signal peptidase in complex with a beta-lactam inhibitor. Nature 396, 186-190.

Pugsley, A. P. (1993). The complete general secretory pathway in

gram-negative bacteria. Microbiol Rev 57,50-108.

van Roosmalen, M. L., Jongbloed, J. D. H., Kuipers, A., Venema,G., Bron, S. & van Dijl, J. M. (2000). A truncated soluble Bacillus

signal peptidase produced in Escherichia coli is subject to self-cleavage at its active site. J Bacteriol182, 5765-5770.

Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). MolecularCloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor,NY: Cold Spring Harbor Laboratory.

Smith, H., Bron, S., van Ee, J. & Venema, G. (1987). Constructionand use of signal sequence selection vectors in Escherichia coliand Bacillus subtilis. J Bacteriol169, 3321-3328.

Smith, H., de Jong, A., Bron, S. & Venema, G. (1988). Charac-terization of signal-sequence-coding regions selected from theBacillus subtilis chromosome. Gene 70,351-361.

Stover, A. G. & Driks, A. (1999a). Secretion, localization, andantibacterial activity of TasA, a Bacillus subtilis spore-associated

protein. J Bacteriol181, 1664-1672.

Stover, A. G. & Driks, A. (1999b). Control of synthesis andsecretion of the Bacillus subtilis protein Y qxM. J Bacteriol181,7065-7069.

Studier, F. W., Rosenberg, A. H., Dunn, J. J. & Dubendorff, J. w.(1990). Use of T7 RNA polymerase to direct expression of clonedgenes. Methods Enzymol6, 60-89.

Thompson, J. D., Higgins, D. G. & Gibson, T. J. (1994). CLUSTAL w:improving the sensitivity of progressive multiple sequence align-ment through sequence weighting, position-specific gap penaltiesand weight matrix choice. Nucleic Acids Res 22, 4673-4680.

Tjalsma, H., Noback, M. A., Bron, S., Venema, G., Yamane, K. &van Dijl, J. M. (1997). Bacillus subtilis contains four closely relatedtype I signal peptidases with overlapping substrate specificities.

Constitutive and temporally controlled expression of different sipgenes. J Biol Chem 272, 25983-25992.

Tjalsma, H., Bolhuis, A., van Roosmalen, M. L., Wiegert, T.,Schumann, W., Broekhuizen, C. P., Quax, W. J., Venema, G.,Bron, S. & van Dijl, J. M. (1998). Functional analysis of the

secretory precursor processing machinery of Bacillus subtilis :

Received 17 Ju I y 2000; revised 13 November 2000; accepted 7 December

2000.