JOURNAL OF BACTERIOLOGY, Dec. 1992, p. 7919-79250021-9193/92/247919-07$02.00/0

Vol. 174, No. 24

Cloning, Sequencing, and High Expression of the ProlineIminopeptidase Gene from Bacillus coagulans

ANA KITAZONO, TADASHI YOSHIMOTO,* AND DAISUKE TSURUSchool ofPharmaceutical Sciences, Nagasaki University, Nagasaki 852, Japan

Received 21 August 1992/Accepted 13 October 1992

The gene coding for proline iminopeptidase in BaciUus coagulans was cloned and expressed in Escherichiacoli. Nucleotide sequencing revealed an 861-bp open reading frame with an unusual TTG initiation codon,encoding a 287-amino-acid protein. The calculated molecular weight of the product was 32,415. The amino acidsequences of the amino-terminal region and those of some peptide fragments obtained by endoproteinase Asp-Ndigestion of the purified enzyme completely coincided with those deduced from the nucleotide sequence. Therare TTG initiation codon that normally codes for leucine was translated as a formal initiation codon; amethionine residue was found at the amino terminus of the enzyme. By using a vector bearing the strong tacpromoter, an expression level as high as 200-fold that of the first clone was achieved. The replacement of theTTG initiation codon with ATG and a simultaneous reduction of the distance to the tac promoter resulted ina further increase of 2.5-fold. The expressed enzyme was easily purified to homogeneity by hydrophobicchromatography on a Toyopearl HW-65C column and crystallization, with a recovery of activity of36%. Themolecular weight was found to be 33,000 by both sodium dodecyl sulfate-polyacrylamide gel electrophoresisand gel filtration on a Hi-Load 16/60 Superdex 200 fast protein liquid chromatography column. The expressedenzyme showed the same catalytic and physicochemical properties as those of the wild type, specifically cleavingthe N-terminal proline from small substrates.

Proline iminopeptidase (EC 3.4.11.5) activity was reportedfor the first time by Sarid et al. (24, 25), who found it in aprolineless mutant of Escherichia coli that could utilizepoly-L-proline in place of the required amino acid. Afterthese two reports, however, there have not been any othersabout proline iminopeptidase activity in E. coli. Recently,Fanghanel et al. (3) reported the use of L-proline iminopep-tidase activity for the identification of Serratia and Hafniaspecies. These strains hydrolyzed L-proline-4-nitroanilide, incontrast to 18 other Enterobacteriaceae species including E.coli. The enzyme activity has also been detected in a varietyof organisms, i.e., some human oral cavity microorganisms(14), Neisseria gonorrhoeae (2), Bacillus megatenium (36),Bacillus coagulans (38), Lyophyllum cinerascens (26), andapricot seeds (18). Furthermore, a proline iminopeptidasewas partially purified from Bacillus brevis and used for theenzymatic synthesis of proline-containing peptides (20).However, only the enzymes from B. coagulans and apricotseeds have been purified homogeneously.

In mammals, an activity that cleaved the N-terminalproline of melanocyte-stimulating-hormone-release-inhibit-ing factor (Pro-Leu-GlyNH2) from pig (19) and bovine kid-ney (10), and recently, the detection of a proline iminopep-tidase activity in rat liver and kidney have also been reported(9). However, the presence of a real proline iminopeptidasein mammalian tissues has remained uncertain, since otheraminopeptidases like leucine aminopeptidase (32), and evencarboxylesterases (8, 16) have a weak activity on the amidesubstrate used in the proline iminopeptidase assay.This paper deals with the cloning, sequencing, and expres-

sion in E. coli of the proline iminopeptidase gene from B.coagulans. The attempts to improve the expression level ofthe gene and the enzymatic characterization of the expressedenzyme are also described.

* Corresponding author.

MATERIALS AND METHODSMaterials. Restriction enzymes, BAL 31 nuclease, T4

DNA ligase, and deletion and M13 sequencing kits andprimers for sequencing were purchased from Takara ShuzoCo. and Toyobo Co., respectively. 35S-dCTP and [32P]dCTPwere from Amersham. Sequenase was obtained from U.S.Biochemical Corp., and agarose I was from Dojin Chemi-cals. Pro- -naphthylamide (Pro-PNA), Fast Garnet GBCsalt, lysozyme, and RNase A were from Sigma. Alkalinephosphatase from calf intestine and Pseudomonas fragiendoproteinase Asp-N were obtained from BoehringerMannheim. Oligonucleotides were synthesized by the phos-phoramidite method with a Pharmacia LKB DNA synthe-sizer.

Bacterial strains, plasmids, and media. E. coli HB101 [F-hsdS20 (rB- mB-) recA13 ara-14proA2 lacYl galK2 rpsL20(Smr) xyl-S mtl-l supE44m X-], XL1-blue [lac (F' TnlOproAB lacIq Z M15)], JM-83 [ara A(lac-proAB) rpsL (= strA)A80rlacZAM15], DH1 [F- recAl gyrA96 thi-1 hsdR17 (rK-MK-) supE44 rell X-], and DH5a [supE44 AlacU169(48OlacZAM15) hsdRI7 recAl endAl gyrA96 thi-1 reLAl]were used as hosts. Plasmids pBR322, pFMOO5, pUC18, andpUC19 were used for cloning and sequencing. PlasmidpFMOO5 a deleted derivative of pKK223-3 (Pharmacia LKBBiotechnology) that bears the strong tac promoter, was agenerous gift of F. Misoka.Phagemids Bluescript SK (+) and (-) were used for

routine cloning procedures, for the construction of deletionmutants, and for nucleotide sequencing with single-strandtemplates. Bacteria were grown in Luria-Bertani broth at300C.

Isolation of plasmid DNA and transformation. PlasmidDNAwas isolated by the alkaline extraction procedure (5) orby CsCl-ethidium bromide equilibrium density gradient cen-trifugation. Competent cells for transformation were pre-pared by using the rubidium chloride treatment (15).

Preparation of gene library and screening. B. coagulans

7919

7920 KITAZONO ET AL.

A

-EcoRI/AccI digestion-Insertion of the adaptor

AATTATA-EcoRI/PstI digestion-Ligation with EcoRI/Pstldigested pFMOO5

SD M Y T BTTTCACACAGGAAACAGAATTCATGTATACGGAAGGAAAGTGTGTCCTTTGTCTTAAGTACATATGCCITCC

FIG. 1. Schematic representation of the construction of pPI-20.The TTG initiation codon was replaced with ATG by using anadaptor as described in Materials and Methods. The final sequenceat the junction of vector with insert in pPI-20 is shown at the bottomof the figure.

chromosomal DNA was isolated by the method of Saito andMiura (22), completely digested with each of the restrictionenzymes HindIII, EcoRI, and PstI, and ligated with asimilarly cleaved and dephosphorylated plasmid, pBR322.Each of the hybrid plasmid mixtures obtained was used totransform E. coli DH1. The transformants, selected in ampi-cillin- or tetracycline-containing Luria-Bertani plates, werescreened for those showing proline iminopeptidase enzymeactivity with Pro-,BNA as a substrate.

Nucleotide sequencing. pPI-10, pPI-14, and the deletedplasmids obtained by exonuclease III digestion (6) weresequenced by the dideoxy-mediated chain terminationmethod (23) by using either double-stranded plasmid orsingle-stranded phagemid DNA as templates. dGTP wasreplaced by dITP in some reaction mixtures to overcomecompression.Replacement of the TTG initiation codon with an ATG

codon by the insertion of an adaptor (Fig. 1). Taking advan-tage of the presence of anAccI site immediately downstreamof the initiation codon, an attempt was made to insert at theAccI an adaptor with a sequence that could change the start

codon from TTG to ATG without causing any shift on thetranslational frame. Two oligonucleotides with sequences5'-GGAATTCATGT-3' and 5'-CClT'AAGTACATA-3' weresynthesized. After the double-stranded adaptor was digestedwith EcoRI, two cohesive ends-one for AccI and one forEcoRI-were obtained. This fragment was inserted intoEcoRI-AccI-digested pPI-14, and the EcoRI-PstI fragmentwas ligated with a similarly digested vector, pFMOO5, result-ing in the formation of pPI-20. Since a ribosome binding siteon the plasmid is located 4 bp upstream of the EcoRI site, itcould direct the translation of the proline iminopeptidasegene insert (4, 7).Enzyme activity assay. The proline iminopeptidase activity

was assayed by measuring the amount of 3NA liberatedfrom Pro-,BNA, as described previously (36). One unit of theenzyme activity was defined as the amount of enzymereleasing 1 ,umol of INA per min. The protein concentrationwas determined by the method of Bradford (1) or by theA280.

Culture of E. coli transformants for the production ofproline iminopeptidase. E. coli DH5a harboring pPI-20 (seeFig. 2) was grown in 12 liters of N-broth (1% meat extract,1% polypeptone, and 0.5% NaCl, pH 7.0, containing ampi-cillin [50 ,ug/ml]) with a fermentor at 30°C. Bacterial growthwas monitored by measurement of A650.

Purification of proline iminopeptidase. All of the purifica-tion procedures were done at 4°C. The washed cells (50 g,wet weight) were suspended in 800-ml 20 mM Tris-HClbuffer, pH 8.0, and disrupted for 10 min with glass beads ina Dyno-Mill. The glass beads were removed by decantation,and the disrupted cell suspension was centrifuged at 8,000 xg for 15 min. After the supernatant was treated with prota-mine sulfate solution (18 mg/g of cells, wet weight), themixture was kept for 30 min and centrifuged at 8,000 x g for15 min. The supernatant was fractionated with ammoniumsulfate from 40 to 80% saturation. The precipitate wasdissolved in about 800-ml 20 mM Tris-HCl buffer, pH 8.0,containing ammonium sulfate at 40% saturation, and theclear solution was applied to a column (6 by 15 cm) ofToyopearl HW-65C equilibrated with the above buffer. Thecolumn was washed with the same buffer, and the adsorbedenzyme was eluted with a decreasing linear gradient ofammonium sulfate concentration from 40 to 0% saturation in20 mM Tris-HCl buffer, pH 8.0. The active fractions werecombined and precipitated with ammonium sulfate at 80%saturation. The resultant precipitate was dissolved in 10-ml20 mM Tris-HCl buffer, pH 8.0, and dialyzed in a cellulosetube against the same buffer containing ammonium sulfate at40% saturation. Needle-like crystals that appeared duringovernight incubation were collected by centrifugation anddissolved in 20 mM Tris-HCl buffer, pH 8.0. After recrys-tallization, the enzyme solution was dialyzed against waterand lyophilized.Enzymatic studies. For substrate specificity studies, the

liberated proline was determined by using ninhydrin (30, 34,35) or high-performance liquid chromatography (HPLC), forsubstrates other than pNAs. The method described by Tsuruet al. (31) was used for the dye-sensitized photooxidationstudies.

Preparation of big crystals by the hanging-drop vapor-diffusion method. The drops were formed by 10 IlI of theprotein and 10 ,ul of the reservoir solution, in which theprotein solution consisted of 10.4 mg of the enzyme per ml in20 mM Tris-HCl, pH 8.0. The reservoir solution contained1.3 M ammonium sulfate, 1 mM EDTA, 1 mM dithiothreitol,0.1% sodium azide, and 120 mM phosphate buffer, pH 6.5.

J. BACTERIOL.

PROLINE IMINOPEPTIDASE GENE FROM BACILLUS COAGUL4ANS 7921

VectorPlasmid promoter

direction

H

pPI-1

pPI-2 l-

pPI-3 -E:

pPI-4 --

BPS AK V A

LiF2~ 'IS

B

P E uI I I

I I I I0 1 2 3 4

pPI-Sa -t:

pPI-b -5>

pPI-10

pPI-14

pPI-18

9.97.5

1.1

1.3

200

pPI-20 -

FIG. 2. Restriction map of the proline iminopeptidase gene from B. coagulans. The solid arrow indicates the direction of transcription ofthe gene. The open affows show the orientation of the promoter on the vector. For each plasmid, the vector was as follows: for pPI-1,pBR322; for pPI-2 to pPI-14, pUC18 or pUC19; and for pPI-18 and pPI-20, pFMOO5. B, BamHI; P, PstI; S, ScaI; A, AccI; V, EcoRV; K,KpnI; E, EcoRI; H, HindIII. E. coli DH1 was the host for plasmids pPI-1 to pPI-4; strain JM-83 was used for pPI-5 to pPI-14, and strain DH5awas used for pPI-18 and pPI-20. No significant variation in the expression level could be observed for any of those strains, but for the lasttwo plasmids, the enzyme production was not steady when strain JM-83 was used as the host.

Small crystals obtained by the dialysis procedure were usedas seeds.SDS-PAGE. Sodium dodecyl sulfate-polyacrylamide gel

electrophoresis (SDS-PAGE) was performed with polyacryl-amide (10%) gels and staining with Coomassie brilliant blueR-250 (13).Amino acid sequence and composition of the expressed

enzyme. The amino-terminal region of the enzyme wassequenced by manual Edman degradation (11, 29). Aminoacid sequences of some endoproteinase Asp-N digests of theexpressed enzyme were also determined after isolation withan HPLC Vydac C18 reversed-phase column (12). Aminoacid composition of the enzyme was determined after hy-drolysis with 6 N HCl for 24, 48, and 80 h.

Nucleotide sequence accession number. The GenBank-EMBL-DDBJ accession number for the B. coagulans pro-line iminopeptidase gene is D11037.

RESULTS

Screening for the enzyme gene and subcloning. Amongapproximately 3,000 transformants screened, one showedenzyme activity and was found to be harboring a plasmidwith a 12-kbp insert in the HindIII site (pPI-1). The pBR322vector in pPI-1 was changed to pUC19 by ligating the4.5-kbp BamHI-HindIII insert fragment with a similarlydigested pUC19, constructing pPI-2. The detailed restrictionenzyme map for pPI-2 is shown in Fig. 2. E. coli cells

harboring this plasmid showed an enzyme activity level1.6-fold that conferred by pPI-1.

Thereafter, pPI-4, pPI-5a, and pPI-5b were obtained byligating the 3.2-kbp fragment to the PstI and SacI sites ofpUC19, respectively. After the disappearance of the enzymeactivity when reducing the length of the insert, the gene wasfound to be located in the fragment harboring the KpnI andEcoRI restriction sites.By digesting pPI-5a with ScaI and pUC19 with HincII and

ligating the blunt ends, a hybrid plasmid (pPI-10) with a

2.1-kbp insert was constructed. By making further unidirec-tional deletions on the insert with exonuclease III digestion(6), pPI-14, a plasmid harboring the smallest insert (1.4 kbp)that still gave constitutive levels of proline iminopeptidaseactivity, was obtained. This plasmid, together with pPI-10,was used for the determination of the nucleotide sequence.

Nucleotide sequence. The nucleotide sequence of the insertin plasmid pPI-10 was determined. Within this sequence,there was an open reading frame consisting of 861 bpbeginning at an unusual TTG codon (Fig. 3). The TTGcodon, which normally codes for leucine, was translated as

a formal initiation codon, and a methionine residue wasfound at the amino terminus of the enzyme by amino acidsequencing. Peptide fragments obtained by endoproteinaseAsp-N digestion of the expressed enzyme gave amino acidsequences that were in agreement with those deduced fromthe nucleotide sequence, as shown in Fig. 3. Similar goodagreement was seen for the values obtained for the deduced

Enz. Activ.Fold

H

1.0I12 kb

1.6kb

0

3.4

I I

I

I I I

i I

VOL. 174, 1992

7922 KITAZONO ET AL.

TGGTAATCTTTGTCGGAATTGGACTAATATTTGGTATAATCGGATGGAAAAAGAAACAATACGCTAAAATGTATGGTGGGTTCGGTCATTTTTTCGGTGTTATCATCAGCTGGATCTTAAGTGTCCTTGTTAGTTTGGATGATCCCAGCAAGGGAATTGCCATTATGGTTCTTTATCCGCTATTTTCTATCGGGGGCTGGTTAATTGGTTTGCACTTTGGGAAAGAAAAGGATAAGCGTACTAGGGATTAACAATAAGGGGGAATGAAGT

1 TTGTATACGGAAGGATTTATCGATGTTACCGGCGGTAGGGTTTCTTTTCAAAAATTTGATGAAAACGGCGGGGGTACCCCGGTTATTGTTTTGCACGGGGGTCCGGGGTCTTCTTGTTAT 120M Y T E G F I D V T G G R V S F Q K F D E N G G G T P V I V L H G G P G S S C Y

121 TCGCTGCTTGGTTTAAAGGCGCTCGCAAAGGATCGGCCAGTAATTTTATATGACCAATTGGGATGCGGAAAGTCCGATCGGCCAATGGATACGACGTTATGGCGCCTCGACCGTTTTGTA 240S L L G L K A L A K D R P V I L Y D Q L G C G K S D R P M D T T L W R L D R F V

241 GAAGAATTAGCACAAATTAGACAGGCGCTTAACCTTGATGAAGTACATATCTCGGTCACTCCTGGGGACTACGCTCGCAGCTGCTTATTGCTTAACGAAGCCAAGCGTGTGAAAAGTGTA 360E E L A Q I R Q A L N L D E V H I S V T P G D Y A R S C L L L N E A K R V K S V

361 ATTTTTTCAAGTCCATGTCTCAGTGCGCCGTTATGGGAACAGGACCAAAAGCGAAATCTAAAGAAATTGCCGCTTGATGTGCAGGAAACCATCAACCGCTGTGAAGAAAATGGCACGACC 480I F S S P C L S A P L W E Q D Q K R N L K K L P L D V Q E T I N R C E E N G T T

481 GATTCAGAAGAATTTGCAGCGGCCATTGAGGTATTTGGAAAGCATTTTGTTAATCGGCTTGAGAAGCAGCCGGAATGGCTGGAGCAAAAACCATCAGGATATCGGAATGCTGACATTTAC 600D S E E F A A A I E V F G K H F V N R L E K Q P E W L E Q K P S G Y R N A D I Y

601 AATATTATGTGGGGACCGTCTGAATTTACCGTGCTTGGAAATTTGAAGAATTTTGATTGCACCACACAATTAAAAGAAATTACCTGCCCCTCTCTATATACGTGTGGGCGTTTTGACGAA 720N I M W G P S E F T V L G N L K N F 0 C T T Q L K E I T C P S L Y T C G R F D E

721 GCTACACCGGAAACGACGGAGTATTACAGCAGCCTGACACCGAAGTCGAAATTCCATGTTTTTGAAAAAAGTGCACACATGCCATACATTGAGGAGCCGGAGGAGTATTTAGCGGTAATT 840A T P E T T E Y Y S S L T P K S K F H V F E K S A H M P Y I E E P E E Y L A V I

841 GGAGATTTTCTTAATTCTATTTAATAATCTTTTAGTTCGAAGCGATTAAAGGAGTGGAGCTGATCCCTTTCAGGTTCCACTCCTTTTTTATGTGAAAATGGGAAAATAAAACCATCTCAGG D F L N S I * *CATCAATTATCAATAATCAGTTAATTATATGGACCAATTAAACCAGGAATAAAGGATTATTTTTGGAAAATATAGGAATCAAACCCTATGAAAAATGTAAGAAAATAGTTATAAAGTTTTATTTTAAACACAAAACAAGGTCTGTCTACATGCACTTGGACCAATTTTCCATTTTTCCCAACAACAACCAAACTATCATTTTCCT

FIG. 3. Nucleotide sequence of the B. coagulans proline imiinopeptidase gene and its deduced amino acid sequence. The anuno acidresidues determined by the manual Edman method are underlined. Dotted underlines show the putative promoter region, ribosome bindingsite, and the inverted repeat sequence downstream of the stop codon.

amino acid composition of the enzyme and for those ob-tained by amino acid analysis (data not shown).

Preceding the TTG initiation codon, there was a putativeribosome binding site. A putative promoter region was alsofound, although it was far upstream from the ribosomebinding site (around 75 bp). At the 3'-flanking side, two TAAstop codons in tandem arrangement and, downstream, apalindromic sequence typical for transcription terminationwere observed.

Expression of the proline iminopeptidase gene. In an at-tempt to effectively improve the expression level, the vectorpUC19 was replaced by pFMOO5 (Fig. 2). After digestion ofpPI-14 with EcoRI and PstI, the 1.4-kbp fragment carryingthe proline iminopeptidase gene was isolated and ligated to asimilarly digested pFMOO5, constructing pPI-18. The trans-formants harboring pPI-18 showed a level of enzyme activitynearly 200-fold higher than that conferred by pPI-1 (Fig. 2and 4). Meanwhile, E. coli DH5a transformed with plasmidpPI-20 (obtained after changing the initiation codon fromTTG to ATG) showed a level of activity 500-fold thatconferred by pPI-1. The effective change of the initiationcodon in pPI-20 was confirmed by nucleotide sequencing.

Purification and crystallization of the enzyme. The purifi-cation procedure is summarized in Table 1. The enzymeeasily crystallized in the presence of ammonium sulfate. Byrepeating the crystallization, the minor contaminants couldbe removed. As shown in Fig. 5, the final preparationpresented a single band on SDS-PAGE. The level of recov-ery of activity was 36%. Through the dialysis procedure(Fig. 6A), needle-shaped crystals were formed after 2 to 3days, and by the hanging-drop vapor-diffusion method,larger prismatic crystals appeared after a few weeks at 25°C(Fig. 6B).

Properties of the expressed enzyme. The optimum pH forthe enzymatic activity was 8.0, and the enzyme was stable ina pH range of 5.5 to 7.5. The maximum activity wasobserved at 40°C, and the enzyme was stable at tempera-tures up to 38°C (50% of the activity remained after 15-minpreincubations at pH 8.0). These determinations were madeunder the conditions described previously (38), except thatPro-PNA was used as the substrate.

The molecular weight was estimated to be 33,000 by bothgel filtration and SDS-PAGE, agreeing with the value de-duced from the nucleotide sequence.

Effects of chemicals and photooxidation on the enzymeactivity and substrate specificity. As shown in Table 2, theenzyme was completely inhibited by p-chloromercuriben-zoate (PCMB) and heavy metal salts. Serine enzyme inhib-itors like diisopropylfluorophosphate and phenylmethylsul-fonyl fluoride, as well as metal chelators such as EDTA ando-phenanthroline and the aminopeptidase inhibitor bestatin,had no effect on the activity. The enzyme was also com-pletely inhibited by methylene blue-sensitized photooxida-tion (data not shown). The enzyme released the amino-terminal proline of dipeptides (Pro-Pro, Pro-Gln, Pro-Trp,

1 2 3 4 5 6 7 8

kDa

- 94

-67

-43

-30

-20.1

FIG. 4. SDS-PAGE of bacterial lysates of E. coli transformantsharboring pPI plasmids. Electrophoresis was done with 10% gels,and proteins were stained with Coomassie brilliant blue R-250. TheE. coli transformants contained the following hybrid plasmids (bylane): 1, pPI-2; 2, pPI-4; 3, pPI-Sa; 4, pPI-5b; 6, pPI-10; 7, pPI-18;and 8, pPI-20. The protein molecular mass standards were run inlane 5 (Calibration kit; Pharmacia).

J. BACTERIOL.

PROLINE IMINOPEPTIDASE GENE FROM BACILLUS COAGULANS 7923

TABLE 1. Purification of proline iminopeptidase fromE. coli DH5o(pPI-20)

Total Purifi-Protein actit Recovery Sp act cati

(g) (U, 10-13) (%) (101) (fold)

Cell extract 2.50 24.0 100 9.6 1.0Ammonium sulfate 1.70 19.0 79 11.0 1.1

fractionationHW-65C chroma- 0.67 12.0 50 18.0 1.9

tographyCrystallization 0.31 8.7 36 28.0 2.9

and Pro-Tyr), amides (Pro-,NA), and oligopeptides such asmelanocyte-stimulating-hormone-release-inhibiting factor 1(Pro-Leu-GlyNH2), Pro-Leu-Gly, and Pro-Phe-Gly-Lys.However, it was inactive against poly-L-proline, Met-Pro,and amino acyl amides other than Pro-PNA (Pyr-,BNA,Phe-PNA, Cys-,NA, Met-,NA, Leu-,NA, Ala-,NA, andZ-Gly-Pro-PNA). No activity was detected onp-nitrophenylacetate.

DISCUSSION

The B. coagulans proline iminopeptidase gene was clonedin E. coli, and its nucleotide sequence was determined.Furthermore, after construction of adequate expression vec-tors, levels of enzyme production almost 1,000-fold higherthan that observed in B. coagulans could be obtained.As far as we know, this is the first report about the

cloning, sequencing, and high levels of expression of aproline iminopeptidase gene. Recently, Matsushima et al.(16) reported the nucleotide sequence of a porcine liverproline-,NAase. This enzyme was, however, finally identi-fied as a carboxylesterase on the basis of its high level ofhomology with rat and rabbit liver carboxylesterases as wellas the strong activity toward p-nitrophenyl acetate (16, 27,28).

1 2 3 4

_... ...

_. . -_1

FIG. 5. SDS-PAGE of preparations obtained through the purifi-cation procedure. Lanes show the steps of the procedure as follows:1, cell extract; 2, after fractionation with ammonium sulfate at 80%saturation; 3, after hydrophobic chromatography (Toyopearl HW-65C); and 4, after recrystallization.

$rK\ kh.ksK XFIG. 6. Proline iminopeptidase crystals. (A) Crystals obtained

by dialysis against 20 mM Tns-HCI buffer containing ammoniumsulfate at 40% saturation. (B) Crystals obtained by the hanging-dropvapor-diffusion method, with ammonium sulfate used as the precip-itating agent.

B. coagulans proline iminopeptidase was found to beencoded by a gene with a size of 861 bp, giving a protein witha size of 287 amino acid residues. The correspondence of thecloned gene with the characterized enzyme was confirmedby the agreement of the determined amino acid sequencesand amino acid composition of the expressed enzyme withthose deduced from the nucleotide sequence.A search for homology with other proline-specific pepti-

dases-aminopeptidase P (37), prolidase, prolyl endopeptid-ase-some aminopeptidases, such as pyroglutamyl ami-nopeptidase, E. coli aminopeptidase N, bovine lens, and E.coli andArabidopsis thaliana leucine aminopeptidases-andcarboxylesterases (16) gave negative results; no enzymewith which B. coagulans proline iminopeptidase sharedsignificant homology could be found.One remarkable feature of the proline iminopeptidase gene

was its unusual TTG initiation codon (Fig. 3). Although thischaracteristic feature is rare among genes derived from E.coli, several cases have been reported on gram-positivebacteria of the Bacillus and Staphylococcus genera (17, 33).Two examples of genes having the triplet UUG as aninitiation codon in E. coli are represented by the adenylatecyclase gene (21) and -y-glutamylcysteine synthetase (33).For these two cases, the replacement of the TTG codon byATG resulted in an increase of the expression level from 1.5-to 4-fold. Similarly, for the gene under study, after changingthe initiation codon to ATG and simultaneously reducing thedistance to the tac promoter, an effective increase of theenzymatic activity of 2.5-fold was achieved.

ia..f..

VOL. 174, 1992

N.B '-,\.-.-f--\ ,l I

i

7924 KITAZONO ET AL.

TABLE 2. Effect of chemicals on enzymatic activity'

Chemical ~~Concn RemaininChemical c(mo(M) activity (%)

Control 100o-Phenanthroline 1.0 103EDTA 1.0 102Phenylmethylsulfonyl fluoride 1.0 102Diisopropylfluorophosphate 1.0 99PCMB 0.5 0

0.1 0lodoacetate 1.0 8Diethylpyrocarbonate 0.1 67Bestatin 2.0 83Leupeptin 1.0 68Guanidine HCI 10.0 26Urea 10.0 66E-64 2.0 77E-475 2.0 104Puromycin 1.0 49Benzamidine HCI 1.0 101Zinc chloride 0.5 0Manganese chloride 0.1 37

a The enzyme was preincubated at 30°C for 10 min with each additive, andthe remaining activity was assayed under the standard conditions.

The overproduction of the enzyme allowed us to purify theenzyme very easily through a single chromatographic stepand crystallization. By repeating the crystallization proce-dure, a preparation showing a single band on SDS-PAGEcould be obtained. Even bigger crystals were obtained byusing the hanging-drop method with samples seeded withpreviously obtained crystals.The physicochemical and catalytic properties of both the

expressed and wild-type enzymes were very similar (38).The enzyme was completely inhibited by PCMB and dye-sensitized photooxidation, suggesting the participation of atleast one of each of the cysteine and histidine residues in thecatalytic function. Almost all of the proline iminopeptidaseshitherto reported were also strongly inhibited by PCMB.Chen and Buchanan (2) described four hydrolases from thecell suspension ofN. gonorrhoeae, one of which had prolineiminopeptidase activity. The enzyme was completely inhib-ited by iodoacetamide and by heavy metal salts. It alsoshowed a high level of sensitivity to photooxidation and todiethyl pyrocarbonate like the present enzyme but could nothydrolyze the Pro-Pro bond, unlike the activities of theenzyme from B. coagulans and B. megaterium (36). Thelatter was also similarly inhibited by PCMB and heavy metalsalts and liberated the amino-terminal proline from peptidesand proline derivatives (36), resembling the B. coagulansenzyme in many aspects. On the contrary, the prolineiminopeptidase from L. cinerascens (shimeji) was resistantto PCMB while being strongly inhibited by prolinol anddiethyl pyrocarbonate (26). The proline iminopeptidase ac-tivity from apricot seeds reported by Ninomiya et al. (18)was strongly inhibited by PCMB and showed a high level ofsensitivity to photooxidation and to diethyl pyrocarbonatebut could not hydrolyze the Pro-Pro bond.

Finally, the identification of the cysteine and histidineresidues that are involved in catalytic function would help usto understand the reaction mechanism and the active sitestructure of the proline iminopeptidases. To accomplish this,site-directed mutagenesis studies are now in progress. Sim-ilarly, since we have obtained large prismatic crystals suit-able for X-ray analyses, these studies would provide valu-

able information about the recognition site for proline thatcauses its specificity and could be even applied to otherproline-specific peptidases.

ACKNOWLEDGMENTS

We are grateful to F. Misoka for the gift of plasmids and thank K.Ito for valuable suggestions and discussions. We also thank S.Morikawa and A. Kanatani for their help in amino acid sequenceanalyses.This work was supported in part by Grants-in-Aid for Scientific

Research from the Ministry of Education, Science and Culture ofJapan. Likewise, A.K. is a recipient of a scholarship from the sameministry and expresses her gratitude.

REFERENCES1. Bradford, M. M. 1976. A rapid and sensitive method for the

quantitation of microgram quantities of protein utilizing theprinciple of protein-dye binding. Anal. Biochem. 72:248-254.

2. Chen, K., and T. Buchanan. 1980. Hydrolases from Neisseriagonorrhoeae. J. Biol. Chem. 255:1704-1710.

3. Fanghanel, S., R. Reissbrodt, and H. Giesecke. 1991. L-Prolineaminopeptidase activity as a tool for identification and differen-tiation of Serratia marcescens, Serratia liquefaciens and Hafniaalvei strains. Zentralbl. Bakteriol. 275:11-15.

4. Goeddle, D. 1990. Gene expression technology. Methods En-zymol. 185:14-36.

5. Hattori, M., and Y. Sakaki. 1986. Dideoxy sequencing methodusing denatured plasmid templates. Anal. Biochem. 152:232-238.

6. Henikoff, S. 1984. Unidirectional digestion with exonuclease IIIcreates targeted break-points for DNA sequencing. Gene 28:351-359.

7. Hershey, J. W. B. 1987. Protein synthesis, 620-627. In F. C.Neidhardt, J. L. Ingraham, K. B. Low, B. Magasanik, M.Schaechter, and H. E. Umbarger (ed.), Escherichia coli andSalmonella typhymurium: cellular and molecular biology, vol.1. American Society for Microbiology, Washington, D.C.

8. Heymann, E., and R Mentlein. 1981. Carboxylesterases-ami-dases. Methods Enzymol. 77:333-344.

9. Hiraoka, B., and M. Harada. 1991. Use of L-prolyl-L-leucylgly-cinamide (MIF-1) for the high-performance liquid chromato-graphic determination of proline iminopeptidase activity in ratliver. J. Chromatogr. 563:142-146.

10. Khilji, M., and G. Bailey. 1978. The purification of a bovinekidney enzyme which cleaves melanocyte-stimulating hormone-release inhibiting factor. Biochim. Biophys. Acta 527:282-288.

11. Kobayashi, R., and G. E. Tarr. 1986. Protein sequencing anal-ysis by manual Edman degradation. Protein Nucleic Acid En-zyme. 31:991-1002. (In Japanese.)

12. Kobayashi, R, T. Yoshimoto, and D. Tsuru. 1989. Completeamino acid sequence of neutral protease from Bacillus subtilisvar. amylosacchanticus. Agric. Biol. Chem. 53:2739-2749.

13. Laemmli, U. K. 1970. Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature (London)227:680-685.

14. Makinen, K. 1969. The proline iminopeptidases of the humanoral cavity-partial purification and characterization. ActaChem. Scand. 23:1409-1438.

15. Maniatis, T., E. Fritsch, and J. Sambrook. 1982. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

16. Matsushima, M., H. Inoue, M. Ichinose, S. Tsukada, K. Mild, K.Kurokawa, T. Takahashi, and K. Takahashi. 1991. The nucleo-tide and deduced amino acid sequences of porcine liver proline-P-naphthylamidase. FEBS Lett. 293:37-41.

17. Mclaughlin, J., C. Murray, and J. Rabinowitz. 1981. Uniquefeatures in the ribosome binding site sequence of the gram-positive Staphylococcus aureus-lactamase gene. J. Biol. Chem.256:11283-11291.

18. Ninomiya, K., K. Kawatani, S. Tanaka, S. Kawata, and S.

J. BAcrERiOL.

PROLINE IMINOPEPTIDASE GENE FROM BACILLUS COAGULANS 7925

Makisumi. 1982. Purification and properties of a proline imi-nopeptidase from apricot seeds. J. Biochem. 92:413-421.

19. Nordwig, A., and H. Mayer. 1973. The cleavage of prolylpeptides by kidney peptidases. Detection of a new peptidasecapable of removing N-terminal proline. Hoppe-Seyler's Z.Physiol. Chem. 354:380-383.

20. Ohshiro, T., K. Mochida, and T. Uwajima. 1992. Enzymaticpeptide synthesis by proline iminopeptidase. Biotechnol. Lett.14:175-178.

21. Reddy, P., A. Peterkofsky, and K. McKenney. 1985. Transla-tional efficiency of the E. coli adenylate cyclase gene: mutatingthe WUG initiation codon to GUG or AUG results in increasedgene expression. Proc. Natl. Acad. Sci. USA 82:5656-5660.

22. Saito, H., and K. Miura. 1963. Preparation of transformingdeoxyribonucleic acid by phenol treatment. Biochim. Biophys.Acta 72:619-629.

23. Sanger, F., S. Nicklen, and A. R Coulson. 1977. DNA sequenc-ing with chain-terminating inhibitors. Proc. Natl. Acad. Sci.USA 74:5463-5467.

24. Sand, S., A. Berger, and E. Katchalski. 1959. Proline iminopep-tidase. J. Biol. Chem. 234:1740-1744.

25. Sarid, S., A. Berger, and E. Katchalski. 1962. Proline iminopep-tidase. II. Purification and comparison with iminodipeptidase(prolinase). J. Biol. Chem. 237:2207-2212.

26. Sattar, A. K. M., T. Yoshimoto, and D. Tsuru. 1989. Purificationand characterization of proline iminopeptidase from Lyophyl-lum cinerascens. J. Ferment. Bioeng. 68:178-182.

27. Takahashi, T., A. Ikai, and K. Takahashi. 1989. Dicyclohexyl-carbodiimide is a specific, reversible inhibitor of proline-3-naphthylamidase. Biochem. Biophys. Res. Commun. 168:1001-1006.

28. Takahashi, T., A. Ikai, and K. Takahashi. 1989. Purification andcharacterization of proline-o-naphthylamidase, a novel enzymefrom pig intestinal mucosa. J. Biol. Chem. 264:11565-11571.

29. Tarr, G. E. 1977. Methods in protein sequencing analysis.Humana Press, Clifton, N.J.

30. Trool, W., and J. Lindsley. 1955. A photometric method for thedetermination of proline. J. Biol. Chem. 215:655-660.

31. Tsuru, D., T. Hirose, and J. Fukumoto. 1971. Studies onbacterial proteases: dye-sensitized photooxidation of neutralsubtilopeptidase amylosacchariticus. J. Biochem. 70:699-705.

32. Turzynski, A., and R. Mentlein. 1990. Prolyl aminopeptidasefrom rat brain and kidney: action on peptides and identificationas leucyl aminopeptidase. Eur. J. Biochem. 190:509-515.

33. Watanabe, K., K. Murata, and A. Kimura. 1987. Effect of achange of unusual initiation codon TTG to ATG by site-directedmutagenesis on -y-glutamyl cysteine synthetase gene expression.Agric. Biol. Chem. 51:1111-1114.

34. Yaron, A., and D. Mlynar. 1968. Aminopeptidase P. Biochem.Biophys. Res. Commun. 32:658-663.

35. Yemm, E. W., and E. C. Cocking. 1955. The determination ofamino acids with ninhydrin. Analyst 80:209-213.

36. Yoshimoto, T., T. Saeki, and D. Tsuru. 1983. Proline iminopep-tidase from Bacillus megaterium: purification and characteriza-tion. J. Biochem. 93:469-477.

37. Yoshimoto, T., H. Tone, T. Honda, K. Osatomi, R. Kobayashi,and D. Tsuru. 1989. Sequencing and high expression of ami-nopeptidase P gene from Escherichia coli HB101. J. Biochem.105:412-416.

38. Yoshimoto, T., and D. Tsuru. 1985. Proline iminopeptidase fromBacillus coagulans. J. Biochem. 97:1477-1485.

VOL. 174, 1992