Portal de Periódicos da CAPES

Three Oligopeptide-binding Proteins Are Involved in the Oligopeptide Transport of Streptococcus thermophilus

2002; Elsevier BV; Volume: 277; Issue: 1 Linguagem: Inglês

10.1074/jbc.m107002200

1083-351X

Peggy Garault, Dominique Le Bars, Colette Besset, Véronique Monnet,

Probiotics and Fermented Foods

The functions necessary for bacterial growth strongly depend on the features of the bacteria and the components of the growth media. Our objective was to identify the functions essential to the optimum growth of Streptococcus thermophilus in milk. Using random insertional mutagenesis on a S. thermophilus strain chosen for its ability to grow rapidly in milk, we obtained several mutants incapable of rapid growth in milk. We isolated and characterized one of these mutants in which anamiA1 gene encoding an oligopeptide-binding protein (OBP) was interrupted. This gene was a part of an operon containing all the components of an ATP binding cassette transporter. Three highly homologous amiA genes encoding OBPs work with the same components of the ATP transport system. Their simultaneous inactivation led to a drastic diminution in the growth rate in milk and the absence of growth in chemically defined medium containing peptides as the nitrogen source. We constructed single and multiple negative mutants for AmiAs and cell wall proteinase (PrtS), the only proteinase capable of hydrolyzing casein oligopeptides outside the cell. Growth experiments in chemically defined medium containing peptides indicated that AmiA1, AmiA2, and AmiA3 exhibited overlapping substrate specificities, and that the whole system allows the transport of peptides containing from 3 to 23 residues. The functions necessary for bacterial growth strongly depend on the features of the bacteria and the components of the growth media. Our objective was to identify the functions essential to the optimum growth of Streptococcus thermophilus in milk. Using random insertional mutagenesis on a S. thermophilus strain chosen for its ability to grow rapidly in milk, we obtained several mutants incapable of rapid growth in milk. We isolated and characterized one of these mutants in which anamiA1 gene encoding an oligopeptide-binding protein (OBP) was interrupted. This gene was a part of an operon containing all the components of an ATP binding cassette transporter. Three highly homologous amiA genes encoding OBPs work with the same components of the ATP transport system. Their simultaneous inactivation led to a drastic diminution in the growth rate in milk and the absence of growth in chemically defined medium containing peptides as the nitrogen source. We constructed single and multiple negative mutants for AmiAs and cell wall proteinase (PrtS), the only proteinase capable of hydrolyzing casein oligopeptides outside the cell. Growth experiments in chemically defined medium containing peptides indicated that AmiA1, AmiA2, and AmiA3 exhibited overlapping substrate specificities, and that the whole system allows the transport of peptides containing from 3 to 23 residues. oligopeptide-binding protein erythromycin optical density chemically defined medium open reading frame insertion sequence Oligopeptide transport systems are key channels between the environment and the inner part of micro-organisms, which have been described in numerous Gram-negative and Gram-positive bacteria. They generally internalize peptides with an ATP-driving force and belong to the ABC transporter family (1Tam R. Saier M.H. Microbiol. Rev. 1993; 57: 320-346Crossref PubMed Google Scholar). They are composed of oligopeptide-binding proteins (OBP),1 which are periplasmic in Gram-negative bacteria and membrane-associated in Gram-positive bacteria; transmembrane proteins that form a channel for the passage of oligopeptides and inner membrane-associated ATPases, which provide the energy for transport. The oligopeptide transport system of Gram-negative bacteria (Escherichia coli and Salmonella typhimurium) transports peptides up to hexapeptides. This size limit seems to be imposed by the outer membrane pores rather than by the transporter itself (2Payne J.W. Smith M.W. Adv. Microb. Physiol. 1994; 36: 1-80Crossref PubMed Google Scholar). In Gram-positive bacteria, the size of the peptides transported is more variable. Peptides from 2 to 7 residues and of 6–7 residues are transported by the oligopeptide transport system ofStreptococcus pneumoniae (3Alloing G. de Philip P. Claverys J.-P. J. Mol. Biol. 1994; 241: 44-58Crossref PubMed Scopus (112) Google Scholar) and Streptococcus gordonii (4Jenkinson H.F. Baker R.A. Tannock G.W. J. Bacteriol. 1996; 178: 68-77Crossref PubMed Google Scholar), respectively. In Lactococcus lactis andListeria monocytogenes, the oligopeptide transport system is capable of internalizing peptides composed of 4–18 amino acids (5Lanfermeijer F.C Picon A. Konings W.N. Poolman B. Biochemistry. 1999; 38: 14440-14450Crossref PubMed Scopus (47) Google Scholar) and of 5–8 amino acids (6Verheul A. Rombouts F.M. Abee T. Appl. Environ. Microbiol. 1998; 64: 1059-1065Crossref PubMed Google Scholar). Thirty years ago, Desmazeaud and Hermier (7Desmazeaud M. Hermier J.H. Eur. J. Biochem. 1972; 28: 190-198Crossref PubMed Scopus (31) Google Scholar) demonstrated that peptides having a mass included between 1000 and 2500 and containing lysine or arginine have a stimulatory effect on the growth of S. thermophilus, whereas larger peptides (masses of 5000) inhibit it. Mixtures of amino acids contained in stimulatory peptides have the same stimulant effect on growth, which demonstrates that stimulatory peptides act as amino acid source. The requirements of S. thermophilus for peptides with specific length were explained by the likely presence of a peptide transport system different from that described in E. coli. The most obvious role of these systems is to supply bacteria with essential amino acids in a low energy fashion. In the case of lactic acid bacteria, which are auxotrophic for several amino acids (8Morishita T. Deguchi Y. Yajima M. Sakurai T. Yura T. J. Bacteriol. 1981; 148: 64-71Crossref PubMed Google Scholar, 9Deguchi Y. Morishita T. Biosci. Biotech. Biochem. 1992; 56: 913-918Crossref PubMed Scopus (45) Google Scholar, 10Cogain-Bousquet M. Garrigues C. Novak L. Lindley N.D. Loubiere P. J. Appl. Bacteriol. 1995; 79: 108-116Crossref Scopus (107) Google Scholar) and which are used to growing in milk, a medium containing a low level of free amino acids (11Thomas T.D. Mills O.E. Neth. Milk Dairy J. 1981; 35: 255-273Google Scholar), the oligopeptide transport system is fundamental to optimal growth. In addition to two di-tripeptide transport systems in L. lactis (a proton motive force-driven di-tripeptide carrier (DtpT) (12Smid E.J. Plapp R. Konings W.N. J. Bacteriol. 1989; 171: 292-298Crossref PubMed Google Scholar, 13Kunji E.R.S. Smid E.J. Plapp R. Poolman B. Konings W.N. J. Bacteriol. 1993; 175: 2052-2059Crossref PubMed Google Scholar, 14Hagting A. Kunji E.R.S. Leenhouts K.J. Poolman B. Konings W.N. J. Biol. Chem. 1994; 269: 11391-11399Abstract Full Text PDF PubMed Google Scholar), and an ATP-driven di-tripeptide transporter (DtpP) (15Foucaud C. Kunji E.R.S. Hagting A. Richard J. Konings W.N. Desmazeaud M. Poolman B. J. Bacteriol. 1995; 177: 4652-4657Crossref PubMed Google Scholar, 16Sanz Y. Lanfermeijer F.C. Konings W.N. Poolman B. Biochem. 2001; 39: 4855-4862Crossref Scopus (14) Google Scholar), an oligopeptide transport system (Opp) (13Kunji E.R.S. Smid E.J. Plapp R. Poolman B. Konings W.N. J. Bacteriol. 1993; 175: 2052-2059Crossref PubMed Google Scholar, 17Tynkkynen S. Buist G. Kunji E. Kok J. Poolman B. Venema G. Haandrikman A. J. Bacteriol. 1993; 175: 7523-7532Crossref PubMed Google Scholar), which internalizes oligopeptides released from caseins by the action of cell wall proteinase, allows L. lactis to grow in milk. In the present work, we have identified and characterized the oligopeptide transport system of S. thermophilus. We demonstrate that it works with three functional oligopeptide-binding proteins, that it is capable of transporting entities as large as 23 amino acid peptides, and that it plays a major role in nutrition. The strains used in this work were described in Table I. All E. coli strains were grown on Luria-Bertani medium (18Sambrook J. Russel D.W. Molecular Cloning: A Laboratory Manual.3rd Ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY2001Google Scholar) at 37 °C with shaking and in the presence of erythromycin (Em, 150 μg/ml) when required. Three media were used for S. thermophiluscultures. The first medium was reconstituted, low heat skim milk (10% w/v) (Nilac, Nederlands Instituut von Zuivelonderzoek, Ede, The Netherlands), autoclaved at 110 °C for 12 min, buffered with 0.75 mm sodium glycerophosphate and, in some cases, containing bactotryptone at 3 g/liter (pancreatic digest of casein; Difco Laboratories, Detroit, MI). Bacterial growth was monitored by measuring optical density (OD) at 480 nm after clarification of the milk by a 10-fold dilution in 2 g/L EDTA pH 12 (19). Two other media were used for general cultures and growth rate experiments. The first was M17Lac medium (20Terzhaghi B.T. Sandine W.E. Appl. Microbiol. 1975; 29: 807-813Crossref PubMed Google Scholar), in which bacterial growth was monitored by measuring the OD at 600 nm. The second was a chemically defined medium (CDM) containing nucleotides, vitamins, salts, potassium phosphate buffer (0.05 mol·liter−1, pH 6.7) and 0.5% lactose (w/v) as described by Letort and Juillard (21Letort C. Juillard V. J. Appl. Microbiol. 2001; 91: 1023-1029Crossref PubMed Scopus (119) Google Scholar), and sterilized by filtration. The nitrogen source of the CDM was provided by amino acids, a mix of amino acids associated with a single peptide (Sigma) or αs2-casein trypsic hydrolysate, sterilized by filtration. Peptides MK, MH, EA, ED, EPET, PQFY, DYM, DYMG, YGGFM, RPKPQQFFGLM, MKRPPGFSPFR, ACTH-(1–17) fragment (SYSMEMFRWGKPVGKKR), ACTH-(1–24) fragment (SYSMEMFRWGKPVGKKRRPVKVYP), and oxidized B chain of insulin (FVNQHLCGSHLVEALYLVCGERGFFYTPKA) were used at rate of 100 μmol·liter−1 in the culture medium. Growth rate experiments were then performed at 37 °C using a Microbiology Reader Bioscreen C (Labsystems, Helsinki, Finland) in 100-well, sterile, covered microplates. Each well contained 200 μl of the culture medium. Overnight M17Lac cultures of S. thermophilus were washed twice and resuspended in a volume of sterile potassium phosphate buffer (0.05 mol·liter−1, pH 6.7) equal to the culture volume. 4 μl of the suspension were used to inoculate each well. The optical density was measured at 600 nm every 20 min, after gentle shaking. The apparent growth rate (μmax) was defined as the maximum slope of a semi-logarithmic representation of growth curves, assessed by OD measurements.Table IStrains and plasmidsStrains E. coliTIL206: TG1 supE hsdΔ5 thiΔ (lac-proAB)F[traD 36− proAB+lac1qlacZΔM15] +repAProvided by P. RenaultTIL401: TIL206 containing pG+h9::ISS1Provided by D. ObisTIL287: TIL206 containing pG+h9Provided by M. Nardi S. thermophilusSt18: wild type strain, PrtS+, plasmid-freeProvided by Rhodia-FoodSt18amiA1−: St18 strainamiA1Δ861 bp AvaII-AvaIIThis studySt18amiA2−: St18 strainamiA2Δ93 bp PshAI-BstSNIThis studySt18amiA3−: St18 strainamiA3Δ313 bp MslI-StuIThis studySt18prtS−: St18 strainprtSΔ2054 bp HpaI-NruIThis studySt18amiA1, prtS−: St18 strainamiA1Δ(AvaII-AvaII);prtSΔ(HpaI-NruI)This studySt18amiA2, prtS−: St18 strainamiA2Δ(PshAI-BstSNI);prtSΔ(HpaI-NruI)This studySt18amiA3, prtS−: St18 strainamiA3Δ(MslI-StuI);prtSΔ(HpaI-NruI)This studySt18amiA1, A2−: St18 strainamiA1Δ(AvaII-AvaII);amiA2Δ(PshAI-BstSNI)This studySt18amiA1, A2, A3−: St18 strain amiA1Δ(AvaII-AvaII);amiA2Δ(PshAI-BstSNI);amiA3ΔMslI-StuIThis studySt18amiA1, A3, prtS−: St18 strain amiA1Δ(AvaII-AvaII) amiA3Δ (MslI-StuI);prtSΔ(HpaI-NruI)This studyIS mutant: St18 insertional mutant containing pG+h9::ISS1 in amiA1geneThis studyPlasmids pGEMtPromega TopoXLInvitrogene pG+h9Ref. 27Biswas I. Gruss A. Ehrlich S.D. Maguin E. J. Bacteriol. 1993; 175: 3628-3635Crossref PubMed Scopus (354) Google Scholar pG+h9ISS1Ref. 23Maguin E. Prévost H. Ehrlich S.D. Gruss A. J. Bacteriol. 1996; 178: 931-935Crossref PubMed Scopus (412) Google Scholar Open table in a new tab Plasmid DNA manipulations and transformations of E. coli were performed as described previously (18Sambrook J. Russel D.W. Molecular Cloning: A Laboratory Manual.3rd Ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY2001Google Scholar). RNA was prepared as previously described from S. thermophilus grown in M17Lac (18Sambrook J. Russel D.W. Molecular Cloning: A Laboratory Manual.3rd Ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY2001Google Scholar). The total DNA of integrants obtained by insertional mutagenesis (see below) were digested byEcoRI or HindIII, and then ligated. TIL206 electrocompetent cells were transformed with ligation products, and EmR colonies were screened by PCR after 24-h incubation at 37 °C. PCR amplifications were performed with the Gene Amp PCR system 2400 (PerkinElmer Life Sciences Inc.) using Taq polymerase (Appligene Oncor, Illkirch, France) and oligonucleotides from pG+h9::ISS1 sequences (5′-ACTACTGACAGCTTCCAAGGA-3′ and 5′-ATAGTTCATTGATATATCCTC-3′ forEcoRI digestion and 5′-GTAAAACGACGGCCAGTG-3′ and 5′-TATCTACTGAGATTAAGGTCT-3′ for HindIII digestion). The Dye Terminator kit and a 310 Genetic Analyzer (Applied Biosystems, Foster City, CA) were used for DNA sequencing; each strand was sequenced twice on independent PCR products. DNA sequences were analyzed with Genetics Computer Group (GCG) sequence analysis software from the University of Wisconsin (22Devereux J. Haeberli P. Smithies O. Nucleic Acids Res. 1984; 12: 387-395Crossref PubMed Scopus (11531) Google Scholar) and Mail Fasta (National Center for Biotechnology Information). Internal amiA2 andamiA3 fragments were amplified using degenerated oligonucleotides (5′-TTGTWTACWTCWGAWGGHGAAGA-3′; 5′-ACTATCWRTYAACCAWGCTTG-3′) corresponding to conserved sequences of streptococcal oligopeptide-binding proteins (Refs. 3Alloing G. de Philip P. Claverys J.-P. J. Mol. Biol. 1994; 241: 44-58Crossref PubMed Scopus (112) Google Scholar and 4Jenkinson H.F. Baker R.A. Tannock G.W. J. Bacteriol. 1996; 178: 68-77Crossref PubMed Google Scholar; this work). Annealing was performed at 54 °C. Additional reverse PCRs were performed to amplify fragments flanking the known parts ofamiA1/amiA2/amiA3 genes. The total DNA of the St18 strain was completely digested by HindIII, EcoRI,TaqI, HaeII, HhaI, PstI, orNsiI, ligated in a dilute form (1 μg/ml), and amplified by PCR. Southern and Northern hybridizations were performed using a Positive nylon membrane for transfer (Appligene Oncor, Illkirch, France) according to the instructions in the ECL detection system (Amersham Biosciences, Inc., Buckinghamshire, United Kingdom). Insertional mutagenesis with pG+h9::ISS1 inS. thermophilus St18 had previously been adapted from the method described by Maguin et al. (23Maguin E. Prévost H. Ehrlich S.D. Gruss A. J. Bacteriol. 1996; 178: 931-935Crossref PubMed Scopus (412) Google Scholar, 24Garault P. Letort C. Juillard V. Monnet V. Appl. Environ. Microbiol. 2000; 66: 5128-5133Crossref PubMed Scopus (75) Google Scholar). Integrants affected for their growth in milk were selected on Fast Strain Differencing Agar medium (25Huggins A.M. Sandine W.E. J. Dairy Sci. 1984; 67: 1674-1679Abstract Full Text PDF Scopus (78) Google Scholar). The genes encoding oligopeptide-binding proteins amiA1,amiA2, amiA3, and the gene encoding cell wall proteinase prtS (26Fernandez-Espla M.-D. Garault P. Monnet V. Rul F. Eur. J. Biochem. 1999; 66: 4772-4778Google Scholar) were inactivated in the St18 strain using the pG+h9 gene replacement system. The mutants obtained are listed in Table I. Deletions were made in the middle of target genes amplified from the DNA St18 strain, as follows. PCR fragments of amiA genes were cloned into the pGEMt easy vector (Promega) according to the manufacturer's instructions. amiA2 deletion was obtained by double digestions with PshAI and BstSNI followed by a ligation step. The partially deleted amiA2 gene fragment was then cloned into pG+h9. Fragments of amiA1 andamiA3 genes were cloned in pG+h9, digested byAvaII for amiA1 and Bsp120I andBstXI for amiA3, and ligated to obtainamiA1 and amiA3 deletions. For theprtS mutant, a PCR fragment gene was first cloned into the TopoXL vector (Invitrogene), according to the manufacturer's instructions. The prtS gene fragment was cloned in pG+h9, and a deletion was obtained by HpaI andNruI digestion followed by ligation. The procedures forS. thermophilus electroporation, pG+h9 integration, and excision were similar to those used for insertional mutagenesis, as described previously (24Garault P. Letort C. Juillard V. Monnet V. Appl. Environ. Microbiol. 2000; 66: 5128-5133Crossref PubMed Scopus (75) Google Scholar, 27Biswas I. Gruss A. Ehrlich S.D. Maguin E. J. Bacteriol. 1993; 175: 3628-3635Crossref PubMed Scopus (354) Google Scholar). The ability of a toxic peptide analog (aminopterin) to inhibit bacterial growth on M17Lac plates was quantified by determining the extent of the inhibitory zone surrounding a filter paper disc saturated with 30 μg of aminopterin (Sigma). Cells were grown in CDM with αs2-casein trypsic hydrolysate as the nitrogen source containing more than 30 different peptides (28Juillard V. Guillot A. Le Bars D. Gripon J.-C. Appl. Environ. Microbiol. 1998; 64: 1230-1236Crossref PubMed Google Scholar). Cells were cultured for 13 h and then centrifuged (5 min, 5000 × g). Culture supernatants were concentrated and desalted with ZipTip (Millipore). Mass spectra were recorded in the positive-ion reflectron mode on a Voyager DE-STR mass spectrometer (Perspective Biosystems, Framingham, MA). All experiments were performed using a 20-kV acceleration voltage, a 337-nm laser, and 100-ns delayed extraction. The matrix solution was prepared freshly by dissolving 10 mg of α-cyano-4-hydroxycinnamic acid (Sigma) in 70/30 acetonitrile/trifluoroacetic acid, 0.3%. 0.5 μl of the sample was mixed with 0.5 μl of the matrix solution, spotted on a stainless steel sample plate, and air-dried. The insertional mutagenesis in S. thermophilus St18 produced 1.183 × 104 EmR integrants. Based on their phenotype on Fast Strain Differencing Agar, we selected 75 of them. After Southern analysis of the digested chromosomal DNAs of integrants growing slowly in milk, we selected 14 clones in which pG+h9::ISS1 was integrated at only one locus, distinct in each one of them. In 12 clones, pG+h9::ISS1 was tandemly integrated, exhibiting two hybridization bands using pG+h9 as a probe, whereas the 2 remaining clones contained only one copy of pG+h9::ISS1. The growth rate of one of the mutants, called the insertion sequence (IS) mutant throughout this paper, was significantly lower in milk (0.19 h−1) than that of the wild type strain (0.79 h−1). Rapid growth was restored by the addition of bactotryptone (growth rate of 0.75 h−1 for the mutant and of 0.85 h−1 for the wild type strain), suggesting that the affected function was related to nitrogen nutrition. The sequence of the interrupted gene of the IS mutant was determined using oligonucleotides from pG+h9::ISS1. We obtained a 392-bp sequence for the IS mutant, which formed part of an ORF exhibiting homologies with fragments of genes encoding oligopeptide-binding proteins (OBPs) from Streptococci, Bacilli, andL. monocytogenes (Refs. 3Alloing G. de Philip P. Claverys J.-P. J. Mol. Biol. 1994; 241: 44-58Crossref PubMed Scopus (112) Google Scholar, 4Jenkinson H.F. Baker R.A. Tannock G.W. J. Bacteriol. 1996; 178: 68-77Crossref PubMed Google Scholar, 29Podbielski A. Pohl B. Woischnik M. Körner C. Schmidt K.H. Rozdzinski E. Leonard B.A.B. Mol. Microbiol. 1996; 21: 1087-1099Crossref PubMed Scopus (87) Google Scholar, and 30Borezee E. Pellegrini E. Berche P. Infect. Immun. 2000; 68: 7069-7077Crossref PubMed Scopus (161) Google Scholar; accession no.AF305387). By applying additional reverse PCRs, we obtained a single 7032-bp DNA fragment containing the entire ORF corresponding to the interrupted gene of the IS mutant, together with four additional ORFs displaying a high level of homology with amiC,amiD, amiE, and amiF from different streptococci. These five ORFs, called amiA, amiC,amiD, amiE, and amiF, constituted the five proteins of ATP-binding cassette transporters (31Higgins C.F. Annu. Rev. Cell Biol. 1992; 8: 67-113Crossref PubMed Scopus (3386) Google Scholar); by homology, we named the S. thermophilus genes amiA1,amiC, amiD, amiE, and amiF. Protein sequences deduced from the entire DNA sequence exhibited the greatest homology with similar proteins from S. pneumoniae(ranging from 62% identity for AmiA1 to 86% identity for AmiE),S. gordonii (56% identity for AmiA1), and S. pyogenes (48% identity for AmiA1). Analysis of the sequence revealed the presence of a putative −10 extended promoter sequence situated 35 bp upstream of the ATG start codon of amiA1, and of a putative terminator situated 8 bp downstream of the stop codon ofamiF. The Southern hybridization under nonstringent conditions of HindIII- and EcoRI-digested St18 strain DNA, using a 1400-bp fragment of amiA1 as a probe, revealed two and three bands, respectively, suggesting the presence of at least two homologous genes (Fig. 1). Using PCR with degenerated oligonucleotides deduced from conserved regions of OBPs from streptococci and DNA from the IS mutant to avoid the amplification of an amiA1 gene fragment, we obtained two 1400-bp PCR products corresponding to two fragments of genes, namedamiA2 and amiA3, homologous to each other and toamiA1. With additional PCRs, we obtained 2938- and 3089-bp sequences containing entire amiA2 and amiA3genes, respectively. Comparisons of AmiA1, AmiA2, and AmiA3 protein-deduced sequences revealed a very strong identity between the three proteins (97.6% identity between AmiA1 and AmiA2, 87.1% identity between AmiA1 and AmiA3). amiA1, amiA2, and amiA3 encode proteins with 655, 655, and 657 residues, respectively. Their primary sequences contain a putative membrane lipoprotein lipid attachment site (VLAACS) (32Sutcliffe I.C. Russel R.R.B. J. Bacteriol. 1995; 177: 1123-1128Crossref PubMed Scopus (332) Google Scholar), an extracellular peptide and nickel-binding protein family signature sequence (A7D2TYYIRKGIKW) (1Tam R. Saier M.H. Microbiol. Rev. 1993; 57: 320-346Crossref PubMed Google Scholar). These features indicate the probable covalent attachment of AmiA proteins to the bacterial membrane. Analysis of the DNA sequences revealed the presence of putative −10 extended promoter sequences upstream of the amiA2 andamiA3 start codons, and of putative terminators downstream of the stop codons of the same genes. No open reading frames homologous to other genes encoding oligopeptide transport components were located either 590 and 1130 bp upstream or 500 and 400 bp downstream of theamiA2 and amiA3 genes, respectively. TheamiA3 promoter region differed from that of the two otheramiA genes because of the presence of four potential −10 extended promoter sequences, including two inverted repeat sequences (Fig. 2). Upstream of the three amiAs, we found part of insertion sequences or transposable elements (Fig.3). A shuffled IS1193(GenBank™ accession no. STIS1193) was found upstream of theamiA1 and the amiA2 sequences. The environment upstream of amiA3 differed from that of the other twoamiA genes because of the presence of a L. lactis IS904 (33Rauch P.J. Beerthuyzen M.M. de Vos W.M. Nucleic Acids Res. 1990; 18: 4253-4254Crossref PubMed Scopus (31) Google Scholar). Downstream of the amiA2and amiA3 genes, we sequenced a part of S. thermophilus IS1193. PCR screening of 21 industrial and three CNRZ collection S. thermophilus strains, using the same degenerated oligonucleotides as those used to search for amiA2 and amiA3, demonstrated the presence of at least one copy of an amiAgene in all strains. Southern analysis of 12 S. thermophilusstrains, using EcoRI-digested DNA and amiA2 as a probe, highlighted the presence of several large hybridization bands (some larger than 6000 bp; data not shown). These results suggested that the presence of several amiA in S. thermophilus is a general characteristic of this species. The first prerequisite for oligopeptide-binding protein to be functional is expression of the corresponding genes. Northern blot analysis revealed the presence of a 7000-bp transcript hybridizing with a 1860-bpamiA2 fragment (Fig. 4). This demonstrated that the potential promoter and terminator sequences identified upstream of amiA1 and downstream ofamiF, respectively, were functional, and that theamiA1, -C, -D, -E, and -F genes were organized into an operon. Northern blot analysis revealed another 2000-bp transcript hybridizing with the 1860-bp amiA2 fragment, indicating that the potential promoter and terminator sequences identified upstream and downstream ofamiA2 and/or amiA3 genes are functional. This result was confirmed by Northern blot analysis after RNA preparation of the IS mutant, which revealed the same 2000-bp transcript hybridizing with the same 1860-bp amiA2 probe. As expected in this case, no 7000-bp transcript corresponding to an ami operon was visible for the RNA preparation of the IS mutant (data not shown). As a second stage, we constructed stable negative mutants for oligopeptide-binding proteins by gene replacement (23Maguin E. Prévost H. Ehrlich S.D. Gruss A. J. Bacteriol. 1996; 178: 931-935Crossref PubMed Scopus (412) Google Scholar) and measured their growth rate in milk. Mutations in the three AmiA-encoding genes were achieved to obtain an AmiA triple negative mutant. Growth of the AmiA1− and AmiA2− mutants was comparable with that of the wild type strain in milk, whereas that of the AmiA3− mutant was significantly lower (Fig.5). The AmiA1/A2/A3− triple mutant exhibited very slow, limited growth in milk, similar to that seen for the IS mutant. This observation confirmed that there are probably no other oligopeptide-binding proteins working with the same system. In addition, we concluded that the mutation in the IS mutant affected the whole operon of oligopeptide transport. This result explained why the AmiA1 mutant and IS mutant exhibited different phenotypes. This was confirmed by the absence of an amioperon transcript on the Northern blot performed with the RNA of the IS mutant. The significant difference in the growth rates of the AmiA3− and AmiA1/A2/A3− mutants indicated that, in addition to AmiA3, at least one other AmiA was functional. At this stage in our work, we concluded that at least two of the three AmiA were functional and functioned with the same permease. The St18 strain is endowed with a cell wall proteinase, PrtS, that degrades proteins and peptides in smaller peptides (26Fernandez-Espla M.-D. Garault P. Monnet V. Rul F. Eur. J. Biochem. 1999; 66: 4772-4778Google Scholar). We constructed each AmiA/PrtS double mutant as well as the AmiA1/A3/PrtS−triple mutant to study the functionality of AmiA proteins independently of extracellular peptide degradation by the cell wall proteinase, PrtS. For technical reasons, we were unable to obtain the AmiA1/A2/A3/PrtS− quadruple mutant. We used CDM containing a trypsic hydrolysate of αs2-casein as the nitrogen source to compare the effects of AmiA mutations on AmiA/PrtS− mutants. The growth of all AmiA/PrtS− mutants was slower than that of the single mutant, PrtS−. More specifically, growth of the AmiA1/PrtS− and AmiA2/PrtS− mutants was half and one third less rapid, respectively, than that of the PrtS− mutant, indicating that AmiA1 and AmiA2 are functional (data not shown). Based on growth experiments in milk and CDM, we concluded that the three AmiA oligopeptide-binding proteins were functional. The simplest way to measure peptide uptake is based on the ability of an auxotrophic strain to utilize peptides as an amino acid source when all the peptidases have an intracellular location, as is the case for S. thermophilus (34Rul F. Monnet V. J. Appl. Microbiol. 1997; 82: 695-704Crossref PubMed Scopus (48) Google Scholar). Internalized peptides are then rapidly hydrolyzed by a battery of highly active intracellular peptidases. The rate-limiting step to peptide utilization in Ami mutants is their transport into the cytoplasm because the St18 strain has the same pool of peptidases as AmiA mutants of the ST18 strain. We studied the specificities of AmiA proteins in two stages. First, they were compared by analyzing the external medium of each AmiA/PrtS mutant in CDM, in which nitrogen was supplied by a mixture of peptides. We grew PrtS and AmiA mutants in CDM with a trypsic hydrolysate of αs2-casein as the nitrogen source. After growth, the culture supernatants were analyzed by mass spectrometry. The presence or absence of a peptide in the supernatant indicated complete or incomplete utilization of a peptide by a mutant. Analysis of the culture supernatants revealed differences in peptide composition. Several peptides were totally consumed by the PrtS− mutant but not by AmiA/PrtS− mutants. Their identification provided an indication of the specificities of OBPs. Most differences were found with AmiA3− mutants where some peptides were still present in the medium after growth, although they had completely disappeared from the culture medium of other AmiA and PrtS mutants. Among the most demonstrative examples, presented in TableII, the 92–114 αs2-casein fragment (FPQYLQYLYQGPIVLNPWDQVKR) was still present in the culture supernatants of AmiA3/PrtS− and AmiA1/A3/PrtS− mutants, but not in that of the PrtS− strain or AmiA1/PrtS− and AmiA2/PrtS− mutants. From these MS analyses, we therefore concluded that large peptides were used by the St18 strain and that the AmiA3 protein was capable of binding the peptides of at least 23 residues.Table IIPresence or absence of peptides after the growth of AmiA and PrtS negative mutants in CDM with peptides from αs2