INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The intracellular hydrolysis of peptides in Salmonella enterica serovar Typhimurium is carried out by at least 12 different enzymes. Some of these enzymes hydrolyze a wide range of peptides, while others are very specific for particular amino acids. Of the well-characterized peptidases in serovar Typhimurium, only three hydrolyze aspartyl (Asp)-X (where X is any amino acid) dipeptides: peptidase B, a broad-specificity aminopeptidase, hydrolyzes both Asp and non-Asp peptides (14); peptidase Q (15), an X-Pro-specific dipeptidase, hydrolyzes Asp-Pro (R. A. Larsen, unpublished results) but not other N-terminal Asp peptides; and peptidase E, an Asp-specific dipeptidase, hydrolyzes almost all Asp-X dipeptides (4, 5).

A strain of serovar Typhimurium (TN1300) lacking four broad-specificity peptidases, peptidases N, A, B, and D, as well as peptidase E and peptidase Q is still able to use Asp-Leu as a leucine source (4). In fact, such a mutant could grow on all of the Asp-X peptides that were tested except Asp-Pro. These results indicate that additional Asp peptide hydrolases are present in serovar Typhimurium. Previous work showed that crude cell extracts of serovar Typhimurium peptidase mutants contain at least two additional Asp-X-hydrolyzing activities (4). Mutants lacking these activities have not been isolated, however, and it is not clear if either or both of them are responsible for the ability of pepB pepE mutants to grow on Asp dipeptides.

This paper reports the characterization of two additional serovar Typhimurium peptidases that are capable of hydrolyzing N-terminal Asp peptides. One of these enzymes is the product of the serovar Typhimurium homolog of the Escherichia coli iadA gene (7). IadA is a dipeptidase that hydrolyzes N-terminal isoAsp dipeptides. Our data suggest that it also has low but detectable activity toward some but not all Asp-X peptides. The other enzyme is an isoAsp-hydrolyzing peptidase encoded by the serovar Typhimurium homolog of the E. coli ybiK gene which can also hydrolyze N-terminal Asp dipeptides at a low but detectable rate. Mutants lacking PepE, PepB, IadA, and YbiK have been constructed and shown to contain an additional Asp-X-specific peptidase.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, and growth conditions. The Salmonella enterica serovar Typhimurium strains used in this work are derivatives of strain LT2 and are listed in Table 1. E. coli DH5 was routinely used for DNA cloning experiments. E. coli strain BL21(DE3) was used for overexpression of plasmid-encoded IaaA. Strains were typically grown at 37°C in LB medium (Lennox L broth; Gibco BRL). E medium (24) containing 0.4% glucose was used as a minimal medium and was supplemented as indicated below with amino acids or peptides at 0.3 mM. When required, either ampicillin or chloramphenicol was added at 50 or 20 µg/ml, respectively.

DNA techniques. DNA manipulations were performed using standard techniques (13). Plasmids were purified using a QiaPrep Spin plasmid kit, and DNA fragments were prepared for cloning by agarose gel extraction using a Qiaquick gel extraction kit (Qiagen). Restriction enzymes were from Gibco BRL or New England Biolabs. PCR was performed using either Taq polymerase (Qiagen) or Pfu polymerase (Stratagene) and deoxynucleoside triphosphates from Qiagen. T4 DNA ligase was from Gibco BRL, and shrimp alkaline phosphatase was from Amersham. Oligonucleotide synthesis and DNA sequencing (on a Perkin Elmer ABI 377A automated DNA sequencer) were carried out by the W. M. Keck Center for Comparative and Functional Genomics at the University of Illinois. Sequencing of pCM499 was carried out using primers BamHI CW and BamHI CCW (New England Biolabs), which anneal near the BamHI sites of pBR322-derived plasmids.

Cloning of peptidase genes. Plasmids containing 8- to 12-kb fragments of serovar Typhimurium DNA from strain TN1246 cloned into pBR328 were introduced by transduction into TN5426 (leuBCD485 pepN90 pepA16 pepB11 pepD3 pepP1 pepQ1 pepT1 iadA1::Kn^r pepE8::MudJ zja::Tn10), selecting for growth on LB ampicillin medium. Transductants were replica plated on to minimal medium containing either leucine or isoAsp-Leu as the leucine source in order to identify colonies able to grow on isoAsp-Leu. Of approximately 1,000 colonies screened, 15 appeared to grow on isoAsp-Leu. Crude extracts of seven of these strains were tested for isoAsp-Leu-hydrolyzing activity after nondenaturing polyacrylamide gel electrophoresis (PAGE). Two of these strains overexpressed IadA, and three overexpressed IaaA. No overexpressed activity was observed for the remaining two clones and they were not further characterized. Plasmids (pCM499, pCM500, and pCM501) isolated from each of the strains overexpressing IaaA were used to transform TN5426, and the ability of the transformants to use isoAsp-Leu was confirmed. Each of these three clones was mapped using restriction endonucleases, but only pCM499 was used for sequencing.

Construction of an iadA overexpression plasmid. Serovar Typhimurium iadA was amplified by PCR using primers designed for the E. coli iadA sequence, iadA-1 (5'-ATGATTGATTATACCGCAGCCGGTTTTAC-3') and iadA-2 (5'-TTAAGCCGTTTCAAACGTTCCTTTCACGCAGGCTT-3'). The 1.1-kb product was cloned into pSE380 under the control of an inducible promoter (Ptrc), resulting in pCM429. To overexpress IadA, strain TN5347 (TN5131/pCM429) was grown in E minimal medium containing 0.4% glucose, 1% Casamino Acids, and ampicillin to mid-exponential phase (optical density at 600 nm = 0.8); 1 mM IPTG (isopropyl--D-thiogalactopyranoside) was added, and growth was continued overnight.

Construction of an iaaA overexpression plasmid. The iaaA open reading frame was amplified using primers ybiK1-3 (5'-GGAACTCCATATGAATAAAGCAGTGATTG-3') and ybiK2-s (5'-ATCGGATCCTATCCAGTTCATCGCTGTG-3'), both derived from the S. enterica serovar Typhi sequence using serovar Typhimurium genomic DNA as the template. The 1-kb product was cloned into the NdeI and BamHI sites of vector pET11b (Novagen), creating pCM532, in which iaaA replaced the phage T7 gene 10 exactly at the translational start site. This construct was transformed into E. coli strain BL21(DE3) containing an inducible T7 RNA polymerase gene (TN5628).

Construction of an iadA disruption. The iadA gene was amplified by PCR as described above using Taq polymerase and cloned directly into pGEM-T (Promega), producing pCM411. A Kn^r cassette, prepared by BamHI excision from pUC4K (Pharmacia), was inserted into a unique BamHI site within the iadA gene. The resulting construct, pCM435, was then transformed into TN2540, from which it was transduced into TN2373 (polA), selecting for resistance to kanamycin. An Ap^s transductant, which had undergone a double-crossover event resulting in a single disrupted copy of iadA in the chromosome, was isolated (TN5378). Amplification of iadA from TN5378 using primers designed for the serovar Typhimurium sequence (which had become available during the course of these experiments), iadA-3 (5'-GCATCTGCGATGTTCTACTCGCGAAT-3') and iadA-4 (5'-TACACCTGCTCAATGCGTAAATCT-3'), resulted in a single 2.3-kb product, which was the expected size of iadA plus the kanamycin resistance gene. Use of the same primers to amplify DNA from strain TN2373 resulted in a 1-kb fragment, which is the predicted size of the wild-type iadA gene.

Construction of an iaaA disruption. A 2.3-kb HindIII/BamHI fragment from pCM499, including the iaaA gene, was cloned into the HindIII and Bam HI sites of pBluescript II KS(+) (Stratagene). A Cm^r gene, which was excised from pKRB9 (G. Philips) using BamHI and had the ends blunted using the Klenow fragment, was then cloned into a unique SmaI site of the iaaA gene. Using the resulting plasmid, pCM531, a disruption of the chromosomal iaaA gene was constructed (TN5538) as described above for iadA. Amplification of iaaA using primers ybiK1-3 and ybiK2-s (described above) led to a 1-kb product from the wild-type strain (TN1379) and a single 2-kb product from the disrupted mutant (TN5538), demonstrating that in the mutant, the disrupted copy had replaced the wild-type gene.

Gel electrophoresis. Sodium dodecyl sulfate-PAGE (SDS-PAGE) was performed as previously described (20). Molecular weight markers were Mark12 wide-range protein standards (Novex). Nondenaturing PAGE and staining for peptidase activity were carried out essentially as previously described (16). The stain solution contained 2 mg of amino acid oxidase, 4 mg of peroxidase, and 1 mg of o-dianisidine in 10 ml of 0.1 M Tris-Cl, pH 8; the solution was mixed with an equal volume of 3.5% agar and was poured directly onto the gel.

Peptidase assays. A semiquantitative L-amino acid oxidase-based assay was used for determining peptidase activity in column fractions and for preliminary characterization of peptidase activities. Assays were carried out in the wells of plastic depression plates as described previously (4). The stain solution was prepared as described above for the staining of PAGE gels.

Purification of IaaA. Strain TN5628 [BL21(DE3)/pCM532] was grown in LB ampicillin medium overnight at 37°C. In preliminary experiments 1 mM IPTG was added at the mid-exponential phase of growth (optical density at 600 nm = 0.8); however, this step was omitted from the final purification protocol. Cells were disrupted by sonication in 0.1 M Tris-Cl, pH 8.0, and the resulting lysate was centrifuged for 30 min at 30,000 × g. The soluble fraction was then subjected to anion-exchange chromatography using a Q Sepharose HiLoad 26/10 column (Pharmacia) and was eluted with a linear gradient of 0 to 1.0 M NaCl in 20 mM Tris-Cl, pH 8.0. isoAsp-Leu-hydrolyzing activity, as determined by the amino acid oxidase assay described above, was eluted at approximately 0.28 M NaCl. Fractions with this activity were pooled and concentrated using a Centriprep 10 concentrator (Amicon) and subjected to gel filtration on a HiPrep Sephacryl HR S-100 (320-ml) column (Pharmacia) in a solution containing 20 mM Tris-Cl, pH 8.0, and 150 mM NaCl. The IaaA activity, which was eluted at an estimated molecular mass of 60 kDa, was concentrated and desalted again using a Centriprep 10 concentrator. Finally, this activity was chromatographed on a DEAE Sepharose fast-flow 10/10 column (Pharmacia) using a linear gradient of 0 to 1.0 M NaCl. IaaA activity was eluted at a concentration of approximately 0.22 M NaCl. The purified enzyme was analyzed by SDS-PAGE and was visually estimated to be greater than 90% pure.

Amino acid sequencing and mass spectrometry. The N-terminal amino acid sequence of each subunit of IaaA was determined using a Perkin-Elmer Applied Biosystems Procise 494 HT protein sequencer. For each polypeptide, an unambiguous sequence of the first 10 residues was obtained. Mass spectrometric analysis of IaaA was performed on a PerSeptive Biosystems Voyager-DE matrix-assisted laser desorption ionization-time of flight spectrometer. Both procedures were carried out at the Protein Science Facility at the University of Illinois Biotechnology Center.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Asp-Leu-hydrolyzing activities in serovar Typhimurium peptidase mutants. When a crude extract of a multiply peptidase-deficient (pepNABDPQE) strain (TN5131) of serovar Typhimurium was subjected to nondenaturing PAGE and the gel was stained to detect Asp-Leu hydrolysis, two distinct bands of activity (R_fs, 0.25 and 0.8) were observed (data not shown). These two activities were not resolved by anion-exchange chromatography (Q Sepharose) alone, but they were separated by subsequent gel filtration chromatography (Superdex 200 HR 16/60). The apparent native molecular masses of each were 200 kDa (R_f, 0.25) and 60 kDa (R_f, 0.8). Partially purified fractions from gel filtration chromatography were assayed in the presence and absence of the metal chelator EDTA. The activity of the 200-kDa enzyme was completely inhibited by EDTA, suggesting that it is a metallopeptidase. The 60-kDa enzyme was not inhibited by EDTA and was presumed to belong to a mechanistic class different from that of the 200-kDa enzyme.

The 200-kDa enzyme with an R_f of 0.25 is IadA. The properties of the enzyme with an R_f of 0.25 led us to suspect that it might be the serovar Typhimurium homolog of E. coli IadA, an isoAsp dipeptidase encoded by iadA and described by Gary and Clarke (7). To test this hypothesis, a clone carrying the serovar Typhimurium iadA gene (pCM429) was constructed. When a crude extract of a strain (TN5347) induced for IadA expression was subjected to nondenaturing PAGE and the gel was stained for isoAsp-Leu-hydrolyzing activity, the intensity of the band corresponding to the R_f of 0.25 was markedly increased (Fig. 1A), suggesting that the observed activity was IadA. Further confirmation of this assignment was obtained by constructing a strain containing an iadA::Kn^r chromosomal disruption (TN5379) and subjecting an extract of this strain to nondenaturing PAGE. The isoAsp-Leu-hydrolyzing activity that migrated at an R_f of 0.25 was absent in the iadA mutant strain (Fig. 1B). Although IadA has been reported to be specific for isoAsp peptides (9), our amino acid oxidase assays suggested that it hydrolyzed Asp-Leu and Asp-His. By HPLC analysis (using the overexpression strain described below) we determined that the rate of Asp-Leu hydrolysis is approximately 1% of the rate of isoAsp-Leu hydrolysis.

View larger version (70K): [in this window] [in a new window]   FIG. 1.   Nondenaturing PAGE of strains either overexpressing or lacking IadA. Soluble cell extracts of each strain were loaded as follows: lane 1, TN5131 (leuBCD485 pepN90 pepA16 pepB11 pepD3 pepP1 pepQ1 pepE8::MudJ); lane 2, TN5347 (leuBCD485 pepN90 pepA16 pepB11 pepD3 pepP1 pepQ1 pepE8::MudJ/pCM429); lane 3, TN1379 (leuBCD485); and lane 4, TN5379 (leuBCD485 iad1::Knr). Hydrolysis of isoAsp-Leu was detected using the amino acid oxidase assay.

Overexpression of IadA allows growth on isoAsp-Leu. Because both IadA and the 60-kDa enzyme with an R_f of 0.8 hydrolyzed isoAsp-Leu, we expected that this peptide could be utilized by a leucine auxotroph as a source of leucine. Surprisingly, neither a wild-type strain, TN1379 (leuBCD485), nor a peptidase-deficient strain, TN5131 (leuBCD485 pepN90 pepA16 pepB11 pepD3 pepP1 pepQ1 pepE8::MudJ), grew on isoAsp-Leu as a leucine source (no visible growth 48 h after streaking on minimal medium with isoAsp-Leu). This suggests that either the rate of peptide uptake or the level of isoAsp-Leu-hydrolyzing activity was too low to support growth. When a strain carrying the IadA overexpression plasmid (TN5347) was streaked onto minimal medium containing isoAsp-Leu as the leucine source, however, small colonies were clearly visible after overnight incubation. This result suggests that peptide hydrolysis is rate limiting for utilization of isoAsp-Leu.

Identification of the gene encoding the 60-kDa isoAsp peptidase with an R_f of 0.8. In order to facilitate further characterization of the enzyme with an R_f of 0.8, it was necessary to clone the gene encoding this enzyme. Because overexpression of IadA allowed growth on isoAsp-Leu as a leucine source, it was reasonable to assume that overexpression of the enzyme with an R_f of 0.8 might also allow growth on this peptide. If so, then it should be possible to clone the gene specifying this activity by screening a genomic library for plasmids that allow growth of a leucine auxotroph on isoAsp-Leu. Libraries of pBR328 carrying chromosomal DNA from TN1246 were screened as described in Materials and Methods. One of the three isolates that overexpressed the enzyme with an R_f of 0.8 was characterized further. The specific activity of isoAsp-Leu hydrolysis of an extract of this strain (TN5549, containing pCM499) was found to be 3.8 nmol of Asp min¹ µg¹, which is 20-fold higher than that of TN5507, a strain containing pBR328 alone (0.182 nmol of Asp min¹ µg¹).

View larger version (8K): [in this window] [in a new window]   FIG. 2.   18-minute region of the E. coli genome. The portion present on plasmid pCM499, which carries iaaA, is indicated by the black bar.

Cloning of iaaA and expression of its gene product. To test the hypothesis that iaaA encodes the peptidase with an R_f of 0.8 observed by native gel analysis, a clone with iaaA under the control of a phage T7 promoter was constructed (pCM532). The levels of IaaA protein expression (determined by SDS-PAGE) in a strain carrying this plasmid (TN5628) and grown in the presence or absence of an inducer (IPTG) were determined (Fig. 3A). Suprisingly, two overexpressed proteins, migrating at approximately the positions predicted for the two subunits of processed IaaA, were observed in extracts of the uninduced cells but not in extracts of induced cells. Correspondingly, the uninduced extract contained 70-fold-greater isoAsp-Leu-hydrolyzing activity: 168 nmol of Asp released min¹ mg¹, compared to 2.5 nmol of Asp min¹ mg¹ for the induced extract. In an attempt to explain these observations, we monitored the time course of protein expression after IPTG induction using whole cells rather than soluble cell extracts (Fig. 3B). In cultures induced with IPTG, high levels of an approximately 30-kDa protein were visible at the first time point (1.5 h) after induction. This protein is presumed to be the unprocessed iaaA gene product, which has a calculated molecular mass of 32.6 kDa. The same protein was visible at all subsequent time points. Proteins of approximately 19 kDa and 10 kDa appeared to increase in intensity after induction but were no longer present in cultures that were incubated overnight. The 30-kDa protein was not observed in soluble cellular extracts of induced cultures (Fig. 3A), suggesting that unprocessed IaaA forms inclusion bodies that are removed by the centrifugation step during extract preparation but are solubilized when whole-cell extracts are boiled in SDS. We hypothesized that low-level expression of the T7 polymerase gene in the absence of an inducer (22) leads to a level of IaaA expression that allows normal maturation of the enzyme. It appears that the high level of IaaA production that follows the induction of the T7 polymerase leads to aggregation rather than proper dimerization and processing.

View larger version (77K): [in this window] [in a new window]   FIG. 3.   SDS-PAGE analysis of IaaA overexpression. SDS-PAGE of soluble protein extracts (A) or whole cells (B) from TN5628 [BL21(DE3)pCM532] was carried out using cultures grown with or without IPTG, as indicated. The time after IPTG addition is indicated in hours. The positions at which unprocessed (30-kDa) and processed (19-kDa and 10-kDa) IaaA migrate are indicated with arrows. Sizes (in kilodaltons) of molecular mass markers are shown to the left of the gel. ON, overnight.

Purification and characterization of IaaA. Based on the results described above, we concluded that overnight growth of TN5628 in the absence of IPTG provides significant overexpression of active IaaA. Extracts of TN5628 grown under these conditions were used to purify IaaA as described in Materials and Methods. SDS-PAGE analysis of samples from each step of the purification is shown in Fig. 4.

View larger version (41K): [in this window] [in a new window]   FIG. 4.   Purification of IaaA. IaaA was overexpressed in strain TN5628 and purified as described in Materials and Methods. Samples from each step in the purification were subjected to SDS-PAGE and loaded as follows: lane 1, soluble cell extract; lane 2, pooled Q Sepharose fractions; lane 3, pooled Superdex fractions:, and lane 4, pooled DEAE fractions. Molecular mass markers were loaded in lane M, and the corresponding sizes are shown to the left of the gel. The positions of each of the two bands corresponding to subunits of IaaA are indicated and are labeled with the approximate molecular mass as calculated from the migration on this gel relative to the mass standards.

Substrate specificity of IaaA. Purified IaaA was used to determine the relative rates of hydrolysis of various peptide substrates. The substrates tested and the rate of hydrolysis relative to that of isoAsp-Leu are listed in Table 2. IaaA hydrolyzed asparagine poorly and did not hydrolyze N-acetylglucosaminylasparagine, indicating that its primary role is neither as an asparaginase nor as a glycosylasparaginase. Of the isoAsp substrates tested, isoAsp-Leu was hydrolyzed most rapidly and isoAsp-Phe-methyl ester (ME) was hydrolyzed least rapidly. The tripeptide isoAsp-Leu-Ala was a good substrate, indicating that IaaA is not strictly a dipeptidase. An internal isoAsp bond, Ala-isoAsp-Leu-Ala, however, was not hydrolyzed. Two chromogenic substrates, isoAsp-pNA and isoAsp-AMC, were both hydrolyzed (data not shown). The apparent K_m of IaaA for the best substrate, isoAsp-Leu, was determined to be 0.3 mM. These results suggest that IaaA is an isoAsp aminopeptidase.

Construction and properties of an iaaA disruption. The iaaA gene was disrupted as described in Materials and Methods. When soluble cell extracts of a strain (TN5538) carrying the iaaA::Cm^r mutation and a wild-type strain (TN2373) were compared by nondenaturing PAGE and staining for isoAsp-Leu-hydrolyzing activity, the band corresponding to IaaA was observed to be absent in the mutant (Fig. 5). This result confirms that the isoAsp peptidase activity is encoded by iaaA.

View larger version (60K): [in this window] [in a new window]   FIG. 5.   Nondenaturing PAGE of strains either overexpressing or lacking IaaA. Soluble cell extracts (10 µg each) were subjected to nondenaturing PAGE, after which the gel was stained for isoAsp-Leu hydrolysis using the amino acid oxidase assay. Strains tested were TN2373, wild type for peptidases (lane A); TN2540/pCM529, carrying an IaaA overexpression plasmid (lane B); TN2540/pCM531, carrying a plasmid with iaaA1::Cmr (lane C); and TN5379, containing a chromosomal iaaA disruption, iaaA::Cmr (lane D).

View larger version (18K): [in this window] [in a new window]   FIG. 6.   Effects of iadA and iaaA mutations on Asp-Leu (A) and isoAsp-Leu (B) hydrolysis in cell extracts. Soluble cell extracts (100 µg) of strains containing no mutations (TN1379 []), mutations in either iadA or iaaA (TN5379 [] and TN5538 [], respectively), or mutations in both iadA and iaaA (TN5617 []) were assayed by HPLC for the rate of Asp-Leu and isoAsp-Leu hydrolysis.

Identification of an additional Asp-X peptidase. Although TN5542 (pepB pepE iadA iaaA) lacks all known enzymes that hydrolyze Asp-Leu, this strain still utilizes this peptide as a source of Leu. In addition, this strain can utilize other Asp-X peptides as nitrogen sources. These results suggest strongly that another Asp-X peptidase is present in serovar Typhimurium. Experiments undertaken during the course of characterizing IaaA demonstrated the existence of such an enzyme. When a crude cell extract of TN5426 was chromatographed on Q Sepharose and the fractions were assayed for Asp-Leu hydrolysis under the standard assay conditions, only one peak of activity corresponding to IaaA was observed. In an effort to test the effects of various divalent cations on IaaA activity, we assayed the column fractions in the presence of the cations and found that when Mn²⁺ (1 mM) was included in the reaction mixture an additional peak of activity distinct from IaaA was observed. This activity was not observed in the presence of other divalent ions (Mg²⁺, Co²⁺, Ca²⁺, or Zn²⁺ as the chloride salt). To provide a preliminary characterization of the substrate specificity of the activity we tested its ability to hydrolyze other peptides. Peptides with residues other than Asp at the N terminus (Glu-Leu, Leu-Leu, Leu-Pro, Asn-Leu, and Lys-Leu) were not hydrolyzed. In addition, no activity toward Asp-pNA, Met-Asp-Phe, or isoAsp-Leu was observed. Based on these limited data the activity appears to be specific for N-terminal Asp peptides and may be another Asp-specific peptidase. Based on gel filtration chromatography, the native molecular mass was estimated to be 70 kDa. When partially purified fractions containing the activity were stained for Asp-Leu hydrolysis after nondenaturing PAGE, only a very faint band could be observed, suggesting that the enzyme is inactivated by this procedure and explaining why it was not observed previously.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The goal of this work was to identify all of the Asp peptide-hydrolyzing activities in S. enterica serovar Typhimurium. We have shown that there are at least three enzymes in addition to peptidases B and E that can hydrolyze N-terminal Asp peptides. Two of these enzymes are more active toward isoAsp (-Asp) peptides than toward peptides with the normal -Asp linkage. One of these enzymes is the serovar Typhimurium homolog of E. coli IadA (isoAsp dipeptidase) (7). This enzyme was originally described by Haley (9), who reported that it did not hydrolyze -Asp peptides. Because of these earlier observations, we considered the possibility that the apparent hydrolysis of Asp-Leu observed after nondenaturing gel electrophoresis might have resulted from contamination of the substrate with the isoAsp peptide. HPLC analysis of the lot of Asp-Leu used in these experiments showed that it contained about 3% isoAsp-Leu. It is possible, therefore, that some of the signal observed is generated by hydrolysis of the isoAsp peptide. The ability of IadA (and IaaA) to hydrolyze Asp-Leu was unambiguously established, however, by using HPLC to monitor the rate of substrate disappearance. Because Asp-Leu and isoAsp-Leu were easily separated by HPLC, we could monitor the rate of loss of each substrate independently. The data concerning the relative rates of hydrolysis of the two peptides by both IadA and IaaA were determined in this way, and these data establish clearly that both enzymes are able to hydrolyze Asp-Leu. The assay method used by Haley was much less sensitive than the HPLC assay used here, and it is not unreasonable to assume that the relatively slow hydrolysis of -Asp peptides by this enzyme might not have been observed under the conditions of these earlier experiments.

The gene encoding the second isoAsp peptidase, iaaA, has been identified for the first time in this work as a homolog of ybiK in E. coli. Although IaaA and IadA share the ability to hydrolyze an N-terminal isoAsp linkage, the two enzymes show no amino acid sequence similarity and clearly belong to different structural and mechanistic families. Based on similarity to certain plant asparaginases, the annotators of the E. coli genome suggested that the ybiK gene product is an asparaginase (1). Although the amino acid sequence similarity between the ybiK gene product and these plant enzymes is significant (44 to 45% identity), our results indicate that IaaA has only a relatively weak L-asparaginase activity. The ybiK gene product also has significant amino acid sequence similarity to glycosylasparaginases from both Flavobacterium and eukaryotes, but it has no activity towards the substrate most rapidly hydrolyzed by glycosylasparaginases, N-acetylglucosaminyl-L-asparagine (Asn-GlcNAc). The human glycosylasparaginase has been shown to hydrolyze isoAsp peptides in addition to Asn-GlcNAc (18), suggesting that the primary difference between this enzyme and IaaA is the more restrictive substrate specificity of IaaA.

Both the lupin asparaginases and the glycosylasparaginases are members of the Ntn hydrolase enzyme family (2). Ntn hydrolases are generated from enzymatically inactive precursors by a proteolytic processing step that generates a new N-terminal serine, threonine, or cysteine residue. This residue functions as the catalytic nucleophile in the hydrolytic reaction. Crystal structures have been determined for several members of this family: Thermoplasma 20S proteasome (12), glutamine phosphoribosylpyrophosphate amidotransferase (17), human glycosylasparaginase (19), Flavobacterium glycosylasparaginase (8), and penicillin acylase (6). These structures reveal a common fold characteristic of the family.

Posttranslational processing of many Ntn hydrolases is thought to occur by autoproteolysis of a homodimeric precursor to generate an ₂₂ heterotetramer as the active species. Our data show clearly that IaaA is a heterotetramer. The active form of IaaA has a native molecular mass of 60 kDa, twice the molecular mass predicted by the open reading frame (32 kDa). SDS-PAGE analysis showed that this species contains two polypeptides, and N-terminal amino acid sequencing indicates that they have been generated by proteolysis at a position equivalent to the processing site in similar enzymes (Gly178-Thr179). Although specific evidence for autoprocessing of IaaA has not been presented, it is likely that it shares this property with other Ntn hydrolases.

The physiological functions of IadA and IaaA are not clear. Both enzymes appear to preferentially hydrolyze isoAsp peptides, although they differ in that IadA is restricted to dipeptides whereas IaaA hydrolyzes tripeptides. It is generally assumed that isoAsp peptidases function in the hydrolysis of peptides generated by the intracellular degradation of isoAsp-containing proteins. Such proteins arise from the isomerization of Asp or Asn peptide bonds. These reactions are thought to occur most often in flexible regions of proteins, and because they introduce an additional CH₂ group into the peptide backbone, they frequently affect the secondary structure and the functioning of the protein. A protein damaged in this manner is either repaired by L-isoAsp methyltransferase (encoded by pcm in E. coli) or degraded. Gary and Clarke (7) have shown that E. coli iadA mutations have no discernible effect on growth or on stationary-phase survival. Although the phenotype of an iaaA iadA double mutant has not been thoroughly characterized, our observations suggest that this strain grows normally.

The location of iaaA as the first gene of an operon that also encodes a putative ATP-binding cassette transporter suggests that peptide catabolism might be an additional function for IadA and IaaA. We speculate that the transporter may function to transport isoAsp peptides, allowing them to be used as sources of amino acids. The third open reading frame in the proposed operon encodes a periplasmic binding protein that on the basis of its sequence belongs to the peptide and nickel binding transporters (family 5 according to Tam and Saier [23] and family 2 according to Linton and Higgins [11]). With the exception of a nickel transporter, all of the characterized members of this family have been shown to bind and import peptides. Although wild-type serovar Typhimurium does not grow on isoAsp-Leu as a leucine source in glucose minimal medium, mutants that do grow can easily be isolated and at least one class of these mutants overproduced IadA (data not shown). It seems possible, therefore, that one or the other of these genes might be regulated to allow isoAsp peptide use under as-yet-unknown environmental conditions.

Serovar Typhimurium contains a seemingly large number of Asp and isoAsp peptidases. In the present study we have shown that the role of IaaA, like that of IadA, is to hydrolyze isoAsp peptides. Peptidase B, peptidase E, and a previously unreported peptidase all hydrolyze -Asp peptides. The third, as-yet-unidentified peptidase has the unusual characteristic of requiring manganese for activity. It will be interesting to define the particular specificity of this enzyme and to determine whether or not it is the only remaining Asp peptidase in serovar Typhimurium.