Aspartic peptide hydrolases in Salmonella enterica serovar typhimurium

Two well-characterized enzymes in Salmonella enterica serovar Typhimurium and Escherichia coli are able to hydrolyze N-terminal aspartyl (Asp) dipeptides: peptidase B, a broad-specificity aminopeptidase, and peptidase E, an Asp-specific dipeptidase. A serovar Typhimurium strain lacking both of these enzymes, however, can still utilize most N-terminal Asp dipeptides as sources of amino acids, and extracts of such a strain contain additional enzymatic activities able to hydrolyze Asp dipeptides. Here we report two such activities from extracts of pepB pepE mutant strains of serovar Typhimurium identified by their ability to hydrolyze Asp-Leu. Although each of these activities hydrolyzes Asp-Leu at a measurable rate, the preferred substrates for both are N-terminal isoAsp peptides. One of the activities is a previously characterized isoAsp dipeptidase from E. coli, the product of the iadA gene. The other is the product of the serovar Typhimurium homolog of E. coli ybiK, a gene of previously unknown function. This gene product is a member of the N-terminal nucleophile structural family of amidohydrolases. Like most other members of this family, the mature enzyme is generated from a precursor protein by proteolytic cleavage and the active enzyme is a heterotetramer. Based on its ability to hydrolyze an N-terminal isoAsp tripeptide as well as isoAsp dipeptides, the enzyme appears to be an isoAsp aminopeptidase, and we propose that the gene encoding it be designated iaaA (isoAsp aminopeptidase). A strain lacking both IadA and IaaA in addition to peptidase B and peptidase E has been constructed. This strain utilizes Asp-Leu as a leucine source, and extracts of this strain contain at least one additional, as-yet-uncharacterized, peptidase able to cleave Asp dipeptides.     

Degradation of proline peptides in peptidase-deficient strains of Salmonella typhimurium.

A mutant strain of Salmonella typhimurium that lacks two proline-specific peptidases (peptidases P and Q) could not complete the degradation of proline peptides formed as intermediates in starvation-induced protein breakdown. The wild-type strain produced free proline as the product of degradation of proline-labeled proteins. The pepP pepQ mutant, however, produced a mixture of small proline peptides. In the absence of peptidase Q only, peptidase P could complete the degradation of most of the proline peptide intermediates formed. In the absence of peptidase P only, about 50% of the proline-labeled, acid-soluble products were proline peptides. These results are consistent with in vitro specificity data indicating that peptidase Q hydrolyzes X-Pro dipeptides only, whereas peptidase P attacks both X-Pro dipeptides and longer peptides with X-Pro at their N-termini. A mutant strain lacking four broad-specificity peptidases (peptidases N, A, B, and D), but containing peptidases P and Q, also produced proline peptides as products of protein breakdown. This observation suggests that broad-specificity peptidases are required to generate the X-Pro substrates of peptidases P and Q. A strain lacking six peptidases (N, A, B, D, P, and Q) was constructed and produced less free proline from protein breakdown than either the pepP pepQ strain or the pepN pepA pepB pepD strain. These observations suggest that the degradation of peptide intermediates involves the sequential removal of N-terminal amino acids and requires both broad-specificity aminopeptidases (peptidases N, A, and B) and the X-Pro-specific aminopeptidase, peptidase P.     

Peptidase-deficient mutants of Escherichia coli.

Mutant derivatives of Escherichia coli K-12 deficient in several peptidases have been obtained. Mutants lacking a naphthylamidase, peptidase N, were isolated by screening for colonies unable to hydrolyze L-alanine beta-naphthylamide. Other mutants were isolated using positive selections for resistance to valine peptides. Mutants lacking peptidase A, a broad-specificity aminopeptidase, were obtained by selection for resistance to L-valyl-L-leucine amide. Mutants lacking a dipeptidase, peptidase D, were isolated from a pepN pepA strain by selection for resistance to L-valyl-glycine. Starting with a pepN pepA pepD strain, selection for resistance to L-valyl-glycyl-glycine or several other valine peptides produced mutants deficient in another aminopeptidase, peptidase B. Mutants resistant to L-valyl-L-proline lack peptidase Q, an activity capable of rapid hydrolysis of X-proline dipeptides. Using these selection procedures, a strain (CM89) lacking five different peptidases has been isolated. Although still sensitive to valine, this strain is resistant to a variety of valine di- and tripeptides. The ability of this strain to use peptides as sources of amino acids is much more restricted than that of wild-type E. coli strains. Strains containing only one of the five peptidases missing in CM89 have been constructed by transduction. The peptide utilization profiles of these strains show that each of the five peptidases can function during growth in the catabolism of peptides.     

Dipeptidyl carboxypeptidase-deficient mutants of Salmonella typhimurium.

Mutants of Salmonella typhimurium deficient in dipeptidyl carboxypeptidase have been isolated by screening for clones unable to use N-acetyl-L-alanyl-L-alanyl-L-alanine (AcAla3) as the sole nitrogen source. An insertion of the transposable element Tn10 near dcp (the locus coding for dipeptidyl carboxypeptidase) has been isolated and used to map the locus in the interval between purB and trp, an otherwise genetically silent region of the S. typhimurium map. All dcp mutants could still grow using N-acetyl-L-alanyl-L-alanyl-L-alanyl-L-alanine (AcAla4) as the sole nitrogen source. Crude extracts from the dcp mutants failed to hydrolyze AcAla3 but retained approximately 80% of the wild-type activity toward AcAla4. Several lines of evidence indicate that hydrolysis of AcAla4 in the dcp mutant results from the action of a new peptidase distinct from dipeptidyl carboxypeptidase. A mutant strain lacking dipeptidyl carboxypeptidase in addition to peptidases N, A, B, and D showed reduced protein breakdown during carbon starvation compared with a strain lacking only peptidases N, A, B, and D.     

Peptidase mutants of Salmonella typhimurium

Six peptidase activities have been distinguished electrophoretically in cell extracts of Salmonella typhimurium with the aid of a histochemical stain. The activities can also be partially separated by chromatography on diethylaminoethyl-cellulose. These peptidases show overlapping substrate specificities. Mutants (pepN) of the parent strain leu-485 lacking one of these enzymes (peptidase N) were obtained by screening for colonies that do not hydrolyze the chromogenic substrate l-alanyl-beta-naphthylamide. The absence of this broad-specificity peptidase in leu-485 pepN(-) mutants allowed the selection of mutants unable to use l-leucyl-l-alaninamide as a leucine source. These mutants (leu-485 pepN(-)pepA(-)) lack a broad-specificity peptidase (peptidase A) similar to aminopeptidase I previously described in Escherichia coli. Mutants (pepD) lacking a dipeptidase (peptidase D) have been isolated from a leu-485 pepN(-)pepA(-) parent by penicillin selection for mutants unable to use l-leucyl-l-glycine as a leucine source. Mutants (pepB) lacking a fourth peptidase (peptidase B) have been isolated from a leu-485 pepN(-)pepA(-)pepD(-) strain by penicillin selection for failure to utilize l-leucyl-l-leucine as a source of leucine. Single recombinants were obtained by transduction for each of the peptidases missing in a leu-485 pepN(-)pepA(-)pepD(-)pepB(-) strain. The growth response of these recombinants to leucine peptides shows that all of these peptidases can function in the catabolism of peptides and that they display overlapping substrate specificities in vivo.     

Isolation and characterization of proline peptidase mutants of Salmonella typhimurium

The proline requirement of Salmonella typhimurium strain proB25 can be satisfied by either of the peptides Leu-Pro or Gly-Pro-Ala. A mutant derivative of strain proB25 isolated by penicillin selection in medium containing Leu-Pro as proline source fails to use either Leu-Pro or Gly-Pro-Ala as a source of proline. This strain is a double mutant that lacks two aminoacyl-proline-specific peptidases. One of these enzymes (peptidase Q) catalyzes the rapid hydrolysis of Leu-Pro but does not hydrolyze Gly-Pro-Ala or poly-l-proline. Mutations at a site (pepQ) near metE lead to loss of this activity. The other peptidase (peptidase P) catalyzes the hydrolysis of Gly-Pro-Ala and poly-l-proline but is only weakly active with Leu-Pro as substrate. This enzyme is similar to aminopeptidase P previously described in Escherichia coli (16). Mutations at a locus (pepP) near serA lead to loss of this enzyme.     

Multiple intracellular peptidases in Neurospora crassa.

Neurospora crassa possesses multiple intracellular peptidases which display overlapping substrate specificities. They were readily detected by an in situ staining procedure for peptidases separated in polyacrylamide gels, within which the auxilliary enzyme, l-amino acid oxidase, was immobilized. Eleven different intracellular peptidases were identified by electrophoretic separation and verified by their individual patterns of substrate specificities. Most peptide substrates tested were hydrolyzed by several different peptidases. The multiple intracellular peptidases may play overlapping roles in several basic cell processes which involve peptidase activity. The amount of peptidase activity for leucylglycine present in crude extracts of cells grown under widely different conditions was relatively constant, suggesting that this enzyme may be constitutive, although alterations in the amounts of individual peptidase isozymes may occur. A single enzyme, designated peptidase II, was partially purified and obtained free from the other peptidase species. Peptidase II was found to be an aminopeptidase with activity toward many peptides of varied composition and size. It was more active with tripeptides than homologous dipeptides and showed strong activity toward methionine-containing peptides. This enzyme, with a molecular weight of about 37,000, was thermolabile at 65 degrees C and was strongly inhibited by p-hydroxymercuribenzoate, Zn(2+), Co(2+), and Mn(2+), but was insensitive to the serine protease inhibitor phenylmethylsulfonyl fluoride. Peptidase II apparently possesses an essential sulfhydryl group and may be a metalloenzyme.     

Isolation and characterization Salmonella typhimurium mutants lacking a tripeptidase (peptidase T).

Salmonella typhimurium contains an enzyme, peptidase T, that hydrolyzes a variety of tripeptides. Specificity studies with a peptidase activity stain after gel electrophoresis of crude cell extracts showed that peptidase T hydrolyzes tripeptides containing N-terminal methionine, leucine, or phenylalanine. Little or no activity could be detected against dipeptides, N-blocked or C-blocked tripeptides, and tetrapeptides. Analysis of reaction products by high-pressure liquid chromatography showed that peptidase T removes the N-terminal amino acid from tripeptides. Mutants lacking peptidase T were isolated by screening microcultures grown in the wells of plastic microtitration plates for hydrolysis of Met-Ala-Ser or Met-Gly-Gly. Mutations (pepT) that eliminate this enzyme were found to be phage P22 cotransducible with purB at approximately 25 map units on the S. typhimurium map. Comparison of the growth properties of mutant and wild-type strains suggests that peptidase T does not function in utilization of tripeptides to provide amino acids during growth.     

Equine peptidases: Correspondence with human peptidases and polymorphism for erythrocyte peptidase a

Equine erythrocyte peptidases were compared to the six human erythrocyte peptidases, A, B, C, D, E, and F, regarding substrate specificity, relative activity, and electrophoretic mobility. Five equine erythrocyte peptidases appeared homologous to human peptidases A, B, D, E, and F. In contrast to human, equine peptidase C was absent in red cells, although it was weakly active in white cells. On the other hand, an equine peptidase, probably homologous to human peptidase S, was weakly active in red cells as well as present in white cells. Polymorphism for equine erythrocyte peptidase A is reported.     

Catalytic triad heterogeneity in S51 peptidase family: Structural basis for functional variability

Peptidase E (PepE) is a nonclassical serine peptidase with a Ser-His-Glu catalytic triad. It is specific for dipeptides with an N-terminal aspartate residue (Asp-X dipeptidase activity). Its homolog from Listeria monocytogenes (PepElm) has a Ser-His-Asn “catalytic triad.” Based on sequence alignment we predicted that the PepE homolog from Deinococcus radiodurans (PepEdr) would have a Ser-His-Asp “catalytic triad.” We confirmed this by solving the crystal structure of PepEdr to 2.7 Å resolution. We show that PepElm and PepEdr lack the Asp-X dipeptidase activity. Our analyses suggest that absence of P1 pocket in the active site could be the main reason for this lack of typical activity. Sequence and structural data reveal that the PepE homologs can be divided into long and short PepEs based on presence or absence of a C-terminal tail which adopts a β-hairpin conformation in the canonical PepE from Salmonella enterica. A long PepE from Bacillus subtilis with Ser-His-Asp catalytic triad exhibits Asp-X dipeptidase activity. Whereas the three long PepEs enzymatically characterized till date have been found to possess the Asp-X dipeptidase activity, the three enzymatically characterized short PepEs lack this activity irrespective of the nature of their catalytic triads. This study illuminates the structural and functional heterogeneity in the S51 family and also provides structural basis for the functional variability among PepE homologs.     

DapE can function as an aspartyl peptidase in the presence of Mn2+

Extracts of a multiply peptidase-deficient (pepNABDPQTE iadA iaaA) Salmonella enterica serovar Typhimurium strain contain an aspartyl dipeptidase activity that is dependent on Mn(2+). Purification of this activity followed by N-terminal sequencing of the protein suggested that the Mn(2+)-dependent peptidase is DapE (N-succinyl-L,L-diaminopimelate desuccinylase). A dapE chromosomal disruption was constructed and transduced into a multiply peptidase-deficient (MPD) strain. Crude extracts of this strain showed no aspartyl peptidase activity, and the strain failed to utilize Asp-Leu as a leucine source. The dapE gene was cloned into expression vectors in order to overproduce either the native protein (DapE) or a hexahistidine fusion protein (DapE-His(6)). Extracts of a strain carrying the plasmid overexpresssing native DapE in the MPD dapE background showed a 3,200-fold elevation of Mn(2+)-dependent aspartyl peptidase activity relative to the MPD dapE(+) strain. In addition, purified DapE-His(6) exhibited Mn(2+)-dependent peptidase activity toward aspartyl dipeptides. Growth of the MPD strain carrying a single genomic copy of dapE on Asp-Leu as a Leu source was slow but detectable. Overproduction of DapE in the MPD dapE strain allowed growth on Asp-Leu at a much faster rate. DapE was found to be specific for N-terminal aspartyl dipeptides: no N-terminal Glu, Met, or Leu peptides were hydrolyzed, nor were any peptides containing more than two amino acids. DapE is known to bind two divalent cations: one with high affinity and the other with lower affinity. Our data indicate that the form of DapE active as a peptidase contains Zn(2+) in the high-affinity site and Mn(2+) in the low-affinity site.     

Purification of Two Dipeptidyl Aminopeptidases II from Rat Brain and Their Action on Proline-Containing Neuropeptides

From the soluble and membrane fractions of rat brain homogenate, two enzymes that liberate dipeptides of the type Xaa-Pro from chromogenic substrates were purified to homogeneity. The two isolated dipeptidyl peptidases had similar molecular and catalytic properties: For the native proteins, molecular weights of 110,000 were estimated; for the denatured proteins, the estimate was 52,500. Whereas the soluble peptidase yielded one band of pI 4.2 after analytical isoelectric focusing, two additional enzymatic active bands were detected between pI 4.2 and 4.3 for the membrane-associated form. As judged from identical patterns after neuraminidase treatment, both peptidases contained no sialic acid. A pH optimum of 5.5 was estimated for the hydrolysis of Gly-Pro- and Arg-Pro-nitroanilide. Substrates with alanine instead of proline in the penultimate position were hydrolyzed at comparable rates. Acidic amino acids in the ultimate N-terminal position of the substrates reduced the activities of the peptidases 100-fold as compared with corresponding substrates with unblocked neutral or, especially, basic termini. The action of the dipeptidyl peptidase on several peptides with N-terminal Xaa-Pro sequences was investigated. Tripeptides were rapidly hydrolyzed, but the activities considerably decreased with increasing chain length of the peptides. Although the tetrapeptide substance P 1-4 was still a good substrate, the activities detected for the sequential liberation of Xaa-Pro dipeptides from substance P itself or casomorphin were considerably lower. Longer peptides were not cleaved. The peptidases hydrolyzed Pro-Pro bonds, e.g., in bradykinin 1-3 or 1-5 fragments, but bradykinin itself was resistant. The enzymes were inhibited by serine protease inhibitors, like diisopropyl fluorophosphate or phenylmethylsulfonyl fluoride, and by high salt concentrations but not by the aminopeptidase inhibitors bacitracin and bestatin. Based on the molecular and catalytic properties, both enzymes can be classified as species of dipeptidyl peptidase II (EC 3.4.14.2) rather than IV (EC 3.4.14.5). However, some catalytic properties differentiate the brain enzyme from forms of dipeptidyl peptidase II of other sources.     

The final N-terminal trimming of a subaminoterminal proline-containing HLA class I-restricted antigenic peptide in the cytosol is mediated by two peptidases

The proteasome produces MHC class I-restricted antigenic peptides carrying N-terminal extensions, which are trimmed by other peptidases in the cytosol or within the endoplasmic reticulum. In this study, we show that the N-terminal editing of an antigenic peptide with a predicted low TAP affinity can occur in the cytosol. Using proteomics, we identified two cytosolic peptidases, tripeptidyl peptidase II and puromycin-sensitive aminopeptidase, that trimmed the N-terminal extensions of the precursors produced by the proteasome, and led to a transient enrichment of the final antigenic peptide. These peptidases acted either sequentially or redundantly, depending on the extension remaining at the N terminus of the peptides released from the proteasome. Inhibition of these peptidases abolished the CTL-mediated recognition of Ag-expressing cells. Although we observed some proteolytic activity in fractions enriched in endoplasmic reticulum, it could not compensate for the loss of tripeptidyl peptidase II/puromycin-sensitive aminopeptidase activities.     

Degradation of Escherichia coli beta-galactosidase fragments in protease-deficient mutants of Salmonella typhimurium.

The degradation rates of several mutationally generated fragments of Escherichia coli beta-galactosidase were determined in wild-type strains of Salmonella typhimurium and in mutant Salmonella strains lacking several proteases and peptidases. Three termination fragments (produced by lacZ545, lacZ521, and lacZX90) and one internal reinitiation (restart) fragment [lacZpi(1)] are degraded in wild-type Salmonella strains at the same rates observed in wild-type Escherichia coli strains. Mutations that lead to loss of peptidases N, A, B, P, and Q or to loss of protease I or II do not affect the decay rates of any of these fragments. In addition, all of these peptidases and proteases are present in E coli mutants carrying deg mutations (deg mutations are known to stabilize beta-galactosidase fragments). Apparently, none of the proteases and peptidases that are currently accessible to direct genetic analysis plays a role in the early steps of the degradation of protein fragments.     

Interaction of purified brush-border membrane aminopeptidase N and dipeptidyl peptidase IV with lectin-sepharose derivatives

The glycoprotein nature of two peptidases purified from the rat intestinal brush-border membrane was examined by their interaction with several lectin-Sepharose derivatives. Aminopeptidase N (EC 3.4.11.2), which contains 20% carbohydrate by weight, was bound minimally (less than 30%) by columns of Con A-, RCAI- and WGA-Sepharose. Alternatively, a greater proportion of dipeptidyl peptidase IV (EC 3.4.14.-) was bound by these immobilized lectins with 50% of the enzyme binding to Con A-Sepharose. Treatment of both enzymes with neuraminidase enhanced the binding of aminopeptidase to RCAI-Sepharose by 4-fold but did not alter the binding patterns of dipeptidyl peptidase IV. A sequential fractionation of the two peptidases with columns of Con A- and RCAI-Sepharose gave four fractions of each enzyme with differing lectin-binding specificities. Approximately 60% of the dipeptidyl peptidase IV interacted with either one or both of the lectins while only 30% of the aminopeptidase N did so. Kinetic analysis of the four isolated fractions revealed some differences, possibly related to variations in the carbohydrate moiety. The findings confirm that these two purified rat intestinal brush-border membrane peptidases are glycoproteins and, while they share a common physiologic function and source, they apparently have very different and possibly unique asparagine-linked oligosaccharide side-chains. In addition, a considerable degree of microheterogeneity exists in the carbohydrate structure of these two enzymes.     

Novel prolyl tri/tetra-peptidyl aminopeptidase from Streptomyces mobaraensis: substrate specificity and enzyme gene cloning

The prolyl peptidase that removes the tetra-peptide of pro-transglutaminase was purified from Streptomyces mobaraensis mycelia. The substrate specificity of the enzyme using synthetic peptide substrates showed proline-specific activity with not only tripeptidyl peptidase activity, but also tetrapeptidyl peptidase activity. However, the enzyme had no other exo- and endo-activities. This substrate specificity is different from proline specific peptidases so far reported. The enzyme gene was cloned, based on the direct N-terminal amino acid sequence of the purified enzyme, and the entire nucleotide sequence of the coding region was determined. The deduced amino acid sequence revealed an N-terminal signal peptide sequence (33 amino acids) followed by the mature protein comprising 444 amino acid residues. This enzyme shows no remarkable homology with enzymes belonging to the prolyl oligopeptidase family, but has about 65% identity with three tripeptidyl peptidases from Streptomyces lividans, Streptomyces coelicolor, and Streptomyces avermitilis. Based on its substrate specificity, a new name, "prolyl tri/tetra-peptidyl aminopeptidase," is proposed for the enzyme.     

Genetics and regulation of peptidase N in Escherichia coli K-12

Escherichia coli K-12 strains contain a cytoplasmic activity, peptidase N, capable of hydrolyzing alanine-p-nitroanilide. Mutations in the structural gene for the enzyme, pepN, were mapped, and the properties of mutant strains were examined. The pepN locus lay between ompF and asnS at approximately 20.8 min on the E. coli chromosome. Loss of peptidase N activity through mutation had no apparent effect on the growth rate or nutritional needs of the cell. Enzyme levels in wild-type strains were constant throughout the growth cycle and were constitutive in all of the growth media tested. Starvation for carbon, nitrogen, or phosphate also did not alter enzyme levels. Constitutive expression of peptidase N is consistent with the idea that the enzyme plays a significant role in the degradation of intracellularly generated peptides.     

Peptidases A,B, C,D and S in the American mink: polymorphism and chromosome localization

Summary An electrophoretic analysis of peptidases was carried out in a population of American mink. Based on substrate and tissue specificities, as well as subunit composition, homologies were established between mink peptidases A, B, C, D and S and human peptidases. Polymorphism for peptidases B and D was demonstrated for minks of three coat colour types. Breeding data indicated that the peptidase variations are under the control of allele pairs at distinct autosomal loci designated as PEPB and PEPD, respectively. Using a panel of American mink-Chinese hamster hybrid clones, the gene for PEPB was assigned to mink chromosome 9.     

Specific inhibition of endopeptidase 24.16 by dipeptides.

The inhibitory effect of various dipeptides on the neurotensin-degrading metallopeptidase, endopeptidase 24.16, was examined. These dipeptides mimick the Pro10-Tyr11 bond of neurotensin that is hydrolyzed by endopeptidase 24.16. Among a series of Pro-Xaa dipeptides, the most potent inhibitory effect was elicited by Pro-Ile (Ki approximately 90 microM) with Pro-Ile greater than Pro-Met greater than Pro-Phe. All the Xaa-Tyr dipeptides were unable to inhibit endopeptidase 24.16. The effect of Pro-Ile on several purified peptidases was assessed by means of fluorigenic assays and HPLC analysis. A 5 mM concentration of Pro-Ile does not inhibit endopeptidase 24.11, endopeptidase 24.15, angiotensin-converting enzyme, proline endopeptidase, trypsin, leucine aminopeptidase, pyroglutamyl aminopeptidase I and carboxypeptidase B. The only enzyme that was affected by Pro-Ile was carboxypeptidase A, although it was with a 50-fold lower potency (Ki approximately 5 mM) than for endopeptidase 24.16. By means of fluorimetric substrates with a series of hydrolysing activities, we demonstrate that Pro-Ile can be used as a specific inhibitor of endopeptidase 24.16, even in a complex mixture of peptidase activities such as found in whole rat brain homogenate.     

Electrophoretic heterogeneity of mouse erythrocyte peptidases

Starch gel electrophoresis in conjunction with a specific staining method revealed the occurrence of five distinct peptidases in mouse red blood cells. These enzymes can be distinguished on the basis of substrate specificity and electrophoretic mobility. They have been designated peptidases A, B, C, D, and E to correspond with the nomenclature adopted for human peptidases with which the mouse enzymes appear to be homologous. Genetically determined variants of peptidase C are described. The phenotype Pep C1 occurs in C57BL/Gr mice and the phenotype Pep C2 in CBA/Gr and Strong A/Gr mice. These phenotypes and the presumed heterozygote, Pep C2-1, appear to be due to the occurrence of codominant autosomal alleles which have been designated Pep-C ¹ and Pep-C ². F₁ and F₂ crosses show segregation in the expected Mendelian ratios. F₂ embryos and their placentae show the same electrophoretic pattern for peptidase C. The occurrence of a separate locus controlling the structure of each distinct peptidase is postulated.