|Home | About | Journals | Submit | Contact Us | Français|
Aeromonas enteropelogenes (formerly A. tructi) was described to be an ampicillin-susceptible and cephalothin-resistant Aeromonas species, which suggests the production of a cephalosporinase. Strain ATCC 49803 was susceptible to amoxicillin, cefotaxime, and imipenem but resistant to cefazolin (MICs of 2, 0.032, 0.125, and >256 μg/ml, respectively) and produced an inducible β-lactamase. Cefotaxime-resistant mutants (MIC, 32 μg/ml) that showed constitutive β-lactamase production could be selected in vitro. The gene coding for the cephalosporinase of A. enteropelogenes ATCC 49803 was cloned, and its biochemical properties were investigated. Escherichia coli transformants showing resistance to various β-lactams carried a 3.5-kb plasmid insert whose sequence revealed a 1,146-bp open reading frame (ORF) encoding a class C β-lactamase, named TRU-1, showing the highest identity scores with A. punctata CAV-1 (75%), A. salmonicida AmpC (75%), and A. hydrophila CepH (71%). The blaTRU-1 locus includes open reading frames (ORFs) showing significant homology with genes found in the genomes of other Aeromonas species, although it exhibits a different organization, as reflected by the presence of additional ORFs located downstream of the β-lactamase gene in the A. hydrophila and A. salmonicida genomes. Specific PCR assays were negative for cphA-like and blaOXA-12-like genes in three A. enteropelogenes ATCC strains. Purified TRU-1 showed a broad substrate profile, efficiently hydrolyzing benzylpenicillin, cephalothin, cefoxitin, and, although with significantly lower turnover rates, oxyiminocephalosporins. Cephaloridine and cefepime were poorly recognized by the enzyme, as reflected by the high Km values observed with these substrates. Thus far, A. enteropelogenes represents the only known example of an Aeromonas species that produces only one β-lactamase belonging to molecular class C.
Resistance to β-lactam antibiotics in Gram-negative bacilli is mediated mainly by the production of β-lactamases, which are divided into four major molecular classes, classes A, B, C, and D (3, 11). Genes encoding AmpC (class C) β-lactamases are generally found on the chromosome of many Gram-negative, aerobic, rod-shaped bacteria and are usually inducible (38). About 20 years ago, plasmid-borne class C enzymes were identified in important clinical bacterial species that do not naturally produce significant amounts of these types of β-lactamase (such as Klebsiella pneumoniae or Escherichia coli) or do not possess an ampC gene (e.g., Salmonella spp.) (10, 34) and actually bear an increasing clinical importance. From a functional standpoint, class C β-lactamases are typically characterized by a very efficient hydrolysis of narrow-spectrum cephalosporins but are also able to degrade more recently found extended-spectrum cephalosporins and cephamycins and, thus, can confer resistance to these agents, especially in strains where the β-lactamase is derepressed or plasmid encoded (9, 11, 34).
Aeromonads are ubiquitous waterborne organisms, and some Aeromonas species are known to cause human gastroenteritis or serious opportunistic infections, such as septicemia and cellulitis (20, 27). Since these species are present in both the environment and the clinical setting, aeromonads have been identified as a reservoir of antibiotic resistance genes, including β-lactamase determinants such as the blaFOX-like and blaCMY-1-like genes (19, 24, 34), which have been successfully transferred to more important opportunistic human pathogens, in particular members of the Enterobacteriaceae.
Aeromonas spp. also show an important diversity in terms of β-lactamase content, and at least two different situations have been reported, consisting of the coordinated production of two or three different naturally occurring β-lactamases. Known Aeromonas β-lactamases include the narrow-spectrum class B metallocarbapenemase, a class C cephalosporinase, and a class D oxacillinase, which provide resistance to carbapenems, cephalosporins, and penicillins, respectively (2, 6, 18, 26).
Aeromonas enteropelogenes (formerly A. trota or A. tructi) recently appeared as a distinct species (12, 25, 33) that presents the same pathogenicity factors featured by A. hydrophila or A. veronii, such as the production of enterotoxins (1). This species was previously described as being susceptible to penicillins and carbapenems, while it exhibits resistance to narrow-spectrum cephalosporins (12, 33), strongly suggesting the production of a cephalosporinase. This study describes detailed genetic and biochemical properties of the endogenous chromosome-encoded class C β-lactamase from A. enteropelogenes type strain ATCC 49803.
Aeromonas enteropelogenes (formerly A. trota or A. tructi) (taxonomy identification number 29489 [http://www.ncbi.nlm.nih.gov/Taxonomy/]) ATCC 49803 (human stool, India), ATCC 49657, and ATCC 49660 and Aeromonas hydrophila ATCC 7966 were purchased from the strain collection of the Pasteur Institute (Paris, France). Escherichia coli TOP10 cells (Invitrogen, Leek, Netherlands) were used as the recipient for cloning experiments. E. coli BL21(DE3) cells (Stratagene Inc., La Jolla, CA) were used as the host for T7 promoter-based expression plasmids for overexpression experiments and large-scale enzyme production.
A. enteropelogenes strains were grown aerobically at 30°C in Mueller-Hinton (MH) broth (Bio-Rad, Marnes-la-Coquette, France). E. coli strains were grown aerobically at 37°C in Luria-Bertani medium (37) for routine culture or Mueller-Hinton plates for antimicrobial susceptibility testing (Bio-Rad, Marnes-la-Coquette, France). Susceptibilities to antimicrobial agents were measured by the microdilution broth method, as recommended by the CLSI (14), at 30°C or 37°C (for A. enteropelogenes and E. coli strains, respectively). Medium for the autoinduction of T7 promoter-based expression systems, as described previously by Studier (40), were used with BL21(DE3) strains for recombinant enzyme production.
The production of β-lactamase in A. enteropelogenes was investigated by measuring spectrophotometrically the initial rate of hydrolysis of a 100 μM cephalothin solution in 50 mM HEPES (pH 7.0). Crude extracts of Aeromonas strains were prepared from a culture grown in brain heart infusion (BHI) broth at 30°C in the presence and absence of 5 μg/ml cefoxitin, from which bacterial cells were collected by centrifugation and disrupted by sonication as previously described (19). Determinations of protein concentrations were performed by using a commercial kit (Bio-Rad, Marnes-la-Coquette, France).
A suspension of A. enteropelogenes containing ~109 CFU was plated onto MH plates containing various inhibitory concentrations of cefotaxime (final concentration, 1, 2, or 10 μg/ml). One cefotaxime-resistant mutant, named M1, was isolated, and the stability of its phenotype was confirmed by repeated subculturing. The production of β-lactamase in the presence and absence of an inducer was investigated as described above (19).
Genomic DNA from A. enteropelogenes ATCC 49803 was prepared as previously described (41), partially digested with HindIII, and ligated into the HindIII site of plasmid pZErO-2 (Invitrogen, Leek, Netherlands). The ligation mixture was transformed into E. coli TOP10 cells, and transformants were successively selected on kanamycin (50 μg/ml) and cefazolin (50 μg/ml). A 3.5-kb BglII restriction fragment was subcloned by using the BamHI site of the same vector, and the cloned DNA fragment was completely sequenced on both strands by using laboratory-designed primers and an Applied Biosystems sequencer (ABI 3130).
The detection of β-lactamase genes encoding class B metalloenzymes and class C or D serine enzymes in A. enteropelogenes strains was achieved by PCR using GoldTaq DNA polymerase (Perkin-Elmer, Waltham, MA) and oligonucleotides that were designed on the basis of available Aeromonas cphA-like (primers MEI-F1 and -R2), ampC (primers AERCP-F1 and -R2), and blaOXA-12-like (primers OXAF and OXAR) genes (Table (Table1).1). Similarly, the region including the moxR-like and the β-lactamase genes was amplified by PCR using primers MOX-F and CEP-R, and the amplification products were sequenced as described above. A. hydrophila ATCC 7966 (39) was used as a positive control for amplification in the various PCR assays.
Sequence analysis, including multiple alignment and homology searches, was performed by using software available at the websites of the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/) and the European Bioinformatics Institute (http://www.ebi.ac.uk/).
The blaTRU-1 open reading frame (ORF) was amplified by PCR using primers TRU-EXP/f and TRU-EXP/r, designed to add NdeI and HindIII restriction sites at the 5′ and 3′ extremities of the amplified fragment, respectively (Table (Table1),1), which was subcloned into expression vector pET-24a, yielding pET-TRU-1. E. coli BL21(DE3) cells were transformed with the latter plasmid and grown aerobically in autoinducing rich medium ZYP-5052 for 48 h.
Cells were harvested by centrifugation (10,000 × g at 10°C for 20 min), and the culture supernatant was concentrated by ultrafiltration using an Amicon 2000A device (Millipore Corp., Billerica, MA) and a YM10 membrane until the sample volume reached 1/40 of the initial volume. The concentrated sample was desalted by using a 26/10 HiPrep desalting column (GE Healthcare, Uppsala, Sweden) using 10 mM HEPES (pH 7.6) buffer for protein elution (buffer H) and loaded onto an XK 16/20 column packed with 25 ml of CM-Sepharose Fast Flow gel (GE Healthcare) previously equilibrated with buffer H. Proteins were eluted by using a linear NaCl gradient (0 to 0.3 M in 400 ml) in buffer H, and fractions containing β-lactamase activity were pooled and concentrated by using an Amicon TCF2A system (Millipore Corp.). The sample was then desalted, as described above, against 20 mM ethanolamine buffer (pH 10) (buffer E) and loaded onto a Mono Q 5/50 GL column (GE Healthcare) previously equilibrated with buffer E. Proteins were eluted by using a linear NaCl gradient (0 to 0.4 M in 40 ml) in buffer E.
The purified TRU-1 β-lactamase preparation (0.9 mg/ml) was stored at −20°C until use. To monitor the presence of β-lactamase during production and purification procedures, the β-lactamase activity was measured as described above. SDS-PAGE analysis, isoelectric focusing (IEF) analysis, and electron spray-ionization mass spectrometry (ESI-MS) were performed as described previously (15, 29).
The steady-state kinetic parameters for the hydrolysis of β-lactam substrates were determined at 30°C, as previously described, with 50 mM HEPES buffer (pH 7.0) (23). Competition experiments with poor substrates were carried out by using 150 μM cephalexin as the reporter substrate. The enzyme concentration in the assay mixtures ranged between 0.2 and 450 nM.
The nucleotide sequence reported in this paper has been submitted to the EMBL/GenBank nucleotide sequence database (accession no. EU046614).
A. enteropelogenes (formerly A. trota or A. tructi) ATCC 49803 exhibits susceptibility to most β-lactam agents (or β-lactam/β-lactamase inactivator combinations), except for cephalothin and, to a minor extent, cefoxitin (Table (Table2).2). These data are in agreement with the previously reported phenotypic features of other isolates of this species (12, 33). In contrast with the other Aeromonas species known for multiple-β-lactamase production and that are usually resistant to several β-lactam agents, including penicillins and carbapenems (e.g., A. hydrophila and A. sobria), the susceptibility profile of A. enteropelogenes is very peculiar and strongly suggests the production of a single β-lactamase with cephalosporinase activity.
An enzymatic assay carried out using a crude extract of A. enteropelogenes ATCC 49803 confirmed the production of a β-lactamase that was produced in an inducible fashion. The basal amount of β-lactamase produced by the wild-type strain was extremely low and could barely be measured (specific activity, ≤0.04 nmol/min·mg of protein). In contrast, the addition of cefoxitin to the culture medium yielded an increased amount of enzyme (induction, >240-fold; specific activity in induced cultures, 9.7 nmol/min·mg of protein), in agreement with the fold induction of cephalosporinase activity previously observed for most Aeromonas species (43). IEF analysis carried out using these crude extracts revealed the presence of a single nitrocefin- and cefazolin-hydrolyzing band at pI 9.0, i.e., notably different from the previously reported value for the class C β-lactamases of Aeromonas punctata (CAV-1) (pI 6.8), A. hydrophila (CepH) (pI 6.6), A. jandaei (AsbA1) (pI 7.0), and A. veronii (CepS) (pI 7.0) (5, 19, 35, 42).
Under the selective pressure created by the presence of inhibitory concentrations of cefotaxime (1, 2, or 10 μg/ml), cefotaxime-resistant mutants could be readily obtained at a frequency of approximately 10−7. Similar frequencies for β-lactamase-hyperproducing mutants of A. hydrophila and A. caviae were previously reported (43). Analysis of one of these mutants, called M1, exhibited an increase of MICs of most antibiotics, except imipenem and, to some extent, amoxicillin (Table (Table2).2). In strain M1, β-lactamase production was no longer dependent on the presence of an inducer, and a high level of β-lactamase activity was constitutively produced (specific activity, 19.3 nmol/min·mg of protein). The derepressed AmpC phenotype of strain M1 was demonstrated to be stable upon repeated subculturing. The determination of the nucleotide sequence of the moxR gene and the intergenic region separating the latter from the β-lactamase gene revealed no mutations compared to the wild-type strain, indicating that the derepression of β-lactamase expression is not due to a modification in the promoter sequence or in the upstream putative transcriptional regulator. Since AmpC production is controlled by a two-component regulatory system in Aeromonas spp., the derepressed phenotype of mutant M1 might likely be caused by remote mutations occurring in genes involved in the regulatory system (2, 4, 5).
A HindIII genomic library of A. enteropelogenes ATCC 48903 was obtained with E. coli TOP10 cells. Plasmid analysis of transformants obtained after selection in the presence of kanamycin and cefazolin showed a cloned insert of approximately 15 kb. A fragment of 3.5 kb, obtained after digestion with BglII, was subcloned into the BamHI site of vector pZErO-II and transformed into E. coli TOP10 cells to yield strain NI540, which showed a production of β-lactamase activity. Analysis of the 3.5-kb insert sequence revealed the presence of four ORFs, including a 1,146-bp ORF encoding a class C β-lactamase (383 residues) that was named TRU-1 (after A. tructi, the former name for A. enteropelogenes in use when the β-lactamase was sequenced) and shows significant homology with class C β-lactamases, in particular with the resident enzymes found in various Aeromonas species and the acquired enzymes likely originating from these organisms (e.g., FOX-type enzymes). The amino acid sequence of TRU-1 contained the conserved motifs typical of serine β-lactamases: S85VSK (motif I), Y171SN (motif II), and K335TG (motif III) (Fig. (Fig.1).1). TRU-1 shows the closest (Fig. (Fig.2)2) ancestry with the acquired FOX-2 and FOX-5 enzymes (76% identity) and with the resident class C β-lactamase of Aeromonas punctata, CAV-1 (75% identity) (19). A sequence comparison of TRU-1 with other resident class C enzymes of Aeromonas species revealed the presence of 35 original substitutions, some of which were located close to catalytically relevant conserved motifs (e.g., Phe90 close to motif I and Glu170 close to motif II) (Fig. (Fig.1).1). The regions that show the highest sequence diversity overall are (i) the N-terminal extremity (residues 1 to 23) and (ii) the residues located immediately before motif III (residues 319 to 330), which corresponds, from a structural standpoint, to the loop that connects the α10 helix and the β11 strand, with the latter bearing the K335TG residues (Fig. (Fig.1).1). The observed sequence heterogeneity in the latter region might be functionally relevant, as amino acid substitutions or deletions in the same region of several other chromosomal or plasmid-borne cephalosporinases are responsible for the extension of their substrate profile toward cefepime and cefpirome (7, 8, 9, 31). This has been particularly well described for CMY-10 and other CMY-1-like variants, which carry a 3-amino-acid (aa) deletion near the so-called α10 helix (α15 in Fig. Fig.1)1) (17, 28).
The presence of a similar class C β-lactamase-encoding gene was confirmed for A. enteropelogenes strains ATCC 49657 and ATCC 49660 by using PCR assays carried out with primers suitable for the amplification of an internal fragment of the blaTRU-1 gene (~850 bp). Sequencing of the amplification product revealed a maximum divergence of 4%, at the level of the nucleotide sequence. However, PCR assays carried out with primers suitable for the amplification of cphA-like and blaOXA-12-like genes yielded, at variance with A. hydrophila ATCC 7966, which was used as a positive control, negative results with all three A. enteropelogenes strains. These data suggest that the genes encoding the resident class B and class D β-lactamase homologues are likely absent in the genome of this species.
The calculated G+C% content of the blaTRU-1 gene and the whole 3.5-kb cloned insert were in agreement with those observed for the available genome sequence of Aeromonas (60.5% and 58.5 to 61.5%, respectively) (A. hydrophila ATCC 7966 [GenBank accession no. CP000462] and A. salmonicida A449 [GenBank accession no. AY301064], respectively).
Four ORFs were found in the 3.5-kb cloned genomic fragment of A. enteropelogenes, including a moxR-like transcriptional regulator gene, the ampC gene, and two additional genes, encoding a putative endonuclease (Holliday junction resolvase, yqgF-like) and another putative transcriptional regulator (Fig. (Fig.3).3). The two latter genes are also found in the published genome sequences of A. hydrophila ATCC 7966 and A. salmonicida subsp. salmonicida strain A449 (protein identities ranging from 87 to 92%), although the overall organization of this locus seems to be different in the three genomes. Indeed, the region between the ampC and the yqgF-like genes (separated by 109 bp in A. enteropelogenes) is much longer in the published Aeromonas genomes (2,614 and 5,286 bp for A. hydrophila and A. salmonicida, respectively) and includes two or three additional ORFs (Fig. (Fig.3),3), only one of which is similar in both organisms (AHA_3133 and ASA_1193). These additional ORFs encoded putative membrane transport proteins or transcriptional regulators (36, 39).
Interestingly, the two genes located downstream of the blaTRU-1 gene and present in the two published genomes were also found downstream of the blaCMY-8 gene, located on the large plasmid pK29 from Klebsiella pneumoniae (13). This finding further supports the origin of some blaCMY genes in Aeromonas-related species, after some chromosomal fragments were mobilized on plasmids that eventually disseminate in the Enterobacteriaceae.
In contrast with most Gram-negative bacilli, in which the transcription of the ampC β-lactamase gene is under the control of a LysR-type transcriptional regulator called AmpR (30), the coordinated expression of the three β-lactamase genes in Aeromonas hydrophila and A. jandaei is apparently regulated by a two-component phosphorylation-dependent regulatory system (2, 4, 5). In these bacteria, the genes encoding the transcriptional regulator and the protein sensor are located on the operon that also encodes a class D β-lactamase (this oxacillinase is absent in A. enteropelogenes) (32), while the ampC-like and cphA-like genes are found at a rather distant location (~1 Mb) in the chromosomes of both A. hydrophila and A. salmonicida (36, 39). It was suggested previously that this transcriptional control involves at least one direct repeat of a promoter-proximal DNA sequence motif, TTCAC, or the “blr tag” (5). Interestingly, the blaTRU-1 promoter region (−50 to −90 bp with respect to the start codon) appeared to be similar to those found upstream of the ampC-like gene in A. hydrophila, A. salmonicida, and A. jandaei, which all present three TTCAC repeats (Fig. (Fig.3).3). This finding implies that the genetic linkage of the two-component regulatory system and the class D β-lactamase is not a prerogative for the presence of a β-lactamase expression regulatory system.
E. coli strain NI540, producing the TRU-1 β-lactamase (as confirmed by enzyme assays and IEF analysis showing a single band at pI 9.0 [data not shown]), presented a susceptibility profile that was similar overall to that of A. enteropelogenes, with decreased susceptibilities for most tested β-lactams (Table (Table2).2). The highest MICs were observed with cephalothin and cefoxitin, which are typically good substrates for class C β-lactamases. Clavulanic acid and tazobactam did not exhibit any synergistic effect, as these compounds are only moderately active on AmpC enzymes, although some inhibitor-sensitive AmpC variants have also been described (16). In comparison with A. enteropelogenes, the MIC values measured with E. coli NI540 are higher globally (especially with amoxicillin), likely reflecting a different level of enzyme production and/or potentially different permeabilities and different sensitivities of the penicillin binding proteins (PBPs).
The TRU-1 β-lactamase was produced by using recombinant E. coli strain BL21(DE3)[pET-TRU-1], and ~1.5 mg of pure enzyme could be obtained per liter of culture by using two chromatographic steps. The purified enzyme preparation was subjected to ESI-MS analysis, which revealed a single protein species with a molecular mass equal to 39,275 Da, in agreement with the theoretical mass obtained after the cleavage of a 23-residue NH2-terminal signal sequence (theoretical mass value, 39,258 Da).
From a functional standpoint and by looking at the catalytic efficiencies, the enzyme showed a rather broad substrate profile. However, if we consider the turnover rates (which here mostly reflect the rates of deacylation of the acyl enzyme adduct, while the kcat/Km reflects the rate at which the acyl enzyme is formed from the free enzyme and the substrate), it efficiently hydrolyzed only penicillin G, most narrow-spectrum cephalosporins (except cephaloridine), and cefoxitin (kcat/Km values of >3.0 × 106 M−1·s−1). It also hydrolyzes, to some extent, cefuroxime, cefotaxime, cefepime, and ceftazidime, with the latter showing a rather high catalytic efficiency, close to 106 M−1·s−1 (Table (Table3).3). Most oxyiminocephalosporins exhibited very low turnover rates overall, although these substrates were often well recognized by the enzyme (as shown by the extremely low Km values measured). On the other hand, the cephalosporins exhibiting a positively charged side chain (due to the presence of a 1-pyridylmethyl, such as in cephaloridine and cefepime) were poorly recognized by the enzyme, as illustrated by the relatively high Km values (≥360 μM). High Km values, compared to those of cephalothin, were also reported for cephaloridine and the class C β-lactamase of Aeromonas sobria (AsbA1) (Km values for cephaloridine and cephalothin of 340 and 25 μM, respectively) (35). It is also interesting that oxyiminocephalosporins showed rather different kinetic behaviors from each other and from other cephalosporins, resulting in wide variations of catalytic efficiencies. Indeed, moderate to very low kcat values were measured with other substrates (cefpirome > cefepime > ceftazidime > ceftriaxone > cefuroxime > cefotaxime), which was overall concomitant with a decrease in Km values, explaining the moderate to high values of catalytic efficiencies observed. Thus, the presence of the oxyimino group in these substrates does not per se totally explain the kinetic differences observed with narrow-spectrum cephalosporins. As mentioned above, the positive side chain present in some substrates (e.g., cefepime) might strongly contribute to a large increase in Km values.
This study demonstrates that the unique pattern of susceptibility to ampicillin and resistance to cephalothin of A. enteropelogenes (A. trota) was due to the production of a single inducible cephalosporinase. This finding further supports the taxonomic differentiation of this species among members of the Aeromonadaceae. It should be noted that blood agar or inositol-starch agar, supplemented with ampicillin, is often used for the recovery of Aeromonas spp. and Plesiomonas spp. from stool culture. This common laboratory isolation technique, in addition to geographical variations, could explain the relatively low prevalence of A. enteropelogenes in most studies, which might be significantly underestimated. A study of Aeromonas sp. isolates recovered from children in Thailand showed that the use of taurocholate-tellurite-gelatin agar (TTGA) medium, which does not include any antibiotic, for the isolation of these organisms resulted in an equal proportion of A. enteropelogenes and A. hydrophila isolates in children from Thailand (1).
From a functional standpoint, TRU-1 is characterized by extremely low turnover rates for most expanded-spectrum cephalosporins and cefoxitin, although kcat/Km values are similar overall to those of penicillins and narrow-spectrum cephalosporins, due to very low Km values. Cefpirome and cefepime differ from the other expanded-spectrum cephalosporins in that they are associated with higher kcat values but also higher Km values.
Of great interest is the ability to obtain cephalosporinase hyperproducer mutants, which could be used as interesting models for further studies on the molecular mechanisms by which β-lactamase expression is regulated in aeromonads. The relatively high frequency of recovery of these mutants indicates that a single mutational event is required for β-lactamase derepression. The nature of this mutation remains to be determined, as at this stage, we were able to exclude only the involvement of the moxR gene and the immediately downstream region where the promoter of the blaTRU-1 gene is found.
This work was partially supported by grants from the European Union (COBRA Project, contract no. LSHM-CT-03-503335) and from the Italian Ministero dell'Istruzione, l'Università e la Ricerca (MIUR) (contract no. 2005061894_004) (G.M.R.).
Published ahead of print on 1 February 2010.