Common mechanism of ampC beta-lactamase induction in enterobacteria: regulation of the cloned Enterobacter cloacae P99 beta-lactamase gene

Expression of the chromosomal beta-lactamase from the ampC gene in inducible in both Enterobacter cloacae and Citrobacter freundii. Cloning of ampC as well as its regulatory gene, ampR, from E. cloacae P99 revealed a gene organization indentical to that of C. freundii in the corresponding region. Although almost no similarities could be found between the restriction maps of ampC and ampR in the two species, the genes cross-hybridize. Also, both ampR gene products have a size of about 31,000. The regulatory features of E. cloacae beta-lactamase induction are very similar to those in C. freundii, i.e., beta-lactamase synthesis is repressed by AmpR in the absence, and stimulated in the presence, of inducer. The AmpR function can be transcomplemented between the two species, but there are quantitative regulatory aberrations in such hybrids, in contrast to the total complementation obtained within each system. These results suggest that the mechanism of beta-lactamase induction is the same in E. cloacae, C. freundii, and other gram-negative bacteria with inducible chromosomal beta-lactamase expression.

P-lactamases can be induced in E. cloacae and C. freundii by ,-lactams in the growth medium (21,28,36). In contrast, P-lactamase synthesis is constitutive in Escherichia coli (36). Although the ,-lactamase gene from E. cloacae 208 has been cloned (33), the mechanism of induction has not been further characterized.
A regulatory gene, ampR, which decreases expression in the absence of inducer and increases it in the presence of inducer has been identified in C. freundii (21). This gene is not present in E. coli. When the ampC gene of C. freundii together with its ampR gene are introduced into E. coli, synthesis of the C. freundii P-lactamase is inducible (21).
Thus, all other factors required for this process can be supplied by the E. coli chromosome.
Chromosomal mutants overproducing 1-lactamase arise at a high frequency in both C. freundii and E. cloacae (10,12,13). In C. freundii it has been shown that these mutations are not closely linked to either ampR or ampC and consequently must affect one or several other loci (21). A similar mutation which leads to ampR-dependent overproduction of the C. freundii P-lactamase from the cloned ampC gene has been described in E. coli (21). The locus for this mutation is designated ampD (F. Lindberg, S. Lindquist, and S. Normark, unpublished data).
To investigate whether the features of ,-lactamase regulation in E. cloacae are similar to those found in C. freundii, we cloned ampC and the genes surrounding it from E. cloacae into E. coli. Owing to the enzymological interest in the P-lactamase produced by E. cloacae P99 (11,16,17), we isolated the genes from this strain. Expression of the cloned ,B-lactamase gene was inducible in E. coli. We show that both the regulation of 1-lactamase synthesis and the gene organization in the ampC region are very similar in E. * Corresponding author. cloacae and C. freundii. Partial complementation of the regulatory factors was possible between the species. MATERIALS AND METHODS Bacteria and plasmids. E. cloacae P99 is a clinical isolate which overproduces a chromosomal class C P-lactamase (11). E. coli SN03 (ampAl ampC8 recA pyrB thi) (29) or its ampD2 derivative SN0302 (21; F. Lindberg, S. Lindquist, and S. Normark, unpublished data) were the hosts in all resistance and 3-lactamase production assays. These strains produce only negligible amounts of P-lactamase activity. As an initial recipient in transformation, E. coli FLO1, a recA derivative of E. coli MM294 (26), was used. This strain was constructed by first transducing MM294 with a P1 lysate of JC10236 (srl::TnJO) (7), selecting for tetracycline resistance, and then transducing the resulting strain with a P1 lysate of HB101 (proA leuA recA56) (5), selecting for sorbitol utilization and screening for UV sensitivity (27). Approximately 50% of the sorbitol+ colonies were UV sensitive, and one of these was picked and named FLOl. In the minicell experiments E. coli ORN103 (31), previously called AA10, was used. Cloning vectors pNU78 (30) and pACYC184 (6) have been described previously. Plasmid pNU305 carries the C. freundii ampR and ampC genes, and pNU314 is an ampR insertion mutant of this plasmid (21). Both are pBR322 derivatives, encoding resistance to tetracycline. pNU311 is a pACYC184 clone of the C. freundii ampR gene (21) and also mediates chloramphenicol resistance.
Determination of ampicillin resistance. Bacteria were grown to an optical density at 420 nm of 0.8 (Zeiss PMQ3 spectrophotometer) in M9CA medium, maintaining the selection for the plasmid mediating tetracycline or chloramphenicol resistance, and then diluted 105-fold. A volume of 0.1 ml (200 to 500 CFU) was spread on each of a series of plates containing ampicillin at increasing concentrations. Resistance was defined as IC50, i.e., the concentration at which 50% of the bacteria were inhibited from forming colonies. Plates were read after 36 to 48 h at 37°C.
Determination of ,B-lactamase production. Bacteria were grown logarithmically for at least eight generations in M9CA medium selecting for plasmids that encode tetracycline and chloramphenicol resistance. At an optical density at 420 nm of 0.8, a 20-ml sample was chilled quickly on ice, and the cells were pelleted, washed twice in 50 mM phosphate buffer (pH 7.0), and resuspended in 0.8 ml of phosphate buffer. Extracts were prepared by adding 0.2 ml of 500-mg/liter lysozyme-50 mM EDTA, followed by a 30-min incubation at 37°C and sonication on ice with a Branson Sonifier B-12 equipped with a microtip (six cycles of 10-s sonication at set 3 with 30-s pauses). A suitable amount of extract (5 to 40 ,ul) was then added to 1 ml of cefalexin solution (170 ,uM in 0.1 M potassium phosphate buffer [pH 7.0]). Hydrolysis was assayed as the rate of decrease in A26o. Specific activity was expressed as micromoles of cefalexin hydrolyzed at 30°C per minute per milligram of protein, as determined by the method of Lowry et al. (22), using bovine serum albumin as the standard. Induction was achieved by two-fold dilution of the culture at an optical density at 420 nm of 0.8 into prewarmed medium containing the inducer 6-aminopenicillanic acid (6-APA). The final 6-APA concentration was 2 g/liter. After 40 min of growth a 20-ml sample was processed as described above.
Restriction mapping and construction of plasmid derivatives. The restriction map of pNU346 was constructed by digesting the plasmid with various restriction endonucleases either alone or in combinations of two or three. These digests were electrophoresed on 0.7 to 2.0% agarose gels in Tris acetate-EDTA buffer as previously described (14). A HindIII digest of bacteriophage A cI857 and a HaeIII digest of bacteriophage 4X184 (new England BioLabs, Inc., Beverly, Mass.) were used as size markers. The order of the Sall fragments of pNU346 was determined by treating a BamHI digest of the plasmid with exonuclease Bal 31 (25). Samples were taken at different times and redigested with SailI. By subsequent agarose gel electrophoresis, the Sall fragments could be ordered from their sequence of disappearance. Plasmid pNU359 was constructed by ligating a Sall-BamHI double digest of pNU346 with plasmid pUC9 (37) digested with the same enzymes, followed by transformation (24) into FLO1, selecting for cefotaxime resistance (0.2 mg/liter). Small-scale DNA preparations (4) were digested with BamHI and Sall, and one of the plasmids with the smallest insert was further characterized. Subsequently, a detailed restriction map of this plasmid was prepared.
To construct a plasmid carrying only ampR, but not ampC, pNU359 was digested with SailI and EcoRV and ligated to SalI-EcoRV-digested pACYC184, selecting for chlorampenicol resistance and screening for tetracycline sensitivity. The structure of the resulting plasmid, pNU362, was verified by restriction analysis. To obtain pNU364, we subcloned the smaller SacII-EcoRI fragment of pNU359 into the vector pNU78 digested with the same enzymes. In the construction of pNU363, a plasmid carrying the ampC gene, the vector pNU78 was digested with PstI followed by the removal of the 3' extension with T4 DNA polymerase (25) and digestion with EcoRI. Subsequently, this vector was ligated to the HincII-EcoRI fragment of pNU359 followed by transformation with FLO1 selecting for tetracycline resistance and cefotaxime resistance (0.1 mg/liter).
Southern blot analysis. Double and triple digests of plasmids pNU346 and pNU359 were separated on 1.5% agarose gels. After denaturation, the DNA was transferred (34) to Zeta-probe membranes (BioRad Laboratories, Richmond, Calif.). Three probe fragments of C. freundii OS60 DNA were isolated from pNU302 and 32p labeled as previously described (2). These fragments covered parts of the coding regions for frdA (EcoRI-HindIII), ampR (BamHIl-SacII2), and ampC (SacII3-BamHI2) (2, 21). The extent of the regions covered by the probes is shown in Fig. 1. Hybridization was done under conditions of both high and lower stringency as previously described (23). After exposure to Du Pont Cronex 4 X-ray film for 2 to 24 h, the filter was incubated for 30 min in 0.1 M NaOH at 40°C, followed by repeated washing in 0.1 M Tris hydrochloride (pH 7.5). After removal of the probe in this manner, the filter was reprobed with the other probes. Analysis of protein expression in minicells. The different plasmids were transformed into the minicell producer ORN103, and minicells were isolated by sucrose gradient centrifugation as previously described (21). Minicell preparations were labeled with a 2-min pulse of [35S]methionine, either directly or after incubation for 10 min in ampicillin (1 g/liter) (21). Equal fractions of the induced and noninduced preparations were separated in parallel on sodium dodecyl sulfate-15% polyacrylamide gels (18).  The smallest fragments hybridizing to the different C. freundii probes are indicated by horizontal bars labeled A, R, and C, indicating hybridization to the frdA, ampR, and ampC probes, respectively. (Middle) More detailed restriction map of pNU359. No sites were found in the insert of pNU359 for DraI (AhaIII), HpaI, NaeI, NcoI, or PstI. Here, the smallest fragments hybridizing to the C. freundii ampR and ampC probes are also indicated. The gene maps were constructed from the results of the hybridization experiments as well as from the assumption that thefrdA gene is followed byfrdB,frdC, and frdD as in E. coli and C. freundii. (Bottom) Schematic maps of the subclones used in the regulatory experiments are also shown. Only the inserts are drawn to scale. bp, Base pairs. GmbH. Phage T4 DNA polymerase and [a-32P]dGTP are products of Amersham Corp. Ampicillin and cefotaxime were gifts from Astra Lakemedel AB and Svenska Hoechst AB, respectively. All chemicals were of the highest grade commercially available.

RESULTS
Cloning of E. cloacae P99 3-lactamase gene. Labeled DNA representing the C. freundii OS60 P-lactamase gene (2, 21) was used to probe Southern blots of chromosomal DNA of E. cloacae P99 DNA cut with various enzymes. It was thereby established that no EcoRI sites are present within the 1-lactamase gene in strain P99 (data not shown). To clone the E. cloacae P99 ampC gene, we therefore digested chromosomal DNA with EcoRI and ligated it into pACYC184. The ligated DNA was transformed into E. coli FLO1, selecting on plates containing tetracycline and ampicillin (10 mg/liter). Five colonies were obtained, and all five mediated resistance to ampicillin (10 mg/liter) and cefotaxime (0.2 mg/liter) and expressed a ,-lactamasecephalosporinase, as assayed with cephalosporin C or cephalexin as substrates. Plasmid DNA from a clone was EcoRI digested, and its insert was recloned into pNU78. The resulting plasmid, pNU346, consists of three EcoRI fragments, the vector and two insert fragments. Repeated exper-iments to subclone separately the larger EcoRI fragment failed. Restriction maps of this plasmid and the SalI-BamHI subclone pNU359 are shown in Fig. 2.
The ,B-lactamase encoded by pNU346 was purified to homogeneity and sequenced by Edman degradation from the amino terminus. The first 20 amino acids were unambiguously identified, and the sequence agrees exactly with the sequence published for the enzyme purified from E. cloacae P99 (16).
Three fragments of C. freundii OS60 DNA from plasmid pNU305 covering, respectively, a part of the C. freundiifrdA gene (Fig. 1, A), a part of ampR (Fig. 1, R), and a segment of ampC (Fig. 1, C) were labeled and used to probe various digests of pNU346 and pNU359 in Southern hybridization experiments. The smallest fragments of pNU346 hybridizing with the three probes are shown in Fig. 2. A subclone carrying SacI-BamHI mediated ,3-lactam resistance, whereas a similar construct carrying SacI-XhoI did not (Fig.  2) (data not shown), which is in agreement with the placement of the ampC gene by hybridization experiments. These results indicate that the gene organization within this region of the C. freundii OS60 and E. cloacae P99 chromosomes is similar. The only similarity between the restriction maps of the frd-amp regions in these two strains are the ApaI and Ball sites which are similarly placed with identical distances between the sites.  . When the gene resides on a pACYC184 derivative it is put within parentheses, otherwise it is located on a pBR322 derivative. If both genes have identical origin codes, they are present on the same plasmid. When two plasmids are present in the same cell, the first is always a pBR322 derivative carrying the ampC gene, and the second is a pACYC184 derivative carrying ampR. b IC50, Concentration at which 50% of the bacteria are inhibited from forming colonies. c Specific P-lactamase expression is given relative to SN03(pNU364) grown under noninducing conditions. The ratio given is that between specific 1-lactamase expression in SN0302 under noninducing conditions and that in SNO3 under the same conditions. d -, No plasmid. e ND, Not determined.
Regulation of E. cloacae ampC P-lactamase. Expression of P-lactamase from the E. cloacae clone pNU346 was inducible by 6-APA in E. coli (Table 1). When the plasmid was transformed into the ampD2 mutant E. coli SN0302, the E. cloacae ,-lactamase from pNU346 was overproduced. These results closely resemble those obtained with the cloned C. freundii ,B-lactamase gene. Three subclones of pNU359 were constructed to further investigate the role of the postulated E. cloacae ampR gene (see Materials and Methods). The first of these, pNU364, extends from SacII through BamHI-2 and should carry frdD, ampR, and ampC (Fig. 2). The second subclone, pNU363, extends from HinclI through BamHI and carries only ampC, whereas pNU362, a subclone from Sall to EcoRV, carries the frdC, frdD, and ampR genes, but lacks ampC.
The ampicillin resistances mediated by the different clones in E. coli SN03 and E. coli SN0302 are shown in Table 1, together with the specific ,B-lactamase activities expressed relative to pNU364 in SNO3. As expected, pNU362 did not mediate any increase in ampicillin resistance or enzyme production. Plasmid pNU364 mediated a 38-fold-higher ampicillin resistance in SN0302 (150 mg/liter) as compared with the resistance mediated in SN03 (4 mg/liter). The 3-lactamase activity produced was about 190 times higher in the former strain than in the latter. The elimination of ampR, such as in pNU363, resulted in a slightly higher resistance and an approximately twofold increase in specific 1Blactamase activity in SN03. In this case both the level of resistance and ,B-lactamase production were equal in SN03 and SN0302. Both of these effects could be complemented in trans by pNU362 (Fig. 2) carrying the E. cloacae ampR gene (Table 1). Expression of 1-lactamase activity in SN03 was higher from pNU346 than from the subclone pNU364. This could be due to an external promoter reading into ampR or ampC or possibly to an additional positive regulatory factor present on pNU346 to the left offrdA or to the right of ampC in Fig. 2. Expression of 13-lactamase activity was induced by growing bacteria in 6-APA (2 g/liter). The relative amounts of specific ,B-lactamase activity were compared between ex-tracts prepared from bacteria before and after 40 min of induction (Table 1). In agreement with the data obtained with the C. freundii ampR and ampC genes (21), 1-lactamase expression was inducible in SNO3 if ampR was present in cis or in trans, and synthesis was low and constitutive if ampR was not present. In SN0302 the cloned 1-lactamase was overexpressed or constitutively expressed at a low level, depending on the presence or absence of ampR.
Identification of E. cloacae ampR and ampC gene products in minicells. Minicells were purified from cultures of E. coli ORN103 carrying the different subclones. These were pulselabeled with [35S]methionine before and after induction with ampicillin. Identical portions of noninduced and induced preparations were electrophoresed on a sodium dodecyl sulfate-15% polyacrylamide gel (Fig. 3). A polypeptide with apparent molecular weight (Mr) of 39,000 and comigrating exactly with the purified E. cloacae ,B-lactamase was expressed from pNU363 and pNU364. This protein was not expressed from pNU362. Plasmids pNU362 and pNU364 expressed an Mr-31,500 polypeptide which is very similar in size to the C. freundii ampR gene product. Plasmid pNU363, which does not carry the ampR gene, did not express the Mr-31,500 protein. Also, an Mr-13,000 polypeptide, probably the frdD gene product, was expressed from pNU364 and pNU363. As is apparent from Fig. 3, the rate of 13-lactamase synthesis increased in minicells induced with ampicillin if the ampR gene was present in cis (Fig. 3, lanes Bi and B2) or in trans (lanes Dl and D2).
Transcomplementation of ampR between C. freundii and E. cloacae. It was possible to complement ampR between the E. cloacae and C. freundii clones in E. coli minicells (Fig. 3, lanes El through F2). To investigate the quantitative aspects of these ampR complementations, E. coli SNO3 and SN0302 were transformed with pNU78-based clones carrying ampC of C. freundii (pNU314) or E. cloacae (pNU363) but not the corresponding ampR gene. In addition, derivatives of pACYC184 carrying only the ampR gene of the respective species were transformed into the strains. These were plasmids pNU311, carrying the C. freundii ampR gene, and pNU362, which carries that of E. cloacae. In SNO3, expres-sion of the E. cloacae p-lactamase was slightly repressed by the addition in trans of ampR from either C. freundii or E. cloacae. In all cases ,B-lactamase production could be induced with 6-APA, showing that ampR could indeed be trans complemented between the species. This was also true in the ampD mutant SN0302. In this strain expression from the C. freundii ampC gene was significantly higher in the presence of the E. cloacae ampR gene than with the homologous ampR gene. In fact, with this combination ,Blactamase production was so high (-5% of total protein) that the doubling time in M9CA medium increased to 70 min from the 60 min normal for SN0302.

DISCUSSION
We cloned the region carrying the p-lactamase gene from E. cloacae P99 into E. coli. The sequence for the 20 amino-terminal residues of the p-lactamase produced from this recombinant plasmid was identical to that published for the E. cloacae P99 enzyme (16). From molecular suboloning experiments and from hybridization studies with C. freundii probes, we conclude that the overall gene organization of the ampC region is identical in these two species, i.e., the genes encoding the fumarate reductase (frdA through frdD) are followed by the ampR regulatory gene and the ampC Plactamase gene (Fig. 2) [355]methionine-labeled proteins expressed in minicells were separated on a sodium dodecyl sulfate-15% polyacrylamide gel. The autoradiogram of this gel is shown. Indicated are the positions of the ampC and ampR gene products. In both cases the topmost marker is aligned with the E. cloacae P99 gene product, and the lower marker is aligned with the C. freundii OS60 polypeptide. Also shown are the positions of the C. freundii OS60frdB, frdC, and frdD gene products as well as that of the chloramphenicol acetyltransferase (CAT) expressed by the pACYC184 derivatives. Each pair of lanes (A lanes, B lanes, etc.) represents proteins labeled without induction (Al, Bi, etc.) and with ampicillin induction (1 g/liter) for 10 min (A2, B2, etc.). The source of the ampC and ampR genes is given below each lane, where C. freundii is indicated as CF and E. cloacae as EC. These letters when within parantheses indicate that the gene is present on a pACYC84 derivative, whereas otherwise they are carried on pNU78 (pBR322) derivatives. Proteins expressed from minicells harboring pNU305 were run in the A lanes, from pNU364 in the B lanes, from pNU363 in the C lanes, from pNU363 together with pNU362 in the D lanes, from pNU363 with pNU311 in the E lanes, and from pNU314 with pNU362 in the F lanes. Lane G contains proteins expressed by pNU362, and lane H contains proteins expressed by pNU311. different isolates of E. cloacae can be divided into two groups on the basis of their isoelectric points (32). The gene cloned from E. cloacae 208 by Seeberg and Wiedemann (33) expresses an enzyme belonging to one group, whereas the P99 enzyme belongs to the other. No evident similarities could be found when comparing the E. cloacae P99 restriction map with that published for E. cloacae 208, except for the similar size of the EcoRI fragments.
Inducibility of the cloned E. cloacae p-lactamase is dependent on the presence of the regulatory gene ampR. A deletion of the E. cloacae ampR gene resulted in a 1.9-foldhigher P-lactamase expression which was not increased by the addition of inducer to the medium. Both this slight increase of expression in the absence of inducer and the loss of inducibility could be complemented by ampR in trans. The same is true for the cloned C. freundii J-lactamase gene, except that the deletion of ampR decreases expression 2.5-fold in the absence of inducer. Although the restriction maps for the two ampR genes differ, the genes hybridize to one another and express similarly sized AmpR proteins (Fig.  3).
Earlier, we described an E. coli mutation, ampD2, which results in constitutive or semiconstitutive ampR-dependent overproduction of the cloned C. freundii 13-lactamase (21). The E. cloacae clone was affected in a similar manner when introduced into such a mutant, i.e., the E. cloacae flactamase was overproduced in an ampR-dependent manner.
The synthesis of C. freundii ,B-lactamase is 11-fold lower in the presence of the homologous ampR gene than when complementing with ampR from E. cloacae. Also, synthesis of E. cloacae 3-lactamase in the presence of inducer or the ampD2 mutation is much higher with the E. cloacae ampR gene than with that of C. freundii. These quantitative differences between homologous and heterologous complementations suggest a physical interaction between the AmpR protein and ampC DNA, or a product thereof. Since the p-lactamase is located in the periplasm, we find it unlikely that the AmpR protein would interact with this enzyme. Also, we have previously shown that a functional flactamase is not required for induction (20). Thus, we favor the view that AmpR interacts directly with either the ampC control region or the ampC transcript.
In conclusion, both gene organization and the regulatory features of the ampC region are almost identical in the two enterobacterial species with inducible p-lactamase production that have been studied. The regulatory function, AmpR, can be partially complemented between the species, in a manner suggesting direct interaction between this protein and the ampC control region or transcript.

ACKNOWLEDGMENTS
We are very grateful to Monica Persson for excellent technical assistance, to Christopher Korch for critically reading this manuscript, and to Ake Engstrdm for performing the amino-terminal sequence analysis. We would also like to thank Jean-Marie Frtre for providing E. cloacae P99, and Astra Lakemedel AB and Svenska Hoechst AB for gifts of P-lactam antibiotics.