INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

-Lactamases catalyze the hydrolysis of the -lactam ring of -lactam antibiotics. Genes that encode -lactamases can be found on the bacterial chromosome or on plasmids (21). Expression of chromosomal -lactamase is either inducible (most chromosomal -lactamases) or constitutive (29). Klebsiella oxytoca is a bacterium that carries a chromosomally encoded -lactamase, which is naturally synthesized constitutively at a low level.

In a recent study, the chromosomal -lactamase genes of K. oxytoca were divided into two main groups: bla_OXY-1 and bla_OXY-2. Some plasmid-encoded -lactamases, such as MEN-1 (3) and TOHO-1 (16), are derived from these K. oxytoca -lactamases. The two -lactamase genes bla_OXY-1 and bla_OXY-2 were found to be very similar (nucleotide identity, 87.3%) (12). However, it has been demonstrated that OXY-2 -lactamase hydrolyzed -lactams differently from OXY-1 -lactamases, and within the OXY-1 group, some differences in the catalytic efficiencies were also observed (11). Several strains can overproduce -lactamase 73- to 223-fold (8, 10). This -lactamase overproduction confers on the bacteria resistance to penicillins and some extended-spectrum -lactams, especially aztreonam. The molecular basis of this overproduction is point mutations in the promoter consensus sequences of the -lactamase gene (10). Three types of mutation were described: a G-to-A transition of the fifth base of the 10 consensus sequence, a G-to-T transversion of the first base of the same hexamer, and a T-to-A transversion of the fourth base of the 35 sequence (9). Rarely, a double mutation is observed (9). However, the amounts of -lactamase in strains carrying the same promoter can be slightly different (9, 10). This could be explained by the difference in -lactamase substrate profile (11).

In order to verify that the mutations previously described increase the promoter strength and to quantify the effects of these mutations on the promoters, we determined the relative strengths of the different versions of the promoters from bla_OXY-1 and bla_OXY-2 in two different species, Escherichia coli and K. oxytoca, using a promoter-probe vector. As OXY-1 and OXY-2 -lactamases have different catalytic efficiencies, the bla_OXY promoters were cloned upstream of the cat gene. A 4- to 31-fold increase was observed when the promoter was carrying one of the previously described mutations. The influence of several factors involved in constitutive promoter regulation (supercoiling, upstream promoter DNA sequence, and gene copy number) was also studied.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. The strains and plasmids used in this study are listed in Table 1. All K. oxytoca strains used for the cloning of promoters were previously identified and analyzed for the promoter sequences and the types (OXY-1 or OXY-2) of their -lactamases (12).

MICs. MICs for K. oxytoca strains of novobiocin were determined by an agar dilution method as previously described (8).

Cloning of the bla promoter in pKK232-8. The 480-bp EcoRI-PstI fragments of pBOF-1 and pBOF-4 (8) were isolated and cut with AluI. The 60-bp EcoRI-AluI fragment was cloned into pBluescript II KS(+) (Stratagene) cut with the EcoRI plus EcoRV enzymes to create pLQ921 and pLQ923, respectively. These clones were recut with BamHI plus HindIII, and the fragments were cloned into pKK232-8 (Pharmacia Canada) cut with BamHI plus HindIII to create pLQ922 and pLQ924, respectively (Table 1) (Fig. 1).

View larger version (8K): [in this window] [in a new window]   FIG. 1.   Maps of DNA fragments cloned in this study. The fragments containing the bla promoters () were cloned upstream of the cat gene (). The EcoRI site and the 35 and 10 regions are also indicated. The numbers refer to the positions of the base pairs according to the sequence of blaOXY-1 (EMBL, GenBank, and DDBJ data library accession no. Z30177).

Cloning of the bla gene into pBGS18+. The entire chromosomal -lactamase genes of SL781 and SL7811 were cloned into pBGS18+. The primer containing a BamHI site described above and another, previously described, primer containing an SmaI site (11) were used. The PCR products from SL781 and SL7811 cut by BamHI and SmaI were cloned into pBGS18+ (31) cut by the same enzymes to give pBF9 and pBF10, respectively (Table 1).

Enzyme assays. Enzyme activities were assayed in crude extracts prepared by sonication in 1 mM Tris-HCl (pH 7.6) (19). Protein concentrations were measured by the Bradford protein assay (Bio-Rad). Chloramphenicol acetyltransferase (CAT) activity was determined with the phase-extraction assay previously described by Seed and Sheen (30). Crude extracts were diluted in order to have an activity within the linear range of the assay, i.e., 0.5 to 19% butyrylated chloramphenicol. -Lactamase activity was determined spectrophotometrically at 25°C by using cephaloridine as the substrate as previously described (20).

Plasmid pLQ943 construction. Plasmid pSC101 (6) was cut by StuI and ligated to eliminate the 1.5-kb region between the two sites. The resulting plasmid, pLQ942, was digested by SphI and EcoRV. The resulting 3.0-kb fragment contained the pSC101 replicon. Plasmid pKK232-8 was digested by AlwNI. The nuclease Bal-31 eliminated about 100 bp of the beginning of RNAII (of the pMB1 replicon). The plasmid was then filled in with the Klenow fragment and cut with NspI. The resulting 3.9-kb fragment containing the bla and cat genes was ligated with the SphI-EcoRV fragment of pSC101 to create the recombinant plasmid pLQ943 (Table 1).

Copy number measurement. Plasmid pACYC177 was used as an internal control and introduced into the strains carrying the different plasmid constructions. Plasmid DNA concentrations were determined in unfractionated detergent sodium dodecyl sulfate lysates by using the method of Chiang et al. (5). The ethidium bromide-stained agarose gels were photographed, and the DNA bands were quantitated by scanning the photographs with a Bioimage, Visage 110S (Millipore). The density of the larger band was normalized to that of the smaller band (from pACYC177) according to their molecular weights.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

-Lactamase promoter strength in pKK232-8. The promoters of different OXY -lactamase genes were cloned upstream of the cat gene of pKK232-8. This plasmid carries a cat gene without a promoter and a bla_TEM-1 gene. The CAT activity is proportional to the bla_OXY promoter strength, and the TEM -lactamase activity is proportional to the plasmid copy number (2): determination of the CAT activity to TEM -lactamase activity ratios was used to avoid effects due to plasmid copy number. These plasmids were introduced into either E. coli HB101 or K. oxytoca SL901. The ratios and the sequences of the promoters are shown in Table 2. The strengths of the wild-type promoters from the -lactamase genes bla_OXY-1 and bla_OXY-2 (from plasmids pLQ925 and pLQ927, respectively) were similar: less than a twofold difference was observed. In E. coli, the GT transversion in the first base of the 10 consensus sequence resulted in a 20-fold increase of the promoter strength of plasmid pLQ925. This increase is greater than that produced by the two other principal mutations. With the same transversion from the bla_OXY-2 gene only a fivefold increase was observed. The two other mutations, a transition (GA at position 5 of the 10 consensus sequence) and a transversion (TA at position 4 of the 35 consensus sequence), had similar effects, resulting in a four- to ninefold increase over wild-type promoter strength. One promoter carries both of these mutations (transition in the Pribnow box and transversion in the 35 hexamer) (plasmid pLQ939). This promoter resulted in a 15-fold increase in CAT expression in E. coli. Secondary mutations (TC and GC at 6 and 4 bp upstream of the 35 consensus sequence, respectively) did not significantly modify the promoter strength either in the wild-type promoter (pLQ926) or in the strong promoters (pLQ929, pLQ930, pLQ933, and pLQ934). However, a GA mutation at 3 bp downstream of the 10 consensus sequence (plasmid pLQ931) seemed to increase the promoter strength by about twofold.

View this table: [in this window] [in a new window]   TABLE 2.   Promoter sequences cloned upstream of the cat gene in pKK232-8 and ratios of CAT activities to TEM -lactamase activities for these different promoters

Effect of supercoiling on promoter strength. In order to study the effect of supercoiling on the bla_OXY promoter, K. oxytoca SL901 containing the plasmid pLQ925 carrying a wild-type promoter was grown for 2 h during the exponential phase in the presence of three different concentrations of novobiocin (2, 6, and 12 µg/ml, corresponding to 1/15, 1/5, and 1/2.5 of the MIC of novobiocin). Novobiocin inhibits DNA gyrase and reduces supercoiling. However, the CAT and TEM -lactamase activities of the cells grown in the presence of novobiocin were similar to those of the cells grown in the absence of novobiocin (data not shown), suggesting that the bla promoter is not controlled by supercoiling.

Effect of an upstream DNA sequence on promoter strength. When we sequenced the region upstream of the -lactamase gene, we found another apparent divergently transcribed gene whose putative promoter overlapped that of the -lactamase gene (data not shown). This new gene had similarity with several other genes for which sequences are deposited in the data libraries. These genes, a dehydrogenase involved in rhizopine catabolism in Rhizobium meliloti and a glucose-fructose oxidoreductase of Zymomonas mobilis, are both regulated. This finding suggested that the -lactamase promoter could be influenced by the regulation of this upstream gene. In order to determine if the DNA located upstream of the promoter influenced bla_OXY promoter strength, the BamHI-HindIII fragment containing the wild-type bla promoter was truncated 160 bp by using an EcoRI site present at 18 bp upstream of the 35 consensus sequence of the bla_OXY promoter. The putative promoter of the divergently transcribed gene was removed from this new fragment. The activity of this plasmid, pBF11, was compared to that of pLQ925, which carries the whole 360-bp promoter fragment. No significant difference between the CAT activity/TEM -lactamase activity ratios of pBF11 and pLQ925 was observed (data not shown), indicating that this putative gene does not affect the expression of -lactamase.

Effect of gene copy number on promoter strength. In order to verify that the observed difference was related to the plasmid copy number, a plasmid with a copy number lower than that of pKK232-8 was constructed. In plasmid pLQ943, the replicon of plasmid pSC101 was substituted for the pMB1 replicon of pKK232-8. The relative plasmid copy number was determined by two methods. First, the plasmid DNA was extracted and analyzed in agarose gels with pACYC177 used as an internal control. The gel photograph was then scanned. pKK232-8 showed a copy number of 25 copies per cell either in E. coli or in K. oxytoca, while pLQ943 showed copy numbers of 10 copies per cell in E. coli and 7 copies per cell in K. oxytoca. Second, the TEM -lactamase activities produced by both plasmids were determined. The TEM -lactamase activity was proportional to the copy number of the plasmid (2). In E. coli, the quantity of -lactamase produced by pLQ943 was 3.9-fold lower than that produced by pKK232-8, and in K. oxytoca, the difference was 4.7-fold. This confirmed that the plasmid pLQ943 has a copy number about fourfold lower than that of pKK232-8.

View this table: [in this window] [in a new window]   TABLE 3.   Ratios of CAT activities to TEM -lactamase activities of cells containing the different promoters cloned into pLQ943

View this table: [in this window] [in a new window]   TABLE 4.   Comparison of OXY -lactamase activities of chromosomal and plasmid-mediated -lactamase genes

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The core promoter is composed of the 35 and 10 hexamers separated by a spacer. The promoter specifies more-rapid transcription as its elements approach the consensus sequences: the 35 hexamer (TTGACA), the 10 hexamer (TATAAT), and the spacer between these two sequences (17 bp) (13, 15). Alterations in the strength of nonconsensus promoters are usually dependent on both the position of the substitution and the particular base substituted (17). The three different up-promoter mutations in the -lactamase promoters cloned into pKK232-8 resulted in a 4- to 31-fold increase of the CAT activity produced by the wild-type (pLQ925) promoter. The GT transversion of bla_OXY-1 in the first base of the 10 consensus sequence causes a greater increase (20-fold in E. coli and 31-fold in K. oxytoca) in the promoter strength for the wild-type promoter than the two other principal mutations. With the same transversion from the bla_OXY-2 gene only a fivefold increase is observed. It is difficult to explain the difference observed between the bla_OXY-1 and bla_OXY-2 promoters carrying the same mutation. However, the base pair just upstream of the 10 consensus sequence is different: it is a G in bla_OXY-1 and an A in bla_OXY-2 (Table 1). The mutation GA at this position was shown to decrease 10-fold the activity of the trp promoter (25). This single base-pair difference just upstream of the first base of the 10 hexamer could perhaps affect the efficiency of the up-promoter mutation in the adjacent base. The GA mutation at 3 bp downstream of the 10 consensus sequence (plasmid pLQ931) seemed to increase the promoter strength about twofold. This mutation is situated between the 10 consensus sequence and the transcriptional start site. The base pair at this position does not seem to be particularly conserved according to the study of compiled promoters (14). However, some up-promoter mutations observed correspond to nonconsensus base pairs (outside the conserved hexamers) altered to other nonconsensus base pairs. These data suggest that a hierarchy of base pair preference could exist at some positions around the hexamers.

The change from 17 to 16 bp in the GT mutant resulted in a ninefold decrease of the promoter strength in comparison with the level observed for the transversion alone. In the promoter compilation study, 92% of promoters had interregion spacing of 17 bp (14). Maximum activity was observed at a spacer length of 17 bp (1).

It is interesting that the expression of the cloned bla_OXY promoters in E. coli is different from that in K. oxytoca, particularly for promoters carrying strong mutations. This might be explained by the fact that E. coli and K. oxytoca RNA polymerases may read the promoter differently. We can speculate that the nature of the RNA polymerases or their relative concentrations within the two cells could affect the relative values of the promoter strength.

The GT mutant promoter is sevenfold less overexpressed, in comparison to the range observed on the chromosome, when the promoters are cloned into a plasmid. Mechanisms, such as gene amplification, that could contribute to disproportionate overexpression of chromosomal -lactamase production were eliminated in previous work (10). In addition, our group showed in previous studies (8, 9) that the spontaneous single-step mutants obtained in vitro always carry one mutation in the chromosomal promoter along with a 70- to 200-fold increase of the -lactamase activity. Finally, we introduced plasmid pLQ925, which carries the wild-type promoter, into K. oxytoca SL781 and SL7811. SL7811, which is a single-step mutant obtained from the wild-type strain SL781, carries the transversion in the 10 consensus sequence and overproduces its chromosomal -lactamase 223-fold (8). The CAT activities measured from these two strains were similar (data not shown), indicating that no factor acting in trans is involved in this overproduction mechanism. These findings indicate that the promoter mutation alone is responsible for the huge increase of the chromosomal bla_OXY production and that the sevenfold difference observed between the cloned and chromosomally located promoters is due only to the former being plasmid borne. Different hypotheses could explain this observation.

The production of -lactamase in K. oxytoca is constitutive (10, 18). Constitutive promoters are defined as carrying in their sequences all instructions specifying the transcriptional efficiency of their complexes with RNA polymerase, in contrast to regulated promoters, which need positive or negative effectors. During the last decade of research on transcription, it has been shown that several factors could modify the activity of a constitutive promoter. One of the most-studied factors is DNA supercoiling: it influences the expression of many bacterial genes, enhancing or repressing the activity of some of them while having no apparent influence on many others (4, 23, 28). Plasmids in general are much more supercoiled than chromosomal DNA. However, we showed that novobiocin concentrations approaching the MIC do not modify bla_OXY promoter strength, suggesting that supercoiling is not a factor involved in expression.

The major determinants of promoter strength are the 35 and 10 hexamers and the conformation of their spacer sequence. However, sequences outside this region but cis to it can strongly influence activity. A survey of nucleotide sequences of promoters indicates a clear relationship between promoter strength and the presence of upstream regions of curved DNA (24, 27). The deletion of the 160-bp sequence upstream of the bla_OXY promoter does not modify its activity, indicating that this upstream sequence is not necessary to the promoter strength. Furthermore, no sequences associated with curved or bent DNA were found in this region (data not shown). Another hypothesis is interaction between promoters due to the presence of another gene transcribed divergently from the -lactamase gene. The promoters of the two genes might have a certain degree of overlap. One of the two promoters may be subject to positive regulation by a factor which represses the activity of the second (and otherwise constitutive) promoter. Although this situation is frequently observed in the regulation of metabolic operons (7, 26), an unrelated gene upstream of the -lactamase gene could influence the expression of the -lactamase promoter. Although we found an unrelated gene transcribed divergently from the -lactamase gene, this gene does not seem to interfere with -lactamase expression.

The most probable explanation is that promoters that are close to the consensus sequence may be limited in their activities by promoter clearance or by competition among strong promoters for the available RNA polymerase. When we cloned the OXY -lactamase gene on a pMB1 replicon plasmid, the mutated promoter was not fully overexpressed in comparison to the wild-type promoter (Table 4). Furthermore, when the plasmid copy number was decreased, the expression of all bla_OXY promoters remained similar, indicating that the strength of this promoter is dependent on the gene copy number. As it is a strong promoter, particularly when it is mutated, the transcriptional machinery might not support a high level of transcription of one kind of mRNA. There is a limitation in the expression of strong promoters cloned into multicopy plasmids.

This study confirmed that the mutations in the promoter are responsible for the overproduction of the chromosomal -lactamases in K. oxytoca. Broadening of the spectrum of resistance caused by an overproduction of chromosomal -lactamases has previously been described. Bacteria increase production of chromosomal -lactamase by several mechanisms: by promoter mutations or acquisition of insertion sequences that take over promoter function, as occurs with the constitutive -lactamase of E. coli; by mutation in a regulator gene, as occurs with the inducible -lactamases of Enterobacter cloacae and Pseudomonas species; or by increased gene copy numbers through gene amplification (22).

In addition, we showed that the three different mutations in the OXY -lactamase promoters resulted in a 4- to 31-fold increase of the strength of the cloned wild-type promoter. The GT transversion in the first base of the 10 consensus sequence causes a greater increase in the promoter strength of the wild-type promoter than the two other principal mutations. The expression of the bla_OXY promoter in E. coli is different from that in K. oxytoca, particularly for promoters carrying strong mutations. Furthermore, the bla_OXY promoter appears not to be controlled by supercoiling or an upstream curved DNA, but it is dependent on the gene copy number.