Promoters P3, Pa/Pb, P4, and P5 Upstream from bla_TEM Genes and Their Relationship to β-Lactam Resistance

Abstract

Using an isogenic system, we have determined the impact that the four promoters known to control bla_TEM gene expression have on β-lactamase activity. For both TEM-1 and TEM-30, this activity gradually increased in relation to the presence of promoters P3, Pa/Pb, and P4 upstream of the corresponding gene. Promoter P5, only found upstream of the bla_TEM-1B gene, was related to the highest expression of this gene.

The extensive nucleotide sequence analysis of plasmid-mediated bla_TEM genes performed over the last 15 years because of their role in resistance to extended-spectrum cephalosporins and to amoxicillin-clavulanate has allowed us to observe that the expression of these genes is controlled by four different promoters (3, 6). The first promoter, called P3 (Fig. 1,) corresponds to the promoter of the bla_TEM-1A gene which is located on a Tn3 transposon and plasmid pBR322 (10). The second promoter, called Pa/Pb and found upstream of the bla_TEM-2 gene, is composed, in fact, by two overlapping promoters resulting from the mutation C→T at position 32 according to the Sutcliffe numbering system (10). The −35 and −10 regions of promoter Pa are TTGAAG and TACGCT, respectively, whereas the −35 and −10 regions of promoter Pb are GTGATA and TAATGT, respectively, the first 2 nucleotides of the −10 region of Pa corresponding to the last 2 nucleotides of the −35 region of Pb (Fig. 1). Chen et al. (2), studying the separately cloned promoters P3 and Pa/Pb, showed that the Pb sequence which was present in the fragment containing the promoter P3 did not work as a promoter and that the overlapping promoter P3 were stronger than promoter P3, with the increase in strength resulting from both the strength of Pa and the cooperative activity of Pa and Pb. The third promoter, called P4 by Goussard et al. (3), differs from promoter P3 by the substitution G162T (Sutcliffe numbering system) corresponding to the first nucleotide of the −10 region of the promoter (TACAAT instead of GACAAT). This substitution makes the sequence of the −10 region of promoter P4 closer to the consensus sequence (TATAAT) (Fig. 1). We suggest that the fourth promoter, which we have recently published, be called P5 (7). This promoter has the same sequence as promoter P3 at the −10 region but has a sequence at the −35 region (TTGAAA) closer to the consensus sequence (TTGACA) than that of promoter P3 (TTCAAA) (Fig. 1). To assess and compare the respective impact of the four promoters on β-lactam resistance, we established an isogenic system in which the only difference consisted of the type of promoter upstream from bla_TEM genes coding for either a β-lactamase susceptible (TEM-1) or resistant (TEM-30 or IRT-2) to class A β-lactamase inhibitors.

FIG. 1.

Nucleotide sequences of promoters P3, Pa/Pb, P4, and P5 of bla_TEM genes. The sequences of the −35 and −10 regions of promoters P3, P4, and P5 are boxed by solid lines, whereas these of the overlapping promoters Pa/Pb are boxed by broken lines. Boldface letters indicate the mutations which led to the creation of promoters Pa/Pb, P4, and P5. The numbers under sequences indicate Sutcliffe numbering positions.

By using primers 5′-ATA AAA TTC TTG AAG AC-3′ and 5′-TTA CCA ATG CTT AAT CA-3′, four bla_TEM-1B genes, each preceded by one of the four promoters, were amplified and cloned from previously published Escherichia coli isolates (6). The bla_TEM-30 gene coding for the inhibitor-resistant TEM-30 or IRT-2 enzyme was also amplified and cloned from previously published E. coli isolates but only with promoter P3, Pa/Pb, or P4 upstream from it, as a bla_TEM-30 gene controlled by promoter P5 has not been discovered thus far (6). Plasmid pPCR Script (Cam^r) (Stratagene, La Jolla, Calif.) and E. coli strain Epicurian Coli XL10-Gold (Stratagene) were used in the first stage of the cloning experiments. After the sequences of inserted fragments (promoter and coding region) were checked (automated cycle sequencing system on a Perkin-Elmer R377 sequencer), the fragments were then cloned into the BamHI-restricted and dephosphorylated plasmid pACYC184 (Tet^r and Cam^r) and expressed in E. coli strain NM554 (8). The size of the inserts and their insertion in the opposite orientation of the promoter of the pACYC184 tetracycline resistance-encoding gene were assessed by restriction analyses with BamHI and AseI enzymes.

The E. coli NM554 transformants containing the different recombinant plasmids were then tested in three independent experiments for β-lactam susceptibility by using the agar dilution method on Mueller-Hinton agar with a Steers multiple inoculator and 10⁴ CFU per spot. The β-lactamase activity of each transformant was also tested from the crude extracts of three independent cultures, as previously described (5). Briefly, β-lactamase was extracted from 100 ml of an exponential-growth-phase culture at 37°C in Trypticase soy broth containing amoxicillin (100 μg/ml) combined with chloramphenicol (30 μg/ml). Bacterial suspensions were disrupted by sonication, and the β-lactamase activity was measured from the crude extracts with a Pharmacia UV2000 spectrophotometer, using benzylpenicillin (100 μM) as the substrate. One unit of β-lactamase activity was defined as the amount of enzyme which hydrolyzed 1 μM benzylpenicillin per min, and the specific activity was defined as the β-lactamase activity per milligram of protein.

As indicated in Table 1, β-lactamase activity as well as MICs of amoxicillin-clavulanate, ticarcillin-clavulanate, piperacillin, and cephalothin gradually increased from the E. coli transformant expressing the bla_TEM-1B gene controlled by the P3 promoter to that expressing the bla_TEM-1B gene controlled by the P5 promoter, with higher values for the two types of parameters when bla_TEM-1B was controlled by the P4 promoter than when it was controlled by the Pa/Pb promoters. Concerning ticarcillin, the concentrations used in this study did not allow us to observe a gradual increase in MICs in relation to the promoter type. For piperacillin-tazobactam, we observed different levels of MICs: low MICs, from 1 to 2 μg/ml for the transformants whose bla_TEM-1B gene was controlled by promoters P3 and Pa/Pb, respectively, and high MICs, from 128 to 512 μg/ml for those whose bla_TEM-1B gene was under the control of promoters P4 and P5, respectively. The first two transformants were categorized as susceptible to piperacillin-tazobactam, whereas the last two were categorized as resistant following the recommendations of the French Antibiogram Committee (9). These results indicate that the C→G substitution, which occurred in the −35 region of promoter P5, is more efficient in terms of promoter strength than the G→T substitution, which occurred in the −10 region of promoter P4. In fact, promoter P5 possesses the trimer TTG in the −35 region that was considered by Kobayashi et al. (4) to be the most essential sequence for determining promoter strength. However, although the overlapping promoters Pa/Pb possess the trimer TTG in the −35 region of promoter Pa, they were related to a weaker β-lactamase activity than that related to promoter P5. This difference suggests that other nucleotide motifs than TTG in the promoter region might be involved in gene expression.

TABLE 1.

β-Lactam susceptibility and β-lactamase activity of TEM-1- and TEM-30 (IRT-2)-producing E. coli transformants according to the type of promoter upstream from the bla_TEM-1B and bla_TEM-30B genes

Strain and blaTEM gene (promoter)	MIC (μg/ml)a						β-Lactamase activity (U/mg) (mean ± SD)
	AMC (≤4->16)	TIC (≤16->64)	TCC (≤16->64)	PIP (≤8->64)	TZB (≤8->64)	CEP (≤8->32)
E. coli
ATCC 25 922	4	4	4	2	1	8	NDb
NM554	2	2	2	1	1	2	ND
Transformant NM554
blaTEM-1B (P3)	32	8,192	32	128	1	8	6.8 ± 3.3
blaTEM-1B (Pa/Pb)	128	>8,192	256	512	2	32	13.8 ± 2.1
blaTEM-1B (P4)	512	>8,192	1,024	1,024	128	64	38 ± 9
blaTEM-1B (P5)	1,024	>8,192	2,048	>1,024	512	128	87.4 ± 8
blaTEM-30B (P3)	256	128	32	8	1	8	14 ± 4.4
blaTEM-30B (Pa/Pb)	2,048	512	128	64	2	8	32.6 ± 5.6
blaTEM-30B (P4)	4,096	2,048	512	64	32	8	120 ± 7.9

^aDrug abbreviations: AMC, amoxicillin with 2 μg of clavulanate per ml; TIC, ticarcillin; TCC, ticarcillin with 2 μg of clavulanate per ml; PIP, piperacillin; TZB, piperacillin with 4 μg of tazobactam per ml; CEP, cephalothin. The MIC breakpoints (in micrograms per milliliter) are shown in parentheses after the drug abbreviations.

^bND, not determined.

Concerning the three transformants containing the bla_TEM-30gene, they differed from each other with regard to the β-lactamase activity and the β-lactam MICs with the exception of cephalothin. The lowest MICs, regardless of the penicillins tested, were found in the transformant harboring the bla_TEM-30B gene preceded by promoter P3, whereas the highest MICs were found in the transformant harboring the bla_TEM-30B gene controlled by promoter P4; intermediate MICs were found in the transformant harboring the bla_TEM-30 gene preceded by promoters Pa/Pb (Table 1). For cephalothin, all of the transformants were susceptible. For piperacillin-tazobactam, only those transformants with promoters P3 and Pa/Pb upstream from the bla_TEM-30B gene were susceptible. For piperacillin, only the transformant with promoter P3 was susceptible. For the transformants with bla_TEM-30B, the lowest β-lactamase activity was observed in the transformant harboring the bla_TEM-30B gene with promoter P3, whereas the highest activity was observed in the transformant harboring the bla_TEM-30B gene with promoter P4.

Overall, the increase in β-lactamase activity observed in relation to the presence of promoters P3, Pa/Pb, and P4 was demonstrated by using either the bla_TEM-1B gene coding for TEM-1 or the bla_TEM-30B gene coding for TEM-30 or IRT-2. Thus, we were able to compare not only the expression of a given bla_TEM gene in relation to different promoters but also the expression of different bla_TEM genes in relation to a given promoter. In this study, it has been unambiguously demonstrated that, for any one promoter, the β-lactamase activity measured by using benzylpenicillin as the substrate was higher for TEM-30 or IRT-2 than for TEM-1 from which the IRT enzyme is derived (1). However, despite this higher β-lactamase activity, the MICs of the penicillins tested were lower, with the exception of amoxicillin-clavulanate, for the transformants producing TEM-30 or IRT-2 than for those producing TEM-1 (Table 1). This result emphasizes that MICs are linked to both the type of promoters upstream of the TEM-encoding gene and to the biochemical properties of the TEM enzyme.

In conclusion, although our isogenic system has probably slightly overestimated MICs and β-lactamase activity for all of the transformants, as the vector used (pACYC184) is a low-number but not a single-copy plasmid, it has allowed us to compare the roles that the four promoters that have been described thus far and that are upstream from the bla_TEM genes have on β-lactam resistance according to the type of TEM enzyme involved.