Recombination and Selection Can Remove bla_TEM Alleles from Bacterial Populations

Abstract

We passaged cells expressing TEM-1 and TEM-12 from a single plasmid through either ampicillin or ceftazidime. We found that the combined effects of recombination and selection removed the bla_TEM-1 allele from the bacterial population when it was passaged through ceftazidime or the bla_TEM-12 allele when cultures were passaged through ampicillin.

The overall abundance of bla_TEM alleles in bacterial populations has caused multiple bla_TEM alleles to coexist in single strains of bacteria (1, 4-7, 9-11, 15, 16), which has created the opportunity for recombination to occur among bla_TEM alleles. The occurrence of recombination between two bla_TEM alleles on a single plasmid strongly resembles gene conversion (14) and will cause the fragment between the recombination sites to be removed if the alleles are oriented in the same direction (3). This will typically result in the removal of one bla_TEM allele from the plasmid while the remaining allele will usually be a chimera (Fig. 1). The potential for the removal of one bla_TEM allele from a plasmid creates a situation in which closely related alleles may compete directly for retention in bacterial plasmids. While the bla_TEM allele that is left on a plasmid following recombination is random, subsequent selection is not. Therefore, the combined effects of recombination and selection could result in the presence of a single allele on each copy of the plasmid and the removal of one of the alleles entirely from the population.

FIG. 1.

Schematic model of the recombination occurring in plasmid containing two regions of high homology. The dashed line marks the region containing two bla_TEM genes (TEM-X and TEM-Y) where recombination takes place (A). As a result, two circular DNA molecules are formed. One molecule (B) contains regulatory elements: the origin of replication (Ori), selection marker, and a chimeric bla_TEM gene (TEM-Z). The second molecule (C) contains the sequence existing between TEM-X and TEM-Y (recombination marker) and another chimeric bla_TEM gene (TEM-Z′). Due to the lack of an Ori element, this molecule is not able to replicate itself and is lost. Recombination can take place at any position within the bla_TEM genes, and either of the chimeras can be retained in the replicating molecule. Each recombination event is random and, depending on selective pressure, a different TEM allele is retained in the bacterial population.

To experimentally investigate whether allelic competition occurs in an environment with potential for both recombination and selection, we cloned the two bla_TEM alleles bla_TEM-1 and bla_TEM-12 into the pAC3 vector (2) in the same orientation (Fig. 1A), validated the construction by sequencing, and expressed them from wild-type E. coli K-12. We passaged these bacteria through L broth containing either 32 μg/ml ampicillin or 4 μg/ml ceftazidime. Every 24 h, the saturated cultures were diluted 100 times in fresh broth, and the remaining bacteria were used to isolate plasmid for quantitative PCR (qPCR) analysis.

Each 25-μl qPCR reaction mixture contained 10⁶ to 10⁷copies (60 pg of DNA) of the plasmid, 1× Brilliant master mixture (Stratagene, La Jolla, CA), 200 nmol of each molecular beacon probe (Biosearch Technologies), and 900 nmol of each primer (Table 1). To measure the ratio between the origin of replication (Ori) and TEM alleles, recombination markers, and TET amplicons (Fig. 1), real-time PCR with TaqMan probes was performed. Each 25-μl qPCR reaction mixture contained 10⁶ to 10⁷copies of the plasmid DNA, 1× Brilliant multiplex master mixture (Stratagene, La Jolla, CA), 200 nmol of each TaqMan probe (Biosearch Technologies), and appropriate concentrations of each primer (Table 1). The specificity of all primers and probes for the appropriate sequence was verified. Real-time PCR experiments were performed on a Stratagene Mx3000 multiplex qPCR system, with the quantitative PCR setting. The cycling conditions were either 1 cycle of 10 min at 95°C, followed by 40 cycles of 30 s at 95°C, 45 s at 55°C, and 30 s at 72°C, or 1 cycle of 10 min at 95°C, followed by 40 cycles of 15 s at 95°C and 45 s at 58°C, for the assay using molecular beacon and TaqMan probes, respectively. The fluorescence signal was collected at the end of each annealing step, using appropriate filters.

TABLE 1.

DNA oligonucleotides used for qPCR assays

Primer or probe	Sequence (5′→3′)	5′ end modification	3′ end modification	Purposea
R164S_MB_F	CCGAAGGAGCTAACCGCTT	None	None	Forward primer
R164S_MB_R	GTGTCACGCTCGTCGTTTG	None	None	Reverse primer
R164	CCGTGGCGCCTTGATCGTTGGGAACCCACGG	FAM	BHQ-1	TEM-1 allele probe
164S	CCGTGGCGCCTTGATAGTTGGGAACCCACGG	HEX	BHQ-1	TEM-12 allele probe
G238G_MB_F	TTGCAGGACCACTTCTGCG	None	None	Forward primer
G238G_MB_R	GATACGGGAGGGCTTACCA	None	None	Reverse primer
G238	GGCTGCTGGAGCCGGTGAGCGTGGATGCAGCC	Quasar 670	BHQ-2	TEM-1 wt allele probe
238G	GGCTGCTGGAGCCGGAGAGCGTGGATGCAGCC	CalFluor Red 610	BHQ-2	TEM-12 allele probe
Ori 3996_F	TTCGTGCATACAGTCCAG	None	None	Forward Ori primer
Ori 4071_R	CCGCGTTTGTCTCATTC	None	None	Reverse Ori primer
Ori	CTTGGAGCGAACTGCCTACCCG	FAM	BHQ-1	Ori probe
Rec 958_F	GGAGCTGAATTACATTCCCA	None	None	Forward Rec primer
Rec 1028_R	CAACGCCAATCAGCAAC	None	None	Reverse Rec primer
Rec	CCGCGTGGCACAACAACTGGC	HEX	BHQ-1	Rec probe
595F_TEM1	ACTGGCGAACTACTTACTC	None	None	Forward TEM primer
634R_TEM1	TGGTCCTGCAACTTTATC	None	None	Reverse TEM primer
TEM	AGCTTCCCGGCAACAATTAATAGACTG	Quasar 670	BHQ-2	TEM probe
710_F_TET	AAAGCGGTCGGACAGT	None	None	Forward tet primer
711_R_TET	GCGCTATATGCGTTGATG	None	None	Reverse tet primer
TET	CTCCGAGAACGGGTGCGCATA	CalFluor Red 610	BHQ-2	TET probe

^aAbbreviations: wt, wild type; Rec, recombination.

Based on results from previous competition experiments (13) and enzymatic studies (8, 12), we predicted that for bacteria passaged through ampicillin, bla_TEM-12 would be eliminated from the plasmid and population and that for bacteria passaged through ceftazidime, bla_TEM-1 would be eliminated from the plasmid and population.

When Escherichia coli K-12 carrying the pAC3/TEM-1_TEM-12 construct was cultured in the presence of 4 μg/ml ceftazidime, the complete removal of one allele from each copy of the plasmid present in the population usually occurred after six passages (Fig. 2A). To determine whether the same allele remained on each copy of the plasmid or if both alleles were still present in the population, we performed single-nucleotide polymorphism (SNP)-specific qPCR as previously described (13) using the probes that distinguish between the wild type and the strain carrying the mutated bla_TEM alleles (Fig. 2B). We found that bla_TEM-12 had swept completely through the bacterial population at the same rate with which recombination had removed one allele. We verified that result by sequencing and also found that the allele present after the experiment was identical in sequence to the allele originally inserted into the construct. To ensure that the bla_TEM-12 allele was retained in the population as a result of recombination rather than the occurrence of a novel point mutation, creating new bla_TEM-12 alleles, we performed qPCR using a set of probes to detect the silent mutation G238* in the bla_TEM-12 allele which was introduced prior to cloning. We followed the frequency of that silent mutation through the population over the course of the experiments and found that its frequency corresponded exactly with the frequency of the R164S substitution (Fig. 2C). These results show that bla_TEM-12 did replace bla_TEM-1 in the population via recombination.

FIG. 2.

Recombination of the pAC3TEM1-TEM12 plasmid in wild-type E. coli K-12 cultured in 4 μg/ml ceftazidime for 9 days (data are shown for a representative culture). On day 0, four independent cultures were inoculated in Luria broth supplemented with antibiotic. Saturated culture (100 μl) was transferred daily to fresh medium, and plasmid was prepared from the remaining culture. (A) qPCR analysis of isolated plasmids. For each reaction, 10⁶ to 10⁷ copies of plasmid was used. Each PCR contained four TaqMan probes labeled with different fluorescence markers specific to either the Ori, the TEM, the recombination (Rec) marker, or the TET region of the plasmid (Fig. 1). Ratio values are the number of each region relative to the number of Ori copies. (B and C) SNP analysis of plasmid DNA. (B) qPCR analysis using probes specific to codon 164 recognizing the R164S substitution present in the allele encoding TEM-12. (C) qPCR analysis using probes specific to the silent mutation present in the codon for G238 of the allele encoding TEM-12.

When we passaged E. coli K-12 carrying the pAC3/TEM-1_TEM-12 vector through ampicillin, we saw a result similar to that described above, except that the recombination removed one allele in fewer replicates than in ceftazidime (75% instead of 100%). As expected for the ampicillin passage, bla_TEM-1 rather than bla_TEM-12 went to fixation in the population (Fig. 3A and B).

FIG. 3.

Recombination of the pAC3TEM1-TEM12 plasmid in wild-type E. coli K-12 cultured in 32 μg/ml ampicillin. For experimental details, see the legend to Fig. 2. (A) qPCR analysis of isolated plasmids using TaqMan probes. (B) SNP analysis of plasmid DNA, using probes specific to codon 164 recognizing the R164S substitution present in the allele encoding TEM-12.

To create controls for both of these sets of experiments, we passaged DH5α containing pAC3/TEM-1_TEM-12 through ceftazidime and ampicillin. We used DH5α because it contains the recA1 mutation, which inhibits the occurrence of recombination. For the DH5α experiments, we never observed the occurrence of recombination. This shows that the potential for recombination is necessary for the affect we observed (Fig. 4).

FIG. 4.

Recombination of the pAC3TEM1-TEM12 plasmid in DH5α cultured in 4 μg/ml ceftazidime. For experimental details see the legend to Fig. 2. qPCR analysis of isolated plasmids using TaqMan probes.

As an additional control, we also passaged K-12 carrying the pAC3/TEM-1_TEM-12 plasmid through broth containing tetracycline. Over the course of the experiment, we did not observe the occurrence of recombination in this control strain.

These results show that the combined effects of recombination and selection can eliminate the cooccurrence of multiple alleles from a single plasmid, even when the alleles confer different resistance phenotypes. Natural plasmids are often larger and have a lower copy number than the constructs we introduced, which may reduce the frequency of recombination among similar alleles. Additionally, selection may be strong enough to maintain the multiple plasmidic alleles of a resistance gene if bacteria are exposed to multiple antimicrobials. However, these results indicate that competition may occur among plasmidic alleles and that resistance phenotypes may be eliminated from bacterial populations by the dynamics resulting from that competition.