Fitness Trade-Offs in bla_TEM Evolution

Abstract

bla_TEM-1 expression results in penicillin resistance, whereas expression of many bla_TEM-1 descendants, called extended-spectrum β-lactamases (ESBLs), results simultaneously in resistance to penicillins and extended-spectrum cephalosporins. Despite the expanded resistance phenotypes conferred by many ESBLs, bla_TEM-1 is still the most abundant bla_TEM allele in many microbial populations. This study examines the fitness effects of the two amino acid substitutions, R164S and E240K, that have occurred repeatedly among ESBL bla_TEM-1 descendants. Using a single-nucleotide polymorphism-specific real-time quantitative PCR method, we analyzed the fitness of strains expressing bla_TEM-1, bla_TEM-10, and bla_TEM-12. Our results show that bacteria expressing the ancestral bla_TEM-1 allele have a fitness advantage over those expressing either bla_TEM-10 or bla_TEM-12 when exposed to ampicillin. This observation, combined with the fact that penicillins are the most prevalent antimicrobials prescribed worldwide, may explain why bla_TEM-1 has persisted as the most frequently encountered bla_TEM allele in bacterial populations.

The β-lactamases are among the best-studied antimicrobial resistance enzymes because they confer resistance to β-lactam antibiotics, which have historically accounted for most of the global consumption of antimicrobials (21). Among the clinical populations of gram-negative microorganisms, the bla_TEM-1 gene is the most frequently detected plasmid-borne antimicrobial resistance gene. The occurrence of bla_TEM-1 was first reported in isolates of Escherichia coli and Salmonella enterica serovar Paratyphi in 1965 shortly after ampicillin was introduced into clinical use (7). In the 1970s, bla_TEM-1 became widespread among Enterobacteriaceae, and by the early 1980s, it was the most prevalent resistance gene in clinical microbial populations throughout the world (23). The TEM-1 β-lactamase primarily confers resistance to penicillins, including ampicillin. However, during the 1980s, novel TEM β-lactamases emerged that were capable of hydrolyzing both penicillins and extended-spectrum cephalosporins. The point mutations that caused this substrate expansion were almost certainly selected in response to heavy usage of extended-spectrum cephalosporins. Since 1983, when the first extended-spectrum β-lactamase (ESBL) bla_TEM allele was isolated (32), ∼160 variants of bla_TEM-1 that differ in amino acid sequence have been identified. The rapid evolution of the sequence and phenotypic diversity of bla_TEM-1 makes it a good model system for studying basic evolutionary-biology principles that have clinical importance.

While ESBL bla_TEM alleles have derived the ability to confer resistance to extended-spectrum cephalosporins, they have also retained the ancestral ability to confer resistance to penicillins (2, 29). The ability of some bla_TEM alleles to confer resistance to both cephalosporins and penicillins suggests that the frequency of those alleles should increase because they confer novel advantageous phenotypes. An obvious prediction for ESBL bla_TEM alleles is that they would either co-occur with bla_TEM-1 or replace it as the most frequently encountered allele in environments where cephalosporins are heavily used. However, neither of these patterns has been observed among clinical populations of microorganisms, and bla_TEM-1 is still the most commonly occurring allele in many microbial populations where cephalosporin resistance has been selected for (1, 3, 9, 12, 13, 17, 19, 20, 30, 34). The high frequency of bla_TEM-1 in microbial populations is counterintuitive, because the alleles descended from bla_TEM-1 confer both the advantageous cephalosporin resistance phenotypes and the ancestral penicillin resistance phenotype, which should promote their fixation in microbial populations. The fact that bla_TEM-1 is the most common gene in numerous microbial populations indicates that there may be a selective advantage for bacteria that express bla_TEM-1 rather than other bla_TEM alleles.

After bla_TEM-1, bla_TEM-12 and bla_TEM-10 are among the most commonly encountered bla_TEM alleles in the United States (28). The enzymes they encode differ from TEM-1 by either one or two amino acid substitutions, respectively. A single amino acid substitution that changes arginine to serine at site 164 of the TEM protein results in the ceftazidime-resistant mutant TEM-12 (2). An additional substitution, in which glutamic acid is replaced by lysine at site 240, gives rise to TEM-10 (33) and further enhances the ceftazidime resistance phenotype. These substitutions result in increased resistance to cephalosporins, and they do not decrease resistance to penicillins within the limits of standard susceptibility-testing assays. Based on the comparison of susceptibility tests, it is unclear why bla_TEM-12 and bla_TEM-10 have not replaced bla_TEM-1 in bacterial populations.

Although susceptibility tests are useful for predicting clinical outcomes during the course of antimicrobial therapy, they may lack the sensitivity to detect phenotypic differences that are important to clinical microbial populations. Fitness competitions between bacteria expressing different bla_TEM alleles provide a more sensitive method for detecting phenotypic differences. It is possible that differences in resistance to penicillins exist between bacteria expressing TEM-1, TEM-12, or TEM-10. While previously undetected, such fitness differences could begin to provide a basis for understanding why bla_TEM-1 has persisted as the most commonly encountered bla_TEM allele in most bacterial populations. Here, we measure the fitness differences of strains expressing TEM-1, TEM-10, or TEM-12 to determine whether expression of TEM-1 confers a fitness advantage over TEM-10 or TEM-12 expression in environments containing ampicillin, which is a commonly used penicillin.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

E. coli strain TP1 [proA23 lac-28 tsx-81 trp-30 his-51 rpsL173(strR) ampCp-1 ampC11] was used as the host for all constructs in the described experiments. This strain was selected because it contains a deletion in the chromosomal ampC gene (4, 16). All bacterial strains were cultured in 10 ml L broth (10 g NaCl, 10 g Bacto tryptone, 5 g Bacto yeast, 1 g glucose per liter of water) containing 15 μg/ml tetracycline with agitation at 37°C unless otherwise noted. The final densities of control cultures and cultures containing antimicrobials were compared.

Plasmids and site-directed mutagenesis.

bla_TEM alleles were expressed from the pBR322 plasmid, which is a moderate-copy-number plasmid containing a Tet^r selection marker and bla_TEM-1. bla_TEM-12 (R164S substitution) and bla_TEM-10 (R164S and E240K substitutions) alleles were generated by QuickChange site-directed mutagenesis (Stratagene, La Jolla, CA) of the pBR322 bla_TEM-1 allele according to the manufacturer's instructions. The mutations were confirmed by sequencing.

Susceptibility testing and growth inhibition assay.

The susceptibilities of the strains to various antimicrobials, expressed as MICs, were determined by broth microdilution according to CLSI guidelines (5). A more sensitive measurement of susceptibility to high concentrations of antimicrobials was obtained by growth inhibition assays in which L broth supplemented with an antibiotic was inoculated with 10⁶ CFU/ml and the absorbance of the culture at 600 nm was measured after 18 h. For each condition, at least three independent samples were measured.

Competitive fitness assays.

Serial-transfer competition experiments were performed to measure the strain fitness in the presence of ampicillin and ceftazidime. All competitions were performed at 37°C. E. coli strain TP1 hosting either a pBR322/TEM-1, pBR322/TEM-10, or pBR322/TEM-12 plasmid were grown to an optical density at 600 nm of ∼0.5 in LB broth supplemented with tetracycline at 15 μg/ml. Two strains expressing different TEM alleles were mixed in equal proportions on day zero, and ∼2 × 10⁶ CFU was inoculated into 10 ml of LB broth supplemented with selective antibiotics. On subsequent days, 1 μl (∼2 × 10⁵ CFU) of each of these cultures was transferred into fresh LB broth containing appropriate antibiotics, and the remaining bacteria were used to isolate plasmids for qPCR analysis as described below. The competitions took place over 1 to 4 days with four independent replicas for each condition. Competition experiments for each condition were performed at least three times. Parallel control experiments in broth without antimicrobials were performed to ensure that no differences in fitness existed between host strains or plasmids in the absence of selective conditions. There was no difference in the yields of plasmids isolated from control samples over the course of the experiment. We determined the coefficients of selection for the strains by computing the slopes of the linear-regression lines of ln(R/1 − R) against time (in generations), where R is the proportion of the population carrying one of the bla_TEM alleles as determined by qPCR, as previously described (8).

Real-Time qPCR.

Real-time qPCR was used to determine the quantities of plasmidic TEM alleles present in bacterial populations at defined time points. Plasmid DNA was isolated using Qiagen miniprep spin columns. Each 25-μl qPCR mixture contained 10⁶ to 10⁷copies (∼60 pg DNA) of the plasmid, 1× Brilliant master mix (Stratagene, La Jolla, CA), 200 nmol of each molecular-beacon probe (Biosearch Technologies), and 900 nmol of each primer (Table 1). Real-time qPCR experiments were performed on a Strategene Mx3000 multiplex qPCR system with the qPCR setting. The cycling conditions were as follows: 1 cycle of 10 min at 95°C and 40 cycles of 30 s at 95°C, 45 s at 55°C, and 30 s at 72°C. The fluorescence signal was collected at the end of each annealing step using appropriate filters. The specificity of the molecular beacons used to detect single-nucleotide polymorphisms between studied TEM alleles was confirmed experimentally. Table 2 presents the results of probe specificity tests.

TABLE 1.

DNA oligonucleotides used for qPCR assays

Primer or molecular beacon	Sequence (5′→3′)	5′-End modificationa	3′-End modification	Purpose
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
E240K_MB_F	TTGCAGGACCACTTCTGCG	None	None	Forward primer
E240K_MB_R	GATACGGGAGGGCTTACCA	None	None	Reverse primer
E240	GGCTGCTGGAGCCGGTGAGCGTGGATGCAGCC	Quasar 670	BHQ-2	TEM-1/12 allele probe
240K	GGCTGCTGGAGCCGGTAAGCGTGGATGCAGCC	CalFluor Red610	BHQ-2	TEM-10 allele probe

^aFAM, 6-carboxyfluorescein; HEX, fluorescent label with excitation wavelength of 535 and emission wavelength of 555.

TABLE 2.

Specificities of qPCR probes

Target present	Probe used	CTa
		Sybrb	HEXc
TEM-12	TEM-12		6.92
TEM-12 (10−1)	TEM-12		9.93
TEM-12 (10−3)	TEM-12		17.9
TEM-12	TEM-1	No CT
TEM-12 (10−1)	TEM-1	No CT
TEM-12 (10−3)	TEM-1	No CT
TEM-1 (10−2)	TEM-1	7.52
TEM-1 (10−4)	TEM-1	14.94
TEM-1 (10−2)	TEM-12		No CT
TEM-1 (10−4)	TEM-12		No CT
TEM-12, TEM-1 (10−2)	TEM-1, TEM-12	7.94	11.84
TEM-12, TEM-1 (10−4)	TEM-1, TEM-12	15.26	19.84
TEM-12, TEM-1 (10−2)	TEM-1	8.24
TEM-12, TEM-1 (10−2)	TEM-12		18.17
Water control	TEM-1	No CT
Water control	TEM-12		No CT

^aC_T (threshold cycle) refers to the first cycle at which the instrument can distinguish the amplification-generated fluorescence as being above the background signal. The C_T value is inversely proportional to the initial target concentration.

^bTEM-1-specific probe was labeled with Sybr green.

^cTEM-12-specific probe was labeled with HEX (fluorescent label with excitation wavelength of 535 and emission wavelength of 555).

RESULTS

Susceptibility testing.

To investigate phenotypic differences resulting from expression of TEM-1, TEM-12, or TEM-10, we performed susceptibility tests on bacteria expressing these genes in ampicillin, which is a penicillin, and in ceftazidime, which is an extended-spectrum cephalosporin. When measured by MIC, there was no decrease in ampicillin resistance associated with elevated ceftazidime resistance. As shown in Table 3, strains expressing TEM-12 and TEM-10 have the same ampicillin MIC as the strain expressing TEM-1. At the same time, TEM-12 and TEM-10, respectively, conferred 64- and over 128-fold increases in ceftazidime resistance relative to the strain expressing TEM-1.

TABLE 3.

MICs

Strain	MIC (μg/ml)a
	AMP	CAZ
TP1	32	1
TP1/pBR322	16,384	8
TP1/pBRTEM-10	16,384	>1,024
TP1/pBRTEM-12	16,384	512

^aAMP, ampicillin; CAZ, ceftazidime.

To measure ampicillin susceptibility with greater sensitivity than an MIC assay, we determined growth inhibition in lower concentrations of ampicillin (Fig. 1A). When bacteria were cultured in media containing 512 or 1,024 μg/ml ampicillin, the differences in the final densities of the cultures did not exceed 5%. After increasing the ampicillin concentration to 2,048 μg/ml, the bacteria expressing TEM-1 reached 95% of their final culture density at lower concentrations (Fig. 1A). However, at the same concentration, the final culture densities of the strains expressing TEM-10 and TEM-12 were significantly lower, reaching 42% and 72% of their control values, respectively. Further increasing the concentration of ampicillin to 4,096 μg/ml resulted in 25% growth inhibition of the strain expressing TEM-1, whereas the growth of the strain expressing TEM-10 was inhibited by 78% and the growth of the strain expressing TEM-12 was inhibited by 36% (Fig. 1A).

FIG. 1.

Growth inhibition assay. E. coli TP1 strains expressing different bla_TEM alleles were cultured with the indicated concentrations of ampicillin (A) or ceftazidime (B). The data presented are the mean values of at least two independent experiments done in triplicate; the error bars represent standard deviations.

When inhibition assays were performed with ceftazidime, the strain expressing TEM-1 was inhibited in the presence of low concentrations of the antibiotic by 99% (Fig. 1B). Bacteria carrying the plasmid with the bla_TEM-10 allele were able to grow to 100% density of the control cultures in ceftazidime concentrations up to 64 μg/ml but experienced growth inhibition at higher concentrations. At the highest tested concentration of 256 μg/ml, the strain was inhibited by 17% (Fig. 1B). The strain expressing TEM-12 showed no noticeable decrease in the final culture density when cultured in ≤16 μg/ml ceftazidime. When the concentration of ceftazidime was increased to 64 μg/ml, growth of the strain was inhibited by 64%. Increasing the concentration of ceftazidime to 256 μg/ml resulted in 88% growth inhibition of that strain.

Competitive fitness assays.

To further investigate the fitness cost of the R164S (TEM-12) substitution alone or in combination with E240K (TEM-10), we competed pairwise combinations of strains expressing either TEM-1, TEM-12, or TEM-10 in an environment containing either 2,048 μg/ml ampicillin, 4 μg/ml ceftazidime, or no antimicrobial.

As shown in Fig. 2 to 4, in a nonselective environment there were no fitness differences among the strains expressing different bla_TEM alleles. In the course of the experiment, the ratios of bacteria expressing TEM-1, TEM-12, and TEM-10 remained near the ratio measured on day 0, with only stochastic fluctuations.

FIG. 2.

Competition between TEM-1 and TEM-12. E. coli TP1 strains carrying either plasmid pBR322/TEM-1 or plasmid pBR322/TEM-12 competed against each other in broth without antibiotics or supplemented with 2,048 μg/ml ampicillin (AMP) or 4 μg/ml ceftazidime (CAZ). The data presented are averages of at least three independent experiments performed in four replicate cultures; the error bars represent standard deviations.

FIG. 4.

Competition between TEM-12 and TEM-10. E. coli TP1 strains that carried either plasmid pBR322/TEM-12 or plasmid pBR322/TEM-10 competed against each other in broth without antibiotics or supplemented with 512 μg/ml ampicillin (AMP) or 256 μg/ml ceftazidime (CAZ). Each line depicts the allele distribution in independent cultures. The data are averaged from at least three independent experiments performed in four replicate cultures. The error bars represent standard deviations.

When mixed bacterial populations were grown in the presence of ampicillin, TEM-1 went to fixation around the 25th generation (Fig. 2 and 3). In competitions between strains expressing TEM-1 and TEM-10 or TEM-1 and TEM-12 in 4 μg/ml of ceftazidime, the strains expressing TEM-10 and TEM-12 went to fixation around the 10th generation (Fig. 2 and 3).

FIG. 3.

Competition between TEM-1 and TEM-10. E. coli TP1 strains that carried either plasmid pBR322/TEM-1 or plasmid pBR322/TEM-10 competed against each other in broth without antibiotics or supplemented with 2,048 μg/ml ampicillin (AMP) or 4 μg/ml ceftazidime (CAZ). The data presented are averages of at least three independent experiments performed in four replicate cultures; the error bars represent standard deviations.

We also performed competitions between bacteria expressing TEM-10 and TEM-12. In 512 μg/ml ampicillin, the TEM-12-expressing strain exhibited a fitness advantage over the strain carrying TEM-10, but the bla_TEM-12 allele never went to fixation. After 50 generations, the TEM-10 allele was present in 10 to 20% of the population (Fig. 4). A similar pattern was observed for cultures grown in 64 μg/ml ceftazidime. The bla_TEM-10 allele, which confers higher ceftazidime resistance, became predominant in the population around the 10th generation, but even after 50 generations, it did not go to fixation (Fig. 4). A possible explanation for these results is that diffusion of the enzyme from the periplasmic space of the fitter strains provided a protective effect for the less fit strain by hydrolyzing the β-lactam in the media.

DISCUSSION

We have examined the effects of R164S and E240K substitutions on ampicillin resistance conferred by β-lactamase enzyme. The arginine at site 164 is located in the omega loop, which is involved in determining the substrate specificity of the enzyme. Substitutions in this residue are frequently observed among TEM variants (http://www.lahey.org/Studies/temtable.asp). Replacement of arginine by serine results in more flexibility within the loop, which opens more space for bulkier side chains in newly developed β-lactams, giving the R164S TEM mutant higher enzymatic activity for ceftazidime (18, 29). An additional mutation resulting in replacement of glutamic acid by lysine at position 240 further enhances this effect. However, as we have demonstrated, these substitutions result in a fitness cost when ampicillin is the selective pressure.

When strains expressing TEM-1 were cocultured with a strain expressing TEM-12 or TEM-10 in broth containing ampicillin, the TEM-1 strain was more fit than strains expressing either TEM-12 or TEM-10 and the TEM-1 allele went to fixation within the first 25 generations (Fig. 2 and 3). In both cases, we observed approximately the same values of the selection coefficient (Table 4), even though growth inhibition values differed significantly between TEM-12 and TEM-10 (Fig. 1A). In other tested concentrations of ampicillin, there was a direct correlation between the selection coefficient and the ampicillin concentration (data not shown).

TABLE 4.

Selection coefficients

TEM pair	Selection coefficient (s/generation)a
	AMP	CAZ	No antibiotic
TEM-1/TEM-10	0.28 ± 0.03b	−0.53 ± 0.18b	−0.0006 ± 0.004b
TEM-1/TEM-12	0.25 ± 0.04b	−0.32 ± 0.16b	−0.00045 ± 0.0052b
TEM-10/TEM-12	0.03 ± 0.06c	−0.221 ± 0.009c	0.00003 ± 0.00164c

^aAMP, ampicillin; CAZ, ceftazidime.

^bValue is relative to TEM-1.

^cValue is relative to TEM-12.

The most likely basis for this fitness difference is the reduced catalytic efficiency of TEM-10 and TEM-12 proteins for ampicillin hydrolysis relative to TEM-1. The TEM-1 enzyme has a k_cat/K_m value of 2.8 × 10⁷ M⁻¹ s⁻¹ for hydrolysis of ampicillin, whereas TEM-12 has a k_cat/K_m of 4.2 × 10⁵ M⁻¹ s⁻¹ for hydrolysis of ampicillin (15). For the TEM-10 enzyme, similar decreases in catalytic efficiency relative to TEM have also been noted for benzylpenicillin and oxacillin (22). The decrease in catalytic efficiency is a more likely cause of the fitness differences than expression differences, because the only differences in the plasmids were the mutations introduced into the bla_TEM alleles. A recent report of antimicrobial consumption in Europe showed that penicillin consumption exceeds cephalosporin consumption in nearly every European country (6, 10, 11). Those reports are consistent with antimicrobial consumption in the United States (Fig. 5) (14, 24-27, 31) and likely reflect global antimicrobial consumption trends.

FIG. 5.

Relative use of β-lactams. Rankings of the top 200 drugs used each year were gathered, and the rankings of all β-lactams were extracted for each year. The β-lactams were divided into three categories: penicillins, penicillin/penicillinase inhibitors (pen/inhibitors), and cephalosporins. Scores indicative of the relative usage of β-lactams were assigned by subtracting each ranking from 201. The scores for all drugs in each category were summed and plotted by year.

The results of our study combined with global antibiotic consumption trends indicate that bla_TEM-1 may be the most common bla_TEM allele in bacterial populations in the United States because it confers a fitness advantage over ESBL bla_TEM alleles in the presence of penicillins. Although the fitness differences between strains expressing bla_TEM-1 and other bla_TEM alleles found outside the United States have not been determined, it is possible that similar phenotypic trade-offs and antimicrobial consumption patterns may be a factor contributing to the high frequency at which bla_TEM-1 is detected throughout the world.