Detection of Nucleotide Variability in rpoB in both Rifampin-Sensitive and Rifampin-Resistant Strains of Chlamydia trachomatis

Abstract

A 656-bp PCR fragment from rpoB was sequenced from five rifampin-resistant Chlamydia trachomatis variants selected in vitro from a wild-type parent with a surprising level of genetic variability in this region. Three variants (MIC, 4 μg/ml) showed Ala522→Val in cluster I (codons 507 to 533), which harbors mutations in most rifampin-resistant bacteria. Two high-level resistance variants (MICs, 64 and 256 μg/ml) showed His526→Tyr in cluster I with additional genetic variation, some of which resulted in amino acid substitutions. None of the latter was situated in clusters related to rifampin resistance in other bacteria.

Rifampin exerts its antibacterial effect through binding to the β-subunit of bacterial RNA polymerase (RNAP). However, resistance to rifampin can develop rapidly and has been described for a number of bacteria, including Escherichia coli (11) and Mycobacterium tuberculosis (18-20). The primary mechanism of resistance is decreased binding capacity of RNAP for rifampin caused by nucleotide changes localized in the central region of the rpoB gene, specifying the β subunit of the enzyme. Although there are three principal clusters harboring nucleotide changes (I to III, amino acid positions 507 to 533, 560 to 572, and 687, respectively), more than 90% of mutations are located in cluster I.

The obligate intracellular bacterial pathogen Chlamydia trachomatis is a major cause of urogenital and ocular infections worldwide. This organism has been shown to be highly sensitive to rifampin in susceptibility studies (e.g., references 4, 5, 7, 17, and 24). However, there are reports of emergence of resistance in in vitro systems (12, 13, 21). For Chlamydia, determinations of MIC for various antibiotics traditionally have been done by adding drug along with elementary bodies to permissive cells. However, this does not represent the in vivo situation, since antibiotics are used against Chlamydia only after intracellular infection has been established. We developed an in vitro cell culture model to investigate the long-term effect of antibiotics on established infection with C. trachomatis (serovar K) in epithelial cells (8). Briefly, antibiotic-free cultured HEp-2 cells were inoculated with C. trachomatis elementary bodies at a multiplicity of infection of 0.075. As previously reported, the effect of rifampin alone was investigated, as well as its combination with azithromycin (9). Treatment, starting 2 days after infection, was done with 0.015 μg of rifampin/ml, which is twofold higher than the MIC of 0.0075 μg/ml. Inhibition given by antichlamydial drugs was monitored in this system over a period of 20 days. Although rifampin alone generally proved able to both inhibit chlamydial growth and suppress de novo synthesis of bacterial rRNA, recurrent infection occasionally occurred in two independent experiments. In those instances, typical chlamydial inclusions were identified and we were able to recover infectious Chlamydia from the cultures. We prepared five variants that were clearly resistant, as determined by susceptibility testing. Three variants showed a MIC of 4 μg/ml, and the remaining two exhibited a higher level of resistance, with MICs of 64 and 256 μg/ml, respectively. Based on previous publications, we assumed that mutations in the rpoB gene would be responsible for the development of resistance to this drug in Chlamydia. We therefore prepared DNA from each of the five variants plus the original stock and amplified a 656-bp PCR product by using the primers rpoB-US 5′-GCGAATGGGCGATGAGAAGA-3′ and rpoB-DS 5′-CCGTACTTGTGTCGGCTTCA-3′; these flank clusters I and II in the central portion of the rpoB gene. High Fidelity Taq polymerase (Invitrogen) was used to avoid amplification errors. The amplified PCR products were cloned into the pGEM-T Easy vector (Promega, Madison, Wis.) and were subsequently sequenced.

As a control for sequencing of the rifampin-resistant variants, the original susceptible strain was analyzed. Analysis of 20 independent clones revealed that the original stock comprised a genetically nonhomogeneous population (Table 1). Although 13 of 20 clone inserts showed 100% identity to the congruent nucleotide sequence region of C. trachomatis serovar D/UW-3/Cx (http://www.stdgen.lanl.gov), seven inserts contained base changes in the central region of rpoB. Three of those seven were silent; two of three inserts had two nucleotide changes each. The remaining four clone inserts carried base differences that resulted in amino acid changes; one clone harbored two such substitutions. No amino acid substitution was shown in either cluster I or II, known to be related to rifampin resistance. Four changes were located downstream of cluster II, whereas one was situated upstream of cluster I. A variety of amino acid substitutions was found, with changes of hydrophobic amino acids to other hydrophobic (Ala617→Val, Ile472→Val, and Ile587→Met), a change of Glu to basic Asp, and a change of polar Thr to Ala. The data suggest that more DNA sequence variation exists in the C. trachomatis genome than previously thought. The stock used in this study was obtained in 1989 (Washington Research Foundation, Seattle, Wash.), and it has been passed in culture since that time by a number of different people in the laboratory. Slightly differential handling of the stock over the years undoubtedly contributed to generation and/or maintenance of such genetic variability, since the original stock was not clonal. Importantly, culture conditions differ among the many laboratories that work with Chlamydia, and selective pressures exerted by such differing conditions over time would also lead to chlamydial populations that are nonhomogenous in nucleotide sequence.

TABLE 1.

Nucleotide variations identified in the rifampin-susceptible wild-type stock of C. trachomatis serovar K

Nucleotide change	Corresponding position in E. coli gene	Location (cluster)a	No. of clones
No mutation			13
GCT→GTT (Ala562→Val)	617	DS of II	1
TGT→TGC (silent)	559	I-II	1
CGT→CGC (silent)	592	DS of II
ATA→ATG (Ile532→Met)	587	DS of II	1
CAG→CGG (Gln563→Arg)	618	DS of II
GTT→GTC (silent)	610	DS of II	1
GAA→GAG (silent)	611	DS of II
ACT→GCT (Thr576→Ala)	631	DS of II	1
ATC→GTC (Ile417→Val)	472	US of I	1
TTA→TTG (silent)	533	I	1

^aDS, downstream; US, upstream.

Each of the five resistant variants carried an amino acid change in cluster I (Table 2). The three variants with a MIC of 4 μg/ml (1-3) each had a change at codon 522 (E. coli numbering system), which led to a change of Ala to Val. The remaining two variants (variants 4 and 5), each of which exhibited high-level rifampin resistance (MICs were 64 and 256 μg/ml, respectively), had a substitution of His526→Tyr. Both amino acid changes correspond to those known in other bacteria and have been shown to play a role in rifampin resistance. Substitutions at position 526 are among the most common in other bacteria, whereas changes at position 522 are less frequent but have been observed (1, 2, 11, 20, 23).

TABLE 2.

Nucleotide differences identified in rifampin-resistant isolates of C. trachomatis

Variant	MIC (μg/ml)	Nucleotide change	Corresponding position in E. coli gene	Location (cluster)a	No. of clones
1	4	GCA→GTA (Ala467→Val)	522	I	1
2	4	GCA→GTA (Ala467→Val)	522	I	1
3	4	GCA→GTA (Ala467→Val)	522	I	8
4	64	CAC→TAC (His471→Tyr)	526	I	13
	64	CAC→TAC (His471→Tyr)	526	I	1
		CCT→TCT (Pro535→Ser)	590	DS of II
	64	CAC→TAC (His471→Tyr)	526	I	1
		CGA→CAA (Arg395→Gln)	450	US of I
	64	CAC→TAC (His471→Tyr)	526	I	1
		ATT→ATA (517 silent)	572	II
	64	CAC→TAC (His471→Tyr)	526	I	1
		GCA→ACA (Ala497→Thr)	552	I-II
5	256	CAC→TAC (His471→Tyr)	526	I	14
	256	CAC→TAC (His471→Tyr)	526	I	1
		AAA→GAA (Lys584→Glu)	639	DS of II
	256	CAC→TAC (His471→Tyr)	526	I	1
		GCA→GCC (392 silent)	447	US of I
	256	CAC→TAC (His471→Tyr)	526	I	1
		GAC→GGC (Asp494→Gly)	549	I-II

^aDS, downstream; US, upstream.

Eight cloned inserts were sequenced from strain 3, all of which showed precisely the same sequence in the 656-bp fragment. The high-resistance variants, i.e., variants 4 and 5, however, both represent a nonclonal population. Four clones of variant 4 and 3 clones from variant 5 harbored a second base change. Each of these was unique, and they were distributed over three different parts of the central region of the rpoB gene. These include the regions upstream of cluster I, between clusters I and II, and downstream of cluster II; none was identified in the known clusters related to rifampin resistance. One base change of each variant was silent. No systematic changes in the charge, polarity, or size of side groups was identifiable for the resulting amino acid substitutions. To our knowledge, no reports exist of genetic variations within one isolate of rifampin-resistant bacteria. Some publications have described double base changes in isolates of M. tuberculosis, Staphylococcus aureus, and Streptococcus pneumoniae (3, 10, 15, 16, 18, 22), all of which were found primarily in cluster I.

Although the number of resistant variants studied here is small, some correlation appears to exist between the level of resistance and the position of the nucleotide changes found in cluster I. The replacement of His526 with Tyr resulted in high-level resistance in variants 4 and 5, whereas substitutions at position 522 engendered lower MICs. These results are consistent with those from reports investigating rifampin resistance in E. coli and M. tuberculosis (11, 14, 18). Aubry-Damon et al. (1) reported some correlation not only with the position of the change but also with the nature of the new amino acid. The two clones of variant 5 harboring double base changes with resulting changes from Lys→Glu and Asp→Gly represent drastic alterations likely to result in conformational changes in the gene product. Structural studies with Thermotoga aquaticus demonstrate that the common amino acid substitutions in cluster I and II in rpoB do not occur at the active site of the enzyme and that most such changes comprise replacement of amino acids having small side chains by those having bulkier side chains (6). It is unclear what impact substitutions in other part of the β subunit have on the binding of rifampin. The high-resistance Chlamydia variants, i.e., variants 4 and 5, harbored the same cluster I His526→Tyr change, but they differ by fourfold in their MIC, indicating that other mechanisms or other alterations in amino acid sequence in some other rpoB region contribute to resistance. None of the nucleotide changes in the parental strain were found in the resistant variants, which might suggest that the additional mutations found in variants 4 and 5 are responsible for the different levels of resistance.