Enterobacterial Repetitive Intergenic Consensus Sequences as Molecular Targets for Typing of Mycobacterium tuberculosis Strains

Abstract

The presence of enterobacterial repetitive intergenic consensus (ERIC) sequences was demonstrated for the first time in the genome of Mycobacterium tuberculosis; these sequences have been found in transcribed regions of the chromosomes of gram-negative bacteria. In this study genetic diversity among clinical isolates of M. tuberculosis was determined by PCR with ERIC primers (ERIC-PCR). The study isolates comprised 71 clinical isolates collected from Sardinia, Italy. ERIC-PCR was able to identify 59 distinct profiles. The results obtained were compared with IS6110 and PCR-GTG fingerprinting. We found that the level of differentiation obtained by ERIC-PCR is greater than that obtained by IS6110 fingerprinting and comparable to that obtained by PCR-GTG. This method of fingerprinting is rapid and sensitive and can be applied to the study of the epidemiology of M. tuberculosis infections, especially when IS6110 fingerprinting is not of any help.

Rapid differentiation of Mycobacterium tuberculosis strains has important public health and therapeutic implications. Methods with arbitrary primers and PCR with specific primers have been reported (6, 9, 12, 16, 19). Amplification of DNA sequences located between frequently occurring restriction enzyme sites and infrequently occurring restriction enzyme sites has been reported for mycobacteria and other strains (11). Random amplified polymorphic DNA analysis is based on low-stringency amplification and a decrease in the temperature of annealing (9). The patterns obtained can vary greatly in response to minimal changes in the amplification; moreover, the reproducibility of the experiment is very low (9). On the other hand, different specific targets have been selected in order to amplify M. tuberculosis DNA: tandem DNA repeats, drug resistance genes, insertion elements, or a combination of these (6, 12, 16, 19, 23). Usually, these targets are considered species specific (12, 16, 23).

In this study we demonstrate the presence of enterobacterial repetitive consensus (ERIC) sequences on the M. tuberculosis genome. ERIC sequences are repetitive elements of 126 bp and appear to be restricted to transcribed regions of the chromosome, either in intergenic regions of polycistronic operons or in untranslated regions upstream or downstream of open reading frames. Their position in the genome seems to be different in different species (10, 14, 15, 20). Up to now the presence of ERIC sequences has been demonstrated only in gram-negative bacteria (14). van Belkum et al. (24) used this fingerprinting technique to type several methicillin-resistant Staphylococcus aureus strains.

Here we report that ERIC sequences can be used to establish clonal relationships between different strains of M. tuberculosis, even within clusters of strains showing identical IS6110 fingerprints.

MATERIALS AND METHODS

We analyzed 71 strains of M. tuberculosis isolated from different specimens from human immunodeficiency virus (HIV)-positive and HIV-negative patients sheltered in two Sardinian Hospitals. The samples were collected from 1992 to September 1996. Of these, 47 were collected from patients in the infectious diseases ward of Sassari University Hospital (northern Sardinia), and 15 of these were described previously (19), while 19 strains were collected from patients in the infectious diseases ward of Cagliari Hospital (southern Sardinia), and 15 of these were also described previously (19). Five strains were isolated in Pakistan and were described previously (19). The M. tuberculosis H37Rv strain, purchased from the American Type Culture Collection, was used as the standard strain. All strains used in this study were identified as M. tuberculosis by a DNA-RNA hybridization method (Gen Probe, Inc., San Diego, Calif.) and the method of Telenti et al. (22). We tested using the method of Fadda et al. (4) the drug susceptibilities of all M. tuberculosis strains isolated (data not shown).

Mycobacterial strains were grown in 10 ml of 7H9 medium (to which albumin, dextrose, and catalase were added), and genomic DNA was extracted and analyzed as described previously (19). For IS6110 fingerprinting the DNA was cut with the restriction endonuclease PvuII, subjected to electrophoresis in a 0.8% agarose gel, blotted onto nylon membranes (21), and hybridized with the plasmid pBK831, containing the 0.45-kb BamHI-SalI fragment of IS6110 (2), previously labelled by using an enhanced chemiluminescence gene labelling kit (5) (Amersham International, Amersham, United Kingdom).

PCR was performed with the primers ERIC1R (5′-ATGTAAGCTCCTGGGGATTCAC) and ERIC2 (5′-AAGTAAGTGACTGGGGTGAGCG) (ERIC-PCR) at a concentration of 1.0 μM as described previously (14). Amplification reactions were performed in a volume of 50 μl with final amounts of 1 U of Taq polymerase, 20 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl₂, and 200 μM deoxynucleoside triphosphate (Gibco, BRL, Life Technology, Paisley, United Kingdom). The reaction mixtures were overlaid with one drop of paraffin oil and were then incubated for 2 min at 94°C, followed by 35 cycles of 94°C for 45 s, 52°C for 1 min, and 70°C for 10 min and a final extension at 70°C for 20 min as described previously (14). The amplification products were visualized after electrophoresis at 90 V for 90 min in a 1.8% Methaphore agarose gel (FMC, Bioproducts, Rockland, Maine), and the gel was stained with ethidium bromide.

PCR-GTG was performed as described previously (19). Briefly, after lysis of the mycobacteria, primers IS2A and GTG1 were used to amplify chromosomal DNA. The amplification products were then visualized after electrophoresis on an agarose gel.

Ribotyping by PCR was performed with two primers complementary to conserved regions near the 3′ end of the 16S and the 5′ end of the 23S rRNA gene reported previously (8, 13). The sequences of the primers are 5′-TTGTACACACCGCCCGTCA for R1 and 5′-GAAACATCTAATACCT for R2. Amplifications were carried out in a final volume of 25 μl. Thirty cycles of amplification were performed, with each cycle consisting of 1 min of denaturation at 94°C, 1 min of annealing at 55°C, and 1 min at 72°C. The last cycle consisted of a 10-min extension at 72°C.

All DNA amplifications were performed in a DNA thermal cycler instrument (model TR3CM220; Hybaid; Omnigene, Teddington, United Kingdom).

The restriction fragment length polymorphisms (RFLPs) obtained by ERIC-PCR and IS6110 fingerprinting were scanned with Screen Machine II software (Fast Multimedia AG, Munich, Germany) and were evaluated with Image Master software (Pharmacia Biotech, Uppsala, Sweden). The similarity rate value (S_AB) among the strains was calculated, and the dendrograms were generated by the unweighted pair group method with arithmetic averages (18).

RESULTS

To show the reproducibility of the ERIC-PCR method, we performed three different experiments, each of which was performed on 3 different days. Figure 1 shows a series of dendrograms constructed on the basis of the patterns and the genetic relationships of five strains (strains H37Rv, SS81, SS82, SS83, and SS84) obtained with the ERIC primers. The DNA extraction, DNA amplification, and electrophoresis gel run were performed on 3 different days for each experiment (Fig. 1, experiments 1 to 3).

FIG. 1.

Results of three different experiments (experiments 1, 2, and 3) represented by three dendrograms associated with the RLFP profiles of five different strains (strains H37Rv, SS81, SS82, SS83, and SS84) obtained by ERIC-PCR. DNA extraction, DNA amplification, and the electrophoresis gel run were performed separately on 3 different days. Increasing similarity resulted in S_AB values ranging from 0 to 1.0.

Figure 2 is a dendrogram constructed on the basis of the different patterns for the 71 M. tuberculosis strains obtained with the ERIC oligonucleotides. The dendrogram shows the genetic relationships among the different strains. The amplified bands varied in size from 140 bp to 2 kb. The 47 strains isolated in Sassari (northern Sardinia) gave 36 different patterns by ERIC-PCR, whereas the 19 strains from southern Sardinia (Cagliari) gave 18 patterns (Table 1). In total, the strains were divided into 59 different family clusters, indicating a high level of polymorphism among these strains (Table 1). The IS6110 copy numbers of the M. tuberculosis Sardinian strains ranged from 5 to 12, with an average of 9 copies per strain. The strains isolated in northern Sardinia produced 29 different patterns by IS6110 fingerprinting, while the strains isolated in Cagliari gave 17 patterns (Table 1). PCR-GTG fingerprinting grouped the 47 strains from Sassari into 36 distinct patterns, whereas the 19 strains isolated in Cagliari were grouped into 17 family clusters (Table 1). All three methods differentiated the five strains isolated in Peshawar, Pakistan, into five different cluster types (Table 1).

FIG. 2.

Dendrogram with RFLP profiles illustrating relationships among the 71 M. tuberculosis strains analyzed by ERIC-PCR. Increasing similarity resulted in S_AB values ranging from 0 to 1.0. The origin of the specimen is also indicated (SS, Sassari; CA, Cagliari; PAK, Peshawar, Pakistan).

TABLE 1.

Number of patterns obtained by IS6110 fingerprinting, GTG-PCR, and ERIC-PCR methods with the 71 M. tuberculosis strains analyzed

Geographic source of isolatesa	No. of isolates	No. of patterns obtained by the following method:
		IS6110 fingerprinting	PCR-GTG	ERIC-PCR
SS	47	29	36	36
CA	19	17	17	18
PAK	5	5	5	5
Total	71	51	58	59

^aSS, Sassari; CA, Cagliari; PAK, Peshawar, Pakistan. 

Different strains were isolated from the same patient during a 2-year period; for instance, strains SS21 to SS25 were isolated from different clinical specimens from an HIV-positive patient. These strains showed the same pattern by both the PCR-GTG method and IS6110 fingerprinting (data not shown). Strains CA11 and CA17 were isolated from two different patients who lived in the same house (similar patterns were also obtained by IS6110 fingerprinting and PCR-GTG).

The ERIC-PCR was also able to divide different families obtained by IS6110 fingerprinting into subclusters; for instance, strains CA4, CA5, and CA6 were grouped together by IS6110 fingerprinting, but ERIC-PCR further differentiated these strains into three different subclusters (Fig. 2).

Figure 3a shows an agarose gel with a representative example of DNA from different strains of M. tuberculosis amplified with the ERIC oligonucleotides. Lanes A to C contain DNAs from different strains of M. tuberculosis (strains SS70, SS71, and SS73, respectively) isolated from the same patient during a 1-year period. These strains were grouped into the same pattern family by IS6110 fingerprinting. The SS70 strain gave a pattern of amplification different from those of the other two strains which were isolated successively. To confirm this hypothesis we amplified the DNAs from the same strains using the PCR-GTG method described previously (19). The result was the same: the pattern obtained for the first strain of M. tuberculosis was different from those for the other two strains (Fig. 3b, lanes A to C, respectively). To have further evidence of the difference between these strains, we amplified their chromosomal DNAs with two primers complementary to the 3′ end of the 16S rRNA gene and the 5′ end of the 23S rrn gene (8, 13). As shown in Fig. 3c, the first strain (SS70; lane A) gave a band with a different molecular size compared to the sizes of the bands from the other M. tuberculosis strains tested. It is interesting that strain SS68 (Fig. 3c, lane D) also produced two bands of 650 and 850 bp. The IS6110 fingerprinting patterns for strains SS70, SS71, and SS73 strains were the same (Fig. 3d). All three strains were susceptible to the different antibiotics tested.

FIG. 3.

(a) Agarose gel electrophoresis (1.8% Methaphore agarose) of amplified DNA from clinical isolates obtained by the ERIC-PCR method. Lanes: A, SS70; B, SS71; C, SS73; D, SS68; E, SS62; F, CA4; G, SS41; MW, bacteriophage λ VI marker (mixture of pBR328 DNA cleaved with BglI and pBR328 DNA cleaved with HinfI; Boehringer Mannheim). (b) Agarose gel electrophoresis (1.8% Metaphore agarose) of amplified DNA from clinical isolates obtained by the PCR-GTG method. Lanes are as described for panel a. (c) Agarose gel electrophoresis (1.8% Metaphore agarose) of amplified DNA from clinical isolates obtained by the PCR-ribotyping method. Lanes are as described for panel a except for lane H, which contains the bacteriophage λ VI marker. (d) Southern blot of chromosomal DNAs from clinical isolates (0.45-kb BamHI-SalI fragment of IS6110 and bacteriophage λ DNA were used as probes). Lanes are as described for panel a except for lane MW, which contains a bacteriophage λ-HindIII marker.

DISCUSSION

A number of different repeated sequences have been used to characterize the genome of M. tuberculosis. Insertion sequences such as IS1081 and IS6110 (1–3, 7, 17, 25–27, 30, 31), as well as direct repeats, major polymorphic repeats, and the GTG cluster, have been of great help in providing an understanding of the clonal relationship among different strains (6, 9, 16, 19, 23, 27, 28, 29). Among these sequences, IS6110 has been used worldwide to classify strains (7, 25). PCR techniques such as RAPD-PCR (9) and GTG-PCR (19) have shortened the time for the detection and classification of M. tuberculosis strains. Furthermore, these methods increase the size of the genome being analyzed in the DNA fingerprint, and the results are a better characterization of the strains tested. A recently reported study used GTG domains as a probe to subcluster the families of patterns obtained by IS6110 fingerprinting (28), and the results are in agreement with those of our previously published study (19), in which we used IS6110 and GTG as molecular targets in a PCR-based method to fingerprint M. tuberculosis strains. Here we report the presence of other possible targets that can be used to study the transmission of tubercular infections. The ERIC sequences have been used to characterize different gram-negative bacterial strains (10, 14, 15, 20). The method has been used to type S. aureus strains (24), but the annealing temperature used was 25°C (against the 52°C used to type gram-negative strains and used in this study), and the investigators used these sequences in a low-stringency amplification. ERIC sequences are reported to be stable in the genome (14). These domains are highly polymorphic, possibly as a result of recombination events between the 126-bp repeats (10, 14, 15).

In this study we used ERIC markers in order to amplify M. tuberculosis chromosomal DNA; we also compared the level of differentiation of this method with those of IS6110 and PCR-GTG fingerprinting. The results obtained by ERIC-PCR were reproducible and comparable to those obtained by PCR-GTG fingerprinting (71 strains grouped into 59 patterns compared with the 58 patterns obtained for the same strains by PCR-GTG), whereas IS6110 fingerprinting was less discriminatory (71 strains were divided into 51 family clusters).

ERIC-PCR was able to differentiate strains of M. tuberculosis that infected the same patient (SS70, SS71, and SS73), whereas IS6110 fingerprinting was not able to discriminate among these strains. These strains were also confirmed to be different by PCR-GTG and PCR-ribotyping methods (8, 13, 19). This example suggests how valuable the use of different methods of following M. tuberculosis infections can be. The results obtained indicate that the ERIC marker could be used as an independent marker, since the relationship between the profiles obtained by ERIC-PCR and those obtained by IS6110 fingerprinting were not always evident. In our experience the best way of differentiating M. tuberculosis strains is to use different markers such as IS6110, GTG, and ERIC sequences to increase the accuracy of epidemiological studies and the correlation between molecular evidence and patient interviews.

A closer relationship among the strains isolated from HIV-positive patients than among those isolated from non-HIV-infected patients was not shown in this study. There is no evidence of any correlation between HIV infection and particular strains of M. tuberculosis, indicating that the disease is due to a possible reactivation of the mycobacteria instead of a new infection.

Finally, we found different sizes of 16S-23S intergenic spacers in two M. tuberculosis strains (strains SS70 and SS68). One of them (strain SS70) produced a band of increased size (about 850 bp), while the other strain (strain SS68) produced two bands, one similar to the normal product of the other M. tuberculosis strains (about 650 bp) and a second band of the same size as the SS70 product (850 bps). This result may suggest the presence of different intergenic spacers in the rrn gene of strain SS70 (which may be due to the presence of a tRNA), whereas the second strain may have two different rrn operons, as reported previously (8).