OUP user menu

Optimization of variable number tandem repeat typing set for differentiating Mycobacterium tuberculosis strains in the Beijing family

Kai Man Kam, Chi Wai Yip, Lai Wa Tse, Ka Ling Leung, Kin Lai Wong, Wai Mei Ko, Wai Sum Wong
DOI: http://dx.doi.org/10.1111/j.1574-6968.2006.00126.x 258-265 First published online: 1 March 2006


Variable number tandem repeat (VNTR) typing of Mycobacterium tuberculosis, as presently used, has often failed to differentiate clonal strains especially in areas where the Beijing family genotype is predominant. We have evaluated mycobacterial interspersed repetitive units (MIRUs), Queen's University of Belfast (QUB) loci and exact tandem repeat (ETR) loci individually for their abilities to differentiate the Beijing and non-Beijing genotype families of M. tuberculosis. The best locus was QUB 3232 followed by QUB 11b. Available MIRU, QUB and ETR loci were then consecutively arranged in declining order of discriminatory power, and combined to form an optimal set of 12-loci VNTR typing (MIRU 10, 26, 39, 40; QUB 11a, 11b, 15, 18, 26, 3232, 3336; ETR A) that possessed comparable discriminatory ability to IS6110 restriction fragment length polymorphism. This optimized 12-loci VNTR typing set can become an important tool for tracking M. tuberculosis of the Beijing family.

  • tuberculosis
  • Beijing family
  • multidrug resistance
  • VNTR
  • IS6110
  • RFLP


Morbidity and mortality associated with Mycobacterium tuberculosis remains a major public health threat and an estimated one-third of the world population has latent infections (Dye et al., 1999). During the last decade, the Beijing genotype, a specific family of M. tuberculosis, has been found to be spreading in various geographic locations (Niemann et al., 1997; Kruuner et al., 2001; Bifani et al., 2002). With the possible associations of the Beijing genotype with drug resistance (Bifani et al., 1999; Drobniewski et al., 2002) and high adaptability to the host intracellular environment (Rad et al., 2003), epidemiology of this special M. tuberculosis genotype has become an increasingly important issue in global tuberculosis (TB) control.

Although restriction fragment length polymorphism (RFLP) analysis based on the insertion sequence IS6110 has been considered as the ‘gold standard’ for the molecular typing of M. tuberculosis, the procedures involved, however, are time consuming and technically demanding. PCR-based typing methods using variable number tandem repeats (VNTRs) on the mycobacterial interspersed repetitive units (MIRUs) in 12 human minisatellite-like regions of the M. tuberculosis genome has been developed and has produced some initial promising results (Kremer et al., 1999; Supply et al., 2000, 2001). In 2002, following adoption of an agreed International Standard Protocol, a consensus was reached in the European Union Concerted Action meeting ‘New genetic markers and techniques for the epidemiology and control of tuberculosis’ (Cascais, Portugal, 2002) that MIRU-based typing methods would supersede IS6110 in the near future following adoption of an agreed International Standard Protocol Studies.

There is an urgent need to develop PCR-based typing method in an easily transportable format for rapid and accurate typing to delineate TB transmission, including drug resistant M. tuberculosis for global TB control. The aim of the present study was to investigate the differentiation ability that can be provided by exact tandem repeat (ETR) (Frothingham & Meeker-O'Connell, 1998) and those loci previously identified at the Queen's University of Belfast (QUB) on the Beijing genotype family strains, as well as the differentiation power of the various combinations of MIRU, ETR and QUB loci. This development of an optimized set of VNTR loci with a resolution comparable to IS6110 RFLP will help assess the feasibility of using these TB epidemiological markers in areas where Beijing family genotype strains are prevalent.

Materials and methods

Bacterial strains

In Hong Kong, all positive Mycobacterium tuberculosis isolates from patients attending public hospitals and TB and Chest clinics are routinely sent to the TB Reference Laboratory for mycobacterial identification and drug susceptibilities testing. Three hundred and eighty non-duplicate drug-resistant strains of M. tuberculosis, for the 4-year period from January 2000 to December 2003, isolated from clinical specimens processed in the Hong Kong TB Reference Laboratory were included in this study. These represented approximately 80% of culture-positive cases in the whole of Hong Kong in the study period. These isolates included all (110 strains) multidrug-resistant (MDR) M. tuberculosis and 270 randomly selected non-MDR resistant strains (out of a total of 814) obtained from clinical specimens during the study period. The strains were recovered from the −70°C stock and grown on Löwenstein–Jensen medium for 3 weeks at 37°C. Mycobacteria identification and drug susceptibilities testing of these strains to first line anti-TB drugs were done as previously described (Kam & Yip, 2001). MDR was defined as resistance to at least isoniazid and rifampicin.

Molecular typing methods

IS6110 RFLP and spoligotyping of the M. tuberculosis strains were performed according to standardized protocols (Niemann et al., 1997; Mazars et al., 2001). Spoligotyping detected the presence of 43 spacer sequences in the direct repeat (DR) region by reversed-line blot hybridization (Kamerbeek et al., 1997). The Beijing family genotype was identified by its unique spoligotyping pattern, in which spacers 35–43 were typically present (van Soolingen et al., 1995).

Variable number tandem repeat typing was performed according to the method as previously described (Frothingham & Meeker O'Connell, 1998). In addition to the 12 MIRUs loci, five exact tandem repeat (ETR – A, B, C, D, E) loci (Frothingham & Meeker-O'Connell, 1998) and twelve loci previously identified at the Queen's University of Belfast (QUB-11a, 11b, 15, 18, 23, 26, 1281, 1451, 1895, 3232, 3336, 4156) (Skuce et al., 2002) were also included. PCRs (HotStarTaq PCR, Qiagen, Hilden, Germany) were run in DNA thermal cyclers (model 9700, Perkin Elmer, Foster City, CA) under the following conditions: (1) ETR loci; 95°C for 15 min, 35 cycles of 94°C for 30 s, 60°C for 1 min and 72°C for 2 min, with final extension at 72°C for 10 min; (2) QUB loci; 95°C for 15 min, 33 cycles of 94°C for 30 s, 59°C for 1 min and 72°C for 1 min, with final extension at 72°C for 8 min. PCR products were analyzed on a 2% NuSieve agarose gel (Applied Biosystems, Foster City, CA). An allele-naming table (Table 1) used in this study was prepared according to the method of Frothingham & Meeker-O'Connell (1998).

View this table:
Table 1

Primer sequence and repeat unit size of the QUB and ETR loci

QUB/ETR locusPCR primer sequence (5′–3′)Predicted size+additional unitAmount in one PCR tube (pmol)
  • QUB, Queen's University of Belfast; ETR, exact tandem repeat.

Computer-assisted and statistical analysis

BioNumerics software (v3.0, Applied Maths, Austin, TX) was used to analyze molecular typing results. IS6110 RFLP patterns were analyzed as fingerprint types while VNTR types were analyzed as a character type. Similarities between VNTR types were calculated by the categorical coefficient in which all VNTR loci were weighted equally, and a dendrogram was constructed according to the unpaired group method using arithmetic averages (UPGMA). The level of discrimination of each typing method was calculated by using the Hunter–Gaston discriminatory index (HGI) as previously described (Hunter & Gaston, 1988).


Spoligotyping and Beijing family genotype

A total of 355 Mycobacterium tuberculosis strains were recovered in this study, including 102 MDR-TB (93%) and 253 non-MDR-TB (94%) strains. Spoligotyping, by examining the deletion of spacers 1–34, identified 243 M. tuberculosis isolates (68.5%) to be members of the Beijing family genotype. As for the 112 non-Beijing M. tuberculosis isolates, IS6110 RFLP study showed 90 isolates as having five or more IS6110 bands and 22 with less than five bands. Proportions of Beijing genotype amongst MDR and non-MDR isolates were found to be similar, with 71 (69.6%) of MDR-TB and 172 (70.0%) of non-MDR strains belonging to the Beijing genotype.

IS6110 RFLP typing

IS6110 RFLP typing of the 355 M. tuberculosis isolates resulted in 302 RFLP patterns (HGI 0.9986) (Table 2). Among the Beijing family isolates, RFLP differentiated 203 patterns (HGI 0.9979), while in the non-Beijing family group, it was able to differentiate 99 types (HGI 0.9961) with a breakdown of 83 types (HGI 0.9987) for isolates having 5 or more IS6110 bands and 16 types (HGI 0.9312) for those with less than five IS6110 bands (Table 2).

View this table:
Table 2

Comparison of discriminatory power of various genetic markers between Beijing and non-Beijing genotype

Non-Beijing genotype
Discriminatory parameterBeijing genotype (n=243)All strains (n=355)(All strains) (n=112)(≥Five bands) (n=90)(<Five bands) (n=22)
(A) QUB (no. of typable strains)212917417303
Total type pattern142826715224
No. of isolates having unique type (%)116 (54.7%)73 (80.2%)60 (81.1%)13 (76.5%)189 (62.4%)
No. of cluster2697235
No. of clustered isolates (%)96 (45.3%)18 (19.8%)14 (18.9%)4 (23.5%)114 (37.6%)
Maximum no. of isolates in a cluster1022210
Hunter–Gaston discriminatory index0.99060.99780.99740.98530.9951
(B) ETR (no. of typable strains)2341038419337
Total type pattern2041331454
No. of isolates having unique type (%)8 (3.4%)20 (19.4%)15 (17.9%)10 (52.6%)24 (7.1%)
No. of cluster122118430
No. of clustered isolates (%)226 (96.6%)83 (80.6%)69 (82.1%)9 (47.4%)313 (92.9%)
Maximum no. of isolates in a cluster17312123185
Hunter–Gaston discriminatory index0.44740.96540.95500.96490.6925
(C) VNTR (12 loci) (no. of typable strains)212897217301
Total type pattern175826814257
No. of isolates having unique type (%)153 (72.2%)75 (84.3%)64 (88.9%)11 (64.7%)228 (75.7%)
No. of cluster2274329
No. of clustered isolates (%)59 (27.8%)14 (15.7%)8 (11.1%)6 (35.3%)73 (24.3%)
Maximum no. of isolates in a cluster62226
Hunter–Gaston discriminatory index0.99750.99820.99840.97790.9986
(D) IS6110 RFLP (no. of typable strains)2431129022355
Total type pattern203998316302
No. of isolates having unique type (%)178 (73.3%)91 (81.3%)78 (88.6%)13 (54.2%)269 (75.8%)
No. of cluster2585333
No. of clustered isolates (%)65 (26.7%)21 (18.8%)10 (11.4%)11 (45.8%)86 (24.2%)
Maximum no. of isolates in a cluster56266
Hunter–Gaston discriminatory index0.99790.99610.99870.93120.9986
  • * Non-Beijing strains are further divided into isolates with five or more IS6110 bands and isolates with less than five bands.

  • QUB, Queen's University of Belfast; ETR, exact tandem repeat; VNTR, variable number tandem repeat; RFLP, restriction fragment length polymorphism.

VNTR typing

All M. tuberculosis isolates were subjected to analysis with 12 QUB and five ETR loci (Table 1). QUB differentiated a total of 224 types (HGI 0.9951) (Table 2) while ETR differentiated only 54 types (HGI 0.6925). Among the Beijing family isolates, QUB differentiated 142 types (HGI 0.9906), while in the non-Beijing group, it was able to differentiate 82 types (HGI 0.9978) with a breakdown of 67 types (HGI 0.9974) for isolates having 5 or more IS6110 bands and 15 types (HGI 0.9853) for those with less than five IS6110 bands. On the other hand, ETR typing can only differentiate 20 types (HGI 0.4474) among Beijing family isolates and 41 types (HGI 0.9654) among the non-Beijing family group (Table 2).

When individual QUB and ETR locus was compared, all these VNTR (QUB and ETR) loci appeared to perform better in differentiating non-Beijing (i.e. a higher HGI) than Beijing isolates (Table 3). Only 3 QUB loci (QUB 3232, QUB 11A, QUB 11B) had HGI of more than 0.5 for Beijing strains, indicating a poor discriminating power for most VNTR loci in this family. Unlike the situation with the Beijing strains, 11 of the 17 VNTR loci showed HGI more than 0.5 for the non-Beijing strains. In this study, the most effective locus was QUB 3232, which had an HGI of 0.8041 to Beijing genotype and 0.9164 to the non-Beijing group. When all VNTR loci (MIRU+QUB+ETR) were considered in use for the differentiation of Beijing strains, a high HGI value of 0.9976 was obtained, which is comparable to that obtained from IS6110 RFLP (0.9979). However, not all M. tuberculosis strains could be typed by all QUB and ETR loci. Despite repeated attempts, no PCR product was found for some VNTR loci when tested on certain bacterial strains (Table 3). QUB 11a, in particular, appeared to be the locus with the highest failure rate, where no PCR product was found in 31 out of 355 strains (8.7%), followed by QUB 11b with a failure rate of 18/355 (5.1%).

View this table:
Table 3

Comparison of Hunter–Gaston discriminative index of various VNTR loci for typing of Mycobacterium tuberculosis strains (Beijing and non-Beijing genotype)

Non-Beijing genotype
MarkersTypable strains (n=355)Beijing genotype(All strains)(≥five bands)(<five bands)
QUB 11a3240.51360.86970.86740.8480
QUB 11b3370.66900.87880.87690.8088
QUB 153550.13230.33890.35040.3043
QUB 183540.48810.86220.85650.6558
QUB 233540.01640.01800.02300.0000
QUB 263480.31440.75680.75250.7645
QUB 12813550.04040.08780.06790.1594
QUB 14513550.00820.62050.63510.4203
QUB 18953520.20570.47220.58510.2899
QUB 32323450.80410.91640.91710.8841
QUB 33363550.21390.84850.83050.9058
QUB 41563550.16730.70650.72360.5761
ETR A3370.18800.67690.64800.7544
ETR B3550.06440.54180.56170.4348
ETR C3550.05680.34170.30070.4855
ETR D3550.07230.49030.33930.7898
ETR E3550.15570.66620.70110.4674
All QUB3030.99060.99780.99740.9853
All ETRs3370.44740.96540.95500.9649
All MIRU3550.88270.99440.99320.9493
All Loci2980.99760.99850.99880.9853
IS6110 RFLP3550.99790.99830.99850.9312
  • * Non-Beijing strains are further divided into isolates with five or more IS6110 bands and isolates with less than five bands.

  • ETR D is also known as MIRU 4, and ETR E is also known as MIRU 31.

  • QUB, Queen's University of Belfast; ETR, exact tandem repeat; VNTR, variable number tandem repeat; RFLP, restriction fragment length polymorphism; MIRU, mycobacterial interspersed repetitive units.

By arranging the HGI values in descending order, a cumulative HGI table was constructed. Stepwise addition of loci was done, such that on each addition, it included the VNTR locus that gave the next highest HGI value on strains of the Beijing family (Table 4). When the top 12 loci (MIRU10, 23, 26, 39, 40; QUB 11a, 11b, 18, 26, 3232, 3336, 4146; ETR A) were selected for typing, the set produced a HGI value of 0.9974, which is almost equal to that obtained in IS6110 RFLP (0.9979). This optimized set of 12 loci differentiated a total of 257 types (HGI 0.9986) for all isolates (Table 2). Among the Beijing family isolates, it differentiated 175 types (HGI 0.9975), while in the non-Beijing family group, it was able to differentiate 82 types (HGI 0.9982) with a breakdown of 68 types (HGI 0.9984) for isolates having five or more IS6110 bands and 14 types (HGI 0.9779) for those with less than five IS6110 bands.

View this table:
Table 4

Optimization of a 12-loci VNTR set: increase in Hunter–Gaston index (HGI) on successive addition of each VNTR loci

HGI value
VNTR lociBeijing genotypeNon-Beijing genotypeAll strains
QUB 32320.80370.91680.8803
QUB 11B0.92640.98670.9602
QUB 180.96060.99540.9799
QUB 11A0.97870.99680.9889
MIRU 390.98830.99730.9938
MIRU 260.99200.99780.9958
QUB 260.99430.99800.9970
MIRU 400.99530.99820.9975
ETR A0.99620.99820.9979
QUB 33360.99680.99820.9983
MIRU 100.99710.99820.9984
QUB 150.99730.99820.9985
QUB 18950.99740.99850.9986
MIRU 230.99750.99850.9986
QUB 41560.99760.99850.9986
MIRU 310.99760.99850.9987
MIRU 20.99760.99850.9987
MIRU 40.99760.99850.9987
MIRU 160.99760.99850.9987
MIRU 200.99760.99850.9987
MIRU 240.99760.99850.9987
MIRU 270.99760.99850.9987
QUB 230.99760.99850.9987
QUB 12810.99760.99850.9987
QUB 14510.99760.99850.9987
ETR B0.99760.99850.9987
ETR C0.99760.99850.9987
  • ETR D, MIRU 4; ETR E, MIRU 31.

  • QUB, Queen's University of Belfast; ETR, exact tandem repeat; VNTR, variable number tandem repeat; MIRU, mycobacterial interspersed repetitive units.

On further analysis using this 12 VNTR loci set, a total of 22 clusters were found amongst the Beijing strains. The largest cluster was (3733–8535–81374) with six members, followed by (3733–5537–81383) four members. There were nine clusters consisted of three members while the remaining 11 clusters were each composed of two members only (Table 5).

View this table:
Table 5

Cluster differentiation using QUB, ETR and MIRU loci typing against IS6110 RFLP

Non-Beijing genotype
Discriminatory parameterBeijing genotypeAll strains(All strains)(≥five bands)(<five bands)
No. of cluster2697235
No. of cluster further differentiated by RFLP (%)18 (69.2%)5 (55.6%)4 (57.1%)1 (50%)24 (68.6%)
No. of cluster122118430
No. of cluster further differentiated by RFLP (%)11 (91.7%)21 (100%)17 (94.4%)2 (50.0%)29 (96.7%)
No. of cluster271815344
No. of cluster further differentiated by RFLP (%)25 (92.6%)15 (83.3%)12 (80.0%)2 (66.7%)39 (88.6%)
(D) VNTR (using optimized 12-loci set)
No. of cluster2274329
No. of cluster further differentiated by RFLP(%)14 (63.6%)3 (42.9%)2 (50.0%)1 (33.3%)18 (62.1%)
(E) IS6110 RFLP
No. of cluster2585333
No. of cluster further differentiated by MIRU (%)10 (40.0%)4 (50.0%)2 (40.0%)2 (66.7%)14 (42.4%)
No. of cluster further differentiated by QUB (%)10 (40.0%)4 (50.0%)2 (40.0%)2 (66.7%)14 (42.4%)
No. of cluster further differentiated by ETR (%)2 (8.0%)2 (25.0%)0 (0.0%)2 (66.7%)4 (12.1%)
No. of cluster further differentiated by VNTR (%) 1717 (68.0%)4 (50.0%)2 (40.0%)2 (66.7%)21 (63.6%)
  • QUB, Queen's University of Belfast; ETR, exact tandem repeat; VNTR, variable number tandem repeat; RFLP, restriction fragment length polymorphism; MIRU, mycobacterial interspersed repetitive units.

Comparison of VNTR and IS6110 RFLP typing

Considering all isolates, 18 of the 30 VNTR clusters could be further subdivided by IS6110 RFLP typing. However, most of these clusters (17/27) shared very similar RFLP patterns, with less than or equal to 3-bands difference (Fig. 1). On the other hand, 17 of 22 (77.3%) RFLP clusters in the Beijing family and 4/7 (57.1%) of non-Beijing RFLP cluster could be further differentiated by VNTR typing. Most of these dispute clusters possessed very similar VNTR patterns. Moreover, it can be seen that for the two large VNTR clusters ((3733–8535–81374) and (3733–5537–81383), all the members shared the same RFLP patterns. No VNTR cluster was found to be shared by both Beijing and non-Beijing genotypes. The HGI values obtained from using both 12-loci VNTR typing and IS6110 RFLP were higher for the non-MDR than the MDR strains in our collection. This is irrespective of whether they belonged to Beijing or non-Beijing family. A high HGI value (HGI=1) was observed in MDR groups with less than five IS6110 RFLP bands. This was probably due to insufficient number of strains (only five isolates) that were involved in this group.

Figure 1

IS6110 restriction fragment length polymorphism of isolates that shared the same variable number tandem repeat (12 loci) patterns amongst some Mycobacterium tuberculosis isolates of the Beijing family.


Addition of QUB and ETR loci substantially increased the discrimination power for the Beijing family of strains. In this study, when 355 drug resistant Mycobacterium tuberculosis isolates (68.5% Beijing) were analysed, a combination of 12 VNTR loci, i.e. the same number of loci used as MIRU typing in various studies (Kremer et al., 1999; Supply et al., 2000; Supply et al., 2001), namely MIRU 10, 26, 39, 40; QUB 11a, 11b, 15, 18, 26, 3232, 3336 and ETR A, produced a comparable level of resolution (HGI=0.9975) as IS6110 RFLP (HGI=0.9979). Apart from the Beijing genotype, this 12-loci VNTR set also appeared to possess good differentiation power for non-Beijing M. tuberculosis (HGI=0.9982) when compared with the 12 MIRU typing (HGI 0.9928) (data not shown). This opens the possibility of replacing the time consuming and technically demanding IS6110 RFLP typing with this more rapid, simple and more easily inter-laboratory comparable VNTR typing method for this closely related M. tuberculosis family in the study of M. tuberculosis molecular epidemiology, especially in areas with high prevalence of Beijing family genotype strains.

For the three sets of VNTR loci that have been tested, QUB loci provided the highest differentiation power for the Beijing family genotype. Previously, Skuce (2002) showed that QUB 11a, QUB 11b and QUB 26 contributed much in differentiating M. tuberculosis strains. In particular, QUB 3232 which was first described to have high discriminatory power in typing M. bovis (Roring et al., 2002), has an exceptionally high HGI values for both Beijing family (0.8041) and non-Beijing (0.9164) M. tuberculosis strains in this study. In a previous smaller scale study of 51 Beijing strains using four QUB, five ETR and five MIRU loci, QUB loci was demonstrated to have high resolving power against Beijing genotype (data not shown). At present, only limited studies on the QUB loci could be found in the literature (Fleche et al., 2002). Since these hold much promise in differentiating closely related members of the Beijing family strains, more studies on these potentially useful loci are urgently needed.

Apart from HGI, the high differentiation power of the optimized 12 loci VNTR typing set was shown by its comparable percentage clustering of isolates (27.8%) to that obtained using IS6110 RFLP (27.4%), as well as the finding that 17/22 (77.3%) IS6110 clusters could be further differentiated by the 12 loci VNTR typing set. However, it must be pointed out that higher differentiation power does not necessarily imply usefulness of such loci in epidemiological studies. IS6110 fingerprinting and VNTR typing are directed at different targets and should provide independent epidemiological information. Although a number of VNTR clusters appeared overlapping with clustering seen with the same or very similar IS6110 RFLP patterns, a considerable number of VNTR clusters showed members having different RFLP patterns. IS6110 RFLP typing had already been proven to have epidemiological relevance in various settings. In Hong Kong, it has been shown that only about 30% of the clustered cases had identifiable epidemiological links (Chan et al., 2003). This apparent difference in clustering results between IS6110 RFLP and VNTR typing means that further studies are required to elucidate the exact role of our 12-loci VNTR typing set in resolving different epidemiological situations (Supply et al., 2000; van Soolingen et al., 2001). Useful genetic loci presumably possess an evolutionary clock that is fast enough to produce variations but slow enough to show the property of clonal relatedness in outbreak investigations. The only way to confirm the usefulness of this 12-loci VNTR typing set is to correlate clustering data with those of clinical and epidemiological data. Further evaluation of these loci is definitely important in areas with high prevalence of Beijing genotype, especially in high incidence countries.

Amplification of tandem repeat loci may at times be difficult. Strand slippage and difficulties in optimizing PCR condition for QUB loci have been observed, in which no PCR product was found for QUB 3232. However, when the PCR reagent specified for amplification reactions involving complex genomic and low copy target (such as HotStarTaq, Qiagen) was used, PCR on QUB loci became possible for most of our isolates. Despite the fact that there may still be failures in VNTR loci amplification, this problem can hopefully be resolved with further advances in PCR reagent development.

When the fingerprints of MDR and non-MDR strains were compared either by RFLP or various VNTR loci, MDR strains appeared to be more clonal than the non-MDR strains (data not shown). This was not unexpected, as the absolute number of MDR strains and thus the number of clones, circulating in a community should be much less than those seen in non-MDR ones. Moreover, different sampling methods, e.g. all MDR strains chosen vs. only a random sample of non-MDR isolates selected, may also contribute to this phenomenon. Apart from this clonal phenomenon of the MDR strains amongst our isolates, there was no virtual difference in distribution of typing patterns between MDR and non-MDR strains. This gives us a reason for performing subsequent analyses regardless of the MDR status of strains.

One significant advantage of rapid typing has been the use in TB laboratories to detect cross-contaminations. This can serve to avoid serious negative impacts on patient management because of laboratory false-positive culture results (Caroline et al., 2004). In the ‘real-time’ high-throughput laboratory situation, a relatively simple, rapid and moderate discriminatory power method is most suitable for routine screening of positive M. tuberculosis isolates. In this respect, a 5-loci format (ETR A to E) has been shown to give good results (Gascoyne-Binzi et al., 2001). In our study, resolution by ETR typing alone was insufficient for the differentiation of Beijing family strains (HGI=0.4475). This is in contrast to the situation with non-Beijing genotype (HGI=0.9654) as shown in previous studies (Gascoyne-Binzi et al., 2001; Caroline et al., 2004). However, using the top 4 discriminatory QUB loci (QUB 3232, QUB 11b, QUB 18, QUB 11a), an HGI value of 0.9787 (Table 4) can be obtained. In geographic areas where Beijing genotype is highly prevalent (van Soolingen et al., 1995), laboratory cross contamination can very often be an important consideration because of the high volume workload in the TB laboratory.

In summary, we have developed a 12-VNTR loci set for typing Beijing family genotype strains. This needs to be further evaluated in the field, especially in population based studies with multidrug resistant strains.


The authors would like to acknowledge C.Y. Mok, S.M. Lam, T.K. Lam and all technical and research staff in the TB Reference Laboratory, Hong Kong, for their assistance in the preparation of bacterial strains as well as electrophoresis works. We also thank the Director of Health, Dr P.Y. Lam, for permission to publish this manuscript.


View Abstract