Transfer of a Mycobacterium tuberculosis Genotyping Method, Spoligotyping, from a Reverse Line-Blot Hybridization, Membrane-Based Assay to the Luminex Multianalyte Profiling System

Abstract

Spoligotyping using Luminex technology was shown to be a highly reproducible method suitable for high-throughput analysis. Spoligotyping of 48 isolates using the traditional membrane-based assay and the Luminex assay yielded concordant results for all isolates. The Luminex platform provides greater flexibility and cost effectiveness than the membrane-based assay.

Spacer oligonucleotide typing (spoligotyping) is a commonly used, amplification-based method for genotyping Mycobacterium tuberculosis isolates (3). It is based on the polymorphisms found in the direct repeat (DR) locus that is present in all M. tuberculosis complex isolates. The DR locus contains multiple 36-bp DRs separated by 30- to 40-bp unique spacer sequences (5). The traditional spoligotyping method detects the presence or absence of 43 different spacer sequences by hybridizing labeled amplicons of the DR locus to oligonucleotide probes for each of the spacers arrayed on a membrane (a reverse line blot hybridization). Because the assay is inexpensive, quick, and robust, it is often used as a first-line genotyping method.

Although the assay is quite functional in its original form, a new technology is available as an alternative to reverse line blot hybridizations. The Luminex (Austin, Tex.) system can rapidly detect multiple DNA sequences. The system uses microspheres (beads) that contain two fluorochromes; a microsphere is given a distinct spectral address by varying the ratio of the concentration of the fluorochromes. Oligonucleotide capture probes are covalently attached to 1 of 100 available microsphere sets. During analysis, individual microspheres are interrogated by two lasers. The first laser excites the fluorochromes within the microsphere and allows identification of the microsphere set. The second laser excites a reporter fluorochrome bound to the hybridized PCR product and allows quantification of the PCR product. A sample can be analyzed in less than 15 seconds. Here, we report the successful transfer of the spoligotyping assay to the Luminex platform.

The 43 spacer oligonucleotide probes used in this study were the same as those used in the traditional spoligotyping assay (3). The probes were synthesized with a 5′-terminal amino group with a six-carbon spacer to allow covalent attachment to membranes or microspheres. Spoligotyping membranes were prepared as previously described (3). To couple the probes to the microspheres (Luminex Corp.), 200 pmol of oligonucleotide, 2.5 × 10⁶ microspheres, and 25 μg of freshly purchased N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide (Pierce Chemical, Rockford, Ill.) were combined in 25 μl of 100 mM 2-(N-morpholino)ethanesulfonic acid (MES), pH 4.5 (Sigma, St. Louis, Mo.); the reaction mixtures were incubated at room temperature in the dark for 30 min. The N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide addition and subsequent incubation were repeated once. After coupling, the microspheres were washed with 0.5 ml of 0.02% Tween 20 followed by 0.5 ml of 0.1% sodium dodecyl sulfate. The prepared microspheres were suspended in 50 μl of Tris-EDTA, pH 8.0, and stored at 4°C in the dark. A mixture of all 43 microspheres was prepared by combining equal volumes of each. The probes and their corresponding microsphere sets are listed in Table 1. The spacer sequences were amplified as originally described for spoligotyping (3). Briefly, primers that hybridize to the 36-bp DR were used to amplify the DR region, simultaneously amplifying all of the spacers present. The average size of the amplicons produced was 75 bp. One primer was biotinylated at the 5′ terminus for subsequent detection using a streptavidin conjugate.

TABLE 1.

Hybridization patterns obtained using the Luminex assay

Spacer no.-bead no.	Negative	Isolate 1a			Isolate 2a			Isolate 3a
	Median (rfu)b	Median (rfu)b	Ratioc	±d	Median (rfu)b	Ratioc	±d	Median (rfu)b	Ratioc	±d
sp01-011	61 ± 2	1,588 ± 50	26.0	+	54 ± 9	0.9	−	2,198 ± 57	36.0	+
sp02-017	46 ± 7	1,383 ± 28	30.1	+	42 ± 8	0.9	−	2,044 ± 123	44.4	+
sp03-018	53 ± 5	55 ± 3	1.0	−	51 ± 8	1.0	−	3,621 ± 70	68.3	+
sp04-019	46 ± 1	2,013 ± 73	43.8	+	48 ± 2	1.0	−	2,591 ± 43	56.3	+
sp05-020	52 ± 3	1,930 ± 43	37.1	+	55 ± 2	1.1	−	2,643 ± 74	50.8	+
sp06-021	63 ± 16	2,623 ± 61	41.6	+	61 ± 10	1.0	−	3,750 ± 140	59.5	+
sp07-024	48 ± 8	1,619 ± 40	33.7	+	43 ± 9	0.9	−	2,160 ± 66	45.0	+
sp08-025	52 ± 4	1,939 ± 27	37.3	+	47 ± 4	0.9	−	2,473 ± 70	47.6	+
sp09-026	55 ± 8	57 ± 5	1.0	−	48 ± 7	0.9	−	2,701 ± 54	49.1	+
sp10-027	54 ± 5	1,505 ± 29	27.9	+	56 ± 3	1.0	−	1,439 ± 14	26.6	+
sp11-028	60 ± 0	1,838 ± 57	30.6	+	54 ± 6	0.9	−	2,584 ± 49	43.1	+
sp12-029	55 ± 2	1,784 ± 49	32.4	+	46 ± 6	0.8	−	2,342 ± 66	42.6	+
sp13-032	51 ± 6	2,840 ± 98	55.7	+	53 ± 2	1.0	−	4,066 ± 51	79.7	+
sp14-033	56 ± 2	909 ± 6	16.2	+	48 ± 9	0.9	−	815 ± 18	14.6	+
sp15-034	60 ± 8	1,068 ± 24	17.8	+	52 ± 3	0.9	−	1,326 ± 54	22.1	+
sp16-035	59 ± 8	49 ± 7	0.8	−	52 ± 2	0.9	−	2,858 ± 136	48.4	+
sp17-036	63 ± 5	1,551 ± 11	24.6	+	52 ± 7	0.8	−	1,740 ± 61	27.6	+
sp18-037	58 ± 0	1,590 ± 49	27.4	+	48 ± 1	0.8	−	2,301 ± 37	39.7	+
sp19-038	53 ± 6	2,146 ± 15	40.5	+	53 ± 8	1.0	−	2,636 ± 65	49.7	+
sp20-039	58 ± 12	2,011 ± 55	34.7	+	52 ± 8	0.9	−	2,738 ± 98	47.2	+
sp21-040	56 ± 4	1,881 ± 71	33.6	+	54 ± 9	1.0	−	54 ± 5	1.0	−
sp22-041	56 ± 13	2,039 ± 11	36.4	+	57 ± 9	1.0	−	55 ± 7	1.0	−
sp23-042	55 ± 8	1,368 ± 28	24.9	+	52 ± 7	0.9	−	54 ± 9	1.0	−
sp24-046	43 ± 4	1,488 ± 29	34.6	+	43 ± 4	1.0	−	49 ± 10	1.1	−
sp25-047	56 ± 11	2,960 ± 84	52.9	+	59 ± 5	1.1	−	3,933 ± 73	70.2	+
sp26-051	71 ± 2	2,101 ± 15	29.6	+	61 ± 9	0.9	−	3,259 ± 120	45.9	+
sp27-052	57 ± 8	1,564 ± 40	27.4	+	52 ± 2	0.9	−	2,229 ± 55	39.1	+
sp28-053	60 ± 9	1,696 ± 68	28.3	+	56 ± 2	0.9	−	2,433 ± 91	40.6	+
sp29-054	63 ± 10	2,504 ± 48	39.7	+	58 ± 4	0.9	−	3,670 ± 49	58.3	+
sp30-055	57 ± 0	2,693 ± 64	47.2	+	59 ± 4	1.0	−	3,747 ± 71	65.7	+
sp31-056	64 ± 16	1,989 ± 21	31.1	+	55 ± 5	0.9	−	2,829 ± 52	44.2	+
sp32-061	64 ± 9	1,617 ± 36	25.3	+	61 ± 3	1.0	−	2,287 ± 65	35.7	+
sp33-062	64 ± 5	1,203 ± 2	18.8	+	63 ± 2	1.0	−	65 ± 6	1.0	−
sp34-063	60 ± 0	2,261 ± 52	37.7	+	58 ± 2	1.0	−	58 ± 5	1.0	−
sp35-064	58 ± 6	2,193 ± 33	37.8	+	3,332 ± 50	57.4	+	60 ± 11	1.0	−
sp36-065	53 ± 2	1,630 ± 36	30.8	+	2,851 ± 83	53.8	+	84 ± 5	1.6	−
sp37-066	62 ± 1	1,978 ± 37	31.9	+	3,534 ± 95	57.0	+	2,879 ± 101	46.4	+
sp38-072	81 ± 11	1,941 ± 30	24.0	+	2,953 ± 116	36.5	+	2,685 ± 151	33.1	+
sp39-073	78 ± 10	69 ± 9	0.9	−	1,532 ± 53	19.6	+	1,291 ± 37	16.6	+
sp40-074	78 ± 15	71 ± 11	0.9	−	2,969 ± 81	38.1	+	2,836 ± 82	36.4	+
sp41-075	79 ± 6	62 ± 3	0.8	−	3,942 ± 76	49.9	+	3,282 ± 119	41.5	+
sp42-076	72 ± 4	66 ± 6	0.9	−	1,540 ± 14	21.4	+	1,448 ± 58	20.1	+
sp43-077	75 ± 8	65 ± 4	0.9	−	2,009 ± 67	26.8	+	1,865 ± 68	24.9	+

^aThe isolates in this table correspond to the isolates in Fig. 1.

^bThe average median (rfu) ± the standard deviation (n = 3).

^cThe ratio is the average median (rfu) divided by the negative average median (rfu).

^dA + indicates that the isolate contains the spacer, and a − indicates that the isolate does not contain the spacer, as determined by the results shown in Fig. 1.

For hybridization, the microsphere mix was diluted 8.7-fold by the addition of 1.5× hybridization buffer (1.5 M tetramethylammonium chloride [Sigma], 75 mM Tris, pH 8.0, 6 mM EDTA, 0.15% Sarkosyl) to a final concentration of approximately 150 microspheres of each set/μl. PCR product (17 μl) and diluted microsphere mix (33 μl) were combined in a Thermowell 96-well plate (VWR International, West Chester, Pa.). The reaction mixtures were incubated for 10 min at 94°C, followed by incubation for 30 min at 52°C, in a GeneAmp 9700 PCR System (Perkin-Elmer, Foster City, Calif.). The plate was centrifuged at 2,250 × g for 3 min, the supernatant was removed by pipette, and the microspheres were resuspended with 75 μl of detection buffer (R-phycoerythrin-conjugated streptavidin [Molecular Probes, Eugene, Oreg.] diluted to 4 μg/ml with 1× hybridization buffer). Following a 5-min incubation at 52°C, the samples were analyzed in the Luminex 100, version 1.7; a minimum of 100 events/microsphere set were analyzed.

Because the basic chemistry involved in conjugating the probes to membranes or to microspheres is very similar, changing to the new substrate was quite simple. Although the manufacturer has shown that larger spacers between the amino group and the complementary sequence on the probe will increase the median signal intensity in Luminex assays, we found that the probes used to prepare membranes, which have only a six-carbon spacer, were sufficient for this assay on the Luminex platform. Spoligotyping membranes are prepared with various probe concentrations to equalize the signal intensities; varying the probe concentrations when preparing microspheres did not appreciably change the signal intensity; therefore, the same concentration was used for all probes (data not shown).

To test the assay on the Luminex platform, three isolates with spoligotypes commonly found for M. tuberculosis complex strains were assayed using the membrane-based assay and were assayed in triplicate using the Luminex method. The average median number of relative fluorescence units (rfu) obtained for each spacer in each isolate is listed in Table 1. Each median was then divided by the median obtained for a background sample (a sample processed in an identical manner but containing water in place of template DNA). When the ratios were compared with the membrane results (Fig. 1), the spacers that were determined to be positive using the membrane platform had ratios of 14.7 to 79.7 (average of 38.4) with the Luminex platform, while those that were negative had ratios of only 0.8 to 1.6 (average of 1.0). A cutoff value of five times the background median was established to score a spacer as positive or negative. The small standard deviations in the median (the average coefficient of variation was 5.8%) indicate that the assay on the Luminex platform is highly reproducible.

FIG. 1.

Hybridization patterns for three isolates obtained using the original, membrane-based assay. Lane 1, isolate 1, M. bovis; lane 2, isolate 2, M. tuberculosis Beijing family; and lane 3, isolate 3, M. tuberculosis Latino-American and Mediterranean family (1).

Next, 48 isolates that had previously been spoligotyped on membranes were typed on the Luminex platform in a blinded fashion. The spoligotypes of all isolates were in perfect agreement between the two platforms. For each spacer, isolates positive for the spacer (as determined by using the membrane-based assay) were separated from those that were negative for the spacer, and the average median divided by the background median was determined for each group. As shown in Fig. 2, there is a clear distinction between positive and negative spacers. The results also suggest that using a cutoff of greater than five times the background for positive spacers is appropriate.

FIG. 2.

Comparison of Luminex-generated spoligotype results with traditional, membrane-based results for 48 clinical isolates. Using the results from the membrane-based assay, spacers in each isolate were identified as positive or negative in each strain. For each spacer, the average median divided by the background median was calculated for positive spacers (•) and for negative spacers (▪).

Spoligotyping on the Luminex platform offers many benefits. Both the turnaround time and the labor involved are significantly decreased, and it is less technically demanding. The Luminex also offers more flexibility because it can be used to assay from 1 to 96 isolates without increasing the labor time or cost per isolate; membranes are most efficiently used for assaying 40 isolates/run. The reagent and supply costs for both platforms are approximately the same. Perhaps the greatest benefit is the increase in reproducibility. The spoligotyping assay is robust and reproducible (>90%) (2, 4), but reproducibility using the membrane-based assay can be affected by the subjective determination of the hybridization signal, which becomes more difficult with repeated use of the membrane. Errors can also occur in recording the 43-digit result (4). The Luminex platform produces numerical data, and the final spoligotype can be generated with a simple Excel spreadsheet without manual data entry. Here, we have shown that the spoligotypes generated on the two platforms are identical, allowing different laboratories to directly compare data regardless of the platform used. The Luminex system is an attractive alternative for laboratories that perform spoligotyping on a high-throughput scale or for those that frequently require a rapid turnaround time for only a few isolates per run.