FEMS Microbiology Letters 98 (1992) 155-160 ~) 1992 Federation of European Microbiological Societies 0378-1097/92/$05.(10 Published by Elsevier
155
FEMSLE (15131
Pulsed field gel electrophoresis of representatives of Mycobacterium tuberculosis and Mycobacterium bovis BCG strains A n n e Varnerot ", Fran~oise Cl6ment ~, Marina Gheorghiu b and Vdronique Vincent-Ldvy-Frdbault " a Centre National de Rdfdrence pour les Mycobactdries, Unitd de la Tuberculose et des Mycobactdries, Paris, France, and b Laboratoire du BCG, Institut Pasteur, Paris, France
Received 6 July 1992 Revision received 1 August 1992 Accepted 17 August 1992
Key words: Pulsed-field gel electrophoresis; Mycobacterium tuberculosis; Mycobacterium boris BCG
1. SUMMARY Using field inversion gel electrophoresis (FIGE), different Mycobacterium tuberculosis strains, such as phage prototypes, exhibit different DNA restriction patterns which are easy to compare. Virulent and avirulent variants of M. tuberculosis H37, as well as daughter strains of M. boris BCG, display characteristic DNA profiles. BCG strains isolated from suppurative adenitis following vaccination of French patients showed patterns identical to the BCG Pasteur strain used for vaccination. These results demonstrate that FIGE of DNA restriction fragments generated by DraI represents a suitable technique for the analysis of mycobacteria at a ge-
Correspondence to." V. Vincent-L6vy-Fr6bault, Centre National de R6f6rence pour les Mycobact6ries, Unit6 de la Tuberculose et des Mycobact6ries, Paris, France.
nomic level. The DraI profiles allow the differentiation and precise identification of the BCG Pasteur, Glaxo, Russian and Japanese strains.
2. INTRODUCTION Pulsed field gel electrophoresis (PFGE) allows the separation of large DNA fragments. The technique, originally developed by Schwartz et al. [14] for the separation of yeast chromosomes, has been applied to the analysis of bacterial genomes. After digestion by low-frequency cleavage restriction endonucleases, bacterial chromosomes provide DNA patterns composed of a few, usually well-separated fragments. Apart from the electrophoresis apparatus, which needs to be especially devised for ensuring electric pulses, DNA has to be extracted from agarose-embedded bacteria, in order to prevent mechanical, non-specific chromosomal fragmentation. Due to this con-
156
straint, usual lysis procedures of mycobacteria, requiring inhibitors of cell wall synthesis and actively growing mycobacteria, could not be used. We have previously applied field inversion gel electrophoresis (FIGE), one of the P F G E techniques, to mycobacteria and developed gentle lysis procedures ensuring suitable yields of entire DNA from immobilized, non-dividing cells [1,17]. The method demonstrated that the two Mycobacterium at,iurn subspecies, subsp, paratuberculosis and subsp, sih~aticurn, presented an homogeneous genomic organization producing subspecies-characteristic profiles, whereas the other M. at'ium strains displayed diverse restriction patterns [17]. Recently, Hector et al. [7] used P F G E in an epidemiological investigation of nosocomial outbreaks due to M. fortuitum. Studying the characteristics of avirulent and virulent strains may be useful for a better knowledge of the still poorly understood virulence of tuberculosis. In the present study, we applied the FIGE technique to some representatives of tubercle bacilli in order to detect differences at the genomic level.
3. M A T E R I A L S AND M E T H O D S
Bacterial strains M. tuberculosis strains used were: H37Rv, type strain of the species; H37Ra, the avirulent variant of the former; strains 965, 1025, 1324, 1381 and H37HS, the phage host prototypes of lysotypes A, Ax, I2, B and C, respectively. The designations of phage types correspond to phage susceptibility previously described [2,13]. Lysotype A corresponds to sensitivity to phage DS6A; lysotype Ax to phages DS6A, DNAIII8 and Legendre; lysotype 12 to phage DS6A, DNAIII8, Legendre, GS4E, BK1, PH, and Sedge; lysotype B to the same phage as lysotype 12 plus BG1; lysotype C to the same phage as lysotype B plus phage 3414. The M. bot:is BCG strains used for vaccine production were Pasteur strain 1173 P2, Glaxo strain 1077, the Russian strain and the Japanese strain 172. Six M. boz,is BCG strains were fresh isolates from suppurative abscesses after BCG vaccination.
Pulsed-field gel electrophoresis DNA preparation, restriction enzyme digestion and electrophoretic conditions were performed as previously described [1,17]. Mycolic acid profiles The determination of mycolate profiles was carried out as previously described [5,10].
4. RESULTS With final forward and reverse pulses of 60 and 20 s, lambda concatemers showed 14 bands (Fig. 1). The upper band corresponds to the limit of separation and thus DNA fragments of less than 630 kb (i.e. 13 lambda concatemers) are separated. With final forward and reverse pulses of 30 and 10 s, fragments up to 10 concatemers, i.e. 485 kb, were well separated and the distance between two concatemers was larger than when using former conditions (Figs. 2 and 3). All strains submitted to the study provided patterns which were easy to compare and were composed of less than 15 bands. Figures 1 and 2 show DraI patterns of M. tuberculosis strains submitted to the study. The phage prototypes displayed very dissimilar patterns (Fig. 2). By contrast, H37-derived strains presented related profiles with main differences encountered in the region over 388 kb. H37Rv strain presented three bands between 400 and 600 kb, not shared by H37Ra which showed only one band of approx. 500 kb. The other bands of the patterns were identical for both strains. H37HS appeared very similar to H37Rv, except that it lacked a band near 400 kb and that all bands presented some shifting down in molecular mass (Fig. 1). DraI patterns of isolates from suppurative adenitis following BCG vaccination of French patients (Fig. 3, lanes 1-7) could not be differentiated from the M. box'is BCG Pasteur strain used for vaccination (Fig. 3, lane 8). This shows the stability of DraI patterns during infection. However, DNA profiles of three other different M. box,is BCG strains used for vaccine production presented strict similarities for bands of molecu-
157
1 2 3 4 5 6
lar mass less than 145 kb, but discrete differences in the upper regions, which allowed differentiation of the four vaccine strains (Fig. 3, lanes 8-11). The mycolic content of these four BCG strains showed that Pasteur and Glaxo strains are devoid of methoxy-mycolate, whereas the Russian and Japanese strains present a mycolate content
1234
388 291
97
582 485 388
194 97
Fig. 1. Field inversion gel electrophoresis of D N A from M. tuberculosis H37-derived strains after DraI digestion. Lanes: l, Lambda concatemers; 2, M. tuberculosis H37Rv; 3, M. tuberculosis H37HS; 4, M. tuberculosis H37Ra. Pulse conditions were set to get forward and reverse pulses of 60 and 20 s respectively at the end of a 36-h run. The molecular masses are indicated in kb.
Fig. 2. Field inversion gel electrophoresis of D N A from prototype strains used for M. tuberculosis phage typing after DraI digestion. Lanes: 1, Lambda concatemers; 2, M. tuberculosis strain 965 (lysotype A); 3, M. tuberculosis strain 1025 (lysotype Ax); 4, M. tuberculosis strain 1324 (lysotype I2); 5, M. tuberculosis strain H37HS (lysotype B); 6, M. tuberculosis strain 1381 (lysotype C). Pulse conditions were set to get forward and reverse pulses of 30 and 10 s, respectively, at the end of a 36-h run. The molecular masses are indicated in kb.
similar to M. boo'is and other tubercle bacilli with a-, methoxy- and keto-mycolates (Fig. 4), confirming previous studies performed on some of these strains [5,11].
158
1 2 34
56
7 8 91011
388 291
97
Fig. 3. Field gel electrophoresis of DNA from M. boL,is BCG strains after Dral digestion. Lanes 1 to 7 correspond to BCG strains isolated from French patients following vaccination. Lanes 8-11 correspond to reference M. bocis BCG strains: 8, Pasteur; 9, Glaxo; 10, Russian; 11, Japanese. Lane 12 corresponds to lambda concatemers. Pulse conditions were set to get forward and reverse pulses of 30 and 10 s, respectively, at the end of a 36-h run. The molecular masses are indicated in kb,
5. DISCUSSION Epidemiological markers of M. tuberculosis strains are of special interest to trace the spread of tuberculosis among patients, study outbreaks and nosocomial infections, and differentiate between relapse and new infection. For many years,
epidemiology of M. tuberculosis was established by phage typing, which provided useful information for studying the propagation of tuberculosis in populations of different geographic origins. However, the method was of limited value for studying outbreaks or mini-epidemics due to the small number of established lysotypes [2,4]. Restriction profiles with standard electrophoretic conditions were then proposed as epidemiological markers but, due to the high number of bands produced, were difficult to interpret [3]. In the present study, tubercle bacilli DNA profiles yielded by digestion with DraI, a restriction enzyme with low-frequency cleavage sites, were composed of less than 15 bands and easy to compare. The epidemiology of M. tuberculosis" has recently been established by study of the different distribution of the multiple copy sequence IS6110 within the genome. M. tuberculosis isolates exhibit a high degree of restriction fragment length polymorphism; IS6110 is thus a very useful tool for the diagnosis and epidemiological studies of tuberculosis [12,16]. It has been demonstrated that a DraI cleavage site is present in IS6110 [15]. As distribution of IS6110 differs from strain to strain, DNA profiles yielded by Dra I digestion were expected to differ within M. tuberculosis'. Patterns displayed by the phage prototype strains confirmed that DNA F I G E patterns were also interesting epidemiological markers of M. tuberculosis" strains. Moreover, DraI F1GE patterns are of special interest for the study of M. bouis BCG strains. It has been demonstrated that BCG strains contain a single IS copy integrated at the same site in the chromosome, and no restriction fragment length polymorphism based on IS6110 could be detected [8]. However, BCG vaccine strains, all derived from the strain maintained and distributed by A. Calmette, evolved as variants differing by immunogenic properties, residual virulence, lipid composition and identification phenotypic tests [6,9,11]. In particular some BCG strains differ from other isolates of tubercle bacilli as they lack the ability to synthetize methoxy-mycolate. DraIdigested DNA profiles allow the differentiation of the BCG daughter strains as F I G E provides
159
,,,
e
W
,, IV V
IV
VI
1
2
3
4
5
6
7
Fig. 4. Mycolic acid patterns of M. bot'& BCG strains used for vaccination. The solvent used was petroleum ether-ether (88/12; v/v). Lanes 1 to 4 correspond to BCG strains: 1, Pasteur; 2, Glaxo; 3, Russian; 4, Japanese. The following lanes correspond to: 5, M. xenopi; 6, M. fortuitum; 7, M. a~'ium. The mycolate designations are as previously described [5]: I, a-mycolate, II, a'-mycolate; Ill, methoxy-mycolate; IV, keto-mycolate; V, epoxy-mycolate; VI, dicarboxy-mycolate.
different profiles, characteristic of each BCG strain, easy to interpret and compare. BCG Pasteur strains isolated from post-vaccinal suppurative adenitis from different patients showed identical patterns to the strain used for vaccine production, demonstrating the high reproducibility and fiability of the FIGE technique. The ability of DraI restriction profiles to show divergence between virulent and avirulent variants originated from the same parental strain may be useful for the construction of physical maps, facilitating further the genetic and molecular studies including the location and cloning of genes involved in virulence. However, it remains to be determined if the differences observed in the FIGE patterns are due to major DNA rearrangements or minor differences in the sequence corresponding to the cleavage site of the restriction enzyme used. Physical maps established for phage prototypes could also be helpful for the cloning of phage receptors. Pulsed-field gel electrophoresis of DNA restriction fragments generated by DraI hydrolysis thus represents a suitable technique for the analysis of mycobacteria at the genomic level. More-
over the DraI profiles constitute valuable epidemiological markers of BCG vaccine strains.
ACKNOWLEDGEMENTS We thank Hugo David for continuous interest and Brigitte Gicquel for reviewing the manuscript.
REFERENCES [1] Charvin, M., Rastogi, N. and Vincent-L6v2¢-Fr6bault, V. (1991) Curr. Microbiol. 22, 327-331. [2] Clavel-S~r~s, S. and Clement, F. (1984) Ann. Microbiol. 135B, 35-44. [3] Collins, D.M. and de Lisle, G.W. (1987) BCG identification by DNA restriction fragment patterns. J. Gen. Microbiol. 133, 1431-1434. [4] Crawford, J. and Bates, J.H. (1984) In: G.P. Kubica and L.G. Wayne (Ed.), pp. 123-132. The Mycobacteria: A Sourcebook. Marcel Dekker, New York, NY. [5] DaffY, M., Lan~elle, M.A., Asselineau, C., L~vy-Fr~bault, V. and H.L. David (1983) Ann. Microbiol. (Paris) 134B, 241-256. [6] Gheorghiu, M. (1990) Biologicals 18, 135-141.
160 [7] Hector, J.S.R., Pang, Y., Mazurek, G.H., Zhang, Y., Brown, B.A. and Wallace, R.J. Jr. (1992) J. Clin. Microbiol. 30, 1250-1255. [8] Hermans, P.W.M., van Soolingen, D., Bik, E.M., de Haas, P.E.W., Dale, J.W. and van Embden, J.D.A. (1991) Infect. Immun. 59, 2695-2705. [9] International Union against tuberculosis (1978) Tubercle 59, 139-142. [10] L~vy-Fr6bault, V., Gob, K.S. and David H.E. (1986) J. Clin. Microbiol. 24, 835-839. [11] Minnikin, D.E., Parlettt, J.H., Magnusson, M., Ridell, M. and Lind, A. (1984) J. Gen Microbiol. 130, 2733-2736. [12] Oral, I., Martin, C., Vincent-L6vy-Fr~bault, V., Thierry, D. and Gicquel, B. (1991) J. Clin. Microbiol. 29, 12521254.
[13] Rado, T.A., Bates, J.H., Engel, H.W.B., Mankiewicz, E., Murohashi, T., Mizubuchi, Y. and Sula, L. (1975) Am. Rev. Resp. Dis. 111,459-468. [14] Schwartz, D.C., Saffran, W., Welsh, J., Hass, R., Goldenberg, M. and Cantor, C.R. (1983) Cold Spring Harbor Symp. Quant. Biol., 189-195. [15] Thierry, D., Cave, M.D., Eisenach, K.D., Crawford, J.T., Bates, J.H., Gicquel, B. and Guesdon, J.L. (1990) Nucleic Acids Res. 18, 188. [16] Van Soolingen, D., Hermans, P.W.M., de Haas, P.E.W. and van Embden, J.D.A. (1991) J. Clin. Microbiol. 29, 2578-2586. [17] Vincent L6vy-Fr~bault, V., Thorel, M.F., Varnerot, A. and Gicquel, B. (1989) J. Clin. Microbiol. 27, 2823-2826.
155
FEMSLE (15131
Pulsed field gel electrophoresis of representatives of Mycobacterium tuberculosis and Mycobacterium bovis BCG strains A n n e Varnerot ", Fran~oise Cl6ment ~, Marina Gheorghiu b and Vdronique Vincent-Ldvy-Frdbault " a Centre National de Rdfdrence pour les Mycobactdries, Unitd de la Tuberculose et des Mycobactdries, Paris, France, and b Laboratoire du BCG, Institut Pasteur, Paris, France
Received 6 July 1992 Revision received 1 August 1992 Accepted 17 August 1992
Key words: Pulsed-field gel electrophoresis; Mycobacterium tuberculosis; Mycobacterium boris BCG
1. SUMMARY Using field inversion gel electrophoresis (FIGE), different Mycobacterium tuberculosis strains, such as phage prototypes, exhibit different DNA restriction patterns which are easy to compare. Virulent and avirulent variants of M. tuberculosis H37, as well as daughter strains of M. boris BCG, display characteristic DNA profiles. BCG strains isolated from suppurative adenitis following vaccination of French patients showed patterns identical to the BCG Pasteur strain used for vaccination. These results demonstrate that FIGE of DNA restriction fragments generated by DraI represents a suitable technique for the analysis of mycobacteria at a ge-
Correspondence to." V. Vincent-L6vy-Fr6bault, Centre National de R6f6rence pour les Mycobact6ries, Unit6 de la Tuberculose et des Mycobact6ries, Paris, France.
nomic level. The DraI profiles allow the differentiation and precise identification of the BCG Pasteur, Glaxo, Russian and Japanese strains.
2. INTRODUCTION Pulsed field gel electrophoresis (PFGE) allows the separation of large DNA fragments. The technique, originally developed by Schwartz et al. [14] for the separation of yeast chromosomes, has been applied to the analysis of bacterial genomes. After digestion by low-frequency cleavage restriction endonucleases, bacterial chromosomes provide DNA patterns composed of a few, usually well-separated fragments. Apart from the electrophoresis apparatus, which needs to be especially devised for ensuring electric pulses, DNA has to be extracted from agarose-embedded bacteria, in order to prevent mechanical, non-specific chromosomal fragmentation. Due to this con-
156
straint, usual lysis procedures of mycobacteria, requiring inhibitors of cell wall synthesis and actively growing mycobacteria, could not be used. We have previously applied field inversion gel electrophoresis (FIGE), one of the P F G E techniques, to mycobacteria and developed gentle lysis procedures ensuring suitable yields of entire DNA from immobilized, non-dividing cells [1,17]. The method demonstrated that the two Mycobacterium at,iurn subspecies, subsp, paratuberculosis and subsp, sih~aticurn, presented an homogeneous genomic organization producing subspecies-characteristic profiles, whereas the other M. at'ium strains displayed diverse restriction patterns [17]. Recently, Hector et al. [7] used P F G E in an epidemiological investigation of nosocomial outbreaks due to M. fortuitum. Studying the characteristics of avirulent and virulent strains may be useful for a better knowledge of the still poorly understood virulence of tuberculosis. In the present study, we applied the FIGE technique to some representatives of tubercle bacilli in order to detect differences at the genomic level.
3. M A T E R I A L S AND M E T H O D S
Bacterial strains M. tuberculosis strains used were: H37Rv, type strain of the species; H37Ra, the avirulent variant of the former; strains 965, 1025, 1324, 1381 and H37HS, the phage host prototypes of lysotypes A, Ax, I2, B and C, respectively. The designations of phage types correspond to phage susceptibility previously described [2,13]. Lysotype A corresponds to sensitivity to phage DS6A; lysotype Ax to phages DS6A, DNAIII8 and Legendre; lysotype 12 to phage DS6A, DNAIII8, Legendre, GS4E, BK1, PH, and Sedge; lysotype B to the same phage as lysotype 12 plus BG1; lysotype C to the same phage as lysotype B plus phage 3414. The M. bot:is BCG strains used for vaccine production were Pasteur strain 1173 P2, Glaxo strain 1077, the Russian strain and the Japanese strain 172. Six M. boz,is BCG strains were fresh isolates from suppurative abscesses after BCG vaccination.
Pulsed-field gel electrophoresis DNA preparation, restriction enzyme digestion and electrophoretic conditions were performed as previously described [1,17]. Mycolic acid profiles The determination of mycolate profiles was carried out as previously described [5,10].
4. RESULTS With final forward and reverse pulses of 60 and 20 s, lambda concatemers showed 14 bands (Fig. 1). The upper band corresponds to the limit of separation and thus DNA fragments of less than 630 kb (i.e. 13 lambda concatemers) are separated. With final forward and reverse pulses of 30 and 10 s, fragments up to 10 concatemers, i.e. 485 kb, were well separated and the distance between two concatemers was larger than when using former conditions (Figs. 2 and 3). All strains submitted to the study provided patterns which were easy to compare and were composed of less than 15 bands. Figures 1 and 2 show DraI patterns of M. tuberculosis strains submitted to the study. The phage prototypes displayed very dissimilar patterns (Fig. 2). By contrast, H37-derived strains presented related profiles with main differences encountered in the region over 388 kb. H37Rv strain presented three bands between 400 and 600 kb, not shared by H37Ra which showed only one band of approx. 500 kb. The other bands of the patterns were identical for both strains. H37HS appeared very similar to H37Rv, except that it lacked a band near 400 kb and that all bands presented some shifting down in molecular mass (Fig. 1). DraI patterns of isolates from suppurative adenitis following BCG vaccination of French patients (Fig. 3, lanes 1-7) could not be differentiated from the M. box'is BCG Pasteur strain used for vaccination (Fig. 3, lane 8). This shows the stability of DraI patterns during infection. However, DNA profiles of three other different M. box,is BCG strains used for vaccine production presented strict similarities for bands of molecu-
157
1 2 3 4 5 6
lar mass less than 145 kb, but discrete differences in the upper regions, which allowed differentiation of the four vaccine strains (Fig. 3, lanes 8-11). The mycolic content of these four BCG strains showed that Pasteur and Glaxo strains are devoid of methoxy-mycolate, whereas the Russian and Japanese strains present a mycolate content
1234
388 291
97
582 485 388
194 97
Fig. 1. Field inversion gel electrophoresis of D N A from M. tuberculosis H37-derived strains after DraI digestion. Lanes: l, Lambda concatemers; 2, M. tuberculosis H37Rv; 3, M. tuberculosis H37HS; 4, M. tuberculosis H37Ra. Pulse conditions were set to get forward and reverse pulses of 60 and 20 s respectively at the end of a 36-h run. The molecular masses are indicated in kb.
Fig. 2. Field inversion gel electrophoresis of D N A from prototype strains used for M. tuberculosis phage typing after DraI digestion. Lanes: 1, Lambda concatemers; 2, M. tuberculosis strain 965 (lysotype A); 3, M. tuberculosis strain 1025 (lysotype Ax); 4, M. tuberculosis strain 1324 (lysotype I2); 5, M. tuberculosis strain H37HS (lysotype B); 6, M. tuberculosis strain 1381 (lysotype C). Pulse conditions were set to get forward and reverse pulses of 30 and 10 s, respectively, at the end of a 36-h run. The molecular masses are indicated in kb.
similar to M. boo'is and other tubercle bacilli with a-, methoxy- and keto-mycolates (Fig. 4), confirming previous studies performed on some of these strains [5,11].
158
1 2 34
56
7 8 91011
388 291
97
Fig. 3. Field gel electrophoresis of DNA from M. boL,is BCG strains after Dral digestion. Lanes 1 to 7 correspond to BCG strains isolated from French patients following vaccination. Lanes 8-11 correspond to reference M. bocis BCG strains: 8, Pasteur; 9, Glaxo; 10, Russian; 11, Japanese. Lane 12 corresponds to lambda concatemers. Pulse conditions were set to get forward and reverse pulses of 30 and 10 s, respectively, at the end of a 36-h run. The molecular masses are indicated in kb,
5. DISCUSSION Epidemiological markers of M. tuberculosis strains are of special interest to trace the spread of tuberculosis among patients, study outbreaks and nosocomial infections, and differentiate between relapse and new infection. For many years,
epidemiology of M. tuberculosis was established by phage typing, which provided useful information for studying the propagation of tuberculosis in populations of different geographic origins. However, the method was of limited value for studying outbreaks or mini-epidemics due to the small number of established lysotypes [2,4]. Restriction profiles with standard electrophoretic conditions were then proposed as epidemiological markers but, due to the high number of bands produced, were difficult to interpret [3]. In the present study, tubercle bacilli DNA profiles yielded by digestion with DraI, a restriction enzyme with low-frequency cleavage sites, were composed of less than 15 bands and easy to compare. The epidemiology of M. tuberculosis" has recently been established by study of the different distribution of the multiple copy sequence IS6110 within the genome. M. tuberculosis isolates exhibit a high degree of restriction fragment length polymorphism; IS6110 is thus a very useful tool for the diagnosis and epidemiological studies of tuberculosis [12,16]. It has been demonstrated that a DraI cleavage site is present in IS6110 [15]. As distribution of IS6110 differs from strain to strain, DNA profiles yielded by Dra I digestion were expected to differ within M. tuberculosis'. Patterns displayed by the phage prototype strains confirmed that DNA F I G E patterns were also interesting epidemiological markers of M. tuberculosis" strains. Moreover, DraI F1GE patterns are of special interest for the study of M. bouis BCG strains. It has been demonstrated that BCG strains contain a single IS copy integrated at the same site in the chromosome, and no restriction fragment length polymorphism based on IS6110 could be detected [8]. However, BCG vaccine strains, all derived from the strain maintained and distributed by A. Calmette, evolved as variants differing by immunogenic properties, residual virulence, lipid composition and identification phenotypic tests [6,9,11]. In particular some BCG strains differ from other isolates of tubercle bacilli as they lack the ability to synthetize methoxy-mycolate. DraIdigested DNA profiles allow the differentiation of the BCG daughter strains as F I G E provides
159
,,,
e
W
,, IV V
IV
VI
1
2
3
4
5
6
7
Fig. 4. Mycolic acid patterns of M. bot'& BCG strains used for vaccination. The solvent used was petroleum ether-ether (88/12; v/v). Lanes 1 to 4 correspond to BCG strains: 1, Pasteur; 2, Glaxo; 3, Russian; 4, Japanese. The following lanes correspond to: 5, M. xenopi; 6, M. fortuitum; 7, M. a~'ium. The mycolate designations are as previously described [5]: I, a-mycolate, II, a'-mycolate; Ill, methoxy-mycolate; IV, keto-mycolate; V, epoxy-mycolate; VI, dicarboxy-mycolate.
different profiles, characteristic of each BCG strain, easy to interpret and compare. BCG Pasteur strains isolated from post-vaccinal suppurative adenitis from different patients showed identical patterns to the strain used for vaccine production, demonstrating the high reproducibility and fiability of the FIGE technique. The ability of DraI restriction profiles to show divergence between virulent and avirulent variants originated from the same parental strain may be useful for the construction of physical maps, facilitating further the genetic and molecular studies including the location and cloning of genes involved in virulence. However, it remains to be determined if the differences observed in the FIGE patterns are due to major DNA rearrangements or minor differences in the sequence corresponding to the cleavage site of the restriction enzyme used. Physical maps established for phage prototypes could also be helpful for the cloning of phage receptors. Pulsed-field gel electrophoresis of DNA restriction fragments generated by DraI hydrolysis thus represents a suitable technique for the analysis of mycobacteria at the genomic level. More-
over the DraI profiles constitute valuable epidemiological markers of BCG vaccine strains.
ACKNOWLEDGEMENTS We thank Hugo David for continuous interest and Brigitte Gicquel for reviewing the manuscript.
REFERENCES [1] Charvin, M., Rastogi, N. and Vincent-L6v2¢-Fr6bault, V. (1991) Curr. Microbiol. 22, 327-331. [2] Clavel-S~r~s, S. and Clement, F. (1984) Ann. Microbiol. 135B, 35-44. [3] Collins, D.M. and de Lisle, G.W. (1987) BCG identification by DNA restriction fragment patterns. J. Gen. Microbiol. 133, 1431-1434. [4] Crawford, J. and Bates, J.H. (1984) In: G.P. Kubica and L.G. Wayne (Ed.), pp. 123-132. The Mycobacteria: A Sourcebook. Marcel Dekker, New York, NY. [5] DaffY, M., Lan~elle, M.A., Asselineau, C., L~vy-Fr~bault, V. and H.L. David (1983) Ann. Microbiol. (Paris) 134B, 241-256. [6] Gheorghiu, M. (1990) Biologicals 18, 135-141.
160 [7] Hector, J.S.R., Pang, Y., Mazurek, G.H., Zhang, Y., Brown, B.A. and Wallace, R.J. Jr. (1992) J. Clin. Microbiol. 30, 1250-1255. [8] Hermans, P.W.M., van Soolingen, D., Bik, E.M., de Haas, P.E.W., Dale, J.W. and van Embden, J.D.A. (1991) Infect. Immun. 59, 2695-2705. [9] International Union against tuberculosis (1978) Tubercle 59, 139-142. [10] L~vy-Fr6bault, V., Gob, K.S. and David H.E. (1986) J. Clin. Microbiol. 24, 835-839. [11] Minnikin, D.E., Parlettt, J.H., Magnusson, M., Ridell, M. and Lind, A. (1984) J. Gen Microbiol. 130, 2733-2736. [12] Oral, I., Martin, C., Vincent-L6vy-Fr~bault, V., Thierry, D. and Gicquel, B. (1991) J. Clin. Microbiol. 29, 12521254.
[13] Rado, T.A., Bates, J.H., Engel, H.W.B., Mankiewicz, E., Murohashi, T., Mizubuchi, Y. and Sula, L. (1975) Am. Rev. Resp. Dis. 111,459-468. [14] Schwartz, D.C., Saffran, W., Welsh, J., Hass, R., Goldenberg, M. and Cantor, C.R. (1983) Cold Spring Harbor Symp. Quant. Biol., 189-195. [15] Thierry, D., Cave, M.D., Eisenach, K.D., Crawford, J.T., Bates, J.H., Gicquel, B. and Guesdon, J.L. (1990) Nucleic Acids Res. 18, 188. [16] Van Soolingen, D., Hermans, P.W.M., de Haas, P.E.W. and van Embden, J.D.A. (1991) J. Clin. Microbiol. 29, 2578-2586. [17] Vincent L6vy-Fr~bault, V., Thorel, M.F., Varnerot, A. and Gicquel, B. (1989) J. Clin. Microbiol. 27, 2823-2826.
