Letters in Applied Microhiology 1990, 11, 189-192
MSI170
Typing of Legionella pneumophila serogroups 2-1 4 strains by analysis of restriction fragment length polymorphisms T.G. H A R R I S O NN.A. , S A U N D E R SA. , H A T H T H O T U W AN, . DOSHI & A . G . T A Y I.OR PHLS Legionella Reference Unit, Division of Microbiological Reagents and Quality Control, Central Public Health Laboratory, 61 Colindale Avenue, London N W 9 5HT, U K Received 25 June 1990 and accepted 27 June 1990 H A K K I S O NT.G., , S A U N D E R SN , . A . , H A T H T H O T U W AA,, , DOSHI,N . & T A YL O K , A . G . 1990. Typing of Legionella pneurnophila serogroups 2-14 strains by analysis or restriction fragment length polymorphisms. Letters in Applied Microbioloqy 11, 189-192. A typing method based on analysis of restriction fragment length polymorphisms has previously been developed for Legionella pneurnophila serogroup 1. Here data are presented demonstrating the utility of this method for typing strains of all other L. pnvurnophila serogroups described to date. The method, which is highly discriminatory, should be of considerable value in epidemiological investigations of legionella infections.
Various typing methods have been developed to compare strains of Legionella pneumophila isolated from environmental sources with clinical strains, as an aid in epidemiological investigations (Harrison & Taylor 1988). Serogrouping, the most widely used technique, divides the species into 14 serogroups (Benson et al. 1988). We have previously reported the development and evaluation of a highly discriminatory and reproducible typing method for L. pneumophila serogroup 1, based on the use of cloned biotinylated DNA probes for the analysis of restriction fragment length polymorphisms (RFLPs) (Saunders et al. 1990). Here we present data demonstrating the utility of this technique for typing strains of L. pneumophila serogroups 2-14. Materials and Methods BACTERIAL STRAINS
Isolates of L. pneumophila serogroups 2-14 were submitted for examination during 1979-88 or obtained from the National Collection of Type Cultures, Colindale. Their identity was confirmed on the basis of nutritional, biochemical and serological characteristics (Harrison &
Taylor 1988). The strains, which were stored on glass beads at -7o"C, were recovered onto BCYE and incubated for 72 h at 37°C before harvesting in saline and pelleting by centrifugation. In this study 147 strains were examined (Table 1). After reviewing all available data strains were classified as 'apparently unrelated' or 'related' to other strains. Apparently unrelated strains were used for estimations of the discrimination of the typing method. RFLP TYPING
RFLP typing was carried out as previously described (Saunders et al. 1990) using DNA purified by the guanidine isothiocyanate method and the biotinylated probes LNS20 and 2NS21.
Results and Discussion The two cloned sequences employed as a probe in this study were selected from a bacteriophage lambda genomic library of L. pneurnophila Knoxville-1 (NCTC 11286) DNA and are specific for L. pneumophila (Saunders et al. 1990). All 147 strains examined in this study were typable, with the exception of one strain, as
190
T . G . Harrison et al. Table 1. Typing of 147 Legionella pneumophila serogroup 2-14 strains
Representative strain*
Related strains?
LC787a LC3 LC789a NCTCl1250 LC224 LC840 LC1160 LC408 LC740 LC861 LC563 LC232 LC580 LC739 LC1058 LC223 LRU72 LRU73 LC82 NCTCl I232 LRU71 LC574
Serogroups
RFLP type
Discrimination7
1 1 1 1 1
2 2 2 2 2
34 43 48 53
15
515
1 1 1 1
5
8/17
1 3 1 1 1 1 1 2 1 1 1 1
3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
42
NCTCl1233
1
4
49
LC356 LC1074d NCTCll417 LC729 LC805c LC608 LC434 LC560
1
5 5 5
44 45 50 50 52 5 5 42
A247181 A89181 LC832 NCTC 11287 LC796 LC883 LC543 NCTC 11406 LC202 LC222 LC807 LC361 LC437 LC566 LClOo LC2 14 LC396 LC439 LC767 LC227
Source1
1
1 1 1 1 2
5
5 5D 5D 5D
1
1 2 1 1 2 1 1 1 1 219 I 1 317 1 1 1 1 812 1 1 2
. ,
RFLP, Resitriction fragment length polymorphism.
6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6
5 15 15 26 26 28 28 28 34 34 34 34 34 37
40
5
11 11 13 15 15
25 26 26 26 26 26 26 26 28 28 28/15** 28 29 34
213
10121
Erratum Typing of Legionella pneumophila serogroups 2-14 strains by analysis of restriction fragment length polymorphisms T. G . H A R R I S O N N,. A . S A U N D E R SA,. H A T H T H O T U WNA. ,DOSHI& A . G . T A Y L O RLetters . in Applied Microbiology 1990, 11, 189-192. Table 1 on page 190 should read as follows. Table 1. Typing of 147 Legionella pneumophila serogroup 2-14 strains Representative strain*
Related strains?
Source1
Serogroupjj
RFLP type
Discrimination1
LC727a LC3 LC789a NCTC I 1230 LC224
1 1 1
2 2 2
1
2
1
2
LC840 LCll60 LC408 LC740 LC861 LC563 LC232 LC580 LC739 LC1058 LC223 LRU72 LRU73 LC82 NCTClI232 LRU71 LC574
1 1 1 1
1 1 1 2 1 I 1 1
3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
15 15 26 26 28 28 28 34 34 34 34 34 31 40 42
NCTCll233
I
4
49
111
LC356 LCI 074d NCTCll417 LC729
I I
5 5
415
1 1 1 2 1 1
5
44 45 50 50 52 5 5 42
2 1 1 2 1 1 1 1 219
6 6 6 6 6 6 6 6 6 6 6 6 6
5 11 11 13
10121
LC805C LC608 LC434 LC560 A24718 1 A8919 1 LC832 NCTCll287 LC796 LC883 LC543 NCTCl1406 LC202 LC222 LC807 LC361 LC437
I 1 3 1 1
1 1
317 1
5 5 SD
5D 5D
(continued)
15
5/5
34 43 48 53 5
8/17
5
15
15 25 26 26 26 26 26 26
213
Erratum Table 1 continued Representative strain.
Related strains?
Source$
Serogroupg
RFLP type
Discrimination7
1 1 1 812 1 1 2 1
6 6 6 6 6 6 6 6
26 28 28 28/15** 28 29 34 56
LC258 NCTCll984 LC254
1 1 1
7
3f 3
7
7 16 55
LC419 LC229 LC812 LC808 LC569 LC1 LC831 LC573 NCTC11985
1
1 1 1 2/12 1 1 1 1
8 8 8 8 8 8 8 8 8
7 10 13 15 25 28 31 42 51
919
NCTC11986 LC802 LC351 LC742
1 1 1
9 9 9 9
1 15 15 57
314
LC811 LC384 LC467 NCTC 12000
1 412 1 1
10 10 10 10
11 15 26 28
414
NCTC12179
1
11
19
111
LC1156 NCTC12180 LC204 LC465 LC744 LC1180
1
5 26 26 26 26 26
216
115 1 1 112
12 12 12 12 12 12
NCTC12181
1
13
7
111
LC572 LC570 LC575
1 1 1
14 14 14
40
34
313
LC566 LCloO LC214 LC396 LC439 LC767 LC227 LC809
I
1
1
U K (E)
42
* All representative strains are apparently unrelated to any other representative strain.
t The number of epidemiologically related strains (including the representative strain). Where two figures are given, the first is the number of clinical strains, the second the number of environmental strains. $ The country of origin (environmental strains) or country in which LD was thought to be contracted (clinical strains). C, Clinical strains; E, environmental strains; C/E, both clinical and environmental strains. $ As defined by Harrison & Taylor (1988). 5D indicates that the strain reacts like a Dallas-1E strain rather than a Cambridge-2 strain. 7 Expressed as the number of RFLP types obtainedlthe number of representative strains of that serogroup. ** Epidemiologically related strains which gave two RFLP types (see text).
RFLP typing of L. pneumophila were all serogroup 1 strains previously examined. Thus, it seems likely that, with few exceptions, any strain of L. pneumophila will be typable by this method irrespective of serogroup. The results obtained for the 147 strains (Table 1) show that 30 distinct RFLP types were identified among the 83 apparently unrelated strains (Fig. 1). The discrimination within each serogroup, where more than one strain was available, was shown to be excellent. It is, however, noteworthy that RFLP types are not serogroup specific. For example, some strains representative of serogroups 2, 3, 6, 8, 9 and 10 were all found to be type 15. Many sets of epidemiologically related isolates were used in this study including representative strains from several documented LD investigations (Ezzeddine et al. 1989; Meigh et
191
al. 1989; Nazareth et al. 1989). In all but one case the RFLP typing results were consistent with the epidemiological and serological data, confirming the reproducibility and utility of the method. The exception was a set of strains from an outbreak in a Belgian hospital (Ezzeddine et al. 1989). Here, clinical and environmental strains of serogroups 6 and 10 (including a dual infection) were made. The serogroup 10 strains (environmental and clinical) and environmental serogroup 6 strains were all RFLP type 15. However the clinical serogroup 6 strains were RFLP type 28. These interesting findings will be studied further. The data presented here show that the RFLP typing method is suitable for the epidemiological investigation of infections caused by L. pneumophila of any serogroup.
Fig. 1. Examples of restriction fragment length polymorphism (RFLP) patterns: Track 1, LC742/ serogroup 9/ RFLP type 56; 2, NCTC1200/sgp lO/type 28; 3, LC1074d/sgp 5jtype 45; 4, LC8ll/sgp lO/type 11; 5, NCTC11232/sgp 3/type 37; 6, NCTC1218l/sgp 13/type 7; 7, NCTCl1984/sgp 7/type 16; 8, LC254/sgp 7/type 55; 9, LC3/sgp 2/type 34; 10, strain not included in study; 11, LC809/sgp ')/type 56; 12, LCl/sgp 8/type 28. Restriction fragments (approximately 1 gjlane) were subjected to electrophoresis at 1.25 V cm for 16 h. The DNA was blotted and hybridized to biotinylated probe (300 ng). Bound probe was visualized by the BluGENE reagents. Lanes marked i, show a mixture of Eco RI and Pst I digests of I phage DNA.
192
T . G . Harrison et al.
We are grateful to Dr C.L.R. Bartlett and the staff of CDSC, Colindale, for providing epidemiological data and to all our microbiologist colleagues who submitted strains t o us for examination.
References EZZEDDINE,H., VAN OSSEL, C., DELMEE,M. & WAUTERS, G. 1989 Legionella spp. in a hospital hot water system: effect of control measures. Journal of Hospital Infection 13, 121-131. HARRISON, T.G. & TAYLOR, A.G. 1988 A Laboratory
Manual for Leqiunnaires ' Disease. Chichester : Wiley. MEIGH,R.E., MAKIN,T., SCOTT, M.H. & HART,C.A. 1989 Legionella pneumophila serogroup 12 pneumonia in a renal transplant recipient: case report and environmental observations. Journal of Hospital Infection 13, 315-319. NAZARETH, B. & SMITH,G.W. 1989 Legionella: a case for culture. Lancet ii, 113. SAUNDERS, N.A., HARRISON, T.G., HATHTHOTUWA, A,, KACHWALLA, N. & TAYLOR,A.G. 1990 A method for typing strains of Legionella pneumophila serogroup 1 by analysis of restriction fragment length polymorphisms. Journal of Medical Microbiology 31,45-55.
MSI170
Typing of Legionella pneumophila serogroups 2-1 4 strains by analysis of restriction fragment length polymorphisms T.G. H A R R I S O NN.A. , S A U N D E R SA. , H A T H T H O T U W AN, . DOSHI & A . G . T A Y I.OR PHLS Legionella Reference Unit, Division of Microbiological Reagents and Quality Control, Central Public Health Laboratory, 61 Colindale Avenue, London N W 9 5HT, U K Received 25 June 1990 and accepted 27 June 1990 H A K K I S O NT.G., , S A U N D E R SN , . A . , H A T H T H O T U W AA,, , DOSHI,N . & T A YL O K , A . G . 1990. Typing of Legionella pneurnophila serogroups 2-14 strains by analysis or restriction fragment length polymorphisms. Letters in Applied Microbioloqy 11, 189-192. A typing method based on analysis of restriction fragment length polymorphisms has previously been developed for Legionella pneurnophila serogroup 1. Here data are presented demonstrating the utility of this method for typing strains of all other L. pnvurnophila serogroups described to date. The method, which is highly discriminatory, should be of considerable value in epidemiological investigations of legionella infections.
Various typing methods have been developed to compare strains of Legionella pneumophila isolated from environmental sources with clinical strains, as an aid in epidemiological investigations (Harrison & Taylor 1988). Serogrouping, the most widely used technique, divides the species into 14 serogroups (Benson et al. 1988). We have previously reported the development and evaluation of a highly discriminatory and reproducible typing method for L. pneumophila serogroup 1, based on the use of cloned biotinylated DNA probes for the analysis of restriction fragment length polymorphisms (RFLPs) (Saunders et al. 1990). Here we present data demonstrating the utility of this technique for typing strains of L. pneumophila serogroups 2-14. Materials and Methods BACTERIAL STRAINS
Isolates of L. pneumophila serogroups 2-14 were submitted for examination during 1979-88 or obtained from the National Collection of Type Cultures, Colindale. Their identity was confirmed on the basis of nutritional, biochemical and serological characteristics (Harrison &
Taylor 1988). The strains, which were stored on glass beads at -7o"C, were recovered onto BCYE and incubated for 72 h at 37°C before harvesting in saline and pelleting by centrifugation. In this study 147 strains were examined (Table 1). After reviewing all available data strains were classified as 'apparently unrelated' or 'related' to other strains. Apparently unrelated strains were used for estimations of the discrimination of the typing method. RFLP TYPING
RFLP typing was carried out as previously described (Saunders et al. 1990) using DNA purified by the guanidine isothiocyanate method and the biotinylated probes LNS20 and 2NS21.
Results and Discussion The two cloned sequences employed as a probe in this study were selected from a bacteriophage lambda genomic library of L. pneurnophila Knoxville-1 (NCTC 11286) DNA and are specific for L. pneumophila (Saunders et al. 1990). All 147 strains examined in this study were typable, with the exception of one strain, as
190
T . G . Harrison et al. Table 1. Typing of 147 Legionella pneumophila serogroup 2-14 strains
Representative strain*
Related strains?
LC787a LC3 LC789a NCTCl1250 LC224 LC840 LC1160 LC408 LC740 LC861 LC563 LC232 LC580 LC739 LC1058 LC223 LRU72 LRU73 LC82 NCTCl I232 LRU71 LC574
Serogroups
RFLP type
Discrimination7
1 1 1 1 1
2 2 2 2 2
34 43 48 53
15
515
1 1 1 1
5
8/17
1 3 1 1 1 1 1 2 1 1 1 1
3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
42
NCTCl1233
1
4
49
LC356 LC1074d NCTCll417 LC729 LC805c LC608 LC434 LC560
1
5 5 5
44 45 50 50 52 5 5 42
A247181 A89181 LC832 NCTC 11287 LC796 LC883 LC543 NCTC 11406 LC202 LC222 LC807 LC361 LC437 LC566 LClOo LC2 14 LC396 LC439 LC767 LC227
Source1
1
1 1 1 1 2
5
5 5D 5D 5D
1
1 2 1 1 2 1 1 1 1 219 I 1 317 1 1 1 1 812 1 1 2
. ,
RFLP, Resitriction fragment length polymorphism.
6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6
5 15 15 26 26 28 28 28 34 34 34 34 34 37
40
5
11 11 13 15 15
25 26 26 26 26 26 26 26 28 28 28/15** 28 29 34
213
10121
Erratum Typing of Legionella pneumophila serogroups 2-14 strains by analysis of restriction fragment length polymorphisms T. G . H A R R I S O N N,. A . S A U N D E R SA,. H A T H T H O T U WNA. ,DOSHI& A . G . T A Y L O RLetters . in Applied Microbiology 1990, 11, 189-192. Table 1 on page 190 should read as follows. Table 1. Typing of 147 Legionella pneumophila serogroup 2-14 strains Representative strain*
Related strains?
Source1
Serogroupjj
RFLP type
Discrimination1
LC727a LC3 LC789a NCTC I 1230 LC224
1 1 1
2 2 2
1
2
1
2
LC840 LCll60 LC408 LC740 LC861 LC563 LC232 LC580 LC739 LC1058 LC223 LRU72 LRU73 LC82 NCTClI232 LRU71 LC574
1 1 1 1
1 1 1 2 1 I 1 1
3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
15 15 26 26 28 28 28 34 34 34 34 34 31 40 42
NCTCll233
I
4
49
111
LC356 LCI 074d NCTCll417 LC729
I I
5 5
415
1 1 1 2 1 1
5
44 45 50 50 52 5 5 42
2 1 1 2 1 1 1 1 219
6 6 6 6 6 6 6 6 6 6 6 6 6
5 11 11 13
10121
LC805C LC608 LC434 LC560 A24718 1 A8919 1 LC832 NCTCll287 LC796 LC883 LC543 NCTCl1406 LC202 LC222 LC807 LC361 LC437
I 1 3 1 1
1 1
317 1
5 5 SD
5D 5D
(continued)
15
5/5
34 43 48 53 5
8/17
5
15
15 25 26 26 26 26 26 26
213
Erratum Table 1 continued Representative strain.
Related strains?
Source$
Serogroupg
RFLP type
Discrimination7
1 1 1 812 1 1 2 1
6 6 6 6 6 6 6 6
26 28 28 28/15** 28 29 34 56
LC258 NCTCll984 LC254
1 1 1
7
3f 3
7
7 16 55
LC419 LC229 LC812 LC808 LC569 LC1 LC831 LC573 NCTC11985
1
1 1 1 2/12 1 1 1 1
8 8 8 8 8 8 8 8 8
7 10 13 15 25 28 31 42 51
919
NCTC11986 LC802 LC351 LC742
1 1 1
9 9 9 9
1 15 15 57
314
LC811 LC384 LC467 NCTC 12000
1 412 1 1
10 10 10 10
11 15 26 28
414
NCTC12179
1
11
19
111
LC1156 NCTC12180 LC204 LC465 LC744 LC1180
1
5 26 26 26 26 26
216
115 1 1 112
12 12 12 12 12 12
NCTC12181
1
13
7
111
LC572 LC570 LC575
1 1 1
14 14 14
40
34
313
LC566 LCloO LC214 LC396 LC439 LC767 LC227 LC809
I
1
1
U K (E)
42
* All representative strains are apparently unrelated to any other representative strain.
t The number of epidemiologically related strains (including the representative strain). Where two figures are given, the first is the number of clinical strains, the second the number of environmental strains. $ The country of origin (environmental strains) or country in which LD was thought to be contracted (clinical strains). C, Clinical strains; E, environmental strains; C/E, both clinical and environmental strains. $ As defined by Harrison & Taylor (1988). 5D indicates that the strain reacts like a Dallas-1E strain rather than a Cambridge-2 strain. 7 Expressed as the number of RFLP types obtainedlthe number of representative strains of that serogroup. ** Epidemiologically related strains which gave two RFLP types (see text).
RFLP typing of L. pneumophila were all serogroup 1 strains previously examined. Thus, it seems likely that, with few exceptions, any strain of L. pneumophila will be typable by this method irrespective of serogroup. The results obtained for the 147 strains (Table 1) show that 30 distinct RFLP types were identified among the 83 apparently unrelated strains (Fig. 1). The discrimination within each serogroup, where more than one strain was available, was shown to be excellent. It is, however, noteworthy that RFLP types are not serogroup specific. For example, some strains representative of serogroups 2, 3, 6, 8, 9 and 10 were all found to be type 15. Many sets of epidemiologically related isolates were used in this study including representative strains from several documented LD investigations (Ezzeddine et al. 1989; Meigh et
191
al. 1989; Nazareth et al. 1989). In all but one case the RFLP typing results were consistent with the epidemiological and serological data, confirming the reproducibility and utility of the method. The exception was a set of strains from an outbreak in a Belgian hospital (Ezzeddine et al. 1989). Here, clinical and environmental strains of serogroups 6 and 10 (including a dual infection) were made. The serogroup 10 strains (environmental and clinical) and environmental serogroup 6 strains were all RFLP type 15. However the clinical serogroup 6 strains were RFLP type 28. These interesting findings will be studied further. The data presented here show that the RFLP typing method is suitable for the epidemiological investigation of infections caused by L. pneumophila of any serogroup.
Fig. 1. Examples of restriction fragment length polymorphism (RFLP) patterns: Track 1, LC742/ serogroup 9/ RFLP type 56; 2, NCTC1200/sgp lO/type 28; 3, LC1074d/sgp 5jtype 45; 4, LC8ll/sgp lO/type 11; 5, NCTC11232/sgp 3/type 37; 6, NCTC1218l/sgp 13/type 7; 7, NCTCl1984/sgp 7/type 16; 8, LC254/sgp 7/type 55; 9, LC3/sgp 2/type 34; 10, strain not included in study; 11, LC809/sgp ')/type 56; 12, LCl/sgp 8/type 28. Restriction fragments (approximately 1 gjlane) were subjected to electrophoresis at 1.25 V cm for 16 h. The DNA was blotted and hybridized to biotinylated probe (300 ng). Bound probe was visualized by the BluGENE reagents. Lanes marked i, show a mixture of Eco RI and Pst I digests of I phage DNA.
192
T . G . Harrison et al.
We are grateful to Dr C.L.R. Bartlett and the staff of CDSC, Colindale, for providing epidemiological data and to all our microbiologist colleagues who submitted strains t o us for examination.
References EZZEDDINE,H., VAN OSSEL, C., DELMEE,M. & WAUTERS, G. 1989 Legionella spp. in a hospital hot water system: effect of control measures. Journal of Hospital Infection 13, 121-131. HARRISON, T.G. & TAYLOR, A.G. 1988 A Laboratory
Manual for Leqiunnaires ' Disease. Chichester : Wiley. MEIGH,R.E., MAKIN,T., SCOTT, M.H. & HART,C.A. 1989 Legionella pneumophila serogroup 12 pneumonia in a renal transplant recipient: case report and environmental observations. Journal of Hospital Infection 13, 315-319. NAZARETH, B. & SMITH,G.W. 1989 Legionella: a case for culture. Lancet ii, 113. SAUNDERS, N.A., HARRISON, T.G., HATHTHOTUWA, A,, KACHWALLA, N. & TAYLOR,A.G. 1990 A method for typing strains of Legionella pneumophila serogroup 1 by analysis of restriction fragment length polymorphisms. Journal of Medical Microbiology 31,45-55.
