Microbiology (2004), 150, 199–204
DOI 10.1099/mic.0.26642-0
A polymorphic tandem repeat potentially useful for typing in the chromosome of Yersinia enterocolitica Ine´s de Benito,1 Marı´a Eliecer Cano,1 Jesu´s Agu¨ero1,2 and Juan M. Garcı´a Lobo2 1
Servicio de Microbiologı´a, Hospital Universitario Marque´s de Valdecilla, Avenida de Valdecilla s/n, 39008-Santander, Spain
Correspondence Juan M. Garcı´a Lobo
2
[email protected]
Departamento de Biologı´a Molecular, Facultad de Medicina, Universidad de Cantabria, Centro Asociado al CIB, CSIC, Cardenal Herrera Oria s/n, 39011, Santander, Spain
Received 11 July 2003 Revised
30 September 2003
Accepted 15 October 2003
The hexanucleotide CCAGCA was found repeated 15 times in tandem on the 59 side of the virginiamycin acetyl transferase gene of Yersinia enterocolitica strain Y56. The corresponding region was analysed by PCR from 54 clinical strains belonging to the same biotype and serotype, and others from this laboratory collection belonging to different biotypes and serotypes. Each strain produced a single amplification product whose size was variable among strains, revealing that the locus was polymorphic. Nucleotide sequence determination of selected PCR products showed that the polymorphism was due to the precise expansion or reduction in the number of hexanucleotide repeats. Analysis of this locus in a few strains showing the same PFGE pattern showed that it was also polymorphic. These results suggest that this method could be valuable to increase the discriminatory power of current Y. enterocolitica typing schemes.
INTRODUCTION Short nucleotide sequences repeated in tandem, called tandem repeats or short sequence repeats, constituting microsatellite elements, are very frequent in eukaryotic organisms. Unstable tandem repeat loci are involved in some human diseases such as fragile X syndrome (Verkerk et al., 1991). The accumulating bacterial genomic data have revealed that tandem repeats are also frequent in bacteria. In some cases these bacterial repetitive loci are polymorphic due to differences in the number of repeated units they contain. These have been called variable number tandem repeats (VNTRs) (van Belkum et al., 1997) and simple sequence contingency loci (SSCL) (Bayliss et al., 2001). The mechanism to generate changes in the number of repeated units in these loci is not very well known but it is suspected to be due to strand slippage during DNA replication (van Belkum et al., 1997). The biological consequence of these polymorphisms depends on the size and position of the repeat. For VNTRs located inside an ORF the effect will depend on the size of the repeated sequence. If it is three or a multiple of three, the change will conserve the sequence of the protein, except for one or a few repeated amino acids. If the repeated sequence is not a multiple of three, the variation can produce a frameshift with complete loss of protein primary structure. For VNTRs outside an ORF the change may be silent unless the region is involved in gene control or expression. This effect has been described in Neisseria as a Abbreviations: PA, proline-alanine; VNTR, variable number tandem repeats.
0002-6642 G 2004 SGM
Printed in Great Britain
major component of the change of phase of molecules involved in pathogenicity (Saunders et al., 2000). Polymorphisms in VNTRs can also affect fine-tuning of the level of gene expression (Kashi et al., 1997). Recently, the value of these bacterial microsatellites has been appreciated as an epidemiological tool (van Belkum, 1999). It has been used to characterize bacterial species that are genetically homogeneous and difficult to type by other means (Adair et al., 2000; Frothingham & Meeker-O’Connell, 1998; Keim et al., 2000). Yersinia enterocolitica, a pathogenic bacterium responsible for gastrointestinal infections carried by contaminated food, has a well-established typing system based on a combination of serological, biochemical and phage susceptibility tests. In addition, several molecular techniques have proved to be adequate to type Y. enterocolitica (Iteman et al., 1996), including the analysis of polymorphisms associated with repeated sequences (Hallanvuo et al., 2002). However, the isolates coming from many outbreaks, and those prevalent in a given geographical area, often share the same phenotype and genomic fingerprint using the currently available methods. Therefore, the availability of new markers in Y. enterocolitica with additional discriminatory power would constitute a very useful epidemiological tool. Here we describe the identification and characterization of a VNTR sequence in Y. enterocolitica. This locus was polymorphic in a sample of clinical isolates belonging to the same biotype and serotype, but not related epidemiologically. The locus was also polymorphic in 199
I. de Benito and others
strains with the same pulsotype. Our preliminary results suggest that the analysis of this VNTR may provide a new method to further discriminate among Y. enterocolitica isolates found to be identical with other epidemiological tools.
METHODS
agarose gels using the Qiagen band purification kit. DNA sequencing was performed with a Perkin Elmer automatic sequencer at the Centro de Investigaciones Biolo´gicas, Madrid, Spain. Sequence comparison with the unfinished Yersinia pseudotuberculosis genome was performed at the web server of the Biology and Biotechnology Research Program of the Lawrence Livermore National Laboratory, CA, USA (http://bbrp.llnl.gov/bbrp/bin/y. pseudotuberculosis_blast).
Bacterial strains. Fifty five isolates of Y. enterocolitica, named YHV1–YHV55, were obtained from the Microbiology Laboratory of the University Hospital ‘Marque´s de Valdecilla’, Santander, Spain, from March 1995 to January 2001. All the strains were isolated from faeces and they did not demonstrate any evident epidemiological relationship. Strain Y56 (biotype 4, serotype 3, lysotype VIII) was from our laboratory collection and has been described previously (Seoane & Garcı´a Lobo, 1991). Strain 8081 (biotype 1B, serotype 8) was generously provided by Dr Ramo´n Dı´az from the University of Pamplona, Spain. Strains representative of other biotypes, P1403 and P219 (Biotype 1A), WA and 13514 (Biotype 1B), My79b and IP97 (Biotype 2), and IP4124 and IP22274 (Biotype 4, serotype 3, lysotype IXb), were from our laboratory collection. Bacteria were routinely grown in TSB or on TSA plates without antibiotics at 37 uC.
The unfinished Y. enterocolitica 8081 genome was accessed at the Sanger Centre web server at http://www.sanger.ac.uk/Projects/ Y_enterocolitica/blast_server.shtml
DNA isolation, amplification and analysis. Total DNA from
Identification of a VNTR in Y. enterocolitica
bacteria for PCR was prepared with the InstaGene matrix (Bio-Rad). DNA obtained by this method from some strains was readily degraded, presumably by the nucleases present in the preparation. In these cases DNA was prepared with the Roche high pure PCR template preparation kit. Around 10 ng total DNA was used as template in PCR with the primers described in the text. The size of the amplification products was evaluated by analytical agarose gel electrophoresis using 3 % gels and a 50 bp ladder (Life Technologies) as a size marker. For DNA sequencing the appropriate bands were excised and purified from
PFGE. The genomic DNA of the different strains embedded in
agarose plugs was prepared essentially as described by Saken et al. (1994). DNA in plugs was digested with NotI for 16 h at 37 uC. The samples were electrophoresed by the contour-clamped homogeneous electric field technique in a CHEF-DR III system (Bio-Rad) in 1 % (w/v) agarose gels made in 0?56 TBE buffer at 14 uC at 6 V cm21. Time pulses were ramped from 8 to 23 s in 20 h.
RESULTS
As part of a sequencing project of the virginiamycin acetyl transferase gene (vat) from the Y. enterocolitica strain Y56 (Seoane & Garcı´a Lobo, 2000) we observed a sequence consisting of 15 perfect repetitions of the hexanucleotide CCAGCA (Fig. 1a). To determine whether this locus was polymorphic among different isolates of Y. enterocolitica, we designed two primers, P1 (59-ACCGGCACTGAATCTGAAGATGCG-39) and P2 (59-CTTATCTTCCATAATCTAGACCTC-39), flanking the CCAGCA repeats.
Fig. 1. (a) Gene map showing the analysed region in Y. enterocolitica 8081 and Y56. The vat gene was absent in strain 8081; it was replaced by a hypothetical ORF represented in grey. The black box at the 39 end of orf528 represents the CCAGCA repeat region. The dotted line in orf528 in Y56 indicates the extent of the available sequence. malI in Y56 is truncated. Sequence similarity is regained 1?8 kb away, suggesting a deletion in Y56. (b) Nucleotide sequence of the 39 end of orf528 in Y. enterocolitica strain Y56. The sequence of the PCR primers is shown in the black boxes. The 15 tandem repeats are underlined. In the bottom line we show a scheme of the sequence containing the tandem repeat in all the Y. enterocolitica strains sequenced. The only difference among the isolates was in the number of repeats. (c) Alignment of the carboxy-terminal regions of Orf528 from the three Yersinia strains. In Y. pestis the sequence PA appears twice near the end. 200
Microbiology 150
Polymorphic tandem repeat in Y. enterocolitica
Table 1. Y. enterocolitica strains grouped according to the number of CCAGCA repeats they contain n*
FrequencyD
Strainsd
14 13
6/55 15/55
11 9 8 7 5 4
6/55 4/55 9/55 9/55 1/55 5/55
YHV2, YHV3, YHV4, YHV19, YHV49, YHV55 YHV1, YHV7, YHV9, YHV10, YHV13, YHV14, YHV15, YHV27, YHV29, YHV30, YHV32, YHV42, YHV44, YHV46, YHV53 YHV5, YHV11, YHV23, YHV24, YHV26, YHV28 YHV18, YHV38, YHV52, YHV54 YHV6, YHV22, YHV31, YHV37, YHV39, YHV41, YHV48, YHV50, YHV51 YHV8, YHV12, YHV17, YHV20, YHV21, YHV34, YHV36, YHV40, YHV43 YHV25 YHV16, YHV33, YHV35, YHV45, YHV47
*Number of CCAGCA repeats. DNumber of occurrences of the allele in the sample analysed. dThe number of repeat units in the underlined strains was determined by DNA sequencing.
These primers amplified a 210 bp DNA fragment from the Y. enterocolitica Y56 chromosome, including the repeated sequence. The same primers were used to amplify the corresponding region from a collection of 55 Y. enterocolitica strains (Table 1). The PCR products from the different isolates were sized by gel electrophoresis and found to be different, indicating that the locus was polymorphic (Fig. 2). To precisely determine the nature of this polymorphism, PCR products representative of the different sizes were purified from gels and sequenced with primer P2. Analysis of the resulting sequences indicated that the size differences were due to the precise change in the number of repeated CCAGCA sequences. The nucleotide sequence of every PCR product was identical except for the number of tandem repeats. All the sequences adhered to the general structure represented in Fig. 1(b). In some cases we sequenced several strains with the same amplicon size, as determined by gel electrophoresis, and obtained the same sequence. Accordingly, we assumed for the remaining strains that the number of repeated copies could be directly deduced from their amplicon size (Table 1). The serotype and biotype of the strains used was determined as indicated in Methods. All the strains belonged to serotype O : 3, biotype 4 (except strain YHV52 which was nonserotypable and belonged to biotype 1A) and produced the Myf factor (Leyva et al., 1995)
two strains belonging to biotype 4, serogroup 3, lysotype IXb contained nine and five repeat copies. Genome sequencing in the genus Yersinia has already
m
1
2
3
4
5
6
7
8
m
bp
350
150
Analysis of the CCAGCA repeat in other Yersinia strains We also determined the status of the locus in some Y. enterocolitica strains from our collection belonging to other biotypes (Fig. 3). Two strains of biotype 1A produced the same amplicon corresponding in size to five copies of the repeat. Two strains of biotype 1B gave bands corresponding to four repeat copies. Two strains of biotype 2 presented, respectively, eleven and four copies of the repeat. Finally, http://mic.sgmjournals.org
Fig. 2. Agarose gel electrophoresis of the amplification products obtained from strains representative of the different alleles observed in the sampled clinical strains. Lanes: 1, strain YHV49 (14 repeats); 2, strain YHV46 (13 repeats); 3, strain YHV28 (11 repeats); 4, strain YHV52 (9 repeats); 5, strain YHV41 (8 repeats); 6, strain YHV40 (7 repeats); 7, strain YHV25 (5 repeats); 8, strain YH16 (4 repeats); m, 50 bp DNA ladder. 201
I. de Benito and others
m
1
2
3
4
5
6
7
8
m
m
1
2
3
4
5
6
7
8
m
kb
bp 242.5
350
100
Fig. 3. Analysis of the status of the polymorphic locus in Y. enterocolitica strains belonging to different biotypes. Lanes: 1, P1403 (Biotype 1A); 2, P219 (Biotype 1A); 3, WA (Biotype 1B); 4, 13514 (Biotype 1B); 5, My79b (Biotype 2); 6, IP97 (Biotype 2); 7, IP4124 (Biotype 4, serotype 3, Lysotype IXb); 8, IP22274 (Biotype 4, serotype 3, Lysotype IXb); m, 50 bp DNA ladder.
produced the complete sequence of Yersinia pestis strains CO-92 (Parkhill et al., 2001) and KIM (Deng et al., 2002). The genomic sequences of Y. pseudotuberculosis IP 32953 and Y. enterocolitica 8081 are close to completion at the time of writing. The sequence of the amplification product of Y. enterocolitica strain Y56 was compared with those from the above-mentioned Yersinia strains. A homologous sequence was found in Y. enterocolitica 8081 containing ten copies of the hexanucleotide repeat. The sequence identified formed part of an ORF encoding a polypeptide of 528 aa (Orf528 was located between coordinates 2 205 137 and 2 203 551 of the Y. enterocolitica 8081 sequence). This protein showed ten copies of a prolinealanine (PA) repeat at its carboxy terminus (Fig. 1c). The two sequenced Y. pestis strains also contained sequences encoding polypeptides homologous to Orf528. In these cases the protein was 514 aa long and contained only two PA copies (Fig. 1c). Y. pseudotuberculosis apparently does not contain a gene homologous to orf528. We investigated the occurrence of other repeated CCAGCA tracts (n>2) in the Y. enterocolitica 8081 unpublished sequence. An ORF of 1320 nt (1 107 347–1 106 028), 202
48.5
Fig. 4. Analysis of the whole chromosome of isolates representing the eight different orf528 alleles digested with NotI and separated by PFGE. Lanes: 1, YHV1 (13 copies); 2, YHV2 (14 copies); 3, YHV5 (11 copies); 4, YHV6 (8 copies); 5, YHV16 (4 copies); 6, YHV25 (5 copies); 7, YHV38 (9 copies); 8, YHV43 (7 copies); m, size marker (l ladder PFG marker; New England Biolabs) contains multimers of l (unit size, 48?5 kb).
encoding a polypeptide of 439 aa containing a relaxase domain, contained 23 repeated CCAGCA copies, which corresponded to 23 PA repeats in the polypeptide. A further case of a gene product containing multiple PA repeats was observed in an uncharacterized membrane protein (3 035 775–3 036 578), which contained 11 PA repeats in its carboxy end. The proline residue in this locus was encoded by a CCT codon, resulting in a CCTGCA hexanucleotide repeat. These two loci were either absent or non-repeated in Y. pestis and Y. pseudotuberculosis. Analysis by PFGE To determine whether the isolates with different numbers of CCAGCA units in the orf528 locus, belonging to the same serotype and biotype, could be distinguished by other molecular typing methods, we analysed by PFGE their chromosomes digested with the restriction endonuclease NotI. Eight strains representative of the different orf528 alleles were found to produce four different restriction patterns (Fig. 4). This indicated that the analysis of the CCAGCA repeat in orf528 had more discriminating power than PFGE. Furthermore, we observed that eight strains with the same PFGE pattern presented seven different orf528 alleles (data not shown), demonstrating that the two analyses were independent and that the study of the orf528 locus can improve greatly the discriminatory ability of the PFGE.
DISCUSSION In this study we describe a locus in Y. enterocolitica Y56 containing 15 tandem copies of the hexanucleotide Microbiology 150
Polymorphic tandem repeat in Y. enterocolitica
CCAGCA. Analysis of this locus in a collection of 55 clinical isolates of Y. enterocolitica showed the existence of eight different alleles, which were easily differentiated by agarose gel electrophoresis (Fig. 2). Nucleotide sequence determination showed that the locus was polymorphic due to the differences in the copy number of the hexanucleotide CCAGCA. The VNTR region found in Y. enterocolitica Y56 was also identified in the unfinished sequence of Y. enterocolitica 8081, located at the 39 side of an ORF named orf528. According to this, the Y. enterocolitica strains analysed containing different numbers of the CCAGCA repeats, carry allelic forms of orf528 and they will presumably produce proteins homologous to Orf528 from strain 8081 with different numbers of the PA dipeptide at their carboxy ends (Fig. 1c). Orf528 is conserved in Escherichia coli (YdgA, SwissProt accession no. P77804) and it has been annotated as a GTP-binding protein, but its function is still unknown. Polymorphisms in VNTR loci inside an ORF usually lead to a change in protein function. Group B streptococci escape from phagocytosis by this mechanism (Madoff et al., 1996) and Listeria controls actin-mediated intracellular motility (Smith et al., 1996). Even more spectacular is the expansion of a TCT triplet in the ahpC gene of E. coli resulting in the conversion of a peroxiredoxin into a disulfide reductase (Ritz et al., 2001). The different alleles of orf528 will probably produce proteins with some functional difference. We do not have any evidence indicating that this polymorphism could be subjected to environmental selection, but this aspect should be further investigated to validate the study of this locus as an epidemiological tool (van Belkum, 1999). Further analysis of the sequences around the orf528 locus, revealed several differences in gene order between Y. enterocolitica strains Y56 and 8081, suggesting that this locus can represent a hotspot for gene rearrangement in this species. Analysis of the locus in our collection of clinical isolates and other Yersinia strains indicated that at least 14 alleles of orf528 were possible, containing from two to fifteen hexanucleotide copies. The already characterized strain Y56, with 15 repeat copies showed the more expanded locus. Y. enterocolitica strain 8081 contained 10 copies and the two sequenced strains of Y. pestis contained two copies of the repeat. Only eight of these possible alleles were represented in the sample analysed (Table 1). The allele with 13 copies was the most frequent among our clinical Y. enterocolitica isolates (15 out of 55 YHV strains analysed), followed in frequency by the eight- and seven-repeat alleles (nine occurrences each). The 55 strains used in this study were characterized after the classical serogrouping and biotyping scheme of Y. enterocolitica and were found to belong (except one) to serotype O : 3 biotype 4. All were pathogenic as they produced the Myf factor. This result agrees with previous data and http://mic.sgmjournals.org
confirmed that the Y. enterocolica strains of biotype 4, serotype O : 3 were prevalent in our area. Our results showed that the analysis of the VNTR locus described here in a sample of 54 Y. enterocolica clinical strains, which were identical according to classical typing schemes, allowed the identification of eight different alleles. Furthermore, the analysis of the locus in Y. enterocolitica strains belonging to other biotypes and serotypes showed that the locus was always present and that it could also be polymorphic. The analysis of the completed sequence of Y. enterocolitica 8081 revealed the existence of a second locus with 23 CCAGCA repeats. This locus is probably also polymorphic and its analysis, in combination with the VNTR in orf528 could be exploited for Y. enterocolitica typing, as has been recently done with a VNTR found at eight different loci in the genus Brucella (Bricker et al., 2003). A sample of the strains representing each one of the different alleles was also analysed by PFGE after genomic digestion with NotI. The eight strains presented four different pulsotypes, indicating that the two systems of analysis were independent and showing that the study of the VNTR in orf528 had a discriminatory power greater than PFGE. Furthermore, we have observed that eight strains belonging to the same biotype and serotype and presenting the same pulsotype after digestion with NotI, presented seven different allelic forms of orf528. In summary, we conclude that the analysis of the VNTR locus described herein may provide a valuable method to discriminate among strains of Y. enterocolitica that seem to be identical according to other typing methods.
ACKNOWLEDGEMENTS We are indebted to Dr Elisabeth Carniel from the Yersinia reference Centre at the Institute Pasteur and to Dr Ramo´n Dı´az from Clı´nica Universitaria-Universidad de Navarra for typing our isolates and providing strains. We are also grateful to the Sanger Centre and the Biotechnology Research Program of the Lawrence Livermore National Laboratory for allowing access to unfinished genome sequences. This work was supported by grant number 99/1224 from the Fondo de Investigaciones Sanitarias of the Spanish Ministry of Health.
REFERENCES Adair, D. M., Worsham, P. L., Hill, K. K., Klevytska, A. M., Jackson, P. J., Friedlander, A. M. & Keim, P. (2000). Diversity in a variable-
number tandem repeat from Yersinia pestis. J Clin Microbiol 38, 1516–1519. Bayliss, C. D., Field, D. & Moxon, E. R. (2001). The simple sequence
contingency loci of Haemophilus influenzae and Neisseria meningitidis. J Clin Invest 107, 657–662. Bricker, B. J., Ewalt, D. R. & Halling, S. M. (2003). Brucella ‘HOOF-
Prints’: strain typing by multi-locus analysis of variable number tandem repeats (VNTRs). BMC Microbiol 3, 15. Deng, W., Burland, V., Plunkett, G., 3rd & 18 other authors (2002). Genome sequence of Yersinia pestis KIM. J Bacteriol 184,
4601–4611. 203
I. de Benito and others Frothingham, R. & Meeker-O’Connell, W. A. (1998). Genetic
Ritz, D., Lim, J., Reynolds, C. M., Poole, L. B. & Beckwith, J. (2001).
diversity in the Mycobacterium tuberculosis complex based on variable numbers of tandem DNA repeats. Microbiology 144, 1189–1196.
Conversion of a peroxiredoxin into a disulfide reductase by a triplet repeat expansion. Science 294, 158–160.
Hallanvuo, S., Skurnik, M., Asplund, K. & Siitonen, A. (2002).
Characterisation of pathogenic Yersinia enterocolitica serogroups by pulsed-field gel electrophoresis of genomic NotI restriction fragments. J Med Microbiol 41, 329–338.
Detection of a novel repeated sequence useful for epidemiological typing of pathogenic Yersinia enterocolitica. Int J Med Microbiol 292, 215–225.
Saken, E., Roggenkamp, A., Aleksic, S. & Heesemann, J. (1994).
Iteman, I., Guiyoule, A. & Carniel, E. (1996). Comparison of three
Saunders, N. J., Jeffries, A. C., Peden, J. F., Hood, D. W., Tettelin, H., Rappuoli, R. & Moxon, E. R. (2000). Repeat-associated phase variable
molecular methods for typing and subtyping pathogenic Yersinia enterocolitica strains. J Med Microbiol 45, 48–56.
genes in the complete genome sequence of Neisseria meningitidis strain MC58. Mol Microbiol 37, 207–215.
Kashi, Y., King, D. & Soller, M. (1997). Simple sequence repeats
Seoane, A. & Garcı´a Lobo, J. M. (1991). Cloning of chromosomal b-lactamase genes from Yersinia enterocolitica. J Gen Microbiol 137,
as a source of quantitative genetic variation. Trends Genet 13, 74–78. Keim, P., Price, L. B., Klevytska, A. M., Smith, K. L., Schupp, J. M., Okinaka, R., Jackson, P. J. & Hugh-Jones, M. E. (2000). Multiple-
locus variable-number tandem repeat analysis reveals genetic relationships within Bacillus anthracis. J Bacteriol 182, 2928– 2936. Leyva, J., Olmo, C., Garcı´a-Jalo´n, I., Irigoyen, A., Oteiza, C., Dorronsoro, I. & Dı´az, R. (1995). Coagglutination with antisera to
141–146. Seoane, A. & Garcı´a Lobo, J. M. (2000). Identification of a
streptogramin A acetyltransferase gene in the chromosome of Yersinia enterocolitica. Antimicrob Agents Chemother 44, 905–909. Smith, G. A., Theriot, J. A. & Portnoy, D. A. (1996). The tandem
repeat domain in the Listeria monocytogenes ActA protein controls the rate of actin-based motility, the percentage of moving bacteria, and the localization of vasodilator-stimulated phosphoprotein and profilin. J Cell Biol 135, 647–660.
MyfA and PsaA to distinguish Yersinia enterocolitica from Yersinia pseudotuberculosis pathogenic isolates. Contrib Microbiol Immunol 13, 158–164.
van Belkum, A. (1999). The role of short sequence repeats in
Madoff, L. C., Michel, J. L., Gong, E. W., Kling, D. E. & Kasper, D. L. (1996). Group B streptococci escape host immunity by deletion of
van Belkum, A., Scherer, S., van Leeuwen, W., Willemse, D., van Alphen, L. & Verbrugh, H. (1997). Variable number of tandem repeats in clinical
tandem repeat elements of the alpha C protein. Proc Natl Acad Sci U S A 93, 4131–4136.
Verkerk, A. J., Pieretti, M., Sutcliffe, J. S. & 7 other authors (1991).
Parkhill, J., Wren, B. W., Thomson, N. R. & 26 other authors (2001).
Genome sequence of Yersinia pestis, the causative agent of plague. Nature 413, 523–527.
204
epidemiologic typing. Curr Opin Microbiol 2, 306–311.
strains of Haemophilus influenzae. Infect Immun 65, 5017–5027. Identification of a gene (FMR-1) containing a CGG repeat coincident with a breakpoint cluster region exhibiting length variation in fragile X syndrome. Cell 65, 905–914.
Microbiology 150
