JOURNAL OF CLINICAL MICROBIOLOGY, Dec. 2008, p. 3935–3940 Vol. 46, No. 120095-1137/08/$08.00�0 doi:10.1128/JCM.00464-08Copyright © 2008, American Society for Microbiology. All Rights Reserved.

Evaluation of a Multilocus Variable-Number Tandem-Repeat AnalysisScheme for Typing Human Brucella Isolates in a Region of

Brucellosis Endemicity�†Mireille M. Kattar,1,2* Rola F. Jaafar,3 George F. Araj,1,2 Philippe Le Flèche,4,5 Ghassan M. Matar,3

Roland Abi Rached,1,2 Simon Khalife,1,2 and Gilles Vergnaud4,6*Department of Pathology and Laboratory Medicine1 and Department of Microbiology and Immunology,3 American University of

Beirut, Beirut, Lebanon; World Health Organization Regional Collaborating Center on Human Brucellosis, Beirut, Lebanon2;Université Paris-Sud 11, CNRS, UMR8621, Institut de Génétique et Microbiologie, Orsay 91405, France4; Department of

Analytical Microbiology, Centre d’Etudes du Bouchet, Vert le Petit 91710, France5; and DGA/D4S-Mission pour laRecherche et l’Innovation Scientifique, Bagneux 92220, France6

Received 9 March 2008/Returned for modification 6 July 2008/Accepted 30 September 2008

Brucellosis remains an important anthropozoonosis worldwide. Brucella species are genetically homoge-neous, and thus, the typing of Brucella species for epidemiological purposes by conventional molecular typingmethods has remained elusive. Although many methods could segregate isolates into the phylogeneticallyrecognized taxa, limited within-species genetic diversity has been identified. Recently, multilocus variable-number tandem-repeat analysis (MLVA) was found to have a high degree of resolution when it was applied tocollections of Brucella isolates from geographically widespread locations, and an assay comprising 16 such loci(MLVA-16) was proposed. This scheme includes eight minisatellite loci (panel 1) and eight microsatellites(panel 2, which is subdivided into panels 2A and 2B). The utility of MLVA-16 for the subtyping of humanBrucella isolates from geographically restricted regions needs to be further evaluated, and genotyping data-bases with worldwide coverage must be progressively established. In the present study, MLVA-16 was appliedto the typing of 42 human Brucella isolates obtained from 41 patients recovered from 2002 to 2006 at atertiary-care center in Lebanon. All isolates were identified as Brucella melitensis by MLVA-16 and were foundto be closely related to B. melitensis isolates from neighboring countries in the Middle East when theirgenotypes were queried against those in the web-based Brucella2007 MLVA database (http://mlva.u-psud.fr/).Panel 2B, which comprised the most variable loci, displayed a very high discriminatory power, while panels 1and 2A showed limited diversity. The most frequent genotype comprised seven isolates obtained over 7 weeksin 2002, demonstrating an outbreak from a common source. Two isolates obtained from one patient 5 monthsapart comprised another genotype, indicating relapsing disease. These findings confirm that MLVA-16 has agood discriminatory power for species determination, typing of B. melitensis isolates, and inferring theirgeographical origin. Abbreviated panel 2B could be used as a short-term epidemiological tool in a small regionof endemicity.

Brucellae are facultative intracellular pathogens that infect awide variety of animal species and humans. Brucellosis is themost common anthropozoonosis, with more than 500,000 casesreported annually worldwide (28). The genus Brucella cur-rently encompasses nine recognized species (seven terrestrialand two marine mammal species) that display animal hostspecificity, among which three present veterinary and publichealth concerns (11, 27, 31). Brucella melitensis predominantlyinfects sheep and goats, B. abortus infects cattle, and B. suisinfects swine and a range of wild animals; but cross-infection of

other mammalian species, including humans, may occur (10).Animal brucellosis causes abortion and infertility in livestock(cattle, goats, and sheep), resulting in serious economic losses.Human brucellosis is a subacute or chronic febrile illness thatcan involve multiple organs and that can result in a wide varietyof manifestations and significant morbidity if the diagnosis isoverlooked and treatment is not promptly initiated (5). Animalbrucellosis has been successfully eradicated in most developedcountries, resulting in the virtual disappearance of the humandisease in North America, Northern Europe, and NorthwestAsia, where most cases are now due to either travel to areas ofendemicity, accidental laboratory exposure, or occasionally,exposure to wild animals (10). However, human brucellosisremains endemic and a major public health problem in manydeveloping countries and some developed countries in LatinAmerica, Southern Europe, Africa, Southeast Asia, and theMiddle East (28). The most common cause of human brucel-losis is B. melitensis, with B. abortus and B. suis accounting forsmaller proportions of cases (19, 28). Brucella spp. are consid-ered potential military, agricultural, and civilian category Bbiological threat agents (30) due to their relative ease of dis-

* Corresponding author. Mailing address for Mireille M. Kattar:Department of Pathology and Laboratory Medicine, American Uni-versity of Beirut Medical Center, Cairo Street, Beirut, Lebanon.Phone: 961-1-374374, ext. 5175. Fax: 961-1-370845. E-mail:mireillekattar@gmail.com. Mailing address for Gilles Vergnaud:Université Paris-Sud 11, CNRS, UMR8621, Institut de Génétique etMicrobiologie, Orsay 91405, France. Phone: 33-1-69156208. Fax: 33-1-69156678. E-mail: gilles.vergnaud@u-psud.fr.

† Supplemental material for this article may be found at http://jcm.asm.org/.

� Published ahead of print on 15 October 2008.

3935

on March 31, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

https://crossmark.crossref.org/dialog/?doi=10.1128/JCM.00464-08&domain=pdf&date_stamp=2008-12-01http://jcm.asm.org/

semination and costly eradication if they were spread in apossible bioterrorism event.

Species identification and subtyping of Brucella culture iso-lates is important for epidemiologic surveillance; investigationof outbreaks in regions of both endemicity and nonendemicity;and distinguishing cases of human reinfection from relapse,thereby influencing clinical therapeutic decisions (3).

Terrestrial Brucella spp. are homogeneous and harbor�80% interspecies homology by DNA-DNA hybridizationstudies (31, 38), identical 16S rRNA sequences (15), and�98% sequence similarity by comparative genomics (18, 29).Molecular methods commonly used for the subtyping of iso-lates of other bacterial species, such as multilocus enzymeelectrophoresis (12), pulsed-field gel electrophoresis (4), ran-dom amplified polymorphic DNA analysis (34), enterobacte-rial repetitive intergenic consensus sequence PCR (26, 35),repetitive intergenic palindromic sequence PCR (25), ampli-fied fragment length polymorphism analysis (42), and mono-locus (such as omp2a and omp2b) sequence analysis (9) ormultilocus sequence typing (40), are able to segregate Brucellaisolates into the recognized species and certain species biovarsat best. However, these methods are not sufficiently discrimi-natory for the routine subtyping of isolates for epidemiologicaltrace-back purposes.

Tandemly repeated sequences used in forensic investiga-tions have also been found in bacteria (for a review, seereference 39). Minisatellites have repeat unit sizes of 9 bp orgreater, and microsatellites have repeat unit sizes of up to 8bp (39). Combinations of minisatellite and microsatelliterepeats in multilocus variable-number tandem-repeat anal-ysis (MLVA) have proven highly discriminatory in the epide-miological subtyping of isolates belonging to monomorphicbacterial species, such as Bacillus anthracis, Yersinia pestis (8,23), and more recently, Brucella spp. In Brucella, MLVAschemes with 21 loci (MLVA-21) and MLVA-16 (1, 22, 41)that use a combination of repeat markers distributed across theBrucella genome were able to distinguish isolates of Brucellaspp. of widespread temporal and geographical origins or ofvery close origins (13). Furthermore, with some rare excep-tions, which might be due to a lack of resolution and incorrectclustering produced by the biotyping methods, the isolatesformed MLVA groups that corresponded to the known Bru-cella species. A study of 128 human isolates of B. melitensis byMLVA-16 identified the Americas, West Mediterranean, andEast Mediterranean groups and detected remarkable degreesof diversity within each group (1). The relative stability ofMLVA loci has been estimated directly by repetitive subcul-turing of a few strains in the laboratory (41) and by the typingof B. melitensis isolates of the Rev.1 vaccine strain obtainedfrom different sources (13), so that the most recent MLVAdata analyses have started to assign different weights to differ-ent markers, taking into account the mutation rate and theassociated levels of homoplasy (1). The MLVA schemes thathave been devised are very promising, but the strength of atyping method depends as much on the technique itself as onthe associated genotyping database. These typing assays stillneed to be applied to larger numbers of isolates from as manycountries as possible to validate their reliability in investigatingstrain relatedness both in countries of endemicity, especiallywhere efforts to eradicate human disease may be attempted,

and in brucellosis-free countries where a cluster of infectionsmay be suspected to be due to a bioterrorism event. Theprimary purpose of high-resolution typing in a format compat-ible with the production of large-scale international databasesis the long-term surveillance of infectious diseases and theearly detection of abnormal events, including temporal or geo-graphical shifts in the population structure, e.g., the emergenceand spread of new clones or lineages. The current Brucel-la2007 MLVA database, hosted at http://mlva.u-psud.fr, con-tains data derived from more than 500 animal and humanBrucella isolates. It is expected that in the future, similar andcompatible databases with data for isolates from wider geo-graphical origins will be developed and made jointly and freelyaccessible to querying via the use of web-based analysis tools(16).

In this study, we applied the MLVA-16 scheme (1, 22) to aseries of clinical isolates of Brucella accrued over a 5-yearperiod at a single tertiary-care center in Lebanon, a smallcountry where human brucellosis is endemic. The genotypes ofthese isolates were compared to the genotypes previously de-termined by MLVA-16 typing and are included in the Bru-cella2007 MLVA database.

(This study was presented in part at the 107th GeneralMeeting of the American Society for Microbiology, Toronto,Ontario, Canada, 2007.)

MATERIALS AND METHODS

Bacterial strains. Forty-two clinical Brucella isolates (isolates AUB-BRUP-S1to AUB-BRUP-S42 [isolates S1 to S42, respectively]) obtained from clinicalspecimens from 41 different patients submitted to the clinical microbiologylaboratory at the American University of Beirut Medical Center from February2002 to December 2006 were investigated. These represented all Brucella isolatescollected during that time period. Thirty-nine of the isolates were isolated fromblood, and three were from tissue specimens. Two of the 39 blood isolates wereobtained from cultures of blood from one patient 5 months apart. All isolateswere identified as Brucella species on the basis of colonial morphology, positiveoxidase and catalase tests, positive agglutination with brucella-specific antiserum,and positive amplification by real-time PCR, as described previously (21). Inaddition, all patients from whom these isolates were obtained had positiveBrucella serology titers by Wright’s standard tube agglutination and the anti-human globulin Coombs test (5).

Brucella MLVA-16 genotyping scheme. MLVA was performed with all 42isolates according to the scheme initially proposed by Le Flèche et al. (22), whichincludes 15 tandem-repeat loci (MLVA-15), and modified by Al-Dahouk et al. toinclude 1 additional locus, bruce19 (MLVA-16) (1). PCR amplification of eightminisatellite loci in panel 1 (the bruce06, bruce08, bruce11, bruce12, bruce42,bruce43, bruce45, and bruce55 loci), three microsatellite loci in panel 2A (thebruce18, bruce19, and bruce21 loci), and five microsatellite loci in panel 2B(the bruce04, bruce07, bruce09, bruce16, and bruce30 loci) was carried out bya protocol similar to the one described previously (1).

DNA preparation. DNA was extracted from one loopful of bacterial cellsgrown for 48 h on chocolate agar, and single colonies were isolated by using thetissue protocol of the QIAamp DNA minikit (Qiagen, Hilden, Germany). DNAconcentrations were measured by UV spectrophotometry (Shimadzu, Japan).

PCR amplification. PCR amplification was performed in a total volume of 15�l containing 5 ng of DNA, 1� PCR buffer, 1 U of Taq DNA polymerase, 200�M of each deoxynucleoside triphosphate, 0.3 mM each flanking primer, and 1M of betaine (catalog no. B2629; Sigma-Aldrich). The PCR cycling parameterswere as follows: initial denaturation at 96°C for 5 min, followed by 30 cycles ofdenaturation at 96°C for 30 s, primer annealing at 60°C for 30 s, and extensionat 70°C for 1 min, with a final extension step at 70°C for 5 min. For detection ofthe PCR products, 5 �l of each of the PCR amplification products was subjectedto gel electrophoresis with a 2% 3:1 HRB agarose gel (catalog no. E776; Am-resco) for panel 1 and a 3% HRB agarose gel for panel 2 in 0.5� Tris-borate-EDTA buffer. Electrophoresis was carried until the bromophenol blue had runat least 20 cm. The size markers used were a 100-bp ladder (EZ Load 100-bp

3936 KATTAR ET AL. J. CLIN. MICROBIOL.

on March 31, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

PCR molecular ruler; Bio-Rad) for panel 1 and a 20-bp ladder (EZ Load 20-bpmolecular ruler; Bio-Rad) for panel 2. Gels that had been stained with ethidiumbromide (0.5 �g/ml) were digitally acquired with a gel documentation system(DigiDoc-It; UVP, Upland, CA).

Analysis of MLVA data. Minisatellite repeat numbers (panel 1 [loci bruce06,bruce08, bruce11, bruce12, bruce42, bruce43, bruce45, and bruce55]) and mic-rosatellite repeat numbers (panel 2A [loci bruce18, bruce19, and bruce21] andpanel 2B [loci bruce04, bruce07, bruce09, bruce16, and bruce30) were calculatedon the basis of the band sizes after the gel images were normalized by use of theBionumerics (version 5.1) software package (Applied Maths, Belgium) and thepreviously published allele numbering convention (22). The reproducibilities ofthe MLVA profiles were assessed by comparing the typing results for experi-ments performed independently at two laboratory testing sites in Lebanon andFrance. The similarity matrix for each panel was calculated with Bionumericssoftware by using the repeat number at each locus as a character type with thecategorical similarity coefficient. This categorical approach considers all alleles ateach locus and within each panel to be equally distant (37). For MLVA-16clustering analysis, an aggregated similarity matrix was produced by using acomposite data set and giving weights of 20, 5, and 1 to panels 1, 2A, and 2B,respectively, according to the principles reported previously (1). Weights wereassigned to the individual loci in each panel because all loci do not evolve at thesame speed but have different mutation rates, as determined empirically. Clus-tering was done by use of the unweighted paired group method for arithmeticaverages algorithm. Genetic diversity was determined by calculating the Hunter-Gaston (or Simpson’s) diversity index (HGDI) of each MLVA marker and forthe combination of markers for each of panel 1, panel 2A, panel 2B, panel 2(panel 2A plus panel 2B), and panels 1 and 2 (20). The 95% confidence intervalsof the HGDIs for indices of �0.85 were calculated by the method of Grundmannet al. (17). The MLVA-16 genotypes, defined by the combination of alleles, ofthe 42 isolates were compared to the corresponding data obtained for thereference strains and field isolates investigated previously (1, 13, 14, 22, 31, 32,33). The present and previous data have been incorporated into a single databasedesignated Brucella2007, which can be queried at http://mlva.u-psud.fr by the useof web-based analysis tools.

RESULTS AND DISCUSSION

The purpose of this study was to evaluate the polymorphismsof the MLVA-16 loci in a series of Brucella isolates and therelatedness of the strains from a geographically restricted re-gion where Brucella is endemic. Furthermore, we sought tovalidate the established Brucella2007 genotyping database as-sociated with MLVA-16 as a reliable investigation tool in thiscontext. In the present study, the MLVA-16 scheme was ap-plied to type 42 human isolates of Brucella obtained over a5-year period at the largest tertiary-care center in Lebanon.Lebanon is a small country in the eastern Mediterranean re-gion with a geographical surface area of 10,452 km2, an esti-mated population of 4,000,000, and a reported incidence ofbrucellosis of 50 per 1 million inhabitants (28).

The typeability of the various MLVA markers in this seriesapproached 1 (37). Only the bruce04 locus in strain S29 andthe bruce09 locus in strain S19 could not be amplified, despitemultiple attempts. All other amplification reactions were suc-cessful. The reproducibility of MLVA-16 was excellent, as ex-periments performed in two different laboratories in Lebanonand France by use of the same protocol and with the resultsinterpreted separately in a blinded fashion yielded 100% con-cordant results. This is not surprising for a sequence-basedapproach and was recently confirmed by a ring trial for thepanel 1 loci among 15 Brucella reference laboratories (http://mlva.u-psud.fr/BRUCELLA/). In that trial, the few discrep-ancies in typing results observed were due to errors in allelecalling (24a). The 42 isolates in the series used in this studyformed a homogeneous group for which the differences did notexceed 15% (Fig. 1). Limited diversity was observed with the

panel 1 and panel 2A loci (Table 1), as would be expected forstrains originating from a very restricted geographical area andby the use of markers that are the most phylogenetically infor-mative. MLVA-16 yielded a total of 29 genotypes with sevenclusters and 22 singleton genotypes. Previous work identified atotal of 68, 52, and more than 300 genotypes for panel 1, panel2A, and panel 2B, respectively (1, 13, 14, 22, 31, 32, 33) (seeFig. S1 in the supplemental material). Twenty-six panel 1 ge-notypes, numbered 41 to 66 (1, 22), corresponded to B.melitensis. No new panel 1 genotypes were identified in thepresent study. Thirty-seven isolates shared panel 1 genotype43. Panel 1 genotypes 42, 44, and 57 were observed in a singleisolate each; and genotype 60 was observed in isolates S18 andS24, which were derived from the same patient. Therefore,only 5 different panel 1 genotypes were observed among the 26panel 1 B. melitensis genotypes observed so far. Accordingly,the bruce06, bruce11, bruce12, and bruce55 loci from panel 1and the bruce21 locus from panel 2 showed no diversity amongthe 42 isolates (HGDI � 0). All five panel 1 genotypes repre-sented in the present study correspond to the East Mediterra-nean group proposed previously (1) (see Fig. S1 in the supple-mental material). The panel 2B markers displayed the mostdiversity in the present collection. This was as expected, giventhe high mutation rate of some panel 2B markers (13), withbruce04 and bruce16 exhibiting the highest HGDIs, followedby bruce30 (HGDIs � 0.7). Bruce07 and bruce09 in panel 2Bhad HGDIs of �0.5. Panel 2B and panels 2A and 2B yielded27 and 28 genotypes, respectively, and had similar HGDIs. Thecombined panel 1 and 2 markers yielded 29 genotypes and anHGDI that was only slightly higher than that for panel 2Balone. The HGDIs of panel 2B, panels 2A and 2B, and panels1 and 2 were all �0.95, which is considered optimal for a typingsystem (37). These data might indicate that in the setting of alocal outbreak investigation, panel 1 and panel 2A could beomitted and the highly polymorphic panel 2B would be suffi-cient. However, typing with panel 1 and 2A is necessary beforetyping with panel 2B is used in order to check that the isolateis indeed of the expected local type, as it is important to keepin mind that the HGDI value is not a sufficient indicator for theselection of tandem repeat loci and validation of an MLVAassay. Three of the five panel 2B loci (bruce04, bruce09,bruce30) are part of the hypervariable octameric oligonucleo-tide fingerprint variable-number tandem repeats (HOOF-Prints) MLVA panel of loci that have been described byBricker et al. (6, 7) and that have proven to be highly mutablein terrestrial Brucella spp. (6, 7, 36, 41). The most frequentgenotype (largest cluster) comprised seven isolates (isolates S3to S9) obtained over 7 weeks in the late spring and earlysummer of 2002, demonstrating an outbreak from a commonsource. Another genotype was shared by isolates S18 and S24,obtained 5 months apart from one patient who had no at-riskoccupational history and who experienced a recurrence of hisfebrile illness. The two isolates from this patient differed fromthe rest of the isolates at the bruce45 minisatellite locus frompanel 1 (panel 1 genotype 60, two versus three repeats), indi-cating relapsing disease in this patient. Two isolates obtained 2months apart (isolates S37 and S41) and four pairs obtainedover 1 year apart (isolates S25 and S30, S12 and S39, S2 andS23, and S28 and S42) had identical genotypes. These fiveseemingly unrelated pairs may represent either epidemiologi-

VOL. 46, 2008 MLVA TYPING OF HUMAN BRUCELLA ISOLATES 3937

on March 31, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

cally unrelated isolates with homoplasy at MLVA-16 loci (mostlikely panel 2B) or persistent circulating strains causing spo-radic infections. All other isolates had distinct genotypes re-flecting sporadic cases. In a previous study (1), two pairs ofepidemiologically related human B. melitensis isolates in Ger-many displayed identical MLVA-16 profiles. One pair of iso-lates was obtained from one patient: one isolate was obtainedbefore treatment for brucellosis and one isolate was obtained3 months later when he experienced a relapse related to inad-equate therapy. Another pair of isolates came from a marriedcouple who visited Turkey, where they contracted their infec-tion from the consumption of unpasteurized cheese. Threeisolates from a small outbreak of B. suis infections in pigs inSpain also shared the same MLVA-16 genotype (14). Theclustering of epidemiologically related isolates identified in thecurrent and previous studies support the use of MLVA-16 as avaluable tool for investigations of outbreaks of both humanand animal brucellosis.

In comparison to previous data and by the use of all 16 loci,

23 of the 29 genotypes obtained were unique (see Fig. S1 in thesupplemental material). Eight isolates had genotypes identicalto those of strains from Syria (n � 2), Turkey (n � 2), Israel (n �1), Lebanon (n � 1), and Germany (n � 2). The last twoisolates probably originated from Turkey (2). Thirty-one (31)isolates were single-locus or double-locus variants of closelyrelated B. melitensis isolates from neighboring countries, in-cluding Syria, Israel, and Turkey (see Fig. S1 in the supple-mental material). The unique fingerprinting profiles for mostof the B. melitensis isolates in this series and their close relat-edness by MLVA compared to the relatedness of other humanB. melitensis isolates (1) probably reflect microevolution due tofew mutational events among indigenous strains derived froma common ancestor rather than the introduction of foreignstrains from other areas, since most of these variations were inpanel 2B loci.

Although phenotypic biovar typing was not performed inthis study, previous studies have shown that the results of theMLVA-16 or the MLVA-21 typing schemes that have been

FIG. 1. Dendrogram based on the MLVA-16 genotyping assay showing the relationships of the 42 human isolates of Brucella. Under panels1, 2A, and 2B are shown the individual MLVA-16 loci and the number of tandem repeats for each isolate. Key, the serial number for the isolatein the Brucella2007 MLVA database (http://mlva-u-psud.fr/) (see Fig. S1 in the supplemental material); strain, the local strain designation; panels1, 2A, and 2B, genotypes corresponding to each isolate in the database for each set of loci; date isolation, the dates (day/month/year) that sampleswere received in the laboratory for culturing. The information under comments provides the specimen source-patient gender (M, male; F,female)/patient age (in years).

3938 KATTAR ET AL. J. CLIN. MICROBIOL.

on March 31, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

devised (1, 41) show no correlation with those of biovar typingfor B. melitensis. Our isolates were closely related to B. meliten-sis isolates from various countries that belonged to all threebiovars (biovars 1, 2, and 3). Neither MLVA nor multilocussequence typing distinguished the B. melitensis biovars, whichare mere serotypes (1, 40, 41). This indicates that variable-number tandem-repeat loci and the single-nucleotide polymor-phisms which provide congruent data evolve independently ofthe putative genetic determinants for these biovars. This is alsothe case for B. abortus, but to a lesser extent. In contrast, bothMLVA-16 and MLVA-21 separate B. suis biovar 2 from theother B. suis biovars, biovars 1, 3, and 4 (see Fig. S1 in thesupplemental material) (14, 41).

Both MLVA-16 and MLVA-21 segregated Brucella strainsinto groups corresponding to the recognized phylogeneticlineages in the genus Brucella (22, 41). When it was appliedto more than 500 strains of terrestrial Brucella, panel 1 inMLVA-16 subdivided these isolates into 65 genotypes thatwere, with one exception, exclusive to each of the Brucellaspecies (see Fig. S1 in the supplemental material). Similarly, byMLVA-21, six minisatellites were found to be sufficient forspecies designation (41). In both schemes, B. canis clusteredwith B. suis biovar 4. These data suggest that MLVA is agenerally useful method for primary species assignment of

terrestrial Brucella spp., especially in regions of endemicitywhere B. melitensis, B. abortus, and B. suis are the predominantcauses of both human and animal brucellosis.

The diversity of the panel 1 loci and most panel 2 loci in ourseries parallels that found in a study of 24 human B. melitensisisolates from Sicily in Italy (24). The two studies are compa-rable in that they focused on a restricted geographic area. Twopanel 1 genotypes (genotypes 49 and 51) were observed in theSicilian study, and both were from the West Mediterraneangroup (see Fig. S1 in the supplemental material). We foundhigher levels of diversity for bruce30 (panel 2B) and bruce19(panel 2A) and a lower level of diversity for bruce08 (panel 1).The Sicilian isolates formed a homogeneous group that couldbe distinguished from the East Mediterranean group at threeof the panel 1 minisatellite loci. These data illustrate the use-fulness of panel 1 not only for the species identification ofBrucella isolates but also for inference of their geographicalorigins. In contrast, panel 2 markers afford a higher discrimi-natory power for investigation of strain relatedness in regionsof endemicity. An abbreviated panel consisting of five micro-satellite loci, panel 2B, can be used as a practical short-termepidemiological tool in this setting, but the species cannot bededuced from the data produced by this panel only. Con-versely, in regions of nonendemicity or whenever the introduc-tion of an exogenous strain might be suspected, all loci shouldbe used to pinpoint the region where the isolates may haveoriginated. For that reason, there is a need to build expandedsearchable and shared international databases of MLVA fin-gerprints to enable such comparisons. The currently existingBrucella2007 MLVA database (http://mlva.u-psud.fr) providesone constituent within the framework of such a concertedeffort (16).

In summary, MLVA-16 displayed a good discriminatorypower for the typing of B. melitensis isolates from a smallregion of endemicity, and either panel 2 or panels 1 and 2 maybe used as epidemiological tools for the resolution of strainsand for distinguishing relapses from reinfections in patientswith brucellosis.

ACKNOWLEDGMENTS

This study was funded by the University Research Board and Med-ical Practice Plan, American University of Beirut, and by the FrenchDélégation Générale pour l’Armement.

REFERENCES

1. Al-Dahouk, S., P. Le Flèche, K. Nöckler, I. Jacques, M. Grayon, H. C. Scholz,H. Tomaso, G. Vergnaud, and H. Neubauer. 2007. Evaluation of BrucellaMLVA typing for human brucellosis. J. Microbiol. Methods 69:137–145.

2. Al-Dahouk, S., H. Neubauer, A. Hensel, I. Shöneberg, K. Nöckler, K. Alpers,H. Merzenich, K. Stark, and A. Jansen. 2007b. Changing epidemiology ofbrucellosis, Germany, 1962–2005. Emerg. Infect. Dis. 13:1895–1900.

3. Al-Dahouk, S., R. M. Hagen, K. Nöckler, H. Tomaso, M. Wittig, H. C. Scholz,G. Vergnaud, and H. Neubauer. 2005. Failure of a short-term antibiotictherapy for human brucellosis using ciprofloxacin. A study on in vitro sus-ceptibility of Brucella strains. Chemotherapy 511:352–356.

4. Allardet-Servent, A., G. Bourg, M. Ramuz, M. Pages, M. Bellis, and G.Roizes. 1988. DNA polymorphism in strains of the genus Brucella. J. Bacte-riol. 170:4603–4607.

5. Araj, G. F. 1999. Human brucellosis: a classical infectious disease withpersistent diagnostic challenges. Clin. Lab. Sci. 12:207–212.

6. Bricker, B. J., and D. R. Ewalt. 2005. Evaluation of HOOF-Print assay fortyping Brucella abortus strains isolated from cattle in the United States:results with four performance criteria. BMC Microbiol. 5:37.

7. Bricker, B. J., D. R. Ewalt, and S. M. Halling. 2003. Brucella ‘HOOF-Prints’:strain typing by multi-locus analysis of variable number of tandem repeats(VNTRs). BMC Microbiol. 3:15.

TABLE 1. Summary of MLVA-16 results for 42 human isolatesof B. melitensis

Panel and locusNo. of

alleles orgenotypes

No. ofrepeats HGDI

a 95% CIb

Panel 1bruce06 1 1 0bruce08 2 4, 5 0.047bruce11 1 3 0bruce12 1 13 0bruce42 2 2, 3 0.047bruce43 2 2, 3 0.047bruce45 2 2, 3 0.093bruce55 1 3 0

Total 5 0.225 NDc

Panel 2Abruce18 2 4, 5 0.047bruce19 3 20, 22, 25 0.181bruce21 1 8 0

Total 4 0.224 ND

Panel 2Bbruce04 8 2, 3, 4, 5, 6,

7, 8, 100.815

bruce07 5 3, 4, 5, 7, 8 0.37bruce09 6 3, 5, 7, 8, 9,

100.274

bruce16 7 4, 5, 6, 7, 8,9, 11

0.814

bruce30 5 3, 4, 5, 6, 9 0.723Total 27 0.957 0.918–0995

Panels 2A and 2B 28 0.958 0.919–0.996

MLVA-16 29 0.961 0.922–0.999

a HGDI, Hunter-Gaston diversity index.b CI, confidence interval.c ND, not done.

VOL. 46, 2008 MLVA TYPING OF HUMAN BRUCELLA ISOLATES 3939

on March 31, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

8. Ciammaruconi, A., S. Grassi, R. De Santis, G. Faggioni, V. Pittiglio, R.D’Amelio, A. Carattoli, A. Cassone, G. Vergnaud, and F. Lista. 2008. Field-able genotyping of Bacillus anthracis and Yersinia pestis based on 25-locimulti locus VNTR analysis. BMC Microbiol. 8:21.

9. Cloeckert, A., N. Vizcaino, J.-Y. Paquet, R. A. Bowden, and P. H. Elzer. 2002.Major outer membrane proteins of Brucella spp.: past, present and future.Vet. Microbiol. 90:229–247.

10. Corbel, M. J. 1997. Brucellosis: an overview. Emerg. Infect. Dis. 3:213–221.11. Foster, G., B. S. Osterman, J. Godfroid, I. Jacques, and A. Cloeckaert. 2007.

Brucella ceti sp. nov., and Brucella pinnipedialis sp. nov. for Brucella strainswith cetaceans and seals as their preferred hosts. Int. J. Syst. Evol Microbiol.57:2688–2693.

12. Gandara, B. A., L. Merino, and M. A. Rogel, and E. Martinez-Romero. 2001.Limited genetic diversity of Brucella spp. J. Clin. Microbiol. 39:235–240.

13. García-Yoldi, D., P. Le Flèche, C. M. Marín, M. J. de Miguel, P. M. Muñoz,G. Vergnaud, and I. López-Goñi. 2007. Assessment of genetic stability ofBrucella melitensis Rev. 1 vaccine strain by multiple-locus variable number oftandem repeat analysis. Vaccine 25:2858–2862.

14. García-Yoldi, D., P. Le Flèche, M. De Miguel, P. M. Muñoz, J. M. Blasco, Z.Cvetnik, C. M. Marín, G. Vergnaud, and I. López-Goñi. 2007. Comparison ofmultiple-locus variable-number tandem-repeat analysis with other PCR-based methods for typing Brucella suis isolates. J. Clin. Microbiol. 45:4070–4072.

15. Gee, E. J., B. K. De, P. N. Levett, A. M. Whitney, R. T. Novak, and T. Popovic.2004. Use of 16S rRNA gene sequencing for rapid confirmatory identifica-tion of Brucella isolates. J. Clin. Microbiol. 42:3649–3654.

16. Grissa, I., P. Bouchon, C. Pourcel, and G. Vergnaud. 2007. On-line resourcesfor bacterial micro-evolution studies using MLVA or CRISPR typing. Bio-chimie 90:660–668.

17. Grundmann, H., H. Satoshi, and G. Tanner. 2001. Determining confidenceintervals when measuring genetic diversity and the discriminatory abilities oftyping methods for microorganisms. J. Clin. Microbiol. 39:4190–4192.

18. Halling, S. M., B. D. Peterson-Burch, B. J. Bricker, R. L. Zuerner, Z. Qing,L.-L. Li, V. Kapur, D. P. Alt, and S. C. Olsen. 2005. Completion of thegenome sequence of Brucella abortus and comparison to the highly similargenomes of Brucella melitensis and Brucella suis. J. Bacteriol. 187:2715–2726.

19. Hubbard, D. J., R. K. Porschen, and K. Van Horn. 2007. Evaluation of dyetolerance media for the identification of Brucella species, abstr. C-170. Abstr.107th Gen. Meet. Am. Soc. Microbiol. American Society for Microbiology,Washington, DC.

20. Hunter, P. R., and M. A. Gaston. 1988. Numerical index of the discrimina-tory ability of typing systems: an application of Simpson’s index of diversityJ. Clin. Microbiol. 26:2465–2466.

21. Kattar, M. M., P. Zalloua, G. F. Araj, J. Samaha-Kfoury, H. Shabklo, S.Khalife, S. S. Kanj, and M. Deeb. 2007. Development and evaluation ofreal-time PCR assays for rapid diagnosis of human brucellosis. Diagn.Microbiol. Infect. Dis. 59:23–32.

22. Le Flèche, P., I. Jacques, M. Grayon, A. Al-Dahouk, P. Bouchon, F. Denoeud,K. Nöckler, H. Neubauer, L. A. Guilloteau, and G. Vergnaud. 2006. Evalu-ation and selection of tandem repeat loci for a Brucella MLVA typing assay.BMC Microbiol. 143:2913–2921.

23. Lista, F., G. Faggioni, S. Valjevac, A. Ciammaruconi, J. Vaissaire, C. leDoujet, O. Gorgé, R. De Santis, A. Carattoli, A. Ciervo, A. Fasanella, F.Orsini, R. D’Amelio, C. Pourcel, A. Cassone, and G. Vergnaud. 2006. Geno-typing of Bacillus anthracis strains based on automated capillary 25-locimultiple locus variable-number tandem repeats analysis. BMC Microbiol.6:33.

24. Marianelli, C., C. Graziani, C. Santangelo, M. T. Xibilia, A. Imbriani, R.Amato, D. Neri, M. Cuccia, S. Rinnone, V. Di Marco, and F. Ciuchini. 2007.Molecular epidemiologic and antimicrobial susceptibility characterization ofBrucella isolates from humans in Sicily, Italy. J. Clin. Microbiol. 45:2923–2928.

24a.Melzer, F., G. Vergnaud, and A. Whatmore. 2008. The results of the firstinternational ring trial of MLVA Brucella typing. Abstr. Int. Brucellosis Res.Conf., London, United Kingdom, abstr. 035.

25. Mercier, E. E. Jumas-Bilak, A. Allardet-Servent, D. O’Callaghan, and M.

Ramuz. 1996. Polymorphism in Brucella strains detected by studying distri-bution of two short repetitive DNA elements. J. Clin. Microbiol. 34:1299–1302.

26. Michaux-Charachon, S., G. Bourg, E. Jumas-Bilak, A. Guigue-Talet, P.Guigue-Talet, A. Allardet-Servent, D. O’Callaghan, and M. Ramuz. 1997.Genome structure and phylogeny in the genus Brucella. J. Bacteriol. 179:3244–3249.

27. Moreno, E., A. Cloeckaert, and I. Moriyón. 2002. Brucella evolution andtaxonomy. Vet. Microbiol. 90:209–227.

28. Pappas, G., P. Papadimitriou, N. Akritidis, L. Christou, and E. V. Tsianos.2006. The new global map of human brucellosis. Lancet Infect. Dis. 6:91–99.

29. Ratushna, V. G., D. M. Sturgill, S. Ramamoorthy, S. A. Reichow, Y. He, R.Lathigra, N. Sriranganathan, S. M. Halling, S. M. Boyle, and C. J. Gibas.2006. Molecular targets for rapid identification of Brucella spp. BMC Mi-crobiol. 6:13.

30. Roth, L. D., A. S. Khan, S. R. Lillibridge, S. M. Ostroff, and J. M. Hughes.2002. Public health assessment of potential biological terrorism agents.Emerg. Infect. Dis. 8:225–230.

31. Scholz, H. C., Z. Hubalek, I. Sedlacek, G. Vergnaud, H. Tomaso, S. Al-Dahouk, F. Melzer, P. Kämpfer, H. Neubauer, A. Cloeckert, M. Maquart,M. S. Zygmunt, A. M. Whatmore, E. Falsen, P. Bahn, C. Göllner, M. Pfeffer,B. Huber, H.-J. Basse, and K. Nöckler. 2008. Brucella microti sp. nov. isolatedfrom the common vole Microtus arvalis. Int. J. Syst. Evol. Microbiol. 58:375–382.

32. Scholz, H. C., Z. Hubalek, J. Nesvadbova, H. Tomaso, G. Vergnaud, P. LeFlèche, A. M. Whatmore, S. Al Dahouk, M. Krüger, C. Lodri, and M. Pfeffer.2008. Isolation of Brucella microti from soil. Emerg. Infect. Dis. 14:1316–1317.

33. Scholz, H. C., E. Hofer, G. Vergnaud, P. Le Fleche, A. Whatmore, S. AlDahouk, M. Pfeffer, M. Krüger, A. Cloeckaert, and H. Tomaso. 30 October2008, posting date. Isolation of Brucella microti from mandibular lymphnodes of red foxes, Vulpes vulpes, in Lower Austria. Vector Borne ZoonoticDis. [Epub ahead of print.]

34. Tcherneva, E., N. Rijpens, B. Jersek, and L. M. F. Herman. 2000. Differen-tiation of Brucella species by random amplified polymorphic DNA analysis.J. Appl. Microbiol. 88:69–80.

35. Tcherneva, E., N. Rijpens, C. Naydensky, and L. M. F. Herman. 1996.Repetitive element sequence based polymerase chain reaction for typing ofBrucella strains. Vet. Microbiol. 51:169–178.

36. Valdezate, S., I. Cervera, P. Hernandez, A. Navarro, and J. A. Saéz Nieto.2007. Characterization of human outbreaks of brucellosis and sporadic casesby the use of hyper-variable octameric oligonucleotide fingerprinting(HOOF) variable number repeats. Clin. Microbiol. Infect. 13:887–892.

37. Van Belkum, A., P. T. Tassios, L. Dijkshoorn, S. Haeggman, B. Cookson,N. K. Fry, V. Fussing, J. Green, E. Feil, P. Gerner-Smidt, S. Brisse, and M.Struelens. 2007. Guidelines for validation and application of typing methodsfor use in bacterial epidemiology. Clin. Microbiol. Infect. 13(Suppl. 3):1–46.

38. Verger, J. M., F. Grimont, P. A. D. Grimont, and M. Grayon. 1985. Brucella,a monospecific genus as shown by deoxyribonucleic acid hybridization. Int. J.Syst. Bacteriol. 35:292–295.

39. Vergnaud, G., and C. Pourcel. 2006. Multiple locus VNTR (variable numberof tandem repeat) analysis (MLVA), p. 83–104. In E. Stackebrandt (ed.),Molecular identification, systematics and population structure of pro-karyotes. Springer-Verlag, Berlin, Germany.

40. Whatmore, A. M., L. L. Perrett, and A. P. MacMillan. 2007. Characterizationof the genetic diversity of Brucella by multilocus sequencing. BMC Micro-biol. 7:34.

41. Whatmore, A. M., S. J. Shankster, L. L. Perrett, T. J. Murphy, S. D. Brew,R. E. Thirlwall, S. J. Cutler, and A. P. MacMillan. 2006. Identification andcharacterization of variable-number tandem-repeat markers for typing Bru-cella spp. J. Clin. Microbiol. 44:1982–1993.

42. Whatmore, A. M., T. J. Murphy, S. J. Shankster, E. Young, S. Cutler, andA. P. Macmillan. 2005. Use of amplified fragment length polymorphism toidentify and type Brucella isolates of medical and veterinary interest. J. Clin.Microbiol. 43:761–769.

3940 KATTAR ET AL. J. CLIN. MICROBIOL.

on March 31, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/