Genotyping of Mycoplasma mycoides subsp. mycoides SC

Vet. Res. (2008) 39:14 www.vetres.orgDOI: 10.1051/vetres:2007052

c© INRA, EDP Sciences, 2008 Original article

Genotyping of Mycoplasma mycoides subsp. mycoides SCby multilocus sequence analysis allows molecular

epidemiology of contagious bovine pleuropneumonia

Aboubakar Yaya1,2, Lucía Manso-Silvan1, Alain Blanchard3,François Thiaucourt1*

1 CIRAD UPR15 “Control of animal diseases”, OIE & FAO reference laboratory for CBPP, Montpellier, France2 LANAVET, Laboratoire national vétérinaire, Garoua, Cameroon

3 INRA, Université de Bordeaux 2, UMR 1090 Génomique, diversité, pouvoir pathogène,Villenave d’Ornon, France

(Received 4 May 2007; accepted 30 October 2007)

Abstract – Mycoplasma mycoides subsp. mycoides SC (MmmSC) is the etiological agent of contagiousbovine pleuropneumonia (CBPP). Although eradicated in most developed countries, the disease reappearedin Europe in the 1990s. This reappearance may have been caused either by importation from sub-SaharanAfrica, where CBPP is still endemic, or by the reemergence of virulent strains in Europe, as suggestedby earlier studies. A multilocus sequence analysis scheme has been developed to address this issue and,most importantly, to be able to monitor new epidemics. The alignment of the full genome sequence ofthe reference strain PG1 and the partial genome sequence of a pathogenic strain allowed the identificationof polymorphic sites. Nineteen initial loci were selected within housekeeping genes, genes of unknownfunction and non coding sequences. The suitability of these loci for genotyping MmmSC strains was firsttested on six strains of diverse geographic origin. The analyses showed that the published PG1 sequencecontained a number of specific polymorphisms that were therefore of no use for molecular typing. Amongthe eight informative polymorphic loci finally selected, only one (ftsY) was positioned within a housekeep-ing gene. Three main groups and 31 different allelic profiles were identified among 51 strains and strainvariants examined. Cluster analysis confirmed that European strains from the 1990s did not originate fromAfrica. It also showed a genetic link between a European strain isolated in 1967 and those found in southernAfrica and Australia. This was in agreement with historical data showing that CBPP was introduced in theseregions during colonisation in the 19th century.

Mycoplasma mycoides / contagious bovine pleuropneumonia / multilocus sequence analysis / molecu-lar epidemiology

1. INTRODUCTION

Mycoplasma mycoides subsp. mycoides SC(MmmSC) is the etiological agent of conta-gious bovine pleuropneumonia (CBPP) [33].This disease belongs to the list of notifiableanimal diseases from the World Organisationfor Animal Health (OIE) because of its eco-

* Corresponding author: [email protected]

nomic importance (http://www.oie.int). CBPPinduces lesions of pneumonia and pleurisyin cattle and domestic buffaloes and, withouttreatment, mortality rates can reach up to 50%.CBPP gained a worldwide distribution as aresult of live animal trade in the middle ofthe 19th century [11]. The disease was pro-gressively eradicated from many countries inEurope and North America on the eve of the

Article available at http://www.vetres.org or http://dx.doi.org/10.1051/vetres:2007052

http://www.vetres.org

http://dx.doi.org/10.1051/vetres:2007052

Vet. Res. (2008) 39:14 A. Yaya et al.

20th century thanks to the strict applicationof stamping out policies. A strategy that com-bined vaccination campaigns, detection andslaughter of infected animals and control ofanimal movements allowed Australia to regaina CBPP-free status in 1973 [8]. In sub-SaharanAfrica, CBPP was eradicated in the extremesouthern part of Africa but remains enzooticelsewhere. The disease was under control inAfrica around 1980, when repeated vaccina-tion campaigns were organised to eradicaterinderpest by use of bivalent vaccines that alsoprovided protection against CBPP. However,its prevalence and distribution have extendeddramatically since the eradication of rinder-pest in 1993, which implies that no morevaccination campaigns have been organisedand subsidised [25]. European countries arenow free of CBPP, though some of them haveexperienced sporadic outbreaks at regular in-tervals. This was the case in Portugal, whereCBPP outbreaks were reported in 1935, 1953,and 1993. Outbreaks were also reported at theborder between Spain and France in 1967 and1982, and in Italy in 1993 [34]. On each oc-casion, field epidemiological inquiries failedto yield clues regarding the origin of the out-breaks and did not permit to elucidate whetherthey were the result of reintroduction or resur-gence of the disease.

MmmSC, the agent of CBPP, belongs tothe so-called Mycoplasma mycoides cluster,which includes six mycoplasma subspeciesand groups of strains that are closely relatedfrom both a phenotypic and a genetic pointof view [9, 10]. The organisms most closelyrelated to MmmSC are M. mycoides subsp. my-coides LC (MmmLC) and M. mycoides subsp.capri (Mmc), which are both etiological agentsof contagious agalactia in small ruminants. Re-cent data have shown that these two subspeciescould be grouped into a single entity [45]. Dif-ferentiation of the strains belonging to thesethree taxons is difficult when relying onlyon the standard growth inhibition technique.Concerning intra-species polymorphism, com-parative studies by SDS-PAGE have alreadyshown that strain variability within MmmSCis much more limited than within MmmLCand Mmc [9]. Although MmmSC is thought to

be a very homogeneous taxon, previous stud-ies have shown that MmmSC strains couldbe differentiated using molecular typing as-says. Analysis of genomic polymorphism byendonuclease restriction of whole genomicDNA showed that different profiles were ob-tained with BamHI and PstI [31] and that theprofiles corresponding to European strains dif-fered from those of African origin. The useof HindIII allowed distinguishing various vac-cine strains [40]. The evidence of multiplecopies of insertion elements present in theMmmSC genome allowed the development ofnew typing tools based on Southern blot hy-bridisation. The use of IS1296 as a probeallowed the clear differentiation of recent Eu-ropean strains from those of African origin [7].The observed difference was explained later onby the identification of an 8.8 kbp deletion inthe genome of most MmmSC strains of Euro-pean origin [43], resulting in a missing IS1296band. The use of another insertion element,IS1634, also led to different Southern blot pro-files [24, 46] although the high copy numberof this insertion sequence (N = 60), as com-pared to that of IS1296 (N = 28), gave rise toprofiles that were difficult to analyse. The useof a technique based on sequencing multipleloci, designated multilocus sequence analysis(MLSA), allowed the identification of 15 dif-ferent allelic profiles within a representativenumber of MmmSC strains (N = 48) of var-ious origins [22]. In 2004 the whole genomesequence of MmmSC reference strain PG1 waspublished [48], opening new opportunities forthe development of typing tools. Analysis ofthe PG1 genome sequence showed that theloci and primer pairs previously selected forMLSA were not the most adequate. Some ofthe primers hybridised on multiple sites, whilstother targeted sequences were duplicated inthe PG1 genome, hampering result interpreta-tion. Comparison of PG1 sequences with thesequences of a pathogenic MmmSC strain, ob-tained through an ongoing genome sequencingproject, allowed the identification of a num-ber of polymorphic sites. The objective of thisstudy was to select a set of polymorphic locito develop a more precise and robust MLSAscheme in order to type MmmSC strains from

Page 2 of 19 (page number not for citation purpose)

Molecular epidemiology of CBPP by MLSA Vet. Res. (2008) 39:14

Table I. List of strains.

Profiles Strains Origin Locality Year of Comments References Provided byisolation

A00 Gemu Gofa Ethiopia Turmi? 1974 NVI Dr FikreA00 83162 Chad Bougor? 1983 CIRADA01 91130 CAR Bouar? 1991 LCV Bangui, Dr CondolasA02 Afadé Cameroon Afade 1965 LaboFarcha Dr ProvostA02 Vom Nigeria Vom? 1965 Sheep orign Vom LaboratoryA03 B17 Chad N’Djamena 1959 LaboFarcha Dr BalisA04 8740 Cameroon Touroua 1987 [50] LANAVET, Garoua Dr TulasneA04 8740-Rita Cameroon Garoua 1987 8740 after LANAVET, Garoua Dr Yaya

passage on cattleA04 8740/11 Cameroon Touroua 1987 11th passage CIRADA04 8740/53 Cameroon Touroua 1987 53rd passage CIRADA04 98029 Benin Kandi? 1998 Dr HendrikxA04 9048 Nigeria Sokoto? 1990 Labocel Niamey Dr BlochA05 Muguga Kenya Athi-river 1997 [47] KARI, Muguga, Dr WesongaA05 95014 Tanzania Shinyanga? 1995 FAO Dr ProvostA05 94111 Rwanda Kigali 1994 Action Nord-Sud, Mme IsakovA06 2000-033 Ethiopia Gimbi 2000 NVRIC Sebata, Dr SentayouA07 7721 Mauritania Tiguent 1997 CNERV, Nouakchott, Dr BrunA08 96010 Mauritania Noukchott? 1996 NVRIC Sebata, Dr SentayouA09 Fatick Senegal Fatick 1968 IEMVT (CIRAD)A10 99042 Mali Kolokani? 1999 LCR Bamako, Dr NiangA11 2000-005 Burkina Faso Ouagadougou? 2000 LDRV, Ouagadougou, Dr BadoA12 2003-036 Mali Segou? 2001 LCV Bamako, Dr NiangA12 87137-10 Burkina Faso Banfora 1987 LDRV, Ouagadougou, Dr RouilleA12 98011 Ivory Coast ? 1998 LCPA Bingerville, Dr CisséA12 9050-529-1 Ivory Coast Korhogo? 1990 LCPA Bingerville Dr DomenechB01 8988-1 Namibia Oudangwa 1989 CVL Windhoek, Dr HehnenB02 97039 Namibia Rundu? 1994 CVL Windhoek, Dr HuebschleB02 97041 Botswana Nata? 1995 CVL Windhoek, Dr Huebschle

all over the world and to elucidate the molecu-lar epidemiology of CBPP.

2. MATERIALS AND METHODS

2.1. MmmSC strains and isolates

Table I lists the names and origin of the51 strains and strain variants of MmmSC used in thisstudy. The strains, selected in order to cover the en-tire geographic distribution of CBPP, were isolatedfrom Africa (N = 41), Europe (N = 5), Australia(N = 2), and India (N = 1), whereas the origin ofstrains Lederle, Asmara, and PG1 is unknown. Theoldest strain of known origin was isolated in 1936(strain V5 from Australia) and the most recent in2004 (strain 2004-003 from Zambia). The date ofisolation of six strains was not precisely known. Al-though the majority of the isolates were of bovineorigin, a few isolates from small ruminants (Vom,94158 and 99048) were also included.

Strains were grown in a modified Hayflickmedium containing sodium pyruvate and glu-

cose [39]. At the end of the exponential phaseof growth, 2 mL of culture were centrifuged at12 000 × g for 15 min at 4 ◦C. The pellet waswashed once in PBS, and re-suspended in 100 µLof distilled water, to which 150 µL of lysis buffer(100 mM Tris HCl, pH 8.5; 0.05% Tween 20;0.24 mg/mL proteinase K) were added. Followingincubation at 60 ◦C for 30 min, proteinase K washeat-inactivated at 95 ◦C for 5 min. Samples in lysisbuffer were stored at –20 ◦C until analysis and theywere added to the PCR mix without further DNAextraction.

2.2. Identification of suitable loci for MLSA

Partial genomic sequences of the MmmSCpathogenic strain 8740-Rita, obtained from an on-going whole genome sequencing project, wereused to evaluate the polymorphism existing be-tween this strain and the reference strain PG1(NC_005364). Although the final assembly of the8740-Rita genome has not yet been obtained, theavailable contigs encompassed most of the DNA

(page number not for citation purpose) Page 3 of 19


Table I. Continued.

Profiles Strains Origin Locality Year of Comments References Provided byisolation

B02 97038 Botswana ? 1995 [1] CVL Windoek, Dr HuebschleB03 T2/34 Tanzania Morogoro? 1956 CIRADB03 99048 India Izatnagar? 1999 Goat origin [38] IVRI, Izatnagar, Dr SrivastavaB03 Gladysdale Australia Gladysdale < 1965 CIRADB03 Pillai Sudan Malakal? 1960 CIRADB04 2004-003 Zambia Sesheke 2004 FAO Dr AmanfuB05 V5 Australia ? 1938 Vaccine [6] CIRADB06 Asmara Eritrea Asmara unknown Vaccine CIRADB07 99021 Tanzania T1/B 1999 T1 variant [47] CIRADB07 T1sr Tanzania Korogwe? 1951 Vaccine CIRADC01 94158 Portugal ? 1994 [5] LNIV, Lisboa Dr BrandaoC01 PO2 France Pyrennees 1981 CIRADC01 99065 Italy Brescia 1992 IZS Teramo Dr PiniC02 9335-170 Italy ? 1992 IZS Teramo Dr SantiniC03 PO1967 France Pyrennees 1967 CIRADD01 97009 Ethiopia Makale 1997 NVI Dr BerheD02 2003-011 Eritrea Assab? 2003 MOA Asmara, Dr TesfaalemE01 DK32 Senegal Dakar unknown Vaccine [32] CIRADE02 9373-804 Guinea Kankan? 1993 LDV Conakry, Dr CondéF01 PG1 Unknown ? < 1931 Reference strain CIRADG01 Lederle Unknown Puigcerda? < 1967 Vaccine CIRADH01 Filfili Senegal Filfili 1965 CIRADI01 KH3J Sudan Juba < 1948 Vaccine [16] CIRAD

Localities followed by a ? were chosen to obtain a geographic positionning. CAR = Central African Re-public; CIRAD = centre de coopération internationale en recherche agronomique pour le développement;CVL = central veterinary laboratory; FAO = food and agriculture organisation; IVRI = Indian veterinary researchinstitute; IZS = instituto zooprofilactico sperimentale; KARI = Kenya agricultural research institute; LABO-CEL = laboratoire central d’élevage; LANAVET = laboratoire national vétérinaire; LCPA = laboratoire centralde pathologie animale; LCV = laboratoire central vétérinaire; LDRV = laboratoire départemental de recherchevétérinaire; LDV = laboratoire de diagnostic vétérinaire; LNIV = laboratorio nacional de investigacao veteri-naria; MOA = ministry of agriculture; NVI = national veterinary institute; NVRI = national veterinary researchinstitute.

sequences homologous to PG1. Only insertion se-quences and large duplicated zones in the PG1genome, unsuitable for MLSA, were excluded. Asa whole, 800 kbp were available for comparison.Alignments were performed using the AlignX pro-gram of Vector NTI software (Vector NTI v10.3.0,Invitrogen corporation, Carlsbad, CA, USA). Poly-morphic sites were recorded and positioned in thePG1 genomic sequence map. Nineteen polymorphicpositions were chosen for initial validation. Sevenof them resided in housekeeping genes, six withinlipoprotein or transmembrane protein genes and afurther six within non-coding sequences and theywere all distributed along the entire PG1 genome, asshown in Figure 1. The suitability of these loci forMLSA was first validated by sequencing the homol-ogous regions of five additional MmmSC strains:the original PG1 strain used for genome sequenc-

ing in Sweden and four other strains from Europe(PO1967), East Africa (94111), West Africa (Fil-fili), and Australia (Gladysdale). Eight potentiallyinteresting loci were then chosen for final validationusing the whole set of strains (Tab. II).

2.3. PCR and sequencing

PCR was carried out in a 50 µL reaction contain-ing: 0.5 unit of Taq DNA polymerase (Qiagen SA,Courtaboeuf, France) in its corresponding buffer(including 1.5 mM MgCl2), 300 µM of dTTP anddATP, 150 µM of dGTP and dCTP, 0.4 µM of eachprimer and 1 µL of sample diluted 1/100. Amplifi-cations were performed using GeneAmp PCR Sys-tems 2400 and 9700 (Perkin Elmer, Courtaboeuf,France). Thermal cycling consisted in an initial de-naturation step at 94 ◦C for 5 min, followed by



Figure 1. Distribution of polymorphic positions in the PG1 genome (NC_005364) and loci used for MLSA.Polymorphic positions are identified giving the corresponding CDS number in the PG1 sequence. The lociretained for the final MLSA scheme are boxed.

35 cycles of denaturation at 94 ◦C for 15 s, an-nealing at 52 ◦C for 30 s and extension at 72 ◦Cfor 90 s. The final extension step was maintainedat 72 ◦C for 7 min. The size and purity of theamplicons were controlled by electrophoretic sepa-ration in 1% agarose gels and samples with relevantfeatures were sent to Genome Express (GenomeExpress Cogenics, Meylan, France) for sequencingwith the corresponding primers. Both forward andreverse strands were sequenced systematically.

2.4. Allelic profile analysis

The sequences obtained from each correspond-ing forward and reverse primers were assembled us-ing Vector NTI SuiteTM (InfoMax, 2001) and the ex-tremities corresponding to single strand sequencesor showing aberrant features were trimmed. Thesequences obtained from different strains for eachlocus were aligned using ClustalW (Vector NTI).Polymorphic sites were recorded, carefully check-ing the corresponding sequence chromatograms. Anew allele number was assigned to each change inthe nucleotide sequence. In the end, each strain was

characterised by an allelic profile, corresponding tothe combination of allele numbers for each of theeight selected loci. This resulted in a matrix com-prising 31 different allelic profiles (Tab. III), whichwas analysed using eBURST V3, a software devel-oped for the analysis of multilocus sequence typing(MLST) datasets1. Standard parameters were used,except for the number of loci, which was fixed ateight. Each strain was then characterised by a letter,corresponding to its group, followed by a number,corresponding to the allelic profile.

2.5. Molecular epidemiology analysis

When the exact origin of the strain was known,the geographic coordinates were inferred from adedicated website2. When this information was notavailable, a putative geographic location was as-signed to the strain, taking into account the knowndistribution of CBPP in the country. For example,strains originating from Namibia were positionednorth of the veterinary fence that separates CBPP

1 http://eburst.mlst.net/v3/2 http://earthsearch.net/intSearch/



Table II. List of primers used for PCR and sequencing. The positions refer to the nucleotide sequenceNC_005364 of reference strain PG1.

Loci names Gene Primers Position Sequences (5’- 3’) Amplicon size(bp)

Loc-PG1-0001 ncLoc-PG1-0001-F 1272 AACAAAAGAGATCTTAAATCACACTTTA

538Loc-PG1-0001-R 1809 CCTCTTGTTTAACTTCTAGATCAGAAT

Loc-PG1-0103 LppLoc-PG1-0103-D 121796 GATGGATATAATCTATACTAGCATTTA

1321Loc-PG1-0103-F 123116 CCTTATATAGATAAAACTCCTCCTTA

Loc-PG1-0287 ncLoc-PG1-0287-F 328750 GATTGCTTTAATCAATTTCTTACTGA

545Loc-PG1-0287-R 329294 GGATAACCTTGATTTTTTATTTGCTTTA

Loc-PG1-0431 LppLoc-PG1-0431-F 485201 CAATTCTTTAAATTTGGGTTTGTT

608Loc-PG1-0431-R 485808 CTTGCAAGAGTATTTAGATTTGATTAAA

Loc-PG1-0489 ftsYLoc-PG1-0489-F 555289 GTTAGTTGTTGAAATGTTAGATAT

756Loc-PG1-0489-R 556044 CCCATATCAGTTTGGATTAA

Loc-PG1-0523 ChpLoc-PG1-0523-F 597315 ACAGCATTTGATCAAGATTTAAGTAGTT

824Loc-PG1-0523-R 598138 TTACCTAGGTGTTTAAAACTTCATTTG

Loc-PG1-0710 HpLoc-PG1-0711 –F 816093 CCAGTTGAACCATTTATTTTATATATACCT

643Loc-PG1-0711-R 816735 AAATATAAGTGGTGCTGGAATAACA

Loc-PG1-0827 ncLoc-PG1-0827-F 940048 AGTTGTACAACTGTATGAATCTATGATTAT

619Loc-PG1-0827-R 940666 CAGGATATACTTCAAAAATTAAAGGTTT

Lpp: Lipoprotein, Chp: conserved hypothetical protein, Hp: hypothetical protein, nc: non coding sequence.

infected regions in the North from CBPP-free zonesin the South. Strains were then positioned on a map,displaying group and allele numbers.

3. RESULTS

3.1. Choice of MLSA loci and initial validation

The first objective of this study was to findsuitable loci for multilocus sequence analy-sis of MmmSC strains. Nineteen polymorphicsites observed by comparing the genome se-quence of MmmSC reference strain PG1 withthe partial genome sequence of pathogenicstrain 8740-Rita were chosen for initial valida-tion. Five strains, including the original PG1used for genome sequencing in Sweden andfour other strains from diverse geographic ori-gins, were used for this initial validation.

Two loci were considered unsuitable be-cause the polymorphisms that were initiallyobserved between PG1 and 8740-Rita wereapparently due to errors in the published PG1sequence. In Loc-PG1-0330 (lepA) both theG in position 379030 and the T in position379050 are actually A. These two modifica-tions result in a single amino acid modificationin the deduced protein sequence. In Loc-PG1-0191 (trsE), the T in position 227175 should

be deleted and a T should be inserted afterposition 227214. These changes modify sig-nificantly the sequence of the trsE product(12 amino acids). Interestingly, when this lo-cus was sequenced in the vaccine strain T1/44,an insertion sequence (IS1634) was found in-terrupting the trsE gene at position 227363.The putative function of the trsE gene productis not well known, though it may be involvedin the transport of large molecules, intracel-lular trafficking and secretion. The disruptionof trsE in strain T1/44 indicates that this geneis not essential, though whether this has con-tributed to the attenuation of the vaccine strainis yet to be elucidated. Four additional lociwere excluded as the polymorphism was foundto be limited to the sequence of strain PG1,whereas the other five MmmSC strains exhib-ited identical sequences. Three of these lociwhere located within the housekeeping genesfusA (Loc-PG1-0159), glpK (Loc-PG1-0258),and rpoB (Loc-PG1-1008), whereas Loc-PG1-0372 corresponded to a non-coding sequence.Another four loci (Loc-PG1-0145, Loc-PG1-0194, Loc-PG1-0456, and Loc-PG1-0750)were eliminated because they did not add tothe discriminatory power of Loc-PG1-0489.Loc-PG1-0233 was also excluded, since only



Tabl

eII

I.M

atri

xsh

owin

gth

eva

riou

sal

leli

cpr

ofile

sob

tain

edth

roug

hM

LS

Aan

alys

isof

the

eigh

tloc

i.

Profi

les

Type

Loc

-PG

1-00

01L

oc-P

G1-

0103

Loc

-PG

1-02

87L

oc-P

G1-

0431

Loc

-PG

1-04

89L

oc-P

G1-

0523

Loc

-PG

1-07

10L

oc-P

G1-

0827

stra

ins

A00

Gem

uG

ofa

12

11

21

11

A01

9113

05

21

12

11

1A

02A

fadé

13

11

21

11

A03

B17

13

11

21

12

A04

8740

43

11

21

11

A05

Mug

uga.

11

11

21

11

A06

2000/0

331

11

13

11

1A

0777

211

24

12

11

1A

0896

010

22

41

21

11

A09

Fati

ck1

24

14

11

1A

1099

042

12

41

22

11

A11

2000/0

052

24

12

21

1A

1287

137

32

41

22

11

B01

C89

881

31

14

32

1B

0297

0039

12

11

43

21

B03

Gla

dysd

ale

12

15

43

21

B04

2004/0

031

41

14

32

1B

05V

51

31

54

32

1B

06A

smar

a1

31

14

31

1B

07T

1SR

15

11

43

11

C01

9415

81

21

15

52

2C

0293

351

21

25

52

2C

03PO

1967

12

11

53

22

D01

9700

94

21

14

31

1D

0220

03/0

114

21

14

31

2E

01D

K32

32

11

11

11

E02

9373/8

043

31

11

11

1F0

1PG

11

25

66

32

2G

01L

eder

lé6

13

44

42

3H

01Fi

lfili

41

21

11

11

I01

KH

3J1

21

34

11

1



two alleles could be defined in this sequence.All strains of European origin, and none of theother, bore allele 2 in this locus, making it agood tag for this origin. It must be noted thatPG1, the origin of which is unknown, also dis-played allele 2 in this locus. On the contrary,strain Lederle, which had been used as a vac-cine in Spain, displayed allele 1, suggestingthat it may not be of European origin. Finally,in Loc-PG1-0194, three alleles were defined,though the results were redundant and did notallow a better differentiation of the strains. Inthe end, eight loci were selected (Tab. II) toperform MLSA on the whole set of MmmSCstrains.

Interestingly, analysis of sequence varia-tions amongst MmmSC strains showed thatmany coding sequences (CDS) were inter-rupted in PG1, whereas larger CDS were iden-tified in the other MmmSC strains. This wasthe case of four out of the eight selected loci.The CDS MSC_0103 and MSC_0104 con-stituted a single CDS of around 370 aminoacids, as well as MSC_0287 and MSC_0288,forming a single CDS of 713 amino acids,and MSC_0710 and MSC_0711, which con-stituted a single CDS of 934 amino acids. TheCDS MSC_0827 was extended from 99 to344 amino acids.

3.2. Robustness of the MLSA

To evaluate the robustness of the method,cultures of pathogenic strain 8740, the samestrain at the 11th and 53rd passages and af-ter a passage in bovines (8740-Rita), wereanalysed. These were considered variants ofthe same strain, differing only by a few, wellcharacterised, in vitro or in vivo passages. Inaddition, three strains (94111, Filfili, Gladys-dale) were tested in duplicate at three monthsof interval. No differences were evidenced ei-ther between the duplicates or between thefour strain variants differing in the number ofpassages. This result was important to showthat the selected loci were stable over a fewpassages, implying that the sequence differ-ences were correlated to the natural evolutionof the strains in vivo. In the case of high muta-tion rates the signal would be blurred or biasedthrough in vitro passage, whereas in the case

of low mutation rates the loci would have beenunsuitable for use as molecular epidemiologytools.

3.3. Alleles defined in the different loci

3.3.1. Alleles defined in non-coding sequences

Five alleles were defined in locus Loc-PG1-0287. Two alleles gathered the majority of thestrains (46 out of 51), with allele 4 found onlyin Western Africa, while allele 1 had no geo-graphic specificity.

Three alleles were defined in locus Loc-PG1-0827. Forty-one strains bore the same al-lele. The other alleles corresponded to strainsfound in Europe and Africa, as well as PG1.

Six alleles were defined in Loc-PG1-0001,with the greatest variability found amongstrains from West Africa.

3.3.2. Alleles defined in genes of unknownfunction: Lipoproteins (Lpp) andConserved hypothetical proteins (Chp)

The smallest number of alleles was 2 forlocus Loc-PG1-0710. The observed mutationinduced an amino acid modification in the pre-dicted putative lipoprotein (L to S). There wasa clear geographic segregation of the alleles(Fig. 2). With the exception of strain Pillai, allstrains found in inter-tropical Africa bore al-lele 1.

Four different alleles were defined in Loc-PG1-0523. However, some MmmSC strainscould not be typed, as this locus was locatedwithin the 8.8 kbp deletion that was foundto be specific to recent European strains. Afifth allele was therefore created in order totype this last group of strains using eBURST.Interestingly, the only strain of European ori-gin for which this locus could be sequenced,PO1967, bore the same allele as the strainsfound in southern Africa and Australia. Thisallele was characterised by six polymorphicpositions that differed in other MmmSC strains(Tab. IV). This gave more strength to the ge-netic link demonstrated by this allele.

Five alleles were defined in locus Loc-PG1-0103. They corresponded to a variable numberof five to nine trinucleotide repeats (AAT) cod-ing for asparagine (N) within a gene coding for



Fig

ure

2.G

eogr

aphi

cdi

stri

butio

nof

alle

les

for

Loc

-PG

1-07

10.E

xcep

tfor

stra

inP

illai

,fro

mS

udan

,all

stra

ins

ofA

fric

anor

igin

disp

lay

alle

le1

inth

islo

cus,

whi

lstt

heot

hers

disp

lay

alle

le2.

Not

eth

atst

rain

PG

1,lo

cate

dby

the

lege

nd,w

asno

tpos

itio

ned

onth

em

apbe

caus

eit

sge

ogra

phic

orig

inis

unkn

own.


Vet. Res. (2008) 39:14 A. Yaya et al.Ta

ble

IV.P

olym

orph

ism

inth

ediff

eren

tall

elic

profi

les.

The

bott

omli

nere

pres

ents

the

cons

ensu

sse

quen

ces.

Nuc

leot

ides

iden

tica

lto

the

cons

ensu

sse

quen

cear

ele

ftbl

ank.

Del

etio

nsar

ein

dica

ted

by][

.The

seco

ndlin

esh

ows

the

posi

tion

ofpo

lym

orph

icsi

tes

inth

eP

G1

geno

me

(seq

uenc

eN

C_0

0536

4).

K:n

umbe

rof

olig

onuc

leot

ide

repe

ats

(TC

AA

CA

AG

A);

LL

:GA

AC

CT

GT

TG

TTA

AT

CC

AA

GA

TAA

AC

AA

GT

TG

AA

ATA

.

Groups

Profile

sTyp

e stra

insLoc

-PG1-

0103

Loc-PG

1-071

0Nu

mber

of stra

ins150

0151

8152

3152

4152

7152

8152

9153

0153

1153

2122

439328

887329

008329

113485

305385

470485

517485

524485

567485

572485

615555

508555

733555

796555

799555

914597

455597

609597

614597

615597

616597

618597

626597

647816

297940

184940

351

AA0

0Ge

mu Go

fa4

2A0

1911

30A

41

A02

Afadé

6ATT

42

A03

B17

6ATT

4C

1A0

4874

0T

6ATT

4G

6A0

5941

114A

TT4

3A0

6200

0/033

4ATT

4A

1A0

7772

1T

41

A08

96010

][][

][][

][][

T4

1A0

9Fa

tickT

31

A10

99042

T4

G1

A11

2000/0

05][

][][

][][

][T

4G

1A1

2200

3/036

TA

T4

G4

BB0

1C8

9886A

TT3

AT

TC

AT

G1

B02

970041

3A

TT

CA

TG

3B0

3Pil

laiC

C3

AT

TC

AT

G4

B04

2004/0

037A

TT3

AT

TC

AT

G1

B05

V56A

TTC

C3

AT

TC

AT

G1

B06

Asma

ra6A

TT3

AT

TC

AT

1B0

7T1

SR8A

TT3

AT

TC

AT

2C

C01

94158

3][

TG

C3

C02

9335

TT

3][

TG

C1

C03

PO196

7C

C3

][T

AT

TC

AT

GC

1D

D01

97009

T3

AT

TC

AT

1D0

2200

3/011

T2

AT

TC

AT

C1

EE0

1DK

32T

51

E02

9373/8

04T

6ATT

51

FF0

1PG

1T

AA

3][

TA

TT

CA

TG

C1

GG0

1Led

erleT

T][

][][

][][

][4A

TT][

G3

AT

TC

AT

TG

A1

HH0

1Filf

iliT

4ATT

A5

1I

I01KH

3JA

31

Conse

nsus

CA

GA

TA

GT

AA

5ATT

AA

GC

AG

GT

GC

TK

LLA

GT

GA

AA

GG

CA

GG

Loc-PG

1-0523

Loc-PG

1-0827

No am

plifica

tionNo

amplif

ication

Loc-PG

1-0001

Loc-PG

1-0287

Loc-PG

A-0431

Loc-PG

1-0489



a putative lipoprotein. There was no apparentgeographic segregation of the various alleles.

3.3.3. Allele defined in the housekeeping geneftsY

Six alleles were defined in Loc-PG1-0489,corresponding to ftsY. The majority of thestrains were segregated in two alleles, com-prising similar numbers (23 and 18 strainsrespectively). Allele 2 was found only onMmmSC strains from sub-Saharan origin. Al-lele 4 was displayed by strains from SouthernAfrica, Australia, and India, as well as by somestrains from East Africa. It was also found onone old strain from West Africa and a strain ofunknown origin (Lederle), which was formerlyused as a vaccine in Europe.

All strains isolated in Europe bore allele 5on this locus (Fig. 3), strain Lederle, formerlyused as a vaccine in Spain, has an unknownorigin.

3.4. Strain groups by MLSA analysis

All 51 strains were characterised by an al-lelic profile and 31 different allelic profileswere identified (Tab. III). The genetic eventsunderlying the description of the various alle-les are described in Table IV. Analysis usingeBURST allowed the identification of threemain groups of strains (comprising 43 strains).Eight strains were included into groups gath-ering only two strains or considered as sin-gletons. The geographic positioning of thevarious groups and allelic profiles is shown inFigure 4.

Group A was the largest, including 25strains (Fig. 5). Thirteen different allelicprofiles were clustered within this group.The largest (A04 and A12) gathered up tofour strains. Group A strains were foundmostly in West and central Africa, thoughsome of them were found in East Africa(Fig. 4). It must be noted that all MmmSCstrains isolated after 1994 in East Africa pre-sented the same allelic profile (A05).

Group B was the second largest, including13 strains that were subdivided into 7 allelicprofiles (Fig. 6). Group B strains were foundin southern Africa, East Africa, Australia and

India (Fig. 4). The most common allelic pro-file (B03) was shared by four strains of variousorigins (Australia, Tanzania, Sudan, and In-dia). Most of group B strains found in EastAfrica had been isolated before 1970 (Pillai,T1sr, T2-34, and a vaccine strain receivedfrom a laboratory at Asmara, the origin ofwhich remains undetermined).

Group C corresponded to strains foundsolely in Europe (Fig. 4 and 6).

Finally, eight strains could not be includedinto any of the three main groups, as they dis-played peculiar allelic profiles. In West Africapeculiar profiles corresponded to old strainsisolated in Senegal and Guinea and in EastAfrica to a vaccine strain (KH3J) also isolatedlong ago. Two strains isolated recently fromnorthern Ethiopia and Eritrea also bore pecu-liar allelic profiles. Finally, PG1 was the strainthat exhibited the most peculiar allelic profile.

4. DISCUSSION

4.1. Genomic analysis

The objective of this study was to estab-lish a more robust MLSA scheme for MmmSCstrains based on whole genome sequence com-parison. Analysis of PG1 sequences showedthat many primers designed on a previousMLSA scheme [22] hybridised at multiplesites or targeted sequences that were dupli-cated in the PG1 genome. The initial valida-tion of 19 selected loci in this study revealedthat sequencing errors may remain in the pub-lished PG1 genome. Out of the 19 initial poly-morphic loci, two corresponded to sequencingerrors. It must be noted that, if this resultwas extrapolated to the 267 polymorphic sitesfound between the sequences of strains PG1and 8740-Rita, there may be around 20 errorsin the published PG1 sequence. Another fourout of the 19 polymorphic positions seemedto be specific to the PG1 sequence. The pe-culiarity of the PG1 genome has already beenshown. For example, this strain possesses alarge DNA repeat of 24 kbp, which is absentin other MmmSC strains [3]. The peculiari-ties of PG1 may be due to a natural diver-gent evolution of the strain. However, it isdifficult to substantiate this hypothesis, since



Fig

ure

3.G

eogr

aphi

cdi

stri

buti

onof

alle

les

for

Loc

-PG

1-04

89.S

ixal

lele

sw

ere

defi

ned

inth

islo

cus.

All

ele

2w

asfo

und

only

onM

mm

SC

stra

ins

from

sub-

Sah

aran

orig

in.A

llel

e4

was

disp

laye

dby

stra

ins

from

Sou

ther

nA

fric

a,A

ustr

alia

,and

Indi

a,as

wel

las

byso

me

stra

ins

from

Eas

tAfr

ica.

Itw

asal

sofo

und

onon

eol

dst

rain

from

Wes

tA

fric

aan

da

stra

inof

unkn

own

orig

in(L

eder

le),

whi

chw

asfo

rmer

lyus

edas

ava

ccin

ein

Eur

ope.

Not

eth

atst

rain

PG

1,lo

cate

dby

the

lege

nd,w

asno

tpos

itio

ned

onth

em

apbe

caus

eit

sge

ogra

phic

orig

inis

unkn

own.



Fig

ure

4.G

eogr

aphi

cdi

stri

buti

onof

alle

lic

profi

les.

The

reis

acl

ear

corr

elat

ion

betw

een

the

alle

lic

profi

leof

Mm

mS

Cst

rain

san

dth

eir

geog

raph

icor

igin

.S

trai

nsof

grou

pA

are

foun

din

sub-

Sah

aran

Afr

ica

and

now

here

else

.Thi

sco

uld

bean

indi

cati

onth

atC

BP

Pw

aspr

esen

tin

Afr

ica

prio

rto

colo

nisa

tion

tim

e.O

nth

eot

her

hand

the

geno

mic

link

betw

een

stra

ins

from

Eur

opea

nor

igin

(gro

upC

)an

dth

ose

foun

din

Sou

ther

nA

fric

a,A

ustr

alia

,Ind

iaan

dso

me

foun

din

Eas

tAfr

ica

(gro

upB

)ca

nbe

corr

elat

edw

ith

hist

oric

alda

tade

scri

bing

CB

PP

expa

nsio

ndu

ring

colo

nisa

tion

inth

e19

thce

ntur

y.N

ote

that

stra

inP

G1,

loca

ted

byth

ele

gend

,was

notp

osit

ione

don

the

map

beca

use

its

geog

raph

icor

igin

isun

know

n.



Figure 5. Tree obtained by eBURST for group A strains. Group A is subdivided into 13 allelic profiles andgathers strains found only in sub-Saharan Africa. The largest allelic profile (A12) comprises four strains.However, these strains were collected in neighbouring countries on well known cattle trade routes. This isa likely explanation for finding strains of similar allelic profiles at different locations.

Figure 6. Tree obtained by eBURST for group B and C strains. Group B is subdivided into seven allelicprofiles and gathers strains found in Southern and East Africa, as well as strains isolated in Australia andIndia. Group C gathers European strains. Strain PO1967 is an intermediate between group B and group C.



the exact origin and year of isolation (be-fore 1931) of the strain are unknown. Morelikely, the divergence of PG1 may result fromthe many in vitro passages. A single house-keeping gene (ftsY) was validated for use inthe new MLSA scheme. It was noteworthy thatpolymorphisms in this gene included the pres-ence of duplicated nonanucleotides coding fora stretch of three amino acids (DQQ) presentin three to five copies, as well as the deletion ofa rather long DNA fragment (35 bp). Since thisgene, coding for a bacterial signal-recognitionparticle receptor, plays a fundamental role inmediating membrane targeting and insertion ofnascent proteins [2], it is likely that the poly-morphic sites are not positioned on the activesites of this molecule.

4.2. Epidemiological analysis

The presence of group A strains only inWest and central Africa may indicate that thesestrains have evolved separately from thosepresent in Europe. This may also indicate thatCBPP existed in the continent prior to coloni-sation and presumed introduction of CBPPwith live cattle. Such a hypothesis was alreadyraised when a traditional “vaccination” pro-cedure, consisting in inoculating pleural fluidsubcutaneously at the bridge of the nose, wasdescribed in West Africa [4]. The local inflam-mation generally led to a reaction of the pe-riosteum, with development of a pseudo-horn,which resulted in the erroneous description ofa new cattle breed called Bos triceros [35].These traditional procedures were unknown inEurope, where the inoculation point, as de-scribed by Willems, was the tip of the tail [49].Furthermore, a possible ancestral allelic pro-file (A00) was identified in strains isolated inChad and the Gemu-Gofa region in WesternEthiopia. In these regions the influx of exoticcattle was minimal in the 19th century. In con-trast, the presence of group B strains in EastAfrica could reflect an influx of affected cat-tle from Europe. This may have taken placedirectly from European countries (i.e. fromItaly to its Eritrean colony) or indirectly, aswhen the 1868 British expedition to Abyssiniaused oxcarts from India to pull artillery in the

mountainous regions [19] and might have thusintroduced CBPP in East Africa.

The geographic location of strains sharingthe same allelic profile agreed with knownepidemiological data. Therefore, the presenceof multiple strains with the same allelic pro-file certainly does not result from an MLSAlack of discriminatory power. For example,the presence of allelic profile A05 in manycountries in Eastern Africa can be explainedby a ‘clonal expansion’ of the initial strain ofKenyan origin that led to the re-contaminationof countries such as Tanzania and Rwanda inthe 1990s [25]. Similarly, the presence of thesame allelic profile (A12) in Mali, BurkinaFaso, and Ivory Coast is in agreement withthe transhumance and trade routes followed bycattle herds that are raised in Sahelian coun-tries and exported to meat markets near largecities in the South.

Although strain PO1967, of European ori-gin, belongs to group C, its sequence in Loc-PG1-0233 showed that it is closely related toGroup B strains. The localisation of GroupB strains in this MLSA scheme corroboratedwhat was known regarding the transmission ofCBPP from Europe to overseas colonies dur-ing the 19th century. It is well known thatCBPP was exported from Europe to South-ern Africa [41], Australia [15, 28], and In-dia [36] at that time. The European strainlinked to this group, PO1967, is the oldeststrain from our collection, which was iso-lated in 1967 on the French-Spanish border.This strain may represent the closest rela-tive to the ancestral strains that circulated inEurope in the 19th century. More recent Eu-ropean strains differ from PO1967 notably bya deletion of 8.8 kbp, which includes genesencoding a glycerol transporter that may be as-sociated with MmmSC virulence [30, 44]. Un-fortunately MmmSC strains circulating prior to1967 in Europe are not available. They couldhave been used to monitor the genetic driftand to check if they were genetically closer toMmmSC strains of Southern Africa.

Recent strains of European origin displayedpeculiar alleles that were not observed innon-European strains, notably on Loc-PG1-0233 which was not retained in the MLSA



scheme. This finding indicates further that re-cent CBPP outbreaks in Europe were not dueto a reintroduction, but were more likely to bea resurgence of the disease. This conclusionshould be of major concern to European coun-tries, particularly with regards to the wideningof EU borders. The unnoticed persistence ofMmmSC in Europe may be explained eitherby the existence of another, yet unidentified,reservoir or by the circulation of strains of low-pathogenicity, which may sometimes regainvirulence. The presence of MmmSC strains inhosts other than cattle and domestic buffaloeshas already been described. This is notablythe case of goats, from which MmmSC strainshave been regularly isolated both in Africaand in Europe and in which lesions of CBPPcan sometimes be reproduced [13]. In an at-tempt to examine whether goats may be usedin place of cattle as models for CBPP, variousinoculation routes were tested by Yaya et al.[51]. Some inoculated goats seroconverted andone animal, inoculated intra-peritoneally, de-veloped lesions. However, the role of goats inthe natural epidemiology of CBPP is certainlymarginal, if not completely irrelevant. Thiswas evidenced in Botswana in 1994, when thiscountry decided to slaughter the entire cattlepopulation from the infected zone [27]. Al-though the goat population was not targetedin the stamping-out campaign, the subsequentreintroduction of naive cattle in this countrywas not followed by CBPP outbreaks, henceshowing that goats did not play a significantrole in CBPP persistence and long term trans-mission, at least in this context.

On the contrary, the unnoticed circulationof low-pathogenic MmmSC strains in Europeremains a possible explanation for the resur-gence of CBPP. It is well known that MmmSCstrains show various degrees of virulence.However, strain virulence is difficult to mea-sure since (i) there is no small animal modelthat can mimic the lesions observed in cat-tle, (ii) endobronchial inoculations in cattle arenot always successful, even when using strainsthat have shown to be highly virulent in thefield, and (iii) in-contact transmission trialsare usually carried out with a reduced num-ber of animals, which may not guarantee the

evaluation of the strain virulence in a singleexperiment. This last point has been verifiedin silico with stochastic transmission modelsshowing that successive trials performed withthe same initial parameters may lead to verydifferent outcomes [21].

MLST is an unambiguous procedure forcharacterising isolates of bacterial speciesbased on the sequences of the internal frag-ments of house-keeping genes [23, 37]. Thesequences of approximately 500 bp are nor-mally used, as these can be accurately se-quenced on both strands using an automatedDNA sequencer. For each house-keeping gene,the different sequences present within a bac-terial species are assigned as distinct allelesand each isolate is characterised by an allelicprofile, corresponding to the alleles definedin each of the loci. In MLST analyses thenumber of nucleotide differences found be-tween alleles is disregarded and sequences aregiven different allele numbers whether theydiffer at a single or at many nucleotide po-sitions. The rationale for this is that a singlegenetic event resulting in a new allele canoccur by a point mutation, altering only a sin-gle nucleotide site, or by a recombinationalreplacement that will often result in the mod-ification of multiple sites. This typing schemehas been applied successfully to a wide varietyof bacterial pathogens such as Staphylococcusaureus [14], Campylobacter jejuni [12], andHaemophilus influenzae [26], to cite only afew. This approach has multiple applicationssuch as phylogeny [18], molecular structureanalysis [42], and molecular epidemiology. Inmolecular epidemiology, MLST can be usedto study the evolution of antibiotic resistantstrains [17], the temporal trends in strain ex-pansion [29], or the distribution of strains ofvarious lineages within a population [20].

In the case of MmmSC, this approach wasnot successful, since the variability withinhousekeeping genes was too limited, if presentat all. This was the case of the genes fusA,lepA, and rpoB. In terms of an evolutionaryperspective this may indicate that MmmSCgenomes are extremely homogeneous, sug-gesting that this mycoplasma has adapted veryrecently to its bovine host. From a practical



point of view, this means that genes of un-known function or non coding sequences hadto be selected in this study to differentiateMmmSC strains. This is why the strategyhas alternatively been named “multilocus se-quence analysis”. This typing scheme is lessuniversal than MLST analyses, since the genetargets used were selected for typing strainswithin a biotype and would not be suitablefor typing strains of higher taxonomic ranks.However, like MLST, MLSA has the advan-tage of being a portable and very robust ap-proach.

This new MmmSC typing scheme provedmore robust than the previous one developedin our laboratory [22]. No double bands wereobtained after PCR amplification, which en-sured that PCR products could be sent withoutany purification step. This was sometimes thecase with the previous MLSA. On the contrary,the new scheme allowed the description of 31profiles instead of only 15 out of a comparablenumber of strains (51 and 48 respectively).

MLSA analysis may open new perspec-tives for a better understanding of MmmSCgenome plasticity. In this respect, it would bevery interesting to obtain ancestral DNA fromsamples of CBPP-affected lungs preserved informalin for histology. This should allow thecharacterisation of the ancestral allelic profilescirculating in Europe and Africa and may per-mit measuring evolution rates for the variousloci.

This MLSA strategy may also be appliedto better understand CBPP epidemiology, im-proving thus the surveillance and control ofthe disease. New loci may be added to in-crease the discriminating power of the MLSAtool on subsets of strains that are very closelyrelated. Including a large number of strainsisolated from goats in various regions of theworld would be very interesting in order to ex-amine whether the MLSA profiles of MmmSCstrains from goat origin are similar or, onthe contrary, more closely related to otherMmmSC strains circulating in the same re-gion. The fact that caprine species may serveas a reservoir should be an incentive to mul-tiply mycoplasma isolation and identificationfrom this host, which is known to harbour my-

coplasmas in the ear canal in the absence ofpathological signs. Such a procedure is alreadyin place in France, conducted by an epidemi-ological network called VIGIMYC, managedby the Agence Française de Sécurité Sanitairedes Aliments in Lyon.

The results obtained in this study will beincluded in a web tool dedicated to molecularepidemiology within an EU-funded projectcalled EPIZONE. This tool should facilitatethe dissemination of typing methods forpathogens threatening Europe, allow anylaboratory to compare in-house data withlarge databank sets and, in turn, generatephylogenetic trees and actualised maps,provided that the geographic coordinates ofall isolates are known. Such a tool shouldnaturally be linked to the websites of the OIE(http://www.oie.int/wahid-prod/) and FAO(http://www.fao.org/ag/againfo/programmes/en/empres/disease_cbpp.asp), which provideupdated information on new outbreaks fornotifiable diseases.

Finally, this new MLSA tool will be par-ticularly useful to all countries at risk ofCBPP reintroduction or resurgence and tothose engaged in eradication campaigns, al-lowing them to trace back the origin of anyremaining CBPP focus.

Acknowledgements. We are indebted to a number ofprojects that made this study feasible: the AU/IBARCBPP PACE research project, funded by the EuropeanUnion (REG/5007/005), the LABOVET PSF project,funded by the French Ministry of Foreign Affairs andthe EPIZONE Rex project, funded by the EuropeanUnion. We are particularly grateful to the French Em-bassy in Cameroon for supporting the PhD thesis workof Yaya Aboubakar.

REFERENCES

[1] Amanfu W., Masupu K.V., Adom E.K.,Raborokgwe M.V., Bashiruddin J.B., An outbreak ofcontagious bovine pleuropneumonia in Ngamilanddistrict of north-western Botswana, Vet. Rec. (1998)143:46–48.

[2] Angelini S., Deitermann S., Koch H.G., FtsY, thebacterial signal-recognition particle receptor, interactsfunctionally and physically with the SecYEG translo-con, EMBO Rep. (2005) 6:476–481.

[3] Bischof D.F., Vilei E.M., Frey J., Genomic dif-ferences between type strain PG1 and field strains ofMycoplasma mycoides subsp. mycoides small-colonytype, Genomics (2006) 88:633–641.



[4] Blancou J., Early methods of surveillance and con-trol for contagious bovine pleuropneumonia, Rev. Sci.Tech. (1996) 15:1241–1282.

[5] Brandao E., Isolation and identification ofMycoplasma mycoides subspecies mycoides SC strainsin sheep and goats, Vet. Rec. (1995) 136:98–99.

[6] Campbell A.D., Contagious bovine pleuropneumo-nia. A report on the use of new antigens for thecomplement fixation and agglutiantion tests, J. Conc.Sci. Ind. Res. (1938) 11:112–118.

[7] Cheng X., Nicolet J., Poumarat F., Regalla J.,Thiaucourt F., Frey J., Insertion element IS1296in Mycoplasma mycoides subsp. mycoides smallcolony identifies a European clonal line distinct fromAfrican and Australian strains, Microbiology (1995)141:3221–3228.

[8] Clay A.L., Lloyd L.C., The eradication of CBPPfrom Australia, Bull. Off. Int. Epizoot. (1974) 91:533–546.

[9] Costas M., Leach R.H., Mitchelmore D.L.,Numerical analysis of PAGE protein patterns and thetaxonomic relationships within the ‘Mycoplasma my-coides cluster’, J. Gen. Microbiol. (1987) 133:3319–3329.

[10] Cottew G.S., Breard A., DaMassa A.J., Erno H.,Leach R.H., Lefevre P.C., Rodwell A.W., Smith G.R.,Taxonomy of the Mycoplasma mycoides cluster, Isr. J.Med. Sci. (1987) 23:632–635.

[11] Curasson G., Péripneumonie bovine, in: Traitéde pathologie exotique vétérinaire et comparée, VigotFrères (Ed.), Paris, France, 1942, pp. 276–353.

[12] Dingle K.E., Colles F.M., Wareing D.R., Ure R.,Fox A.J., Bolton F.E., Bootsma H.J., Willems R.J.,Urwin R., Maiden M.C., Multilocus sequence typingsystem for Campylobacter jejuni, J. Clin. Microbiol.(2001) 39:14–23.

[13] Dujardin-Beaumetz E., Transmission de la périp-neumonie des bovidés aux espèces ovine et caprine,Ann. Inst. Pasteur (Paris) (1906) 20:449–466.

[14] Enright M.C., Day N.P.J., Davies C.E., PeacockS.J., Spratt B.G., Multilocus sequence typing for char-acterization of methicillin-resistant and methicillin-susceptible clones of Staphylococcus aureus, J. Clin.Microbiol. (2000) 38:1008–1015.

[15] Fisher J., The origins, spread and disappear-ance of contagious bovine pleuro-pneumonia in NewZealand, Aust. Vet. J. (2006) 84:439–444.

[16] Gambles R.M., Studies on the contagious bovinepleuropneumonia with special reference to the comple-ment fixation test, Br. Vet. J. (1956) 112:34-40, 78-86,120-127, 162–169.

[17] Gherardi G., Fallico L., Del Grosso M., BonanniF., D’Ambrosio F., Manganelli R., Palu G., DicuonzoG., Pantosti A., Antibiotic-resistant invasive pneu-mococcal clones in Italy, J. Clin. Microbiol. (2006)45:306–312.

[18] Hanage W.P., Kaijalainen T., Herva E.,Saukkoriipi A., Syrjanen R., Spratt B.G., Usingmultilocus sequence data to define the Pneumococcus,J. Bacteriol. (2005) 187:6223–6230.

[19] Hozier H.M., British expedition to Abyssinia,MacMillan (Ed.), London, UK, 1869, p. 271.

[20] Lacher D.W., Steinsland H., Blank T.E.,Donnenberg M.S., Whittam T.S., Molecular evo-lution of typical enteropathogenic Escherichia coli:clonal analysis by multilocus sequence typing andvirulence gene allelic profiling, J. Bacteriol. (2007)189:342–350.

[21] Lesnoff M., Laval G., Bonnet P., Abdicho S.,Workalemahu A., Kifle D., Peyraud A., LancelotR., Thiaucourt F., Within-herd spread of contagiousbovine pleuropneumonia in Ethiopian highlands, Prev.Vet. Med. (2004) 64:27–40.

[22] Lorenzon S., Arzul I., Peyraud A., Hendrikx P.,Thiaucourt F., Molecular epidemiology of contagiousbovine pleuropneumonia by multilocus sequence anal-ysis of Mycoplasma mycoides subspecies mycoidesbiotype SC strains, Vet Microbiol. (2003) 93:319–333.

[23] Maiden M.C.J., Bygraves J.A., Feil E., MorelliG., Russell J.E., Urwin R., Zhang Q., Zhou J., ZurthK., Caugant D.A., Feavers I.M., Achtman M., SprattB.G., Multilocus sequence typing: a portable approachto the identification of clones within populations ofpathogenic microorganisms, Proc. Natl. Acad. Sci.USA (1998) 95:3140–3145.

[24] March J.B., Clark J., Brodlie M., Characterizationof strains of Mycoplasma mycoides subsp. mycoidessmall colony type isolated from recent outbreaksof contagious bovine pleuropneumonia in Botswanaand Tanzania: evidence for a new biotype, J. Clin.Microbiol. (2000) 38:1419–1425.

[25] Masiga W.N., Domenech J., Windsor R.S.,Manifestation and epidemiology of contagious bovinepleuropneumonia in Africa, Rev.-Off. Int. Epizoot(1996) 15:1283–1308.

[26] Meats E., Feil E.J., Stringer S., Cody A.J.,Goldstein R., Kroll J.S., Popovic T., Spratt B.G.,Characterization of encapsulated and noncapsulatedHaemophilus influenzae and determination of phylo-genetic relationships by multilocus sequence typing, J.Clin. Microbiol. (2003) 41:1623–1636.

[27] Mullins G.R., Fidzani B., Kolanyane M., Atthe end of the day. The socioeconomic impacts oferadicating contagious bovine pleuropneumonia fromBotswana, Ann. N.Y. Acad. Sci. (2000) 916:333–344.

[28] Newton L.G., Contagious bovine pleuropneumo-nia in Australia: some historic highlights from entry toeradication, Aust. Vet. J. (1992) 69:306–317.

[29] Perez-Losada M., Crandall K.A., Zenilman J.,Viscidi R.P., Temporal trends in gonococcal populationgenetics in a high prevalence urban community, Infect.Genet. Evol. (2007) 7:271–278.



[30] Pilo P., Vilei E.M., Peterhans E., Bonvin-KlotzL., Stoffel M.H., Dobbelaere D., Frey J., A metabolicenzyme as a primary virulence factor of Mycoplasmamycoides subsp. mycoides small colony, J. Bacteriol.(2005) 187:6824–6831.

[31] Poumarat F., Solsona M., Molecular epidemiol-ogy of Mycoplasma mycoides subsp. mycoides biotypesmall colony, the agent of contagious bovine pleurop-neumonia, Vet. Microbiol. (1995) 47:305–315.

[32] Provost A., Prophylaxis and vaccination in bovinepleuropneumonia. Evolution of technics and their cur-rent practical applications, Rev. Elev. Med. Vet. PaysTrop. (1974) 27:145–161.

[33] Provost A., Perreau P., Breard A., Le Goff C.,Martel J.L., Cottew G.S., Contagious bovine pleurop-neumonia, Rev.-Off. Int. Epizoot.(1987) 6:625–679.

[34] Regalla J., Caporale V., Giovannini A., Santini F.,Martel J.L., Goncalves A.P., Manifestation and epi-demiology of contagious bovine pleuropneumonia inEurope, Rev.-Off. In. Epizoot. (1996) 15:1309–1329.

[35] Rochebrune A.T. (de), Formation de races nou-velles. Recherche d’ostéologie comparée, sur une racede boeufs domestiques observée en Sénégambie, C. R.Acad. Sci. (Paris) (1880) 91:304–306.

[36] Shirlaw J.F., Observations on the existence of con-tagious bovine pleuropneumonia in British India, withan account of preliminary pathological investigation ofcases of this disease reported from Assam, Indian J.Vet. Sci. (1939) 9:139–150.

[37] Spratt B.G., Multilocus sequence typing: molec-ular typing of bacterial pathogens in an era ofrapid DNA sequencing and the Internet, Curr. Opin.Microbiol. (1999) 2:312–316.

[38] Srivastava N.C., Thiaucourt F., Singh V.P., SunderJ., Isolation of Mycoplasma mycoides small colonytype from contagious caprine pleuropneumonia inIndia, Vet. Rec. (2000) 147:520–521.

[39] Thiaucourt F., Di Maria A., A new microtitrationmethod for the enumeration of contagious bovine pleu-ropneumonia (CBPP) vaccines, Biologicals (1992)20:11–13.

[40] Thiaucourt F., Lorenzon S., David A., TulasneJ.J., Domenech J., Vaccination against contagiousbovine pleuropneumonia and the use of moleculartools in epidemiology, Ann. NY Acad. Sci. (1998)849:146-151.

[41] Thiaucourt F., Van der Lugt J.J., Provost A.,Contagious bovine pleuropneumonia, in: Infectiousdiseases of livestock., J.A. Coetzer, R.C. Tustin (Eds.),

Oxford University Press Southern Africa, Cape Town,pp. 2045–2059.

[42] Tourasse N.J., Helgason E., Okstad O.A., HegnaI.K., Kolstø A.B., The Bacillus cereus group: novel as-pects of population structure and genome dynamics, J.Appl. Microbiol. (2006) 101:579–593.

[43] Vilei E.M., Abdo E.M., Nicolet J., Botelho A.,Goncalves R., Frey J., Genomic and antigenic differ-ences between the european and African/Australianclusters of Mycoplasma mycoides subsp. mycoides SC,Microbiology (2000) 146:477–486.

[44] Vilei E.M., Frey J., Genetic and biochemicalcharacterization of glycerol uptake in Mycoplasmamycoides subsp. mycoides SC: its impact on H2O2production and virulence, Clin. Diagn. Lab. Immunol.(2001) 8:85–92.

[45] Vilei E.M., Korczak B.M., Frey J., Mycoplasmamycoides subsp. capri and Mycoplasma mycoidessubsp. mycoides LC can be grouped into a single sub-species, Vet. Res. (2006) 37:779–790.

[46] Vilei E.M., Nicolet J., Frey J., IS1634, a novelinsertion element creating long, variable-length di-rect repeats which is specific for Mycoplasma my-coides subsp. mycoides small-colony type, J. Bacteriol.(1999) 181:1319–1323.

[47] Wesonga H., Thiaucourt F., Experimental studieson the efficacy of T1sr and T1/44 vaccine strains ofMycoplasma mycoides subsp. mycoides SC against afield isolate causing contagious bovine pleuropneumo-nia in Kenya- effect of a revaccination, Rev. Elev. Med.Vet. Pays Trop. (2000) 53:313–318.

[48] Westberg J., Persson A., Holmberg A., GoesmannA., Lundeberg J., Johansson K.E., Pettersson B.,Uhlen M., The genome sequence of Mycoplasma my-coides subsp. mycoides SC type strain PG1T, thecausative agent of contagious bovine pleuropneumonia(CBPP), Genome Res. (2004) 14:221–227.

[49] Willems L., Mémoires sur la pleuropneumonieépizootique du gros bétail, Recl. Méd. Vét. Pratique,Maisons Alfort, France, (1852) 3:401–434.

[50] Yaya A., Golsia R., Hamadou B., Amaro A.,Thiaucourt F., Essai comparatif d’efficacité de deuxsouches vaccinales T1/44 et T1sr contre la péripneu-monie contagieuse bovine, Rev. Elev. Med. Vet. PaysTrop. (1999) 52:171-179.

[51] Yaya A., Hamadou B., Yaya D., Abdoulkadiri S.,Thiaucourt F., Inoculation expérimentale de l’agent dela péripneumonie contagieuse bovine à des chèvres,Rev. Elev. Med. Vet. Pays Trop. (2000) 53:319–324.


Documents

Genotyping of Mycoplasma mycoides subsp. mycoides SC