145J Clin Pathol 1993;46:145-148

New phenotypic typing scheme for group Bstreptococci

S R Heard, J A Mawn

AbstractAims: To develop a new typing system forgroup B streptococci based on 35S-methio-nine-labelled protein profiles of bacterialproteins.Methods: 377 clinical isolates of group Bstreptococci were examined by incorpora-tion of 35S-methionine into bacterial pro-teins under strict anaerobic conditions.After sodium dodecylsulphatepolyacryl-amide gel electrophoresis, autoradi-ography was performed. The patternsproduced were visually analysed and cate-gorised into clusters of organisms basedon the pattern of band productionbetween 32-46 kilodaltons.Results: 294 of the typed strains classifiedinto seven different groups designated a-g. 32 strains failed to incorporate 35S-methionine sufficiently to be grouped and11 strains did not fall into one of the sevenidentified groups. Typability, reproduc-ibility, and discrimination of the systemwas evident.Conclusions: This typing system may helpto distinguish between colonising andinvasive strains ofthe organism.

( Clin Pathol 1993;46:145-148)

Department ofMedical Microbiology,St Bartholomew'sHospital MedicalCollege, WestSmithfield, LondonECIA 7BES R HeardJ A MawnCorrespondence to:Dr S R HeardAccepted for publication7 August 1992

Group B streptococci are commensals of thegastrointestinal and female genital tracts. Upto 28% of pregnant women may carry thisorganism in the vagina.' It is an importantneonatal pathogen, responsible for both earlyand late onset neonatal sepsis which is char-acterised by a potentially life-threateningpneumonia, septicaemia, or meningitis. Theincidence of neonatal disease varies, with twoto three cases per 1000 live births beingreported from the United States2 and 0 3 per1000 live births in the United Kingdom.4Prematurity and low birthweight, prolongedrupture of membranes, and maternal pyrexiaare recognised as important risk factors insusceptible mother-infant pairs.' Maternalsepsis accounts for 10-20% of septicaemiasafter delivery in the United States: group Bstreptococci is the second most commonblood culture isolate in this setting.6 Thepathogenesis of early onset neonatal disease isrelated to the degree of genital colonisation inthe mother7 which may be reduced by pro-phylaxis with chlorhexidene during delivery,8and to additional risk factors which con-tribute to the potential for the organism tocause invasive disease.9 The epidemiology of

late onset disease is less well defined but bothmaternal and nosocomial sources have beenimplicated.'0 Cross-infection resulting innosocomial outbreaks in nurseries has alsobeen well documented."I

Current typing methods for the organismare based on a reference laboratory typingscheme using serotyping followed, if neces-sary, by phage typing. This combinedapproach may be required because althoughmost clinical strains can be serotyped, thissystem lacks sufficient discrimination, andless than 80% of strains can be typed usingbacteriophages alone.'2 Serological classifica-tion is based on Lancefield's original descrip-tion of two cell wall carbohydrate antigens,the group specific common antigen carried byall strains, and the type specific capsular anti-gen. Clinical strains are currently dividedserologically according to the presence of spe-cific antigen into seven groups (Ia, Ib, Ia/c, II,III, IV, VI).13 Groups I, II, and III are associ-ated with early onset neonatal disease, whiletype III causes 90% of all cases of both earlyand late onset meningitis and 90% of all casesof late onset disease.'4 In addition to the poly-saccharide antigens, C proteins on the surfaceof group B streptococci form a complex com-posed of at least four different components(a, fi, y and 5) which are found on all Ibstrains and in up to two thirds of type II andIII strains.'5 Additional protein antigens,termed R and X, have also been described.These are primarily associated with type IIand III strains.'6 An epidemiological studyusing restriction endonuclease analysis hasrecently demonstrated the usefulness of thistechnique in examining recurrent invasivedisease'7 and in outbreaks,'8 but it has notbeen formulated into a typing system.

Neither serological nor phage typing sys-tems can distinguish between colonisingstrains and those which have the potential tocause invasive disease in neonates. It is clearthat host immunological factors, includingthe amount of type specific protective anti-body that is present, are of critical importancein the development of invasive disease,'9 butpathogenic factors related to particular strainsof invasive organisms are still not welldefined.We were specifically interested in investi-

gating the role of bacterial proteins in strainspeciation as there is evidence that at least theC protein is a likely virulence factor and may,in fact, be a potentially good candidate for avaccine.15 20 21 The immunological study of thedisease has certainly identified one aspect of

on April 26, 2022 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.46.2.145 on 1 February 1993. D

ownloaded from

Heard, Mawn

invasive disease, but it may be that character-istics of the organism itself can account, inpart, for its clinical behaviour. To date, sero-logical, phage, and molecular attempts- atclassification systems have failed to be of ben-efit in identifying potentially "high-risk" inva-sive strains. The typing system presented hereattempts to classify the organism at the phe-notypic level and is based on the developmentof a typing scheme using 35S-methionineincorporation into bacterial proteins.

MethodsSequential group B streptococci isolatesgrown under standard laboratory conditionsfrom high vaginal swabs received in the clini-cal laboratory were purified and stored inglycerol broth at - 70°C. After detailedexperimental conditions for the optimalincorporation of 35S-methionine into group Bstreptococci had been ascertained, the mostimportant of which is the requirement forstrict anaerobic conditions, the stored strainswere inoculated on to pre-reduced blood agarand Islam's media (Oxoid, Basingstoke) andwere incubated in an anaerobic chamber(10% H2, 10% CO2, 80% N2; Don WhitleyInstrumentation) for 24 hours. 5S-methion-ine (Amersham International) at a final con-centration of 10 p Ci was added to 100 ul

.

Figure 1 Autoradiographs of seven standard strains ofgroup B streptococci. Lstrains a, b, c; lane 4: molecular weight markers; lanes 5-8: strains d, e, f, g.

aliquots of sterile pre-reduced methioninefree modified Eagle's medium (FlowLaboratories) which was inoculated with astandard loop of group B streptococci calcu-lated to 20 pg of protein., The mixture wasthen incubated for two to eight hours at 37°Cunder strict anaerobic conditions. An equalvolume of double strength cracking buffer(2% 2-mercaptoethanol, 20% sodium dode-cylsulphate (SDS), and 40% glycerol) wasthen added to stop the incorporation. Afterboiling for 5 minutes and cooling, 35,pl sam-ples were loaded into sample lanes of a 4%stacking gel and electrophoresed into a 10%SDS-acrylamide (BDH Laboratories) resolv-ing gel. '4C-methylated molecular weightmarkers (Range: 14-3-200 kilodaltons) to afinal activity of 0-1 p Ci were run with eachgel, and as the system became clarified,appropriate specific controls were also used.Eight standard NCTC strains were alsoanalysed using this typing methodology.Twenty five per cent of the strains were typedbetween two and four times to assess thereproducibility of the method.

Electrophoresis was performed using aProtean II dual vertical slab gel with 16 cmplates (Biorad Laboratories, Richmond,California) at 4°C. A constant current of 35mA was applied to run the samples throughthe stacking gel, after which the current wasincreased to 55 mA for 3-4 hours. The gelswere fixed (10% acetic acid, 10% iso-propanol), dried, and autoradiographed.Analysis of the autoradiographs was per-formed independently by two researchers(SRH and JM) by visual analysis of the bandswhich were within the molecular weight range30-46 kilodaltons.

ResultsThree hundred and thirty seven strains werestudied. A comparative visual analysis of thebands in the molecular weight range 30-46kilodaltons showed that 294 (87%) of strainscould be classified into seven groups desig-nated a-g. Consideration of bands above orbelow this region complicated the classifica-tion and did not contribute to reliable dis-crimination among strains. Thirty.two strains(9%) failed to incorporate the 35S-methioninesufficiently to be grouped and 11 strainscould not be classified into one of the sevengroups. The 73 strains typed on more thanone occasion gave the same result each time.The groups distinguished by a distinctive

pattern on 35S-methionine protein analysis aredesignated as a, b, c, d, e, f and g.Autoradiographs showing these seven groupsare shown in fig 1. The distribution of the294 typed strains within the seven strains isshown in the table. Figure 2 comprisesautoradiographs of eight standard NCTCserotyped strains of group B streptococci.These show that different serogroups may fallwithin the same 35S-methionine group (strainstyping as serogroup VI (lane 3) and III (lane

anes 1-3: 10) both type as f, and conversely, that thesame serogroup may have different 35S-

146

....

on April 26, 2022 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.46.2.145 on 1 February 1993. D

ownloaded from

New phenotypic typing scheme for group B streptococci

MW(KD) 1 2

200

97-4

69

46

30

Figure 2 Autoradiograp8182 (Ic), type f; lane 3:lane 5: NCTC 8185 (III,8: NCTC 8180 (Ib), typo(III), type f.

to take a small number of strains classified by3 4 5 6 7 8 9 10 existing systems and apply the new system to

these as a means of assessment.2" Our view,however, is that a systematic approach ofinvestigating a large number of clinical iso-

A lates should be undertaken to define a com-prehensive typing system with the potential touse it for both epidemiological and analyticalpurposes.

In the case of group B streptococci, exist-ing classifications are based on serotyping sys-tems which identify type specific cell wallcarbohydrate groups with some protein anti-gens, and on a defined phage typing system.24Restriction endonuclease analysis (REA) hasrecently been performed on 54 isolates andseems to help in discriminate between sero-logically similar strains, but produced 28 dif-ferent REA patterns.'8 A second study notedtwo particular restriction digest patterns fromnine of 10 strains of neonates with early onsetdisease.25 This methodology has not, how-ever, been developed into a coherent typing

__ system and its very complexity suggests that itwould have limited use for this purpose.

As there is evidence that bacterial proteinsare involved in the pathogenesis of invasivegroup B streptococci disease, a typing schemewhich is based on protein analysis is appro-priate.'5 35S-methionine analysis has been

)hs of serotyped NCTC strains. Lane 1: type d; lane 2: NCTC usiaed"forepiemhionoiceaallprpses tosienvsiNCTC 8188(VI), typef; lane 4: molecular weight markers; used for epidemiological purposes to investi-), type c; lane 6: type e; lane 7: NCTC 9993 (Ia), type e; lane gate outbreaks, but only in the case ofe b; lane 9: NCTC 8100 (Ic), type b; lane 10: NCTC 8184 Clostridium difficile has the technique been

used to develop a comprehensive typingscheme26 which has been applied clinically.27

Less than 10% of strains we tested couldnot be typed due to failure of incorporation of

methionine types (strains 8182 (lane 2) and the "S-methionine. Group B streptococci8100 (lane 9), which are both Ic strains, fall grows better under anaerobic conditions andinto groups f and b, respectively). Figure 3 also produces both increased haemolysis andshows previously untyped strains in relation more of its characteristic tan pigment onto the seven standard strains in order to starch medium if it is grown anaerobically.demonstrate the typing system "in use". Most Initial difficulties with 35S-methionine incor-strains were readily classifiable, but 3% of poration into the bacterial proteins were over-strains could not be classified within the come by performing the incorporation underseven existing standard groups, as shown in strict anaerobic conditions.lanes 7 and 9, a mother and baby pair. There was good discrimination among

the groups with only 3% of strains not allo-cated to one of the seven identified groups.

Discussion These are likely to form the basis for furtherThe development of a new typing scheme for extension and refinement of the system espe-an organism is time consuming and labori- cially as it is applied over a wider geographi-ous. For epidemiological purposes, many cal area.28 The system also allows for goodworkers now use a variety of comparative reproducibility: the 73 strains which weremethods to assess whether outbreak strainsare the same or different, without placingthem within the context of a well defined typ-ing system.'822 Another possible approach is

Distribution of typed group B streptococci strains in each35S-methionine group

3SS Group No of strains (%)

a 54 (18-4)b 48 (16-3)c 44 (14-9)d 35(119)e 30 (10-2)f 46 (157)g 37 (12-6)

typed on more than one occasion producedthe same result. Reproducibility is more read-ily achieved by using appropriate controls onthe typing gel. With the current complementof seven standard strains, seven standardstrains, a molecular weight marker, and seventest strains can all be run on a single gel (fig3). The strains analysed in this report havenot been subject to serotyping because thebasis of the serotyping system is the polysac-charide component of the organism. As anti-cipated, there was no relation between theserogroup and the 35S-methionine groupamong the serotyped NCTC strains exam-ined. In 35 additional strains which havebeen both serotyped and typed by the re-

147

on April 26, 2022 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.46.2.145 on 1 February 1993. D

ownloaded from

Heard, Mawn

Figure 3 Autoradiographs of standard group B streptococci strains (lanes 1, 3, 5, 8, 10,

12) and test strains (lanes 2, 4, 7, 9, 11). Lane 1: type b; lanes 2-4: type ; lane 5: type

d; lane 6: molecular weight markers; lanes 7 and 9: untypable; lane 8: type e; lane 10:

type f; lanes 11 and 12: type g.

ported system, again no relation was apparent

(personal observations).The development of a new typing scheme

based on protein analysis is needed to furtherour understanding of the epidemiology ofgroup B streptococci and to attempt to iden-tify strains which could be potentially patho-genic. Existing typing schemes are neithersufficiently discriminatory nor comprehensiveto offer a solution to this problem. If the fun-damental clinical issue is How can we deter-mine which of the 30% of pregnanciescolonised with group B streptococci are atrisk from invasive disease in the neonate or

mother?, then several questions, includingthose related to the immunology of the dis-ease,19 pathogenicity related to structure,

adhesion,29 and possibly toxin production,'0and the identification of strains which can

both colonise and invade, will need to be con-

sidered. The35S-methionine system reportedhere is currently being used to study invasivestrains of the organism to consider whetherthis phenotypic classification can be of clini-cal value in determining which pregnanciesrequire early intervention for carriage of inva-sive group B streptococci strains.

We gratefully acknowledge the support of the Joint ResearchBoard and the Special Trustees of St Bartholomew's Hospital.

1 Hastings MJG, Easmon CSF, Neill J, Bloxham B, RiversRP. Group B streptococcal colonization and the out-come of pregnancy. J Infect 1986;12:23-9.

2 Anthony BF, Okada DM. The emergence of group Bstreptococci in infections of the newbomn infant. AnnRev Med 1977;28:355-69.

3 Dillon HC Jr, Khare S, Gray BM. Group B streptococcalcarriage and disease: a 6-year prospective study. J7Paediatr 1987;110:31-6.

4 Maynon-White RT. The incidence of neonatal group Bstreptococcal disease in Britain. In: Holm SE,Christensen P, eds. Basic concepts of streptococci and strep-tococcal disease. Chertsey: Reedbooks 1982;305-6.

5 Boyer KM, Gadzala CA, Kelly PD, Gotoff SP. Selectiveintrapartum chemoprophylaxis of neonatal group Bstreptococcal early-onset disease. Jf Infect Dis 1983;148:810-16.

6 Faro S. Group B beta-hemolytic streptococci and puer-peral infections. AmJ Obstet Gynecol 1981;139:686-9.

7 Ancona RJ, Ferrieri P, Williams PP. Maternal factors thatenhance the acquisition of group B streptococci by new-born infants. J Med Microbiol 1980;13:273-80.

8 Burman LG, Christensen P, Christensen K, et al.Prevention of excess neonatal morbidity associated withgroup B streptococci by vaginal chlorhexidine disinfec-tion during labour. Lancet 1992;340:65-9.

9 Baker CJ, Edwards MS. Group B streptococcal infections.In: Remington JS, Klein JO, eds. Infectious diseases of thefetus and newborn infant. Third edn Philadelphia: WBSanders, 1990:742-811.

10 Anthony BF, Okada BM, Hobel CJ. Epidemiology of thegroup B streptococcus: maternal and nosocomialsources for infant acquisitions. J Pediatr 1979;95:431-6.

11 Easmon CSF, Hastings MJG, Clare AJ, et al. Nosocomialtransmission of group B streptococci. Br Med Jf 1981;283:459-61.

12 Colman G. Typing of Streptococcus agalactiae(Lancefield Group B). Eur J Clin Microbiol Infect Dis1988;7:226-31.

13 Henrichen J, Ferrieri P, Jelinkova J, Kohler W, MaxtedWR. Nomenclature of antigens of group B streptococ-cus. IntJ Syst Bactyid 1984;34:500.

14 Milder CJJ, Aanen HC. Neonatal group B streptococcalmeningitis. Arch Dis Child 1984;59:439-43.

15 Chun CSY, Brady U, Boyle MDP, Dillon HC, AyoutEM. Group B streptococcal C-protein associated anti-gens: association with neonatal sepsis. J Infect Dis 1991;163:786-91.

16 Flores AE, Ferrieri P. Molecular species of R-protein pro-duces by clinical isolates of group B streptococci.JfClinMicrobiol 1989;27: 1050-4.

17 Denning DW, Bressack M, Troup NJ, Tompkins LS.Infant with two relapses of group B streptococcal sepsisdocumented by DNA restriction enzyme analysis.Paediatr Infect Dis 1988;7:729-32.

18 Denning DW, Baker CJ, Troup NJ, Tompkins LS.Restriction endonuclease analysis of human and bovinegroup B streptococci for epidemiologic study.Jf ClinMicrobiol 1989;27: 1352-6.

19 Feldman RG, Fleer A.The immune response to the groupB streptococcus. Rev Med Microbiol 1992;3:52-8.

20 Lancefield RC, McCarty M, Everly WN. Multiple mouse-protective antibodies directed against group B strepto-cocci. JExp Med 1975;142:165-79.

21 Madoff LC, MichelJL, Kasper DL. A monoclonal anti-body identifies a protectiveC-protein alpha-antigen epi-tope in group B streptococci. Infect Immun 1991;59:204-10.

22 Poxton IR, Aronsson B, Molby R, NordCE,Collee JG.Immunochemical fingerprinting ofClostridium difficilestrains isolated from an outbreak of antibiotic-associatedcolitis and diarrhoea.3MedMicrobiol 1984;17:317-24.

23Cookson BD, Stapleton P, Ludlam H. Ribotyping ofcoagulase-negative staphylococci. J Med Microbiol 1992;36:414-19.

24 Stringer J. The development of a phage-typing system forgroup B streptococci. J Med Microbiol 1980;13: 133-43.

25 Nagano Y, Nagano K, TakahashiS, et al. Restrictionendonuclease digest patterns of chromosomal DNAfrom group B beta-haemolytic streptococci. J MedMicrobiol 1991 ;35:297-303.

26 TabaqchaliS, O'FarrellS, Holland D, Silman R. Typingscheme forClostridium difficile. Lancet 1984;ii:935-8.

27 HeardS, O'FarrellS, Holland D,CrookS, Barnett MJ,Tabaqchali S.The epidemiology ofClostridium difficilewith use of a typing scheme: nosocomial acquisition andcross-infection among immunocompromised patients. JInfect Dis 1986;153:159-62.

28 Tabaqchali S. Epidemiologic markers ofClostridium diffi-cile. Rev Infect Dis 1990;12(Suppl 2):S192-S9.

29 Broughton RA, BakerCJ. Role of adherence in the patho-genesis of neonatal group B streptococcal infection.Infect Immun 1983;39:837-43.

30 HellerquistCG, Sundell H, Gettins P. Molecular basis forgroup B beta-haemolyticstreptococcal disease. Proc NatlAcad Sci USA 1987;84:51-5.

148

on April 26, 2022 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.46.2.145 on 1 February 1993. D

ownloaded from