JOURNAL OF CLINICAL MICROBIOLOGY,0095-1137/98/$04.0010

Apr. 1998, p. 866871 Vol. 36, No. 4

Copyright 1998, American Society for Microbiology

Molecular Identification of Gemella Species fromThree Patients with Endocarditis

BERNARD LA SCOLA AND DIDIER RAOULT*

Unite des Rickettsies, CNRS UPRESA 6020, Faculte de Medecine, Universite de laMediterrannee, 13385 Marseille Cedex 05, France

Received 31 July 1997/Returned for modification 10 October 1997/Accepted 30 December 1997

Gemella morbillorum and Gemella haemolysans are opportunistic pathogens which cause endocarditis andother severe infections. We report on three patients with endocarditis, one with endocarditis caused by G.haemolysans and two with endocarditis caused by G. morbillorum. The paucity of reports concerning thesebacteria is probably related to the difficulties associated with their identification. For example, one of thestrains reported in this study was originally sent to our laboratory with a preliminary characterization as ashort gram-negative coccobacillus, highlighting the specific problem associated with Gram staining of thesebacteria. The usefulness of 16S rRNA gene amplification, partial sequencing, and comparison of the nucleotidesequence to those in databases when standard phenotypic identification schemes are not helpful is emphasized.We also suggest that the use of simple tests, such as testing susceptibility to vancomycin for gram-negativebacteria and colistin for gram-positive bacteria, could prevent misinterpretation of Gram staining in gram-variable bacteria such as Gemella spp.

Gemella morbillorum and Gemella haemolysans are gram-positive coccal commensal organisms of the mucous mem-branes of humans and other warm-blooded animals. However,as opportunistic pathogens, gemellae are able to cause se-vere localized and generalized infections.

The cases of Gemella infection reported to date have beenpredominantly endovascular infections (1, 6, 812, 15, 22, 25,28, 29, 31, 3335, 39, 40, 43, 45). Among these cases of Gemellainfection, most are endocarditis, usually associated with previ-ous valvular damage and/or poor dental state. Central nervoussystem and skeletal infections have also been described (4, 13,23, 36, 38, 48). In a study of 52 cases of streptococcal endo-carditis, gemellae represented 6% of the viridans group strep-tococci and 5% of all isolates (17). In our hospital center,gemellae were the cause of 6% of the cases of endocarditis dueto nonstaphylococcal gram-positive cocci, were the cause of13.3% of the cases of endocarditis caused by viridans groupstreptococci, and represented 2.9% of all bacterial isolatescausing 70 cases of endocarditis diagnosed during a 3-yearperiod (unpublished data).

These bacteria are easily decolorized during Gram stainingand sometimes appear as elongated cells, explaining why theywere first described as Neisseria (46). They may also be moreinvolved in clinical disease than is presently recognized, be-cause they can be incorrectly identified as viridans group strep-tococci or left unidentified (42).

In this paper we report on two patients with G. morbillorumendocarditis and one patient with G. haemolysans endocarditis.The identification of one G. morbillorum strain was achievedwith the help of partial 16S rRNA gene sequencing, since thebacterial isolate appeared to be gram negative. Partial 16SrRNA gene sequencing was also used as a reference techniqueto confirm the identities of a further two strains of Gemella.The results obtained by electron microscopy and analysis of

cell wall fatty acids are reported, and the previous cases ofendocarditis caused by G. haemolysans and G. morbillorum arereviewed.

CASE REPORTS

Patient 1. A 63-year-old man who was a heavy smoker withchronic obstructive bronchitis and a very poor dental state wasadmitted to hospital complaining of intermittent fever, loss ofweight, and anorexia over a 1-month period. He had no historyof rheumatic disease, although a moderate heart murmur hadbeen discovered 1 year previously, and at that time an echo-cardiogram had yielded moderate mitral valve regurgitation.On initial examination, the patient had an increased heartmurmur, a temperature of 38.5C, and a pulse of 100 beats/min. An ejection systolic murmur was heard at the apex of theheart. Laboratory investigations showed a hemoglobin concen-tration of 100 g/liter, an erythrocyte sedimentation rate of 97mm/h, and a leukocyte count of 8.53 3 109/liter (77% neutro-phils). A transesophageal echocardiogram demonstrated thepresence of vegetation on the mitral valve with moderate mi-tral valve regurgitation. Six sets of blood samples for cultureswere taken on admission and the following day, and the bloodsamples were inoculated into BACTEC aerobic (NR 6-A*)and anaerobic (NR-7A*) bottles in a BACTEC NR-860 auto-mated instrument (Becton Dickinson Diagnostic InstrumentSystems, Sparks, Md.). All yielded slowly growing, gram-posi-tive cocci, subsequently identified as G. haemolysans. A kidneyechography, carried out for microscopic hematuria, yielded alesion inside of the left kidney which was compatible with anabscess. The patients treatment began the day after his admis-sion, in which treatment with amoxicillin (4 g intravenously at6-h intervals) and amikacin (5 mg/kg of body weight intrave-nously at 8-h intervals) was begun. His condition improvedrapidly, and after 2 weeks of this regimen, he underwent car-diac surgery in order to remove the motile vegetation. Oneweek after surgery antibiotic therapy was discontinued. After 2years of follow-up he remains well.

Patient 2. A 74-year-old man with a history of Potts diseasein infancy, chronic alcoholism, and a poor dental state was

* Corresponding author. Mailing address: Unite des Rickettsies,CNRS UPRESA 6020, Faculte de Medecine, Universite de la Medi-terrannee, 27 Blvd. Jean Moulin, 13385 Marseille Cedex 05, France.Phone: (33).4.91.38.55.17. Fax: (33).4.91.83.03.90. E-mail: Didier.Raoult@medecine.univ-mrs.fr.

866

on June 9, 2015 by guest

http://jcm.asm.org/D

ownloaded from

admitted to hospital complaining of intermittent fever, sweat-ing, loss of weight, and basithoracic pain over a 3-month pe-riod. He had neither a history of rheumatic disease nor aprevious heart murmur. On initial examination, the patient hada diastolic heart murmur, a temperature of 38C, and a pulse of100 beats/min. Laboratory investigations showed a hemoglobinconcentration of 95 g/liter, an erythrocyte sedimentation rateof 63 mm/h, and a leukocyte count of 12.8 3 109/liter (87%neutrophils). A transesophageal echocardiogram demon-strated aortic valve incompetence but failed to show any veg-etation. Six sets of blood samples for culture were taken onadmission and the following day, and the blood samples wereinoculated into BACTEC aerobic (NR 6-A*) and anaerobic(NR-7A*) bottles in a BACTEC NR-860 automated instru-ment. All yielded slowly growing, gram-variable cocci and coc-cobacilli subsequently identified as G. morbillorum. The pa-tients treatment began the day after his admission andcomprised amoxicillin (4 g intravenously at 6-h intervals) andgentamicin (1 mg/kg intravenously at 8-h intervals). Althoughhis condition improved rapidly, it was necessary to transfer himto a cardi-thoracic unit because he was also suffering fromsignificant aortic valve regurgitation and a dilated left ventricle.The aortic valve was successfully replaced with a prostheticdevice, and the patient made an uneventful postoperative re-covery. Gentamicin was discontinued 1 week later, and amoxi-cillin was discontinued 3 weeks later. After 1 year of follow-uphe remains well.

Patient 3. A small, fastidious, gram-negative rod which hadbeen isolated from the blood of a male patient with infectiousendocarditis by using BACTEC aerobic (NR 6-A*) and anaer-obic (NR-7A*) bottles was sent to our laboratory for identifi-cation, because it was thought that it could be a Bartonella sp.The patient was a farmer and suffered from a bicuspid aorticvalve. Clinical symptoms included intermittent fever and aweight loss of 12 kg over a period of several months. Attemptsto amplify citrate synthase and 16S rRNA genes with specificprimers for Bartonella (27) were unsuccessful. Standard phe-notypic characterization methods for gram-negative rods alsofailed to provide an identity. The bacterium was identified asG. morbillorum by 16 rRNA gene sequencing.

MATERIALS AND METHODS

Phenotypic identification. Presumptive identification of the Gemella organ-isms was achieved through assessment of colonial morphology, hemolysis onColumbia agar with 5% sheep blood (BioMerieux, Marcy lEtoile, France),microscopic appearance after Gram staining, and the results of biochemical tests(with the API 20 Strep and the API 20A systems according to the manufacturersinstructions). Growth was also attempted in broth containing 6.5% NaCl. Sus-ceptibility to vancomycin and colistin was assessed with 30-mg vancomycin and50-mg colistin disks (Sanofi Diagnostic Pasteur, Marnes la Coquette, France) andMueller-Hinton broth with 5% sheep blood agar (BioMerieux) by the conven-tional disk diffusion test method (2). The strain was considered to be susceptibleto vancomycin and to colistin when inhibition zone diameters of $12 and $15mm, respectively, were observed.

Cellular fatty acids. Colonies of the three isolates and of the type strains G.haemolysans ATCC 10379 and G. morbillorum ATCC 27824 were grown onTrypticase soy agar with 5% sheep blood (Becton Dickinson) at 38C for 48 h.They were then saponified, and cell wall fatty acids were extracted and analyzedby gas chromatography as reported previously (37).

Processing for electron microscopy. Bacterial cells were harvested from colo-nies grown on Columbia agar and for fixation were suspended in 2.5% glutaral-dehyde in 0.1 M cacodylate buffer (pH 7.2) containing 0.1 M sucrose. The fixedcells were washed overnight with the same buffer and were then fixed for 1 h atroom temperature with 1% osmium tetroxide in 0.1 M cacodylate buffer. Dehy-dration was performed by washing the cells in gradually increased concentrations(25 to 100%) of ethanol. The cells were then embedded in Epon 812. Thinsections were cut from embedded blocks with an LKB Ultratome III microtomeand were poststained with a saturated solution of methanol-uranyl acetate andlead citrate in water before examination on a Jeol JEM 1200 EX electronmicroscope.

PCR amplification and sequencing of the PCR product. DNA extracts, suit-able for use as templates in PCRs, were made from 200 ml of a bacterialsuspension by using a QIA ampBlood kit (Qiagen, Hilden, Germany) accordingto the manufacturers instructions. All primer sets used in this study are listed inTable 1. DNA extracts were amplified by a PCR incorporating universal primersfD1 and rP2 (49) (Eurogentec, Seraing, Belgium). PCR amplifications wereperformed in 100-ml volumes incorporating 10 ml of extracted DNA, 10 ml of 103reaction buffer, 100 mM (each) deoxynucleoside triphosphate, 0.2 mM (each)primer, 2 U of Taq polymerase (Perkin-Elmer Cetus, Norwalk, Conn.), andsterile, distilled water. Amplifications were carried out in a Perkin-Elmer 9600thermal cycler for 35 cycles, with each cycle consisting of denaturation at 95C for30 s, primer annealing at 55C for 30 s, and extension at 72C for 60 s. Thesuccess of the amplification was determined by ethidium bromide staining fol-lowing the resolution of products by 1% agarose gel electrophoresis. Each ex-periment included sterile water (no DNA) as a negative control and Escherichiacoli DNA as a positive control. The products of 16S rRNA gene amplificationwere purified for sequencing by using Microspin S-400 HR columns (PharmaciaBiotech, Uppsala, Sweden) according to the manufacturers instructions. Se-quencing reactions were prepared by using the Amplicycle sequencing kit (Per-kin-Elmer), according to the manufacturers instructions, and one of six 59fluorescein-labelled primers was incorporated into the reaction mixture (Table1). The thermal cycle used in each sequencing reaction was dependent on thebase sequence of the primer (Table 1). When the primer annealing temperaturewas $50C, an initial denaturation step at 95C for 1 min was followed by 25cycles, with each cycle comprising denaturation at 95C for 30 s, annealing for30 s, and extension at 72C for 1 min. Amplification was completed by anextension step of 5 min at 72C to allow complete extension of the amplifiedproducts. When the primer annealing temperature was ,50C, an initial dena-turation step at 95C for 1 min was followed by 30 cycles, with each cyclecomprising denaturation at 95C for 30 s, annealing for 30 s, and extension at60C for 2 min. Ten supplementary cycles of denaturation at 95C for 10 s andextension at 60C for 90 s were performed in order to increase the amount ofpolymerization. Sequencing reactions were carried out on a Perkin-Elmer 9600thermal cycler, and reaction products were resolved by electrophoresis on a 6%polyacrylamide gel incorporated into an ALF Automatic Sequencer (PharmaciaBiotech).

Sequence analysis. Partial 16S rRNA gene sequences derived from each re-action mixture with the ALF manager were combined in a complete sequence byusing PC Gene software (IntelliGenetics Inc.). The complete sequence was thencompared with all bacterial sequences available in the GenBank database, byusing the multisequence comparison program FASTA (part of the BISANCEsoftware package) (14).

RESULTSBacterium from patient 1. The bacterium isolated from pa-

tient 1 was thought to be a strain of G. haemolysans on thebasis of conventional phenotypic identification. Growth waseasily achieved on Columbia agar with incubation at 37Cunder 5% CO2, but only poor growth was obtained underanaerobic conditions. Colonies were small and weakly beta-hemolytic after 5 days of incubation. Microscopic examinationafter Gram staining yielded gram-positive cocci which became

TABLE 1. Summary of primers used for PCR amplification andsequencing in this study

Primerpurpose and

primerSequences (59 to 39)

Primerspecificity orhybridization

temp (C)

PCRfD1 -AGAGTTTGATCCTGGCTCAG- Eubacterial 16S

rRNA generP2 -ACGGCTACCTTGTTACGACTT- Eubacterial 16S

rRNA gene

Sequencinga

fD1 -AGAGTTTGATCCTGGCTCAG- 55rP2 -ACGGCTACCTTGTTACGACTT- 55800f -ATTAGATACCCTGGTAG- 50800r -CTACCAGGGTATCTAAT- 501050f -TGTCGTCAGCTCGTG- 431050r -CACGAGCTGACGACA- 43

a Primers 59 labelled with fluorescein.

VOL. 36, 1998 MOLECULAR IDENTIFICATION OF GEMELLA SPECIES 867

on June 9, 2015 by guest

http://jcm.asm.org/D

ownloaded from

gram-variable after subculture. Cocci were arranged in pairs,short chains, or clusters and were of various sizes. The resultsof biochemical reactions and enzyme analyses are summarizedin Table 2. Sequence data for 1,022 bp of the 16S rRNA genewere obtained. After removing ambiguities, the sequence dem-onstrated 99.9% similarity to the 16S rRNA gene sequence ofG. haemolysans (GenBank accession no. L14326) (50).

Bacterium from patient 2. The bacterium isolated from pa-tient 2 was thought to be a strain of G. morbillorum on the basisof conventional phenotypic identification. Growth was easilyobtained on Columbia agar with incubation at 37C under 5%CO2 and under anaerobic conditions. Colonies were weaklyalpha-hemolytic only when they were incubated under 5%CO2. Microscopic examination after Gram staining yieldedgram-variable cocci and coccobacilli arranged in pairs, shortchains, or clusters and of various sizes. The results of biochem-ical reactions and enzyme analyses are summarized in Table 2.Sequence data for 948 bp of the 16S rRNA gene were ob-tained. After removing ambiguities, the sequence demon-strated 99.8% similarity to the 16S rRNA gene sequence of G.morbillorum (GenBank accession no. L14327).

Bacterium from patient 3. Conventional phenotypic charac-terization of the gram-negative coccobacilli yielded no definiteidentification. Visible growth was obtained after 3 or 4 days onColumbia agar with incubation at 37C under 5% CO2. Thecolonies were small and nonhemolytic. Sequence data for 949bp of the 16S rRNA gene were obtained. After removingambiguities, the sequence demonstrated 99.68% similarity tothe 16S rRNA gene sequence of G. morbillorum (GenBankaccession no. L14327). Conventional biochemical identifica-tion was reassessed. The results are summarized in Table 2 andare compatible with the identification of the bacterium as G.morbillorum with the exception of the results obtained byGram staining. Subsequently, a more complete 16S rRNAgene sequence was determined. This 1,474-bp sequence dem-onstrated 99.73% similarity with the 16S rRNA gene sequenceof G. morbillorum.

Cellular fatty acids. The major fatty acids in cells wereidentified as C16 and C18, but definite identification was notachieved by this technique.

Electron microscopy. All bacteria studied presented with atypical gram-positive membrane organization. Some cells ap-peared as short bacilli, usually when they were in the process ofdividing.

DISCUSSION

G. morbillorum was originally proposed as Diplococcus ru-beolae (47), and a second member of the genus named Diplo-coccus morbillorum was added soon after (41). However, onreappraisal the two Diplococcus species were considered to beidentical and were unified under the name D. morbillorum.Although D. morbillorum could be grown under aerobic con-ditions, it was primarily anaerobic, and on this basis it wastransferred to the genus Peptostreptococcus (44), only to bereclassified into the genus Streptococcus (26). G. haemolysanswas first described in 1938 as Neisseria haemolysans (46) (dem-onstrating its indeterminate Gram staining characteristic), butthe species was reclassified into a new genus, Gemella (7),following demonstration of biochemical differences with otherNeisseria species. The nucleotide sequence of the 16S rRNAgene of Streptococcus morbillorum was found to closely resem-ble that of G. haemolysans (32), and on the basis of DNA-DNAhybridization results and G1C content analysis, it was pro-posed that S. morbillorum be transferred to the genus Gemellaas G. morbillorum comb. nov. (30). The phylogenetic relation-ships of Gemella spp. to other gram-positive bacteria are pre-sented in Fig. 1.

G. morbillorum and G. haemolysans are uncommon causesof infectious endocarditis; a review of the literature revealsonly 10 cases caused by G. morbillorum (1, 6, 10, 12, 29, 33, 35,40, 45) and 12 cases caused by G. haemolysans (8, 9, 11, 15, 22,25, 28, 31, 34, 39, 43). Two additional cases of G. morbillorumendocarditis have been recorded among 52 apparent failures ofendocarditis prophylaxis (17), and 8 cases have been recordedamong 364 cases of streptococcal endocarditis (19). The pa-tients described in Table 3 and Table 4 demonstrate that poordental state and/or dentistry are predisposing factors, as inpatients with endocarditis caused by viridans group strepto-cocci. Previous valvular damage is also a common occurrence.In most cases, the infections were successfully treated withantibiotic therapy, usually benzylpenicillin or amoxicillin asso-ciated with gentamicin.

Gemella spp. possess a typical gram-positive cell wall struc-ture, as confirmed by electron microscopy in this study. How-ever during Gram staining, cells are easily decolorized and maytherefore may appear to be gram variable and even gramnegative. It is likely that Gram staining abnormality and mor-phological polymorphism are responsible for the misidentifi-cation of Gemella spp. and, thus, perhaps for the fact that sofew cases of Gemella infection are reported. Rapid phenotypicidentification systems are unable to identify accurately allstrains of these species (3, 20), even though manufacturershave significantly improved their databases over recent years(e.g., the Rapid ID 32 Strep database [21]).

Nevertheless, for two cases of infection described in thisreport, identification of G. morbillorum or G. haemolysans asthe causative agents was achieved by standard phenotypic iden-tification schemes. Identification of the organism causing in-fection in patient 3 was difficult, however, even though a largenumber of biochemical tests were performed. After its identi-fication had been achieved by the PCR-based approach withthe 16S rRNA gene and the bacterium had been recognized asbeing gram positive, reassessment of phenotypic characteristicsrevealed a correct identity. The same problem had alreadyoccurred with a short, gram-negative bacillus, isolated froma patient with infectious arthritis, which had been sent to ourlaboratory for identification; analysis of its 16S rRNA gene alsorevealed that this isolate was G. morbillorum (unpublisheddata).

In our laboratory, we now routinely use cell wall fatty acid

TABLE 2. Results of phenotypic characterization of strains isolatedin this study

Feature Patient 1,G. haemolysansPatient 2,

G. morbillorumPatient 3,

G. morbillorum

Voges-Proskauer 1 2 2Alkaline phosphatase 1 2 2Leucine aminopeptidase 1 1 1 (weak)Pyrrolidone arylamidase 1 2 2Esculin hydrolysis 2 2 2Acid from glucose 1 1 1Acid from sucrose 1 1 1Acid from maltose 1 1 1Acid from mannitol 1 2 1Acid from mannose 1 1 1Acid from trehalose 2 1 2Growth in 6.5% NaCl 2 2 2Vancomycin susceptibility 1 1 1

868 LA SCOLA AND RAOULT J. CLIN. MICROBIOL.

on June 9, 2015 by guest

http://jcm.asm.org/D

ownloaded from

analysis and/or 16S rRNA gene analysis when a bacterium isnot easily identified by phenotypic schemes. About 0.05 to0.1% of all isolates fall into this category. Amplification withuniversal primers, partial sequencing (about 900 bp) with con-

served primers, and sequence comparison take less than 3 days.If the partial 16S rRNA gene sequence obtained shares morethan 99% similarity with a specific 16S rRNA gene sequence inthe database, the bacterium may be considered to be identi-

FIG. 1. Phylogenetic tree based on 16S rRNA gene sequences available in the GenBank database obtained by the neighbor-joining method. The phylogeneticrelationships of G. haemolysans and G. morbillorum with selected gram-positive bacteria with low G1C contents are shown. The bar represents a 1% difference innucleotide sequence.

TABLE 3. Summary of patients with reported cases of G. morbillorum endocarditis and features of our two patients with G.morbillorum endocarditisa

Patient no./age (yr)/sexa Underlying conditions or source of infection

Infectedvalve Therapy

Valvereplacement Outcome

Reference,year

1/60/m Dental manipulation NA Penicillin, rifampin No Cure 12, 19842/38/m Anal subcutaneous sphincterotomy,

sigmoidoscopyMi Benzylpenicillin,

gentamicin,erythromycin

No Cure 35, 1989

3/39/m Dental manipulation, bicuspid aortic valve Ao Benzylpenicillin,gentamicin

Yes Cure 35, 1989

4/42/m Mi regurgitation Mi Benzylpenicillin,streptomycin

No Cerebral mycotic aneurysm, cure 10, 1990

5/19/m Intravenous drug abuser Tri Flucloxacillin,benzylpenicillin,gentamicin

No Cure (?; reduction in the size ofvegetation)

6, 1992

6/48/m Poor dentition, systolic heart murmur,end-stage renal disease

Mi Vancomycin No Wrist arthritis, cure 40, 1993

7/64/m Dental manipulation Mi, Ao Benzylpenicillin Yes Acute appendicitis, cure 45, 19948/29/f Hypertrophic obstructive cardiomyopathy,

dental manipulationNa Benzylpenicillin,

gentamicin,erythromycin,rifampin

No Cure 29, 1994

9/74/m Left colonic resection for rectal neoplasia Tri Cefuroxime,amikacin,benzylpenicillin

No Pulmonary embolism, cure 1, 1995

10/75/m Rheumatic fever in infancy Mi Benzylpenicillin,gentamicin,teicoplanin,rifampin,erythromycin

No Allergic reaction toantimicrobial agents, cure

33, 1995

11/74/m Poor dentition, chronic ethylism Aob Amoxicillin,gentamicin

Yes Cure Patient 2

12/NA/m Bicuspid aortic valve Ao NA NA NA Patient 3

a Abbreviations: Ao, aortic valve; Mi, mitral valve; Tri, tricuspid valve; NA, not available; m, male; f, female.b No evidence of vegetation on echocardiogram.

VOL. 36, 1998 MOLECULAR IDENTIFICATION OF GEMELLA SPECIES 869

on June 9, 2015 by guest

http://jcm.asm.org/D

ownloaded from

fied, provided that the genes of closely related species share asignificantly lower degree of similarity. However, if this is notthe case, complete identification can be achieved in two ways.First, confirmation can be achieved by a few simple additional

biochemical tests, as described for patient 3 of this study. Thismethod is quick and cheap, but it requires that a bacterium beeasily subcultured. If this is not possible, additional sequencingcan be used in order to determine a complete 16S rRNA genesequence. This second approach is more expensive and time-consuming, but it avoids the need for additional phenotypic orspecific antibody tests, both of which may not be routinelyavailable in the diagnostic laboratory and thus would requirereferral to a reference laboratory. Furthermore, this approachalso allows the description of new pathogens (16) which mayhave been misidentified by standard phenotypic identificationschemes. Finally, the use of this technique for identificationdoes not require an experienced microbiologist and gives auniversal bacterial identification ability to all laboratories, pro-vided that they are equipped with an automated sequencer.With increasing availability and decreasing costs, such equip-ment is likely to become a feature of more and more routinelaboratories in the years to come.

However, for the present, phenotypic characterization re-mains the standard approach to bacterial identification, withGram staining being one of the most important first steps ofmost routine identification schemes (5). A mistake at this stagecan lead to the application of inappropriate tests and thereforeunnecessary delays in processing; thus, it is important that allefforts be made to ensure correct interpretation of the Gramstaining result. The reference method for assessing cell walltype is electron microscopy, and although it is accurate, themethod is expensive, is time-consuming, and requires special-ized equipment. As an alternative we have added vancomycinand colistin to our standard antibiotic susceptibility tests. Mostgram-positive bacteria are susceptible to vancomycin, whereasmost gram-negative bacteria are susceptible to colistin. Such amethod has previously been shown to be beneficial for thedetermination of cell wall type among nonenterobacterial rods(24). The test is easy to perform (Fig. 2) and is inexpensive. Itmust be noted, however, that this test is not definitive, since a

FIG. 2. Examples of the usefulness of vancomycin test to determine theGram reaction for gram-variable bacteria. The vancomycin-susceptible gram-positive bacterium is G. haemolysans (A), and the vancomycin-resistant gram-negative bacteria are Acinetobacter (B), Moraxella sp. (C), and Kingella kingae(D). Arrowheads indicate the limits of the inhibition zone around vancomycin(VA) and colistin (CS) disks.

TABLE 4. Summary of reported cases of G. haemolysans endocarditis and features of our patient with G. haemolysans endocarditisa

Case/age(yr)/sex

Underlying conditions orsource of infection

Infectedvalve Therapy

Valvereplacement Outcome

Reference,year

1/62/m Dental manipulation Mi Cefamandole, gentamicin, benzylpenicillin,amoxicillin

No Spondylodiscitis, cure 11, 1982

2/48/m Poor dentition, respiratory viralsyndrome

Ao Benzylpenicillin, gentamicin Yes Phlebitis, cure 11, 1982

3/56/m Mi insufficiency, dental manipulation Mi Benzylpenicillin, gentamicin No Cure 11, 19824/68/m Aortic insufficiency Ao Benzylpenicillin, streptomycin Yes Cure 31, 19845/47/m Bioprosthetic aortic valve,

periodontal diseaseAo Erythromycin, rifampin Yes Cure 39, 1985

6/62/f None Mi Benzylpenicillin, tobramycin, clindamycin No Cure 28, 19897/74/m Mi prolapse Mi Benzylpenicillin, gentamicin, amoxicillin No Cure 8, 19918/42/m Aortic insufficiency, scalp wound Aob Vancomycin, fusidic acid, amoxicillin,

gentamicinNo Cure 22, 1993

9/71/m Colonic carcinoma Mi Amoxicillin 1 clavulanic acid, gentamicin No Cure 25, 199310/53/f Mi regurgitation Mi Benzylpenicillin, gentamicin Yes Cure 15, 199311/20/m None Ao Benzylpenicillin, gentamicin Yes Right dorsalis pediculis

territory embolus, cure34, 1994

12/34/m Multiple valve replacements withprosthetic valve, infectiveendocarditis, asymptomatic dentalsepsis, benzylpenicillin allergy

Ao Cefuroxime, tobramycin, ciprofloxacin,erythromycin

No Allergic neutropenia due tocefuroxim, cure

43, 1995

13/63/m Chronic obstructive bronchitis, poordentition

Mi Amoxicillin, amikacin Yes Kidney abscess, cure Patient 1

a Abbreviations: Ao, aortic valve; Mi, mitral valve; Tri, tricuspid valve; NA, not available; m, male; f, female.b No evidence of vegetation on echocardiogram.

870 LA SCOLA AND RAOULT J. CLIN. MICROBIOL.

on June 9, 2015 by guest

http://jcm.asm.org/D

ownloaded from

vancomycin-resistant strain of G. haemolysans has been en-countered (18).

In conclusion, for bacteria which are easily identified byphenotypic schemes (especially with the help of commercialidentification kits or simple additional tests) and which repre-sent more than 99.9% of the bacteria isolated in clinical mi-crobiology laboratories, no additional identification techniqueis required. However, 16S rRNA gene analysis is becomingmore competitive in terms of efficiency and accuracy for fas-tidious bacteria or those not easily identified by phenotypicschemes. The use of 16S rRNA gene analysis should also leadto an increase in the number of descriptions of new pathogensand in the recovery of unexpected pathogens.

ACKNOWLEDGMENT

We thank Richard Birtles for correcting the manuscript.

REFERENCES

1. Abboud, R., and A. Friart. 1995. Deux cas dendocardite tricuspidienneisolee apre`s intervention colique. Acta Clin. Belg. 50:242245.

2. Acar, J. F., and F. W. Goldstein. 1996. Disk susceptibility test, p. 151. In V.Lorian (ed.), Antibiotics in laboratory medecine, 4th ed. The Williams &Wilkins Co., Baltimore, Md.

3. Appelbaum, P. C., M. R. Jacobs, J. L. Heald, W. M. Palko, A. Duffett, R.Crist, and P. A. Naugle. 1984. Comparative evaluation of the API 20S systemand the AutoMicrobic System Gram-Positive Identification Card for speciesidentification of streptococci. J. Clin. Microbiol. 19:164168.

4. Aspevall, O., E. Hillebrandt, B. Linderoth, and M. Rylander. 1991. Menin-gitis due to Gemella haemolysans after neurosurgical treatment of trigeminalneuralgia. Scand. J. Infect. Dis. 23:503505.

5. Baron, E. J., A. S. Weissfeld, P. A. Fuselier, and D. J. Brenner. 1995.Classification and identification of bacteria, p. 249264. In P. R. Murray, E. J.Baron, M. A. Pfaller, F. C. Tenover, and R. H. Yolken (ed.), Manual ofclinical microbiology, 6th ed. American Society for Microbiology, Washing-ton, D.C.

6. Bell, E., and A. C. McCartney. 1992. Gemella morbillorum in an intravenousdrug abuser. J. Infect. 25:110112.

7. Berger, U. 1961. A proposed new genus of gram negative cocci: Gemella. Int.Bull. Bacteriol. Nomencl. Taxon. 11:1719.

8. Brack, M. J., P. G. Avery, P. J. B. Hubner, and R. A. Swann. 1991. Gemellahaemolysans: a rare and unusual cause of infective endocarditis. Postgrad.Med. J. 67:210212.

9. Buu-Hoi, A., A. Sapoetra, C. Branger, and J. F. Acar. 1982. Antimicrobialsusceptibility of Gemella haemolysans isolated from patients with subacuteendocarditis. Eur. J. Clin. Microbiol. 1:102106.

10. Calopa, M., F. Rubio, M. Aguilar, and J. Peres. 1990. Giant basilar aneurysmin the course of subacute bacterial endocarditis. Stroke 21:16251627.

11. Chatelain, R., J. Croize, P. Rouge, C. Massot, H. Dabernat, J. C. Auvergnat,A. Buu-Hoi, J. P. Stahl, and F. Bimet. 1982. Isolement de Gemella haemo-lysans dans trois cas dendocardites bacteriennes. Med. Mal. Infect. 1:2530.

12. Coto, U., and S. L. Berk. 1984. Endocarditis caused by Streptococcus mor-billorum. Am. J. Med. Sci. 287:5457.

13. Debast, S. B., R. Koot, and J. F. Meis. 1993. Infections caused by Gemellamorbillorum. Lancet 342:560.

14. Dessen, P., C. Frondat, C. Valencien, and G. Munier. 1990. BISANCE: aFrench service for access to biomolecular sequences databases. Comput.Appl. Biosci. 6:355356.

15. Devuyst, O., P. Hainaut, J. Gigi, and G. Wauters. 1993. Rarity and potentialseverity of Gemella haemolysans endocarditis. Acta Clin. Belg. 48:5253.

16. Drancourt, M., L. Niel, and D. Raoult. 1995. Unknown multiresistant gramnegative bacterium causing meningitis in HIV-positive patient in Africa.Lancet 346:1168.

17. Durack, D. T., E. L. Kaplan, and A. L. Bisno. 1983. Apparent failures ofendocarditis prophylaxis. JAMA 250:23182322.

18. Efstratiou, A., D. Morrisson, and N. Woodford. 1993. Glycopeptide-resistantGemella haemolysans from blood. Lancet 342:927928.

19. Facklam, R. R. 1977. Physiological differentiation of viridans streptococci.J. Clin. Microbiol. 19:184201.

20. Facklam, R. R., D. L. Rhoden, and P. B. Smith. 1984. Evaluation of theRapid Strep system for the identification of clinical isolates of Streptococcusspecies. J. Clin. Microbiol. 20:894898.

21. Freney, J., S. Bland, J. Etienne, M. Desmonceaux, J. M. Boeufgras, and J.

Fleurette. 1992. Description and evaluation of the semiautomated 4-hourRapid ID 32 Strep method for identification of streptococci and members ofrelated genera. J. Clin. Microbiol. 30:26572661.

22. Fresard, A., V. P. Michel, X. Rueda, G. Aubert, G. Dorche, and F. Lucht.1993. Gemella haemolysans endocarditis. Clin. Infect. Dis. 16:586587.

23. Garavelli, P. L. 1990. Meningite da Streptococcus morbillorum. MinervaMed. 81:69.

24. Graevenitz, A., and C. Bucher. 1983. Accuracy of the KOH and vancomycintests in determining the Gram reaction of nonenterobacterial rods. J. Clin.Microbiol. 18:983985.

25. Helft, G., X. Tabone, J. P. Metzger, and A. Vacheron. 1993. Gemella hae-molysans endocarditis with colonic carcinoma. Eur. J. Med. 2:369370.

26. Holdeman, L. V., and W. E. C. Moore. 1974. New genus Coprococcus, twelvenew species, and emended description of four previously described speciesfrom human feces. Int. J. Syst. Bacteriol. 24:260277.

27. Joblet, C., V. Roux, M. Drancourt, J. Gouvernet, and D. Raoult. 1995.Identification of Bartonella (Rochalimea) species among fastidious gram-negative bacteria on the basis of partial sequence of the citrate-synthasegene. J. Clin. Microbiol. 33:18791883.

28. Kaufhold, A., D. Franzen, and R. Lutticken. 1989. Endocarditis caused byGemella haemolysans. Infection 17:385387.

29. Kerr, J. R., C. H. Webb, J. G. McGimpsey, and N. P. S. Campbell. 1994.Infective endocarditis due to Gemella morbillorum complicating hypertro-phic obstructive cardiomyopathy. Ulster Med. J. 63:108110.

30. Kilpper-Balz, R., and K. H. Schleifer. 1988. Transfer of Streptococcus mor-billorum to the genus Gemella as Gemella morbillorum, comb. nov. Int. J.Syst. Bacteriol. 38:442443.

31. Laudat, P., P. Cosnay, B. Icole, A. Raoult, P. Raynaud, and M. Brochier.1984. Endocardite a` Gemella haemolysans: une nouvelle observation. Med.Mal. Infect. 4:159161.

32. Ludwig, W., M. Weizenegger, R. Kilpper-Balz, and K. H. Schleifer. 1988.Phylogenetic relationships of anaerobic streptococci. Int. J. Syst. Bacteriol.38:1518.

33. Martin, M. J., D. A. Wright, and A. R. R. Jones. 1995. A case of Gemellamorbillorum endocarditis. Postgrad. Med. J. 71:188.

34. Matsis, P. P., and R. N. Easthorpe. 1994. Gemella haemolysans endocarditis.Aust. N. Z. J. Med. 24:417418.

35. Maxwell, S. 1989. Endocarditis due to Streptococcus morbillorum. J. Infect.18:6772.

36. May, T., C. Amiel, C. Lion, M. Weber, A. Gerard, and P. Canton. 1993.Meningitis due to Gemella haemolysans. Eur. J. Clin. Microbiol. Infect. Dis.12:644645.

37. Miller, L., and T. Berger. 1985. Bacterial identification by gas chromatog-raphy of whole cell fatty acids. Hewlett-Packard application note 228241.Hewlett-Packard, Avondale, Pa.

38. Mitchell, R. G., and P. J. Teddy. 1985. Meningitis due to Gemella haemo-lysans after radiofrequency trigeminal rhizotomy. J. Clin. Pathol. 38:558560.

39. Morea, P., M. Toni, M. Bressan, and P. Stritoni. 1991. Prosthetic valveendocarditis by Gemella haemolysans. Infection 19:446.

40. Omran, Y., and C. A. Wood. 1993. Endovascular infection and septic arthritiscaused by Gemella morbillorum. Diagn. Microbiol. Infect. Dis. 16:131134.

41. Prevot, A. R. 1933. Etudes de sytematique bacterienne. I. Lois generals. II.Cocci anaerobies. Ann. Sci. Nat. 15:23260.

42. Ruoff, K. L. 1990. Gemellae: a tale of two new species (and five genera). Clin.Microbiol. Newsl. 12:14.

43. Samuel, L., and P. Bloomfield. 1995. Gemella haemolysans prosthetic valveendocarditis. Postgrad. Med. J. 71:188.

44. Smith, L. 1957. Peptostreptococcus Kluyver and van Niel (1936), p. 533541.In R. S. Breed, E. G. D. Murray, and N. R. Smith (ed.), Bergeys manual ofdeterminative bacteriology, 7th ed. The Williams & Wilkins Co., Baltimore,Md.

45. Terada, H., K. Miyahara, H. Sohara, M. Sonoda, H. Venomachi, J. Sanada,and T. Arima. 1994. Infective endocarditis caused by an indigenous bacte-rium (Gemella morbillorum). Intern. Med. 33:628631.

46. Thjota, T., and J. Boe. 1938. Neisseria haemolysans. A haemolytic species ofNeisseria Treviscan. Acta Pathol. Microbiol. Scand. Suppl. 37:527531.

47. Tunnicliff, R. 1917. The cultivation of a micrococcus from blood in pre-eruptive and eruptive stage of measles. JAMA 68:10281030.

48. Von Essen, R., M. Ikavalko, and B. Forsblom. 1993. Isolation of Gemellamorbillorum from joint fluid. Lancet 342:177178.

49. Weisburg, W. G., S. M. Barns, D. A. Pelletier, and D. J. Lane. 1991. 16Sribosomal DNA amplification for phylogenic study. J. Bacteriol. 173:697703.

50. Whitney, A. M., and S. P. OConnor. 1993. Phylogenetic relationship ofGemella morbillorum to Gemella haemolysans. Int. J. Syst. Bacteriol. 43:832838.

VOL. 36, 1998 MOLECULAR IDENTIFICATION OF GEMELLA SPECIES 871

on June 9, 2015 by guest

http://jcm.asm.org/D

ownloaded from