Characterization of Nutritionally Variant Streptococci by Biochemical

JOURNAL OF CLINICAL MICROBIOLOGY, Dec. 1985, p. 1030-10340095-1137/85/121030-05$02.00/0Copyright C 1985, American Society for Microbiology

Characterization of Nutritionally Variant Streptococci byBiochemical Tests and Penicillin-Binding Proteins

ANNE BOUVET,* FRANÇOISE VILLEROY, FRANÇOISE CHENG, CATHERINE LAMESCH,RUSSELL WILLIAMSON, AND LAURENT GUTMANN

Laboratoire de Microbiologie Médicale, Université Pierre et Marie Curie, Paris, France

Received 5 June 1985/Accepted 21 August 1985

A comparative study of 60 strains of nutritionally variant streptococci (NVS) with 34 strains of Streptococcusmitis and 37 strains of Streptococcus sanguis Il showed the presence of a red chromophore which was absentin the other streptococcal species. By using the conventional microbiological tests, only small differences were

found between the NVS and the two other related species. In contrast a clear-cut delineation was found by theAPI 20 Strep system of identification. Ali NVS contained pyrrolidonylarylamidase, an enzyme which was

absent in S. mitis and S. sanguis Il strains, and lacked the alkaline phosphatase enzyme which was present in56% of S. mitis strains and 62% of S. sanguis II strains. According to the additional enzymatic and biochemicaltests of the API 20 Strep system, there were three biotypes among NVS. The major biotype included 33 of 60strains which were characterized by the presence of both alpha- and beta-galactosidases and the capacity tohydrolyze trehalose. This biotype also showed a specific pattern of penicillin-binding proteins. These resultsshow that NVS are recognized as a separate variety distinct from S. mitis and S. sanguis II species, despite somecommon biochemical properties. Moreover, the delineation of 33 strains with a specific biotype and a specificpenicillin-binding protein pattern strongly suggests that a large part of NVS strains belong to an individualspecies.

Nutritionally variant streptococci (NVS) are associatedwith various pathological situations including septicemia andendocarditis (2, 4, 5, 11, 12, 14, 19, 20, 25). Since their firstdescription as ungroupable viridans streptococci that growas satellite colonies around colonies of other bacteria andthat require sulfhydryl compounds for normal growth (11),they have been called either pyridoxal-dependent strepto-cocci or NVS (4, 8, 14). Recently, we described a semisyn-thetic medium for the growth of NVS that gave highergrowth yields and normal ultrastructure (3) when comparedwith growth in cysteine- or pyridoxal-supplemented complexmedia (1).The taxonomical position of NVS has not been clearly

established, in spite of a number of studies attempting toclassify these streptococci. By using physiological tests,Cooksey et al. (8) have divided 25 strains of NVS among atleast five different species of viridans streptococci, whereasRoberts et al. (20) identified 12 strains of NVS as vitaminB6-dependent "Streptococcus mitior". Colman and Ballused the API 20 Strep system to study 18 NVS strains (6) andconcluded that their strains formed a reasonably homoge-neous collection. In previous studies we have demonstratedthat all NVS and Streptococcus mitis strains produce a

pH-dependent chromophore that is located in their cell wall,which was the first positive characteristic for identificationof NVS and S. mitis (3, 23). However, results of ourpreliminary studies, using the patterns of penicillin-bindingproteins (PBP) of the bacteria as a taxonomic tool, show thatNVS differs from S. mitis (15).

In this investigation we studied a larger number of NVSstrains for their physiological and biochemical characteris-tics determined by the conventional tests (9) and the API 20Strep system (26) to determine if a relationship exists be-tween NVS and the two closest viridans streptococcalspecies S. mitis and Streptococcus sanguis Il or if they

* Corresponding author.

belong to new viridans species, and we correlated theseresults with the electrophoretic pattern of the bacteria PBPs.The differences between NVS and S. mitis and S. sanguis Ilspecies were sufficient to support their distinctness and todiscriminate a part of NVS as a new streptococcal species.

MATERIALS AND METHODS

Strains. Ofthe 60 strains used in this study, 25 were isolatedfrom patients hospitalized in Broussais and Saint-JosephHospitals (Paris, France). Eight strains isolated in the UnitedStates were kindly provided by R. B. Roberts, New YorkHospital, Cornell Medical Center (New York, N.Y.), and byW. R. Wilson, Mayo Clinic (Rochester, Minn.). Six strainswere sent to us by J. Vandepitte from St.-Raphael UniversityHospital (Leuven, Belgium). The other strains were referredto us for identification from different parts of France. Themajority of strains (36) were isolated from patients withendocarditis, and the others were from patients withbacteremia (16 strains), from patients with sepsis (2 strains),and from other sources (6 strains). Stocked cultures were keptlyophylized or frozen at -80°C. All NVS strains remainedpyridoxal dependent in complex media, and each subculturewas tested for growth as satellite colonies around Staphylo-coccus epidermidis on 5% horse blood Trypcase-soy agar(BioMerieux, Charbonnieres les Bains, France). The refer-ence strains of S. mitis (NCTC 3168, CDC 429, 20279) and S.sanguis II (NCTC 10712, NCTC 7864, SsB, 063-07-78) wereoriginally obtained from R. R. Facklam, Centers for DiseaseControl (Atlanta, Ga.), K. C. Gross, Cornell Medical Center,T. Horaud, Institut Pasteur (Paris, France), and J. P. Gayral,API S.A. (La Balme les Grottes, France). The otherstreptococcal strains used in this study were clinical isolatesfrom our laboratory culture collection. They included 31strains of S. mitis and 33 strains of S. sanguis II that we

compared with the NVS and the 7 reference strains. We alsoconducted an investigation for the presence of thechromophore in 384 strains representative of the streptococ-

1030

Vol. 22, No. 6

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

cm o

n 21

Feb

ruar

y 20

22 b

y 12

5.20

5.14

4.24

2.

CHARACTERIZATION OF NUTRITIONALLY VARIANT STREPTOCOCCI

cal species isolated in humans (Table 1) and routinelyidentified by serological and physiological procedures (17).All the strains were resistant to optochin and susceptible tovancomycin.Chromophore assay. The chromophore test was done on

all the streptococcal strains. Bacteria were grown overnightin 30 ml of a chemically defined medium enriched with 2%Todd-Hewitt dialysate, as described previously (3), andharvested by centrifugation at 3,000 x g for 15 min. Thebacteria were suspended in a volume of 2 N HCI equal to thepellet and heated at 100°C. The appearance of a pink colorwithin 5 min indicated the presence of the chromophore (3,23).

Conventional microbiological tests. A reduced number oftests required for viridans streptococcal species identifica-tion was performed by the method of Facklam (9). Theyincluded tolerance to bile esculin and to 6.5% NaCl; hydro-lysis of arginine, esculin, hippurate, and starch; and fermen-tation of mannitol, sorbitol, inulin, lactose, raffinose,trehalose, salicin, melibiose, and arabinose (Table 2). For allNVS, S. mitis, and S. sanguis II strains tested, the basalmedium was enriched with pyridoxal hydrochloride (SigmaChemical Co., St. Louis, Mo.) prepared in a 67 mM sodiumphosphate buffer (pH 7.8) at a final concentration of 10iLg/ml. Test tubes were inoculated with 0.1 ml of an over-night culture in pyridoxal-supplemented Todd-Hewitt broth(Difco Laboratories, Detroit, Mich.) and incubated in acandle extinction jar. The addition of buffered pyridoxalhydrochloride did not alter the reactions, when examinedwith the reference strains of S. mitis and S. sanguis Il and innoninoculated test tubes incubated for 48 h under the sameconditions. Tests that were negative after 18 h were reeval-uated after another 24 h of incubation. Production of dextranwas assessed on 5% sucrose agar, enriched with 100 ,ug ofpyridoxal hydrochloride per ml for NVS strains.API 20 Strep identification system. The API 20 Strep

system (API System S.A., La Balme les Grottes, France) isidentical to the API Rapid Strep system (Analytab Products,Plainview, N.Y.). The strips are each composed of 20cupules containing dehydrated substrates used for the deter-mination of the following reactions: acetoin production;hydrolysis of hippurate and esculin; production of the en-zymes pyrrolidonylarylamidase, alpha-galactosidase, beta-glucuronidase, beta-galactosidase, alkaline phosphatase,leucine arylamidase, and arginine dihydrolase; and fermen-tation of ribose, L-arabinose, mannitol, sorbitol, lactose,trehalose, inulin, raffinose, starch, and glycogen. The man-ufacturers instructions were followed with modifications.Briefly, growth from two sheep blood Columbia agar (BioMerieux) plates enriched with 100 ,ug of pyridoxal hydro-chloride per ml was suspended in 0.85% saline to yieldsufficient inocula (-4 McFarland turbidity standard). Threedrops of this suspension were placed into cupules number 1to 10 with a Pasteur pipette. For all streptococcal strainstested, the medium provided by the manufacturer for carbo-hydrate fermentation was enriched with buffered pyridoxal(100 ,ug/ml in sodium phosphate buffer with a final concen-tration of 67 mM [pH 7.8]), and the growth of two otherpyridoxal-enriched blood agar plates was suspended in thismedium and inoculated into cupules number 11 to 20. Afterincubation of the strips for 18 h, reagents were added to theacetoin, hippurate, and enzyme cupules, and results for all20 reactions were recorded. Positive and negative reactionswere not altered by the addition of buffered pyridoxalsupplement as tested with reference strains and a controlnoninoculated strip. Since the leucine arylamidase is a

positive marker for streptococci, we considered that a neg-ative test was an indication of a poor inoculum. Conse-quently, four NVS strains and one S. mitis strain with initialnegative reactions were retested with success. A purity platewas prepared from each streptococcal inoculum, and asatellite test was done in addition for NVS.Nomenclature. The name of species that are not in the

Approved List of Bacterial Names (21) are placed in quota-tion marks. We use these species names when we refer totheir use in studies of other authors. When comparing theBritish and the American Streptococcus taxonomy (10) thestrains called "S. mitior" according to the classificationscheme of Colman and Williams (7) will be called S. mitis, ifthey do not ferment raffinose, or S. sanguis II, if theyferment raffinose according to the classification of Facklam(9).

Detection of PBP. Approximately 5 x 107 CFU of bacteriagrown to log phase in the semisynthetic medium (3) werecentrifuged at 4,000 x g for 5 min at 4°C and suspended in 50,tl of 50 mM sodium phosphate buffer (pH 7.0). This suspen-sion of whole bacteria was incubated at 37°C for 10 min with[3H]benzylpenicillin at a concentration of 1 ,ug/ml (15). The[3H]benzylpenicillin, ethylpiperidinium salt (25 Ci/mmol)was a generous gift of Rhone Poulenc (Vitry sur Seine,France) and of the Service des Molecules Marquées, Com-missariat à l'Energie Atomique (Gif sur Yvette, France). Thereaction was halted by the addition of excess unlabeledpenicillin and the addition of 2 ml of phosphate buffer. Thesamples were centrifuged at 4,000 x g for 10 min at 4°C andsuspended in 50 ,ul of 50 mM phosphate buffer (pH 7.0)containing 0.6% Sarkosyl (wt/vol; CIBA-GEIGY Corp.,Summit, N.J.) and 30 ,ug of mutanolysin (28). Incubation at37°C for 15 min resulted in complete lysis of the bacteria.The lysates were prepared for sodium dodecyl sulfate-polyacrylamide slab gel electrophoresis as described previ-ously (16) and run on a 6.5% acrylamide gel prepared by themethod of Laemmli (18). Electrophoresis was carried out at60 V for 90 min and then at 120 V for 3 h. Fluorography wasperformed as described previously (16). Fluorograms wereexposed to the dried gel for 8 days at -80°C. Molecularweights of the PBPs were estimated by comparison withthose of a Streptococcus faecium reference strain (27).

TABLE 1. Analysis of streptococci for the presence of apH-dependent chromophore

Streptococcal species No. of strains No. of strains withexamined chromophore (%)

NVS 61 61 (100)S. mitis 34 29 (4.5)S. sanguis Il 37 31(84)S. mutans il 0 (0)S. sanguis I 74 0 (0)S. salivarius 24 0 (0)S. intermedius 54 0 (0)S. constellatus il 0 (0)S. morbillorum 7 0 (0)S. acidominus 5 O (0)S. uberis 1 0 (0)Lancefield-groupable 197 0 (0)

streptococciaa The Lancefield-groupable streptococci included 4 group D streptococcal

species, S. faecalis (50 strains), S. faecium (12 strains), S. bovis (20 strains),and S. durans (8 strains), and 10 other groupable streptococcal species, groupA (20 strains), group B (20 strains), group C (6 strains), group F (32 strains),group G (5 strains), group H (3 strains), group L (10 strains), group M (3strains), group N (7 strains), and group P (1 strain).

VOL. 22, 1985 1031

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

cm o

n 21

Feb

ruar

y 20

22 b

y 12

5.20

5.14

4.24

2.

1032 BOUVET ET AL.

TABLE 2. Conventional microbiological tests

No. (%) of the following strains givingTest positive reactions:

NVS S. mitis S. sanguis Il

Tolerance toBile esculin O O O6.5% NaCI O O O

Hydrolysis ofArginine 0 2 5Esculin 1 0 0Hippurate 2 0 0Starch 3 4 3

Fermentation ofMannitol O O OSorbitol O O OInulin 17 (28) 0 0Lactose 27 (44) 34 (100) 37 (100)Raffinose 3 0 37 (100)Trehalose 35 (58) 5 (15) 10 (27)Salicin 3 7 5Melibiose 4 0 30 (81)Arabinose 7 0 0

Dextran production on 0 6 (18) 15 (41)5% sucrose

RESULTS

Chromophore assay. The chromophore test was done on alarge panel of groupable and ungroupable streptococcalspecies (Table 1). Among the ungroupable streptococci, thechromophore was found in 100% of NVS, in 85% of S. mitis,in 84% of S. sanguis Il, and in none of the 187 other strains.The 197 Lancefield-groupable streptococcal strains testeddid not contain the pH-dependent chromophore.

Conventional microbiological réactions. For the majority ofthese reactions, the results obtained with the NVS strainswere negative or variable, and therefore they most closely

TABLE 3. API 20 Strep system of identification tests

No. (%) of the following strains givingpositive reactions:

Test

Acetoin 44 (72) 0 0Hippurate 2 0 0Esculin hydrolysis 12 (20) 5 4Pyrrolidonylarylamidase 60 (100) 0 0alpha-Galactosidase 34 (56) 0 37 (100)beta-Glucuronidase 12 (20) 0 0beta-Galactosidase 33 (54) 11 (32) 15 (41)Alkaline phosphatase 0 19 (56) 23 (62)Leucine aminopeptidase 60 (100) 34 (100) 37 (100)Arginine hydrolysis 0 2 4Ribose 1 0 2Arabinose 0 0 0Mannitol 0 0 0Sorbitol 0 0 0Lactose 25 (41) 34 (100) 37 (100)Trehalose 35 (57) 5 (15) 10 (27)Inulin 17 (28) ORaffinose 12 (20) 0 37 (100)Starch 33 28 32Glycogen 1 0 3

TABLE 4. Comparison of the three NVS biotypes to S. mitis andS. sanguis Il as determined by the API 20 Strep system

Test Biotypes of NVS S. sanguis1 2 3

Pyrrolidonylarylamidase + + + - -Alkaline phosphatase - - - v+ v+alpha-Galactosidase + - - - +beta-Glucuronidase - + -

beta-Galactosidase + - - v vTrehalose + - - v vInulin - v v -

a Abbreviations: +, 100% positive results; v+, 50 to 75% positive results;v-, 10 to 45% positive results; -, 0% positive results.

resemble those of S. mitis or S. sanguis Il (Table 2).However, none ofNVS strains produced dextran on sucroseagar, and a noticeable percentage of NVS strains were foundto be positive for inulin fermentation (28%) or negative forlactose fermentation (56%), in contrast with S. mitis and S.sanguis Il. In addition a high percentage (58%) of NVSstrains fermented trehalose as compared with S. mitis (15%)and S. sanguis Il (27%).API 20 Strep identification system. By using the API 20

Strep system, important differences were found betweenNVS and S. mitis or S. sanguis II (Table 3). The bestdelineation was found for enzymatic activities: pyr-rolidonylarylamidase was produced by all the NVS strainsand was absent in all S. mitis and S. sanguis II strains;alkaline phosphatase was absent in all NVS strains and wasproduced by 56% of S. mitis and 62% of S. sanguis II strains.Acetoin was produced from glucose by 72% of NVS strainsand by none of the S. mitis or the S. sanguis Il strains.

In addition, among the NVS, 57 of 60 strains could beseparated into three biotypes (Table 4), and three strainswith disparate biotypes remained unclassified. The mostcommon pattern shown by 33 strains of biotype 1 was theproduction of alpha- and beta-galactosidases, the fermenta-tion of trehalose, the absence of beta-glucuronidase, and thelack of fermentation of inulin. Biotype 2 included 12 strainsthat produced beta-glucuronidase, did not produce alpha-and beta-galactosidases, did not ferment trehalose, andsometimes fermented inulin. Biotype 3 included 12 strainsthat did not produce either alpha- and beta-galactosidases orbeta-glucuronidase, did not ferment trehalose, and some-times fermented inulin. None of these NVS corresponded toa previously named viridans streptococcal species.PBPs. Different electrophoretic patterns of PBPs were

found in NVS, S. mitis, and S. sanguis Il strains. Among theNVS, 55 of 60 strains were separated into group I and groupII (Fig. 1). Group I included 33 strains with identical PBPs:each strain had five major bands with molecular sizes of 188,96, 88, 77, and 62 kilodaltons. This group was very homo-geneous (Fig. 1). Group II included 22 strains with fourmajor bands with molecular sizes of 85, 83, 72 to 75, and 68to 70 kilodaltons (Fig. 1) and a minor band with a lowmolecular mass. Noticeable variations were observed in theposition of the third and fourth bands. The five remainingNVS strains had PBP electrophoretic patterns variable fromone strain to another and different from those of group I andgroup II (Fig. 2). PBPs of the NVS strains differed fromthose of S. mitis and S. sanguis Il. Although the PBPpatterns of these species were not homogeneous (data notshown), they had three to five major bands with molecularsizes between-75 and 96 kilodaltons that differed from the

J. CLIN. MICROBIOL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

cm o

n 21

Feb

ruar

y 20

22 b

y 12

5.20

5.14

4.24

2.

CHARACTERIZATION OF NUTRITIONALLY VARIANT STREPTOCOCCI

1 2 3 4 5 6 7 8 TABLE 5. Correlation between biotypes and PBP groups in NVS

Biotype No. of strains in the following PBP groups:(no. of strains) I II Ungroupable

1 (33) 332 (12) 123 (12) 9 3Unclassified (3) 1 2

92--_~~FIG. 1. Fluorogram showing the PBPs of NVS. Lanes 1 to 4,

PBP group I; lanes 5 to 8, PBP group Il. Numbers to the left aremolecular sizes in kilodaltons.

five major bands ofNVS group I and the four major bands ofNVS group II (Fig. 2).A good correlation was found between NVS biotypes and

PBP analyses (Table 5). The 33 biotype 1 strains were all ofPBP group I. The 12 biotype 2 strains were all of PBP groupII, and biotype 3 included 9 strains of PBP group Il and 3ungroupable strains. The three strains of unclassifiedbiotypes belonged to the PBP group Il (one strain) or were

ungroupable (two strains).

DISCUSSIONTo delineate the NVS from other related viridans strepto-

coccal species, we investigated 60 strains of these patho-genic, fastidiously growing microorganisms and comparedthem with 34 strains of S. mitis and 37 strains of S. sanguisIl. The presence of a red chromophore was shown in all 60strains of NVS and in 85% of the S. mitis and 84% of the S.sanguis II strains. It was not found in 187 strains of otherviridans streptococci. Although the chemical nature of thispigment has not been yet characterized (23), it appears as abiochemical marker of particular interest for the character-ization of the whole set of NVS, S. mitis, and S. sanguis Il

1 2 3 4 5 6 7 8 9 10

126 -*

92-_86-'

- _

dé amO_

FIG. 2. Fluorogram comparing the PBPs of groupable andnongroupable NVS with those of S. mitis and S. sanguis II. Lane 1,NVS PBP group 1; lane 2, NVS PBP group Il; lanes 3 to 7,ungroupable NVS; lanes 8 and 10, S. mitis; lane 9, S. sanguis Il.Numbers to the left are molecular sizes in kilodaltons.

species. It is a good positive criterion for distinguishing theseorganisms from the other streptococci, which is in contrastwith the conventional microbiological tests that mainly pro-vide negative reactions (9).Among these chromophore-positive species, the conven-

tional tests did not allow characterization of NVS from theother viridans streptococcal species. Most of NVS strainswould have been identified as S. mitis or S. sanguis Il (Table2). The same conclusion arose from the results of the studyof Roberts et al. (20), who identified 12 strains of NVS asvitamin B6-dependent "S. mitior," according to the classi-fication of Colman and Williams (7).

In contrast to this, the API 20 Strep system provided aclear-cut delineation among NVS and the other possiblyrelated streptococcal species (Tables 3 and 4). All NVSstrains produced pyrrolidonylarylamidase, an enzyme that isnot produced by any of the S. mitis and S. sanguis II strains.None of the NVS strains produced alkaline phosphatase,which was produced by 56% of the S. mitis and 62% of theS. sanguis Il strains. With this identification system we wereable to divide the 60 NVS strains into three biotypes, ofwhich biotype 1 included the majority of the strains and wasthe best characterized. We defined biotype 1 by the produc-tion of both alpha- and beta-galactosidases, together with thecapacity for trehalose fermentation and the absence ofbeta-glucuronidase production and inulin fermentation. Ofthe 60 NVS strains studied, 33 belonged to biotype 1, whichseems to correspond to the API profile of 6 of 18 NVS strainsstudied by Colman and Ball (6). We defined biotype 2, shownby 12 of our NVS strains, by the production of beta-glucuronidase, the absence of production of alpha- andbeta-galactosidases, the absence of trehalose fermentation,and the inconstant ability to ferment inulin. The five strainsstudied by Colman and Ball which produced beta-glucuronidase could belong to this biotype. We definedbiotype 3 by the absence of production of alpha- andbeta-galactosidases and beta-glucuronidase, the inability toferment trehalose, and an inconstant ability to fermentinulin. Twelve of our strains exhibited these characteristics.The analysis of the PBP pattern of NVS and the two other

related viridans streptococci confirmed our previous resultsobtained with 10 NVS strains (15). These proteins whichspecifically bind beta-lactam antibiotics have been shown tohave an electrophoretic pattern that varies from one bacte-rial species to another (13) and have been proposed as areliable genetic marker in bacterial taxonomy (13, 22). In thepresent study a good correlation was found between PBPpatterns and our biotype classification (Table 5), since allstrains of biotype 1 belonged to PBP group I and all strainsof biotype 2 belonged to the PBP group II. Strains of biotype3 and unclassified biotypes belonged either to PBP group Ilor to the unclassified group of PBP. All these strains hadPBP patterns different from those of the S. mitis and S.sanguis II strains we examined.These data strongly suggest that NVS could be considered

126 -a

1033VOL. 22, 1985

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

cm o

n 21

Feb

ruar

y 20

22 b

y 12

5.20

5.14

4.24

2.

1034 BOUVET ET AL.

as a separate variety of streptococci, with at least one novelspecies. This finding is in agreement with the results of vande Rijn and George (24), who showed that NVS can besubdivided into serotypes by rocket line immunoelectropho-resis based on antigenic components not cross-reactive withother viridans streptococci, although there was no evidencethat all NVS have a common serogroup antigen. From theresults of our study we postulate that biotype 1-PBP group I,which includes the majority of the strains we studied, likelybelongs to a new individual streptococcal species. DNAhybridization studies are in progress to confirm this hy-pothesis.

ACKNOWLEDGMENTSWe thank M. C. Chantoiseau for secretarial assistance.This work was supported by grant CRE 1985 from the Institut

National de la Santé et de la Recherche Médicale and by funds fromthe Caisse Nationale d'Assurance Maladie des Travailleurs Salariés,France.

LITERATURE CITED1. Bouvet, A., A. Ryter, C. Frehel, and J. F. Acar. 1980. Nutrition-

ally deficient streptococci: electron microscopic study of 14strains isolated in bacterial endocarditis. Ann. Microbiol. (Paris)131B:101-120.

2. Bouvet, A., I. van de Rijn, and J. F. Acar. 1982. Nutritionallyvariant streptococcal endocarditis, p. 66-67. In S. E. Holm andP. Christensen (ed.), Basic concepts of streptococci and strep-tococcal diseases. Reedbooks Ltd., Chertsey, England.

3. Bouvet, A., I. van de Rijn, and M. McCarty. 1981. Nutritionallyvariant streptococci from patients with endocarditis: growthparameters in a semisynthetic medium and demonstration of achromophore. J. Bacteriol. 146:1075-1082.

4. Carey, R. B., K. C. Gross, and R. B. Roberts. 1975. VitaminB6-dependent Streptococcus mitior (mitis) isolated from pa-tients with systemic infections. J. Infect. Dis. 131:722-726.

5. Cayeux, P., J. F. Acar, and Y. A. Chabbert. 1971. Bacterialpersistence in streptococcal endocarditis due to thiol requiringmutants. J. Infect. Dis. 124:247-254.

6. Colman, G., and L. C. Ball. 1984. Identification of streptococciin a medical laboratory. J. Appl. Bacteriol. 57:1-14.

7. Colman, G., and R. E. O. Williams. 1972. Taxonomy of somehuman viridans streptococci, p. 281-299. In L. W. Wannamakerand J. M. Matsen (ed.), Streptoçocci and steptococcal disease.Academic Press, Inc., New York.

8. Cooksey, R. C., F. S. Thompson, and R. R. Facklam. 1979.Physiological characterization of nutritionally variant strepto-cocci. J. Clin. Microbiol. 10:326-330.

9. Facklam, R. R. 1977. Physiological differentiation of viridansstreptococci. J. Clin. Microbiol. 5:184-201.

10. Facklam, R. R. 1984. The major differences in the American andBritish Streptococcus taxonomy schemes with special referenceto Streptococcus milleri. Eur. J. Clin. Microbiol. 3:91-93.

11. Frenkel, A., and W. Hirsch. 1961. Spontaneous development ofL forms of streptococci requiring secretions of other bacteria orsulphydryl compounds for normal growth. Nature (London)191:728-730.

12. George, R. H. 1974. The isolation of symbiotic streptococci. J.Med. Microbiol. 7:77-83.

13. Georgopapadakou, N. H., and F. Y. Liu. 1980. Penicillin bindingproteins in bacteria. Antimicrob. Agents Chemother.18:148-157.

14. Gephart, J. F., and J. A. Washington Il. 1982. Antimicrobialsusceptibility of nutritionally variant streptococci. J. Infect.Dis. 146:536-539.

15. Gutmann, L., A. Bouvet, and J. F. Acar. 1982. Penicillin bindingprotein of nutritionally variant streptococci. FEMS Lett.14:11-14.

16. Gutmann, L., R. Williamson, and A. Tomasz. 1981. Physiolog-ical properties of penicillin binding proteins in group A strepto-cocci. Antimicrob. Agents Chemother. 19:872-880.

17. Horodniceanu, T., and F. Delbos. 1982. Les streptocoques nongroupables dans les infections humaines: identification etsensibilité aux antibiotiques. Ann. Microbiol. (Inst. Pasteur)133B:255-269.

18. Laemmli, U. K. 1970. Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature (London)227:680-685.

19. McCarthy, L. R., and E. J. Bottone. 1974. Bacteremia andendocarditis caused by satelliting streptococci. Am. J. Clin.Pathol. 61:585-591.

20. Roberts, R. B., A. G. Krieger, N. L. Schiller, and K. C. Gross.1979. Viridans streptococcal endocarditis: the role of variousspecies, including pyridoxal-dependent streptococci. Rev. In-fect. Dis. 1:955-965.

21. Skerman, V. B. D., V. McGowan, and P. H. A. Sneath. 1980.Approved list of bacterial names. Int. J. Syst. flacteriol.30:225-420.

22. Tomasz, A. 1982. Penicillin binding proteins in bacteria. Ann.Intern. Med. 96:502-504.

23. van de Rijn, I., and A. Bouvet. 1984. Characterization of apH-dependent chromophore for nutritionally variant strepto-cocci. Infect. Immun. 43:28-31.

24. van de Rijn, I., and M. George. 1984. Immunochemical study ofnutritionally variant streptococci. J. Immunol. 133:2220-2225.

25. Verhaegen, J., and J. Vandepitte. 1983. Endocarditis caused bynutritionally deficient streptococci. Acta Clin. Belgica 38:12-18.

26. Waitkins, S. A., D. R. Anderson, and F. K. Todd. 1981. Anevaluation of the API STREP identification system. Med. Lab.Sci. 38:35-39.

27. Williamson, R., C. Le Bouguenec, L. Gutmann, and T. Horaud.1985. One or two low affinity penicillin-binding proteins may beresponsible for the range of susceptibility of Enterococcusfaecium to benzylpenicillin. J. Gen. Microbiol. 131:1933-1940.

28. Yokogawa, K., S. Kawata, S. Nishimura, Y. Ikeda, and Y.Yoshimura. 1974. Mutanolysin bacteriolytic agent for cariogenicstreptococci. Partial purification and properties. Antimicrob.Agents Chemother. 6:156-165.

J. CLIN. MICROBIOL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

cm o

n 21

Feb

ruar

y 20

22 b

y 12

5.20

5.14

4.24

2.