Antonie van Leeuwenhoek 44 (1978) 1-14

Quantitative gas chromatography of Bacteroides species under different growth conditions J. G . E. M . LINDNER AND J. H . MARCELIS

Anaerobe section of the Laboratory of Microbiology, State University Utrecht, Catharijnesinge159, Utrecht, The Netherlands

LINDNER,J. G. E. M. and MARCELIS,J. H. 1978. Quantitative gas chromatography of Bacteroides species under different growth conditions. Antonie van Leeuwenhoek 44: 1-14. From 56 strains of strictly anaerobic gram-negative rods isolated from stool and purulent lesions the fermentation products in the presence and absence of hemin were determined by quantitative gas-solid chromatography, using a simple and more rapid chromatographic procedure. With hemin the fermentation products were propionic, acetic, lactic and succinic acid. Without hemin no or little succinic acid was formed and the main products were lactic and acetic acid. In both groups the distribution of subspecies was determined and the production of fatty acids measured quantitatively. Fourteen strains of the lesion group showed a higher metabolic activity, resulting in an increased total acid production caused by an excessive production of acetic and lactic acid. This characteristic is probably a virulence factor in these strains. All strains were protoporphyrin- and oxgall-dependent. It is postulated that these substances are used for the production of cytochromes which permits the formation of succinic acid by a fumarate reductase resulting in an increased growth rate and growth yield.

INTRODUCTION In recent years Bacteroides species have received more attention and have been related to various disorders. By improved methods for isolation and identification many species and subspecies can be distinguished. The most widely used procedures for identification by biochemical tests were described by Holdeman and Moore (1972). Also serological methods were used to identify Bacteroides species (Werner, 1969; de la Cruz and Cuadra, 1969; Reinhold, 1971 ; Beerens et al., 1971 ; Lambe and Moroz, 1976).

J. G. E. M. LINDNER AND J. H. MARCELIS Werner et al. (1975) showed that isolates from clinical material differ from strains usually found in stools. B. fragilis ssp. fragilis seems to be the most virulent, whereas B. fragilis ssp. vulgatus and ssp. distasonis are commonly found in stools. It is not clear why one subspecies of B.fragilis is more virulent, as the differences in biochemical characteristics of the subspecies are very slight. At present, identification is based additionally on the detection of fermentation products by paper chromatography (Guillaume, Beerens and Osteux, 1956; Charles and Barrett, 1963; Slifkin and Hercher, 1974) or by gas chromatography (Moore, 1970; Holdeman and Moore, 1972; Cartsson, 1973; Hauser and Zabransky, 1975). The latter method requires separate runs for volatile and nonvolatile fatty acids. Methods for detection by one run take a lot of time for sample preparation (Salanitro and Muirhead, 1975). Therefore, these methods are often considered too complex for routine diagnostic use. For the study of large numbers of strains a simplified, short time method would be advantageous. We modified existing methods for quantitative gas chromatography and developed a procedure with short periods of sample preparation and analysis (Materials and Methods). With this method the fermentation products of 56 isolates of Bacteroides (32 from pus, wounds and abdominal abcesses, 22 from stools and 2 reference strains) were studied. As certain Bacteroides species require hemiia for growth (as indicated for B. ruminicola (Caldwell et al., 1965), for B. melaninogenicus (Rizza et al., 1968) and for 4 strains of B. fragilis (Varel and Bryant, 1974; Macy, Probst and Gottschalk, 1975)), we also studied the influence of hemin on growth and fatty acid production. More detailed information about the characteristics of "lesion" strains and fecal strains in various conditions might contribute to the understanding of the differences in pathogenicity.

MATERIALS AND METHODS

Strains. Clinical isolates mainly from purulent lesions were obtained at the University Hospital of Utrecht. Fecal strains were isolated from normal adults. Reference strains were received from the National Collection of Type Cultures, London. Cultural methods. About 1 g of stool was put in 19 ml of a prereduced transport medium (van der Wiel-Korstanje and Winkler, 1970). After weighing and homogenization tenfold dilutions from 103 up to 108 were made in the same medium. From appropriate dilutions one drop was used to inoculate (1) a Reinforced Clostridial Agar plate (RCA Oxoid cm 151) with 7 . 5 ~ sheep blood and (2) a Viande Levure plate (VL) (Beerens and Fievez, 1971), containing

QUANTITATIVE GAS CHROMATOGRAPHY OF BACTEROIDES

neomycin 100 mg/litre and vancomycin 7.5 mg/litre, supplemented with agar 1.5 ~o according to Finegold et al. (1971). Both media had been kept prereduced in a nitrogen atmosphere. Pus was inoculated on these media without dilution. Both media were incubated anaerobically at 37 C for 72 h in an atmosphere of 90 ~ H 2 and 10 ~ CO 2. The jars used for this purpose were checked weekly for leaks and catalyst performance according to Watt, Collee and Brown (1976). Total counts of stool samples on RCA sheep blood agar were 2-5 • 101~bacteria per gram wet weight. On VL agar with neomycin and vancomycin the total count was slightly lower. Small translucent colonies of gram-negative rods with blunt ends were tested for growth under aerobic and anaerobic conditions on prereduced RCA bloodagar. Subcultures were inoculated with platinum loops in an anaerobic cabinet. Only obligate anaerobic gram-negative rods were used for further investigation. Identification procedures. VL broth with glucose 0.25 ~ and supplemented with a filter-sterilized solution of hemin 2 mg/litre and menadione 0.5/~g/litre was used as a basal medium (HVL) final pH 6.9. Basal medium with 1 filter-sterilized carbohydrate instead of glucose was used for fermentation reactions. After 6 days of growth at 37 C under anaerobic conditions the pH was measured with a glass electrode (Philips PW 9418); a pH below 5.6 was considered to be positive for acid production. Production of H2S and gas was tested in HVL with ferrosulphate 2 g/litre, sodium thiosulphate 3 g/litre and Bacto-agar 5 g/litre (Difco, Detroit, MI). Glucuronidase production was determined in trypton broth (Difco, Detroit, MI) with 4 Methyl-umbeUiferyl-glucuronide (Koch-Lights Laboratories, Colnbrooks, England) 0.15 g/litre (Dahlen and Linde, 1973). Fluorescence was tested at pH 10.6. Colistin sensitivity was tested in HVL with colimycin sulphate (Pfizer) 10 mg/litre. The schemes for identification used refer to the Anaerobic Laboratory Manual (Holdeman and Moore, 1972) and to Bergey's Manual (8th ed., 1974). Growth yield and growth stimulation were studied in VL broth, in HVL and VL with protoporphyrin or oxgall at various concentrations. The optical density was measured with a Klett Summerson Extinctiometer. Gas chromatography. Seven-days cultures unless otherwise mentioned were centrifugated at 3000 g for 15 rain. The supernates were used for quantitative determination of short-chain fatty acids and stored at -20 C if necessary. The volatile fatty acids (VFA): acetic acid (A), propionic acid (P), iso- and butyric acid (IB and B), iso- and valeric acid (IV and V) and also the nonvolatile lactic acid (L) (see below) were determined after deionization, by gas chromatography over a column of chromosorb 101 (80/100 mesh, Johns Manville, Denver Co U.S.A.) by a modification of the method of Carlsson (1973). Since VFA are strongly adsorbed to glass, polypropylene columns

4

J. G. E. M. LINDNER AND J. H. MARCELIS

(0.7 • 4 cm, no. 371-110, Biorad, Richmond, Ca, U.S.A.) has to be used for deionization. One ml of supernate was passed through 1-cm columns of Dowex 50 W x 4 (200-400 mesh, Fluka, A.G., Buchs, Switzerland). The retention of VFA in the column required fourfold elution with 0.5 ml of twice distilled water. Supernate and elution fluids were collected into 10 ml polyethylene bottles. The non-volatile fatty acids (non-VFA) lactic acid (L) and succinic acid (S) were determined after esterification according to Holdeman and Moore (1972). However, since the lactic-ethylester peak is better separated from chloroform than the methylester, ethanol was used for esterification. The separation of fumaric and succinic acid is poor. However it was shown by IB B P

t.~ t/l z cD

V

S

III

t/%

rv c~ CD U t,l

ill o

L

_A

!

I

2 Cl

357 b

TIME (M I N.) Fig. 1. Gas-chromatographic separation of a. fatty acids from a standard mixture of non esterified acids in aqueous solution at 10 mM

of each component. A = acetic, P = p r o p i o n i c , IB = i s o b u t y r i c , B = butyric, V ~ valeric and L = lactic

acid. b. the ethylesters of lactic acid (L) and succinic acid (S) after esterification of a standard

mixture containing 10 mmoles of each component.

Q U A N T I T A T I V E GAS C H R O M A T O G R A P H Y

OF BACTEROIDES

methylation that not one out of nine strains produced fumaric acid. With a Hamilton syringe (10 pl, type 701 N) 1/A eluent (VFA) or 5/A of the chloroform layer (non-VFA) was injected straight into the column above the packing. A dual glass-column Packard Becker gas chromatograph type 417 with a flame ionization detector (45 ml Hz/min, 450 ml air/min, 240 C, sensitivity 2.5 • 10 -12 Amp. for V F A and 2.5 • 10 -11 Amp. for non-VFA) was used, linked to a 0.5-mV span recorder (TE 200, T e k m a n Ltd, U.K.). The second column was used as a blank for n o n - V F A runs and for V F A analysis. Nitrogen (40 ml/min) was used as carrier gas. F o r V F A runs the temperature of inlet and oven was 210 C. Attenuation was 2 • and analysis-time 4 min. V F A ' s were detected at concentrations > 1 m ~ and lactic acid > 2.5 mM. A c h r o m a t o g r a m of test solutions of V F A and nonV F A (10 raM) is shown in Fig. 1. Peak heights of standard solutions were used for calibration. A calibration curve was included in each run, accepting

500 k

kW

500

/

~00

SE ~00

h

//

3OO

300

21= 200

~20o

E

LE

..1.-

100 k

0u 0 a

//~

110

/

~ L

2'0

3'0

t-'0

5'0 CONC.mMot/t

~

0

0 b

1'0

2'0

3'0

40

510

EONC. mMot/t

Fig. 2. Linear calibration curves (peak heights versus concentrations) between 5-50 mM for: a. standards of fatty acids and b. standards of lactic and succinic acid after esterification.

J. G. E. M. LINDNER AND J. H. MARCELIS only curves with a correlation coefficient above 0.990 (Fig. 2). Relative retention times expressed in relation to acetic acid (A, 52 sec. = 1) were 1.5 for P; 1.9 for IB; 2.2 for B; 3.5 for V and 4.8 for L. For non-VFA determinations the inlet and oven temperature was 210 C. After 1 min the oven temperature was raised with 15 ~ to 250 C, kept at this level for 3 min and cooled for 2 min to obtain the initial temperature. Attenuation was 4 • and analysis time 8 min. Calibration was as for VFA. The relative retention of S with regard to L (168 sec. = 1) was 2.6. The limit of sensitivity for S and L was 1 mM. Biochemical determinations of the same lactic acid standard solutions using L-lactic dehydrogenase (Hohorst, 1970) with a Boehringer test kit (15972) corresponded well with both gas-chromatographic results. All results are expressed in mmoles per litre of culture supernate.

RESULTS

Fermentation products. Of all obligate anaerobic gram-negative rods isolated from stool or pus, presumptively identified as Bacteroides species (Holdeman and Moore, 1972) only the saccharolytic strainswere kept for further investigations. Thirty-two lesion strains (LS) and 22 fecal strains (FS) were biochemically identified using the procedures of the Virginia.Polytechnic Institute. The results of the tests together with those of two reference strains are shown in Table 1. All strains except one were identified as B. fragilis. The distribution of the subspecies differed in the two groups (FS and LS). In the isolates from feces only B.fragilis ssp. vulgatus and ssp. distasonis could be found. Twelve strains (37.5 ~), isolated from lesion, were B.fragilis ssp.fragilis. Characterization of subspecies by fermentation reactions often depends on many weak and variable reactions to various substrates (see Table 1). Quantitative GSC adds a new dimension to this characterization. In our case GSC enables a division within the groups of FS and LS based on the formation of propionic acid in this medium. The results are shown in Table 2. Seven strains of group FS produced propionic acid (FS 1), 15 strains did not (FS2). Production of total acid was the same in both groups (mean 17.7 mmole). The same biochemical subspecies were found in both groups. From the 32 strains isolated from lesions, 23 strains produced propionic acid (LS 1), 9 did not (LS2). These 9 strains belong to all five different subspecies. The total acid production was comparable with that of group FS1 and FS2. The group that formed propionic acid consisted of two subgroups LSla and LSlb. Group LSla (9 strains belonging to the subspeciesfragilis, vulgatus and distasonis) is comparable with group FS 1 but group LS 1b (14 strains, belonging to the subspeciesfragilis, vulgatus, distasonis and thetaiotaomicron) is different.

_ _

.

.

_ _

-

_

.

+ •

+~

+w

+w

+ w

~

_ +

v

v

+w

__

i

~

+ ~ +w

v

+w

~w

+w

t w

o

~

+ w +

-

+

+

~

_

~

+ q-

-

+

+

+

+

o

+ +

4-

+

+

+

+

o

_ -

-

+

+*

_

_

~

_ -

-

-

v

v

_

+ +

4-

+

+

+

+

~

_ -

-

-

-

_

~

ssp. v u l g a t u s B . f r a g i l i s ssp. d i s t a s o n i s B . f r a g i l i s spp. t h e t a i o t a o m i c r o n

B . f r a g i l i s ssp. f r a g i l i s

18.7 • 4.2

1.7 1.3 1.7 2.7

Total acid

• A • •

7 2 8.4 1.4 2.4 6.4 16.8 s 3.1

15 2 7.5 • 2.9 3.5 i 2.4 5.7 s 2.2 20.3

8.8 2.4 9.1

NCTC 10583

3.1

• 2.0 i 1.0 i 1.6 • 2.7 17.0 i

9 3 6.8 1.1 2.4 6.7

LS 1a

FS2

FS 1

N u m b e r of strains N u m b e r of subspecies Acetic acid Propionic acid Lactic acid Succinic acid

Group/strain

Lesion strains

Fecal strains

• 7.1 i 2.5 • 3.7 • 1.8 42.7 ~z 12

14 4 25.4 3.2 9.7 4.4

LS 1b

15.6 •

5.1

9 5 3.9 • 1.2 2.7 • 3.4 8.7 i 1.8

LS2

N C T C 9344 B . f r a g i l i s - o t h e r N C T C 10583 B . f r a g i l i s ssp. v u l g a t u s

B. capillosus

B.fragilis

Table 2. Fatty acids end p r o d u c t s (in mM) o f 56 strains o f B a c t e r o i d e s , g r o w n in the presence o f hemin

+ w: weak reaction; v: variable reaction.

+ •

.

+

+

.

_

+

_

_

_

+_

+

~

~ :

:

15

7

10.4

9.4 2.5 2.1 8.4

NCTC 9344

1

3

8

8

12

,.d

Table 1. Biochemical reactions o f 56 strains o f B a c t e r o i d e s , (32 f r o m lesions (LS), 22 f r o m stools (FS) and 2 reference strains), according to the scheme for anaerobic identification of the Virginia Polytechnic Institute

t7

,-]

O

,-1 0

O

o0

Z

M

J. G. E. M. LINDNER AND J. H. MARCELIS Whereas the growth yield was comparable with the other groups the total amount of fatty acid of these strains was 2.5 times higher (mean 42.7 mmole). The pattern differed as well. The production of lactic and acetic acid was 3 times higher than in the other groups. The patterns of the two reference strains are also given in Table 2. N C T C 10583 isolated from feces was completely comparable with the strains of group FS1. Total acid was 20.3 mu. Strain NCTC 9344 isolated from a septic operation wound produced propionic acid and must be compared with group LS1. The relative proportion KLETT VALUES

KLETT

VALUES

,sc

150

100

100f soF 20 10

20

o

30 TIHE

150t

90

10

20

30

90

b

(HOURS)

9 /

//

100

t li

// / c Fig. 3. Growth of a representative strain in the absence and presence of various concentrations of heroin, protoporphyrin and oxgall. a. Hemin: 9 0.5 #g/ml;/~ 1 #g/ml;& 2 #g/ml; [] 4 #g/rnl. b. Protoporphyrin: 9 0.6 #g/ml;~ 1.25 #g/ml;& 2.5/~g/ml; E35 #g/ml. c. Oxgall: 9 0.125~o;/k 0.25~;& 0.5~o; [] 1~; 110.3~. (3 control

QUANTITATIVE GAS CHROMATOGRAPHYOF BACTEROIDES as well as the amount of total acid were similar to group LS 1a. Influence of hemin on (a) the growth of bacteria. Since certain Bacteroides species require hemin for growth (de Vries, van Wijck-Kapteyn and Stouthamer, 1973; Macy et al., 1975) the growth of nine given strains from lesions and feces was studied with various concentrations of heroin in VL medium (Fig. 3). Growth without hemin was very slow and poor. With hemin the rate of growth as well as the yield increased 2-3 times. Maximal stimulation was obtained at a concentration of at least 2 ktg/ml heroin. Stimulation in this concentration was seen with all the strains used in this study including the reference strains. Protoporphyrin can replace hemin as a growth factor (Fig. 3). The optimal concentration is also 2-3 #g/ml medium. The stimulatory effect on growth of these strains by bile was the same (Fig. 3). Optimal concentration was 0.3 Voo oxgall, higher concentrations were often inhibitory. Previous adaptation in a medium containing oxgall (0.3 ~o) promoted growth to a level comparable to that of growth in the presence of hemin. Influence of hemin on ( b) the pattern of fermentation products. Fermentation products were measured by GSC for 10 hemin-dependent strains and the two reference strains grown in the presence and absence of hemin after 7 days. A representative graph is given in Fig. 4. In the absence of hemin, lactic acid was the predominant product (mean 67 ~ of total acid) with some acetic acid

Z i.w oc

u

\ Io

Iio

llb

TIME

Fig. 4. Example of the gas-chromatographic profile of the fermentation products in a culture grown without hemin (I) or with hemin (II), analysis after ion exchange (a) and after esterification (b).

32 2 2 2 0 2 3 3 2 16 6.4

B18 BB2o BB27 BB28 Az13 D27 Dzt BB3o 153 137 Mean NCTC 9344 NCTC 10583

6

12

8

55 14 7 10 5 2 7 1 2 39 14

(63) (85) (74) (77) (83) (44) (67) (25) (44) (71) 5_ 6 (67)

Lactic

3

(37) (12) (21) (15) (0) (44) (29) (75) (44) (29) • 3 (31)

Acetic

Strain number

Without hemin

0.5

0

0 0.5 0.5 1 1 0.5 0.5 0 0.5 0 0.45

(0) (3) (5) (8) (17) (11) (5) (0) (11) (0) -i- 0.1

Succinic

(2)

9

9

46 7 11 7 0 7 8 6 7 9 11

(74) (40) (45) (33) (0) (42) (43) (40) (34) (50) • 4 (48)

Acetic

With hemin

0

2

3 2 2 0 0 1 0 0 0 1 0.9

(5) (11) (8) (0) (0) (6) (0) (0) (0) (6) 4- 0.3

Lactic

(4)

9

8

13 (21) 8.5 (49) 11.5 (47) 14.5 (67) 15 (100) 8.5 (52) 10.5 (57) 9 (60) 13.5 (66) 8 (44) 11 • 0.8 (48)

Succinic

Table 3. Fatty acids in mM and as percentage of total acid ( ), produced by 12 strains of Bacteroide., ssp., determined after 7 days of growth in the absence or presence of hemin

o~

K ;~

~Z

Z ~ Z ,~

o

QUANTITATIVE GAS CHROMATOGRAPHY OF BACTEROIDES

II

mM/L

15

S I

i

1

2

_

J

I

I

i

3

4

5

6

TIME (DAYS)

Fig. 5. Mean quantities of acetic (e), lactic (zZ),succinic acid (~) and of total acid (&) formed by six strains after various incubation times.

being formed (Table 3). Only small amounts of succinic acid were formed (2 % of total acid). With hemin added to the medium, succinic and acetic acid were the major products (means of both acids 48 Vo of total acid). For each strain the total amount of acid was also increased. Formation of propionic acid was not influenced by hemin and is therefore not mentioned in Table 3. The patterns of the two reference strains were completely comparable with those of our strains. The time required for establishing the final pattern of fatty acids is 6-7 days, as given by Holdeman and Moore (1972). As the growth rate is increased in the presence of hemin we expected that the final pattern of fatty acids would be reached in less time. For 6 hemin-dependent strains the fatty acids and the total acid were determined on various days (Fig. 5). The amounts of succinic and acetic acid remained constant after 2 days, lactic acid production was maximal after 3 days.

12

J. G. E. M. LINDNERAND J. H. MARCELIS DISCUSSION

In contrast with the existing methods the described method for GSC does not require elaborate sample preparation and permits immediate sample analysis, as about 35 esterified samples or about 80 unesterified samples could be tested on the same day. The results were reproducible within 10 ~ for unesterified acids and within 5 ~ for esterified acids. The sensitivity of 1 m i for VFA and non-VFA seems sufficient. The use of hemin in the medium greatly facilitated all the work with Bacteroides and permitted the determination of fermentation products already after 3 days as shown in Fig. 4. The GSC analysis of fermentation products enables the division into strains that do or do not form propionic acid. The majority of the lesion strains forms propionic acid (72 ~, LS1 + LS2). Besides, a number of strains (61 ~ ) within this group can be distinguished by a very high production of acid. This high production of total acid probably contributes to virulence. The excessive production of acid, especially lactic and acetic acid might be responsible for the development of persistance of the inflammatory reaction in tissues. This characteristic is not restricted to one biochemical subspecies. By routine biochemical tests the majority of our lesion strains are identified as Bacteroidesfragilis ssp. fragilis. The isolates from feces belong to the subspecies vulgatus and distasonis. These results confirm those of Wemer et al. (1975) about Bacteroides in clinical material. The influence of hemin on growth was significant. With 2 #g/ml hemin in the medium both the growth rate and the growth yield of all our strains was 2-3 times higher. Also protoporphyrin showed the same effect, indicating the presence of a ferrochelatase in our strains. The hemin requirement shown by Varel and Bryant (1974) for 3 reference strains of B. fragilis suggests that also our 54 strains of Bacteroides use hemin or protoporphyrin to synthesize cytochromes. The production of succinic acid in the presence of hemin in all our strains may indicate that our strains contain also a functional fumarate reductase as demonstrated for 2 strains of B.fragilis ssp.fragilis by Macy et al. (1975). There is evidence that many anaerobes including Bacteroides spp. contain a type of electron transport system with one or more cytochromes (de Vries et al., 1973, Macy et al., 1975). The stimulation by bile might be due to either the presence of free hemin or preoursors in bile or the capacity of the strains to convert bile pigments into hemin and cytochromes. If no hemin or related substances can be found in oxgall a study with pure substances (as biliverdin bilirubin, etc.) is indicated. We wish to thank Dr. J. A. A. Hoogkamp-Korstanje, Prof. Dr. K. C.

QUANTITATIVE GAS CHROMATOGRAPHY OF BACTEROIDES

13

W i n k l e r , M r s . A . J. d e n D a a s - S l a g t a n d M r . N. M. d e V o s f o r h e l p f u l a d v i c e and technical assistance.

R e c e i v e d 8 A u g u s t 1977

REFERENCES

BEERENS,H., WATTRE,P., SHINJO,T. et ROMOND,C. 1971. Premiers r6sultats d'un essay de classification s6rologique de 131 souches de bacteroides du groupefragilis (Eggerthella). Ann. Inst. Pasteur 121: 187-198. BEERENS, H. and FIEVEZ, L. 1971. Isolation of Bacteroidesfragilis and Sphaerophorus - Fusiformis groups, p. 109-113. In D. A. Shapton and R. G. Board, (eds.), Isolation of anaerobes. - - Academic Press, London and New York. Bergey, 8th ed. 1974. Edited by R. E. Buchanan and N. E. Gibbons. - - Williams and Wilkins, Baltimore. CALDWELL, D. R., WHITE, D. C., BRYANT, M. P. and DOETSCH, R. N. 1965. Specificity of the heme requirement for growth of Bacteroides ruminicola. - - J. Bacteriol. 90:1645-1654. CARLSSON,J. 1973. Simplified gas chromatographic procedure for identification of bacterial metabolic products. - - Appl. Microbiol. 25: 287-289. CHARLES,A. I . and BARRETT,F. C. 1963. Detection of volatile fatty acids produced by obligate gram-negative anaerobes. - J. Med. Lab. Technol. 20: 266-268. DAHEEN,G. and LINDE,A. 1973. Screening plate method for detection of bacterial fl-glucuronidase. - - Appl. Microbiol. 26: 863-866. DE LA CRUZ, E. and C~ADRA, C. 1969. Antigenic characteristics of five species of human Bacteroides. - - J. Bacteriol. 100:1116-1117. FINEGOLD, S. M., SUGIHARA,P. T. and SUTTER, V. L. 1971. Use of selective media for isolation of anaerobes from humans, p. 99-108. In D. A. Shapton and R. G. Board, (eds.), Isolation of anaerobes. - - Academic Press, London and New York. GUILLAUME,J., BEERENS, H. et OSTEUX, R. 1956. La chromatographie sur papier des acides aliphatiques volatiles de C a h C~. Son application/a la d6termination des bact6ries ana6robies. - - Ann. Inst. Pasteur Lille 8:13 23. HAUSER, K. J. and ZA~RANSKY,R. J. 1975. Modification of the gas - liquid chromatography procedure and evaluation of a new column packing material for the identification of anaerobic bacteria. - J. Clin. Microbiol. 2: 1-7. HOHORST, H,-J. 1970. L-(+)-Lactat Bestimmung mit Lactat-Dehydrogenase und NAD, S. 1425 1429. In H. U. Bergmeyer, Methoden der enzymatischen Analyse, I1. Bd. Verlag Chemie, Weinheim. HOLDEMAN, L. V. and MOORE, W. E. C. 1972. Anaerobic Laboratory Manual. - - Virginia Polytechnic Institute and State University, Blacksburg, Virginia. LAMBE, D. W., JR. and MOROZ, D. A. 1976. Serogrouping of Bacteroides fragilis subsp. fragilis by the agglutination test. - - J. Clin. Microbiol. 3: 586-592. MACY, J., PROBST, I. and GOTTSCHALK,G. 1975. Evidence of cytochrome involvement in fumarate reduction and adenosine 5"-triphosphate synthesis by Bacteroidesfragilis grown in the presence of heroin. - - J. Bacteriol. 123: 436-442. MOORE, W. E. C. 1970. Relationships of metabolic products to taxonomy of anaerobic bacteria. - - Int. J. Syst. Bacteriol. 20: 535-538. REINHOLD,L. 1971. Serologische Untersuchungen an St/immen von Bacteroides Thetaiotaomicron und Bacteroides Fragilis im Agargel pr/izipitations Test. - - Zbl. Bakt. I. Orig. 216: 219-227. RIZZA, V., SINCLAIR, P. R., WHITE, D. C. and CUORANT, P. R. 1968. Electron transport

14

J. G . E. M . LINDNER AND J. H . MARCELIS

system of the protoheme-requiring anaerobe Bacteroides melanogenicus. - - J. Bacteriol. 96:665-671 SALANITRO,J. P. and MUmHEAD, P. A. 1975. Quantitative method for the gas chromatographic analysis of short-chain monocarboxylic and dicarboxylic acids in fermentation media. - - Appl. Microbiol. 29: 374-381. SUFKIN, M. and HERCHER, H. J. 1974. Paper chromatography as an adjunct in the identification of anaerobic bacteria. - - Appl. Microbiol. 27: 500-505. VAgEL, V. H. and BRYANT, M. P. 1974. Nutritional features of Bacteroides fragilis subsp. f r a g i l i s . - Appl. Microbiol. 28: 251-257. DE VRIES, W., VAN WIJCK-KAPTEYN,W. M. C. and STOUTHAMER,A. H. 1973. Generation of ATP during cytochrome-linked anaerobic electron transport in propionic acid bacteria. - J. Gen. Microbiol. 76: 31-41. WATT, B., COLLEE,J. G. and BROWN, R. 1976. Tests of performance of anaerobic jars. - J. Clin. Pathol. 29: 534-536. WERNER, H. 1969. Das serologische Verhalten von St~immen der Species B. Convexus, Thetaiotaomicron, Vulgatus und Distasonis. - - Zbl. Bakt. I, Orig. 210: 192-201. WERNER, H., KUN~TEK-SANTOS,H., SCHOCKEMOBLE,C. und GUNDUREWA, M. 1975. Bakteroides und Appendizitis. - - Pathol. Microbiol. 42:110-118. VAN DER WIEL-KORSTANJE,J. A. A. and WINKLER, K. C. 1970. Medium for differential count of the anaerobic flora in human feces. - - Appl. Microbiol. 20:168-169.