The Journal of Nutrition Vol. 128 No. 1 January 1998, pp. 91-96

NMR Detection of ¹³CH₃¹³COOH from 3-¹³C-Glucose: A Signature for Bifidobacterium Fermentation in the Intestinal Tract¹

Meyer J. Wolin^{*, 2}, Yongchao Zhang, Shelton Bank, Susan Yerry^*, and Terry L. Miller^*

^* Wadsworth Center for Laboratories and Research, New York State Department of Health, Empire State Plaza, Albany, NY 12201-0509 and  Department of Chemistry, State University of New York-Albany, Albany, NY 12222

	ABSTRACT

Abstract Introduction Methods Results Discussion References

The gastrointestinal tracts of breast-fed infants are colonized more easily with bifidobacteria than are those of formula-fed infants. Colonization is thought to reduce infant diarrhea. Amendments to formulas that improve colonization by bifidobacteria are being actively investigated. Colonization studies almost invariably require measurements of the concentration of the bifidobacteria in feces to assess their importance in the colon. We investigated the use of nuclear magnetic resonance (NMR) analysis of products of fermentations of 1- and 3-¹³C-glucose to evaluate the importance of bifidobacteria in the colonic ecosystem. Bifidobacteria use a unique pathway of hexose catabolism to produce primarily acetate and lactate. The fermentation yields 3 mol of acetate from 2 mol of glucose. Two of the acetates are formed from C1 and C2 of glucose and the third is formed entirely from C3 of glucose. We first employed high resolution NMR to verify the pathway used by a pure culture of Bifidobacterium bifidum. The major products of fermentation of 1- and 3-¹³C-glucose were acetate and lactate. Most of the ¹³C from 3-¹³C-glucose was in ¹³CH₃¹³COOH with equal enrichment in the methyl and carboxyl groups. The ¹³C-acetate from 1-¹³C-glucose was almost entirely enriched in the methyl of acetate and no ¹³CH₃¹³COOH was produced. NMR analysis of glucose fermentation by the colonic flora of a 158-d-old strictly breast-fed infant showed production of ¹³CH₃¹³COOH from 3-¹³C-glucose. The amount of ¹³CH₃¹³COOH formed established that the Bifidobacterium pathway was the major pathway used for glucose fermentation by this infant's colonic microbes.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

Breast-fed infants are more easily colonized with bifidobacteria than are formula-fed infants (Benno et al. 1984). Colonization with bifidobacteria is thought to minimize infant diarrhea (Bullen et al. 1971). Addition of Bifidobacterium bifidum and Streptococcus thermophilus to infant formula diminished the occurrence of acute diarrhea and the shedding of rotavirus in infants in a pediatric hospital (Saavedra et al. 1994). The pursuit of amendments to infant formulas that improve colonic colonization by bifidobacteria and minimize the occurrence of diarrhea is an area of active investigation (Langhendries et al. 1995, Roberts 1986, Roberts et al. 1992). Establishment of bifidobacteria in the adult colon to reduce infectious and noninfectious diseases of the colon is also being studied (Gibson and Roberfroid 1995).

Studies of colonization of bifidobacteria in the colon almost invariably require measurement of the concentration of the organisms in feces to estimate their concentrations in the colon. Most studies use classical dilutions and plating in agar media that allow selective growth of bifidobacteria colonies (Sutter et al. 1985). Plating methods are labor intensive, time consuming, and subject to the difficulty of developing absolutely selective media. Recently, 16S ribosomal RNA gene probe analysis was successfully used to enumerate bifidobacteria in fecal suspensions (Langendijk et al. 1995).

We considered using analysis of fermentation products formed from 1- and 3-¹³C-glucose by fecal suspensions to determine the significance of bifidobacteria in the colonic ecosystem. Bifidobacteria use a unique fermentation pathway to produce primarily acetate and lactate (Scardovi and Trovatelli 1965, de Vries and Stouthamer 1967). Small amounts of ethanol and formate are also formed. Fermentation of 2 mol of glucose yields 2 mol of acetate with methyl groups from C1 and carboxyl groups from C2 of glucose. One mole of acetate is also formed with both the methyl and carboxyl groups arising from C3 of glucose. Two moles of glyceraldehyde-3-phosphate also are produced and come from C4, C5 and C6 of glucose. Glyceraldehyde-3-phosphate is converted to lactate, acetate, ethanol and formate.

NMR analysis can unambiguously detect acetate formed from C3 of glucose. Acetate formed from 3-¹³C-glucose will contain ¹³C in both the methyl and carboxyl groups and produce a unique NMR signal. Therefore, measurement of the amount of ¹³CH₃¹³COOH produced from fermentation of 3-¹³C-glucose by fecal suspensions can determine the contribution of bifidobacteria to the total fermentation of the microbial community.

The purpose of this study was to confirm that the predictions of the pathway were correct, i.e., that fermentation of 3-¹³C-glucose does yield ¹³CH₃¹³COOH. We studied the fermentation of 1- and 3-¹³C-glucose with ¹²CO₂ and ¹²C-glucose and ¹³CO₂ by B. bifidum. The results in this report prove the prediction. We included 1-¹³C-glucose and ¹³CO₂ as substrates to provide additional evidence for the origin of the carbons of acetate. We used the method to show that the microbial community in the fecal suspensions of a 158-d-old breast-fed infant fermented glucose by the Bifidobacterium fermentation pathway.

	MATERIALS AND METHODS

Abstract Introduction Methods Results Discussion References

Organism and growth.  B. bifidum ATCC 29521 was obtained from the American Type Culture Collection (Rockville, MD). It was grown anaerobically by use of the serum bottle modification of the Hungate technique (Miller and Wolin 1974). The medium used for growth of pure cultures contained (per liter): NaHCO₃, 7.5 g; K₂HPO₄, 0.3 g; KH₂PO₄, 0.3 g; (NH₄)₂SO₄, 0.3 g; NH₄Cl, 1 g; NaCl, 0.6 g; MgSO₄·7H₂O, 0.15 g; CaCl₂·2H₂O, 80 mg; MnSO₄·H₂O, 4.5 mg; FeSO₄·7H₂O, 3.0 mg; CoSO₄·7H₂O, 1.8 mg; ZnSO₄·7H₂O, 1.8 mg; CuSO₄·5H₂O, 100 µg; AlK(SO₄)₂·12H₂O, 180 µg; Na₂MoO₄·2H₂O, 100 µg; H₃BO₃, 100 µg; Na₂SeO₄, 1.9 mg; NiCl₂·6H₂O, 92 µg; nitrilotriacetic acid, 15 mg; thiamine·HCl, 2.0 mg; D-pantothenic acid, 2.0 mg; nicotinamide, 2.0 mg; riboflavin, 2.0 mg; pyridoxine·HCl, 2.0 mg; biotin, 10.0 mg; cyanocobalamin, 20 µg; p-aminobenzoic acid, 100 µg; folic acid, 50 µg; cysteine·HCl·H₂O, 0.5 g; rumen fluid, 100 mL; sodium formate, 2.5 g; sodium acetate, 2.5 g; glucose, 10.0 g; and yeast extract, 2.0 g. Resazurin (1 mg/L) was added as an oxidation-reduction potential indicator. Growth was at 37°C in an atmosphere of 100% CO₂ that was passed through a heated column containing copper filings to remove trace amounts of oxygen.

For studies of the fermentation of ¹³C substrates, 25 mg of 1-¹³C-glucose and 3-¹³C-glucose, respectively, were autoclaved in 10-mL serum bottles with 80% N₂:20% CO₂. After cooling, 5 mL of sterile medium without added glucose was injected into each bottle. The medium was reduced by the addition of 0.15 mL of a sterile solution containing 71 mmol/L cysteine and 52 mmol/L sodium sulfide. The bottles were gassed with 100% CO₂ for 1 min and then inoculated with 0.25 mL of a 24-h culture grown in the same medium with 56 mmol/L glucose. The same amount of culture was added to a third bottle containing 5 mL of medium plus 0.15 mL of the cysteine/sulfide solution. It was acidified by addition of 0.5 mL of 2.5 mol/L H₂SO₄ and served as the control for the concentration of fermentation products in the unincubated medium. After incubation of the bottles with labeled glucose for 72 h, gas samples were removed for gas chromatography and mass-spectrometric analyses. The cultures were acidified, as described above, and centrifuged along with the acidified control at 16,000 × g for 15 min. The supernatants were analyzed for fermentation products and residual glucose.

Fecal suspension.  A fecal suspension from a 158-d-old exclusively breast-fed infant was prepared in anaerobic dilution solution under CO₂ as described previously (Miller and Wolin 1982, Weaver et al. 1986). The suspension was held at 4°C and used within 24 h of collection. Anaerobic conditions were maintained by use of the serum bottle-modification of the Hungate technique (Miller and Wolin 1974). Duplicate portions (5 mL) were dried to constant weight to determine dry matter content. Human fecal fermentation protocols were reviewed and approved by the New York State Department of Health Institutional Review Board

Fermentations by a fecal suspension.  Fermentation of glucose was performed in 10-mL serum bottles. Glucose (25 mg) was added as a dry powder before insertion of butyl rubber stoppers and sealing with aluminum seals. The bottles were gassed with 80% N₂:20% CO₂. A portion (2.5 mL) of 76 mmol/L NaHCO₃ from a serum bottle with 100% N₂ was injected and the bottles were cooled to 0°C in ice water. After the addition of 2.5 mL of the suspension and incubation for 24 h at 37°C, the bottles were boiled for 10 min. The contents were either analyzed immediately or frozen at 20°C and thawed before removing gas samples for gas chromatography and mass-spectrometric (GC-MS) analyses. Suspensions were then acidified by the addition of 0.5 mL of 2.5 mol/L H₂SO₄ and centrifuged at 16,000 × g for 15 min. Supernatants were analyzed for fermentation products and residual glucose.

Chromatographic methods.  Soluble fermentation products and glucose were determined by HPLC procedures (Ehrlich et al. 1981). We used an Aminex ion-exclusion column (HPX-87H, Bio-Rad, Richmond, CA) at 35° C and eluted with 0.0065 mol/L H₂SO₄ (0.55 mL/min, 77 kg/cm²). Compounds were detected by refractive index and identified and quantified by using a Shimadzu C-R3A integrator (Shimadzu Scientific, Baltimore, MD) and the absolute calibration curve method. H₂ and CH₄ were quantified using described gas chromatographic procedures (Pavlostathis et al. 1988).

Mass spectrometry.  Mass spectral analyses were conducted on a Hewlett-Packard 5890A Gas Chromatography-Hewlett-Packard 5970 Series Mass Selective Detector (Hewlett-Packard, Palo Alto, CA). Helium was the carrier gas. Temperatures were as follows: GC oven 50°C, transfer line 200°C, source 200°C. Electron ionization was 70 eV, and the mass range was 40-50. Five microliters of gas sample was used for each GC-MS analysis.

Nuckear magnetic resonance (NMR) methods.  NMR spectra were acquired on a Varian XL-300 spectrometer (Varian Associates, Walnut Creek, CA) operating at 75.43 MHz. Pulses of 36° were used and the delay time was 1 s. The number of transients ranged from 5000 to 40,000. The NMR locking material was deuterium oxide. All spectra were recorded at 25°C with a spectral width of 16,000 Hz.

Chemical shifts (ppm) for acetate can be assigned directly by using literature values and samples of labeled acetate. A pattern of two coupled doublets (J = 52 Hz) at 177.6 ppm and 21.3 ppm, respectively, shows the presence of double-labeled acetate. Chemical shifts for other compounds were determined by preparing solutions of authentic materials under the same conditions as the fermentation samples. Doping experiments, however, were needed when some products had close and/or variable resonance signals; that is, a predetermined amount of one of the candidates was added into the sample and the enhanced signals were therefore due to the species added.

Measurement of the percentage enrichment of ¹³C used dioxane as an external standard; it gives a strong singlet at 67.4 ppm that is separate from the signals of the species under study. Dioxane (5 µL) was sealed in a capillary with D₂O, and this capillary was used as reference each time a spectrum was acquired. Three standard samples of labeled acetate were prepared with identical concentrations of 53.63 mmol/L. Two were single-labeled with ¹³C at either C-1 or C-2; the third was double-labeled. All three samples had an enrichment of 31.01%. Their spectra were acquired under identical conditions with the above reference capillary in the NMR tube. All three samples showed agreement in the ratios of peak intensities of C-1 or C-2 to dioxane, determined as 2.1 or 3.1 (symbolized as R_o). Thus the spectrum of a fermentation sample with a known mmol/L concentration (C) was acquired with the same capillary and the ratio of its peak intensity to dioxane was obtained (R). To convert this ratio R to percentage enrichment, we normalize with the standard sample of enrichment 31.01% and concentration of 53.63 mmol/L, leading to the following equation:

NMR analysis of fermentation samples.  Acidified fermentation supernatant was added to the NMR tube without adding any solvents. The same reference capillary was also put into the tube each time. Identification of products present and enrichment of ¹³C were determined by the methods described above.

Materials.  ¹³C-Labeled C-1, C-2 and doubly labeled sodium acetate were from MSD Isotopes (Montreal, Canada) and were 99% enriched. ¹³C-Labeled glucose and sodium bicarbonate (99% enriched) were purchased from Cambridge Isotope Laboratories, Woburn MA. All other chemicals were of reagent grade or better.

	RESULTS

Abstract Introduction Methods Results Discussion References

Bifidobacterium pathway.  Bifidobacteria produce 3 mol of acetic acid and 2 mol of lactic acid from 2 mol of glucose as shown in Figure. 1 (de Vries et al. 1967). The major carbon transformations involve the enzymes transketolase, transaldolase and phosphoketolase to form the two- and three-carbon precursors of acetate and lactate. Formation of ¹³CH₃¹²COOH from C1 and C2 of 1-¹³C-glucose and ¹³CH₃¹³COOH from 3-¹³C-glucose is predicted from the pathway. Table 1 shows the expected labeling of acetate when the different known microbial fermentation pathways ferment 1-¹³C- and 3-¹³C-glucose.

View larger version (29K): [in this window] [in a new window]   Fig 1. Bifidobacterium pathway. Glu, glucose; Fru, fructose; F-6-P, fructose-6-phosphate; Pi, inorganic phosphate; Ac~P, acetyl~phosphate; E-4-P, erythrose-4-phosphate; S-7-P, sedoheptulose-7-phosphate; G-3-P, glyceraldehyde-3-phosphate; X-5-P, xylulose-5-phosphate.

Products of glucose fermentation.  We measured the products of fermentation of 1- and 3-¹³C-glucose and unlabeled CO₂ and unlabeled glucose and ¹³CO₂ by B. bifidum. HPLC analysis showed that the major products were acetate and lactate with small amounts of ethanol and formate (Table 2). Gas chromatography showed that H₂ and CH₄ were absent. The concentration of acetate (40.8 mmol/L) in the inoculated medium before incubation was approximately the same as the amount produced during incubation.

NMR and mass spectrometric analyses.  Figure 2 shows the acetate methyl region of the NMR spectra of authentic ¹³CH₃¹²COOH,¹²CH₃¹³COOH and ¹³CH₃¹³COOH. A singlet signal at 21.3 ppm is produced when ¹³C is only in the methyl group, and a singlet is produced at 177.6 ppm when only the carboxyl group contains ¹³C. When both the methyl and carboxyl groups are labeled with ¹³C, doublet signals are produced at both 21.3 and 177.6 ppm. The appearance of these coupled doublets in a fermentation sample proves that double-labeling of acetate has occurred. Figure 3 shows the NMR spectra of the supernatants obtained after fermentation of 1-¹³C-glucose and 3-¹³C-glucose by B. bifidum. When 1-¹³C-glucose was fermented, a singlet was detected at 21.3 ppm but not at 177.6 ppm and no doublet signals were found. ¹³CH₃¹²COOH was the only labeled species formed. The pronounced doublet signals found at 21.3 ppm and 177.6 ppm show that ¹³CH₃¹³COOH was the only labeled species formed from 3-¹³C-glucose.

View larger version (9K): [in this window] [in a new window]   Fig 2. Nuclear magnetic resonance (NMR) spectra of authentic acetate species [53.63 mmol/L, enrichment (31.01%)]: (a) 13CH312COOH, (b) 12CH313COOH and (c) 13CH313COOH. The singlet at 21.3 ppm is from 13CH312COOH. The singlet at 177.6 ppm is from 12CH313COOH. The doublet signals at 21.3 ppm and 177.6 ppm are from 13CH313COOH.The chemical shift reference peak at 67.4 ppm is from the methylene of dioxane.

View larger version (9K): [in this window] [in a new window]   Fig 3. Nuclear magnetic resonance (NMR) spectra of the supernatants obtained after fermentation of 13C-glucose by B. bifidum: (a) 1-13C-glucose fermentation supernatant and (b) 3-13C-glucose supernatant. The singlet at 21.3 ppm is from 13CH312COOH. The doublet signals at 21.3 ppm and 177.6 ppm are from 13CH313COOH. The chemical shift reference peak at 67.4 ppm is from the methylene of dioxane.

Table 3 shows the amount of enrichment of the carbon atoms of acetate formed from 3-¹³C-glucose and 1-¹³C-glucose. Most of the ¹³C from 3-¹³C-glucose was in ¹³CH₃¹³COOH that was equally enriched in the methyl and carboxyl groups (Table 3). Almost all of the acetate formed from 1-¹³C-glucose was enriched in the methyl of acetate and no ¹³CH₃¹³COOH was produced. The final concentrations of the different labeled species of acetate (Table 3) were calculated by multiplying the percentage of enrichment of the relevant carbon by the total acetate concentration measured at the end of the fermentation period. Expected enrichments and concentrations of the labeled species (shown in parentheses in Table 3) are the amounts expected if acetate was formed by cleavage of the fructose-6-P and xylulose-5-P of the bifidobacteria pathway (Fig. 1).

A small amount of ¹³C from 3-¹³C-glucose was found as a single enriched methyl group (¹³CH₃¹²COOH) and as a single enriched carboxyl group (¹²CH₃¹³COOH) (Table 3). Mass spectrometric analysis showed a small enrichment of CO₂ formed from 1-¹³C-glucose and 3-¹³C-glucose. The percentage of enrichment was 2.8 from 1-¹³C-glucose and 1.5 from 3-¹³C-glucose. ¹³CO₂ was not incorporated into acetate.

Fermentation by suspensions of infant feces.  Glucose (1-¹³C and 3-¹³C) was fermented by a suspension of a sample collected from a 158-d-old exclusively breast-fed infant. Net products, after subtracting products formed in the absence of glucose, were as follows (mmol/L): acetate (32.3); ethanol (8.5); and formate (1.2).

Figure 4 shows the NMR spectra of the supernatants obtained after fermentation of 3-¹³C-glucose by the microorganisms in the infant's feces. The doublets found at 21.3 ppm and 177.6 ppm demonstrated that acetate labeled in ¹³CH₃¹³COOH was formed from 3-¹³C-glucose. Table 4 shows the enrichment of the carbon atoms of acetate formed from 3-¹³C-glucose and 1-¹³C-glucose. Only a singlet was detected at 21.3 and 177.6 ppm when 1-¹³C-glucose was fermented, i.e., only ¹³CH₃¹²COOH but not ¹³CH₃¹³COOH was produced. The results demonstrate that the microbial community uses the Bifidobacterium pathway for the fermentation of glucose.

View larger version (15K): [in this window] [in a new window]   Fig 4. Nuclear magnetic resonance (NMR) spectra of the supernatant obtained after fermentation of 3-13C-glucose by fecal suspensions of a 158-d-old exclusively breast-fed baby. From chemical shift region of (a) 0-65 ppm and (b) from 4 to 208 ppm. Peak 1 is the doublet of the methyl C of 13CH313COOH. Peak 2 at 67.4 ppm is the chemical shift reference peak of the methylene of dioxane. Peak 3 is the doublet of the carboxyl C of 13CH313COOH.

The sum of the final concentrations of the different labeled species of acetate (Table 4) was 28.4 mmol/L or 88% of the net acetate produced. The concentration of the double-labeled species used for the calculation was the average estimated from the enrichment of the methyl and carboxyl carbons (6.4 mmol/L). According to the Bifidobacterium pathway, 12.8 mmol/L of methyl-labeled acetate should have been formed from 1-¹³C-glucose. Use of another pathway, e.g., the Embden-Meyerhof-Parnas path, by some of the microbes in the suspension would account for the 6 mmol/L of methyl-labeled acetate that are not associated with the Bifidobacterium fermentation.

Enrichment in acetate with single-labeled methyl or carboxyl carbons was 1.1 and 0.7%, respectively, and no double-labeled acetate was formed when ¹³CO₂ was used in the presence of ¹²C-glucose. Enrichment in CO₂ formed from 1-¹³C- and 3-¹³C-glucose was 1.7 and 1.8%, respectively.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

Production of ¹³CH₃¹³COOH from 3-¹³C-glucose provides a clear signal for the unique fermentation pathway of bifidobacteria. The amounts of ¹³CH₃¹³COOH and ¹³CH₃¹²COOH produced from 3-¹³C-glucose and 1-¹³C-glucose by the B. bifidum strain we examined are the amounts expected if acetate is formed by cleavage of the fructose-6-P and xylulose-5-P of the bifidobacteria pathway (Fig. 1). The results provide additional confirmation of the pathway originally proposed by de Vries et al. (1967) and Scardovi and Trovatelli (1965). The pathway yields 2 mol of xylulose-5-phosphate for 2 mol of hexose used. One mole of the xylulose-5-phosphate is enriched in both the C-1 and C-2 positions when 3-¹³C-glucose is fermented. Cleavage to acetyl-phosphate precedes the formation of ¹³CH₃¹³COOH. The unique doublet signal of the double-labeled molecule is easily detected by NMR analysis (Fig. 2).

Additional acetate is also made from acetyl-phosphate formed from the C-1 and C-2 positions of 1 mol of fructose-6-phosphate and the remaining mole of xylulose-5-phosphate produced from 2 mol of hexose (Fig. 1). Carbon 1 of these sugar phosphates is enriched when 1-¹³C-glucose is fermented and is the origin of acetate labeled solely in the methyl group.

The amounts of enriched acetate formed from 1-¹³C-glucose and 3-¹³C-glucose by B. bifidum were slightly lower than the amounts expected if all of the acetate was formed from the C-1 and C-2 of the sugar phosphates. However, many strains of bifidobacteria, including the one used in this study, produce some acetate and ethanol from pyruvate (Lauer and Kandler 1976, de Vries and Stouthamer 1968). Glyceraldehyde-3-phosphate, produced from C4, C5 and C6 of glucose (Fig. 1) is converted to pyruvate, which is then converted to lactate, acetate, ethanol and formate. Pyruvate carbons are not enriched by fermentation of 1-¹³C-glucose and 3-¹³C-glucose. Pyruvate is the precursor of the unlabeled acetate formed in the Bifidobacterium pathway.

Using enzymatic assays, de Vries and Stouthamer (1968), showed that 17 different strains of Bifidobacterium used the pathway shown in Fig. 1. Because all members of the genus use the pathway, fermentation of 3-¹³C-glucose and detection of ¹³CH₃¹³COOH by NMR spectroscopy provide a distinct signature for bifidobacterial fermentation in colonic or other ecosystems. Other fermentation pathways yield acetates that have the same origin in C1 and C2 of glucose as the acetate formed by bifidobacteria (Table 1). However, only the bifidobacterial pathway yields an acetate in which both carbons originate from C3 of glucose.

The Bifidobacterium pathway is the major fermentation path used by the colonic microbial community of the 158-d old-infant we examined. Approximately 70% of acetate formed by fermentation of glucose can be accounted for by the pathway. Because ¹³CO₂ was not incorporated into double-labeled acetate, the possibility that acetate was formed by reduction of CO₂ is ruled out. Our results with the baby fecal fermentation system show that NMR analysis can be used to measure the contribution of the Bifidobacterium fermentation pathway to the fermentation in intestinal tract ecosystems. Analysis consists of incubation with 3-¹³C-glucose, measurement of acetate production and NMR analysis. Very little sample preparation is necessary for either analysis. We added ¹²CH₃¹²COOH to the medium and showed that ¹³CH₃¹³COOH could be detected in a background of acetate with natural abundance of ¹³C (1% enrichment). A specific NMR method to assess bifidobacterial activity provides a useful adjunct to selective colony counts and to ribosomal RNA gene probes for detection of bifidobacteria.

	FOOTNOTES

Manuscript received 11 March 1997. Initial reviews completed 25 April 1997. Revision accepted 26 September 1997.

	LITERATURE CITED

Abstract Introduction Methods Results Discussion References

• Benno Y., Sawada K., Mitsuoka T. The intestinal microflora of infants: composition of fecal flora in breast-fed and bottle-fed infants. Microbiol. Immunol. 1984; 28:975-986 [Medline]
• Bullen C. L., Tearle P. V., Stewart M. C. Resistance of breast fed infants to gastroenteritis. Br. Med. J. 1971; 3:338-344
• Ehrlich G. G., Goerlitz D. F., Bourell J. H., Eisen G. V., Godsy E. M. Liquid chromatographic procedure for fermentation product analysis in the identification of anaerobic bacteria. Appl. Environ. Microbiol. 1981; 42:878-885 ^{[Abstract/Free Full Text]}
• Gibson G. R., Roberfroid M. B. Dietary modulation of the human colonic microbiota: introducing the concept of prebiotics. J. Nutr. 1995; 125:1401-1412
• Langendijk P. L., Schut F., Jansen G. J., Raangs G. C., Kamphuis G. R., Wilkinson M.H.F., Welling G. W. Quantitative fluorescence in situ hybridization of Bifidobacterium spp. with genus-specific 16S rRNA-targeted probes and its application in fecal samples. Appl. Environ. Microbiol. 1995; 61:3069-3075 ^[Abstract]
• Langhendries J. P., Detry J., Van Hees J., Lamboray J. M., Darimont J., Mozin M., Mc Secretin, J., Senterre J. Effect of a fermented infant formula containing viable bifidobacteria on the fecal flora composition and pH of healthy full-term infants. J. Pediatr. Gastroenterol. Nutr. 1995; 21:177-181 [Medline]
• Lauer E., Kandler O. Mechanismus der Variation des Verhältnisses Acetat/Lactat bei der Vergärung von Glucos durch Bifidobacterien. Arch. Microbiol. 1976; 110:271-277 [Medline]
• Miller T. L., Wolin M. J. A serum bottle modification of the Hungate technique for cultivating obligate anaerobes. Appl. Microbiol. 1974; 27:985-987 [Medline]
• Miller T. L., Wolin M. J. Enumeration of Methanobrevibacter smithii in human feces. Arch. Microbiol. 1982; 131:14-18 [Medline]
• Pavlostathis S. G., Miller T. L., Wolin M. J. Fermentation of insoluble cellulose by continuous cultures of Ruminococcus albus. Appl. Environ. Microbiol. 1988; 54:2655-2659 ^{[Abstract/Free Full Text]}
• Roberts, A. K. (1986) Prospects for further approximation of infant formulas to human milk. Human Nutr. Appl. Nutr. 40 (Suppl. 1): 27-37.
• Roberts A. K., Chierici R., Sawatzki G., Hill M. J., Volpato S., Vigi V. . Supplementation of an adapted formula with bovine lactoferrin: 1. Effect on the infant faecal flora. Acta Paediatr. 1992; 81:119-124 [Medline]
• Saavedra J. M., Bauman N. A., Oung I., Perman J. A., Yolken R. H., Robert H. Feeding of Bifidobacterium bifidum and Streptococcus thermophilus to infants in hospital for prevention of diarrhoea and shedding of rotavirus. The Lancet 1994; 344:1046-1049 [Medline]
• Scardovi, V. & Trovatelli, L. D. ( 1965) The fructose-6-phosphate shunt as peculiar pattern of hexose degradation in the genus Bifidobacterium. Ann. Microbiol. 15: 19-29.
• Sutter, V. L., Citron, D. M., Edelstein, M. A. C.& Finegold, S. M. (1985) Wadsworth Anaerobic Bacteriology Manual, 4th ed. Star Publication, Belmont, CA.
• de Vries W., Gerbrandy S. J., Stouthamer A. H. Carbohydrate metabolism in Bifidobacterium bifidum. Biochim. Biophys. Acta 1967; 136:415-425 ^[Medline]
• de Vries W., Stouthamer A. H. Pathway of glucose fermentation in relation to the taxonomy of bifidobacteria. J. Bacteriol. 1967; 93:574-576 ^{[Abstract/Free Full Text]}
• de Vries W., Stouthamer A. H. Fermentation of glucose, lactose, galactose, mannitol, and xylose by Bifidobacteria. J. Bacteriol. 1968; 96:472-478 ^{[Abstract/Free Full Text]}
• Weaver G. A., Krause J. A., Miller T. L., Wolin M. J. Incidence of methanogenic bacteria in a sigmoidoscopy population: an association of methanogenic bacteria and diverticulosis. Gut 1986; 27:698-704 [Abstract/Free Full Text]

0022-3166/98 $3.00 ©1998 American Society for Nutritional Sciences