Quantification of Human Fecal Bifidobacterium Species by Use of Quantitative Real-Time PCR Analysis Targeting the groEL Gene

Abstract

Quantitative real-time PCR assays targeting the groEL gene for the specific enumeration of 12 human fecal Bifidobacterium species were developed. The housekeeping gene groEL (HSP60 in eukaryotes) was used as a discriminative marker for the differentiation of Bifidobacterium adolescentis, B. angulatum, B. animalis, B. bifidum, B. breve, B. catenulatum, B. dentium, B. gallicum, B. longum, B. pseudocatenulatum, B. pseudolongum, and B. thermophilum. The bifidobacterial chromosome contains a single copy of the groEL gene, allowing the determination of the cell number by quantification of the groEL copy number. Real-time PCR assays were validated by comparing fecal samples spiked with known numbers of a given Bifidobacterium species. Independent of the Bifidobacterium species tested, the proportion of groEL copies recovered from fecal samples spiked with 5 to 9 log₁₀ cells/g feces was approximately 50%. The quantification limit was 5 to 6 log₁₀ groEL copies/g feces. The interassay variability was less than 10%, and variability between different DNA extractions was less than 23%. The method developed was applied to fecal samples from healthy adults and full-term breast-fed infants. Bifidobacterial diversity in both adults and infants was low, with mostly ≤3 Bifidobacterium species and B. longum frequently detected. The predominant species in infant and adult fecal samples were B. breve and B. adolescentis, respectively. It was possible to distinguish B. catenulatum and B. pseudocatenulatum. We conclude that the groEL gene is a suitable molecular marker for the specific and accurate quantification of human fecal Bifidobacterium species by real-time PCR.

INTRODUCTION

Microbial colonization of the initially sterile intestine starts with the exposure of the newborn to microbes from the mother and the environment. In this phase, Bifidobacterium species become predominant, representing 60 to 91% of fecal bacteria in breast-fed infants and 28 to 75% in formula-fed infants (17). At 2 years of age, when the complex gut microbiota is fully established, the proportion of fecal bifidobacteria decreases to 1 to 3% (41). Even lower levels or absence of bifidobacteria have been reported for the elderly (3, 15, 19, 38).

The establishment of the intestinal microbiota in early life is a critical phase that affects the maturation of the immune system. Numerous studies show that breast-fed infants have a lower incidence of gastrointestinal infections and atopic diseases than formula-fed infants (20, 46). It has been proposed that the health-promoting effect of breast milk is in part mediated by the intestinal microbiota, which is characterized by low diversity and a high proportion of bifidobacteria (11).

Common Bifidobacterium species of the human intestinal microbiota include Bifidobacterium adolescentis, B. angulatum, B. bifidum, B. breve, B. catenulatum, B. dentium, B. longum, and B. pseudocatenulatum (7). B. gallicum has rarely been detected (6, 7). Recently, B. pseudolongum and B. thermophilum were isolated from human adult and baby feces, respectively (44, 53). These species have previously been considered to be of animal origin (7). B. animalis subsp. lactis is commonly used as probiotic and thus may be detected in human feces.

Bifidobacteria are considered beneficial members of the gut microbiota. The assessment of their diversity and population size in the gastrointestinal tract is therefore important. However, only a few studies have enumerated bifidobacteria to the species level, and the majority of them used cultivation-based techniques (7). The latter are not only laborious but also hampered by partly fastidious growth conditions of bifidobacteria and by the lack of growth medium selectivity (2).

Since the advent of cultivation-independent methods, the 16S rRNA gene has been widely used as a valuable tool for bacterial identification (12). However, the resolution power of the 16S rRNA gene among closely related species is limited. Isolates displaying more than 97% 16S rRNA gene sequence identity are usually considered the same species. Since Bifidobacterium species reveal a relatively high 16S rRNA gene sequence identity (mean, 95% [31, 47]), more discriminative identification markers are needed.

Alternative target genes for the differentiation of Bifidobacterium species include housekeeping genes, such as atpD (49), dnaK (52), groEL (21, 27, 47, 50, 55), groES (50), recA (23, 51), tal (37), tuf (48, 51), and xfp (5, 47, 54). All these genes except groES were demonstrated to have similar or even higher discriminating power for bifidobacteria than the 16S rRNA gene. Unfortunately, only a limited number of sequences of these marker genes are available. The most sequences exist for the groEL gene. The Chaperonin Sequence Database (http://www.cpndb.ca; 18) currently contains more than 13,000 entries for prokaryotes, eukaryotes, and archaea, 121 of which belong to 27 Bifidobacterium species. The groEL gene encodes the chaperonin GroEL (synonyms are Cpn60, GroL, Hsp60, and MopA), which plays an essential role in the handling of cellular stress. For example, it promotes refolding of misfolded polypeptides. Southern blot experiments and sequence analysis of whole genomes have revealed that there is just one groEL copy per genome in bifidobacteria (14, 42, 50, 56). This facilitates quantitative analyses of bifidobacteria.

Here, we report the quantification of Bifidobacterium species in human feces with quantitative real-time PCR (qPCR) using groEL as a discriminative marker. Ninety-seven partial (∼600-bp) or complete (∼1,600-bp) groEL sequences of 12 Bifidobacterium species (Table 1) were used to design species-specific primers. They were applied to SYBR green I chemistry-based qPCR for quantification of Bifidobacterium species. The targeted species include members of the human gut microbiota and B. animalis, some strains of which are used as probiotics.

Table 1.

Bacterial strains used for testing the specificity of the developed primers by qPCR

Bifidobacterium strain(s)b	qPCR results with primersa:
	B_ado-f/B_ado-r	B_ang-f/B_ang-r	B_ani-f/B_ani-r	B_bif-f/B_bif-r	B_bre-f/B_bre-r	B_cat-f/B_cat-r	B_den-f/B_den-r	B_gal-f/B_gal-r	B_lon-f/B_lon-r	B_pcat-f/B_pcat-r	B_plon-f/B_plon-r	B_the-f/B_the-r
B. adolescentisDSM 20083T, DSM 20086	+	−	−	−	−	−	−	−	−	−	−	−
B. angulatumATCC 27535T, DSM 20225	−	+	−	−	−	−	−	−	−	−	−	−
B. animalis DSM 20105	−	−	+	−	−	−	−	−	−	−	−	−
B. animalis subsp. animalis DSM 20104T	−	−	+	−	−	−	−	−	−	−	−	−
B. animalis subsp. lactisDSM 10140T	−	−	+	−	−	−	−	−	−	−	−	−
B. bifidumDSM 20456T, DSM 20215, DSM 20239	−	−	−	+	−	−	−	−	−	−	−	−
B. breveDSM 20213T, DSM 20091	−	−	−	−c	+	−	−	−	−	−	−	−
B. catenulatumATCC 27539T, DSM 20224	−	−	−	−	−	+	−	−	−	−	−	−
B. dentiumATCC 27534T, ATCC 27678, DSM 20084	−	−	−	−	−	−	+	−	−	−	−	−
B. gallicumDSM 20093T	−	−	−	−	−	−	−	+	−	−	−	−
B. longum subsp. longumATCC 15707T	−	−	−	−	−	−	−	−	+	−	−	−
B. longum subsp. suis DSM 20211T	−	−	−	−	−	−	−	−	+	−	−	−
B. pseudocatenulatumATCC 27919T, DSM 20439	−	−	−	−	−	−	−	−	−	+	−	−
B. pseudolongum subsp. pseudolongumATCC 25526T, DSM 20094	−	−	−	−	−	−	−	−	−	−	+	−
B. pseudolongum subsp. globosum DSM 20092T	−	−	−	−	−	−	−	−	−	−	+	−
B. thermophilumDSM 20210T, DSM 20209, DSM 20212	−	−	−	−	−	−	−	−	−	−	−	+

^aIn addition to the bifidobacterial strains listed, negative qPCR results were obtained for the following bacterial strains: B. producta DSM 2950^T, B. thetaiotaomicron DSM 2079^T, B. vulgatus DSM 1447^T, Clostridium histolyticum DSM 2158^T, Escherichia coli DSM 30083^T, Faecalbacterium prausnitzii DSM 17677, Lactobacillus acidophilus DSM 20079^T, Lactobacillus gasseri DSM 20243^T, Ruminococcus albus DSM 20455^T, and Ruminococcus gauvreauii DSM 19829^T. +, specific PCR product; −, no PCR product.

^bThe underlined strains served as representative strains for the respective species and were used to generate qPCR standards.

^cIn the case of B. breve DSM 20091, a nonspecific PCR product was observed.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

The strains listed in Table 1 were obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ), Braunschweig, Germany, and the American Type Culture Collection (ATCC), Manassas, VA. All bacteria were cultured anaerobically under a gas phase of N₂-CO₂ (80:20 [vol/vol]) in 15-ml Hungate tubes containing 10 ml ST medium, except for Faecalibacterium prausnitzii, which was grown on YCFA GSC medium (NCIMB Ltd., Aberdeen, Scotland) (9). The cultures were incubated overnight at 37°C and grown to a density of 9 to 10 log₁₀ cells/ml.

The purity of bacterial cultures was checked by Gram staining and plating on Columbia agar with 5% sheep blood (bioMérieux, Vienna, Austria). The identities of the Bifidobacterium strains underlined in Table 1 (used to generate qPCR standards) were confirmed at the species level by sequencing either the 16S rRNA gene or the groEL gene (Eurofins MWG Operon, Ebersberg, Germany).

Subjects and fecal samples.

Five healthy, full-term, exclusively breast-fed infants (4 males and 1 female) up to 3 months old with birth weights of 2,500 g to 4,500 g were included in the study. The infants had been born by caesarean section, except INF-4, who had been born by vaginal delivery. Infants undergoing antibiotic therapy during the first 14 days of life were excluded. Newborns were enrolled in a clinical study whose protocol was reviewed and approved by an independent ethics committee, and informed written consent was obtained from a legal representative(s). Fecal samples from infants were collected within 30 min after defecation and were kept in an anaerobic jar with a reduced atmosphere created by AnaeroGen (Oxoid, United Kingdom) at 4°C for up to 8 h until they were stored at −20°C. Ten healthy adults from our institute staff (5 males and 5 females), 20 to 40 years old, voluntarily provided fecal samples. They did not have any gastrointestinal disorders during sample collection and did not undergo antibiotic treatment in the 6 months prior to sampling. The fecal samples from the adults were stored at −20°C within 2 h after defecation.

Extraction of DNA from bacterial cultures.

Genomic DNA of Bifidobacterium strains was extracted from 1-ml overnight cultures with the DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany) following the manufacturer's instructions for Gram-positive bacteria. Since genomic DNA of B. longum subsp. longum ATCC 15707^T was used as a standard for quantification of total bifidobacteria by qPCR, high yields of DNA were required and were obtained with the RTP Invitek Bacteria Mini Kit (Invitrogen, Darmstadt, Germany). The manufacturer's protocol for Gram-positive bacteria was followed with some modifications. After enzymatic cell lysis, an additional mechanical lysis was performed in a 2-ml tube containing 0.75 g of sterile zirconium/silica beads (0.1 mm in diameter; Roth, Karlsruhe, Germany) by running the Fastprep FP120 Instrument (Thermo Electron Corp., Waltham, MA) at speed level 4 for 8 min. Genomic DNA of non-Bifidobacterium strains was extracted with the RTP Invitek Bacteria Mini Kit (Invitrogen, Darmstadt, Germany) following the manufacturer's instructions.

Extraction of bacterial DNA from fecal samples.

Bacterial DNA was extracted from frozen fecal material using the QIAamp DNA stool kit (Qiagen, Hilden, Germany) with a modified protocol for cell lysis. After homogenization of 220 mg feces with 1.2 ml lysis buffer from the kit by vortexing for 2 min in a 2-ml tube containing 0.75 g of sterile zirconium/silica beads (0.1 mm in diameter; Roth, Karlsruhe, Germany), the final suspension was incubated at 95°C for 15 min with continuous shaking (1,400 min⁻¹; Thermomixer 5436; Eppendorf, Hamburg, Germany). The sample was allowed to cool on ice for 2 min. Cells were mechanically lysed by Fastprep treatment for 8 min and 15 s as described above. After cooling on ice for 2 min, coarse particles, cell debris, and the zirconium/silica beads were spun down by centrifugation (20,000 × g; 4°C; 1 min), and the supernatant was transferred to a 2-ml tube. The pellet was mixed with 350 μl lysis buffer from the kit, vortexed for 1 min, and incubated at 95°C for 5 min with continuous shaking as described above. After centrifugation at 20,000 × g and 4°C for 1 min, the supernatants were combined. DNA-damaging substances and PCR inhibitors present in the stool samples were removed by their adsorption to the InhibitEX matrix provided in the kit. The InhibitEX matrix was separated by centrifugation at 20,000 × g for 6 min, and the supernatant was collected and filled up to 1 ml with sterile phosphate-buffered saline (pH 7). DNA was purified with the QIAcube (Qiagen, Hilden, Germany) and eluted from the silica-based membrane with 200 μl ultrapure water.

groEL sequence analysis and design of species-specific primers.

groEL sequences belonging to target Bifidobacterium species (Table 2) and available in GenBank release 185.0 (4) were subjected to a multiple alignment with the program ClustalW2 version 2.1 (24) provided by the European Bioinformatics Institute (http://www.ebi.ac.uk). A region of approximately 600 bp located at positions ∼250 to 840 of the complete groEL gene of ca. 1,600 bp was selected to identify discriminative target sites for the species-specific detection of bifidobacteria. Primer pairs for 12 Bifidobacterium species (Table 3) were manually designed on the basis of 97 partial and complete groEL sequences (Table 2).

Table 2.

groEL sequences used for the design of Bifidobacterium species-specific primers

Bifidobacterium species	Strain	GenBank accession no.	Size (bp)
B. adolescentis	JCM 1275T	AF210319	1,749a
	ATCC 15703T	AP009256	1,617a
B. angulatum	JCM 7096T	AF240568	590b
B. animalis	B83	AY166539	590b
	D1	AY166540	590b
	III-1	AY166541	590b
	III-6	AY166542	590b
subsp. animalis	JCM 1190T	AY004273	590b
	ATCC 25527T	AY488178	1,155b
	TEEV 1/10	HQ851043	587b
	TEEV 3/10	HQ851044	598b
	TEEV 1/11	HQ851047	586b
	TEEV 4/9 AG	HQ851048	608b
subsp. lactis	HN019	ABOT01000006	1,614a
	9952	AF286735	590b
	DSM 10140T	AY004282	1,614a
	9950	AY004287	590b
	JB-1	AY166543	590b
	NCC 363	AY488176	1,155b
	NCC 402	AY488177	1,155b
	DSM 27674	AY488179	1,155b
	DSM 10140T	AY488180	1,155b
	ATCC 27536	AY488181	1,155b
	NCC 239	AY488182	1,155b
	ATCC 27672	AY488183	1,155b
	LMG 18906	AY586539	1,614a
	AD011	CP001213	1,614a
	Bl-04 ATCC SD5219	CP001515	1,614a
	DSM 10140T	CP001606	1,614a
	BB-12	CP001853	1,614a
	V9	CP001892	1,614a
	CNCM I-2494	CP002915	1,614a
	BLC1	CP003039	1,614a
B. bifidum	JCM 1255T	AY004280	590b
	PRL2010	CP001840	1,626a
	S17	CP002220	1,626a
B. breve	DSM 20213T	ACCG02000012	1626a
	DSM 20213T	AF240566	584b
B. catenulatum	JCM 1194T	AY004272	590b
	JCM 7130	AY166565	590b
B. dentium	ATCC 27679	AEEQ01000009	1,617a
	JCVIHMP022	AEHJ01000021	1,617a
	JCM 1195T	AF240572	590b
B. gallicum	DSM 20093T	ABXB03000002	1,611a
	JCM 8224T	AF240575	592b
B. longum	N10	FJ577573	552b
	N103	FJ577576	552b
	N81	FJ577625	552b
	N86	FJ577629	552b
	N88	FJ577630	564b
subsp. infantis	JCM 1210	AF240576	591b
	JCM 1222T	AF240577	591b
	ATCC 15697T	AP010889	1,626a
	157F	AP010890	1,626a
	9901	AY004286	591b
	bb52	AY004288	591b
	6w-50	AY166564	590b
	D6-9-6	AY166566	590b
	D30-3-3	AY166567	590b
	D30-3-10	AY166568	590b
	D30-3-11	AY166569	590b
	L2	AY166570	590b
subsp. longum	ATCC 15697T	CP001095	1,626a
	ATCC 55813	ACHI01000023	1,626a
	NCC 2705	AE014295	1,626a
	JCM 1217T	AF240578	590b
	L4	AY166571	590b
	TJ	AY166572	593b
	JCM 7052	AY166573	590b
	JCM 7053	AY166574	590b
	ATCC 15707T	AY835622	552b
	DJO10A	CP000605	1,626a
	JDM 301	CP002010	1,626a
	BBMN 68	CP002286	1,626a
	KACC 91563	CP002794	1,626a
	F8	FP929034	1,626a
	R0175	HM009034	1,626a
subsp. suis	JCM 1269T	AY013248	591b
	JCM 7139	AY166575	590b
	TEEV 4/9	HQ851041	590b
B. pseudocatenulatum	DSM 20438T	AY004274	589b
	JCM 7040	AY166552	590b
	T1-3-15	AY166553	590b
	T2-3-4	AY166554	590b
	Z2-3	AY166555	590b
B. pseudolongum subsp. globosum	JCM 5820T	AF286736	591b
	1-23-3	AY166544	590b
	1-25-3	AY166545	590b
	02-2	AY166546	590b
	7#-3-7	AY166547	590b
	7#-10-11	AY166548	590b
	09-25-2	AY166549	590b
	Fb9	AY166550	590b
	MU8	AY166551	590b
subsp. pseudolongum	JCM 1205T	AF240573	591b
B. thermophilum	JCM 1207T	AF240567	591b
	RBL67	DQ340558	590b

^aComplete coding sequence.

^bPartial coding sequence.

Table 3.

Bifidobacterium species-specific primers based on the groEL gene^a

Target species	Primerb	Sequence (5′ to 3′)c	Primer concn (nM)	Annealing temp (°C)	Amplicon
					Size (bp)	Melting temp (°C)
B. adolescentis	B_ado-f	CTCCGCCGCTGATCCGGAAGTCG	300	75	268	88.0
	B_ado-r	AACCAACTCGGCGATGTGGACGACA
B. angulatum	B_ang-f	CTGTCCTCCCAGCAGGACGTGGTC	300	80	97	85.5
	B_ang-r	GCGCTTCGCCGTCAACGTCTTCGG
B. animalis	B_ani-f	CACCAATGCGGAAGACCAG	250	64	184	87.5
	B_ani-r	GTTGTTGAGAATCAGCGTGG
B. bifidum	B_bif-f	CTCCGCAGCCGACCCCGAGGTT	300	64	233	88.5
	B_bif-r	TGGAAACCTTGCCGGAGGTCAGG
B. breve	B_bre-f	GCTCGTCGTTGCCGCCAAGGACGTT	300	72	272	88.0
	B_bre-r	ACAGAATGTACGGATCCTCGAGCACG
B. catenulatum	B_cat-f	GGCTATCGTCAAGGAGCTCA	300	64	188	87.0
	B_cat-r	AGTCCAGATCCAAACCGAAAC
B. dentium	B_den-f	GGCCCAGTCTTTGGTGCATGAAGGCC	300	75	364	89.5
	B_den-r	GTCTTCGAGCACCGCGGTCTGGTCC
B. gallicum	B_gal-f	AGCTCGTCAAGTCCGCCAAGC	300	68	188	87.5
	B_gal-r	CATACCTTCGGTGAACTCGAGG
B. longum	B_lon-f	CGGCGTYGTGACCGTTGAAGAC	250	70	259	88.0
	B_lon-r	TGYTTCGCCRTCGACGTCCTCA
B. pseudocatenulatum	B_pcat-f	AGCCATCGTCAAGGAGCTTATCGCAG	250	68	325	87.5
	B_pcat-r	CACGACGTCCTGCTGAGAGCTCAC
B. pseudolongum	B_plon-f	CRATYGTCAAGGAACTYGTGGCCT	300	66	312	89.0
	B_plon-r	GCTGCGAMGAKACCTTGCCGCT
B. thermophilum	B_the-f	ACTGGTCGCTTCCGCCAAGGATG	300	66	326	88.0
	B_the-r	CCARGTCAGCMAGGTGRACGATG

^aThe target sites are located at positions ∼250 to 840 of the corresponding nucleotide sequence.

^bf, Forward primer; r, reverse primer.

^cIUPAC ambiguity codes: M (A or C), K (G or T), R (A or G), Y (C or T).

Quantitative real-time PCR. (i) Standards.

For quantification of the investigated Bifidobacterium species, represented by the strains underlined in Table 1, PCR products containing the target sequence were used as qPCR standards. Therefore, part of the genomic groEL gene was amplified with primers binding to a sequence at least 17 bp up- and downstream of the species-specific primer-binding sites to prevent their degradation during storage. The 50-μl PCR mixture consisted of 1× TopTaq PCR buffer (Qiagen, Hilden, Germany), 200 μM (250 μM for B. pseudocatenulatum) of each deoxynucleoside triphosphate (dNTP) (Invitek, Berlin, Germany), 200 nM each forward and reverse primer (Table 4), 1.25 U of TopTaq DNA polymerase (Qiagen, Hilden, Germany), and 1 μl genomic DNA (30 to 110 ng) of one of the representative strains (Table 1). The reactions were performed in a thermal cycler (Thermo Hybaid MultiBlock System, Ulm, Germany) under the conditions given in Table 4. The amplification products were purified with the High Pure PCR Product Purification Kit (Roche, Mannheim, Germany) following the manufacturer's instructions. The concentration of the amplified DNA (concentration_amplicon [g/liter]) was determined spectrophotometrically at 260 nm (NanoDrop; Peqlab, Erlangen, Germany). The number of PCR products (or partial groEL gene copies) per volume was calculated as follows: (i) concentration_amplicon (g/liter)/molecular mass_amplicon (g/mol) = concentration_amplicon (mol/liter); (ii) concentration_amplicon (mol/liter) × N_A = concentration_amplicon (molecules/liter), where N_A is the Avogadro constant (6.022 × 10²³ molecules/mol). The molecular mass_amplicon values are listed in Table 4.

Table 4.

Primers used to generate PCR products containing Bifidobacterium species-specific primer-binding sites (qPCR standards)^a

Target species	Primerb	Sequence (5′ to 3′)	PCR conditions	Amplicon
				Size (bp)	Molecular mass (g/mol)c
B. adolescentis	B_ado-std-f	GATCAGATCGCTGCCACCGCAAC	94°C, 3 min; 30 times (94°C, 30 s; 69°C, 30 s; 72°C, 30 s); 72°C, 10 min	320	197,799.00
	B_ado-std-r	ATCAGCAGCGGCTTGCCGGTCTT
B. angulatum	B_ang-std-f	AGTGCTCGAAGACCCGTACATCCT	94°C, 3 min; 30 times (94°C, 30 s; 65°C, 30 s; 72°C, 30 s); 72°C, 10 min	164	101,387.60
	B_ang-std-r	GGATGTTGTTCAGGATCAGGGTCG
B. animalis	B_ani-std-f	GTCACCGTCGAAGACAACAACC	94°C, 3 min; 30 times (94°C, 30 s, 65°C, 30 s; 72°C, 30 s); 72°C, 10 min	299	184,812.60
	B_ani-std-r	CAGGACTTGAAGGTGCCACGG
B. bifidum	B_bif-std-f	CAAGGACGTGGAGACCAAG	94°C, 3 min; 30 times (94°C, 30 s; 65°C, 30 s; 72°C, 30 s); 72°C, 10 min	391	241,687.40
	B_bif-std-r	CTTGTTCAGGATGAGGGTCG
B. breve	B_bre-std-f	CCACCGAGGTCATCGTC	94°C, 3 min; 30 times (94°C, 20 s; 58°C, 20 s; 72°C, 20 s); 72°C, 10 min	322	199,032.80
	B_bre-std-r	TGCTGGGAGGAGACCTTG
B. catenulatum	B_cat-std-f	GTCGTGGTATCGAGAAG	94°C, 3 min; 30 times (94°C, 30 s; 47°C, 30 s; 72°C, 30 s); 72°C, 10 min	244	150,819.60
	B_cat-std-r	TAGCCCTTGTCGAAACG
B. dentium	B_den-std-f	CCACTACCGCAACTGTGCTG	94°C, 3 min; 30 times (94°C, 50 s; 62°C, 50 s; 72°C, 50 s); 72°C, 10 min	400	247,244.00
	B_den-std-r	GTCAGGAGGATGTACGGGTC
B. gallicum	B_gal-std-f	TGAGAAGGCTTCCGACG	94°C, 3 min; 30 times (94°C, 30 s; 57°C, 30 s; 72°C, 30 s); 72°C, 10 min	242	149,590.80
	B_gal-std-r	GGGAGATGTAGCCCTTG
B. longum	B_lon-std-f	CTGAGGCTCTGGACAAGGTCG	94°C, 3 min; 30 times (94°C, 30 s; 66°C, 30 s; 72°C, 30 s); 72°C, 10 min	323	199,651.20
	B_lon-std-r	GGTGCCACGGATGTTGTTCAGG
B. pseudocatenulatum	B_pcat-std-f	CGTCGTGGTATCGAGAAGGCTTCCG	94°C, 3 min; 30 times (94°C, 50 s; 64°C, 50 s; 72°C, 50 s); 72°C, 10 min	434	268,244.60
	B_pcat-std-r	AAGGTCGGCAGAGCCTCGCCAT
B. pseudolongum	B_plon-std-f	CCTCAAGAACGTTGTGGC	94°C, 3 min; 30 times (94°C, 30 s; 49°C, 30 s; 72°C, 30 s); 72°C, 10 min	420	259,607.00
	B_plon-std-r	CCGGTCTTCATCACGAG
B. thermophilum	B_the-std-f	GAAGGCCTGAAGAACGTGGTCG	94°C, 3 min; 30 times (94°C, 40 s; 65°C, 40 s; 72°C, 40 s); 72°C, 10 min	449	277,522.60
	B_the-std-r	TCAGCCACGATCAGCAGCGGA

^aThe target sites are located at least 17 bp up- and downstream of the Bifidobacterium species-specific primer-binding sites to prevent their degradation during storage.

^bf, forward primer; r, reverse primer.

^cRefers to the double-stranded amplicon.

The identities of the generated qPCR standards were confirmed by sequencing (Eurofins MWG Operon, Ebersberg, Germany). For the quantification of total bifidobacteria, genomic DNA from B. longum subsp. longum ATCC 15707^T was used as the qPCR standard. After extracting genomic DNA, its concentration was determined spectrophotometrically as described above. Assuming a mean genome size for Bifidobacterium of 2.25 Mb (GenBank [4]), the average mass of a single Bifidobacterium genome would be 2.47 fg: [2.25 Mb × average molecular mass_{base pair} (660 g/mol)]/N_A = 2.47 fg/molecule. The number of bifidobacterial genomes per volume was calculated as follows: concentration_{genomic DNA} (g/liter)/(2.47 fg/molecule) = concentration_{genomic DNA} (molecules/liter). The standards were stored in 10 mM Tris-HCl (pH 8) at −20°C and subjected to 10-fold serial dilutions in ultrapure water prior to each measurement.

(ii) Conditions.

qPCR based on the fluorescent dye SYBR green I was performed with 7500 Fast Real-Time PCR Systems using the 7500 software v2.0.5 (Life Technologies, Darmstadt, Germany). Reactions were measured in triplicate. Each reaction mixture of 25 μl contained 1× QuantiFast SYBR green PCR Master Mix (HotStarTaq Plus DNA polymerase, QuantiFast SYBR green PCR buffer, dNTP mixture, and the ROX passive reference dye; Qiagen, Hilden, Germany), 250 to 300 nM each forward and reverse primer (Table 3), and 1 μl template DNA. Bifidobacterium species were quantified with the following temperature program (Table 3): 95°C for 5 min, 40 cycles at 94°C for 15 s, 64 to 80°C for 15 s, 72°C for 15 s, and 83°C for 15 s. The fluorescence of SYBR green I was measured after each amplification cycle and during the last step at 83°C to get rid of nonspecific fluorescence signals caused by primer dimers or nonspecific minor products (36). Total bifidobacteria were quantified using the 16S rRNA gene-targeting primer pair g-Bifid-F (5′-CTC CTG GAA ACG GGT GG-3′)/g-Bifid-R (5′-GGT GTT CTT CCC GAT ATC TAC A-3′) with the corresponding temperature program as described elsewhere (29). Postamplification melting-curve analysis was performed by slowly increasing the temperature from 68°C to 95°C (increments of 1%, holding for 10 s), while fluorescence was measured continuously. Thus, the specific PCR product was verified based on the specific melting temperature (Table 3) and distinguished from possible nonspecific products. Background fluorescence was calculated during the initial stage of the qPCR corresponding to cycles 3 to 10. The threshold was set within the exponential phase of the amplification plot at 0.01 (single Bifidobacterium species) or 0.05 (total bifidobacteria) relative fluorescence. Threshold cycles (C_T) (the PCR cycle numbers at which the fluorescence exceeds the threshold above the calculated background) of less than 11 and more than 33 were excluded from the analysis. The number of groEL copies or bifidobacterial genomes in a fecal sample was calculated using a standard curve, which was generated by plotting the C_T values obtained for 10-fold serial dilutions of the qPCR standards as a linear function of the base 10 logarithm of known concentrations. All primers used in this study were commercially synthesized by Eurofins MWG Operon (Ebersberg, Germany).

Spiking experiments.

Fresh fecal samples were collected from two healthy human adults initially tested for the absence of the investigated Bifidobacterium species by the developed qPCR assays. Aliquots of 220 mg fecal material were spiked with 10-fold serial dilutions of known amounts of the representative strain of each Bifidobacterium species (Table 1) ranging from 4 to 9 log₁₀ cells. The concentration of bacterial cells was estimated microscopically by using the Thoma-Zeiss counting chamber (chamber depth, 0.01 mm; small square, 0.0025 mm²). Microscopic counts were determined at least in duplicate. Bacterial DNA from spiked feces was extracted as described above.

Statistics.

Data analysis of the comparison of cell numbers determined with qPCR assays and the Thoma-Zeiss counting chamber was done with GraphPad Prism version 5.0 (San Diego, CA). Significance was tested by one-way analysis of variance (ANOVA) with Bonferroni's posttests. A P value of <0.05 was considered significant.

RESULTS

Specificities of primer pairs.

The properties and specificity of the primer pair for each Bifidobacterium species were checked in silico with OligoAnalyzer 3.1, provided by Integrated DNA Technologies (35), and the Basic Local Alignment Tool (1), provided by the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/). Each primer pair perfectly matched the groEL sequence of the targeted Bifidobacterium species but had several mismatches with those of the nontargeted Bifidobacterium species investigated in the study (Table 1). In the case of B. thermophilum, the designed primer pairs also matched the groEL of B. thermacidophilum, which, however, is not found in the human intestinal tract (7) and was therefore disregarded. The specificity of each primer pair was experimentally tested by performing qPCR on genomic DNA extracted from 28 Bifidobacterium strains and 10 non-Bifidobacterium strains representing dominant intestinal bacterial species (Table 1). At the appropriate annealing temperature and the chosen primer concentrations, the primer pairs were specific for their respective target species and yielded an amplicon of the expected size (data not shown). Application of the primer pairs to nontargeted bacterial species tested in this study never led to an amplification product. This conclusion was supported by the following observations: C_T values obtained for nontarget organisms were close to the C_T values of the no-template control, and there was no peak at the melting temperature of the specific amplicon as determined by postamplification melting-curve analysis. However, application of primer pair B_bif-f/B_bif-r to genomic DNA of B. breve DSM 20091 led to a PCR product that, based on the melting-curve analysis, could clearly be identified as nonspecific.

Bifidobacterium species qPCR assays.

For all Bifidobacterium species represented by the strains underlined in Table 1, 10-fold serial dilutions of qPCR standards (PCR products containing the respective partial groEL sequence) were used to generate a standard curve. The standard curves for all Bifidobacterium species were highly linear (R² > 0.99), at least in the range of 10 to 100,000 groEL copies per PCR, corresponding to 6 to 10 log₁₀ groEL copies/g feces, except for B. angulatum and B. pseudolongum, for which linearity started at 100 groEL copies per PCR. The PCR efficiency (E) was calculated by the slope of the standard curve as follows: E = 10^(−1/slope) − 1 (e.g., E = 1, or 100%). For B. adolescentis, B. animalis, B. bifidum, B. breve, B. catenulatum, B. dentium, B. gallicum, and B. pseudocatenulatum, E was at least 90%. For B. angulatum, B. longum, B. pseudolongum, and B. thermophilum, E was somewhat lower but at least 85%.

Reproducibility.

The reproducibility of the qPCR assays was determined for DNA extracts from three adult fecal samples spiked with different cell numbers of the representative strain of each Bifidobacterium species (Table 1) in two replicate runs. The resulting coefficients of variation (CV) based on the groEL copy number were less than 10%, indicating that the developed qPCR assays are precise. The reproducibility between different DNA extracts was determined by extracting DNA three times from adult or infant fecal samples and quantifying B. breve, B. longum, and total bifidobacteria. The CV based on groEL copies varied less than 23%, demonstrating that the DNA extraction method was highly reproducible.

Quantification limit.

To determine the lowest groEL copy number in feces detectable with the developed qPCR assays (quantification limit), adult fecal samples were spiked with the representative strain of each of the Bifidobacterium species under study. The quantification limit in feces was 1 to 10 groEL copies per PCR, corresponding to 5 to 6 log₁₀ groEL copies/g feces (Table 5).

Table 5.

Quantification of Bifidobacterium species in feces of human adults and infants

Bifidobacterium species	Log10groEL copies/g feces in samplea:
	Adults (n = 10)										Infants (n = 5)
	AD-1	AD-2	AD-3	AD-4	AD-5	AD-6	AD-7	AD-8	AD-9	AD-10	INF-1 (wk 1)	INF-2 (wk 1)	INF-3 (wk 6)	INF-4 (mo 3)	INF-5
															wk 6	mo 3
B. adolescentis	9.7	9.7	−	9.9	9.8	−	8.9	9.3	9.3	−	−	−	−	−	−	−
B. angulatum	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−
B. animalis	−	7.0	−	−	−	−	−	−	−	−	−	−	−	−	−	−
B. bifidum	8.2	8.6	−	9.1	−	−	−	−	−	8.6	−	−	9.7	−	−	−
B. breve	−	−	−	−	−	−	−	−	−	−	−	6.2	10.4	10.1	−	8.3
B. catenulatum	−	8.1	−	−	8.1	−	−	−	−	−	−	−	−	−	−	−
B. dentium	−	6.4	6.7	−	7.1	−	−	−	−	−	−	−	8.8	−	−	−
B. gallicum	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−
B. longum	7.2	9.1	−	8.9	9.2	9.3	6.6	9.1	−	9.0	−	6.4	−	10.1	−	10.0
B. pseudocatenulatum	−	−	9.3	−	8.9	−	−	8.7	−	−	−	−	−	−	−	−
B. pseudolongum	−	−	8.0	−	−	−	−	−	−	−	−	−	−	−	−	−
B. thermophilum	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−	−
Sum of species	9.8	9.9	9.3	10.0	10.0	9.3	8.9	9.6	9.3	9.1	−	6.6	10.5	10.4	−	10.1
Total bifidobacteriab (log10 bifidobacterial genomes/g feces)	10.3	10.0	10.1	10.5	10.7	10.0	9.5	10.3	9.8	9.8	−	7.6	11.1	11.0	8.1	11.1

^aInfants were around 1 week (wk 1), 6 weeks (wk 6), and 3 months (mo 3) old; adults were 20 to 40 years old. −, not quantified (the quantification limits were 5 log₁₀ groEL copies/g feces for B. adolescentis, B. breve, B. catenulatum, B. dentium, B. longum, B. pseudocatenulatum, and B. thermophilum and 6 log₁₀ groEL copies/g feces for B. angulatum, B. animalis, B. bifidum, B. gallicum, and B. pseudolongum).

^bThe total number of bifidobacteria was determined with the 16S rRNA gene-targeting primer pair g-Bifid-F/R by Matsuki et al. (29). The quantification limit was reported to be 6 log₁₀ cells/g feces.

Comparison of cell numbers determined with qPCR assays and the Thoma-Zeiss counting chamber.

Fecal samples were spiked with cells of six Bifidobacterium species (B. adolescentis, B. angulatum, B. animalis, B. breve, B. pseudolongum, and B. thermophilum). The number of groEL genes measured by the developed qPCR assays was compared with the number of bifidobacteria added to the fecal sample and determined with the Thoma-Zeiss counting chamber prior to spiking. Independent of the Bifidobacterium species, the proportion of groEL copies recovered from fecal samples spiked with 5 to 9 log₁₀ cells/g was 45 to 48% (Fig. 1A and B). These results indicated that the cell numbers of a given Bifidobacterium species present in a fecal sample and determined with the developed qPCR assays were comparable to those obtained by microscopy, suggesting sufficiently effective cell lysis. The recovery of groEL copies was considered in all calculations.

Fig 1.

Comparison of cell numbers determined with qPCR assays and the Thoma-Zeiss counting chamber using human adult feces spiked with Bifidobacterium species cells. (A) Spiking was performed with 6 log₁₀ (▲), 7 log₁₀ (●), and 8 log₁₀ (■) cells of six Bifidobacterium species. (B) Spiking was performed with 5 log₁₀ (n = 3), 6 log₁₀ (n = 6), 7 log₁₀ (n = 6), 8 log₁₀ (n = 6), and 9 log₁₀ (n = 5) cells of B. adolescentis (⊙), B. angulatum (⊡), B. animalis (), B. breve (♢), B. pseudolongum (▽), and B. thermophilum (○). The values are means. Significance was estimated by one-way ANOVA with Bonferroni's posttests. Differences of means were not significant. GroEL copy recovery was 48% ± 6% (A) and 45% ± 6% (B) (means ± SD).

Presence of PCR inhibitors.

To verify that PCR inhibitors present in feces had been completely removed during DNA extraction, B. longum was quantified with the developed qPCR assay using DNA obtained from four fecal samples (INF-4, INF-5 month 3, AD-5, and AD-6) at dilutions of 0 to −3 log₁₀. Since the cell numbers (groEL copies/g feces) calculated for a given sample were nearly the same for all four dilutions (10.09 ± 0.03 in INF-4, 9.86 ± 0.12 in INF-5 month 3, 9.17 ± 0.02 in AD-5, and 9.28 ± 0.05 in AD-6 [mean ± standard deviation {SD} log₁₀; n = 4]), it was concluded that PCR inhibitors had been completely removed.

Quantification of Bifidobacterium species in fecal samples.

The designed and validated primers were subsequently applied to fecal samples from five healthy, full-term, breast-fed infants and 10 healthy adults (Table 5). Fecal samples from breast-fed infants contained no more than three Bifidobacterium species. B. breve was not detected in adult feces but was found in 4/6 infant samples at up to 10.4 log₁₀ groEL copies/g feces. B. bifidum and B. longum were present at high concentrations (9.7 log₁₀ groEL copies/g feces in INF-3 and 10.1 log₁₀ groEL copies/g feces in INF-4, respectively). B. adolescentis, which is a typical Bifidobacterium species of the adult fecal microbiota, was not detected in infant feces. Most of the adult fecal samples contained between one and three Bifidobacterium species (one species, 2/10; two species, 2/10; three species, 4/10; five species, 1/10; and six species, 1/10). B. adolescentis and B. longum were frequently found in adults (7/10 and 8/10, respectively). Their concentrations were relatively high, with 8.9 to 9.9 log₁₀ groEL copies/g feces for all but two subjects (AD-1 and AD-7), who harbored B. longum at much lower concentrations (7.2 log₁₀ and 6.6 log₁₀ groEL copies/g, respectively). B. pseudocatenulatum, even though detected in just 3/10 samples, also displayed high counts of 8.7 to 9.3 log₁₀ groEL copies/g feces. In adult feces, the concentration of B. bifidum, a predominant Bifidobacterium species in the feces of infants, was in the range of 8.2 to 9.1 log₁₀ groEL copies/g feces. B. catenulatum and B. pseudolongum were found rarely (2/10 and 1/10, respectively) at 8.0 to 8.1 log₁₀ groEL copies/g feces. The lowest numbers, ranging from 6.4 to 7.1 log₁₀ groEL copies/g feces, were measured for B. dentium and B. animalis. The latter was detected only once (AD-2). In adults, total bifidobacteria varied between 9.5 and 10.7 log₁₀ bifidobacterial genomes/g feces. In contrast, 1-week-old infants harbored relatively low levels of total bifidobacteria (7.6 log₁₀ bifidobacterial genomes/g feces in INF-2) or even none (INF-1). However, bifidobacteria were detected in higher numbers in 3-month-old infants than in adults (11.0 log₁₀ bifidobacterial genomes/g feces in INF-4 and 11.1 log₁₀ bifidobacterial genomes/g feces in INF-5). The sum of groEL copies of all targeted Bifidobacterium species was slightly lower than the number obtained with the genus-specific primer pair g-Bifid-F/R targeting the 16S rRNA gene (0.1 to 1.0 log₁₀). The only exception was the sample collected from donor INF-5 at 6 weeks of age, in which none of the investigated Bifidobacterium species could be detected but the concentration of total bifidobacteria was 8.1 log₁₀ bifidobacterial genomes/g feces. B. angulatum, B. gallicum, and B. thermophilum were not detected in any fecal sample from adults or infants.

DISCUSSION

The groEL gene is a powerful phylogenetic marker.

The phylogenetic tree based on fragments of the groEL gene (538 bp, located at positions ∼250 to 840 of the corresponding nucleotide sequence) or larger (∼1,200-bp) or even complete (∼1,600-bp) groEL sequences of Bifidobacterium species (27, 47, 50) was similar to the one based on 16S rRNA gene sequences (31). However, the genetic distances of the groEL sequences were greater than those of the 16S rRNA gene sequences (21, 55). The partial groEL sequence (∼550 bp) identity among different Bifidobacterium species was 80 to 96%, providing considerable discriminative power at the interspecies level (55). The analysis of larger (∼1,200-bp) and complete (∼1,600-bp) groEL sequences even allowed the differentiation of B. animalis and B. longum at the subspecies level (27, 50). The ubiquitous GroEL-encoding gene nevertheless reveals a high degree of sequence conservation, in agreement with the central cellular function of the encoded chaperonin. All in all, the groEL gene fulfills several prerequisites to serve as a reliable alternative phylogenetic marker for Bifidobacterium species.

Advantages of the groEL gene compared to the 16S rRNA gene.

Although the 16S rRNA gene, with a mean sequence identity of 95% (31, 47), allows the discrimination of most Bifidobacterium species, its discriminative power among closely related species is limited. Thus, it is not possible to distinguish B. catenulatum and B. pseudocatenulatum, which share 99.5% sequence identity in their 16S rRNA genes (31, 47). These species have therefore been detected together (16, 29, 33). As demonstrated here, B. catenulatum and B. pseudocatenulatum could easily be distinguished because the two organisms share only 91 to 93% sequence identity in the groEL sequence of ∼550 bp (21, 55). Owing to more than 96% 16S rRNA gene sequence identity, it has not been possible to determine the evolutionary distance between strains belonging to B. breve and B. longum (25). This also applies to B. adolescentis/B. dentium, B. breve/B. pseudolongum, B. catenulatum/B. dentium, and B. longum/B. pseudolongum (25). In contrast, the selected region of the groEL gene used in this study revealed a considerably higher resolution power between these species than the 16S rRNA gene (21, 47, 55).

Limitations.

Initially, phylogenetic analysis of bifidobacteria based on the groEL gene was constrained because the number of available groEL sequences was limited to 73 (21, 55). With increasing interest in the gene as a powerful phylogenetic marker and as a result of the sequencing of bifidobacterial genomes, the database of groEL sequences has continuously grown (18). Although groEL sequences are available for almost all 36 described Bifidobacterium species (http://www.dsmz.de), the majority of these sequences are incomplete; they consist of an approximately 600-bp region located at positions ∼250 to 840 of the gene. For some Bifidobacterium species, only one or two sequences are available (e.g., B. angulatum and B. catenulatum, respectively).

Quantification of Bifidobacterium species in fecal samples.

The colonization of the gastrointestinal tract starts at birth and is characterized by different microbial successions. Within the first days of life, facultative anaerobes, such as coliforms and enterococci, dominate the microbiota. They create a reduced environment, enabling the establishment of strict anaerobes, including bifidobacteria, by days 4 to 7 (26).

In our study, at 1 week after birth, Bifidobacterium species were absent or present at low concentrations. In contrast, at 3 months of age, bifidobacteria had advanced to become one of the dominant population groups in the infant gut. Differences in the patterns and sizes of bifidobacterial populations as reported here have been observed previously (19, 30, 38). Our study revealed low Bifidobacterium diversity and the presence of three or less than three species in a given sample from infants or adults (8/10). This result is similar to a previous study with higher numbers of subjects (30).

High cell counts of B. bifidum, B. breve, and B. longum (including B. longum subsp. infantis) were frequently found in feces from breast-fed infants, suggesting that these three species play a crucial role in gut colonization (16, 30, 39). In accordance with previous studies (15, 37, 38), B. longum was the most common species in human adult feces, followed by B. adolescentis, B. bifidum, and B. pseudocatenulatum. A high incidence of B. longum, B. adolescentis, B. bifidum, and members of the B. catenulatum group was observed in Japanese adults (29, 30). Although we were able to distinguish B. catenulatum and B. pseudocatenulatum in our study, the prevalence of both species together was considerably lower than that reported for these Japanese subjects (40% versus 89% [29] and 92% [30], respectively).

As an inhabitant of the oral cavity (7) B. dentium may pass into the intestine, where it is infrequently detected at low cell numbers (29, 30, 39). In contrast to Matsuki et al. (29, 30), who occasionally detected B. breve, the species was mostly below the quantification limit in the adult samples we analyzed. The donor of sample AD-2 was not aware of having consumed a probiotic food prior to the collection of the sample. However, B. animalis subsp. lactis may be present in normal yogurts that are not explicitly labeled as containing bifidobacteria (3).

None of the few fecal samples tested in our study contained B. angulatum, B. gallicum, or B. thermophilum, all of which have rarely been detected. B. angulatum was occasionally found in infant and adult feces (29, 30), in the latter at relatively low concentrations (∼6.6 log₁₀ cells/g feces [29] and 7.5 to 8.3 log₁₀ CFU/g [3]). B. gallicum was mostly absent (15, 30, 33), except for a baby delivered by caesarean section (6). B. thermophilum, first isolated from baby feces (53), was recently detected in the feces of a breast-fed infant at 6.7 log₁₀ cells/g feces (28). Since B. pseudolongum has so far been associated with animals, it has rarely been investigated in humans. However, in agreement with our study, it was recently found in adults, but not in infants (44).

The numbers of Bifidobacterium species that we detected in adults and infants are essentially in agreement with previous findings (30). Matsuki et al. (30) mostly observed between one and four species in adults (29% harbored three and 31% harbored four species) and up to three species in infants (33% harbored one and 22% harbored three species). However, Matsuki et al. (30) were not able to further discriminate between members of the B. catenulatum group.

The total counts of bifidobacterial genomes determined with the genus-specific primers g-Bifid-F/R targeting the 16S rRNA gene were slightly higher than those resulting from the sum of groEL copies detected with the species-specific primers. It is well known that bifidobacteria possess two to five rrn loci (8, 10, 40). Differences in the 16S rRNA gene copy numbers are adequately considered when the qPCR standard curve uses genomic DNA from a bacterium that represents the predominant species in the sample (34). Taking this into account, B. longum subsp. longum ATCC 15707^T, with four 16S rRNA gene copies (13), was used as a standard for quantification of total fecal bifidobacteria in adults and infants. However, Candela et al. (10) have demonstrated that the rrn copies do not differ only at species level. Even different strains of B. adolescentis may harbor between two and five copies (10, 40). The genome size of bifidobacterial species varies between 1.92 and 2.83 Mb (GenBank; corresponding to 2.11 to 3.11 fg/cell). Assuming a mean genome size of 2.25 Mb (2.47 fg/cell) would result in overestimation of the cell number if the predominant Bifidobacterium strain in the sample has a genome size smaller than 2.25 Mb and/or a number of 16S rRNA gene copies higher than four. This would apply to B. adolescentis ATCC 15703^T (GenBank accession no. AP009256), B. bifidum PRL2010, and B. bifidum S17 (43, 56). These problems can be avoided by targeting a single-copy gene, such as groEL. However, a multiple alignment of all complete groEL sequences listed in Table 2 did not reveal consensus regions of sufficient length to design genus-specific primers. Quantification limits of qPCR with species-specific primers that target the single-copy groEL gene are comparable with the one targeting the multicopy 16S rRNA gene (Table 5), indicating that the qPCR assays developed in this study are highly sensitive.

During this study, two novel Bifidobacterium species were isolated from human feces, namely, B. stercoris (22) and B. kashiwanohense (32), for which at least partial groEL sequences are available. It is interesting that recent microbiomic analysis suggested the presence of previously undescribed Bifidobacterium species, which, however, have not been identified so far (45).

Conclusion.

This study describes the use of the groEL gene as a target for the identification and quantification of human fecal Bifidobacterium species by qPCR. The method developed allows the distinction between closely related Bifidobacterium species, enabling monitoring of the bifidobacterial population in response to age, antibiotic treatment, pro- and prebiotic intake, or disease.