Characterisation of intestinal bacteria in infant stools using real-time PCR and northern hybridisation analyses


*Corresponding author. Tel.: +44 1382 632535; fax: +44 1382 633952, E-mail address:


Real-time PCR and northern hybridisations were used to quantify bacterial populations in the large gut of infants. PCR primers for rapid, sensitive, high throughput detection of bifidobacteria, bacteroides, sulphate-reducing bacteria and Enterococcus faecalis, based on analysis of 16S rRNA genes were used. Bacterial populations were analysed in faeces from 40 infants aged 0–6, 7–12 and 13–24 months. The effects of breast versus bottle feeding was also investigated. Real-time PCR indicated that bacteroides and desulfovibrio numbers increased markedly in the 7–12 and 13–24 month age groups, and that the reverse occurred with Ent. faecalis. With the exception of desulfovibrios, this was seen with northern hybridisations, which also showed increased colonisation by the Clostridium coccoides group and Faecalibacterium prausnitzii after 6 months. Both methodologies indicated increased bifidobacteria in breast-fed babies, and higher levels of desulfovibrios in bottle-fed children.


Bacterial colonisation of the infant large intestine is influenced by many factors including diet, environment, antibiotic treatment, and age [1]. Faeces from breast-fed babies have been shown to contain significantly higher numbers of bifidobacteria, and a less complex microbiota than formula-fed infants, although as the microbiota develops during weaning, these differences become less evident [2–5]. The composition of the microbiota in babies has been the subject of a number of investigations, aimed at identifying how breast-feeding confers protection against intestinal diseases [6–8]. Moreover, probiotic studies have shown that dietary supplements containing lactobacilli and bifidobacteria can help in the management of diarrhoea related to rotavirus infection and antibiotic therapy, and also influence the outcome of allergic disease [9–11].

Early studies on the gut microbiota in infants were done using conventional culture methodologies, however, it has been shown that variable regions of 16S rRNA gene sequences provide the basis for development of specific primers or probes that can be used for culture-independent analyses of bacterial populations [12–14]. Real-time PCR assays are increasingly being used for detecting and quantitating bacterial populations that are present in relatively low numbers, which is not possible with northern hybridisations, using the slot blot technique. Their application in many ecosystems has been reported, and this has recently included studies on the large intestine [15–18]. Sensitive, high throughput, and quantitative molecular assays for the detection of these organisms in large scale studies on the gut microbiota of infants would therefore be of considerable value.

The objectives of this study were to quantitate common groups of microorganisms in the faeces of young children, using two different methods of molecular analysis (northern hybridisations, real-time PCR), and to investigate the effects of different types of feeding on bacterial populations.

Bacteroides and bifidobacteria play an important role in carbohydrate metabolism in the large bowel [19,20] and bifidobacteria are also of interest because of their numerical predominance in some infants, as well as their widespread use as probiotics [21].

Facultative anaerobes are early colonisers in the infant large intestine, and in later life, Ent. faecalis is the cause of a significant number of opportunistic infections including bacterial endocarditis, and infections of the urinary tract and surgical wounds [22]. A particular cause for concern is the progressive emergence of antibiotic resistance which places a number of constraints on antibiotic therapy [23]. Colonisation with this organism could have important implications for the dissemination of antibiotic resistance, because enteroccoci exchange genetic information in a promiscuous way in the gut [24].

Dissimilatory sulfate-reducing bacteria (SRB) were investigated in this study because these metabolically specialised organisms form an integral part of the adult colonic microbiota, and may be involved in inflammatory bowel disease (IBD), due to the effects of sulphide toxicity in the gut mucosa [25]. The developmental stage at which these organisms become major colonisers in the large intestine, and factors affecting this process are currently unknown.

2Materials and methods

2.1Subjects and specimens

A single sample of fresh faecal material was collected from 40 children living in the Dundee area and transported to the laboratory within 1 h of defaecation. The average age of the children was 11.5 months (range three weeks to 24 months). The current method of feeding was also noted, whether breast-fed (n= 7), bottle-fed (n= 15) or weaned (n= 18). Upon arrival in the laboratory, stool samples were processed immediately for bacterial isolation, while aliquots were stored at −80 °C for DNA and RNA analyses. Parental consent was obtained for the collection of infant stool samples, which was approved by the Tayside Committee on Medical Research Ethics, Dundee. Subject consent and ethical approval was also granted for the analysis of these samples.

2.2Bacterial strains

DNA extracted from the following strains, obtained from the National Collection of Industrial and Marine Bacteria (Aberdeen, UK), the National Collection of Type Cultures (London, UK), Deutsche Sammlung von Mikroorganismen und Zellkulturen (Braunscheig, Germany), the American Type Culture Collection (Manassas, VA, USA), and laboratory stock cultures at the University of Dundee (DUN-), were used to confirm the specificities of PCR primer sets and oligonucleotide probes used in this study: Aeromonas hydrophilia (DUN-43), Bacteroides fragilis (ATCC 25285), Bacteroides thetaiotaomicron (NCTC 10582), Bacteroides vulgatus (DUN-116), Bifidobacterium adolescentis (NCIMB 702231), Bifidobacterium adolescentis (DSM 20083), Bifidobacterium angulatum (DUN-87), Bifidobacterium longum (NCTC 11818), Clostridium butyricum (NCIMB 7423), Clostridium difficile (NCTC 11223), Clostridium histolyticum (DSM 2158), Clostridium innocuum (DUN-52), Clostridium perfringens (NCIMB 8533), Clostridium septicum (NCTC 282), Clostridium tetani (NCTC 5404), Desulfovibrio desulfuricans (NCIMB 12833), Desulfovibrio desulfuricans (NCIMB 8307), Enterococcus faecalis (ATCC 51299), Enterococcus faecalis (NCDO 611), Enterococcus faecium (NCTC 12202), Escherichia coli (ATCC 11775), Faecalibacterium prausnitzii (NCIMB 13872), Faecalibacterium prausnitzii (ATCC 27768), Lactobacillus acidophilus (DSM 20079), Lactobacillus acidophilus (DUN-28), Methanobrevibacter smithii (DUN-501), Peptostreptococcus anaerobius (DUN-167), Ruminococcus productus (DSM 2950), Salmonella typhimurium (DUN-44), Shigella sonnei (DUN-51), Streptococcus pyogenes (DUN-37), Veillonella parvula (NCDO 1214), Vibrio parahaemolyticus (DUN-59).

Desulfovibrios were grown in Postgate medium B in universal bottles [18]; all other organisms were cultured on Wilkins–Chalgren agar plates or broth and incubated in an anaerobic chamber (in an atmosphere of 80% N2, 10% CO2, and 10% H2) at 37 °C.

2.3DNA extraction from bacterial cultures

DNA was isolated from pure cultures of bacteria using a modified method based on the Qiagen DNA Blood Mini Kit (Qiagen Ltd., West Sussex, England). Bacterial cells were resuspended in 450 μl of deionised water to which 50 μl of lysozyme (50 mg ml−1) was added and incubated at 37 °C for 30 min. Twenty-five microlitres of proteinase K (20 mgml−1), 50 μl of 20% sodium dodecyl sulphate and 0.1 mm glass beads (350 mg) were added prior to mechanical disruption (2 min) using a Mini Beadbeater-8 (Biospec Products Bartlesville, OK, USA). The bead beating step was repeated after 10 min incubation at 60 °C and the cell debris was removed by centrifugation (5000g, 3 min). DNA in the supernatants was then purified using the Qiagen DNA Blood Mini Kit according to the manufacturer’ s instructions.

2.4DNA extraction from faecal material

Fresh faeces were homogenised (10% w/v) in 0.1 M sodium phosphate buffer (pH 6.5). DNA was extracted from 2 ml of the faecal slurry using the Qiagen DNA Stool Kit (Qiagen Ltd., West Sussex, UK). The cells were collected by centrifugation at 20,000g (10 min) and resuspended in 1.4 ml of ASL lysis buffer. Glass beads (350 mg) were added and beaten for 2 min at maximum speed. using a Mini Bead-Beater-8. The suspension was incubated at 95 °C for 5 min, followed by an additional bead-beating step of 2 min. After centrifugation (5000g, 2 min) to remove cell debris, the supernatant was transferred to a clean vial and an Inhibitex (Qiagen) tablet added to remove DNA-damaging substances and PCR inhibitors. The tablet was dissolved with 3 s vigorous agitation, using the bead-beater. DNA was then purified using QIAamp spin columns (Qiagen) as per the manufacturers instructions.

2.5PCR primers

The primers sets used in this study are listed in Table 1. Determining the specificity of each primer set and the optimum annealing temperature was done using a Mastercycler gradient PCR machine (Eppendorf, Hamburg, Germany) and a Techne Genius PCR machine (Duxford, UK). Each PCR mixture (50 μl) contained 5 μl (10 × PCR buffer, Promega, Madison, WI, USA), 1.5 mM MgCl2, 0.25 mM concentrations of deoxynucleoside triphosphates, 0.5 μM primer, 2 μl of bacterial template DNA, and 1 U of Taq DNA polymerase (Promega, Madison, WI, USA). The PCR program consisted of 35 cycles with a DNA denaturation step at 95 °C (1 min), followed by an annealing step (1 min) and elongation step at 72 °C (45 s). The PCR was completed with a final elongation step at 72 °C (10 min). The primers were synthesised commercially by Invitrogen Life Technologies (Paisley, UK).

Table 1.  Target groups and sequences of the PCR primers used in this study
Primer setTarget groupSequence (5′–3′)Annealing temperature (°C)Product size (bp)References
  1. aModified from reference.

Uni 331FBacteriaTCCTACGGGAGGCAGCAGT58466[26]
Bac 708FBacteroides–Porphyromonas–PrevotellaCACGAAGAACTCCGATTG55385[27]
Bif 164FBifidobacteriaGGGTGGTAATGCCGGATG58442[13]
Efs 130FEnterococcus faecalisAACCTACCCATCAGAGGG57360[29]
Dsv 691FDesulfovibrio genusCCGTAGATATCTGGAGGAACATCAG63135[18]

2.6Culturing experiments

To further test the efficiency of real-time PCR, 2 ml of a bacterial suspension (Addition A) of B. fragilis and B. longum (9.8 × 109 and 4.4 × 109, respectively) and 2 ml of a bacterial suspension (Addition B) of Ent. faecalis and Bif. longum (2.4 × 108 and 4.4 × 109, respectively) were each added separately to 2 ml of a 10% faecal slurry and the DNA extracted for analysis. Viable counts from bacterial suspensions and faecal material were made as described previously [30,31]. Bacteria were serially diluted (10-fold) using anaerobic, half-strength peptone water (peptone, 5; NaCl, 2.5 g l−1 distilled water) and aliquots were then spread, in duplicate, onto the following pre-reduced agar plates: Wilkins–Chalgren agar (total bacteria), Wilkins–Chalgren agar with GN antibiotic supplements (bacteroides), and Beerens agar (bifidobacteria) [32] which were incubated in an anaerobic chamber (10% H2, 10% CO2, 80% N2) for 48 h, prior to colonies being counted. Enterococcosal agar plates (Becton Dickinson Diagnostic systems, Sparks, MD, USA), which contain bile, aesculin and azide for the selective isolation of Ent. faecalis were incubated aerobically.

2.7Real-time PCR

Quantitation of bacterial DNA was done using an iCycler Real-Time PCR Detection System (BioRad, Hercules, CA, USA) associated with iCycler Optical System Interface software (version 2.3). Each reaction was done in duplicate in a volume of 20 μl in 96-well optical-grade PCR plates, sealed with optical sealing tape (BioRad). Amplification reactions were done with iQ SYBR Green Supermix (BioRad) containing 3 mM MgCl2, 20 mM Tris–HCl (pH 8.4), 50 mM KCl, each deoxynucleotide triphosphate at a concentration of 200 μM, SYBR Green 1, 10 nM fluorescein, 0.625 U of iTAQ DNA polymerase mixed with the selected primer set at a concentration of 0.5 μM for each primer, and 1.6 μl of the respective DNA. Amplifications were done with the following temperature profiles: one cycle at 95 °C for 3 min, 38 cycles of denaturation at 95 °C (30 s), followed by 30 s at the appropriate annealing temperature, and one final cycle at 95 °C (30 s). Melt curve analyses were done by heating the PCR mixtures from 55 to 95 °C slowly (1 °C per cycle of 10 s), as end point assays to confirm PCR specificity. Quantitation of unknowns was achieved by using standard curves made from known concentrations of plasmid DNA containing the respective amplicon for each set of primers [18].

2.8Northern hybridisations – 16S rRNA analysis

Ribosomal RNA was extracted, blotted, and hybridised using procedures described previously [14]. Bacteria were disrupted mechanically prior to phenol:chloroform extraction. Nucleic acid samples were denatured and diluted before blotting, in triplicate, onto Hybond XL nylon hybridisation membranes (Amersham Pharmacia Biotech Inc., Bucks, England), together with a reference RNA series for normalisation. Table 2 shows details of the oligonucleotide probes used in this study. These were 5′ end-labelled with 32P using {γ32P ATP} and purified using Microspin G-25 columns (Amersham Pharmacia). The membranes were then pre-hybridised for 2 h before labelled probe was added in fresh buffer, and hybridised for a further 16 h. Following this, the membranes were washed for 2 h in SDS/SSC solution, and then twice at the experimentally determined wash temperatures. The hybridisation signals were quantitated, and rRNA abundance expressed as a percentage of total bacterial rRNA.

Table 2.  Target groups and sequences of the oligonucleotide probes used in this study
ProbeTarget groupSequence (5′–3′)TwReferences
Bacto 1080Bacteroides–Porphyromonas–PrevotellaGCACTTAAGCCGACACCT46/50[3]
Lacto 722Lactobacillus–Streptococcus–EnterococcusYCACCGCTACACATGRAGTTCCACT54[34]
Erec 482Clostridium coccoides groupGCTTCTTAGTCARGTACCG47[35]
Enter 1418EnterobacteriaCTTTTGCARCCCACT44[36]
Dsv 1292Desulfovibrio genusCAATCCGGACTGGGACGC43[37]
Bif 1412BifidobacteriaCGGGTGCTRCCCACTTTCATG52[38]
Efs130Enterococcus faecalisCCCTCTGATGGGTAGGTT46[29]
F. prau 0645Faecalibacterium prausnitziiCCTCTGCACTACTCAAGAAAAAC50[39]
Strc 493Streptococcus–LactococcusGTTAGCCGTCCCTTTCTGG54[35]


Unless stated otherwise, all chemicals were obtained from Sigma (Poole, Dorset, UK). Bacteriologic culture media were purchased from Oxoid (Basingstoke, Hamps, UK).


Copy numbers of 16S rRNA genes per mg of sample were transformed into logarithms and the non-parametric Kruskal–Wallis and Mann–Whitney U tests were used for pair wise comparison of the different infant groups. Probability values ≤0.05 were taken as statistically significant.


For real-time PCR analysis of faecal material, four primers for the groups of bacterial 16S rRNA of interest were used, in conjunction with an assay for total bacteria. All primer sets were highly specific and gave positive results with only target bacteria when tested against a range of intestinal isolates. Melt curve analysis demonstrated that when used with the stool extraction procedure, these assays could detect as low as 10 copies of specific bacterial 16S rRNA genes per gram wet weight faeces, in the presence of large amounts of non-specific background DNA. Conversion of bacterial 16S rRNA genes to the number of bacterial cells in a sample was done by dividing gene numbers by the values that represent the average rRNA gene copy number for a cell belonging to that particular bacterial group, according to their respective numbers of operons recorded in the Ribosomal RNA Operon Copy Number Database ( Comparison of bacterial cell numbers obtained by culture and real-time PCR showed that with the exception of Ent. faecalis, the numbers of cells detected by PCR were greater than the values obtained by cell culture (Table 3).

Table 3.  Comparison of bacterial populations in faecal material determined by culture and real-time PCR
 Log10 cells (gram faeces)−1 determined by:
 CultureReal-time PCR
  1. aNot analysed.

Addition A
Total bacteria9.7, 9.510.6, 10.4
Bacteroides group9.4, 9.810.4, 10.3
Bifidobacterium genus8.9, 9.310.0, 10.4
Enterococcus faecalisNAa4.2, 4.1
Addition B
Total bacteria9.7, 9.510.1, 9.9
Bacteroides groupNAa6.1, 6.4
Bifidobacterium genus8.9, 9.29.7, 10.0
Enterococcus faecalis9.2, 9.08.7, 8.8

3.1Age related changes in faecal microbiotas in young children

Analysis of infant stools by real-time PCR showed that total numbers of bacteria were relatively constant in the 0–6, 7–12 and 13–24 month age groups, as were bifidobacterial numbers (Fig. 1). Organisms belonging to the bacteroides-group had reached very high numbers between the ages 7 and 12 months, where values were on average about one 1000-fold higher than in children aged 0–6 months. Desulfovibrios were detected at all ages, and showed a similar colonisation pattern to the bacteroides group, increasing by 200-fold after 6 months of age. Conversely, the amount of Ent. faecalis DNA was markedly higher in the youngest infants, decreasing to average values of less than 10 copies per gram of faeces in the 7–12 month age group. Numbers of these bacteria increased in children aged 13–24 months, but this was not significant.

Figure 1.

Analysis of faecal bacterial populations in infants using real-time PCR. Black bars, 0–6 months old (n= 8); open bars, 7–12 months old (n= 5); grey bars, 13–24 months old (n= 9). Results are mean values ± standard error of the mean. *Significant difference P < 0.05.

Compared to real-time PCR, a wider range of 16S rRNA probes were used for analysing faecal bacterial populations using northern hybridisations. Fig. 2 shows that amounts of bacterial RNA in infant faeces varied considerably with age. Stools from babies aged 0–6 months contained significantly lower proportions of C. coccoides-group RNA, which increased dramatically in the 7–12 month age group (P= 0.018). Faecalibacterium prausnitzii was below the level of detection of the assay in 0–6 months stools, but gradually increased with age. Conversely, bifidobacterial RNA decreased as the children got older. Large inter-individual variations were observed in bacteroides-group RNA, and no significant difference was detected with increasing age. The proportion of RNA from facultatively anaerobic organisms was greatest in the 0–6 month group, especially enterobacteria. These organisms and Ent. faecalis accounted for 13.1% and 9.3%, respectively, of total bacterial RNA in stools of 0–6 month babies, which dropped to 0.8% and 2.0% between 7 and 12 months of age. Desulfovibrio RNA was detected in all age groups, especially in the 0–6 month age group.

Figure 2.

Analysis of faecal bacterial populations in infants using northern hybridisations. Black bars, 0–6 months old (n= 9); open bars, 7–12 months old (n= 6); grey bars, 13–24 months old (n= 11). LAB, lactic acid bacteria. *Significant difference P < 0.05.

3.2Effect of breast-feeding on infant faecal microbiotas

Data from the 0–6 month age group was further sub-divided, based on whether the children had been breast- or bottle-fed (Table 4). The proportion of bifidobacterial RNA in stools of breast-fed babies was twice that of bottle-fed children (8.6% and 4.4%, respectively), whereas the latter cohort had higher concentrations of bacteroides (6.8% and 20.9%, respectively), C. coccoides-group and enterobacterial RNA. Bifidobacterial populations were also higher in the breast-fed group when assessed by real-time PCR (log10 9.1, compared to log10 7.9 per gram faeces), while lower concentrations of bacteroides were found in the bottle-fed group. However, none of the results between these two groups were found to be statistically significant.

Table 4.  Real-time PCR and northern hybridisation analysis of faecal bacterial populations in breast and bottle-fed infantsa
Primers/probe targetsPercentage of total RNALog1016S rRNA genesb
  1. aFaeces were obtained from 11 children aged 0–6 months, of which five were breast-fed and six were bottle-fed.

  2. bResults are mean values per gram of faeces ± standard error of the mean.

  3. cValues in parenthesis are the percentage of the total population based on the number of 16S rRNA genes.

  4. dNot analysed.

  5. eNot detected.

Total bacteria1001009.8 ± 0.4 (100)c9.9 ± 0.2 (100)
Bifidobacterium genus8.6 ± 1.24.4 ± 3.09.1 ± 0.7 (19.8)7.9 ± 0.6 (1.0)
Bacteroides group6.8 ± 0.720.9 ± 11.87.5 ± 0.9 (0.5)5.6 ± 1.3 (<0.1)
Enterococcus faecalis5.2 ± 4.79.8 ± 7.47.2 ± 0.4 (0.25)5.4 ± 1.2 (<0.1)
Desulfovibrio genus0.5 ± 0.10.7 ± 0.13.7 ± 1.3 (<0.1)4.5 ± 0.1 (<0.1)
Lactic acid bacteria2.3 ± 0.61.3 ± 0.4d
Clostridium coccoides group5.0 ± 3.411.2 ± 5.3
Enterobacteria6.5 ± 3.216.5 ± 6.5
Faecalibacterium prausnitziiNDeND
Streptococci1.4 ± 0.81.1 ± 0.6


Using real-time PCR, the bacteroides-group, bifidobacteria and Ent. faecalis DNA could be detected against a background of nucleic acid from faecal material (Table 3). Previous studies have shown that viable counting can underestimate bacterial population sizes when compared to other methods of analysis [40], and results in Table 3 show that this occurred when assessing total bacterial counts and populations of bacteroides and bifidobacteria by real-time PCR. Using the total bacterial real-time PCR assay, a greater population size was also found in carious dentine compared to viable counts, although the extraction procedures and PCR platforms used differed from those in the present work [26]. These differences also highlight the difficulties in comparing real-time PCR and culture, since the average number of rrna operons is used to convert gene copy numbers to viable counts, and varies with the different species present in the faecal sample [41].

Studies involving viable counting have shown that one of the major events occurring during bacterial succession in the infant colon is the gradual replacement of facultative anaerobes, which are the primary colonisers, with strict anaerobes [2,5]. Results obtained in this investigation support these observations, since the proportion of rRNA specific to facultative anaerobes (enterobacteria, streptococci, lactic acid bacteria, Ent. faecalis) decreased after 6 months of age, with the Ent. faecalis results being verified by the real-time PCR method.

Faecalibacterium prausnitzii, previously designated Fusobacterium prausnitzii is a numerically important species in adults [42], and did not become a major constituent of the colonic microbiota until after 6 months of age. Previous work using the F. prausnitzii probe showed that this organism accounted for approximately 5% of total bacterial RNA in adult faeces [39], and similar values were detected in the 13–24 month old age group in this investigation. Other strict anaerobes that became more dominant with age were members of the C. coccoides group, although these organisms stabilised earlier, with adult-like levels of specific RNA being detected in the 7–12 month old group (Fig. 2). The bacteroides-group 16S rRNA probe indicated that the lowest numbers of these bacteria occurred in 0–6 month infants, and this was confirmed by significantly low levels of bacteroides DNA monitored by real-time PCR (Fig. 1).

Desulfovibrios are the predominant SRB in adult stools, and these organisms are of particular interest due to their potential link with inflammatory bowel disease [43], although it was thought that they might not be able to colonise the gut until late childhood. These bacteria are difficult to quantitate reliably using culturing methods, and are often ignored in studies on the colonic microbiota. To our knowledge, this investigation is the first to demonstrate that SRB can be detected in the stools of very young infants (0–6 months of age). With respect to DNA quantitation, desulfovibrio populations were found to be significantly lower in the stools of this group than in elder infants, although this was not consistent with the RNA analysis, which also gives an indication of metabolic activity [44]. This emphasises an important difference between the two techniques: with real-time PCR, a specific measure of cell number is achievable, while northern hybridisations give a value that is a percentage of total bacterial 16S rRNA in a sample. The presence of desulfovibrios in an infant microbiota containing large numbers of bifidobacteria might be explained by the large amounts of lactate available, which together with hydrogen, is an important electron donor for these SRB in the gut [45,46].

Contrary to previous work, significant differences in bifidobacterial populations were not observed with either increasing age or breast-feeding, while the greater proportion of lactic acid bacteria RNA detected in the 0–6 month age group was found to be approaching significance (P= 0.059). Both methods of analysis showed that breast-fed infants generally had higher bifidobacterial populations compared to bottle-fed children (log10 9.1 ± 0.7, and 7.9 ± 0.6 16S rRNA gene copy number, 8.6 ± 2.0 and 4.4 ± 3.0% of total RNA), although this only became statistically relevant when all cases were included in the analysis, irrespective of age matching. Under these circumstances, breast-fed infants had higher proportions of bifidobacterial RNA compared to those that were bottle-fed (P= 0.024), and those weaned to solid food (P= 0.002). The reason for the discrepancy might be related to the small sample size of the very young babies used in the investigation.

Bacteroides and Ent. faecalis were found to comprise a greater proportion of the microflora in bottle-fed infants than breast-fed infants by RNA analysis, however, the opposite was found by real-time PCR. Bifidobacteria accounted for only 8.6% and 4.4% by slot–blot hybridisation and 20% and 1% by real-time PCR, respectively. These differences may reflect the more sensitive direct measure of cell numbers by real-time PCR.

Using a larger set of probes for the northern hybridisations, only 33–66% of the total microflora was still recoverable, this may be due to the presence of other bacteria in faeces that these probes did not detect. Harmsen et al. [47], developed oligonucleotide probes to target the Atopobium cluster, which includes the Coriobacterium group, and used them for FISH analysis of faecal samples from breast and formula-fed infants. The findings of that study indicated that formula-fed babies had higher numbers and greater prevalence of coriobacteria in their stools. Large numbers of these organisms were found, reaching up to 39% of all bacteria in faecal samples, which also hybridised with the Atopobium probe.

In conclusion, this investigation has shown that facultative anaerobes in infant stools were greatly decreased after the age of 6 months, and that these bacteria were replaced by organisms such as those belonging to the C. coccoides-group which rapidly increased to adult levels by 7–12 months. Faecalibacterium prausnitzii is not a constituent of the colonic microflora in early infancy, whereas SRB are a part of the normal microbiota at this time. Real-time PCR provided a highly sensitive means of identifying and quantitating faecal bacteria. The technique is more sensitive than northern hybidisations, and the absence of a requirement for a radioactive label makes the method more suitable for large scale studies on the ecology of gut bacteria. However, real-time PCR gives no information on how metabolically active the bacteria are in a population, therefore, developing an increased number of probes for real-time PCR in conjunction with northern hybridisation analysis could provide new insights into the structure and metabolic activities of microbial communities in the gastrointestinal tract.


This work was funded by the Medical Research Council.