Chang, Ju Y.*; Shin, Son M.†; Chun, Jongsik‡; Lee, Jae-Hak§; Seo, Jeong-Kee||
Intestinal microbiota are essential in the regulation of metabolism in the human body and are related to the pathophysiology of various diseases. Because most previous studies investigating the composition of human intestinal microbiota, including those of preterm infants, have been performed using culture-based methods, information about unculturable components of the intestinal microbiota is still limited (1). Since the 1990s, molecular methods targeting 16S rDNA have been actively used in the investigation of intestinal microbiota. According to the results of the previous studies, only 20% to 40% of human adult intestinal microbiota are culturable (2).
Among the molecular methods used to analyze intestinal microbiota, DNA sequencing is considered the most useful tool to uncover information about the composition of intestinal microbiota. Sequencing typically identifies most intestinal bacteria to the species level and sometimes to the strain level (3). Sequencing also can be used to discover previously unidentified bacterial groups and to perform ecological analysis of microbial communities at each taxonomic level; however, classical cloning sequencing methods yield an inadequate number of sequences to represent entire bacterial communities because of time and cost limitations.
Recently, new DNA sequencing technology combined with 16S rDNA-based molecular methods have emerged as the most powerful tool to analyze microbial communities. Roche 454 FLX–based sequencing (a high-throughput pyrosequencing method) yields results that represent a more detailed and unbiased view of complex microbial communities (4). Many studies have used this tool to investigate environmental or animal microbial communities, and newer studies are beginning to investigate human intestinal microbiota (5).
Early intestinal bacterial colonization patterns in preterm infants have been reported to be different from those of healthy full-term infants. Both immaturity of vital functioning, including gut immunity, and exposure to specialized medical environments have been thought to be responsible for this difference, but further studies are still needed (6,7). Perturbation of intestinal microbiota has been recognized as a risk factor for the development of necrotizing enterocolitis (NEC) and other gastrointestinal illnesses, such as feeding intolerance (8). The possibility also exists that intestinal bacteria are the causative agents of systemic infection in preterm infants because of the immaturity of mucosal and systemic immunity. In addition, obesity and allergic diseases have been reported to be related to certain patterns of intestinal microbiota (9–11). Therefore, the establishment of healthy intestinal microbiota during the early colonization period is likely important in preventing certain diseases.
According to the results of previous studies using culture-based methods, the intestines of preterm infants have a more limited number of species and the colonization of bifidobacteria is delayed compared with full-term infants. In addition, potentially pathogenic bacteria such as Escherichia coli, enterococci, and staphylococci are dominant (12–14). Although the number of studies using molecular methods is still small compared with those of culture-based studies, the results of both methods are compatible in many aspects. To the best of our knowledge, only 3 studies to date have investigated intestinal microbiota using sequencing methods and only 1 study has used high-throughput technology (7,15,16). Recently, strict anaerobic bacteria such as Ruminococcus spp were reported to colonize during the initial stage of succession of neonatal intestinal microbiota (17). Therefore, it is possible that other unculturable bacterial species also may be involved during the initial colonization process, although previously aerobic bacteria are known to predominate during this period. The previous study using a classical cloning sequencing method to explore the intestinal microbiota of preterm infants did not reveal any unidentifiable bacterial species (7). Molecular methods revealing unculturable or undocumented organisms such as deep pyrosequencing are needed to investigate thoroughly the composition of intestinal microbiota (18).
Here, we investigated the intestinal microbiota of preterm infants using high-throughput sequencing technology to explore the diversity of these bacterial communities and monitor early intestinal colonization patterns in Korean preterm infants.
SUBJECTS AND METHODS
Preterm infants were recruited from a single neonatal care unit at Cheil General Hospital and Women's Healthcare Center for 6 months. Fecal samples that were defecated spontaneously were serially collected within 72 hours after birth, at 2 weeks, and at 1 month of age from all of the preterm infants of birth weight <1500 g. Preterm infants who were treated with antibiotics after the first week, could not be fed for >5 successive days, or had congenital heart disease, chromosomal anomalies, or severe birth asphyxia were excluded. The protocols for the present study were approved by the institutional review board for Seoul metropolitan government, Seoul National University Boramae Medical Center, and Cheil General Hospital and Women's Healthcare Center.
Stool Collection and DNA Isolation
Fecal samples were collected in sterile tubes and immediately frozen at −20°C. They were transferred to the laboratory on dry ice within 72 hours for further processing. Total bacterial DNA was extracted from each fecal sample using a commercial DNA isolation kit (QIAamp DNA stool mini kit; QIAGEN, Germantown, MD) with some modifications as described previously (19). Briefly, approximately 1 mL volume of 0.1 mm zirconia/silica beads (Biospec Products, Bartlesville, OK) was added to fecal slurries after the addition of ASL buffer. The mixtures were processed for 1 minute on a Mini Beadbeater-8 (BioSpec Products) and incubated at 95°C for 10 minutes, instead of the manufacturer-recommended 70°C.
Amplification of 16S rRNA Genes
The 16S rRNA-encoding gene sequences were amplified from each DNA sample using a pair of universal bacterial fusion primers that target the V2 region of the molecule. For pyrosequencing and tagging each polymerase chain reaction (PCR) product, 454 Life Sciences primer A and the unique 8-base barcode (NNNNNNNN) with a linker (AC) were ligated to the reverse primer 338R (5′-GCCTCCCTCGCGCCATCAGCANNNNNNNNGCTGCCTCCCGTAGGAGT-3′). The forward primer 27F was also fused with 454 Life Sciences primer B (5′-GCCTTGCCAGCCCGCTCAGTCAGAGTTTGATCMTGGCTCAG-3′) (20,21). PCR was performed using PuRe Taq Ready-To-Go PCR beads (GE Healthcare, Piscataway, NJ). Reaction mixtures were set up with 50 ng of template DNA, 10 pmol of each primer, and water to a total volume of 25 μL. This yielded a reaction mixture containing 1.5 U Taq polymerase, 10 mmol/L Tris-HCl (pH 9.0 at room temperature), 50 mmol/L KCl, 1.5 mmol/L MgCl2, a 100-μmol/L concentration of each nucleotide, and stabilizers, including bovine serum albumin. The reaction mixtures were subjected to amplification in a DNA thermal cycler (Mastercycler gradient; Eppendorf AG, Hamburg, Germany) with the following cycling conditions: initial denaturation at 94°C for 5 minutes followed by 20 cycles of denaturation at 94°C for 30 seconds, annealing at 55°C for 30 seconds, and extension at 72°C for 30 seconds. A final extension at 72°C for 7 minutes was performed. Two independent PCR reactions were performed for each sample.
Amplicon Purification and Pyrosequencing
Amplicons were isolated using QIAquick Gel extraction kits (QIAGEN, Valencia, CA) followed by QIAquick PCR purification kits (QIAGEN) according to the recommendations of the manufacturer. For library quality assessment and quantitation, each purified PCR product was analyzed with a Bioanalyzer DNA1000 LabChip (Agilent, Palo Alto, CA) and a Quant-iT PicoGreen dsDNA Assay Kit (Invitrogen, Carlsbad, CA). Equal amounts of each PCR product were added to a master pool of DNA, which was then subjected to pyrosequencing with the Genome Sequencer FLX system (Roche, Basel, Switzerland) according to the recommendations of the manufacturer (GS FLX amplicon DNA library preparation method manual; GS FLX sequencing method manual).
Taxonomic Assignment of Individual Sequencing Reads
Preprocessing of sequencing reads was performed as described previously (22). After preprocessing, the EzTaxon-extended database (http://www.eztaxon-e.org) was used for taxonomic assignment of each pyrosequencing read as described previously (22,23). Individual pyrosequencing reads were compared with sequences in the EzTaxon-extended database using the combination of BLASTN-based search and pairwise sequence similarity comparison. The following criteria were used for taxonomic assignment of each read (x = similarity): species (x ≥97%), genus (97> x ≥94%), family (94> x ≥90%), order (90> x ≥85%), class (85> x ≥80%), and phylum (80> x ≥75%). If the similarity was below the cutoff of a given rank, then the read was assigned as an unclassified group under the next assigned taxonomic rank. For example, a sequence read that shows 95% similarity to E coli, is assigned as unclassified Escherichia spp, that is, Escherichia_uc.
Calculation of Species Richness and Diversity Indices
The diversity and species-richness indices were calculated using the Ribosomal RNA Database Project's pyrosequencing pipeline (http://pyro.cme.msu.edu). The cutoff value for assigning a sequence to the same group was ≥97% similarity.
The differences between different stool collection times were analyzed with nonparametric methods using SPSS 16.0 (SPSS Inc, Chicago, IL) software. The statistical significance was set at P <0.05.
Characteristics of the Infants
A total of 30 fecal samples from 10 Korean preterm infants were included for bioinformatic processing. Clinical information about the subjects is shown in Table 1.
Information Related to 16S rDNA Sequences
A total of 46,369 partial 16S rRNA-encoding gene sequences passed quality control and were used for final analysis (Table 2). The average number of sequences per fecal sample was about 1500. The numbers of sequences of individual microbiome were not significantly different between fecal collection times (P > 0.05). To show sequencing depth, the rarefaction curves for the preterm infants at 1 month of age (at 3% sequence dissimilarity) are shown in supplemental Figure S1 (http://links.lww.com/MPG/A64).
Diversity of Fecal Microbiota
A total of 27 molecular phyla, 216 molecular genera, and 393 molecular species were identified. Among the total molecular species, 86 (21.9%) were not identified at the species level and were therefore considered unclassified.
The number of molecular phyla, genera, and species showed significant variation among individual infants and the 3 postnatal ages at which samples were taken (Table 3). The subjects did not significantly differ in gestational age or birth weight. The total number of molecular phyla and species were greatest in samples taken within 72 hours after birth. This also was the collection time at which variation between individuals was the greatest. The majority of the molecular species were transient colonizers; 226 (65.9%) of the 343 total molecular species found within 72 hours after birth were not observed in the latter 2 periods. The number of major molecular species (arbitrarily defined as the molecular species representing >1% of the total sequences of the individual microbiome) was not significantly different from those of the latter 2 periods. In contrast, the proportions of rare and unclassified molecular species were significantly higher than the latter 2 periods. The number of both total and major molecular species and the diversity indices were significantly greater at 1 month of age than at 2 weeks of age. The median number of molecular species in individual infants was 21 at 2 weeks of age and 35 at 1 month of age. Most of the molecular species observed at 2 weeks were persistent colonizers; 94 (83.2%) of the 113 total molecular species from the 10 preterm infants also were observed at 1 month of age.
Taxonomic Composition of Fecal Microbiota
At the division level, nearly 90% of the total sequences belonged to the phyla Proteobacteria and Firmicutes during the whole observation period (Table 4). The classes Gammaproteobacteria, Bacilli, and Clostridia were the most predominant (Fig. 1). The most abundant and ubiquitous major molecular genera or species (defined as those representing >1% of the total sequences in at least 1 individual microbiome and found in more than half of the preterm infants) common to all of the postnatal ages were E coli and species belonging to the genera Enterobacter, Enterococcus, Veillonella, Serratia, Staphylococcus, Roseburia, Acinetobacter, Citrobacter, Bacteroides, Faecalibacterium, Blautia, and Streptococcus (see Supplemental Tables S1 [http://links.lww.com/MPG/A65], S2 [http://links.lww.com/MPG/A66], and S3 [http://links.lww.com/MPG/A67]).
Although there was high interindividual variation, the proportion of certain molecular phyla, genera, or species in the fecal microbiome of each individual preterm infant was significantly varied based on postnatal age. At the division level, the proportion of the phylum Bacteroidetes increased significantly at 1 month of age compared with 2 weeks of age (Table 4); however, the proportion of the 2 predominant phyla, Firmicutes and Proteobacteria, did not vary significantly over time. Many other minor molecular phyla observed within 72 hours of age were not found in the fecal microbiota of the preterm infants at 2 weeks or 1 month of age. At the class level, the total proportion of Gammaproteobacteria and Bacilli, which contained most of the ubiquitous pathogenic or potentially pathogenic molecular species, tended to decrease at 1 month of age compared with within 72 hours after birth or 2 weeks of age in many of the preterm infants, although the change was not statistically significant (Fig. 1A, B). On the contrary, the proportion of the class Clostridia except C difficile and C perfringens increased significantly at 1 month of age compared with 72 hours after birth (Fig. 1A, C).
At the genus and species levels, among the most abundant and ubiquitous major molecular genera or species, certain molecular species were especially ubiquitous and/or abundant only at a specific postnatal age or their proportions varied significantly between ages. Within 72 hours after birth, many of these molecular species were pathogenic or potentially pathogenic bacterial groups (Table 5). Pseudomonas spp were found ubiquitously and abundantly only at this age. The proportion of the sequences corresponding to Acinetobacter spp was significantly lower at 1 month of age compared with within 72 hours after birth. The proportion of sequences corresponding to C difficile was far higher than at the other 2 ages in 2 of the preterm infants (P6: 87.9%, P7: 30.2%) (see Supplemental Tables S1 [http://links.lww.com/MPG/A65], S2 [http://links.lww.com/MPG/A66], and S3 [http://links.lww.com/MPG/A67]). At 1 month of age, the proportion of certain nonpathogenic or anaerobic molecular species including Veillonella spp, Eubacterium spp (Roseburia genus), Bacteroides spp, Phascolartobacterium spp, and Lactobacillus spp was significantly greater compared with within 72 hours after birth and/or 2 weeks of age (Table 6). The pathogenic or potentially pathogenic bacterial groups such as Haemophilus spp also were found ubiquitously and abundantly only at this age (Table 5).
The primary objective of the present study was to evaluate the utility of the pyrosequencing method with barcoded primers for comparative analysis of human intestinal microbiota at neonatal age. We observed many molecular species previously unknown to adult fecal microbiome. Furthermore, the results of the present study also are compatible with those of previous studies using culture or other molecular methods. Sequences corresponding to both ubiquitous core pathogenic or potentially pathogenic species and common nonpathogenic or anaerobic species were identified, allowing us to compile important medical and ecologic information.
To minimize the influence of environmental factors that are known to affect the normal development of intestinal microbiota, we included relatively homogeneous subjects. The preterm infants examined here were very-low-birth weight infants from 1 neonatal intensive care unit (NICU), no subjects received antibiotics after the first week, and all of them were delivered via cesarean section. Although we did not analyze the influence of other factors such as feeding method, including breast-milk feeding or formula feeding, or gestational age because of the small number of subjects, the results of our study do provide a basis for future large-scale research using this pyrosequencing method.
According to the results of previous studies on the preterm intestinal microbiota, the diversity of intestinal microbiota of the preterm infants was far lower than that of healthy human intestinal microbiota (14,24). Using cloning and Sanger sequencing metagenomics, Magne et al (7) observed an average of 3.25 (range 1–8) molecular species from the intestinal microbiota of 16 preterm infants and concluded that no unknown species were detectable. In our study, we obtained approximately 10 times more molecular species, including those corresponding to anaerobic bacterial groups. New molecular species, which were unknown previously, represented approximately 20% of the intestinal microbiota in the preterm infants. These results demonstrate that even the diversity of intestinal microbiota of preterm infants, which may be the lowest among all of the human intestinal microbiota, cannot be investigated thoroughly with classical cloning sequencing methods.
According to our results, the diversity of intestinal microbiota varied according to postnatal age. The total diversity of the intestinal microbiota was generally greater at 1 month of age compared with 2 weeks after birth. In contrast to samples taken from infants at 2 weeks after birth and 1 month of age, a majority of the diversity within 72 hours after birth was caused by a large amount of transient and/or rare molecular species. The proportion of the unclassified molecular species, which did not correspond to known species, was significantly higher and environmental bacteria such as Aquabacterium spp were ubiquitous and abundant only in this period (see Supplemental Tables S1 [http://links.lww.com/MPG/A65], S2 [http://links.lww.com/MPG/A66], and S3 [http://links.lww.com/MPG/A67]). The diversity of the initial stage of intestinal bacterial colonization in preterm infants exposed to specialized medical environments seemed vulnerable and different from the colonization patterns of the later neonatal period.
In samples from all of the preterm infants during the entire observation period, most 16S rDNA sequences belonged to 2 major bacterial phyla, Proteobacteria and Firmicutes. The class Bacilli represented a significant portion within the phylum Firmicutes. These data are similar to those from the intestinal microbiota of the full-term infants (25) and is thought to be a characteristic pattern of the neonatal intestinal microbiota. It is now well known that in the intestinal microbiota of adults, the phyla Firmicutes and Bacteroidetes are codominant and most bacterial species within Firmicutes belong to the class Clostridia. Therefore, compared with adult intestinal microbiota, the proportion of anaerobic bacterial species was lower and the proportion of potentially pathogenic bacteria was higher in the neonatal intestinal microbiota. Recently, the most extensive metagenomic analysis on adult intestinal microbiome was published (26). The present study reported that the most abundant and common 75 bacterial species present in >50% of 124 European individuals with genome coverage >1% were identifiable. Although direct comparison with our data is not possible because of the different methodologies used in the 2 studies, only 3 species belonging to the phylum Firmicutes among those 75 bacterial species belonged to the abundant and ubiquitous major species in our study.
In preterm infants, the immaturity of the gut mucosal barrier and systemic immunity can allow potentially pathogenic bacterial species to translocate to intestinal mucosa and enter blood vessels and other tissues, leading to sepsis and systemic infection (27). Consequently, there is a strong practical medical motivation to explore the intestinal microbiota of preterm infants. In the present study, we detected a large number of the sequences belonging to the groups of pathogenic or potentially pathogenic bacterial species in the intestines of preterm infants. The sequences corresponding to certain pathogenic or potentially pathogenic bacterial species, including C difficile, Pseudomonas, and Acinetobacter spp, were more abundant and/or ubiquitous within 72 hours after birth compared with the other 2 ages. Although we cannot definitely conclude that this feature is solely caused of the effect of antibiotics, this notion is supported by the fact that this observation was not present in the fecal microbiota within 72 hours after birth in healthy full-term infants who were not treated with antibiotics (unpublished data). For example, the marked transient dominancy of C difficile within 72 hours after birth in 2 of the preterm infants (patients 6 and 7) has been observed in adult intestinal microbiota under antibiotic stress in previous studies (28), although less prominently. There is a possibility that in addition to the treatment with antibiotics, the combination of multiple factors including the immaturity of vital organs of the preterm infants, cesarean section delivery, and the special NICU environment may bring about the perturbation of intestinal microbiota during this vulnerable period. Among the ubiquitous major potential pathogens, Haemophilus spp (mostly H parainfluenza in our study) and Cronobacter spp were rarely reported as important colonizers of preterm infants in previous culture-based studies. Haemophilus spp are mostly fastidious Gram-negative bacteria that are usually not easily culturable on regular agar plates. These bacteria have been reported as causative agents of infective endocarditis in young children (29). In our study, Haemophilus spp were significantly more ubiquitous and abundant at 1 month of age compared with the former 2 postnatal ages. Cronobacter spp were also recognized as causative agents of neonatal bacteremia, meningitis, and NEC (30,31).
In our study, the molecular species belonging to anaerobic bacteria also were ubiquitous even within 72 hours after birth. During the initial stage of bacterial colonization, it is thought that aerobic bacteria create a reduced environment favorable to anaerobic bacteria, which colonize the gut at the end of the first week of life, taking over from the aerobic bacteria (18). With the introduction of molecular methods, however, there were a limited number of case reports suggesting that the certain anaerobic bacterial species can be the first dominant colonizers (17). The results of our study support this idea in that the intestinal environment is already sufficiently anaerobic to support the growth of anaerobic bacteria during the first 3 days of life even in preterm infants.
Many anaerobic bacterial species were identified as rare species within 72 hours after birth and their proportion increased thereafter. The proportion of the sequences corresponding to the class Clostridia, except C difficile and C perfringens, increased significantly at 1 month of age compared with within 72 hours after birth. Among the bacterial species belonging to the class Clostridia, Veillonella spp were most abundant. They are well known as lactate-fermenting bacteria and have been reported to be important colonizers in the intestinal microbiota of both adults and infants in a few previous studies (17,32). Previous research investigating the composition of anaerobic intestinal microbiota of preterm infants focused mainly on Bacteroides spp and Clostridium-related species. Both of these species have been known to be predominant bacterial species in the adult fecal microbiota and perform important metabolic functions. According to the results of previous studies, Bacteroides spp are relatively ubiquitous and are core species in the intestinal microbiota of preterm infants. They increased over time and were relatively more common in formula-fed infants compared with breast-fed infants, which is consistent with the results of our study (see Supplemental Tables S1 [http://links.lww.com/MPG/A65], S2 [http://links.lww.com/MPG/A66], and S3 [http://links.lww.com/MPG/A67]).
Clostridium-related species have been investigated as part of the C coccoides-Eubacterium rectale group in previous molecular studies. In our study, a total of 13 Clostridium spp were identified. Most of these species were rare, except C difficile. In contrast, Eubacterium rectale was the most abundant and ubiquitous bacterial species among the C coccoides-E rectale group. Ruminococcus spp, which were reported as initial colonizers of neonates in a previous molecular study (17,18), also were found ubiquitously in our preterm infants.
Bifidobacterium spp were sparse in the fecal microbiota of the preterm infants in our study, which is consistent with the results of previous studies (33). Bifidobacterium spp were observed less frequently in studies involving sequencing-based metagenomics research compared with those that used culture-based methods (16). Because the sequences of Bifidobacterium spp were detected in up to 5% of the total sequences of the full-term infants using the same pyrosequencing method (unpublished data), it is likely that the sparsity of the sequences corresponding to Bifidobacterium spp in preterm infants was not solely the result of the sequencing method. We postulate that a combination of factors, including use of empirical antibiotics during the first several days of life, the special environment of the NICU, cesarean section delivery method, and formula feeding also may influence the relative sparsity of Bifidobacterium spp colonization. There was also a recent report that a high Bacteroides count was associated with a low bifidobacterial count and vice versa (34). In our study, Bacteroides spp were one of the most predominant and ubiquitous bacteria common to all of the postnatal ages and were especially abundant at 1 month of age. Additionally, the sequences corresponding to Lactobacillus spp, which also are important candidates for biotherapeutic agents, were identified as rare species in most preterm infants. At 1 month of age, they were observed ubiquitously and had increased significantly in number, which also is consistent with the results of previous studies (12).
Preterm infants are greatly heterogeneous in nature with various factors such as gestational age, duration and type of antibiotic treatment, mode of delivery, and feeding type, all of which may affect the composition of the intestinal microbiota. Therefore, the type of the influencing factor should be taken into consideration in the interpretation of the results of the studies investigating the intestinal microbiota of preterm infants; however, the results of previous studies investigating these factors have been inconsistent in some aspects. It has been reported that the composition of intestinal microbiota of preterm infants became more homogenous as time progressed, regardless of different influencing factors (24). In our study, interindividual similarity of the intestinal microbiota was not significantly different between the 3 postnatal ages (data not shown).
In conclusion, we have described the preterm intestinal microbiota more thoroughly than previously possible with a novel pyrosequencing method with barcoded primers. Although pathogenic or potentially pathogenic bacterial species were the predominant species, a diverse array of anaerobic or nonpathogenic bacterial species also was found ubiquitously during the neonatal period. The composition and diversity of intestinal microbiota differed based on the postnatal ages and were distinct from that of healthy full-term infants. These results provide important data in interpreting the results of bacteriological tests in preterm infants. These data also may be used as a framework for further research to study the role of intestinal microbiota in certain diseases affecting preterm infants, such as NEC, or to manipulate the intestinal flora to establish a healthy intestinal microbiota.
1. Furrie E. A molecular revolution in the study of intestinal microflora. Gut 2006; 55:141–143.
2. Zoetendal EG, Collier CT, Koike S, et al. Molecular ecological analysis of the gastrointestinal microbiota: a review. J Nutr 2004; 134:465–472.
3. Dethlefsen L, Eckburg PB, Bik EM, et al. Assembly of the human intestinal microbiota. Trends Ecol Evol 2006;21:517–23.
4. Sogin ML, Morrison HG, Huber JA, et al. Microbial diversity in the deep sea and the underexplored “rare biosphere”. Proc Natl Acad Sci U S A 2006; 103:12115–12120.
5. Andersson AF, Lindberg M, Jakobsson H, et al. Comparative analysis of human gut microbiota by barcoded pyrosequencing. PLoS One 2008; 3:e2836.
6. Fanaro S, Chierici R, Guerrini P, et al. Intestinal microflora in early infancy: composition and development. Acta Paediatr Suppl 2003; 91:48–55.
7. Magne F, Abely M, Boyer F, et al. Low species diversity and high interindividual variability in faeces of preterm infants as revealed by sequences of 16S rRNA genes and PCR-temporal temperature gradient gel electrophoresis profiles. FEMS Microbiol Ecol 2006; 57:128–138.
8. Millar MR, Linton CJ, Cade A, et al. Application of 16S rRNA gene PCR to study bowel flora of preterm infants with and without necrotizing enterocolitis. J Clin Microbiol 1996; 34:2506–2510.
9. Ley RE, Backhed F, Turnbaugh P, et al. Obesity alters gut microbial ecology. Proc Natl Acad Sci U S A 2005; 102:11070–11075.
10. Kalliomaki M, Collado MC, Salminen S, et al. Early differences in fecal microbiota composition in children may predict overweight. Am J Clin Nutr 2008; 87:534–538.
11. Penders J, Thijs C, Vink C, et al. Factors influencing the composition of the intestinal microbiota in early infancy. Pediatrics 2006; 118:511–521.
12. Blakey JL, Lubitz L, Barnes GL, et al. Development of gut colonisation in pre-term neonates. J Med Microbiol 1982; 15:519–529.
13. Sakata H, Yoshioka H, Fujita K. Development of the intestinal flora in very low birth weight infants compared to normal full-term newborns. Eur J Pediatr 1985; 144:186–190.
14. Gewolb IH, Schwalbe RS, Taciak VL, et al. Stool microflora in extremely low birthweight infants. Arch Dis Child Fetal Neonatal Ed 1999; 80:F167–F173.
15. Wang Y, Hoenig JD, Malin KJ, et al. 16S rRNA gene-based analysis of fecal microbiota from preterm infants with and without necrotizing enterocolitis. ISME J 2009; 3:944–954.
16. Mshvildadze M, Neu J, Shuster J, et al. Intestinal microbial ecology in premature infants assessed with non-culture-based techniques. J Pediatr 2010; 156:20–25.
17. Favier CF, Vaughan EE, De Vos WM, et al. Molecular monitoring of succession of bacterial communities in human neonates. Appl Environ Microbiol 2002; 68:219–226.
18. Morelli L. Postnatal development of intestinal microflora as influenced by infant nutrition. J Nutr 2008; 138:1791S–1795S.
19. Li F, Hullar MA, Lampe JW. Optimization of terminal restriction fragment polymorphism (TRFLP) analysis of human gut microbiota. J Microbiol Methods 2007; 68:303–311.
20. McKenna P, Hoffmann C, Minkah N, et al. The macaque gut microbiome in health, lentiviral infection, and chronic enterocolitis. PLoS Pathog 2008; 4:e20.
21. Hamady M, Walker JJ, Harris JK, et al. Error-correcting barcoded primers for pyrosequencing hundreds of samples in multiplex. Nat Methods 2008; 5:235–237.
22. Chun J, Kim KY, Lee JH, et al. The analysis of oral microbial communities of wild-type and Toll-like receptor 2-deficient mice using a 454 GS FLX titanium pyrosequencer. BMC Microbiol 2010; 10:101.
23. Chun J, Lee JH, Jung Y, et al. EzTaxon: a web-based tool for the identification of prokaryotes based on 16S ribosomal RNA gene sequences. Int J Syst Evol Microbiol 2007; 57:2259–2261.
24. Schwiertz A, Gruhl B, Lobnitz M, et al. Development of the intestinal bacterial composition in hospitalized preterm infants in comparison with breast-fed, full-term infants. Pediatr Res 2003; 54:393–399.
25. Palmer C, Bik EM, DiGiulio DB, et al. Development of the human infant intestinal microbiota. PLoS Biol 2007; 5:e177.
26. Qin J, Li R, Raes J, et al. A human gut microbial gene catalogue established by metagenomic sequencing. Nature 2010; 464:59–65.
27. Westerbeek EA, van den Berg A, Lafeber HN, et al. The intestinal bacterial colonisation in preterm infants: a review of the literature. Clin Nutr 2006; 25:361–368.
28. Chang JY, Antonopoulos DA, Kalra A, et al. Decreased diversity of the fecal microbiome in recurrent Clostridium difficile-associated diarrhea. J Infect Dis 2008; 197:435–438.
29. Feder HM Jr, Roberts JC, Salazar JC, et al. HACEK endocarditis in infants and children: two cases and a literature review. Pediatr Infect Dis J 2003;22:557–62.
30. Muytjens HL, Zanen HC, Sonderkamp HJ, et al Analysis of eight cases of neonatal meningitis and sepsis due to Enterobacter sakazakii. J Clin Microbiol 1983;18:115–20.
31. Friedemann M. Epidemiology of invasive neonatal Cronobacter (Enterobacter sakazakii) infections. Eur J Clin Microbiol Infect Dis 2009; 28:1297–1304.
32. Rinttila T, Kassinen A, Malinen E, et al. Development of an extensive set of 16S rDNA-targeted primers for quantification of pathogenic and indigenous bacteria in faecal samples by real-time PCR. J Appl Microbiol 2004; 97:1166–1177.
33. Hall MA, Cole CB, Smith SL, et al. Factors influencing the presence of faecal lactobacilli in early infancy. Arch Dis Child 1990; 65:185–188.
34. Harmsen HJ, Wildeboer-Veloo AC, Raangs GC, et al. Analysis of intestinal flora development in breast-fed and formula-fed infants by using molecular identification and detection methods. J Pediatr Gastroenterol Nutr 2000; 30:61–67.