Beyond bacteria: a study of the enteric microbial consortium in extremely low birth weight infants

Abstract

Extremely low birth weight (ELBW) infants have high morbidity and mortality, frequently due to invasive infections from bacteria, fungi, and viruses. The microbial communities present in the gastrointestinal tracts of preterm infants may serve as a reservoir for invasive organisms and remain poorly characterized. We used deep pyrosequencing to examine the gut-associated microbiome of 11 ELBW infants in the first postnatal month, with a first time determination of the eukaryote microbiota such as fungi and nematodes, including bacteria and viruses that have not been previously described. Among the fungi observed, Candida sp. and Clavispora sp. dominated the sequences, but a range of environmental molds were also observed. Surprisingly, seventy-one percent of the infant fecal samples tested contained ribosomal sequences corresponding to the parasitic organism Trichinella. Ribosomal DNA sequences for the roundworm symbiont Xenorhabdus accompanied these sequences in the infant with the greatest proportion of Trichinella sequences. When examining ribosomal DNA sequences in aggregate, Enterobacteriales, Pseudomonas, Staphylococcus, and Enterococcus were the most abundant bacterial taxa in a low diversity bacterial community (mean Shannon-Weaver Index of 1.02 ± 0.69), with relatively little change within individual infants through time. To supplement the ribosomal sequence data, shotgun sequencing was performed on DNA from multiple displacement amplification (MDA) of total fecal genomic DNA from two infants. In addition to the organisms mentioned previously, the metagenome also revealed sequences for gram positive and gram negative bacteriophages, as well as human adenovirus C. Together, these data reveal surprising eukaryotic and viral microbial diversity in ELBW enteric microbiota dominated bytypes of bacteria known to cause invasive disease in these infants.

Figure 1. Lower eukaryotic sequence abundance.

Panel A: Order level data per sample is shown as a percentage of the total sequences. Panel B: Predominant yeast genera and species. Best match genus-species estimates per infant's sample are shown. Where only a genus is indicated, species level data could not be confidently determined.

Figure 2. Rarefaction curves for lower eukaryotic sequences begin to plateau.

Curves are identified by infant (first number) and sample number (second number). Phylogenetic cutoff distances for 0.03 (approximately species level) and 0.1 (approximately genus level) are shown in Panels A and B respectively.

Figure 3. Bacterial phyla and orders in aggregate.

Sequences from 16S rDNA were combined among all samples and analyzed in aggregate. Only taxa with greater than 50 sequences are shown. Panel A: Phyla-level sequence analysis. Panel B: Most-abundant orders (>0.5% of total sequences). Panel C: Low-abundance orders (<0.5% of total sequences).

Figure 4. Bacterial genera by infant sample.

Panel A: Percentages of the total sequences are shown for each sample. Shannon-Weaver Index (cutoff distance of 0.1) and Chao1 (cutoff distance of 0.1) are shown. Panel B: A graphical representation of the change in the Shannon-Weaver Index between samples per infant.

Figure 5. Rarefaction curves for 16S rDNA sequences.

Curves are identified by infant (first number) and sample number (second number). Phylogenetic cutoff distances for 0.03 (approximately species level) and 0.1 (approximately genus level) are shown in Panels A and B, respectively.