- Record: found
- Abstract: not found
- Article: not found
Peripartum Antibiotics Promote Gut Dysbiosis, Loss of Immune Tolerance, and Inflammatory Bowel Disease in Genetically Prone Offspring
Author(s):
Jun Miyoshi ,
Alexandria M. Bobe ,
Sawako Miyoshi ,
Yong Huang ,
Nathaniel Hubert ,
Tom O. Delmont ,
A. Murat Eren ,
Vanessa Leone ,
Eugene B. Chang
Publication date Created:
July 2017
Publication date (Print):
July 2017
Journal:
Cell Reports
Publisher:
Elsevier BV
Read this article at
There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.
Abstract
<p id="P4">Factors impacting the developing neonatal gut microbiome and immune networks
may increase
risk for developing complex immune disorders such as inflammatory bowel diseases (IBD).
In particular, peripartum antibiotics have been suggested as risk factors for human
IBD, although direct evidence are lacking. We therefore examined the temporal impact
of the commonly used antibiotic, cefoperazone, on both maternal and offspring microbiota
when administered to dams during the peripartum period in the IL-10 deficient murine
colitis model. By rigorously controlling for cage, gender, generational, and murine
pathobiont confounders, we observed that offspring from cefoperazone-exposed dams
develop a persistent gut dysbiosis into adulthood associated with skewing of the host
immune system and increased susceptibility to spontaneous and chemically (DSS)-induced
colitis. Thus, early life exposure to antibiotic-induced maternal dysbiosis during
a critical developmental window for gut microbial assemblage and immune programming
elicits a lasting impact in genetically susceptible offspring to increase IBD risk.
</p><p id="P5">
<div class="figure-container so-text-align-c">
<img alt="" class="figure" src="/document_file/7f2bdee0-c51b-41a6-bb17-d9afcecdc22a/PubMedCentral/image/nihms888805u1.jpg"/>
</div>
</p>
Related collections
Most cited references25
- Record: found
- Abstract: found
- Article: not found
Exploring Microbial Diversity and Taxonomy Using SSU rRNA Hypervariable Tag Sequencing
Susan M. Huse, Les Dethlefsen, Julie A. Huber … (2008)
Introduction The biosphere contains between 1030 and 1031 microbial genomes, at least 2–3 orders of magnitude more than the number of plant and animal cells combined [1]. Microbes control global utilization of nitrogen through nitrogen fixation, nitrification, and nitrate reduction, and drive the bulk of sulfur, iron and manganese biogeochemical cycles [2]. They regulate the composition of the atmosphere, influence climates, recycle nutrients, and decompose pollutants. Without microbes, multi-cellular life on earth would not have evolved and biology as we know it would not be sustainable. The diversity of microbial communities and their ecologic and metabolic functions are being explored across a great range of natural environments: in soils [3]–[5], air [6] and seas [7]–[10], on plants [11] and in animals [12],[13] and in extreme environments such as the arctic [14], deep-sea vents [15], uranium-contaminated soil [16], and waste-water treatment discharge [17]. In recognition of the role marine microbes play in the biogeochemical processes that are critical to life in all environments on Earth including carbon and nitrogen cycling, the International Census of Marine Microbes (ICoMM: http://icomm.mbl.edu) has launched an international effort to catalogue the diversity of microbial populations in the oceanic, coastal, and benthic waters. Microbes associated with human health will be intensely studied through two recent large-scale initiatives: the Human Microbiome Project sponsored by the NIH (http://nihroadmap.nih.gov/hmp/) and MetaHIT sponsored by the EU (http://www.metahit.eu), which seek to characterize the composition, diversity and distribution of human-associated microbial communities. Other recent human health studies include microbes in breast milk [18], chronic wounds [19], human gut [20], dental caries [21], and childcare facilities [22]. Microbes associated with the human body outnumber human cells by at least a factor of ten [23]. Some microbes cause disease, but the overwhelming majority are either innocuous or play a role in human physiology, including immune response, digestion and vitamin production. As recently as the late 1980's, descriptions of human-associated microbiota were constrained by cultivation technologies. Over the last twenty years, sequencing surveys of amplified regions of small subunit ribosomal RNA (SSU rRNA) genes have revealed that microbial diversity is much greater than the 5,000 microbial species described using phenotypic features in Bergey's taxonomic outline [24], and that microbial communities are far more complex than initially thought. For instance, E. coli, once thought to be a dominant species in the human gut, is clearly a minor member relative to various members of the phyla Bacteroidetes and Firmicutes. It is now evident that microbiologists have been successful in culturing fewer than one percent of the different kinds of single cell organisms from most microbial communities [25]. Even for well-studied communities, such as the human distal gut, only 20–40% of the microbes have been cultured. Deeper surveys with new approaches are revealing ever-greater diversity. Even these studies, with hundreds of thousands of microbes sampled, have not been extensive enough to provide a complete picture of the diversity (richness) and relative abundance (evenness) of microbial communities. To explore these questions, microbiologists must be able to compare microbial communities within and across individuals, in different states of health and disease, and over time. The first step in these community analyses is to develop detailed descriptions of each population, including low abundance taxa that comprise the rare biosphere [9]. Exploration of the human microbiome can leverage methods used to explore the microbiome of other environments such as soil, the deep sea, and other vertebrate microbiomes. By necessity, microbiologists have historically focused their efforts on the dominant components of microbial communities. Recognizing the importance of gathering information about high, medium and very low abundance taxa, Sogin et al. [9] introduced the use of massively-parallel DNA sequencing of short hypervariable regions of SSU rRNA to characterize microbial populations. In a subsequent study that collected nearly one million short hypervariable region tags, Huber et al. [15] demonstrated that there are over ∼40,000 different kinds of bacteria and archaea in a few liters of hydrothermal vent fluid. Rarefaction data from this study and others show that in many environments, even this level of sequencing is insufficient to fully describe microbial diversity [5],[26]. The lower cost and higher throughput of pyrosequencing employed in these studies allows for sampling efforts that are orders of magnitude greater than traditional capillary dideoxy sequencing of cloned SSU rRNA amplicons [27]. With the recently announced capability to sequence >400 nt, it will be possible to span most hypervariable regions, multiple adjacent hypervariable regions, or possibly combinations of non-adjacent hypervariable regions through paired-end sequencing strategies. However, comparisons of 400 nt reads from rapidly evolving rRNA regions do not contain sufficient evolutionary information to infer robust phylogenetic relationships and the hundreds of thousands of reads produced in a single experiment far exceeds the limitations of current phylogenetic software. Tags from the V6 region are also too short for current implementations of Bayesian classifiers such as the Ribosomal Database Project Classifier (RDP) [28]. However, each read represents a hypervariable region tag of an SSU rRNA gene present in the sample. We developed a tag search engine, Global Alignment for Sequence Taxonomy [9], GAST, which utilizes existing databases of full-length SSU rRNA genes and their pre-computed phylogeny for high-throughput taxonomic analysis of microbial communities using hypervariable region tag sequences. The use of a single, small hypervariable region tag for assigning taxonomy presents several challenges. The information content in a short hypervariable region sequence may not be sufficient for inferring taxonomic affinity. BLAST [29] alone is insufficient to consistently identify the sequences in molecular databases that are the closest matches to tag queries along their full length. Here we analyze the reliability of assigning taxonomic identifiers based solely on tags, specifically using the V3 and the V6 hypervariable regions of SSU rRNA. Using SSU rRNA genes from the human gut and deep-sea vents, we compare the taxonomic assignments of the full-length sequences with the taxonomic assignments of their V3 and V6 regions, excised in silico. We then examine microbial populations of the human gut in greater detail, using both massively-parallel pyrosequencing of hypervariable region tag and Sanger-generated full-length sequences, to determine if any differences in sampling and taxonomic assignment exist with these two sequencing strategies. In a companion paper [30], we used GAST to examine the impact of the antibiotic ciprofloxacin on population structures of the human microbiome. Results Assessment of Hypervariable Region Specificity in the RefSSU Database Tags from a hypervariable region must map to a full-length SSU rRNA with minimal ambiguity to serve as reliable phylogenetic markers; i.e. phylogenetically distinct lineages should not contain identical tags. Most of the redundant SSU rRNA sequences containing identical sequences for a particular hypervariable region in the database are from the same genus. This redundancy does not interfere with the use of the hypervariable region tags for taxonomy, but rather strengthens the assignment. For each unique V3 and V6 sequence, we examined the number and taxonomy of all the source full-length sequences. We treat each hypervariable region independently, since two full-length rRNA sequences with identical V3 regions may differ at the V6 region or another hypervariable region. Of the 59,830 unique V6 reference sequences, 74% mapped to one SSU rRNA sequence, 10% mapped to 2 sequences, and only 5% mapped to 7 or more. The V3 region, which is longer, showed slightly better resolution: 82% mapped to one SSU rRNA, 8% mapped to 2 sequences, and only 3% mapped to 7 or more (results not shown). Although only a small percentage of tags map to more than a few SSU rRNA sequences, this included some that mapped to a very large number of different SSU rRNA sequences. For example, 3 V6 reference sequences and 8 V3 reference sequences each mapped to more than 1000 different SSU rRNA sequences. In almost all cases where a hypervariable tag sequence maps to more than one full-length SSU rRNA sequence, the overwhelming majority of full-length sequences still map to only one genus (Table 1). Since RefSSU sequences that give rise to identical hypervariable region tags are generally from the same or highly similar organisms, V6 and V3 tags can be unambiguously mapped to the genus level 97% and 99% of the time, respectively. Even if we examine only the subset of V3 sequences that mapped to multiple SSU rRNA sequences, 95% map uniquely to genus, 98% to family, and 99% to order, class and phylum. Similarly for the V6 region, 91% of tags derived from multiple SSU rRNA entries map uniquely to genus, 96% to family, 97% to order and 99% to class and phylum. Even in those cases where reference tags mapped to more than 1,000 full-length sequences, in 9 of the 11 cases there was one dominant taxon. Of the other two cases, one had complete consensus to the order level and the other had all but two matching at family level. Not only do most of the reference tags in the database have only one SSU rRNA source, even those from multiple SSU rRNA sources still represent almost exclusively one taxon. 10.1371/journal.pgen.1000255.t001 Table 1 Percent of hypervariable region tags from the RefSSU database that map to one or more taxa. Hypervariable region V3 Number of Taxa 1 2 3 4 5+ Phylum 99.96% / 114328 0.04% / 42 0.00% / 42 0 0 Class 99.93% / 109352 0.07% / 77 0.00% / 2 0 0 Order 99.88% / 99682 0.12% / 113 0.00% / 4 0 0 Family 99.62% / 88015 0.34% / 297 0.04% / 34 0.00% / 3 0.00% / 3 Genus 99.11% / 69686 0.07% / 495 0.09% / 64 0.05% / 35 0.00% / 29 Hypervariable region V6 Number of Taxa 1 2 3 4 5+ Phylum 99.83% / 54728 0.17% / 94 = 0.15, insufficient data to analyze the accuracy of the GAST process for these larger distances. 10.1371/journal.pgen.1000255.t002 Table 2 Comparison of taxonomic assignments using full-length and in silico generated V3 and V6 hypervariable region tags. Human Gut Microbiome V3 region Deep-sea Vents Microbiome V3 region Count Same %Same Count Same %Same Superkingdom 7208 7208 100.00% / 100.00% 963 963 100.00% / 100.00% Phylum 7168 7145 99.68% / 99.82% 920 901 97.93% / 97.68% Class 7033 7002 99.56% / 99.52% 884 857 96.95% / 99.55% Order 7019 6987 99.54% / 99.57% 807 784 97.15% / 100.00% Family 5726 5688 99.34% / 98.93% 764 746 97.64% / 98.44% Genus 5178 5153 99.52% / 97.99% 701 686 97.86% / 99.10% Human Gut Microbiome V6 region Deep-sea Vents Microbiome V6 region Count Same %Same Count Same %Same Superkingdom 7215 7215 100.00%/ 100.00% 1058 1058 100.00% / 99.53% Phylum 7175 7152 99.68% / 99.82% 1008 986 97.82% / 90.00% Class 7040 7009 99.56% / 99.52% 970 939 96.80% / 90.08%% Order 7026 6994 99.54% / 99.56% 881 847 96.14% / 87.06% Family 5731 5693 99.34% / 98.93% 833 814 97.72% / 87.23% Genus 5183 5158 99.52% / 97.97% 766 749 97.78% / 88.52% Treating the V3 and the V6 regions independently, we counted the number of assignments the GAST process made at each taxonomic rank and the number and percent of times those assignments were the same as the assignment given to the full-length source sequence. The second percent value is the rate at which the top BLAST match predicted the same assignment as the full-length source. We compared the use of GAST to assign taxonomy with the use of the top BLAST match (Table 2). While the top BLAST match was consistent with GAST to the order level, it was less accurate for family and genus for the V6 region, especially for the deep-sea vent sequences. In addition, the top BLAST match was often not the best GAST match in cases where the taxonomic assignment was the same (Table 3). A comparison of the BLAST rank vs. GAST distance did not show any significant correlation (results not shown). 10.1371/journal.pgen.1000255.t003 Table 3 BLAST ranks for top GAST hits. Human Microbiome Deep-Sea Vents BLAST Rank V3 region V6 region V3 region V6 region 1 83.70% 94.90% 75.17% 81.89% 2 13.48% 4.34% 5.27% 7.16% >2 2.82% 0.76% 19.57% 10.95% The percentages reported represent the frequency with which the top GAST match corresponds to the top BLAST match, the second best BLAST match, or any other BLAST match. Comparison of Population Sampling Using Sanger-Generated Full-Length and Pyrosequencing-Generated V3 and V6 Tags for the Human Gut Microbiome Dataset We compared taxonomic assignments and their frequencies for the human gut microbiome data sampled with full-length SSU rRNA (n = 7215, length = 1300–1450 nt), V3 tags (n = 422,992, trimmed length = 100–200 nt) and V6 tags (n = 441,894, trimmed length = 50–70 nt) [30]. Full-length sampling detected a total of 43 genera, V3 pyrosequencing detected 116 genera and V6 pyrosequencing detected 103 genera (Figure 2). V3 sampling detected 74 genera that were not detected using the full-length sequencing and V6 sampling detected 60 genera (102 different genera combined). No genera were detected by the full-length sequencing alone. V3 sampling missed one taxon represented in the full-length sequences: Escherichia, detected 5 times with the full-length and once with the V6. V6 sampling missed three taxa, Hespellia, Klebsiella, and TM7 detected by the full-length with only one sequence each (detected 12, 16 and 4 times with V3, respectively). Only the TM7 sequences could be affected by a primer bias; the other two taxa have exact matches to forward and reverse primers as represented in the reference database of full-length SSU rRNA. The TM7 includes several different primer region sequences in the reference database each of which is one or two bases from its nearest primer, but only at the 5′ end, which should still detect abundant genera. The lack of TM7 in the V6 pyrosequencing data could be caused by a primer bias acting on a rare population, or could simply be the undersampling of rare organisms. In sum, less than 1% of the taxa identified by full-length sequences, representing 1) that were lost in the full-length sequencing (y = 0). The x-intercept comparing the two tag sequencing experiments (Figure 3C) is at the origin, implying that the two hypervariable regions were comparable in elucidating the rare biosphere. Primer bias was shown to be negligible in most cases, and undersampling is the most likely cause for the differences in small populations detected. Both tag sequencing experiments found similar numbers of taxa and substantially more than in the full-length sequencing. For tag sequencing of hypervariable regions to be effective for mapping taxonomy, specific sequences must match unambiguously to source organisms. If it were common for two divergent organisms to have the same or highly similar V3 or V6 regions, tag sequencing would not be an accurate means for assigning taxonomy. In our reference database of over 500,000 rRNA sequences, we found that hypervariable region tags map to individual taxa with high fidelity. In only a few cases for either the V3 or the V6 region did we find sequences that exactly match two or more distinct taxa (Table 1). The reference databases are replete with highly studied bacteria, and multiple copies of a hypervariable region for these organisms do not imply any taxonomic ambiguity. The V3 region was 99% accurate to the genus level. The V6 region was only slightly less resolved than the V3, still providing a 97% accuracy in assignment to the genus level, 98+% accuracy to the level of family, and 99% accuracy at the level of order. The V3 region is longer than the V6, which should have a positive effect on its specificity. The RefV6 database contains about half the number of reference sequences as the RefV3, and the accuracy of the V6 may be even greater as the reference databases grow. These levels of accuracy show that hypervariable region tags contain adequate information to accurately map taxonomy of both bacterial and archaeal organisms. To assign taxonomy to our long SSU rRNA gene sequences, both in our RefSSU database and in our experimental sequences, we used the Ribosomal Database Project classifier (RDP). RDP is not 100% accurate and some of the ambiguities in the reference database could be attributable to limitations with the RDP classifier. For tag sequencing with the GAST process, however, we are assessing the utility of a tag as a surrogate for the longer SSU rRNA sequences via a look-up and distance matrix. Can we consistently assign the correct taxonomy to both a SSU rRNA sequence and its constituent hypervariable regions independently? The RDP taxonomy provides a consistent and high-quality taxonomic classification (Bergey's taxonomy), facilitating our analysis. Slight inaccuracies in RDP are not an important factor in whether a tag sequence can be used as a surrogate for a full-length SSU rRNA sequence. We conducted an in silico experiment assessing taxonomic assignment using tags of hypervariable regions extracted from full-length sequences and compared the tags directly to the full-length sequences. We used two independent datasets, a human gut microbiome SSU rRNA dataset, which should be relatively well represented in the reference databases of SSU rRNA genes, and one of deep-sea vent microbes, which are less well studied and therefore less well represented in the reference databases. We examined the use of both the V3 and the V6 region tags to assign taxonomy for both groups of microbes. In all four cases, we found excellent correspondence between the use of the GAST process for assigning taxonomy to short hypervariable region tags and the use of RDP for assigning taxonomy to the full-length sequences. For both variable regions from the human microbiota, the taxonomic assignments of the tags agreed with the long sequences at a rate consistently greater than 99%. The deep-sea vent data agreed in over 97% of instances at the genus level. The two variable regions mapped taxonomy with virtually identical fidelity. The greater difference was not in choice of variable region, but in the environment examined. Of the human gut microbes sampled, 91% of all tags had exact matches in the reference database and virtually all had matches within a 10% sequence match. Of the deep-sea vent microbes, which have not historically been as well studied, only 51% had exact matches in the reference database and only 90% were within a 10% sequence match of a reference sequence. Despite the fact that as many as 10% of the deep-sea vent tags did not have a close match in the reference database, the GAST process only mis-assigned 3% of the tags. The GAST process accurately assigns taxonomy to tags diverging as much as 15% from their nearest reference match (Figure 1). Although our data included insufficient tags with GAST distances greater than 0.15 to fully assess their accuracy, the ability to transfer taxonomic information from the reference database is will certainly be less accurate at greater GAST distances. The deep-sea vent mismatch rate was similar for both hypervariable regions and at each taxonomic level from genus to phylum, and less than one-third of the mismatches were for tags >15% divergent. A possible explanation is the deep-sea vent environment may contain phyla that are not yet adequately described, and whose full-length SSU rRNA sequences are therefore not well classified by the RDP. These results imply that hypervariable region tag sequencing and the GAST process are excellent tools for assigning taxonomy, but they cannot overcome basic gaps in knowledge of the under-explored areas of the microbiome. As more is learned about these organisms, and their full-length SSU rRNA genes are added to the reference databases, hypervariable region tags sequencing projects will directly benefit, and taxonomy will improve. The top BLAST matches for the human gut microbiome mapped taxonomy better than the top BLAST matches for the deep-sea vents. This is likely because the human microbiome data are better represented in the reference database. Since 91% of the human microbiome tags had exact matches in the reference database these should be consistently identified by BLAST as the best match. For the deep-sea samples where only 51% had exact matches, a top BLAST hit may find only a local match within the sequence rather than a globally-weighted match as with GAST. Reliance on local alignments can be misleading: 80% of the deep-sea vent V6 tags that did not rank among the top two BLAST hits showed a BLAST alignment length shorter than the tag sequence for the top BLAST hit. BLAST failed to identify the correct genus for 11% of these V6 tag sequences, whereas GAST which failed for only 2%. An increase in distance from the tag to the nearest reference sequence using GAST did not correlate with a lower BLAST rank. The magnitude of divergence from the reference database does not explain the difference between V3 and V6 regions. The V3 tags were noticeably more divergent from their top BLAST hit than were the V6 tags, despite the larger dataset of reference V3 sequences. This did not adversely affect the ability to identify tags to the genus level. Since the RDP taxonomy is restricted to the genus level, we could not review the BLAST ranks for species-level. Full-length sequencing missed 63% of the genera identified by V3 and 58% of the genera identified by V6 (for more details, see Dethlefsen et al. [30]). The full-length sequencing uncovered only four rare taxa missed by one or the other hypervariable region, but no taxa missed by both of the hypervariable regions. Primer mismatches were minimal, but may be relevant when combined with a very low abundance. The hypervariable region tag sequencing did not introduce any strong biases against the discovery of common taxa or the relative abundance of these taxa in this experiment. As predicted, the hypervariable region tag sequencing provided a much greater breadth and depth of sampling. Although the level of sampling with tag sequencing is orders of magnitude greater than with traditional methods, a single pyrosequencing run (with >400,000 sequences) is still insufficient to fully sample the rare biota in the human distal gut. The sampling limitation of this experiment can be seen in the small but distinct number of taxa that appeared in one but not the other tag sequencing experiment. All were of low abundance and are dispersed throughout the microbial world rather than clustering in one specific taxon. Since no common taxa were omitted by sequencing of either hypervariable region, any effects of primer bias are limited to rare taxa and cannot be discerned from the effects of undersampling. Other methods such as Greengenes [31] and SeqMatch [32] use short tag sequences to determine phylogenetic affinity through comparisons to reference data sets of nearly full-length SSU rRNA sequences without requiring the inference of phylogenetic trees from the hypervariable regions. Greengenes (http://greengenes.lbl.gov) uses NAST [33] to align tags by inserting them into a pre-existing database of >10,000 aligned full-length sequences, and then assigns taxonomy based on the nearest neighbor in the database. Liu et al [34] used NAST on simulated tag sequences and found similar Unifrac clustering results from the tags as from their full-length source sequences. The Greengenes website is limited to 500 sequences and uses a database of only 10,000 sequences for comparison and does not perform a consensus of multiple taxonomy matches. SeqMatch uses a k-nearest-neighbor, word-matching algorithm rather than a multiple sequence alignment to display nearest matches in the RDP dataset, and uses the lowest common taxon for consensus (essentially a unanimous consensus rather than 66% used by GAST). The RDP dataset is smaller than SILVA database but more selective. The website tool is limited to 2,000 sequences. Conclusions/Significance Hypervariable region tag sequencing using either the V3 or the V6 region, and presumably other hypervariable regions, is an effective means for assigning taxonomy and provides great advantages over traditional sampling. A tag mapping process such as GAST with an extensive database of rRNA genes such as our RefSSU derived from SILVA can map tag sequences to the same taxonomy as their source genes at better than a 99% correlation rate for commonly studied environments such as the human microbiome and better than 96% for less commonly studied environments such as deep-sea vents. The V3 and V6 regions have only minimal ambiguity in mapping to the SSU rRNA gene all the way to the genus level. While tags can map to more than one SSU rRNA source, these cross-mappings between taxa are infrequent and do not compromise the overall methodology. We show that these short hypervariable region tags contain adequate information to uniquely and accurately map the phylogeny with a 98% or greater fidelity even without an exact match in the reference database or with potential multiple copies in the database. GAST is accurate for tags as much as 15% divergent from their nearest reference match, although there were very few tags that far from the current set of reference SSU rRNA sequences. The GAST distance, like BLAST scores or e-values, should be maintained and used as an assessment for the likely reliability of the GAST assignment of more divergent tags. The consistently high correspondence of the hypervariable region tags vs. long SSU rRNA taxonomies shows the robustness of the GAST process and the use of tags as surrogates for the full-length rRNA genes even for microbial environments that are not well-represented in the reference databases. Massively-parallel pyrosequencing of tags can be used to great advantage over traditional sequencing of full-length rRNA genes to explore both the diversity and relative abundance of microbial populations. Further research into hypervariable region tag sequencing may uncover advantages of one region over another, such as the relative levels of microvariation, length of sequence, density of homopolymers (which can lead to pyrosequencing errors), ability to identify to the species level, or the merits of different amplification primers. Tag sequencing yields similar taxonomy and relative abundance values as conventional sequencing of full-length SSU rRNA genes, but provides more reads, uncovers more organisms, avoids assembly, and costs less per read than conventional sequencing of full-length SSU rRNA genes. As the technology continues to improve, yielding greater read counts and longer sequences, pyrosequencing will provide even greater opportunities for tag sequencing, such as the use of longer hypervariable regions or combinations of variable regions, and ever-greater sampling depth. This process will also improve as reference databases of SSU rRNA genes continue to grow. The great advantage of hypervariable region tag sequencing is that it can take advantage of massively-parallel pyrosequencing, sampling to depths several orders of magnitude greater than previously achieved, facilitating the exploration of the vast diversity of microbial populations and the rare biosphere. Materials and Methods Creating the Reference Database of Full-Length, V3, and V6 SSU rRNA Sequences We downloaded 503,971 aligned small subunit rRNA sequences from the SILVA database, version 92 [35]. Using the SILVA quality assessments, we eliminated low-quality sequences (sequence quality = 80%. If the bootstrap value was = 80) and generated a consensus taxonomy. If two-thirds or more of the full-length sequences shared the same assigned genus, the tag was assigned to that genus. If there was no such agreement, we proceeded up one level to family. If there was a two-thirds or better consensus at the family level, we assigned this taxonomy to the tag, and if not, we continued to proceed up the tree. Occasionally, a tag could not be assigned taxonomic classification at the domain level. This was because the RDP Classifier could not assign a domain with an adequate bootstrap value, rather than a tag mapping to full-length sequences from different domains. These may represent novel organisms whose taxonomy has not yet been determined. Sample tags that did not have a single BLAST match in the RefSSU database also were not given a taxonomic assignment. We chose to use a 66% (two-thirds) majority although other values or a distributional vs. strict percentage approach can be implemented. We reviewed nearly 17 million tags in our sequencing database (primarily of the V6 region) from a wide range of studies using the 66% majority as the threshold for assignment. A distribution curve of voting majority did not show any obvious break points (graph not shown), although 95% of the tags had a voting majority of 75% or better, and 90% had a voting majority > = 83%.
- Record: found
- Abstract: found
- Article: found
Mom Knows Best: The Universality of Maternal Microbial Transmission
Summary The sterile womb paradigm is an enduring premise in biology that human infants are born sterile. Recent studies suggest that infants incorporate an initial microbiome before birth and receive copious supplementation of maternal microbes through birth and breastfeeding. Moreover, evidence for microbial maternal transmission is increasingly widespread across animals. This collective knowledge compels a paradigm shift—one in which maternal transmission of microbes advances from a taxonomically specialized phenomenon to a universal one in animals. It also engenders fresh views on the assembly of the microbiome, its role in animal evolution, and applications to human health and disease. Introduction While the human microbiota comprises only 1–3% of an individual's total body mass, this small percentage represents over 100 trillion microbial cells, outnumbering human cells 10 to 1 and adding over 8 million genes to our set of 22,000 [1],[2]. This complexity establishes a network of interactions between the host genome and microbiome spanning gut development [3], digestion [4],[5], immune cell development [6]–[8], dental health [9],[10], and resistance to pathogens [11],[12]. Recent studies have also provided a greater understanding of how the composition of an individual's microbiota changes throughout development, especially during the first year of life [3],[13]. While the general dogma is that the placental barrier keeps infants sterile throughout pregnancy, increasing evidence suggests that an infant's initial inoculum can be provided by its mother before birth [14]–[18] and is supplemented by maternal microbes through the birthing [19] and breastfeeding [20],[21] processes. While maternal transmission of microbes in humans has attracted considerable attention in the last few years, nearly a century's worth of research is available for vertical transmission of symbionts in invertebrates [22]. Similar to gut bacteria in humans that assist nutrient intake, many insect-associated bacteria function as nutritional symbionts that supplement the nutrient-poor diet of their host with essential vitamins or amino acids [23],[24]. Since these indispensable symbionts cannot live outside of host cells, they cannot be acquired from the environment and are faithfully transferred from mother to offspring [22],[25]. Maternal transmission in invertebrates has been reviewed elsewhere [22],[26],[27], and Box 1 and Box 2 highlight examples of heritable symbioses across invertebrate phyla. Box 1. Examples of Maternal Transmission in Marine Invertebrates Marine Sponges (Phylum Porifera) Sponges are ancient metazoans that evolved over 600 million years ago as one of the first multicellular animals [83]. In marine sponges, a remarkably large consortium of extracellular microbial symbionts thrives within the sponge's mesohyl, a gelatinous connective tissue located between the external and internal cell layers. Many of these bacterial residents are found in diverse species of sponges with nonoverlapping distributions but not in the surrounding seawater [84]–[86]. These “sponge-specific” microbes are hypothesized to have originated from ancient colonization events before the diversification of marine sponges and are maintained as symbionts through vertical transmission [87]. Independent studies have estimated that up to 33 phylogenetically distinct microbial clusters spanning ten bacterial phyla and one archaeal phylum are vertically transmitted in sponges [41],[84],[86],[88]. Both transmission electron microscopy (TEM) and fluorescent in situ hybridization (FISH) studies have confirmed the presence of microorganisms of different shapes and sizes in the oocytes of oviparous sponges [41] and in the embryos of viviparous sponges [43]–[45]. Vesicomyid Clams (Phylum Mollusca) Deep-sea hydrothermal vent communities rely upon chemosynthetic bacteria to harness chemical energy stored in reduced sulfur compounds extruding from the vents. Metazoans that live in this extreme environment harbor chemosynthetic endosymbionts in their tissues that provide most, if not all, of the host's nutrition [89]. Somewhat surprisingly, most invertebrates that live near hydrothermal events acquire their endosymbionts anew from the environment each generation [90],[91], even though chemosynthetic bacteria are crucial for survival in such a harsh habitat. A major exception to this trend is found in the Vesicomyidae family of clams [92]. Vesicomyid clams retain a rudimentary gut and rely primarily on sulfur-oxidizing bacteria sequestered intracellularly within specialized host cells called bacteriocytes in the clam's large, fleshy gills [93]. Vertical transmission via transovarial transmission appears to be the dominant mechanism for maintenance of these thioautotrophic bacterial symbionts given that follicle cells surrounding an oocyte and the oocyte itself are heavily infected with the chemosynthetic bacteria [46],[94]. Box 2. Examples of Maternal Transmission in Terrestrial Invertebrates Insects (Phylum Arthropoda) Insects that thrive on unbalanced diets such as plant sap, blood, or wood depend upon microbial symbionts for the provision of essential amino acids or vitamins lacking in their food source. In turn, hosts provide a wide range of metabolites to their symbionts as well as protection from environmental stressors. This codependence requires faithful transfer of symbionts to all offspring, usually through transovarial transmission [23],[24]. Reproductive parasites, such as the obligate, intracellular bacteria Wolbachia, are also widespread in insects and hijack maternal transmission routes to ensure their spread within an insect population (reviewed in [95],[96]). Pea Aphid (Acrythosiphon pisum) The pea aphid Acrythosiphon pisum (Figure 2A) and its nutritional endosymbiont Buchnera aphidicola are a preeminent example of obligate mutualism in insects. The ancestral Buchnera gammaproteobacteria was acquired by aphids between 160 and 280 million years ago [97] and has since diverged in parallel with its aphid hosts through strict vertical transmission [26],[97]. Buchnera are housed within the cytoplasm of bacteriocytes arranged into dual bacteriome structures located in the aphid body cavity adjacent to the ovaries [98], allowing efficient transfer of Buchnera symbionts to developing oocytes or embryos during the sexual and asexual phases of aphid reproduction, respectively. At the cellular level, symbiont transfer occurs when maternal bacteriocytes release Buchnera symbionts through exocytosis into the extracellular space between the bacteriocyte and oocyte or embryo, which then actively endocytoses the extracellular Buchnera symbionts [51]. Cockroaches (Order Blattodea) Just as insects are morphologically diverse, the mechanisms by which insects transport symbionts to oocytes are highly varied. In cockroaches, Blattabacterium-filled bacteriocyte cells migrate from the abdominal fat body to the distantly located ovarioles where they adhere to the oocyte membrane [99],[100]. Interestingly, the bacteriocytes remain associated with the oocyte for eight to nine days before finally expelling their symbionts through exocytosis. The Blattabacterium cells then squeeze between the follicle cells surrounding the oocyte and are engulfed into the oocyte cytoplasm via endocytosis just prior to ovulation [100]. Whiteflies (Family Aleyrodidae) The whitefly circumvents exocytosis of its intracellular nutritional symbiont, Portiera aleyrodidarum, by depositing entire bacteriocytes into its eggs. These maternal bacteriocytes remain intact yet separate from the developing embryo until the embryonic bacteriomes form, at which point the maternal bacteriocytes deteriorate [22]. Tsetse Flies, Bat Flies, and Louse Flies (Superfamily Hippoboscoidea) Members of the Hippoboscoidea superfamily (Order Diptera) are obligate blood feeders that have developed a unique reproductive strategy termed adenotrophic viviparity that offers a different solution to internal maternal transfer of symbionts. Females of this superfamily develop a single fertilized embryo at a time within their uterus (modified vaginal canal) until it is deposited as a mature third instar larva immediately preceding pupation. During their internal development, the larvae are nourished with milk produced by modified accessory glands that empty into the uterus [101]. The milk primarily consists of protein and lipids [102], but it also serves as a reservoir for maternally transmitted microbial symbionts [103]. For example, the obligate mutualistic symbiont of tsetse flies, Wigglesworthia glossinidia, is absent from the female germ line and surrounding reproductive tissues but is found extracellularly in the female milk glands and is first detected in tsetse offspring once milk consumption begins during the first larval stage [103]. Stinkbugs (Superfamily Pentatomoidea) One of the most common mechanisms of external maternal transmission in insects is that of “egg smearing,” which occurs when a female contaminates the surface of her eggs with symbiont-laden feces during oviposition. Upon hatching, offspring probe or consume the discarded egg shells to acquire the maternal bacteria. This mode of transmission is commonly found in plant-sucking stinkbugs, including the Pentatomidae and Acanthosomatidae families [104]. In the Cynidae family of stinkbugs, along with the Coreidae family of leaf-footed bugs, gut symbionts are transferred maternally via coprophagy, in which offspring consume maternal feces, sometimes directly from the mother's anus [22],[104]. Stinkbugs of the Plataspidae family, on the other hand, have developed a unique mode of transmission via a maternally provided “symbiont capsule” deposited on the underside of the egg mass [56]. These capsules are comprised of bacterial cells dispersed throughout a resin-like matrix surrounded by a brown, cuticle-like envelope that protects the symbionts from environmental stressors such as UV irradiation or dissection [57]. After hatching, plataspid nymphs immediately probe the capsules to ingest the symbionts [56],[59]. European Beewolf (Philanthus triangulum) While nutritional symbionts appear to be the most common type of bacteria transmitted via external maternal transmission in insects, the European beewolf (Philanthus triangulum) instead cultivates a symbiotic bacteria that protects offspring against microbial infection during development. Beewolves are solitary digger wasps that deposit their offspring in moist, underground nests, making them susceptible to fungal and bacterial infections [105]. To combat these pathogens, female beewolves cultivate Streptomyces philanthi bacteria in specialized glands in their antennae, which they copiously spread on the ceiling of the brood cell before oviposition [106]–[108]. After hatching, the larvae take up the bacterial cells and incorporate them into their cocoon that they build before pupation. When adult beewolves emerge from their cocoon in the summer, female beewolves acquire the maternally provided Streptomyces symbiont and house them in the female-specific gland reservoirs along each antenna [108],[109]. By integrating previous studies in invertebrates with recent evidence for maternal microbial transmission in humans and other vertebrates, we contend that maternal provisioning of microbes is a universal phenomenon in the animal kingdom. As a result, a considerable new phase of study in heritable symbiont transmission is underway. Thus, this essay presents current evidence for maternal microbial transmission and provides new insights into its impact on microbiome assembly and evolution, with applications to human health and disease. Internal Maternal Transmission At the turn of the twentieth century, French pediatrician Henry Tissier asserted that human infants develop within a sterile environment and acquire their initial bacterial inoculum while traveling through the maternal birth canal [28]. More than a century later, the sterile womb hypothesis remains dogma, as any bacterial presence in the uterus is assumed to be dangerous for the infant. Indeed, studies of preterm deliveries have found a strong correlation between intrauterine infections and preterm labor, especially when birth occurs less than 30 weeks into the pregnancy [29],[30]. Since preterm birth is the leading cause of infant mortality worldwide [31], much attention has focused on identifying the bacterial culprits responsible for spontaneous preterm labor. Surprisingly, most of the bacteria detected in intrauterine infections are commonly found in the female vaginal tract [29], and risk of preterm birth is markedly increased in women diagnosed with bacterial vaginosis during pregnancy [32]. Interestingly, the vaginal microbial community varies significantly among American women of different ethnicities (Caucasian, African-American, Asian, or Hispanic), with African-American and Hispanic women more likely to have a microbiota traditionally associated with bacterial vaginosis (predominance of anaerobic bacteria over Lactobacillus species) [33] and a higher rate of spontaneous preterm deliveries (reviewed in [34]). While intrauterine infection and inflammation is important in understanding the etiology of preterm birth, relatively few studies have examined the uterine microbiome of healthy, term pregnancies owing to the sterile womb paradigm. Investigations into the potential for bacterial transmission through the placental barrier have detected bacteria in umbilical cord blood [17], amniotic fluid [14],[35], and fetal membranes [35],[36] from babies without any indication of inflammation (Figure 1). Furthermore, an infant's first postpartum bowel movement of ingested amniotic fluid (meconium) is not sterile as previously assumed, but instead harbors a complex community of microbes, albeit less diverse than that of adults [18],[37]. Interestingly, many of the bacterial genera found in the meconium, including Enterococcus and Escherichia, are common inhabitants of the gastrointestinal tract [18],[37]. To test whether maternal gut bacteria can be provisioned to fetuses in utero, Jiménez et al. [18] fed pregnant mice milk inoculated with genetically-labeled Enterococcus faecium and then examined the meconium microbes of term offspring after sterile C-section. Remarkably, E. faecium with the genetic label was cultured from the meconium of pups from inoculated mothers, but not from pups of control mice fed noninoculated milk. Meconium from the treatment group also had a higher abundance of bacteria than that of the control group. Importantly, the study controlled for potential bacterial contamination from contact between skin and the meconium by sampling an internal portion of the meconium [18]. Thus, this study provides foundational evidence for maternal microbial transmission in mammals. 10.1371/journal.pbio.1001631.g001 Figure 1 Sources of microbial transmission in humans from mother to child. Cut-away diagram highlighting the various internal and external sources of maternal microbial transmission as well as the species that are commonly associated with transfer from those regions. Regions discussed include the oral cavity [14],[16], the mammary glands [40],[78],[79], the sebaceous skin surrounding the breast [78],[80], the vaginal tract [19],[33],[73], and the intrauterine environment [14],[15],[17],[18],[29],[35]–[37],[40]. Illustration by Robert M. Brucker. Other than ascension of vaginal microbes associated with preterm births, the mechanisms by which gut bacteria gain access to the uterine environment are not well understood. One possibility is that bacteria travel to the placenta via the bloodstream after translocation of the gut epithelium. While the intestinal epithelial barrier generally prevents microbial entry into the circulatory system, dendritic cells can actively penetrate the gut epithelium, take up bacteria from the intestinal lumen, and transport the live bacteria throughout the body as they migrate to lymphoid organs [38],[39]. Interestingly, microbial translocation may even increase during pregnancy, as one study showed that pregnant mice were 60% more likely to harbor bacteria in their mesenteric lymph node (presumably brought there by dendritic cells) than nonpregnant mice [40]. Bacterial species normally found in the human oral cavity have also been isolated from amniotic fluid and likely enter the bloodstream during periodontal infections, facilitated by gingiva inflammation [14],[16] (Figure 1). Overall, the study of internal maternal transmission of microbes in mammals is in its infancy due to the enduring influence of the sterile womb paradigm and to the ethical and technical difficulties of collecting samples from healthy pregnancies before birth. Thus, we still know very little about the number and identity of innocuous microbes that traverse the placenta, whether they persist in the infant, or whether their presence has long-term health consequences for the child. Similarly, we know almost nothing about nonpathogenic viruses or archaea that may be transferred from mother to child alongside their bacterial counterparts. Fortunately, the advent of culture-independent, high-throughput sequencing will serve as a tremendous resource for this field and will hopefully lead to a characterization of the “fetal microbiome” in utero. Maternal provisioning of microbes to developing offspring is widespread in animals, with evidence of internal microbial transmission in animal phyla as diverse as Porifera [41]–[45] (Box 1), Mollusca [46]–[49] (Box 1), Arthropoda [50]–[52] (Box 2, Figure 2), and Chordata [19],[53],[54] (Box 3, Figure 2). The presence of maternal transmission at the base of the Animalia kingdom and the surprising plasticity by which microbes gain access to germ cells or embryos in these systems signifies that maternal symbiont transmission is an ancient and evolutionarily advantageous mechanism inherent in animals, including humans. Therefore, we can no longer ignore the fact that exposure to microbes in the womb is likely and may even be a universal part of human pregnancy, serving as the first inoculation of beneficial microbes before birth. 10.1371/journal.pbio.1001631.g002 Figure 2 Examples of animals that exhibit microbial maternal transmission. (A) Pea aphid (Acyrthosiphon pisum), photo credit: Whitney Cranshaw, Colorado State University/©Bugwood.org/CC-BY-3.0-US; (B) Domesticated chicken hen (Gallus gallus domesticus), photo credit: Ben Scicluna; (C) Sockeye salmon (Oncorhynchus nerka), photo credit: Cacophony; (D) South American river turtle (Podocnemis expansa), photo credit: Wilfredor. All photos were obtained from Wikimedia Commons (www.commons.wikimedia.org). Box 3. Examples of Maternal Transmission in Vertebrates Aside from studies in human and mouse models, very little is known about maternal transmission of microbial communities in vertebrates, especially outside Class Mammalia. Furthermore, research on vertical transmission in nonmammalians has largely focused on maternally transmitted pathogens, especially in animals of agricultural importance like chickens and fish. Domesticated Chickens (Gallus gallus domesticus) Zoonotic Salmonella infections acquired from contaminated chicken eggs is estimated to cause more than 100,000 illnesses each year in the United States [110]. In addition to horizontal transmission of Salmonella on eggs through surface contamination, direct transovarial transmission also occurs when Salmonella colonizes the reproductive tissues of hens (Figure 2B). Depending on the infection location within the female reproductive tract, the bacteria are deposited into the yolk, albumen, eggshell membrane, and/or eggshell of the developing egg before oviposition (reviewed in [111]). Other poultry pathogens, such as Mycoplasma synoviae in chickens [112] and M. gallisepticum, M. cloacale, and M. anatis in ducks [113], have also been cultured from the yolk of embryonated eggs, though whether commensal flora are incorporated into the egg is not known. Ray-Finned Fish (Class Actinopterygii) Several bacterial pathogens of economically important fish are transmitted transovarially in the egg yolk including Renibacterium salmoninarum, the agent of bacterial kidney disease in salmonids (Figure 2C), and Flavobacterium psychrophilum, which causes bacterial cold water disease in salmonids and rainbow trout fry disease in trout (reviewed in [114]). F. psychrophilum has also been found in ovarian fluid and on the surface of eggs of steelhead trout [115]. Additionally, an obligate, intracellular eukaryotic parasite, Pseudoloma neurophilia, is a common pathogen found in zebrafish (Danio rerio) facilities and has been observed in spores of the ovarian stroma and within developing follicle cells of spawning females, suggesting that it can be vertically transmitted, though it is primarily spread from fish to fish in contaminated water (reviewed in [116]). Turtles (Order Chelonii) The formation of egg components in the uterine tube and uterus of turtles takes approximately two weeks, providing ample opportunity for maternal transmission of intestinal or reproductive microbes to the egg [117]. One study of unhatched (dead) eggs from loggerhead sea turtle (Caretta caretta) nests found several potential pathogens, including Pseudomonas aeruginosa and Serratia marcesans, in fluid from the interior of the eggs, though environmental contamination of the eggs cannot be ruled out [118]. A similar study of eggs from two species of South American river turtles, Podocnemis expansa (Figure 2D) and P. unifilis, identified several Enterobacteriaceae species, including Escherichia coli, Shigella flexneri, and Salmonella cholerasuis, in the eggs but not in the environmental samples taken from the turtle nests [119], suggesting that they may have a maternal origin. In support of this hypothesis, a separate study in green turtles (Chelonia mydas) that collected eggs directly from the maternal cloacal opening during egg laying isolated Pseudomonas, Salmonella, Enterobacter, and Citrobacter from the eggshell, albumen, and yolk. In fact, the yolk was the egg component most heavily infected with bacteria [120]. Altogether, many potentially pathogenic species have been isolated from turtle eggs, but whether these bacteria actually cause disease in turtles or are part of their natural flora remains to be determined. External Maternal Transmission External maternal transmission encompasses any transfer of maternal symbionts to offspring during or after birth. In invertebrates, it is often accomplished by “egg smearing,” in which females coat eggs with microbes as they are deposited [55], or through the provision of a microbe-rich maternal fecal pellet that is consumed by larval offspring upon hatching [56]–[59] (see Box 2). Similarly, human infants are “smeared” with maternal vaginal and fecal microbes as they exit the birth canal [60]–[62] (Figure 1). Several studies have shown that the human neonatal microbiota across all body habitats (skin, oral, nasopharyngeal, and gut) is influenced by their mode of delivery [19],[63]–[65], with infants born vaginally acquiring microbes common in the female vagina while C-section infants display a microbiota more similar to that of human skin [19]. Furthermore, while the microbiota of a vaginally delivered infant clusters with the vaginal bacteria of its mother, the microbiota of C-section babies is no more related to the skin flora of its mother than that of a stranger, indicating that most microbes are transmitted to the neonate from those handling the infant [19]. Importantly, epidemiological data suggest that a Cesarean delivery can have long-term consequences on the health of a child, especially concerning immune-mediated diseases. For example, children born via C-section are significantly more likely to develop allergic rhinitis [66], asthma [66], celiac disease [67], type 1 diabetes [68], and inflammatory bowel disease [69]. These statistics are alarming given that 32.8% of all births in the United States in 2010 were delivered via C-section with similar rates on the rise in most developed countries [70]. The higher rate of immune-mediated diseases in C-section children may indicate that maternally transferred vaginal or fecal microbes are unique in their ability to elicit immune maturation in the neonate. Development of the intestinal mucosa and secondary lymphoid tissues in the gut is contingent upon recognition of microbial components by pattern-recognition receptors on intestinal epithelial cells (reviewed in [71],[72]). It is possible that these receptors cannot properly interact with the community of microbes acquired during Cesarean deliveries, leading to disrupted immune development and an increased risk for immune-mediated disorders in C-section children. Conversely, transmission for thousands of years of vaginal and fecal microbes at birth has likely produced specific human-microbe interactions important for neonatal gut development. In fact, a recent study found that the vaginal microbial community changes during pregnancy, becoming less diverse as the pregnancy progresses [73]; yet, in spite of the general decrease in richness, certain Lactobacillus bacterial species are enriched in the vaginal community during pregnancy and are hypothesized to be important for establishing the neonatal upper GI microbiota after vaginal delivery [73]. Breastfeeding provides a secondary route of maternal microbial transmission as shown in humans (reviewed in [74], Figure 1) and nonhuman primates such as rhesus monkeys [75]. In humans, maternal milk microbes are implicated in infant immune system development [76], resistance against infection [77], and protection against the development of allergies and asthma later in childhood [74]. High-throughput sequencing of breast milk from 16 healthy women identified 100–600 species of bacteria in each sample with nine genera present in every sample: Staphylococcus, Streptococcus, Serratia, Pseudomonas, Corynebacterium, Ralstonia, Propionibacterium, Sphingomonas, and Bradyrhizobiaceae [78]. This “core” milk microbiome represented approximately 50% of all bacteria in each sample, with the other half representing individual variation in microbial composition [78]. A similar study found that the bacterial composition in breast milk changes over time: milk produced immediately after labor harbored more lactic acid bacteria along with Staphylococcus, Streptococcus, and Lactococcus, while breast milk after six months of lactation had a significant increase in typical inhabitants of the oral cavity, such as Veillonella, Leptotrichia, and Prevotella [79], perhaps to prime the infant for the switch to solid food. However, as with any DNA-based, culture-independent study that does not discriminate between live and dead bacteria, the number and identity of bacteria detected in these studies should be interpreted with some caution. Given that milk is only produced temporarily in a woman's life, the origin of milk microbes is still somewhat of a mystery. Breast milk was traditionally thought to be sterile; however, colostrum (the first milk produced after delivery) collected aseptically already harbors hundreds of bacterial species [79]. Breast milk does share many taxa with the microbiota found on sebaceous skin tissue around the nipple [78],[80], and high levels of Streptococcus in breast milk may be a result of retrograde flow from an infant's oral cavity back to the milk ducts during suckling [81] since Streptococcus is the dominant phylotype in infant saliva [82]. However, the presence of anaerobic gut bacteria in human milk suggests that an entero-mammary route of transfer also exists that may utilize phagocytic dendritic cells to traffic gut microbes to the mammary glands, similar to microbial transfer to amniotic fluid as discussed earlier. To support this hypothesis, Perez et al. [40] found identical strains of bacteria in milk cells, blood cells, and fecal samples from lactating women, but more work is needed to directly connect bacterial translocation in the gut to incorporation in breast milk. Overall, maternal transmission of beneficial microbes in humans has widespread relevance for human health. Evolution with these microbes has resulted in our dependence on them for the proper maturation and development of the immune system and gastrointestinal tract. Somewhat paradoxically, modern medicine designed to prevent infant mortality (such as emergency Cesarean sections and formula feeding) has likely contributed to the rise in immune-mediated diseases in developed countries due to the inherent lack of exposure to maternal microbes associated with these practices. Fortunately, biomedicine is also making strides in finding effective probiotic supplements to promote immune development and ameliorate some of the risks that C-section or formula-fed infants face as children and adults. Hopefully, as we gain understanding of the diversity and function of maternally transmitted microbes in humans, more complete and effective probiotic blends will recapitulate the microbial communities found in vaginally delivered, breast-fed infants and restore the microbe-host interactions that humans depend upon for proper development. Conclusions Since the early twentieth century, the study of maternal microbial transmission has focused heavily on animal systems in which maternal transmission maintains sophisticated partnerships with one or two microbial species. However, with the development of high-throughput sequencing technologies, it is now possible to identify entire microbiomes that are transferred from mother to offspring in systems not traditionally considered to exhibit maternal transmission, such as humans. By expanding the definition of maternal transmission to include all internal and external microbial transfers from mother to offspring, we contend that maternal transmission is universal in the animal kingdom and is used to provision offspring with important microbes at birth, rather than leave their acquisition to chance. Finally, with microbes contributing 99% of all unique genetic information present in the human body, maternal microbial transmission should be viewed as an additional and important mechanism of genetic and functional change in human evolution. Similar to deleterious mutations in our genetic code, disruption of maternal microbial acquisition during infancy could “mutate” the composition of the microbial community, leading to improper and detrimental host-microbe interactions during development. Maternal transmission is also a key factor in shaping the structure of the microbiome in animal species over evolutionary time, since microbes that promote host fitness, especially in females, will simultaneously increase their odds of being transferred to the next generation. Thus, whether internal or external, the universality and implications of maternal microbial transmission are nothing short of a paradigm shift for the basic and biomedical life sciences.
- Record: found
- Abstract: found
- Article: found
Oligotyping: differentiating between closely related microbial taxa using 16S rRNA gene data
A. Eren, Lois Maignien, Woo Jun Sul … (2013)
Bacteria comprise the most diverse domain of life on Earth, where they occupy nearly every possible ecological niche and play key roles in biological and chemical processes. Studying the composition and ecology of bacterial ecosystems and understanding their function are of prime importance. High-throughput sequencing technologies enable nearly comprehensive descriptions of bacterial diversity through 16S ribosomal RNA gene amplicons. Analyses of these communities generally rely upon taxonomic assignments through reference data bases or clustering approaches using de facto sequence similarity thresholds to identify operational taxonomic units. However, these methods often fail to resolve ecologically meaningful differences between closely related organisms in complex microbial data sets. In this paper, we describe oligotyping, a novel supervised computational method that allows researchers to investigate the diversity of closely related but distinct bacterial organisms in final operational taxonomic units identified in environmental data sets through 16S ribosomal RNA gene data by the canonical approaches. Our analysis of two data sets from two different environments demonstrates the capacity of oligotyping at discriminating distinct microbial populations of ecological importance. Oligotyping can resolve the distribution of closely related organisms across environments and unveil previously overlooked ecological patterns for microbial communities. The URL http://oligotyping.org offers an open-source software pipeline for oligotyping.
Author and article information
Journal
Title:
Cell Reports
Abbreviated Title:
Cell Reports
Publisher:
Elsevier BV
ISSN
(Print):
22111247
Publication date Created:
July
2017
Publication date
(Print):
July
2017
Volume: 20
Issue: 2
Pages: 491-504
Article
DOI: 10.1016/j.celrep.2017.06.060
PMC ID: 5667669
PubMed ID: 28700948
SO-VID: 36099ba4-e7ba-44e2-8fdb-f14ee4443d56
Copyright © ©
2017
History
Data availability: