There were flies in space. We looked at their gut microbes. Because, why not?
The work presented here is the result of a serendipitous encounter between researchers involved in two independent projects that were winners of an a ISS Research Competition hosted by Space Florida. These projects were flown to the International Space Station as part of the payload on the SpaceX CRS-3 mission, which launched from Cape Canaveral on April 18, 2014. The goal of Project HEART FLIES was to examine the effects of microgravity on the heart development and function in five strains of Drosophila melanogaster, including three normal strains and two mutant strains that are predisposed to heart malfunction. Upon return to the Earth’s surface, these flies were dissected to assay heart function, morphology, and tissue-specific gene expression. Ground-based control populations were assayed at the same time. One of the goals of Project MERCCURI was to examine the growth kinetics of several bacterial strains that were collected from the surfaces of various public built environments throughout the United States, including sporting venues and national monuments. A second goal was to examine the microbial communities present on surfaces in the International Space Station, using DNA-based censusing techniques.
The fruit fly Drosophila melanogaster, has a storied history of over a century as a model organism for genetic and behavioral studies. In more recent decades, it has also been used as a model organism for immune system and cardiac function, and it is emerging as a model for host-gut microbiome interactions. While Drosophila cardiac anatomy is quite different from that of a human, it does share some functionality (e.g., ion channels) and many of the genes involved in building a heart are shared between flies and humans (Cammarato 2011)(Wolf 2011). As in humans, cardiac function declines over the life span of the fly (Ocorr 2007). However, the typical lifespan of a fly is 45-60 days, allowing for quick study of age-dependent decline in organ function. Drosophila melanogaster has also been a frequent flier to low Earth orbit, where it has been used to study the effects of spaceflight and microgravity on innate immunity (Taylor 2014) (Marcu 2011), DNA mutation (Vaulina 1982), and development (Marco 1992) (Abbott 1992).
Numerous recent studies have suggested a relationship between the animal gut microbiome and both brain (Foster 2013) and cardiac function (Vinje 2014), including both indirect (Reardon 2014) and direct (Lam 2012) effects on the risk and severity of heart attacks. As new correlations between the gut microbiome composition and disease states in various other organ systems emerge, it becomes increasinly important to take advantage of model systems in which correlations can be tested further for causation (Fritz 2013); see (Baxter 2014) for a good example. Currently, the ratio of mouse microbiome to Drosophila microbiome publications is greater than 25:1, but there are many advantages to the use of Drosophila as a model for microbiome studies (Erkosar 2014), including the relative ease and low-cost of (especially axenic) rearing (Charroux 2012)(Ridley 2013).
Here, we present the first look at the differences in microbiome composition, in the context of a controlled experiment, between animals reared in a laboratory on the International Space Station and reared in a laboratory on the surface of the Earth. To do this, we employed 16S rDNA PCR surveys of dissected fly guts, swabs of feces from the surface of fly vials, and the post-dissection carcasses.
Basic setup of the experiment. Which genotypes and why, what kind of food, which developmental stage went up in the rocket, how many went up, how many came down alive, etc...
Description of the dissections.
Tissue samples were transferred directly into the Bead Tubes provided in the MoBio Power Soil Microbial DNA Isolation Kit (MoBio Laboratories, Inc.), and stored at -20degC until the DNA extractions were performed.
DNA was extracted from all samples with the MoBio Power Soil kit, according to the manufacturer’s instructions, with a final elution of 50uL. The 16S rRNA genes were targeted for amplification via PCR, using primer constructs that consisted of the “universal” bacterial primers 515F and 806R, both modified by the addition of Illumina adaptor and an 8bp barcode sequence. (All primer sequences used for this study are shown in table XXX). The PCR amplification protocol included a 3 minute denaturation step at 94degC, 30 cycles of (30sec at 94degC, 45sec at 55degC, 90sec at 72degC), and a final 10min extension step of 72degC. PCR reactions were visualized on an agarose gel and purified using AMPure Beads (Agilent Technologies, Inc.) Purified PCR products were quantified with the Qubit® 2.0 Fluorometer (Life Technologies) HS dsDNA assay. All samples were pooled in eqiumolar ratios prior to sequencing on the Illumina MiSeq sequencer.
The pooled amplicon libraries were de-multiplexed using a custom script (available in a GitHub repository at https://github.com/gjospin/scripts/blob/master/Demul_trim_prep.pl). This script allows for 1 base pair difference per barcode to accommodate for sequencing errors. Paired reads are aligned and a consensus is computed using FLASH (Magoc 2011) with maximum overlap of 120 and a minimum overlap of 70 (other parameters are left as default). FLASH output is reformatted (as separate FASTA files,) with appropriate naming conventions for QIIME v. 1.8.0 (Caporaso 2010).
Unless stated otherwise, all microbial diversity analyses were performed using python scripts included in the QIIME v. 1.8.0 analytical pipeline (Caporaso 2010). An annotated IPython notebook is provided linktext.
Sequences for which the forward and reverse reads had enough overlap to be aligned to create a single consensus sequence were clustered against the greengenes (gg_13_8_otus) OTUs clustered at 97% similarity using the pick_open_reference_otus.py workflow. All reads that failed to hit the reference database were clustered de novo. An OTU table in the .biom format (McDonald 2012) was then constructed and used as the starting point for all downstream analyses.
OTUs that were classified as mitochondrial, chloroplast, or “Unassigned” at the Domain level were filtered from the OTU table using the filter_taxa_from_otu_table.py script. Chimera Slayer (Haas 2011) as implemented in the identify_chimeric_seqs.py script was used to identify putative chimeric sequences, which were then filtered from both the .biom table and from the set of representative sequences used to represent each OTU. A new phylogenetic tree, with chimeric OTUs removed, was produced with the make_phylogeny.py script. Because we wish to compare the microbiomes with and without the inclusion of the intracellular bacterium, Wolbachia, the filter_taxa_from_otu_table.py script was also used to produce an additional .biom table with OTUs classified as Wolbachia removed.
The core_diversity_analyses.py script is a QIIME workflow script that implements a large suite of alpha and beta diversity analyses, including taxonomic composition, diversity and richness estimates, rarefaction curves, both phylogenetic (UniFrac) and distance-based (Bray-Curtis) clustering of samples in Principlal Coordinates Analyses, as well as several hypothesis tests about the differences between groups of samples and the individual microbial taxa that contribute to those differences. Only analyses that are relevant to the discussion will be highlighted in the Results/Discussion.
Is there a difference between the space flies and the ground flies? Is that true for all genotypes? Is that true for all sources (poop, guts, bodies?) Is there a difference between genotypes? Is there a difference between guts and poop? Bodies and guts? Bodies and poop? Are there any taxa unique to the space samples? to the ground samples?