High-throughput DNA sequencing of the moose rumen from different geographical location reveals a core ruminal methanogenic archaeal diversity and a differential ciliate protozoal diversity.

This project expanded upon my work with moose bacteria from three geographic locations, to explore whether there were differences in methanogenic archaea or ciliated protozoa based on location.

Archaea are microorganisms in their own Domain, as they are neither Bacteria nor Eukaryota, although they often have similarities to organisms found in the other two domains.  Archaea are found in many extreme environments, but those found in the digestive tract of animals and humans come from the phylum Euryarchaeota.   Methanogens require hydrogen to make energy for themselves, and in that process (methanogenesis) methane is created as a byproduct.  In the digestive tract, especially in ruminants where the fermentation of plants creates a lot of hydrogen, the presence of methanogens acts a hydrogen sink and can prevent the build up of hydrogen which would otherwise lower the gut pH and be detrimental to both host and microbes.  To date, it is unclear if methanogens have any other health effect.

Protozoa are single-celled eukaryotes, and depending on which species they are, can be beneficial or pathogenic.  Typically, protozoa in the digestive tract of humans or other monogastrics are pathogens obtained from drinking contaminated water.  However, the digestive tracts of monogastrics (ex. humans) and ruminants (ex. moose) are very different, and the later can support a much different microbial community.  Specifically, protozoa found in ruminants that have cilia to move around (i.e. ciliated protozoa or ciliates) can have a number of roles, including fermentation of fiber or starch, or predation of bacteria and fungi.  As they are so difficult to maintain in culture and study in the lab, the role of protozoa in contributing to host health or methanogenesis is understudied.

Moose methanogen communities were significantly different between moose in Vermont, Norway, and Alaska, but maintained a core of shared taxa across all populations.  This implies that the moose rumen environment (pH, salt content, turnover, host-microbe interactions, etc.) is suitable for only a small number of methanogen species, and that this regulates the community as much as diet might.  Methanogen communities were also different based on sex of the moose, and age/weight.

Figure 2: Diversity of moose rumen methanogens. Members of the RO clade are coloured in blues; members of the SGMT clade are coloured in reds. Mbr., Methanobrevibacter.

On the other hand, protozoal communities were dramatically different between moose in Vermont, Norway, and Alaska, and shared far fewer taxa.  This was surprising, as previous studies on deer had shown a core protozoal community across multiple geographically-separated populations.  These moose populations had not been geographically isolated long, but we hypothesized that diet was a much stronger driver of rumen protozoal diversity than previously thought.

Figure 3: Diversity of the moose rumen protozoa.

Ishaq, S.L., Sundset, M.A., Crouse, J., Wright, A-D.G. 2015. High-throughput DNA sequencing of the moose rumen from different geographical location reveals a core ruminal methanogenic archaeal diversity and a differential ciliate protozoal diversity. Microbial Genetics, 2015(1).  Article


Featured Image; Figure 1: PCoA for moose methanogens (A, C, E) and protozoa (B, D, F). PCoA is coloured by (A, B) gender: female, red; male, blue; (C, D) location: Alaska, red; Norway, green; Vermont, blue; and (E, F) weight class: 1–100 kg, red triangle; 101–200 kg, yellow triangle; 201–300 kg, green down-facing triangle; 301–400 kg, green right-facing triangle, >400 kg (live weight), light blue circle; not available, blue square.

Dissertation: A Comparative Analysis Of The Moose Rumen Microbiota And The Pursuit Of Improving Fibrolytic Systems

As a Ph.D. student, I worked in the laboratory of Dr. André-Denis Wright in the Department of Animal Science at the University of Vermont. My thesis work investigated the microorganisms (bacteria, archaea, and protozoa) in the digestive tract of the moose from several geographical locations. In addition to identifying bacteria and ciliate protozoa using high-throughput sequencing, […]

Design and validation of four new primers for next-generation sequencing to target the 18S rRNA gene of gastrointestinal ciliate protozoa.

If the research tools you require don’t exist- then you must create them yourself.  Such is often the case in working with microbial genomics.  In order to adapt sequencing technology to identify rumen ciliate protozoa, we needed to first design primers which to be used for Polymerase Chain Reaction (PCR) in order to amplify enough copies of the 18S rRNA gene for laboratory work.  This involved designing primers in silico, by aligning sequences from the few protozoal 18S rRNA genes publicly-available at the time, in order to identify short sections which were identical across protozoal species.  We then added 18S rRNA gene sequences to our alignment from other Eukaryotes, such as fungi and plants, which we did not want to amplify, to ensure that our primers would target only the desired taxa.  We also needed to design a primer set which would work well in the laboratory; in particular which had an optimal size for the sequencing technology on hand, and which would provide enough information in the portion of the gene capture to identify which protozoal species the DNA in our rumen samples originated from.

Ishaq, S.L., Wright, A-D.G. 2014. Design and validation of four new primers for next-generation sequencing to target the 18S rRNA gene of gastrointestinal ciliate protozoa. Applied and Environmental Microbiology, 80(17):5515-5521.  Article

Four new primers and one published primer were used to PCR amplify hyper-variable regions within the protozoal 18S rRNA gene to determine which primer pair provided the best identification and statistical analysis. PCR amplicons of 394 to 498 bases were generated from three primer sets, sequenced using Roche 454 pyrosequencing with Titanium, and analyzed using the BLAST (NCBI) database and MOTHUR ver. 1.29. The protozoal diversity of rumen contents from moose in Alaska was assessed. In the present study, primer set 1, P-SSU-316F + GIC758R (amplicon = 482 bases) gave the best representation of diversity using BLAST classification, and amplified Entodinium simplex and Ostracodinium spp., which were not amplified by the other two primer sets. Primer set 2, GIC1080F + GIC1578R (amplicon = 498 bases), had similar BLAST results and a slightly higher percentage of sequences that identified with a higher sequence identity. Primer sets 1 and 2 are recommended for use in ruminants. However, primer set 1 may be inadequate to determine protozoal diversity in non-ruminants. Amplicons created by primer set 1 were indistinguishable for certain species within the genera Bandia, Blepharocorys, Polycosta, Tetratoxum, or between Hemiprorodon gymnoprosthium and Prorodonopsis coli, none of which are normally found in the rumen.


Figure 1: A map of the full-length protozoal 18S rRNA gene, including variable (V1-V9) and rumen ciliate signature regions (SR1-SR4), and showing the respective amplicons of the three primer sets used in the present study.

Figure 2: Taxonomy and proportion of unique pyrosequences using NCBI (BLAST), by forward primers P-SSU-316F (Sylvester et al., 2004), GIC1080F (present study), and GIC1184F (present study). All sequences used passed all quality assurance steps outlined in Methods.

High-throughput DNA sequencing of the ruminal bacteria from moose (Alces alces) in Vermont, Alaska, and Norway.

For the second project of my Ph.D., I expanded upon my findings in Vermont moose.  Following the collection of samples from moose in Vermont, I was able to obtain samples from moose in Alaska and Norway, as well.  The Alaskan moose were part of the Moose Research Station herd in Soldotna, Alaska, where they were allowed to roam and graze freely in a large 1 mi sq enclosure.  During sample collection for another project, Dr. John Crouse and Dr. Kimberlee Beckmen, both of the Alaska Department of Fish & Game, were able to intubate the sedated moose and collect rumen digesta samples.

A colleague in Norway, Dr. Monica Sunset, of the University of Tromsø, was able to facilitate sample collection and storage of moose rumen samples from two hunters; Drs. Even Jørgensen and Helge K. Johnsen, of the University of Tromsø.  As mailing rumen samples across country lines is restricted to prevent the potential spread of livestock diseases, it was actually easier to send me to Norway to extract DNA to ship home.  While in Norway, Dr. Sundset taught me how to culture microorganisms anaerobically – without the presence of oxygen.

For this project, we used high-throughput sequencing using the Roche 454 pyrosequencing platform, required me to learn the fine art of bioinformatics.  We were surprised to find that rumen bacterial communities in moose were different for each geographical location.  While we did not identify the diet that moose were eating, we speculated that these differences were driven by slightly different diets at the time points that each location was sampled in.  Plants often become more fibrous and less nutritious as the growing season develops and passes, and this nutritional change in substrate can select for different bacterial communities.  It has since been confirmed by a number of studies that geographic differences exist in the microbiome, driven by changes in site-specific diet, as well as food- and waterbourne microbial influences.

Ishaq, S.L., Wright, A-D.G. 2014. High-throughput DNA sequencing of the ruminal bacteria from moose (Alces alces) in Vermont, Alaska, and Norway. Microbial Ecology, 68(2):185-195. Article


In the present study, the rumen bacteria of moose (Alces alces) from three distinct geographic locations were investigated. Moose are large, browsing ruminants in the deer family, which subsist on fibrous, woody browse, and aquatic plants. Subspecies exist which are distinguished by differing body and antler size, and these are somewhat geographically isolated. Seventeen rumen samples were collected from moose in Vermont, Alaska, and Norway, and bacterial 16S ribosomal RNA genes were sequenced using Roche 454 pyrosequencing with titanium chemistry. Overall, 109,643 sequences were generated from the 17 individual samples, revealing 33,622 unique sequences. Members of the phylum Bacteroidetes were dominant in samples from Alaska and Norway, but representatives of the phylum Firmicutes were dominant in samples from Vermont. Within the phylum Bacteroidetes, Prevotellaceae was the dominant family in all three sample locations, most of which belonged to the genus Prevotella. Within the phylum Firmicutes, the family Lachnospiraceae was the most prevalent in all three sample locations. The data set supporting the results of this article is available in the Sequence Read Archive (SRA), available through NCBI [study accession number SRP022590]. Samples clustered by geographic location and by weight and were heterogenous based on gender, location, and weight class (p < 0.05). Location was a stronger factor in determining the core microbiome than either age or weight, but gender did not appear to be a strong factor. There were no shared operational taxonomic units across all 17 samples, which indicates that these moose may have been isolated long enough to preclude a core microbiome among moose. Other potential factors discussed include differences in climate, food quality and availability, gender, and life cycle.


Proteobacteria Unique Sequence Firmicutes Bacteroidetes Weight Class 

Ishaq poster FEMS 2013

Insight into the bacterial gut microbiome of the North American moose (Alces alces).

My first research project as a Ph.D. student was investigating the differences in the rumen and colon bacterial communities in moose from Vermont.  To obtain samples, I needed to secure permission from VT Department of Fish and Wildlife, as moose hunting and moose sample collection are tightly regulated; hunting is only permitted for approximately one week a year, and the licenses are awarded on a lottery system.  I was able to contact several permitted hunters for the 2010 season, and passed on sample collection kits to them, complete with gloves, jars, ice packs, a cooler, and instructions with photographs.  A total of 14 samples were collected, 8 from the rumen and 6 from the colon – a small sample group in general, but a good-sized one for a wild animal study.  If a hunter was successful, they would contact me to collect the samples as soon as possible after the carcass was reported to the state at one of a number of weigh stations.  Between dropping off and collecting samples, I drove a total of 1500 miles in about 9 days, and was able to see some beautiful parts of Vermont.

Once back at the lab, the samples were frozen at -80 degrees Celsius to protect the microbial cells and DNA from degradation- I knew I may not get another chance to collect rumen samples from moose.  For this project, I used a small amount of each sample and extracted the total mass of nucleic material available – the DNA and some RNA from all the cells present, including microbial, moose, and plant.  I then used Polymerase Chain Reaction (PCR) to amplify, or make copes of, just the bacterial DNA.  Specifically, I was looking for the 16S rRNA gene, which I would use for amplification.  I used a DNA microarray, Phylochip, which had fragments of DNA from known bacteria bound to the chip.  Once I applied my sample DNA, if there was a DNA- DNA match, my sample would bind to the chip as well as would fluoresce under ultraviolet light.  A sophisticated computer program would read the light signals and interpret then as the presence and rough abundance of that particular species of bacteria.

The most important finding of the study was that the rumen and colon hosted sufficiently different bacterial communities, such that colon or fecal samples were not a good proxy for what was happening in the rumen.  While this had been shown once before in sheep, it was not common practice to sample from more than one gastrointestinal location in ruminants, particularly in wild ones, because the cost of sequencing was high and the logistics of sample collection were difficult.  Thus, it was common practice to collect fecal samples from wild ruminants and speculate about the rumen microbial communities.


Ishaq, S.L., Wright, A-D.G. 2012. Insight into the bacterial gut microbiome of the North American moose (Alces alces). BMC Microbiology, 12:212.  Article



The work presented here provides the first intensive insight into the bacterial populations in the digestive tract of the North American moose (Alces alces). Eight free-range moose on natural pasture were sampled, producing eight rumen samples and six colon samples. Second generation (G2) PhyloChips were used to determine the presence of hundreds of operational taxonomic units (OTUs), representing multiple closely related species/strains (>97% identity), found in the rumen and colon of the moose.


A total of 789 unique OTUs were used for analysis, which passed the fluorescence and the positive fraction thresholds. There were 73 OTUs, representing 21 bacterial families, which were found exclusively in the rumen samples: Lachnospiraceae, Prevotellaceae and several unclassified families, whereas there were 71 OTUs, representing 22 bacterial families, which were found exclusively in the colon samples: Clostridiaceae, Enterobacteriaceae and several unclassified families. Overall, there were 164 OTUs that were found in 100% of the samples. The Firmicutes were the most dominant bacteria phylum in both the rumen and the colon. Microarray data available at ArrayExpress, accession number E-MEXP-3721.


Using PhyloTrac and UniFrac computer software, samples clustered into two distinct groups: rumen and colon, confirming that the rumen and colon are distinct environments. There was an apparent correlation of age to cluster, which will be validated by a larger sample size in future studies, but there were no detectable trends based upon gender.


ColonGut microbiomeRumenVermont16S rRNA


Figure 2 Breakdown of unclassified families by phylum. (a) OTUs present in all 14 samples. There were 41 OTUs found exclusively in the rumen that were not classified down to the family level. (b) OTUs found exclusively in the rumen. There were 22 OTUs found exclusively in the rumen that were not classified down to the family level. (c) OTUs found exclusively in the colon. There were 19 OTUs found exclusively in the colon that were not classified down to the family level. Several are candidate phyla and are named by where they were discovered: AD3, soil in Virginia and Deleware, USA; OP3 and OP10, now Armatimonadetes, Obsidian Pool hot spring in Yellowstone National Park, USA; NC10, Null Arbor Caves, Australia; TM7, a peat bog in Gifhorn, Germany; WS3, a contaminated aquifer on Wurtsmith Air Force Base in Michigan, USA.


Figure 3A comparison of the OTUs exclusive to the rumen or the colon. A comparison of the 73 OTUs exclusive in the rumen (n = 8) or 71 OTUs exclusive in the colon (n = 6), by family. Families with three or more associated OTUs are labeled in the chart; all other families with two or fewer OTUs are labeled via the legend. The Unclassified sections are broken down by phyla in Figure2b, and2c, respectively.