Identification of the major bacterial groups in the digestive tract of cod and haddock Neil McEwan Institute of Biological, Environmental and Rural Sciences, Llanbadarn Campus, Aberystwyth University, Aberystwyth, Ceredigion, SY23 3AL, Wales nrm@aber.ac.uk Jim Treasurer Vikingfish, Ardtoe Marine Laboratory, Ardtoe, Acharacle, Ardnamurhcan, Argyll, PH38 4LZ Scotland jim.treasurer@vikingfish.com Summary The bacterial diversity contained within samples collected from the stomach of both haddock and cod was examined. In both cases, examples of microbes identified in the digestive tract of other vertebrates were observed. However, the majority of sequences identified were most similar to sequences which had previously been found in environmental samples (e.g. soil samples). This is suggests that the bacterial community within the stomach of these fish is directly influenced by microbes being swallowed. This is different from the observations which have been reported for mammalian gut environments, where the food ingested (as opposed to the bacteria themselves) influence the bacterial population. This point is re-iterated by the relative abundance of sequences which were found which had high similarity to chloroplasts, which previously have constituted a minor level of contamination in other gut systems studied, suggestive of there not being a particularly strongly established bacterial population within the stomach of these fish.
Identification of the major bacterial groups in the digestive tract of cod and haddock Neil McEwan (Aberystwyth University, Aberystwyth, Wales) Jim Treasurer (Vikingfish, Ardtoe, Scotland) General background to the work undertaken Food is ingested into the digestive tract of all vertebrates. Research in both mammals and birds has shown that the process of breaking down the food to the basic constituents needed by the host animal generally cannot be achieved by the enzymes encoded by the vertebrate itself. Instead it is clear that, to varying degrees, the microbes of the gut play a role in maximising the breakdown of food sources, thereby making nutrients available to the host animal. Given the importance attached to recognising the microbes in the digestive tract of both mammals and birds, the current work was undertaken to initiate a similar level of understanding for the bacterial composition of the digestive tract of two economically important marine fish. In the first instance, it was proposed that examples of some of the more abundant groups of organisms present in the digestive tract be identified for both haddock and cod. Methodology Five haddock (90-110g) and five cod (50-90g) which were around 8 months old and had routinely been fed a diet of 3 mm size Biomar Biomarine were used in the experiment. They had been maintained in seawater tanks at Vikingfish and following killing were transported on ice to Aberystwyth for analysis. The intestine and stomach from each fish was removed and the digesta extracted immediately. DNA was extracted from all samples using a stool kit (Qiagen) which has previously proved a successful resource for harvesting DNA from a number of sources (e.g. rumen contents, horse faecal samples, rabbit caecal samples). The quality of DNA extracted was verified by two different techniques; spectrophometric analysis to assess purity levels (i.e. check that contamination from protein sources was low) and also by electrophoresis to check for size integrity of the extracted DNA (i.e. check that the DNA has not become fragmented during the extraction process). All samples showed DNA which appeared to have high purity, and also was relatively intact (i.e. 23kb or greater on an agarose gel).
DNA samples from each source type (e.g. haddock intestine, cod stomach, etc) were pooled for molecular biological analysis. Regions of the bacterial 16S ribosomal genes from total DNA extracts were amplified by PCR, using primers previously shown to amplify sequences from other gut environments, and other ecosystems such as soil. The PCR step was designed to produce an amplicons of around 900 nucleotides in length, which is generally accepted as being sufficiently long to permit taxonomic and/or phylogenetical analysis to take place. Analysis of the amplicons generated following PCR by electrophoresis showed that samples from the stomach of both cod and haddock produced strong signals of the predicted size. In contrast, the results from intestinal samples were less successful, with weak product samples being obtained from the haddock material, and those from cod failing to produce an amplicon. Despite repeating the PCR process, neither intestinal sample gave a better product and so a decision was made to concentrate on the stomach samples in order that comparisons between samples from both species would be possible. PCR products from stomach samples from both species were cloned into a pcr 4- TOPO vector (Invitrogen) in preparation for cloning in E. coli strain (MACH 1 TM T 1 R ). Resulting colonies were subjected to a preliminary screen for inserts, based on blue/white colouration. All colonies which remained white after a brief incubation at 4 C were selected as candidate clones for sequencing. Sequencing was performed in both directions using the M13 Forward and M13 Reverse primers recommended for use with this cloning vector. Resulting DNA sequences were then compared by BLAST (Basic Local Alignment Search Tool) using the NCBI / GenBank database. For each sample, the organism with the highest identity level for the sequence being investigated was noted, together with the type of environment in which this organism was first identified. These results are reported in Table 1. General observations arising from experimental work Previously we have found that the pcr 4-TOPO vector in conjunction with E. coli strain (MACH 1 TM T 1 R ) has proved a highly reliable method of cloning PCR products in preparation for DNA sequencing. However, in the current investigation, numbers of
colonies were considerably less than those we have normally found. Typically less than 10 colonies of cloned cells were identified on any particular agar plate, which is in sharp contrast to the number of colonies which we would normally anticipate (i.e. 200-500). This resulted in multiple attempts at generating sufficient clones, in each case resulting in considerably lower numbers of clones than would normally be expected. All clones which remained white after a brief incubation at 4 C were selected for sequencing. Normally at this point a very small number (typically only 1-2%) of the PCR products which have been cloned are identified as having come from chloroplasts. This is due to the fact that chloroplasts also have 16S ribosomal genes, and some of the chloroplast DNA present in the plant material used as a food source for the animals we are working with has been amplified. In the current work, around half of the sequences being analysed had to be removed from the study as they were clearly of chloroplast origin. The reason for this abnormally high number of chloroplast sequences being obtained is unclear. Most probably, it is either due to the DNA in the stomach bacteria being particularly well-suited to amplification by this specific pair of primers, or alternatively the number of bacteria in the stomach is relatively low in comparison to that seen in other gut systems with which we have worked. Either of these explanations is feasible, and the two are not mutually exclusive, with a combination of the two being possible. Although the primers used for PCR in this investigation have been used with considerable success for gut samples from other sources, it is well-known that some primer pairs just appear to be unsuited for specific types of work. Troubleshooting of primer optimisation has the potential to be a lengthy process, and was unlikely to be realistically achievable within the timescale allocated to this work.
Results of sequence analysis Analysis of the sequences obtained is shown in Table 1 below. Source Most similar species GenBank Number Type of bacterial group Identity (%) Source of best identity / comments Haddock Microbacterium hominis AM181504 Actinobacteria 98 lung aspirate Haddock Propionibacterium acnes DQ672261 Actinobacteria 99 soil Haddock Propionibacterium acnes DQ672261 Actinobacteria 99 soil Haddock Propionibacterium acnes AY642054 Actinobacteria 98 Haddock Staphylococcus epidermidis AF269314 Firmicutes 99 skin Haddock Propionibacterium acnes DQ672261 Actinobacteria 99 soil Haddock Enterobacter intermedius AF310217 Gammaproteobacteria 99 99% identical to sample Haddock Enterobacter intermedius AF310217 Gammaproteobacteria 99 from rainbow trout gut Haddock Clostridia sp. EU551097 Firmicutes 95 cow manure Haddock Nitrosomonas sp. EU155074 Betaproteobacteria 97 prawn farm sediment Haddock Uncultured Lachnospiraceae EF700905 Firmicutes 94 human intestinal tract Haddock Uncultured Mycobacterium sp EU341163 Actinobacteria 99 aircraft cabin air Haddock Uncultured bacterium DQ824525 Firmicutes 94 human faeces Haddock Burkholderia sp AF052387 Betaproteobacteria 99 volcanic deposits Cod Uncultured Firmicute AY913399 Firmicutes 99 forest floor Cod Propionibacterium sp EU531789 Actinobacteria 99 mud volcano Cod Morganella morganii AB099406 Gammaproteobacteria 98 yellowtail muscle samples Cod Vibrio sp AB038029 Gammaproteobacteria 99 marine bacterium Cod Vibrio sp AB038029 Gammaproteobacteria 99 marine bacterium Cod Uncultured Firmicute AY913399 Firmicutes 97 forest floor Cod Leuconostoc citreum AB362721 Firmicutes 100 French cheeses Cod Burkholderia fungorum EF650018 Betaproteobacteria 99 soil Cod Uncultured Firmicute EF402828 Firmicutes 95 human faeces Cod Propionibacterium sp AF443576 Actinobacteria 99 soil Table 1. Summary of the relationship between the sequences generated in this work and those already present in the GenBank public database. The number of sequences generated by this work was lower than initially anticipated. Two primary reasons were attributed to this; the low number of clones generated in the first instance, and the unexpectedly high number of these clones which contained DNA which had originated from chloroplasts. All identity values are based on preliminary sequence analysis which would have to be verified more closely before data were made public. However, identity levels are unlikely to change more than 1-2%, meaning that although the species might change in some cases, the broader classification (e.g. firmicutes) will remain unchanged.
However, inspection of the range of organism which were detected is also very interesting, as it demonstrates that, as with most environments, there is a wide range of different bacterial species present (e.g. firmicutes, proteobacteria, etc). The type of environment in which the best hit species was first identified also provides interesting information. There are a few sequences which have their origin in a gut environment either in the form of samples collected directly from intestinal material (including two with high similarity to an isolate from the digestive tract of a trout), or indirectly from faecal samples. However, the majority of the samples fall into the category of what can best be described as environmental samples, being typical of those seen in sediments, soil or air. This observation may also be explained in one of two ways. Firstly, it is possible that these results are a reflection of the successful nature of relatively closely related bacterial species in colonising very different environmental niches. Thus, although related to an organism which occupies a different niche, the sequences identified in this work may be representative of those found in this particular gut environment. Secondly, the bacteria from which the DNA had been extracted may include a number of organisms which genuinely do fall into the category of environmental samples. If this is the case, then their presence in the stomach may be a reflection of the lifestyle of the fish. Instead of being truly organisms which have evolved for existence in the digestive tract, it is possible that these individuals are actually transient in nature and have been swallowed by the fish. This is another likely scenario, as these bacteria could be present both in water entering the mouth, or even in any sediments which might be engulfed. Concluding remarks The number of sequenced obtained in the current work was less than had originally been expected. This could be attributed to two different observations which both contributed to this low number. Firstly there was the apparent recalcitrance of the PCR products to clone, and secondly there was unexpectedly high level of chloroplast sequences observed, which had to be removed from the subsequent analysis.
A few of the sequences obtained showed similarity to sequences previously obtained from gut samples. However, in most cases, the best hits identified were from samples obtained from environmental niches. It is unclear if these are genuinely environmental bacteria which were the source of these sequences. However, by pursuing the samples from the intestine in more detail may provide an answer to this problem, as it is less likely that an environmental sample is going to persist and survive further down the digestive tract.