To our knowledge, this is the first study into the intestinal microbiota of triploid Atlantic salmon, and the first into any species of triploid fish. The objective was to compare the GI microflora and the antibiotic resistance profiles of the intestinal microflora of Atlantic salmon after chromosome set manipulation (triploidy) to see if it resulted in an altered physiological profile. No differences were observed in the diversity of bacteria isolated from the gut of triploid or diploid fish, however, triploid fish were found to contain greater total bacterial counts within each gut section [Figure 2], although this was not due to a general increase in all the isolated bacteria [Table 3]. For specific bacteria, triploids were found to contain significantly more Pseudomonas sp., Pectobacterium carotovorum, Psychrobacter spp., Bacillus spp., and Vibrio spp., whereas Carnobacterium spp. was close to being significantly lower in triploids compared to diploids. These results suggest the physiological differences associated with triploidy have an influence on the gut microbiota.
Bacterial culturing and 16S rRNA PCR were used in this study to identify and quantify the bacteria within the fish gut. It is acknowledged that traditional culturing may only identify 11-50% of the present bacteria [46, 47]. Additionally, competitive differences between the bacteria on the culture medium compared to the fish gut may lead to misrepresentation of the microbiota community , and the media used in this study, blood agar and Brocalin agar, are not traditionally used in fish studies and may not have supported growth of Lactobacillus spp., previously found to be dominant in the GI tract of Atlantic salmon . Metagenomic analysis may have led to a more realistic representation of the GI microflora in this study, however, this method is relatively underdeveloped in Atlantic salmon and not comparable to the existing literature that uses methods that are still routinely, and historically, used in the field of microbiology. As such, previously identified increases in the level of bacteria from the foregut to the hindgut were observed in this study [23, 42], along with the appearance of certain bacteria inhabiting the different gut sections within salmonids [20, 42, 49]. Therefore, the authors believe the results of this study are reliable for making an accurate comparison related to ploidy within this study.
Differences were observed in the culturable bacteria levels between the triploid and diploid tank water. Increases in Bacillus spp. and Vibrio spp. within triploid tanks were consistent compared to diploid tanks, however the levels of the other isolated bacteria were similar. All tanks were supplied with the same inflow water that contained neither Vibrio spp., nor Bacillus spp., however, feed did contain Bacillus spp. Therefore, increases in Bacillus spp. could be explained by an increase in uneaten feed in the triploid tanks; however, feed intake was not assessed and uneaten feed was removed by the self-cleaning design of the tank, and this would not explain the increase in Vibrio spp. Alternatively, the increases in Vibrio spp. and Bacillus spp., could be explained by the release of allochthonous bacteria that are only transient in the gut and are released over time in faeces. Therefore the increase in Bacillus spp., and Vibrio spp., in triploid compared to diploid tank water could be expected due to the increases in these two bacteria observed within the triploid gut. However, this was not the case for Pseudomonas spp., Pectobacterium carotovorum, and Psychrobacter spp. that were found in equal quantities in tank water from both ploidies, despite being found to have increased numbers within the triploid gut. The reasons for this are unknown, but may be related to the ability of different bacterial species to adhere to the intestinal mucosa as part of the autochthonous flora and therefore not be released in such great quantities in the faeces, or the ability of these bacteria to survive and compete within the tank water.
Limitations in our study include no control for the effect of pressure treatment independent of ploidy, and the effect of genetic variation within the study fish. Firstly, the current evidence would suggest that triploid induction procedures (when optimized) do not influence triploid performance [12, 13, 16], although Malison et al.  concluded otherwise, this study failed to have an adequate control. With regards to the genetic composition of the study population, it is recognized family effects are observed on triploid survival and physiological performance  and this may influence gut microbiota, however we did not test for this. Therefore, the study should be repeated with an emphasis on the family effect on gut microbiota in fish of both ploidy.
Previously, 10 of the 11 isolated bacteria have been recorded in Atlantic salmon [27, 42, 52, 53]. In addition, the levels of bacteria isolated in this study agree with those previously found in Atlantic salmon . However, we could not find any previous report of Pectobacterium carotovorum subsp. carotovorum, previously isolated from Chinese cabbage [Zhang et al. 2008, unpublished data], in Atlantic salmon. Its source and biological significance in this study are unknown.
Our results suggest the triploid gut provides a different environment to that of the diploid gut, based on bacterial levels. Unfortunately, no literature exists on the triploid gut, or triploid physiology that may affect bacterial growth, therefore some of the major factors that influence gut microbiota cannot be discussed (i.e. gut mucous, physiochemical environment) [52, 54, 55]. Although no significant differences on weight, length, or condition factor were seen within this study, significant differences were seen in the growth (triploids were heavier and longer), condition factors (triploids had a lower value), and the hepatosomatic index (higher in triploids) in pre-smolts. Previously, differences in gut microbiota have been identified between fast and slow growing members of fish from the same population , and may explain our results. Alternatively, we also observed differences in the leucocyte composition (neutrophilic and B-cells proportions) in the same study fish three weeks after seawater transfer (Fraser et al, in preparation). It is well established that the gut microbiota influences the development of the immune system in mammals  and evidence suggests the same occurs in fish . Therefore, it is possible that differences in the immune system related to ploidy (i.e. leucocyte proportions) are influencing the size and composition of the gut microbiota through interactions with mucosal immune cells, or the mechanisms by which the microbiota influences the immune development in fish may be altered by the ploidy status. Cortisol levels and female sex hormone levels have been found to be lower in triploids compared to diploids [18, 57, 58], and these may also influence the immune system in fish [59–65]. However, the majority of studies have observed no differences in cortisol [18, 62–65] due to ploidy and, up until the time of maturing sex hormone levels are typically below detectable levels in salmon and therefore unlikely to influence the immune system in this study. A final explanation is that of differences in cell size and number relating to ploidy. It is well established that cell size increases with increasing numbers of chromosome sets . This phenomenon is more apparent in cells where the nucleus occupies a greater proportion of the total cell volume, so may have little effect on some intestinal cells (i.e. enterocytes). However, triploids have organs of equal sizes to diploids due to reductions in cell number [6, 7], and it may be that triploids have fewer cells along the gastrointestinal tract, such as goblet cells, altering the conditions within the gut. Therefore, differences in the ontogenetic growth, the immune system, or gut environment in the triploids may explain our results.
A difference in the gut microbiota between the ploidies could have a number of implications on triploid aquaculture practices. For instance, Vibrio spp., are known to be opportunistic pathogens and causative agents of disease and mass mortalities , Pseudomonas spp., Bacillus spp, and Carnobacterium spp., are capable of inhibiting pathogenic bacteria and improving survival in vivo after challenge tests [47, 66], Psychrobacter spp., and Arthrobacter agillis have been linked to enteritis in Atlantic salmon , and Bacillus spp., and Carnobacterium spp., are commonly used in fish pro- and pre-biotics [23, 30]. These results suggest that the susceptibility of triploids to opportunistic pathogens (important in both fish and human health) or potential probiotic treatments may be increased in comparison to diploids. Indeed, several studies have reported triploids to be more prone to disease than diploids [12–15].
Despite no known antibiotic exposure within our study population, we detected a variety of drug resistance patterns within bacterial isolates. Previous studies have reported antibacterial drug resistance in various aquatic environments [67–69] at levels higher than those found within our study, possibly as a result of antibiotic use that exerts an ecological pressure on bacteria . However, increased levels of antibiotic resistant bacteria, over that of the local environment, have been reported in aquaculture facilities that are not utilizing antibiotics , and the reasons for this remain unclear.
Triploids were found to have consistently higher levels of antibacterial resistance in their bacterial isolates than in diploids; with 3 of the 6 effective antibacterials significantly affected by ploidy [Figure 3] but no ploidy effect was seen on the occurrence of multiple drug resistance. We suggest two possible explanations for the first observation i) possible physiological differences between the two ploidies (as discussed above) impacting upon the host-bacterium interaction, and ii) the significantly greater levels of bacteria found within the triploid compared to diploid intestine might provide better conditions for more efficient bacterial colonization, which is very critical for horizontal gene transfer . Unfortunately, both of these hypotheses were outside the scope of this study.