Protective role of the vulture facial skin and gut microbiomes aid adaptation to scavenging

Background Vultures have adapted the remarkable ability to feed on carcasses that may contain microorganisms that would be pathogenic to most other animals. The holobiont concept suggests that the genetic basis of such adaptation may not only lie within their genomes, but additionally in their associated microbes. To explore this, we generated shotgun DNA sequencing datasets of the facial skin and large intestine microbiomes of the black vulture (Coragyps atratus) and the turkey vulture (Cathartes aura). We characterized the functional potential and taxonomic diversity of their microbiomes, the potential pathogenic challenges confronted by vultures, and the microbial taxa and genes that could play a protective role on the facial skin and in the gut. Results We found microbial taxa and genes involved in diseases, such as dermatitis and pneumonia (more abundant on the facial skin), and gas gangrene and food poisoning (more abundant in the gut). Interestingly, we found taxa and functions with potential for playing beneficial roles, such as antilisterial bacteria in the gut, and genes for the production of antiparasitics and insecticides on the facial skin. Based on the identified phages, we suggest that phages aid in the control and possibly elimination, as in phage therapy, of microbes reported as pathogenic to a variety of species. Interestingly, we identified Adineta vaga in the gut, an invertebrate that feeds on dead bacteria and protozoans, suggesting a defensive predatory mechanism. Finally, we suggest a colonization resistance role through biofilm formation played by Fusobacteria and Clostridia in the gut. Conclusions Our results highlight the importance of complementing genomic analyses with metagenomics in order to obtain a clearer understanding of the host-microbial alliance and show the importance of microbiome-mediated health protection for adaptation to extreme diets, such as scavenging. Electronic supplementary material The online version of this article (10.1186/s13028-018-0415-3) contains supplementary material, which is available to authorized users.

2 methionine and cysteine degradation. This likely explains the identification of the metabolism of cysteine and methionine as the two most abundant subclasses from the amino acids metabolism in both facial skin and gut microbiomes. A carcass also produces volatile organic compounds [7], such as acetone, methyl ethyl ketone, toluene, ethylbenzene, m,p-xylene, styrene, and o-xylene. In this regard, toluene degradation was one of the subclasses not driving variation in the facial skin functional intra samples comparison, and the MOCAT facial skin cores had more genes related to xenobiotics biodegradation metabolism than those of the gut microbiome (Additional file 3).

Intestinal microbiome related to digestion
Among the taxa present in greater abundance in the facial skin microbiome than in the gut microbiome, we identified taxa and functions that are usually part of the gut microbiome of mammals. These bacteria could be derived from the carrion but be removed from the vulture gut microbiome. For example, present only in the facial skin dataset was Cellulophaga lytica, which is capable of degrading proteins and polysaccharides [8], as well as Flavobacterium columnare, which produces gelatin-degrading and chondroitin sulfate-degrading enzymes [9,10]. This is relevant given that chondroitin sulfate is one of the main structural components of cartilage.
Although we did not identify these genes from F. columnare, we identified chondroitin sulfate ABC lyase genes in both facial skin (4 genes from Bacteroides spp. and Proteus spp.) and gut microbiomes (13 genes from Bacteroides spp., Edwardsiella spp., and Proteus spp.).

Fusobacterium digestive roles
It has been proposed that the abundance of Fusobacterium in the gut could aid in the digestion of meat, given their ability to metabolize amino acids [11,12]. This suggestion is supported by the 3 finding of F. nucleatum and F. varium in the vulture's gut microbiome. One of the most abundant genes in the gut microbiome codes for an alpha-2-macroglobulin family protein from F. mortiferum (the most abundant Fusobacterium in the gut). This protein has been suggested to be used in bacteria as a colonization rather than a virulence factor [13]. Besides, eukaryotic alpha-2macroglobulin, produced by the liver, binds to and removes MMP-2 and MMP-9 (active forms of the gelatinase), which is produced in the stomach to digest gelatin [13,14]. However, the gelatin colloidal properties aid in the digestion of various types of food [15,16]. Furthermore, bacterial alpha-2-macroglobulin can be structurally very similar to that of eukaryotes [17]. This suggests that Fusobacterium could also be playing digestive aiding roles in the vulture gut.

Pathogenic characterization
Looking at the identified bacteria taking into account the strain information, the maximum number of potentially pathogenic bacteria identified in a sample (a facial skin sample) was 482, and the minimum was 10, with x ̅ = 159.8. Each pathogen was present with x ̅ = 11.98, a minimum of 1, and a maximum of 75 (Clostridium perfringens ATCC 13124 and C. perfringens str. 13). We found that the facial skin has more different species of potential pathogens than the gut (P= 0.036, x ̅ facial skin = 189.8, x ̅ gut = 137.3). Present in at least 90% of the samples are three C. perfringens strains which produce gas gangrene [18], and one Stenotrophomonas maltophilia, which produces bacteraemia, bronchitis, pneumonia, and urinary tract infection [19]. In the gut samples, the most abundant hosts for the potentially pathogenic bacteria are human, chicken, turkey, cattle, pigs, and mouse. For those in the facial skin, the most abundant reported hosts are human, followed by cattle, and plants. Among

Fusobacteria and Clostridia pathogenicity
It has been speculated that the large abundance of Fusobacteria and Clostridia in the vulture gut outcompetes other more virulent and toxic relatives, being harmless pathogenic versions that occupy the space and resources that more pathogenic versions would occupy otherwise [20]. To examine this hypothesis, we searched for toxin-related genes from these taxa in the gut functional core. We identified two putative enterotoxins from C. perfringes, and interestingly, also a protein in two gut samples from the bacteriocinogenic plasmid pIP404 from C. perfringes [21]; less toxinrelated genes were found for Fusobacterium.

Probiotics and beneficial functions
One of the most abundant taxa in the facial skin microbiome was Pseudomonas fluorescens, which can produce the antibiotic mupirocin [22]. This antibiotic is used for treating skin, ear, and eye disorders by interfering with isoleucyl-tRNA synthetase activity of pathogens, suggesting it may play a role in treating illnesses that some bacteria could cause on the vulture. Among the potentially health-beneficial bacteria identified in higher abundance in the facial skin microbiome was 5 Arthrobacter phenanthrenivorans (max. coverage in the facial samples= 4.4%, max. coverage in the gut samples= 0.4%, with 163 annotated genes in the facial skin and 66 annotated genes in the gut), which is able to degrade phenanthrene [23], a skin-irritating poly-cyclic aromatic hydrocarbon. Also, part of the facial functional core that takes taxonomy into account is Hylemonella gracilis (max. coverage= ~4%, 69 assembled genes), which has been shown to prevent long term colonization by Yersinia pestis [24]. Among the plasmids present only or most abundantly in the gut microbiome, we identified those of probiotic bacteria Lactobacillus brevis KB290 [25], L. casei W56 [26], L. paracasei [27], L. salivarius CECT 5713 [28], an L. reuteri SD2112 [29], which produces the antimicrobial reuterin [30]. We also identified plasmids from Lactobacillus sakei (max. coverage in the gut samples= 91.5%, max. coverage in the facial skin samples= 26.85%), from which we also identified a putative bacteriocin immunity protein and a type II secretion system protein coding gene in the NR gene catalogue. Folate has been related to skin cancer prevention [31]. Notably, one of the most abundant sub-pathways in the facial skin intra-sample comparison was the folate biosynthesis. In this regard, we also identified the gene dihydropteroate synthase type-2 from Acinetobacter sp. NIPH 899 in the facial microbiome.
Furthermore, aminobenzoate is used to treat skin disorders, and we found that the aminobenzoate degradation subclass from the xenobiotics degradation metabolism was the most abundant in both facial skin and gut microbiomes.

Antibiotics
We identified several genes for the biosynthesis of antibiotics. Among the most abundant metabolic subclasses in both facial skin and gut microbiomes was the biosynthesis of carbapenem.
From the metabolism of terpenoids and polyketides, the most abundant subclasses in the facial skin and gut microbiome were the biosynthesis of tetracycline, macrolides, and ansamycins. Also, 6 among those subclasses not driving functional variation in the facial skin microbiome were the biosynthesis of monobactam, anthocyanin, and ansamycins. Furthermore, in the facial skin MOCAT strict core, we identified a gene involved in the production of naphthocyclinones antibiotics [32] from Herbaspirillum frisingense [33].

Phages
In accordance to a potential phage therapy strategy, among the phages present only in the facial skin microbiome we identified the phage phi MR11, which eliminates multidrug resistant Staphylococcus aureus [34]. We also identified Acinetobacter phage Petty, which infects Acinetobacter baumanii [35], a multidrug resistant pathogen usually isolated from wounds. The phage Acibel004, active against A. baumanii [36], was also present only in the facial skin microbiome. From the facial skin functional strict core, the most abundant phage was BPP-1, which infects pathogenic Bordetella spp. [37]. Among the identifications was the Enterobacteria phage P22 present only in the gut samples, which infects the pathogenic S. typhimurium [38].
Furthermore, we identified the phage L-413C, which is specific for Y. pestis [39] and identified as more abundant in the gut samples. Also present was the phage phi CD119, which reduces toxin production in C. difficile [40]; notably, we also identified genes related to C. difficile virulence.
Also, the Enterobacteria phage HK620 was identified as more abundant in the gut microbiome, this phage absorbs the O-antigen of E. coli H [41]. The gut functional strict core contains the phage phiCD6356, which infects C. difficile [42], and phage SPN3US, which has shown effective inhibition of S. enterica [43]. Notably, we identified putative virulent genes from S. enterica in the gut microbiome (Additional file 7). 7 Besides bacterial killing strategies, we also identified insecticide, fungicide, and antiparasitic related taxa and genes. Among those taxa significantly more abundant in the facial skin microbiome we identified Lysinibacillus sphaericus, which produces insecticidal toxins that control mosquito growth [44], and for which we assembled the gene coding for sphaericolysin, an insecticidal pore-forming toxin. We also identified Pseudomonas entomophila (8 genes in the gut samples, max. coverage in the gut samples= 0.2%, max. mapping reads in the gut samples= 216; 175 genes in the facial skin samples, max. coverage in the facial skin samples= 0.5%, max. mapping reads in the facial skin samples= 542) which infects insects causing lethality in fly larvae and adults [45,46]. Notably, we identified in the NR gene set catalogue a gene annotated as derived from P. entomophila coding for an insecticidal toxin SepC/Tcc class. We also identified Streptomyces violaceusniger (present in 20 facial skin samples, 0.1% max. coverage, 214 max mapping reads, 28 assembled genes; 11 gut samples, 0.03% max. coverage, 58 max mapping reads, 15 assembled genes), which is an antifungal for various plant fungal pathogens [47,48]. Among the antiparasitic taxa in the facial skin relaxed core, we identified Kitasatospora setae, which is capable of producing the antitrichomonal setamycin [49], and Streptomyces bingchenggensis, which produces the anthelmintic macrolide milbemycin [50]. We also identified Heterorhabditis bacteriophora (found in 34 gut samples and 24 facial skin samples), which kills pests like fleas, ants, and flies by releasing Photorhabdus luminescens bacteria from their digestive tract [51].

Defence versus eukaryotes
Interestingly, we also identified P. luminescens (max. mapping reads in the gut samples= 462, max. mapping reads in the facial skin= 20). Although present in low amounts, we also identified in 21 of the gut samples the Dictyostelium genus (D. intermedium and D. citrium, max. mapping reads in the gut samples= 596, max. mapping reads in the facial skin samples= 36), which is a bacteriovorous protozoa present in the soil, where they keep bacterial populations in balance [52]. 8 We also identified Adineta vaga, which feeds on dead organic matter, mainly dead bacteria and protozoans [53], in 93.6% of the intestinal samples and 54.5% of the facial skin samples (gut normalized abundance= 54,230 mapping reads, facial skin normalized abundance= 7,972 mapping reads).

Non-antibiotic mechanisms
Among the taxa more abundant in the facial skin microbiome, it is interesting to note the identification of bacteria capable of growing in cancerogenous substances and producers of anticancer immunosuppressant substances. This is of relevance given that such bacteria might produce antimicrobial alternatives or products beneficial to the vulture to aid in fighting the constant aggression of the toxins present in the carcasses. Present in the relaxed facial skin core we identified Chromobacterium violaceum [54] and Janthinobacterium sp. HH01 [55], which produce violacein, an anticancer, antibacterial, antifungal, and antiviral compound. Janthinobacterium sp.
HH01 was also present in the facial skin functional strict core. The bacteria Polaromonas naphthalenivorans, capable of degrading the potentially carcinogenic naphthalene [56], was present more abundantly in the facial skin microbiome (max. mapping reads in facial skin samples= 3,194, max. mapping reads gut samples= 60). This is of relevance given that 1-methyl naphthalene is produced in carrion decomposition [57,58].
Biosurfactants represent a strong antimicrobial means of blind killing, including bacteria with antibiotic resistance that would otherwise be difficult to treat. Interestingly, we identified the biosurfactant producer fungi Yarrowia lipolytica [59] and the bacteria Rhodococcus erythropolis [60] in both facial skin and gut microbiomes. We also identified surfactin biosynthesis regulatory proteins from Flavobacteriaceae. Furthermore, annotation of the non-mapping reads with DIAMOND identified various surfactin synthetase proteins from various genera in the facial skin 9 microbiome. Surfactin is a very powerful surfactant that serves as antibacterial, antiviral, antifungal, and attacks red blood cells with deadly efficiency [61,62].

Biofilm and colonization resistance
The presence of biofilm forming bacteria has been suggested to play a protective role for the host [63]. We identified the biofilm-forming bacteria P. fluorescens [64] in higher abundance in the facial skin and as part of the strict taxonomic facial skin microbiome core. Interestingly, in the gut functional core, we identified genes coding for biofilm formation promoter proteins from F. mortiferum, such as sialic acid-binding periplasmic protein [65], and rubrerythrin [66]. And from C. perfringens, such as UDP-glucuronic acidepimerase [67], putative alginate biosynthesis protein AlgI [68], and fibronectin-binding protein [69], as well as toxin-antitoxin biofilm protein from E.
coli [70]. These results suggest that potentially pathogenic bacteria could form biofilms which allow them to thrive in the gut.

Pathogenic biofilm formation
Notably, a biofilm-mediated protection scenario requires a special interaction with the vulture's immune response, otherwise a scenario such as that in the biofilm formation in patients with cystic fibrosis (CF) would develop. The biofilm in CF patients results in clinical symptoms due to the host immune response producing tissue damage as a result of the chronic inflammation mediated by the immune complex that is trying to attack the highly resistant bacteria in the biofilm [71].
Thus, the colonization resistance mechanism of the vulture microbiome mediated by the biofilm formation requires that the vulture's immune system does not react against with a chronic inflammatory response. Interestingly, the PIK3AP1 and TNFAIP3 genes, involved in B-cell development, antigen presentation, auto-inflammation, and NF-kappa B activation, have been found to contain potentially functional altering amino acid changes in the cinereous vulture 10 (Aegypius monachus) [72]. Even more, in CF patients it has been shown that sub-minimal inhibitory concentrations of some antibiotics, such as erythromycin (from which we identified related genes in our NR gene set) and azithromycin, suppress the production of exoproducts, such as proteases and phospholipase C [73][74][75]. The inhibition of these exoproducts reduces the antigenic load and thus could lead to the decrease of immune system response. Similar modulatory mechanisms could be taking place in the vulture gut microbiome, where we identified various antibiotics.

Resistance genes
In the ResFinder database search, we identified resistance genes in 17 (52) is tetracycline, and there is no substance with resistance genes in more than 90% of the samples. In at least 50% of the facial skin and gut samples there were resistance genes for aminoglycoside, lincosamide, macrolides, and tetracycline. Only facial skin samples were found to contain resistance genes against 11 sulphonamide. Interestingly, in the gut dataset there was an abundance of genes resistant to lincosamide, which is used to treat pseudomembranous colitis caused by C. difficile [76]. the phosphotransferase enzyme family, which confers resistance to various aminoglycosides [77].
The resistance genes identified from the ResFams search can be classified as resistant to the following types of drugs: i) for treatment of various diseases, such as urinary and respiratory diseases, meningitis, tuberculosis, and against Staphylococcus and Streptococcus, ii) for the treatment of enteric diseases, and iii) for the treatment of other diseases caused by fungi or protozoa. Interestingly, there were also genes resistant to indiscriminate antibiotics, such as surfactants, organic solvents, heavy metal ions, antifolates, and carcinogens and anticarcinogens.
Among the most abundant genes from the facial skin MOCAT NR gene set we found antibiotic resistance genes for aminoglycoside from A. baumanii, and kanamycin from Staphylococcus epidermidis. In the search of the facial skin dataset against the ResFinder database we identified a differentially abundant number of resistance genes to macrolide. Given that some macrolides have antibiotic or antifungal activity [78,79], their higher abundance in the facial skin is expected taking into account that the facial skin has a significantly greater fungal diversity (P= 0.029), with many of them being plant pathogens. The facial skin microbiome also contained more resistance genes towards phenicol than the gut microbiome (x ̅ facial skin = 112,149.94, x ̅ gut = 2,177.54). Their use against infections in body parts such as eye and ear [80] could explain their higher abundance in the facial skin microbiome. In the ResFams database search we also identified resistance genes to drugs for the treatments of diseases such as enteric diseases (e.g. streptogramin [81] and bicyclomycin [82]), and tuberculosis (e.g. aminoglycoside [83] and oxazolidinones [84]).
Interestingly, many of the antibiotics with a resistance gene would also pose serious adversities to the vulture, such as macrolides [85] and cephalosporin [86], which cause digestive disturbances to humans.