During the last 30 years, a variety of extremophiles have been isolated from abyssal and hadal environments in diverse locations [1
]. These habitats are of particular interest since they are characterized by high hydrostatic pressure, low temperature, lack of nutrients and absence of light. Recently, deep sea adaptation has been investigated by sequencing two genomes of psychropiezophilic bacteria which are considered in this study. These cold and pressure-loving microorganisms are Photobacterium profundum
strain SS9 [2
] and Shewanella benthica
], hereafter called SS9 and KT99. SS9 belongs to the Vibrionaceae family and is a moderately piezophilic, γ-proteobacterium, extensively studied for its importance as a model organism for deep-sea adaptation [4
]. It was isolated from an amphipod in the Sulu Sea at 2551 m depth and displays optimum growth at 28 MPa and 15°C. On the other hand, KT99 is a deep-sea obligate piezophile heterotroph, isolated from a sample taken at 9000 m depth in the Tonga-Kermadec Trench in the Pacific Ocean [3
]. Both bacteria have a remarkable number of phylogenetically closely-related species that are adapted to shallow water conditions. In order to perform a comparative genome analysis we considered only the species for which complete genomes are available, namely V. parahaemolyticus
, V. fisheri, V. vulnificus
for the Vibrionaceae family and S. baltica
, S. oneidensis
, S. frigidimarina for
the Shewanellaceae family, all of which are mesophilic aquatic bacteria. The Shewanella genus can be subdivided into two major branches, the first one characterized by high-pressure adapted species includes S. benthica
and the other group characterized by pressure-sensitive species contains S. baltica
, S. oneidensis
and S. frigidimarina
]. A bioinformatic procedure was applied independently on both families, revealing a valuable number of orthologous genes in each of the two bacterial taxa.
Genetic and biochemical experiments have revealed that both physiological and structural adaptations are essential for high-pressure life. It has been demonstrated that membrane lipids, proteins and solutes accumulation (piezolytes) can influence bacterial growth in deep sea environments [4
]. Pressure effects on DNA replication and topology, as well as on cell division, have also been widely discussed as have the nature and regulation of genes that are important for pressure-sensing and the relevance of the transport process in piezophilic bacteria [6
The object of this study is to further characterize the effect of abyssal conditions on the evolution of SS9 and KT99 genomes. It is not obvious to establish the biological, physical and chemical parameters that are relevant to protein adaptation because abyssal and shallow water environments are extremely different in term of pressure, temperature, light and nutrients. This is an important point to be considered when the two abyssal species described in this paper are compared to phylogenetically related mesophilic species for which our knowledge on their optimal growth conditions is not always fully established. In fact, we are not able to exclude that other parameters, different from hydrostratic pressure, play a role in protein adaptation. Temperature, for example, could influence our analysis as it is one of the parameters that separates P. profundum
SS9 from the other shallow-water bacteria considered in our comparison. This parameter has probably less influence in shewanellaceae comparison since growth curves determined for shewanellaceae [5
] indicate that only S. oneidensis
can be considered a mesophilic bacterium, whereas S. benthica
, S. frigidimarina
and S. baltica
are psychrophiles. Furthermore all shallow-water bacteria considered in our comparison contain the deoxyribodipyrimidine photo-lyase gene involved in repair of UV radiation-induced DNA damage, whereas the SS9 and KT99 genomes do not code for this enzyme, which is consistent, since there is a distinct absence of sunlight in their normal deep sea environment.
In Table are summarized the main features of the bacteria considered in our study. Growth optimum was obtained from [7
] for P. profundum
SS9, from [8
] for V. parahaemolyticus
, from [5
] for S. baltica
, S. benthica
e S. frigidimarina
, while growth data for S. oneidensis
were kindly provided by Daniel I. of the University Claude Bernard (Lyon).
Features of the bacteria considered in this paper
We investigated the presence of positively selected genes in the two bacteria adapted to an abyssal environment, using as a control the phylogenetically related mesophilic species. The computation most frequently used to carry out this analysis is the non-synonymous to synonymous substitution rate ratio, defined omega (ω) [9
]. In general, an excess of non-synonymous substitutions (dN) over synonymous substitutions (dS) is considered a clear indicator of positive natural selection, because non-synonymous mutations are typically subject to strong purifying selection, whereas synonymous changes are typically neutral.
It should be considered that if we take any two species we would expect to find some genes that underwent positive selection due to the respective evolutionary history of each species. In this study we consider two families of bacteria, each represented by four species. We can assume that some genes underwent positive selection during the evolution of each species, whereas other genes underwent positive selection in clades including more than one species. To identify the genes that underwent positive selection during the evolution of a species we designed the following approach. Firstly, we chose the species for which we wanted to detect genes that underwent positive selection, in our case the deep-sea bacteria. Secondly, we compared the chosen species with the other three control species of the same family, calculating the ω values for each orthologous pair. Thirdly, we performed the calculation of the ω values also within the three control species. Finally, we selected the genes that resulted positive (i.e. underwent positive selection) in each of the three comparisons of the chosen species versus the control species but were negative in the three comparisons within the control species.
It is important to consider that it is likely that the genes selected by this approach specifically underwent positive selection in the chosen species, but this does not automatically imply that they are involved in deep-sea adaptation. This point should be considered for a correct interpretation of the results.
It is known from the literature that evolution of bacterial genes is influenced by several factors, such as protein expression level, functional class and metabolic cost of amino acid residues [11
]. Therefore we further considered the distribution of Codon Adaptation Index (CAI) versus
ω values for each orthologous gene, both in Shewanellaceae and Vibrionaceae. It had been previously defined that CAI is negatively correlated with dN [11
]. In fact, proteins expressed at high levels are generally associated with the usage of the "best" synonymous codons, resulting in a direct influence of the codon usage bias on the rate of non-synonymous substitutions in bacteria.
Genome-wide studies have already identified a number of biological processes involved in high pressure adaptation [12
]. We used genes that we obtained from the comparison between piezophiles and mesophiles (PS genes) to establish which Gene Ontology (GO) and Cluster of Ortologous Groups (COG) functional classes are targeted by natural selection in bacterial evolution. Moreover PS proteins, belonging to particular categories, can easily show PS sites mainly localized in specific domains. In order to better define the role of amino acid substitutions on deep sea adaptation, we mapped PS sites on the 3D structure as well as on protein topology in respect of trans-membrane regions.