The mosaicism of plasmids revealed by atypical genes detection and analysis
© Bosi et al; licensee BioMed Central Ltd. 2011
Received: 17 March 2011
Accepted: 8 August 2011
Published: 8 August 2011
Skip to main content
© Bosi et al; licensee BioMed Central Ltd. 2011
Received: 17 March 2011
Accepted: 8 August 2011
Published: 8 August 2011
From an evolutionary viewpoint, prokaryotic genomes are extremely plastic and dynamic, since large amounts of genetic material are continuously added and/or lost through promiscuous gene exchange. In this picture, plasmids play a key role, since they can be transferred between different cells and, through genetic rearrangement(s), undergo gene(s) load, leading, in turn, to the appearance of important metabolic innovations that might be relevant for cell life. Despite their central position in bacterial evolution, a massive analysis of newly acquired functional blocks [likely the result of horizontal gene transfer (HGT) events] residing on plasmids is still missing.
We have developed a computational, composition-based, pipeline to scan almost 2000 plasmids for genes that differ significantly from their hosting molecule. Plasmids atypical genes (PAGs) were about 6% of the total plasmids ORFs and, on average, each plasmid possessed 4.4 atypical genes. Nevertheless, conjugative plasmids were shown to possess an amount of atypical genes than that found in not mobilizable plasmids, providing strong support for the central role suggested for conjugative plasmids in the context of HGT. Part of the retrieved PAGs are organized into (mainly short) clusters and are involved in important biological processes (detoxification, antibiotic resistance, virulence), revealing the importance of HGT in the spreading of metabolic pathways within the whole microbial community. Lastly, our analysis revealed that PAGs mainly derive from other plasmid (rather than coming from phages and/or chromosomes), suggesting that plasmid-plasmid DNA exchange might be the primary source of metabolic innovations in this class of mobile genetic elements.
In this work we have performed the first large scale analysis of atypical genes that reside on plasmid molecules to date. Our findings on PAGs function, organization, distribution and spreading reveal the importance of plasmids-mediated HGT within the complex bacterial evolutionary network and in the dissemination of important biological traits.
Comparative whole-genome analyses have demonstrated that horizontal gene transfer (HGT) provides a significant contribution to prokaryotic genome evolution/innovation. In fact, it is very likely that a significant proportion of the genetic diversity exhibited by extant bacteria might be the result of the acquisition of sequences from more or less distantly related organisms . Indeed, HGT gives a venue for bacterial diversification by the reassortment of existing capabilities  and this formidable sexual promiscuity has given bacteria a great advantage, providing an awesome mechanism for ongoing adaptive evolution, a sort of permanently and rapidly evolving communal genome .
During evolution, HGT and recombination have shaped bacterial genomes, which today appear as complex mosaics of genes from different lineages, species, and genera .
In this picture, plasmids (collections of functional genetic modules that are organized into a stable, self-replicating entity or 'replicon'), might have played (and might still play) a major role because they can be transferred between microorganisms, thus representing natural vectors for the transfer of genes and the functions they code for . Moreover, it can be suggested that, during their evolutionary history, plasmids can undergo genetic rearrangements with either plasmids and/or cromosome(s) residing the same cytoplasm and/or with phages infecting the same cell. As a consequence, newly acquired genes can be integrated on plasmids and (eventually) be maintained.
In general, three major processes can mediate HGT among bacteria: transformation (the uptake of free DNA), transduction (DNA transfer mediated by bacteriophages) and conjugation (DNA transfer by means of plasmids or integrative conjugative elements) . However, regardless of the transfer mechanism, once that DNA has entered the recipient cell it can undergo homologous recombination or homology-facilitated illegitimate recombination and can be successfully integrated into the genome of the new host. Lastly, if the newly acquired DNA confers a selective advantage to the host, it can be maintained and, possibly, spread again through the bacterial population. Importantly, it can be surmised that, at least in the first stages following the integration event (before the amelioration process can start), exogenous sequences maintain their own peculiar compositional features [e.g. GC% and dinucleotide relative abundance difference (δ*)] that usually differ from the rest of the "new" hosting molecule; for this reason these sequences are often defined "atypical".
Atypical (and, possibly, horizontally transferred) genes detection can be pursued by composition-based methods [5, 6] that involve alignment-free features, such as GC% content and/or δ* . Compositionally oriented methods rely on the observation that some genome features (including GC% content and δ*) are typical for a given bacterial genome and similar between closely related genomes. Accordingly, recently acquired genes are likely to display anomalous composition, especially when they originated in distantly related species; moreover a different composition will also be observed in those cases in which amelioration process has been retarded. Interestingly, it has been proposed that the genome signature (a compositional parameter reflecting the dinucleotide relative abundance values between two different DNA strands) of plasmids does not resemble that of their host genome, probably indicating either absence of amelioration or a less stable relationship between plasmids and their host .
Based on composition-oriented strategies, recent analyses on large sets of bacterial and archaeal chromosomes have revealed their mosaic structure, since considerable proportions of most of them consist of horizontally acquired genes [8–10]. For example, applying a Bayesian method on 116 prokaryotic complete genomes, Nakamura et al (2004) found that the average proportion of horizontally transferred genes per genome was about 12% of all ORFs, ranging from 0.5% to 25%. Similarly Cortez et al. (2009), analysing a set of 119 bacterial and archaeal chromosomes (351111 ORFs), found that a large fraction of them was populated by atypical genes (defined as clusters of atypical genes, CAGs) (58487, 16% of all genes). Hence, this strongly indicates that archaeal and bacterial chromosomes contain an impressive proportion of recently acquired foreign genes (including ORFans, that is open reading frames without matches in current sequence databases) coming from a still largely unexplored reservoirs . Finally, the same authors found that among the identified CAGs, a large number were likely of plasmid origin . These lines of evidence suggest that genetic mobility should not be merely interpreted in terms of transportation of genes bypassing the cell barriers of prokaryotes, but rather as a perpetual flow between discrete reproductive units , i.e chromosomes and/or MGE, including plasmids. In fact, an emerging view suggests that plasmids (and MGEs in general) should be considered as mosaics of functional blocks (modules) of genes [12, 13]. Remarkably, in a few cases, the mosaic structure of plasmids (according to compositional criteria) has been demonstrated [9, 14, 15] revealing interesting insights on the dissemination of key biological traits such as antibiotic resistance, virulence and heavy metal detoxification. Accordingly, it is reasonable that the identification and the analysis of plasmid atypical genes (hereinafter PAGs) might reveal interesting insights in the (probably complex) network of intra- and inter-cellular gene transfer(s) that plasmids can face during their evolution. Indeed, PAGs might be either the outcome of (one or more) HGT(s) or of internal recombination events with chromosome(s) residing the same cytoplasm but possessing different compositional signatures.
However, to the best of our knowledge, a massive analysis of alien modules that may reside on plasmid molecules has not been undertaken up to now. Therefore, the aim of this work was to develop a statistically validated computational strategy that integrates two distinct compositional measures [GC% content and δ*] to scan nearly 2000 archaeal and bacterial plasmids for the presence of PAGs. Finally retrieved PAGs datasets have been analyzed, revealing interesting trends in the overall plasmid gene exchange network.
All the available complete plasmids, phages and chromosome sequences were downloaded from NCBI ftp site at (http://www.ncbi.nlm.nih.gov/Ftp/, as on February the 1st 2010). Concerning plasmids, we focused our attention only on those longer than 3 kb and harboring at least 2 ORFs, in order to be able to detect δ* and differences in GC% content among all the genes. This allowed to assemble a dataset of 1853 plasmids for a total of 128.569 ORFs. The complete list of plasmids analyzed in this work (together with other information such as their size, their accession codes etc.) is available as Additional file 1.
Hence, for each of the 128.569 ORFs of the 1853 plasmids (Figure 1a), we estimated both the δ* and the GC content difference (ΔGC) in respect to the corresponding source plasmid (Figure 1b and 1c). Since the distributions of these values did not follow a normal distribution (according to Kolmogorov-Smirnov test with a p-value threshold of 0.05), we used a distribution-independent procedure to evaluate the probability of each point of these distributions. performing a bootstrap sampling (Figure 1d and 1e) of all the obtained values. In other words, a probability was assigned to each of the values [e.g. P(A)] of these two distributions computing it as P(A) = n(A)/N, where n(A) is the number of times in which the observed value (of ΔGC and δ*, respectively) was greater than the other (128.569) values after N samplings. By doing so, two distinct p-values (Figure 1f and 1g) were associated to each sequence of the dataset: the first accounting for the probability of a gene to be atypical in terms of δ∗ and the other accounting for the probability of a gene to be atypical in terms of ΔGC. Further on, these two distinct p-values were integrated in a single one according to the Fisher method . In its basic form, the Fisher method is used to combine the results from two (or more) tests bearing upon the same overall null hypothesis. In other words, Fisher's method combines p-values into one test statistic (X2) using the formula: X2 = -2∑loge(pi), where pi is the p-value for the i th hypothesis test. Accordingly, (Figure 1h) only those sequences identified at a confidence interval (CI) greater than 95% were considered PAGs. Moreover, in order to explore different CI thresholds, we also collected gene sets that were identified as atypical with lower confidence values (i.e. 70%, 80% and 90%). The whole pipeline has been implemented in Perl codes and is available upon request. As it might be expected, lower CI thresholds allowed the assembly of larger PAGs dataset, ranging from 14731 (with a CI of 90%) to almost 40000 (with a CI of 70%) (see Additional file 1).
In order to identify the most likely source molecule of identified PAGS we developed a similarity-oriented computational pipeline according to which each of the identified PAG was used as a query for a BLAST  search against three different databases, each of which embedding 30000 sequences retrieved from NCBI plasmid, phage and chromosome databases (see Methods), respectively. For each of the BLAST searches, only the best BLAST hit was considered, in order to reduce any possible bias due to the presence of closely related sequences in the database that would falsely increase the number of homologs for a given ORF. This strategy was repeated 1000 times for each PAG and, for each of the 1000 runs, new plasmid, chromosome and phage databases were assembled, randomly sampling 30000 sequences from the NCBI databases. Finally, the putative source molecule was identified according to the database (plasmid, phage or chromosome) that produced the highest number of best hits after 1000 BLAST probings.
All statistical tests were performed with the R package . All other statistical analyses were performed using in-house developed Perl scripts.
PAGs general features
N. of analyzed plasmids
N. of analyzed sequences
N. of retrieved PAGs
Percentage of PAGs
Average PAGs for plasmid
PAGs in clusters (≥ 2 genes)
Cluster of PAGs (≥ 2 genes)
PAGs per clusters (on average)
Overall we found that 71 (out of 106) genera possessed an amount of PAGs that was higher (55 PAGs enriched genera) or lower (16 PAGs depleted genera) than that expected to occur by chance (p-value < 10-4).
The two best-scoring genera in terms of PAGs content were Acaryochloris and Shigella (22.5% and 17.9% of all the plasmids encoded proteins, respectively); in both cases PAGs enrichment was shown to be statistically significant. Interestingly, plasmids from A. marina were already shown to possess metabolic capabilitities that were probably acquired HGT transfer , thus confirming their mosaic structure assessed by PAGs analysis. Similarly, the mosaicism of Shigella plasmids is a well known issue [30–32] and has been demonstrated to be biologically relevant since it has probably allowed these strains to acquire pathogenic adaptation [31, 33].
Both genera resulted PAGs enriched also when lower CI thresholds were applied (respectively 48.5, 74.3 and 103.1 PAGs/plasmid at 90%, 80% and 70% threshold for Acaryochloris and 31.1, 52.4 and 66.8 in the case of Shigella representatives) although lowering CI values below 80% resulted in statistical inconsistency, likely due to the inclusion of too many false positives in the dataset.
At all CI thresholds analyzed, also PAGs depleted plasmids span over a large taxonomic range, comprising Actinobacteria, Firmicutes, Proteobacteria and Cyanobacteria. Interestingly, we found that α-proteobacterial plasmids (mainly from Sinorhizobium, Agrobacterium and Rhizobium) are higlhy represented within PAGs-depleted plasmids, suggesting that plasmids hosted by representatives of this taxonomic unit might undergo recombination/HGT events less frequently than the others. The finding that these bacteria harbor a lower number of PAGs than that expected by chance, might be accounted for by the fact that these are mainly soil inhabiting microorganisms. Indeed, it has been proposed that bacteria inhabiting this ecological niche might represent a less connected component of the overall plasmids-mediated HGT network . Accordingly, this might partially explain their lower number of PAGs. Alternatively, since it has been suggested that there is considerable gene flow between replicons in the rhizobiaceae [35, 36], it can be surmised that these bacteria frequently undergo recombination with the chromosomes of their hosting cells. However it is noteworthy that, in most cases, compositional features of rhizobiaceae replicons are pretty similar (as, for example, the GC% content (around 60%) in Sinorhizobium and Rhizobium representatives along all the replicons inhabiting the same cell). Thus, it is absolutely possbile that a fraction of this internal recombination event(s) may remain obscure due to the composition-oriented pipeline developed in this work. Interestingly, for what concerns identified PAGs (for whose identification another compositional measure was addedd to GC% content, i.e. δ*) we found that in alpha-proteobacteria chromosomal origin PAGs are more represented in respect to the whole dataset (10% and 5%, respectively, see below), suggesting that plasmids from these microorganisms might really have more genes of chromosomal origin that what is seen in other species.
As shown in Figure 5, most of the PAGs encoded proteins (around 22% of all the annotated ones) are associated to those molecular functions that are able to catalyze the movement of DNA among and within informative molecules, that is DNA integration, transposition and conjugation (7.9%, 7.3% and 7%, respectively, see red bars in Figure 5). Notably, this result partially validates the applied approach for PAGs detection, since genes that are able to move across different molecules are also expected to differ from a compositional viewpoint from the correspondig hosting molecule. Moreover, we found that another important fraction (303 sequences, corresponding to 10.9% of all the PAGs for which a putative function was retrieved) is represented by proteins involved in transcription regulation (DNA-mediated). A further investigation revealed that these proteins are mainly involved in important biological processes that are usually associated to plasmids and\or trasposons and that have a (more or less) long documentated history of HGTs, such as mercury detoxification (e.g. MerD transcriptional regulator from plasmid pEC-IMP, GI: 226807665), tetracycline resistance (e.g. TetR trascriptional regulator of Salmonella typhimurium R64 plasmid, GI: 32470145) and virulence (e.g. VirF trascriptional regulator in plasmid pSS_046 of Shigella sonnei Ss046, GI: 74314878). The presence of such proteins within the assembled PAGs dataset is intriguing. Indeed, it might be expected that the introgression of proteins capable of interfering with the overall (complex) regulatory network of the cell might be (quite) "dangerous" (from a biological viewpoint) and prone to be counterselected by the novel hosting cell. However, further analyses (see below) revealed that, a considerable amount of PAGs are embedded in more or less compact clusters, involved in processes that are known to be often spread by HGT (including virulence, antibiotic resistance and heavy metals detoxification). Accordingly, atypical trascriptional regulators might be part of this gene clusters and, consequently, might be involved in the regulation of the flanking regions. Alternatively, it can be surmised that, after the introgression of the atypical transcriptional regulator, some modifications (i.e. mutations) might have occurred, rendering the newly acquired sequence compatible with the overall "new" regulatory network, that is more easily recognizable by the transcriptional apparatus of the host cell, as experimentally demonstrated . After these two biological processes, we found that pathogenesis- and antibiotic resistance-related sequences are the most abundant among annotated PAGs (6.8% and 5.2%, respectively). Remarkably, the finding that these two biological processes are highly represented among atypical genes underlines the key role that HGT possesses in the spreading of these two important biological features within the microbial world.
As previously pointed out, PAGs may derive from homologous recombination or homology-facilitated illegitimate recombination between a given plasmid and/or other informative molecules, that is phages, chromosomes or other plasmids. Hence, we aimed at identifying the putative source molecule of PAGs. To this purpose, we have developed a computational approach (see Methods) that, on the basis of the number of orthologs present in different (and randomly assembled) plasmid, chromosome and phage databases allows to assign the most likely source molecule to a set of sequences (in our case the 8065 PAGs). It must be stated clear that this kind of strategy has only an explorative purpose and might be strongly influenced by the present content of public databases that, undoubtedly, represents just a glimpse of the real biodiversity present in nature. For this reason, the reliability of the developed approach was firstly revealed by a test on a set of 1000 likely chromosomal native sequences retrieved by Cortez et al.  from a set of 119 bacterial and archaeal chromosomes. Indeed, results of this preliminar screening showed that almost 90% of the probed sequences were correctly annotated, i.e. resulted to possess a putative chromosome origin. Furthermore, we sought to test the implemented strategy on a set of sequences of plasmid and phage origin. However, as already pointed out, in these cases a set of "core" sequences is very difficult (if not impossible) to be retrieved. Hence, the previously described pipeline was applied on two distinct randomly assembled datasets of viral and plasmids sequences, embedding 1000 sequences each. Results similar to those obtained with (likely) chromosome native sequences were obtained (84% and 82% of "correct" identifications in the case of plasmdis and phages, respectively) thus suggesting that, in most cases, the implemented strategy is able to detect the correct source molecule of a given sequence.
Among identified PAGs clusters, some possess a partially documented evolutionary history, mainly driven by HGT/recombination events as, for example, mercury resistance gene cluster(s). In fact, mercury resistance genes (mer) have been usually found embedded in a single compact operon  that, in turn, has been suggested to represent an aberrant mercury resistance transposon (namely TndPKHLK2) that, in some cases, has lost those genes responsible for its transposition . The analysis of PAGs clusters allowed the identification of (at least) 10 different mer clusters (see, Additional file 10) that showed a different composition in respect to the source molecule, thus revealing the pivotal role of HGT in the spreading of this metabolic ability across bacteria belonging to (sometimes) very different taxonomical units and inhabiting separate ecological niches. Moreover, other gene clusters (or part thereof) coding for important biological traits (i.e. antibiotic resistance, host invasion and cobalamin biosynthesis) were retrieved. For example invasion-associated genes encoding proteins involved in invasion of mammalian cells were found in atypical clusters on plasmids retrieved from different specie of Shigella genus (see Additional file 10) providing further support to the idea that one (or more) HGT envent(s) played a role in spreading this feature within representatives of this genus [32, 44]. Similarly, plasmid mediated HGT seems to have contributed to the spreading of other key metabolic traits in microbial representatives, as, for example, cobalamin biosynthesis  and tetracycline resistance for which very similar (atypical) gene clusters were retrieved from very distantly related microorganisms, including Geobacter, Halorubrum, Methylibium, Methylobacterium and Deinococcus representatives in the case of cobalamin biosynthesis and Escherichia, Klebsiella, Aeromonas, Serratia, Yersinia and Enterobacter in the case of tet genes (see Additional file 10).
The high number of PAGs retrieved in CMPs provide strong support to the idea that plasmids have played (and are still playing) a central role in microbial evolution . In fact, our data suggest that, by visiting different cells, CMPs can undergo recombination event(s) with the host's DNA molecules more frequently than NCMPs; consequently, they can probably acquire pieces of exogeneuos DNA that, in turn, can be further spread within the whole microbial communities. This idea is also partially supported by the finding that CMPs, on average, possess more and longer PAGs clusters in respect to NCMPs as reported in Additional file 12.
Within the complex evolutionary network of plasmids, new functional blocks are added and exchanged. However, to date, information on plasmids atypical regions is available only for a very limited number of plasmids and/or microorganisms.
In this work we have developed a computational pipeline to detect compositionally atypical ORFs that reside on plasmids and performed a large scale analysis of them. Implementing our strategy on a dataset of nearly 2000 plasmids we have identified 8065 PAGs, almost 6% of all the analyzed ORFs. Accordingly, these PAGS are likely the outcome of (one or more) HGT event(s), although it must be mentioned the hypothesis that, at least part of them, may derive from events of internal recombination with chromosome(s) inhabiting the same cytoplasms but that, in some cases, may possess different compositional features in respect to the corresponding plasmids (as suggested in ).
It is worth of noticing that the total amount of retrieved PAGs is, on average, lower than that estimated for chromosomes (10-15%) [10, 20]. This might be due to the high confidence interval (C.I. 95%) applied during our PAGs retrieval pipeline (Figure 1), that might have led to a partial understimatation of the actual amount of atypical ORFs that are integrated on plasmids. Indeed, applying the same pipeline for PAGs retrieval with lower CI thresholds allowed to assemble larger PAGS datasets (embedding 14731, 27201 and 38950 sequences at 70%, 80% and 90% CIs, respectively). Nevertheless analyses on these (lower confidence) assembled datasets revealed overlapping trends, suggesting that the possible exclusion of some false negatives did not influence the general conclusions that can be drawn on PAGs and, more in general, on plasmid-mediated HGT. Overall, we found that PAGs are not uniformly distributed among the sampled plasmids dataset. Indeed most of the plasmids harbor a few atypical genes or do not possess any atypical gene at all, wehreas PAGs enriched plasmids are progressively more rare. This finding is in partial agreement with previous findings on horizontal flow of plasmid genes  and suggests that all plasmids may not contribute equally to the overall horizontal flow of genes but, instead, some of them may occupy more central positions in the overall network of HGT events. In particular, this role might be covered by conjugative plasmids that have been shown to possess, on average, a higher amount of atypical regions in respect to not mobilizable ones.
Interestingly, plasmids size does not sensibly correlate with PAGs content. This result may provide important evolutionary insights, suggesting that the acquisition of exogeneous DNA (i.e. HGT) might not be the only force driving the plasmids assembly and the expansion of their coding capabilities. Indeed, it might be possible that other molecular mechanisms play a role in this process, such as gene duplication (possibly followed by evolutionary divergence) (Maida et al. unpublished data). Moreover, the fact that for a fraction of the identified PAGs it was not possible to retrieve an associated function, together with the observation that PAGs encoded proteins are, on average, shorter than not-atypical ones, points towards the presence of a fraction of pseudogenes within PAGs. Accordingly, these might have originated from unsuccessful HGT events, one of the most likely source of pseudogenes within prokaryotic genomes .
The automated functional annotation we have performed has revealed that, among the annotated PAGs, most are involved in the overall process of DNA mobilization, although other biologically relevant functions have been identified, such as transcription, pathogenesis and antibiotic resistance. The fact that we have retrieved the genes encoding for these functions associated to atypical DNA regions has important biological drawbacks, underlining the important role of HGT in the bacterial sharing of these key traits. Importantly PAGs are often found in multi-cistronic clusters embedding two or more genes. However, the fact that shorter plasmids (embedding 2 or 3 genes) are much more frequent than longer ones, probably indicates that PAGs clusters are fragmented following thier integration or that, alternatively, the transfer of shorter clusters is favoured in respect to longer gene arrays.
Finally, our analysis revealed that most of the PAGs might be of plasmid origin suggesting that plasmid-plasmid gene exchange might be favoured in respect to phage-plasmid and chromosome-plasmid ones. This is in partial agreement with previous findings  and reveals that a sort of preferential gene flow between vehicles of the same type (in our case plasmids) might exist.
MF is supported by a postdoctoral fellowship from "Fondazione Adriano Buzzati-Traverso". We are grateful to three anonymous reviewers for their useful comments and suggestions that greatly improved the manuscript.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.