Horizontal gene transfer greatly facilitates rapid genetic adaptation of bacteria to shifts in environmental conditions and colonization of new niches by allowing one-step acquisition of novel functions. Conjugation is a major mechanism of horizontal gene transfer mediated by conjugative plasmids and integrating conjugative elements (ICEs). While in most bacterial conjugative systems DNA translocation requires the assembly of a complex type IV secretion system (T4SS), in Actinobacteria a single DNA FtsK/SpoIIIE-like translocation protein is required. To date, the role and diversity of ICEs in Actinobacteria have received little attention. Putative ICEs were searched for in 275 genomes of Actinobacteria using HMM-profiles of proteins involved in ICE maintenance and transfer. These exhaustive analyses revealed 144 putative FtsK/SpoIIIE-type ICEs and 17 putative T4SS-type ICEs. Grouping of the ICEs based on the phylogenetic analyses of maintenance and transfer proteins revealed extensive exchanges between different sub-families of ICEs. 17 ICEs were found in Actinobacteria from the genus Frankia, globally important nitrogen-fixing microorganisms that establish root nodule symbioses with actinorhizal plants. Structural analysis of ICEs from Frankia revealed their unexpected diversity and a vast array of predicted adaptive functions. Frankia ICEs were found to excise by site-specific recombination from their host's chromosome in vitro and in planta suggesting that they are functional mobile elements whether Frankiae live as soil saprophytes or plant endosymbionts. Phylogenetic analyses of proteins involved in ICEs maintenance and transfer suggests that active exchange between ICEs cargo-borne and chromosomal genes took place within the Actinomycetales order. Functionality of Frankia ICEs in vitro as well as in planta lets us anticipate that conjugation and ICEs could allow the development of genetic manipulation tools for this challenging microorganism and for many other Actinobacteria.
Citation: Ghinet MG, Bordeleau E, Beaudin J, Brzezinski R, Roy S, Burrus V (2011) Uncovering the Prevalence and Diversity of Integrating Conjugative Elements in Actinobacteria. PLoS ONE 6(11): e27846. https://doi.org/10.1371/journal.pone.0027846
Editor: Niyaz Ahmed, University of Hyderabad, India
Received: September 9, 2011; Accepted: October 26, 2011; Published: November 16, 2011
Copyright: © 2011 Ghinet et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: Financial support for this work was provided by Centre d'étude et de valorisation de la diversité microbienne (Université de Sherbrooke), Centre SÈVE (Fonds québecois de la recherche sur la nature et les technologies), and Discovery Grants from the Natural Sciences and Engineering Research Council of Canada (NSERC) to RB, SR, and VB. EB is the recipient of an Alexander Graham Bell Canada Graduate Scholarship from NSERC. VB holds a Canada Research Chair in molecular biology, impact and evolution of bacterial mobile elements. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Actinobacteria are found in many different ecological niches. These high G+C Gram-positive bacteria are soil and aquatic inhabitants (e.g., Streptomyces, Micromonospora, Rhodococcus), plant symbionts (e.g., Frankia), plant and animal pathogens (e.g., Corynebacterium, Mycobacterium, Nocardia), or gastrointestinal commensals (e.g., Bifidobacterium). Nitrogen-fixing actinobacteria belonging to genus Frankia live as soil saprophytes and as endophytic symbionts in over 200 plant species . These host plants (actinorhizal plants) are found on nearly all continents, and the contribution of Frankia and actinorhizal plants to global nitrogen fixation is estimated at 25% , . Frankiae are understudied compared to other actinomycetes and this is largely attributable to difficulties in isolation and maintenance of actively proliferating cultures. Nevertheless, the global environmental importance of Frankiae command more research to better understand their interactions with other microorganisms, and host plants.
Frankiae have circular chromosomes, in contrast to streptomycetes. Genome sizes are highly variable, as exemplified by strains Cci3, ACN14a and EAN1pec (5.4, 7.5, and 9.4 Mbp, respectively) . This diversity suggests the existence of a spectrum of lifestyles ranging from quasi-obligatory symbiosis (small genome size) to free-living status with the possibility of symbiosis in a wide array of host plants (large genome size). The size of the larger genomes of Frankiae is also reminiscent of that of related, GC-rich, biotechnologically important genus Streptomyces. However, in great contrast to what is known of streptomycete genetics, much remains to be learned about the horizontal exchange of genetic material in Frankiae. Such knowledge is essential to decipher the evolution of members of this genus and to achieve stable genetic manipulation in these strains – a goal that has eluded researchers to date.
A growing number of reports indicate that horizontal gene transfer is an essential mechanism by which some actinobacterial species acquired pathogenesis-related functions , , . Mobile genetic elements are also responsible for extensive genomic rearrangements revealed by structural genomics and observed even among related actinobacterial species . Actinomycete integrative and conjugative elements (AICEs), aka. integrating conjugative plasmids, are particularly prevalent in the genomes of several species of Streptomyces . AICEs belong to the broader class of integrative and conjugative elements (ICEs) usually described in Gram-negative bacteria and in the Firmicutes. Like temperate bacteriophages, ICEs integrate into and replicate with the chromosome of the bacterial host , , . The site-specific integration of ICEs in the host chromosome is usually mediated by a tyrosine recombinase or occasionally by a serine recombinase, encoded by an int gene. Site-specific recombination occurs between two short identical or nearly identical sequences located in attachment sites, one on the element (attP) and the other one on the chromosome (attB). Upon various conditions, ICEs can excise from the chromosome to form circular covalently closed molecules. Like most conjugative plasmids, ICEs disseminate via conjugation. In Gram-negative bacteria and Firmicutes, conjugative DNA transfer typically requires the assembly of a type IV secretion system (T4SS) which is involved in translocation of the DNA to recipient cells . Biochemical processing of the DNA molecule to transfer is initiated at a specific cis-acting site called the origin of transfer (oriT), which is bound by a DNA relaxase (Mob protein) and other auxiliary proteins, forming altogether a nucleoprotein complex called the relaxosome . The phosphodiesterase activity of the relaxase mediates a strand-specific cleavage within oriT, allowing the unwinding and 5′ to 3′ transfer of a single-stranded DNA (T-strand) to the recipient cell. A coupling protein (T4CP) links the relaxosome to the T4SS and is thought to form the conjugative pore through which the T-strand is translocated to the recipient cell. While in the donor cell the complementary strand is used as a template for replacement strand synthesis, within the recipient cell host the T-strand is converted into double-stranded DNA that can be either re-circularized and/or recombined into the recipient chromosome.
Interestingly, inter-mycelial conjugative transfer of ICEs identified in actinomycetes significantly differs from this general mechanism as it proceeds via translocation of unprocessed double-stranded DNA. This mechanism is reminiscent of chromosome partitioning during sporulation and cell division (reviewed by te Poele et al. ). Furthermore, once excised from the chromosome, actinomycete ICEs replicate autonomously like genuine plasmids , . They encode a replication initiator protein and carry an origin of replication. Mutations in either of these two elements has been shown to lead to loss of transfer of the prototypical actinomycete ICE pSAM2, indicating that autonomous replication is required for conjugative transfer to recipient cells , , . AICEs have been developed into useful tools for genetic engineering of actinobacteria .
In this study, we investigated the prevalence and diversity of ICEs in Actinobacteria, and focused particularly on ICEs in six Frankia strains, EAN1pec, ACN14a, Cci3, EuI1c, EUN1f and Frankia symbiont of Datisca glomerata. Prior to this study, three AICEs, Fean5323, Fean6303 and Faln5456 originating from two members of the genus Frankia, have been briefly described . Our study reports the frequent occurrence of AICEs in Frankia genomes. Their ability to excise from the chromosome as circular molecules both in vitro and in planta was examined as excision is the initial step preceding their eventual transfer to a recipient cell via conjugation.
Results and discussion
Identification of putative AICEs in the genomes of Actinobacteria
To address the prevalence, repartition and diversity of AICEs in Actinobacteria, we carried out a large-scale in silico analysis using data extracted from the RefSeq database . The predicted proteomes encoded by 275 genomes of Actinobacteria were screened using HHM profiles of protein orthologs associated with the functions of core modules of ICEs previously identified in actinomycetes  (Figure 1 and Table 1): (i) integrases (Int) of the tyrosine or serine family of recombinases that are involved in ICE integration and excision, (ii) replication initiator proteins (Rep) of the RepSA- and RepAM-type, and putative polymerases of the Prim-Pol-type (bifunctional DNA primase/polymerase), and (iii) FtsK/SpoIIIE domain-containing proteins (Tra) that are involved in translocation of double-stranded chromosomal DNA. This analysis revealed a large number of possible AICE-associated proteins (Table 2); yet many of these hits may be part of other types of mobile elements (transposons, prophages, integrated plasmids) or have functions unrelated to mobile elements. For instance, FtsK/SpoIIIE domains are known to be found in proteins rescuing the remainder of chromosomal DNA into the appropriate daughter-cell compartments during constriction of the septal membranes in prokaryotic cell division (FtsK) , ,  and in asymmetric division of sporulating Bacillus subtilis (SpoIIIE) . Similarly, subsets of tyrosine recombinases such as MrpA encoded by the Streptomyces coelicolor A3(2) plasmid SCP2  catalyze the resolution of plasmid multimers into the monomeric form after replication. Therefore, to limit our investigation to genuine putative AICEs, 30- to 60-kb chromosomal regions containing genes encoding at least one of each of the three components (Int, Tra and Rep) were further considered in our analyses (Figure 1). However, we excluded the very few cases for which the int gene was located in between the rep and tra genes as recombinase genes are typically located adjacent to one of the two att sites flanking an integrated ICE or prophage. The chromosomal fragments coding for two or less of these components were also excluded and likely correspond to AICE remnants, prophages, transposons and other genomic islands. A total of 139 putative AICEs were identified in chromosomal sequences (Table 2 and S1). In addition, the analysis of 176 actinobacterial plasmids revealed 4 additional ICEs: the mega plasmid pNDAS01 contains a putative ICE, while plasmids pSA1.1, pSLS and pWTY27 likely correspond to ICEs in their excised form (Table S1).
Possible combination of integration/excision, replication and conjugative transfer modules are represented. For each functional module, specific domain-containing proteins where searched for as follow: Tra (FtsK/SpoIIIE, PF01580), Int-Tyr (Phage integrase, PF00589), Int-Ser (Recombinase, PF07508), RepSA (this study), RepAM (DUF3631, PF12307) and Prim-pol (Prim-pol, PF09250). AICEs Prim-pol proteins are encoded along with putative replication proteins (RepPP) or can be associated with RepAM replication initiator proteins.
This first in silico analysis unambiguously showed that based on the criteria indicated above, AICEs are constrained to members of the Actinomycetales order (Figure 2A). On the contrary, the closely related Bifidobateriales order, and the other Actinobacteria subclasses (Acidimicrobidae, Coriobacteridae and Rubrobacteridae) appeared to be completely devoid of AICEs. This observation echoes a similar conclusion that was drawn by te Poele et al.  from a smaller genome sample. However, since this initial analysis was based on specific assumptions concerning the nature and gene conservation among known AICEs, it was inherently biased. To circumvent this bias, we extended our analysis to include 15 other families of replication initiator proteins of plasmid or viral origin. Surprisingly, using this approach, we were able to identify only one additional putative ICE, which relies on a Rep2-type replication initiator protein, in a strain of Bifidobacterium longum (Table S1).
(A) Distribution per suborder of actinobacteria. (B) Distribution per environmental niche of the original host. (C) Distribution as a function of genome size. (A and B) Median, standard deviation (rectangles) and extreme values (brackets) are shown. (B and C) Only the Actinomycetales are shown.
We then analyzed the distribution of AICEs among Actinobacteria as a function of their host's environmental niche. In general, species isolated from plant, soil and water samples tend to contain a significantly larger number of AICEs per genome than those isolated from dairy products, animal, insects or sediments (Figure 2B). This observation suggests that AICEs could confer specific advantages to their respective bacterial hosts in these specialized ecological niches. However, additional extensive sequence analyses and functional characterization of the gene cargo borne by AICEs will be necessary to determine conclusively what benefits these mobile elements confer to their host. As expected, large actinomycete genomes tend to bear significantly more AICEs than small genomes (Figure 2C). It is not clear to date whether the acquisition of multiple AICEs by a single genome is constrained by specific barriers, such as the impossibility for two different AICEs to occupy the same integration site and form tandem arrays, or by the existence of specific AICE-encoded immunity- or exclusion-like systems. For example, Pif, a NUDIX hydrolase domain-containing protein encoded by pSAM2, has been shown to act in the recipient cell to prevent redundant exchange between two cells harboring the same conjugative element . According to our analysis, AICEs often bear genes encoding NUDIX hydrolase domain-containing proteins, suggesting that conjugal immunity is widespread (data not shown). However, the molecular mechanism of conjugal immunity mediated by proteins such as Pif and the range of AICEs that a single Pif-like protein can target for immunity remain unknown.
AICE recombination functions and integration sites
Among the 144 AICEs that we detected in our analysis, only 19 (∼14%) were found to rely on an integrase of the serine recombinase family for their integration into and excision from the chromosome (Figure 3A and Table S1). All the other AICEs appear to rely on integrases of the tyrosine recombinase family instead (Figure 3B and Table S1). This imbalanced proportion correlates rather well with what is generally observed for ICEs and temperate bacteriophages; yet surprisingly, more than half of the AICEs encoding serine recombinases were found in the order Corynebacterineae. In fact, all Mycobacterium AICEs but Mkan24688 encode serine recombinases, which clustered in two groups. The first group contains the recombinases encoded by Mmar2129 and Mav3790 as well as Srot1958 from Segniliparus rotundus, another member of Corynebacterineae. These AICEs are integrated into protein-encoding genes. The second group has 8 recombinases which originate from different species of Mycobacterium yet exhibit high sequence similarities, and catalyze the integration into the same tRNA Leu gene (Figure 3A). Serine recombinase-mediated integration into tRNA genes is rare . In fact to our knowledge, only φRSM1, a filamentous phage infecting Ralstonia solanacearum, was found to encode a serine recombinase catalyzing its integration into the 3′ end of a tRNA Ser gene . Interestingly, no AICE was detected in any of the 30 genomes of Mycobacterium tuberculosis or in the 3 genomes of Mycobacterium bovis that were part of our analysis.
(A) Serine recombinases. (B) Tyrosine recombinases. Sub-families were defined as clades containing at least 4 members with a distance cut-off of 1 and with a >80 bootstraps support. Label colors indicate the suborders of Actinomycetales within which the AICEs were found (see legend). Colors in the outer wheel in panel A and the dark green strip in panel B indicate the type of tRNA genes into which each AICE is integrated (see Table S1 for details). Black dots in panel A for Ssrg02924 and Sslg03435 indicate that while no tRNA-encoding gene was detected as their integration sites, the att sites nearest to their respective int genes were identical to the ones of their closest relatives, i.e. pSAM2 and Ssbg00935, respectively.
Phylogenetic analyses of 125 AICE integrases of the tyrosine recombinase family revealed several groups of proteins clustering into subfamilies which possess more extensive amino acid sequence similarities (Figure 3B). Eight such groups containing four or more integrases were identified in the phylogenic tree of tyrosine recombinases. One of these subfamilies, pSE211, was previously identified by Williams . Eight new subfamilies clearly emerge from our analysis, SLP1, pSLS, pMR2, pSAM2, Fcci3390, pSA1.1, Feun3577 and Fean5518. No single subfamily was found to be exclusive of a particular species, genus or even sub-order. One of the possible benefits for AICEs of the availability of such large pool of different integrase subfamilies is that they are more likely to target different integration sites, allowing them to stably coexist within the same genome. Different mobile genetic elements sharing the same integrase and integration site have been shown to form tandem arrays that are inherently unstable, leading to recombination between elements composing the array, incompatibility and/or loss , , , , .
We observed that the vast majority of AICE encoding tyrosine recombinases were integrated in the 3′ end of a tRNA gene, with the exception of all AICEs encoding integrases belonging to the pSLS sub-family. Several past reports revealed a strong preference for tRNA and tmRNA gene sequences as prophage integration sites , , . In at least two instances, i.e. Sslg03435 and Ssrg2924, no tRNA gene could be readily identified as the integration site by tRNAscan-SE; yet the att sequences flanking these two AICEs were found to be identical to the att sites flanking Ssbg03435 and pSAM2 respectively, which encode closely related integrases and integrate into the 3′ end of tRNA genes (Figure 3B). This observation indicates that probable tRNA gene remnants can be suitable integration sites. As expected, closely related integrases seem to target the same tRNA gene in their respective hosts. For example, the 6 closely related pSE211-type integrases encoded by the group Ssqg04320/Sslg03699 catalyze integration into a tRNA-Thr gene. Closely related integrases can also catalyze integration into different tRNA genes, as observed for Fcci1033/Fean5534 (tRNA-Lys) and Faln5456 (tRNA-Met). Mobile elements such as prophages and ICEs likely take advantage of the frequent association between tyrosine recombinase with tRNA genes as integration sites. While the benefits of such associations have not yet been fully elucidated, several suitable hypotheses have been proposed including sequence reliability over protein-coding sequences, transcriptional coupling and so on .
Replication functions of AICEs
Replication of the first described AICE, pSAM2, was found to be dependent upon the protein RepSA , . As other replication initiator proteins involved in rolling-circle replication (RCR), RepSA was shown to bind to the binding region of the double-stranded origin (dso) and to introduce a site-specific nick in the nick region of dso, leaving a 3′-OH end used for priming the elongation by host encoded replication proteins (DNA polymerase III, helicase) . Surprisingly, phylogenetic analysis of replication initiator proteins detected with the RepSA HMM profile that we generated, revealed clustering in two major distantly related subfamilies that were not previously reported as phylogenetically distinct by te Poele et al. . The first subfamily, RepSASAM2, counts 69 proteins and is represented by the RepSA protein of pSAM2. The second subfamily, RepSAMR2, counts 30 proteins and is represented by the RepSA protein of pMR2 (Figure 4A). Interestingly, all RepSA proteins encoded by Streptomycineae AICEs belong to the RepSASAM2 subfamily. RepSA proteins encoded by Mmu2220/Mmul0108 and Avis01490 do not seem to belong to either subfamily; instead they cluster into a distant monophyletic clade of Actinomycineae replication initiator proteins.
(A) Phylogenetic tree of RepSA proteins. Sub-families were defined as clades containing at least 4 members with a distance cut-off of 1 and with a >80 bootstraps support. Label colors indicate the suborders of Actinomycetales within which the AICEs were found (see legend). (B) Comparison of the conserved motifs found in RepSA-like proteins from AICEs with the conserved motifs found in the catalytic domain of rolling-circle initiator proteins of plasmids and single-stranded DNA coliphages. The consensus sequences shown are as previously reported by Del Solar et al. . Motif 2 (HUH or His-hydrophobic-His) corresponds to the putative divalent cation-binding site and motif 3 contains the tyrosine catalytic residue. Mmu2220, Mmul0108 and Avis01496 were not included in the alignment used to generate the logo sequence of the RepSASAM2 conserved motifs.
Replication initiator proteins from distant RCR replicons (i.e., bacteriophages, plasmids, eukaryotic viruses) contain three conserved amino acids motifs , . Motif 1 was initially of unknown function. Motif 2 contains a HUH motif (histidine-hydrophobic-histidine) important for metal-binding whereas motif 3 contains the catalytic tyrosine residue. Motif 3 was used by Koonin and Ilyina  to define 2 superfamilies of RCR replication initiator proteins; superfamily I contains two conserved tyrosine residues whereas superfamily II contains only one. Based on the conservation of these motifs, the pSAM2 RepSA protein was initially found to be related to replication initiator protein RepA of coliphage φX174 by Hagège et al. . Later te Poele et al.  reported that the replicon pSAM2 belongs to the pC194 family based on the conserved motifs in RepSA and nick site at dso. To further examine the functional relatedness of the RepSA-like proteins of the AICEs detected in our analysis, we generated amino acid logo sequences for the 3 conserved motifs. While the RepSASAM2 and RepSAMR2 protein motifs 2 and 3 exhibit some similarity with the φX174 Rep family, they clearly diverge at specific positions (Figure 4B). Solely based on motif 3, RepSASAM2 proteins are clearly related to the φX174-group of RCR initiators defined by Wigel and Seitz  while the RepSAMR2 proteins would define an entirely new group of RCR initiators. This hypothesis is supported by motif 1 of the RepSAMR2 group which clearly differs from motif 1 of φX174 or pC194, while motif 1 of RepSASAM2 group is very similar to these RCR initiators (Figure 4B). A sequence analysis of other RCR initiator proteins and their putative DNA binding sequences, albeit from the distant plant infecting geminiviruses family, suggests that motif 1 is involved in the dso binding region specificity , . Consequently, we predict that RepSA proteins cut the dso nick region using a similar mechanism but the RepSAMR2 would recognize discrete dso binding regions. Such recognition differences may facilitate the coexistence of similar yet different AICE within the same genome.
Other less-well characterized proteins likely involved in AICE replication were also sought for in order to unearth additional putative AICEs. The RepAM proteins, first demonstrated to be responsible for autonomous replication in AICE pMEA300 from Amycolatopsis methanolica , are predicted to function as RCR initiator proteins. Although they lack similarity to other known replication initiator protein, their putative nick sites were shown to be similar to those of the pC194 family of RCR plasmids . Identification of 14 new AICEs relying on RepAM proteins reveals such AICEs are not constrained to the Pseudonocardinae unlike previously reported  (Figure 5A). Phylogenetic analysis revealed 3 clades of RepAM initiators, one of which grouping AICEs only found in Streptomyces. Interestingly, RepAM initiator genes of all three clades are occasionally associated with a gene coding for a Prim-Pol domain-containing protein (Figure 5B and data not shown).
(A) RepAM (DUF3631) domain-containing proteins. (B) Prim-Pol (bifunctional DNA primase/polymerase) domain-containing proteins Label colors indicate the suborders of Actinomycetales within which the AICEs were found (see legend). The color strip in panel B indicates the type of associated replication protein: dark blue, RepAM; light blue, RepPP.
Prim-Pol proteins could be involved in AICE replication as well and were identified within one of the two region important for the replication of AICE SLP1 from S. coelicolor A3(2) . This region contains a small operon consisting of a Prim-Pol gene, SCO4617, followed by a putative replication gene, SCO4618, different from repAM. Our analysis revealed 40 AICEs encoding a Prim-Pol domain-protein (Figure 3A). 12 of these AICEs were found to bear a repAM gene immediately adjacent and 16 were found to carry a SCO4618-like gene (Figure 3B). Two ICEs, Sgha02332 and Fdat2245, were found to respectively encode a Poxvirus_D5 (PF03288, PF08706) domain- and a DNA_pol_A (PF00476) domain-containing protein. Other proteins encoded by genes associated to Prim-Pol genes in the remaining AICEs do not exhibit any similarity with any replication or DNA processing proteins known to date, nor contain domains related to this function (data not shown). Yet together with the Prim-Pol genes, these genes likely are unforeseen AICE replication modules.
AICEs replication modules are diverse. Interestingly, none of them seem to be exclusive to phylogenetically related bacteria. Consequently, some of these mobile elements may have been exchanged between distant bacteria. Additionally, AICEs from distant bacteria containing similar replication modules may have arisen from distinct recombination events with other mobile elements (e.g., plasmids, phages), although this seems less likely.
Unlike most known conjugation systems, AICEs and Streptomyces conjugative plasmids transfer as double stranded circular DNA molecules and do not require a genuine T4SS . AICEs conjugative transfer seems to only require a single transfer protein, Tra or TraB, which recognize and bind the cis-acting locus (clt) . Recently, these Tra proteins were shown to be structurally and functionally related to septal DNA translocators of the FtsK/SpoIIIE family, gathering into hexameric channel structures and forming pores in lipid bilayers .
AICEs Tra proteins identified by our analysis clustered in several monophyletic clades (Figure 6). The pSAM2 and pMR2 clades, named after the eponym elements, contain the largest number of AICEs, which interestingly all encode RepSA-like RCR initiators. Three other clades, pMEA100, pSE211 and SLP1, each groups a smaller number of AICEs, which exclusively rely on RepAM and/or Prim-Pol proteins for replication. The pSAM2 and pMR2 clades are devoid of Tra protein encoded by Streptomyces conjugative plasmids (pSG5, pSVH1, pJV1, pIJ101, pJV1, pFP1 and SLP2), which were also included in our phylogenetic analyses. This suggests that exchanges of genetic material between AICE and Streptomyces conjugative plasmids are rare events. A noticeable exception is the Tra protein of pJV1, which belong to the SLP1 clade. pJV1 Tra is very similar to the Tra protein of AICE pSLS (76% identity). In addition to harboring orthologous intermycelial transfer genes, pJV1 and pSLS also carry similar intramycelial spread genes (spdB123) contiguous to tra (Figure S1). However, plasmid pJV1 does not code for an integrase and does not code for Rep proteins similar to those encoded by pSLS. This example clearly indicates that some of the AICEs and conjugative plasmids could have arisen from common ancestors and/or have exchanged functional modules.
Sub-families were defined as clades containing at least 4 members with a distance cut-off of 2 and with a >80 bootstraps support. Label colors indicate the suborders of Actinomycetales within which the AICEs were found (see legend). Outer color strips indicate AICEs encoding Prim-Pol (red), RepAM (dark blue), RepPP (light blue), and RepSASAM2 (green), RepSAMR2 (light green) and other RepSA-like (dark green) replication proteins. Black dots indicate AICEs encoding serine site-specific serine recombinases.
Conjugative plasmids in Streptomyces have been shown to mediate chromosomal gene transfer. pIJ101-induced mobilization of chromosomal genes in Streptomyces lividans has been shown to occur without prior plasmid integration into the chromosome . Recently, Vogelmann et al  reported that TraB of the Streptomyces venezuelae conjugative plasmid pSVH1 binds several clt-like regions in the whole chromosome of S. coelicolor suggesting a yet elusive mechanism that likely involves Tra binding onto clt-like sequences on the chromosome eventually promoting transfer of chromosomal DNA fragments into the recipient cells. Consequently, the numerous Tra-dependent AICEs identified herein represent a huge potential for exchange of chromosomal genes within the Actinomycetales order in addition to the AICE-borne cargo of auxiliary genes.
Identification of putative T4SS-based ICEs in the genomes of actinobacteria
Furthermore, we also looked for putative ICEs relying on a T4SS-like conjugative machinery similar to the one encoded by conjugative elements of Gram-negative bacteria  and of Firmicutes such as ICEBs1 from Bacillus subtilis  or the enterococcal plasmid pIP501 . For that purpose, in addition to an int gene, we sought for (i) a gene coding for a putative T4CP, (ii) a gene coding for a VirB4/TraC-like T4SS component, and (iii) a gene coding for a putative DNA relaxase (Mob) (Table 1). T4CP proteins are VirD4 homologs of the Agrobacterium T-DNA transfer system, which is a distant relative of the FtsK/SpoIIIE domain . VirB4/TraC-like T4SS components usually have ATPase activity and are involved in the assembly of the pilus in Gram-negative bacteria . Relaxases have been shown to be related to several types of DNA-processing proteins including RCR proteins , . For this study, we limited our analysis to relaxases of the MOBF family (TrwC relaxase domain) whose members are well represented in the Actinobacteria . With a few known exceptions , , it is generally assumed that such ICEs do not replicate autonomously. Therefore the presence of a replication initiator protein was not considered as a requirement for this identification as this role can be eventually played by the Mob protein . Based on these criteria, we identified 17 putative ICEs in all major Actinobacteria suborders (Table S2). Among them, one was found in Frankia sp. Cci3 stain. Analysis of the molecular structure of Fcci3350 revealed that this ICE encodes a lantibiotic dehydratase-like protein. Proteins from this family participate to the biosynthesis of lantibiotic, a ribosomally synthesised antimicrobial agent . This T4SS-like ICE also encodes genes that may participate or limit inter-bacterial transfer of mobile genetic elements (Table S3). Several components of a putative CRISPR-Cas system, an RNA-based immune system targeting bacteriophages and plasmids in bacteria , are encoded by Fcci3350. However, this system is likely not functional as shown for the linear plasmid pSHK1 from Streptomyces . Additional characterization of these new ICEs is beyond the scope of this article and will be the topic of another publication.
AICEs and related elements in Frankia: general structure, synteny and gene conservation
Frankiae and their symbioses are present on all continents except Antarctica, and it is estimated that Frankia sp. is responsible for 15–25% of global nitrogen fixation , . The critical role of this microorganism and its symbiosis in the global nitrogen cycle underscores the importance of deciphering their genetic inheritance, diversity, genetic plasticity and evolution. Research to date has barely touched these questions, often limited by the fastidious nature of this microorganism in the laboratory and the complete lack of genetic tools. AICEs as a whole and their building blocks, used to generate new genetic tools, could significantly facilitate the study of Frankia and their symbiotic interactions with plants.
Frankia strains are among the species of Actinomycetales counting the highest number of AICEs per genome on average (Figure 2A). In this study 13 new AICEs were found in six genomes of Frankia in addition to the 3 AICEs identified by te Poele et al.  (Figure 7 and Table S1). The 16 Frankia AICEs have an average G+C content of 68%; yet a subset carries genes with a G+C content below 55% (dashed black arrows in Figure 7) suggesting that significant amounts of genetic material have been acquired outside the Actinobacteria. The synteny between the 16 Frankia AICEs was examined to better understand the relationships linking AICEs found in the same or different Frankia strains. This sequence analysis revealed a clear segregation of AICEs into three families (Figure 7 and S2).
The name, size and type of tRNA gene within which AICEs are inserted are indicated. Colour coding: orange, recombination; red, recombination directionality; yellow and gray, replication; blue, intermycelial transfer; lime green, regulation; dark green, transcriptional regulator LacI; pink, IS elements; mauve and fuchsia, conserved hypothetical proteins; dashed black arrows, ORFs with G+C contents <55%. Arrows indicate the noncoding sequence regions that were deleted in order to simplify the schematic representation of Fean6303. pept, peptidase C14 caspase catalytic subunit P20; HD, metal dependent phosphohydrolase; hap, Peptidase S1 and S6 chymotrypsin/Hap; LigT, 2′–5′ RNA-ligase; PhyH, phytanoyl-CoA dioxygenase; MTase_11, methyltransferase, domain MTase_11; NodU, carbamoyltransferase; PurN, phosphoribosylglycinamide formyltransferase; akr, aldo/keto reductase; amt, aminotransferase class I and II ; P450, cytochrome P450; hap, 2-alkenal reductase; NB-ARC, NB-ARC domain protein; p-kin, putative protein kinase; pcmt, protein-L-isoaspartate(D-aspartate) O-methyltransferase; php, PHP -like protein; MT_19, putative S-adenosyl-L-methionine-dependent methyltransferase, pks, Putative modular polyketide synthase; cbiA, CobQ/CobB/MinD/ParA nucleotide binding domain protein; hth, helix-turn-helix domain protein; phz, phenazine biosynthesis PhzC/PhzF protein; nbp, nucleotide-binding protein; ssb, single-strand binding protein; C2H2-lp, zinc finger - C2H2-like protein; if-2, translation initiation factor IF-2; whiB, transcription factor WhiB; cutA1, divalent ion tolerance protein CutA1; C5-MT, C-5 cytosine-specific DNA methylase; NTP-ase, putative signal transduction protein with Nacht domain – NTPase; DNA_pol3_beta_2, DNA polymerase III beta subunit, central domain.
The first family regroups ten AICEs (Fean5323, Fean5518, Fean5534, Fean6303, Faln1739, Faln5456, Fcci1033, FeuI0027, Feun0941 and Feun3577) that are characterized by a conserved structural module containing pra, tra, rep, xis and int (Table S4). Phylogenetic analysis showed that all encode a pMR2-type Tra protein and a RepSAMR2 RCR initiator protein (Figure 4 and 6). Although int was found to be the most divergent gene (Figure 3B), these observations suggest that these 10 AICEs all derive from a common ancestor. Interestingly, Xis and Pra encoded by AICEs originating from the same strain share up to 97% identity suggesting potential cross-talks between elements of the same family. For instance, transactivation of excision and conjugative transfer was reported in cells harbouring multiple copies of the ICE Tn916 from Enterococcus faecalis . Likewise, coordinated induction of the lytic cycle of Gifsy prophages was recently reported in polylysogenic strains of Salmonella .
The second family of AICEs includes the two smallest AICEs identified in Frankia, Fcci4274 and Fdat4298. While both encode a pMR2-type Tra protein, unlike AICEs of the previous family, both encode a RepSASAM2 RCR initiator (Figure 4 and 6) suggesting that module exchange took place between AICEs.
The third family includes Faln2929 and Fcci3390. Like AICEs of the second family, both elements retain pMR2-type tra genes and RepSASAM2 RCR initiator rep genes. However, they encode completely different int genes (Figure 3, 4 and 6). In Faln2929, int encodes a serine recombinase, which is convergent with tra and rep. Fcci3390 appears to correspond to a circular permutation of the classical AICE att-tra-rep-xis-int-att structure into an att-int-tra-rep-xis-att structure.
Finally, FeuI6863 and Fdat2245 do not seem to share any extended sequence identities with each other or with any AICEs of the 3 previous families. While the Tra proteins encoded by these two ICEs are distantly related to those encoded by pMEA300 and pSE222, they encode different replication initiator proteins and Fdat2245 is the only Frankia AICE encoding a Prim-pol replication protein (Figure 4, 5 and 6). Altogether, these observations suggest that FeuI6863 and Fdat2245 are recent acquisitions in Frankia species.
Fean1457, Fcci0407, Fcci1144 and Feul5809 form a group apart as they lack either a tra or rep gene (Figure S3 and Table S5). Therefore, this group may be considered as remnant AICEs or mobilizable integrating elements.
Integration and excision of Frankia AICEs
With the exception of Faln2929, which encodes a serine recombinase and disrupts a gene encoding a putative endonuclease (Fraal2928, data not shown), all Frankia AICEs are site-specifically integrated into the 3′ end of tRNA genes, a process likely mediated by the tyrosine recombinase that they encode (Figure 3 and 7). These integrases grouped within pSE211, Fean5518, pSA1.1, Fcci3330 and SLP1 sub-families (Figure 3B). While integration of AICEs is mediated by the integrase alone, excision requires in most cases the concerted expression of a recombination directionality factor (RDF), usually called excisionase or Xis protein , . All the AICEs identified but Faln 2929 carry a gene coding for a putative RDF (Figure 7).
Most Frankia AICEs exhibit the typical att-tra-rep-int-att structure, with a few exceptions. Fean5518 and Fean5534 are integrated in a tandem fashion into a tRNA-Lys gene in Frankia sp. EAN1pec. Such organization is also found in Streptomyces avermitilis for the AICEs Sav3708 and Sav3728 (te Poele et al. 2008 and this study), which are also integrated in tandem fashion in a tRNA-Arg gene. Interestingly, while Fean5518 and Fean5534 encode two closely related integrases, Sav3708 and Sav3728 encode integrases belonging to two different sub-families (Figure 3B). Another anomalous organization is found for Fean6303 which contains three direct repeated sequences extending beyond the 3′ end of the tRNA-Arg gene. Two of these repeated sequences correspond to two different attL sites, attL and attL2, and the third one is attR (Figure 7). The first attL (81 bp) defines, together with attR, an AICE 5 kb shorter than the structure previously described by te Poele et al. . These findings suggest that Fean6303 is able to excise as two different circular molecules of 15.3 kb and 32.3 kb. A similar organization was reported for ICEs found in the lactic acid bacteria Streptococcus thermophilus . Finally, Fcci3390 bears two different int genes separated by an attL site and a gene coding for a hypothetical protein. Interestingly, unlike most AICEs, these two int genes are located upstream of the tra gene instead of downstream of rep.
Excision of ICEs from the chromosome is the first step preceding their conjugative transfer. Together, the integrase and the RDF Xis catalyze a site-specific recombination event between the attL and attR attachment sites flanking the ICE to generate an attB site on the chromosome and an attP site on the circularized ICE. The latter can readily be detected using a nested PCR assay. To investigate whether the putative AICEs detected in the Frankia genomes were functional mobile elements, we carried out nested PCR experiments to detected attP for all of the putative AICEs detected in Frankia strains ACN14a, Ean1pec and Cci3. Nine out of ten tested AICEs were found to excise when the strains were grown in liquid cultures. Excision of Fcci4274 could not be detected in our assay.
Interestingly, Fean5518 and Fean5534, which are inserted in a tandem fashion, were found to excise both independently and associatively within a large ∼31.8 kb circular molecule. Fean6303 was found to be able to excise by recombination between attR and either attL (15.3 kb form) or attL2 (32.3 kb form) (Figure 8A, lanes 6 and 7). Similar observations have been reported for ICESt1 and related ICEs from S. thermophilus  as well as for the High Pathogenicity Island (HPI) in ICEEc1 from Escherichia coli .
Formation of attP sites was detected by nested PCR in circular excised elements. (A) Detection in vitro. 1, Fean1457; 2, Fean5323; 3, Fean5518; 4, Fean5534; 5, Fean5518–5534; 6, Fean6303; 7, Fean6303–6336; 8, Faln1739; 9, Faln5456; 10, Fcci0407; 11, Fcci1033; 12, Faln2929; 13, Fcci3388. (B) Detection in planta. 1, Faln1739; 2, Faln5456. M: marker DNA fragments.
Surprisingly, excision of Faln1739 was detectable despite the absence of a functional recombinase (Figure 8A, lane 8). The int gene of Faln1739 encodes a predicted truncated protein due to the presence of a stop codon at position 291. Another integrase encoded by the genome of ACN14a could mediate the integration and excision of this AICE; yet we failed to detect any closely related int gene in this genome. Furthermore, no mutation was detected in the tRNA genes for Tyr or Gln codons or other tRNA genes, ruling out the involvement of a suppressor tRNA in the translation of this gene into a functional protein. We hypothesize that slippage of the ribosome through the TAA stop codon during translation of Faln1739′s int allows sufficient production of a functional integrase.
Excision of Fcci3390 was observed only for the segment located between attL and attR (Figure 8A, lane 13). Phylogenetic analysis showed that the two integrases encoded by this AICE (see Fcci3388 and Fcci3390 in figure 3B) belong to two different sub-families. Excision of the segment located between attL2 and attR2 which should be mediated by int2 is likely prevented by the lack of a cognate RDF.
Excision of Faln2929, which encodes a serine recombinase but no identifiable RDF, was also detected. This suggests that this recombinase is able to catalyze both integration and excision of this AICE without any requirement for an auxiliary AICE-encoded Xis protein. A similar observation has been reported for Streptomyces phage BT1 .
Finally, excision of Faln1739 and Faln5456 was also detected in Alnus viridis ssp. crispa nodules infected by Frankia alni ACN14a and grown in controlled laboratory conditions (Figure 8B, lanes 1 and 2). These results demonstrate the possibility of genetic exchange in planta involving Frankia and other endophytes occupying this same, narrow, ecological niche.
Putative regulatory functions in Frankia ICEs
Conjugative transfer of most natural ICEs identified to date is tightly regulated and can be induced by ICE-specific environmental stimuli, such as the presence of antibiotics (tetracycline), DNA-damaging agents (UV light, mitomycin C) or quorum sensing . These diverse mechanisms of induction employ a vast array of ICE-encoded regulator proteins. Some of these proteins can interact positively or negatively with other mobile genetic elements . Frankia AICEs also include putative regulators such as orthologs of Pra encoded by pSAM2 as well as putative proteins containing XRE or GntR domains. In pSAM2 from Streptomyces ambofaciens, int, xis and repSA form an operon that is positively regulated by the product of pra . All but three Frankia AICEs carry a pra ortholog.
GntR-family transcriptional regulators encoded by Faln1739, Fcci4274, Fdat4298, and Feul5809 are related (35%–48% identity) to KorSA, a protein encoded by pSAM2. KorSA belongs to a kil-kor system and negatively regulates the replication and transfer of pSAM2 by repressing pra . Interestingly, the gene korSA of Fdat4298 has a low G+C content compared to the neighbouring ORFs suggesting that this gene had been acquired from bacteria with lower G+C content.
Eleven out of the sixteen studied AICEs contain two different types of ORFs coding for XRE-family transcriptional regulators (Figure 7). XRE proteins have an N-terminal helix-turn-helix DNA-binding motif and belong to the xenobiotic response element family of transcriptional regulators. The first type of gene, mostly located near pra encodes small (129 to 180 amino acid residues) XRE proteins. The second type encodes larger proteins (281 to 441 amino acid residues) sharing 27–36% identity with NsdA proteins from various Streptomyces strains, such as SCO5582 from S. coelicolor. SCO5582 is a negative regulator of antibiotic production and sporulation . FeuI6863 encodes a putative WhiB-like transcription factor. WhiB has an essential role in sporulation in S. coelicolor .
The presence of such transcriptional regulators in AICEs suggests that acquisition of specific AICEs by Frankia strains could eventually alter particular pathways or functions in Frankia such as secondary metabolite production, establishment of symbiosis or nitrogen fixation.
Auxiliary functions encoded by AICEs in Frankia
Examination of the general molecular structure of AICEs from the six strains of Frankia revealed the presence of an impressive number of genes coding for hypothetical proteins. A search with batch Blast combined with an HMMsearch against all Pfam-A families from Pfam 25.0 database allowed us to attribute putative functions to a significant number of hypothetical genes (Table S6). Two main groups of genes carried by Frankia AICEs can be distinguished based on the predicted functions: (i) genes coding for proteins involved in nucleic acids and amino acids metabolism and (ii) genes coding for adaptive functions.
In the first group, a cluster of 3 genes is shared by four Frankia AICEs and by Sco5349, pMEA100 and pSE211. This gene cluster contains mdp (metal-dependent phosphohydrolase), nud (Nudix hydrolase) and xre (XRE-family transcriptional regulator). Metal-dependent phosphohydrolases are HD hydrolases usually involved in the nucleic acid metabolism and signal transduction in bacteria and archaea . Nudix hydrolases such as MutT in E. coli have the ability to degrade potentially mutagenic oxidised nucleotides , . The Nudix hydrolase Pif of pSAM2 is involved in conjugal immunity . Fcci1033 encodes a predicted phosphoribosylglycinamide formyltransferase (EC:18.104.22.168). Proteins from this family are involved in the biosynthesis of secondary metabolites and nucleotide metabolism.
The second group of genes code for adaptive functions such as heavy metals resistance, antibiotic biosynthesis and response to different environmental stressors. Fean6303 encodes a carbamoyltransferase belonging to the same family as the Rhizobium protein NodU, which is involved in the synthesis of nodulation factors, and CmcH, a protein from Nocardia lactamdurans, which is involved in the biosynthesis of the antibiotic cephamycin (Coque et al. 1995; Jabbouri et al. 1995).
Faln5456 encodes a 150-amino acid protein with predicted modular polyketide synthase activity. In streptomycetes, polyketide synthases participates to the synthesis of molecules with antibiotic and antifungal properties , . Interestingly, Faln5456 and Faln5323 encode two distantly related putative S-adenosyl-L-methionine-dependent methyltransferases. Such enzymes are known to be important for polyketide activation . Finally, FeuI6863 encodes a putative CutA1-like divalent ion tolerance protein, which may be involved in resistance toward heavy metals.
Together with other mobile elements such as plasmids and prophages, ICEs are important catalysts for bacterial genomes evolution. In addition to their own dissemination, ICEs of different families are also mobilizing chromosomal DNA by an increasing number of ways such as hfr-like transfer , mobilization of related genomic islands ,  and possibly by recognition of clt-like regions by Tra proteins in Streptomyces , . In contrast to T4SS ICEs which were recently determined to be the most abundant conjugative elements in prokaryotes , AICEs have received little attention. The systematic in silico identification of 144 AICEs in the whole Actinobacteria phylum reveals the diversity of possible canonical AICEs modules (integration, replication, transfer) and but mostly their surprising combination plasticity. As an example, Int-Tyr family proteins cluster in several sub-families which are not necessarily associated to the same replication protein families. Both tyrosine and serine integrase family proteins promote the specific integration in a vast array of different sites, mostly tRNA genes. These elements are promising substrates for the development of molecular genetic tools for gene expression as well as gene interruption. Moreover, the observation of many AICEs in the same genome, as well as putative T4SS based ICEs in Actinobacteria, further expands the possibility for large genome modification by these possible molecular tools in diverse species, many of which, such as Frankia spp, are currently lacking effective genetic tools for their study.
A total of sixteen different AICEs were identified in the genomes of six Frankia strains. Nine of these AICEs were shown to excise in vitro from the chromosome by site-specific recombination, the initial prerequisite step for their conjugation. Additionally, excision in planta of two of these AICEs was shown to occur while in symbiotic association in root nodules of Alnus viridis. Frankia AICEs harbor diverse sets of genes that do not seem to be involved in their mobility, but rather encode for putative proteins likely involved in AICEs compatibility and exclusion, nodulation or secondary metabolism. Although the possible advantages conferred by AICE-borne genes have yet to be determined, our findings suggest that AICEs play an adaptive role and establishes new grounds for evolutionary and molecular study of Actinobacteria.
Materials and methods
The predicted proteomes of 275 genomes (Table S7) and 176 (Table S8) plasmids of Actinobacteria available in June 2011 were recovered from the RefSeq database  and analyzed to identify putative ICEs. Integrases, and putative transfer and replication proteins were sought for in these proteomes using profile hidden Markov models (HMMs) with HMMsearch from the HMMER v3.0 software package . The HMM profiles that were used for data mining were recovered from the Pfam 25.0 database  and are summarized in Table 1. Since no HMM profile for the RepSA-like replication proteins was available at the time of this study, it was built using HMMbuild from the HMMER v3.0 software package from a C- and N-terminal trimmed multiple sequence alignment of the RepSA proteins described by te Poele et al  using MUSCLE multiple sequence alignment software  (File S1). Maximum-likehood phylogenies of AICE conserved proteins were generated using the PhyML v3.0 program  with the LG substitution model. Tree topologies were optimized by PhyML using both the NNI and SPR methods and the starting tree was estimated using BioNJ. Branch support of the phylogenies was estimated using non-parametric bootstrap (100 replicates). Phylogenetic analyses were computed from reliable amino acid alignments built by MUSCLE. Removal of poorly aligned regions from amino acid alignments was carried out by trimAl v1.2 software using the automated heuristic approach  prior to phylogenetic analyses. Phylogenetic trees were viewed using iTOL v2  and are available as a shared project on the iTOL website (http://itol.embl.de/shared/VincentBurrus). The search for tRNA genes, located near integrase genes, was performed using tRNAscan-SE webserver .
Large genome segments (30 to 60 kbp) containing at least a site-specific recombination gene and a DNA translocation and replication gene were analysed with Yass  to identify the direct repeats of the attL and attR attachment sites flanking Frankia ICEs. The direct repeat sequence corresponding to the attR site was usually found to be the 3′ end of a tRNA-encoding gene. Predictions of domains and functions of Frankia ICE-encoded proteins were determined using HMMsearch against all Pfam-A families from Pfam 25.0 database. Comparative analyses of AICEs were carried out using Mauve .
Bacterial strain and culture methods
Frankia alni sp. ACN14a was isolated in 1982 from nodules of Alnus viridis subsp. crispa plants in Tadoussac, QC, Canada  (catalog registry number ULQ010201401) . Frankia sp. EAN1pec was isolated in 1978 from nodules of Elaeagnus augustifolia growing in Ohio, U.S.A.  (catalog registry number ULQ13100144). Frankia sp. CcI3 was isolated in 1983 from nodules of Casuarina cunninghamiana (MA, U.S.A.) (catalog registry number HFP020203) . Frankia alni ACN14a and Frankia sp. Cci3 were grown at 30°C (obscurity, static) in BAPS medium: propionate-BAP medium  supplemented with 5 g/L sodium succinate. Frankia sp. EAN1pec was grown at 25°C (obscurity, static) in MP medium supplemented with 20 mM fructose as described by Tisa and Ensign .
DNA extraction from liquid culture and nodules
Total genomic DNA from Frankia strains grown in liquid cultures or present in Alnus viridis subsp. crispa nodules grown in controlled laboratory conditions was purified using REDExtract-N-Amp™ Plant or Seeds PCR Kit (Sigma-Aldrich; St-Louis, MO, USA).
ICE excision assays
ICE excision was determined by nested PCR . Primers used for the detection of circularized forms of ICEs are described in Table S9. REDExtract-N-Amp™ PCR ReadyMix™ (Sigma-Aldrich) was used to perform amplification in 20 µl reactions in the following conditions: 95°C for 5 min, 35 cycles of 95°C for 30 s, 57–67°C for 1 min (depending on the set of primers used) and 72°C for 30 s, with a final elongation step at 72° for 10 min. Then 1 µl of amplification product, obtained following the first round of PCR, was used for a second round of PCR using different sets of primers identified as inner primers in Table S9. PCR conditions were the same as described above. Amplicons from the second round of PCR were sequenced by Genome Quebec Innovation Center of McGill University (Montreal, QC, Canada).
Selected AICEs and plasmids synteny. The synteny maps were generated by Mauve. Colored blocks correspond depict collinear and homologous sequences or regions between AICEs DNA sequences. A similarity profile is represented inside each block. The height of the similarity profile corresponds to the average level of conservation in that region of AICEs sequence. Regions outside blocks lack detectable homology among the studied AICEs.
Synteny of Frankia AICEs. Synteny between Frankia AICEs was determined as indicated in Figure S2. attL and attR sites are indicated by flanking red vertical lines.
Genetic organization of putative Frankia AICE remnants. The genetic organisation of the elements is depicted as in Figure 7. GT2, putative glycosyl transferase family 2; HATP-ase, putative signal transduction histidine kinase; GP49, GP49-like protein; apc5, anaphase-promoting complex subunit 5; s-kin, serine/threonine protein kinase; lt, lytic transglycosylase; PALP, pyridoxal-5′-phosphate-dependent protein subunit beta.
T4SS ICEs identified in this study.
Predicted functions of putative proteins encoded by Frankia remnant AICEs or mobilizable integrating elements.
Predicted functions of putative proteins encoded by Frankia AICEs.
Actinobacterial genomes analysed in this study.
Actinobacterial plasmids analysed in this study.
The authors acknowledge the support of Université Laval for granting access to the Frankia collection of Pr. Maurice Lalonde.
Conceived and designed the experiments: MG EB VB. Performed the experiments: MG EB VB. Analyzed the data: MG EB VB. Contributed reagents/materials/analysis tools: JB RB SR VB. Wrote the paper: MG EB RB SR VB.
- 1. Roy S, Khasa DP, Greer CW (2007) Combining alders, frankiae, and mycorrhizae for the revegetation and remediation of contaminated ecosystems. Can J Bot 85: 237–251.
- 2. Dawson JO (2008) Ecology of actinorhizal plants. In: Pawlowski K, Newton WE, editors. Nitrogen-fixing actinorhizal symbioses. Dordrecht: Springer. pp. 199–234.
- 3. Dixon ROD, Wheeler CT (1986) Nitrogen Fixation in Plants. Glasgow: Biackie.
- 4. Normand P, Lapierre P, Tisa LS, Gogarten JP, Alloisio N, et al. (2007) Genome characteristics of facultatively symbiotic Frankia sp. strains reflect host range and host plant biogeography. Genome Res 17: 7–15.
- 5. Lerat S, Simao-Beaunoir AM, Beaulieu C (2009) Genetic and physiological determinants of Streptomyces scabies pathogenicity. Mol Plant Pathol 10: 579–585.
- 6. Rosas-Magallanes V, Deschavanne P, Quintana-Murci L, Brosch R, Gicquel B, et al. (2006) Horizontal transfer of a virulence operon to the ancestor of Mycobacterium tuberculosis. Mol Biol Evol 23: 1129–1135.
- 7. Kers JA, Cameron KD, Joshi MV, Bukhalid RA, Morello JE, et al. (2005) A large, mobile pathogenicity island confers plant pathogenicity on Streptomyces species. Mol Microbiol 55: 1025–1033.
- 8. Jayapal KP, Lian W, Glod F, Sherman DH, Hu WS (2007) Comparative genomic hybridizations reveal absence of large Streptomyces coelicolor genomic islands in Streptomyces lividans. BMC Genomics 8: 229.
- 9. te Poele EM, Bolhuis H, Dijkhuizen L (2008) Actinomycete integrative and conjugative elements. Antonie Van Leeuwenhoek 94: 127–143.
- 10. Burrus V, Waldor MK (2004) Shaping bacterial genomes with integrative and conjugative elements. Res Microbiol 155: 376–386.
- 11. Wozniak RA, Waldor MK (2010) Integrative and conjugative elements: mosaic mobile genetic elements enabling dynamic lateral gene flow. Nat Rev Microbiol 8: 552–563.
- 12. Burrus V, Pavlovic G, Decaris B, Guedon G (2002) Conjugative transposons: the tip of the iceberg. Mol Microbiol 46: 601–610.
- 13. Christie PJ, Atmakuri K, Krishnamoorthy V, Jakubowski S, Cascales E (2005) Biogenesis, architecture, and function of bacterial type IV secretion systems. Annu Rev Microbiol 59: 451–485.
- 14. de la Cruz F, Frost LS, Meyer RJ, Zechner EL (2010) Conjugative DNA metabolism in Gram-negative bacteria. FEMS Microbiol Rev 34: 18–40.
- 15. Sezonov G, Hagege J, Pernodet JL, Friedmann A, Guerineau M (1995) Characterization of pra, a gene for replication control in pSAM2, the integrating element of Streptomyces ambofaciens. Mol Microbiol 17: 533–544.
- 16. Sezonov G, Duchene AM, Friedmann A, Guerineau M, Pernodet JL (1998) Replicase, excisionase, and integrase genes of the Streptomyces element pSAM2 constitute an operon positively regulated by the pra gene. J Bacteriol 180: 3056–3061.
- 17. Hagege J, Boccard F, Smokvina T, Pernodet JL, Friedmann A, et al. (1994) Identification of a gene encoding the replication initiator protein of the Streptomyces integrating element, pSAM2. Plasmid 31: 166–183.
- 18. Hagege J, Pernodet JL, Friedmann A, Guerineau M (1993) Mode and origin of replication of pSAM2, a conjugative integrating element of Streptomyces ambofaciens. Mol Microbiol 10: 799–812.
- 19. Smokvina T, Boccard F, Pernodet JL, Friedmann A, Guerineau M (1991) Functional analysis of the Streptomyces ambofaciens element pSAM2. Plasmid 25: 40–52.
- 20. Raynal A, Karray F, Tuphile K, Darbon-Rongere E, Pernodet JL (2006) Excisable cassettes: new tools for functional analysis of Streptomyces genomes. Appl Environ Microbiol 72: 4839–4844.
- 21. Pruitt KD, Tatusova T, Klimke W, Maglott DR (2009) NCBI Reference Sequences: current status, policy and new initiatives. Nucleic Acids Res 37: D32–36.
- 22. Biller SJ, Burkholder WF (2009) The Bacillus subtilis SftA (YtpS) and SpoIIIE DNA translocases play distinct roles in growing cells to ensure faithful chromosome partitioning. Mol Microbiol 74: 790–809.
- 23. Kaimer C, Gonzalez-Pastor JE, Graumann PL (2009) SpoIIIE and a novel type of DNA translocase, SftA, couple chromosome segregation with cell division in Bacillus subtilis. Mol Microbiol 74: 810–825.
- 24. Begg KJ, Dewar SJ, Donachie WD (1995) A new Escherichia coli cell division gene, ftsK. J Bacteriol 177: 6211–6222.
- 25. Burton BM, Marquis KA, Sullivan NL, Rapoport TA, Rudner DZ (2007) The ATPase SpoIIIE transports DNA across fused septal membranes during sporulation in Bacillus subtilis. Cell 131: 1301–1312.
- 26. Warth L, Haug I, Altenbuchner J (2011) Characterization of the tyrosine recombinase MrpA encoded by the Streptomyces coelicolor A3(2) plasmid SCP2*. Arch Microbiol 193: 187–200.
- 27. Possoz C, Gagnat J, Sezonov G, Guerineau M, Pernodet JL (2003) Conjugal immunity of Streptomyces strains carrying the integrative element pSAM2 is due to the pif gene (pSAM2 immunity factor). Mol Microbiol 47: 1385–1393.
- 28. Williams KP (2002) Integration sites for genetic elements in prokaryotic tRNA and tmRNA genes: sublocation preference of integrase subfamilies. Nucleic Acids Res 30: 866–875.
- 29. Askora A, Kawasaki T, Fujie M, Yamada T (2011) Resolvase-like serine recombinase mediates integration/excision in the bacteriophage phiRSM. J Biosci Bioeng 111: 109–116.
- 30. Hochhut B, Beaber JW, Woodgate R, Waldor MK (2001) Formation of chromosomal tandem arrays of the SXT element and R391, two conjugative chromosomally integrating elements that share an attachment site. J Bacteriol 183: 1124–1132.
- 31. Garriss G, Waldor MK, Burrus V (2009) Mobile antibiotic resistance encoding elements promote their own diversity. PLoS Genet 5: e1000775.
- 32. Burrus V, Waldor MK (2004) Formation of SXT tandem arrays and SXT-R391 hybrids. J Bacteriol 186: 2636–2645.
- 33. Pembroke JT, Murphy DB (2000) Isolation and analysis of a circular form of the IncJ conjugative transposon-like elements, R391 and R997: implications for IncJ incompatibility. FEMS Microbiol Lett 187: 133–138.
- 34. Murphy DB, Pembroke JT (1999) Monitoring of chromosomal insertions of the IncJ elements R391 and R997 in Escherichia coli K-12. FEMS Microbiol Lett 174: 355–361.
- 35. Fouts DE (2006) Phage_Finder: automated identification and classification of prophage regions in complete bacterial genome sequences. Nucleic Acids Res 34: 5839–5851.
- 36. Panis G, Duverger Y, Courvoisier-Dezord E, Champ S, Talla E, et al. (2010) Tight regulation of the intS gene of the KplE1 prophage: a new paradigm for integrase gene regulation. PLoS Genet 6:
- 37. del Solar G, Giraldo R, Ruiz-Echevarria MJ, Espinosa M, Diaz-Orejas R (1998) Replication and control of circular bacterial plasmids. Microbiol Mol Biol Rev 62: 434–464.
- 38. Ilyina TV, Koonin EV (1992) Conserved sequence motifs in the initiator proteins for rolling circle DNA replication encoded by diverse replicons from eubacteria, eucaryotes and archaebacteria. Nucleic Acids Res 20: 3279–3285.
- 39. Koonin EV, Ilyina TV (1993) Computer-assisted dissection of rolling circle DNA replication. Biosystems 30: 241–268.
- 40. Weigel C, Seitz H (2006) Bacteriophage replication modules. FEMS Microbiol Rev 30: 321–381.
- 41. Orozco BM, Hanley-Bowdoin L (1998) Conserved sequence and structural motifs contribute to the DNA binding and cleavage activities of a geminivirus replication protein. J Biol Chem 273: 24448–24456.
- 42. Orozco BM, Gladfelter HJ, Settlage SB, Eagle PA, Gentry RN, et al. (1998) Multiple cis elements contribute to geminivirus origin function. Virology 242: 346–356.
- 43. Omer CA, Cohen SN (1989) SLP1: a paradigm for plasmids that site-specifically integrate in the actinomycetes. In: Berg DE, Howe MM, editors. Mobile DNA. Washington, DC: American Society for Microbiology. pp. 289–296.
- 44. Vogelmann J, Ammelburg M, Finger C, Guezguez J, Linke D, et al. (2011) Conjugal plasmid transfer in Streptomyces resembles bacterial chromosome segregation by FtsK/SpoIIIE. EMBO J 30: 2246–2254.
- 45. Pettis GS, Cohen SN (1994) Transfer of the plJ101 plasmid in Streptomyces lividans requires a cis-acting function dispensable for chromosomal gene transfer. Mol Microbiol 13: 955–964.
- 46. Berkmen M, Laurer SJ, Giarusso BK, Romero R (2011) The Integrative and Conjugative Element ICEBs1 of Bacillus subtilis. In: Roberts AP, Mullany P, editors. Integrative Mobile Gnetic Elements: Landes Bioscience. in press.
- 47. Abajy MY, Kopec J, Schiwon K, Burzynski M, Doring M, et al. (2007) A type IV-secretion-like system is required for conjugative DNA transport of broad-host-range plasmid pIP501 in gram-positive bacteria. J Bacteriol 189: 2487–2496.
- 48. Gomis-Ruth FX, Sola M, de la Cruz F, Coll M (2004) Coupling factors in macromolecular type-IV secretion machineries. Curr Pharm Des 10: 1551–1565.
- 49. Garcillan-Barcia MP, Francia MV, de la Cruz F (2009) The diversity of conjugative relaxases and its application in plasmid classification. FEMS Microbiol Rev 33: 657–687.
- 50. Lee CA, Babic A, Grossman AD (2010) Autonomous plasmid-like replication of a conjugative transposon. Mol Microbiol 75: 268–279.
- 51. Mohd-Zain Z, Turner SL, Cerdeno-Tarraga AM, Lilley AK, Inzana TJ, et al. (2004) Transferable antibiotic resistance elements in Haemophilus influenzae share a common evolutionary origin with a diverse family of syntenic genomic islands. J Bacteriol 186: 8114–8122.
- 52. Mavaro A, Abts A, Bakkes PJ, Moll GN, Driessen AJ, et al. (2011) Substrate Recognition and Specificity of the NisB Protein, the Lantibiotic Dehydratase Involved in Nisin Biosynthesis. J Biol Chem 286: 30552–30560.
- 53. Terns MP, Terns RM (2011) CRISPR-based adaptive immune systems. Curr Opin Microbiol 14: 321–327.
- 54. Guo P, Cheng Q, Xie P, Fan Y, Jiang W, et al. (2011) Characterization of the multiple CRISPR loci on Streptomyces linear plasmid pSHK1. Acta Biochim Biophys Sin (Shanghai) 43: 630–639.
- 55. Flannagan SE, Clewell DB (1991) Conjugative transfer of Tn916 in Enterococcus faecalis: trans activation of homologous transposons. J Bacteriol 173: 7136–7141.
- 56. Lemire S, Figueroa-Bossi N, Bossi L (2011) Bacteriophage crosstalk: coordination of prophage induction by trans-acting antirepressors. PLoS Genet 7: e1002149.
- 57. Raynal A, Tuphile K, Gerbaud C, Luther T, Guerineau M, et al. (1998) Structure of the chromosomal insertion site for pSAM2: functional analysis in Escherichia coli. Mol Microbiol 28: 333–342.
- 58. Lewis JA, Hatfull GF (2001) Control of directionality in integrase-mediated recombination: examination of recombination directionality factors (RDFs) including Xis and Cox proteins. Nucleic Acids Res 29: 2205–2216.
- 59. Pavlovic G, Burrus V, Gintz B, Decaris B, Guedon G (2004) Evolution of genomic islands by deletion and tandem accretion by site-specific recombination: ICESt1-related elements from Streptococcus thermophilus. Microbiology 150: 759–774.
- 60. Schubert S, Dufke S, Sorsa J, Heesemann J (2004) A novel integrative and conjugative element (ICE) of Escherichia coli: the putative progenitor of the Yersinia high-pathogenicity island. Mol Microbiol 51: 837–848.
- 61. Zhang L, Ou X, Zhao G, Ding X (2008) Highly efficient in vitro site-specific recombination system based on Streptomyces phage phiBT1 integrase. J Bacteriol 190: 6392–6397.
- 62. Daccord A, Ceccarelli D, Burrus V (2010) Integrating conjugative elements of the SXT/R391 family trigger the excision and drive the mobilization of a new class of Vibrio genomic islands. Mol Microbiol 78: 576–588.
- 63. Sezonov G, Possoz C, Friedmann A, Pernodet JL, Guerineau M (2000) KorSA from the Streptomyces integrative element pSAM2 is a central transcriptional repressor: target genes and binding sites. J Bacteriol 182: 1243–1250.
- 64. Li W, Ying X, Guo Y, Yu Z, Zhou X, et al. (2006) Identification of a gene negatively affecting antibiotic production and morphological differentiation in Streptomyces coelicolor A3(2). J Bacteriol 188: 8368–8375.
- 65. Davis NK, Chater KF (1992) The Streptomyces coelicolor whiB gene encodes a small transcription factor-like protein dispensable for growth but essential for sporulation. Mol Gen Genet 232: 351–358.
- 66. Aravind L, Koonin EV (1998) The HD domain defines a new superfamily of metal-dependent phosphohydrolases. Trends Biochem Sci 23: 469–472.
- 67. Bessman MJ, Frick DN, O'Handley SF (1996) The MutT proteins or “Nudix” hydrolases, a family of versatile, widely distributed, “housecleaning” enzymes. J Biol Chem 271: 25059–25062.
- 68. Kamiya H, Ishiguro C, Harashima H (2004) Increased A:T-->C:G mutations in the mutT strain upon 8-hydroxy-dGTP treatment: direct evidence for MutT involvement in the prevention of mutations by oxidized dGTP. J Biochem 136: 359–362.
- 69. Zotchev S, Haugan K, Sekurova O, Sletta H, Ellingsen TE, et al. (2000) Identification of a gene cluster for antibacterial polyketide-derived antibiotic biosynthesis in the nystatin producer Streptomyces noursei ATCC 11455. Microbiology 146(Pt 3): 611–619.
- 70. Hopwood DA (1997) Genetic Contributions to Understanding Polyketide Synthases. Chem Rev 97: 2465–2498.
- 71. Hochhut B, Marrero J, Waldor MK (2000) Mobilization of plasmids and chromosomal DNA mediated by the SXT element, a constin found in Vibrio cholerae O139. J Bacteriol 182: 2043–2047.
- 72. Osorio CR, Marrero J, Wozniak RA, Lemos ML, Burrus V, et al. (2008) Genomic and functional analysis of ICEPdaSpa1, a fish-pathogen-derived SXT-related integrating conjugative element that can mobilize a virulence plasmid. J Bacteriol 190: 3353–3361.
- 73. Guglielmini J, Quintais L, Garcillan-Barcia MP, de la Cruz F, Rocha EP (2011) The Repertoire of ICE in Prokaryotes Underscores the Unity, Diversity, and Ubiquity of Conjugation. PLoS Genet 7: e1002222.
- 74. Eddy SR (2009) A new generation of homology search tools based on probabilistic inference. Genome Inform 23: 205–211.
- 75. Finn RD, Mistry J, Tate J, Coggill P, Heger A, et al. (2010) The Pfam protein families database. Nucleic Acids Res 38: D211–222.
- 76. Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 32: 1792–1797.
- 77. Guindon S, Dufayard JF, Lefort V, Anisimova M, Hordijk W, et al. (2010) New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Syst Biol 59: 307–321.
- 78. Capella-Gutierrez S, Silla-Martinez JM, Gabaldon T (2009) trimAl: a tool for automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics 25: 1972–1973.
- 79. Letunic I, Bork P (2011) Interactive Tree Of Life v2: online annotation and display of phylogenetic trees made easy. Nucleic Acids Res.
- 80. Schattner P, Brooks AN, Lowe TM (2005) The tRNAscan-SE, snoscan and snoGPS web servers for the detection of tRNAs and snoRNAs. Nucleic Acids Res 33: W686–689.
- 81. Noe L, Kucherov G (2005) YASS: enhancing the sensitivity of DNA similarity search. Nucleic Acids Res 33: W540–543.
- 82. Darling AE, Mau B, Perna NT (2010) progressiveMauve: multiple genome alignment with gene gain, loss and rearrangement. PLoS One 5: e11147.
- 83. Normand P, Lalonde M (1982) Evaluation of Frankia strains isolated from provenances of two Alnus species. Can J Microbiol 28: 1133–1142.
- 84. LeChevalier MP (1985) Catalog of Frankia strains, second edition. Actinomycetes 19: 131–162.
- 85. Lalonde M, Calvert HE, Pine S (1981) Isolation and use of Frankia strains in actinorhizae formation. In: Gibson AH, Newton WE, editors. Current perspectives in nitrogen fixation. Canberra: Australian Academy of Science. pp. 296–299.
- 86. Zhang Z, Lopez MF, Torrey JG (1984) A comparison of cultural characteristics and infectivity of Frankia isolates from root nodules of Casuarina species. Plant Soil 78: 79–90.
- 87. Igual JM, Dawson JO (1999) Stimulatory effects of aluminum on in vitro growth of Frankia. Can J Bot 77: 1321–1326.
- 88. Tisa LS, Ensign JC (1987) Comparative physiology of nitrogenase activity and vesicle development for Frankia strains CpI1, ACN1Ag, EAN1pec and EUN1f. Arch Microbiol 147: 383–388.