Isolation, nucleotide sequencing and genomic comparison of a Novel SXT/R391 ICE mobile genetic element isolated from a municipal wastewater environment

Integrative Conjugative Elements (ICE’s) of the SXT/R391 family have largely been detected in clinical or environmental isolates of Gammaproteobacteria, particularly Vibrio and Proteus species. As wastewater treatment plants accumulate a large and diverse number of such species, we examined raw water samples taken from a municipal wastewater treatment plant initially using SXT/R391 family integrase gene-specific PCR probes to detect the presence of such elements in a directed approach. A positive amplification occurred over a full year period and a subsequent Restriction Fragment Length Polymorphism (RFLP) analysis revealed a very limited diversity in the treatment plant examined. Samples demonstrating positive amplification were cultured using Vibrio and Proteus selective media and PCR amplification tracking was utilized to monitor SXT/R391-ICE family containing strains. This screening procedure resulted in the isolation and identification of a Proteus mirabilis strain harbouring an ICE. Whole-genome sequencing of this ICE containing strain using Illumina sequencing technology revealed a novel 81 kb element that contained 75 open reading frames on annotation but contained no antibiotic or metal resistance determinants. Comparative genomics revealed the element contained a conserved ICE core with one of the insertions containing a novel bacteriophage defence mechanism. This directed isolation suggests that ICE elements are present in the environment without apparent selective pressure but may contain adaptive functions allowing survival in particular environments such as municipal wastewater which are reservoirs for large bacterial phage populations.

insert at the 3′ end of the multi-copy tRNA-Ser gene and have been found solely in Vibrio species 7,10 . Another type of SXT/R391 ICE integrates into the pabA (para-aminobenzoate synthase) gene as with ICESh95 11 .
The integrase encoding gene of the tyrosine recombinase class is a disguising feature of Type 1 elements. Integrase genes from several different SXT/R391 ICE's were compared to design specific primers to amplify the gene 12 . Type 1 SXT/R391 ICEs show site-specific integration into the 17 bp integration site at the 5′ end of the prfC gene 9,13 and contain a core of 52 conserved genes that make up the 'backbone' of the element with the key functions being integration, excision and regulation 14 . Type 1 elements contain five hotspot regions and five variable regions which allow for gene or gene segments to be inserted, giving the element an array of potential adaptive functions 14 . All ICES characterised to date have been isolated serendipitously or identified from genome sequences There have been reports of SXT/R391-like ICEs from wastewater environments, ICEVchMex1 (GQ463143.1) isolated from sewage in San Luis Potosi, Mexico 15 and ICESheChnS12 (AXZL01000060.1), isolated from a wastewater treatment plant in China, 2012 15 . However, these were discovered following analysis rather than using a directed approach. ICEVchMex1 contains genes in all its hotspots and in one variable region and encodes ampicillin resistance 15 . ICESheChnS12 was found to contain genes in all its hotspot but no genes were contained in the variable regions. The hotspots contained accessory genes such as AAA ATPase and a Restriction Modification System 16 .
In this study, a directed approach was used to detect and isolate ICEs of the SXT/R391 family from a municipal wastewater treatment plant and to characterise one of these isolates as the first ICE detected and isolated from Ireland.

Materials and Methods
Bacterial strains, elements and media. The bacterial strains and SXT/R391 elements utilised as part of this study are listed in Table 1. Strains were stored at −80 °C in either Luria-Bertani (LB) broth or M9 minimal media containing 50% (v/v) glycerol.

collection and processing of wastewater sample for detection of ice containing strains.
Composite 24 hr wastewater samples were collected from the inlet works at a municipal wastewater treatment plant, Ireland in May and November 2018.
Bacterial screening and molecular identification of ICE containing strains. Serial dilutions of the collected and processed samples isolated from wastewater were carried out using a 0.85% Saline solution. Serial dilutions from10 −2 to 10 −8 were prepared and samples spread on selective media. CLED media was used for the detection of Proteus species and TCBS media for the detection of Vibrio species, plates were incubated at 37 °C for 24 hours. Colonies were routinely isolated and inoculated using LB broth, incubated at 37 °C for 24 hours at 200 rpm. The remaining overnight culture was used to prepare glycerol stocks and stored at −80 °C. Replica plating was carried out to identify single colonies containing ICE-like elements via PCR. The streak plate technique was used to isolate a single pure colony from the replica plate.
pcR. DNA extractions from isolates were performed using the DNeasy Blood & Tissue extraction kit (Qiagen) under the Gram-Negative bacterial protocol for liquid cultures. A PowerSoil DNA extraction kit (Qiagen) was used for solid materials. PCR was carried out using specifically designed primers (Table 2) to amplify the characteristic integrase gene of SXT/R391 family ICE MGEs 12 . www.nature.com/scientificreports www.nature.com/scientificreports/ community analysis. PCR-RFLP (Restriction Fragment Length Polymorphism) was carried to determine whether more than one ICE type might be present in the isolated samples. NEB_Cutter 17 (http://www.labtools. us/nebcutter-v3-0/) was used to select restriction endonucleases that could be used to digest modified integrase genes. The nucleotide sequences of sequenced SXT/R391 ICEs present in the ICEBerg database 8 and the respective integrase gene sequences were selected to determine what enzymes might be utilised to determine RFLP variation. As a result, three restriction endonucleases were selected, HpaII, ApoII and TauI. An ApoII site was not present in ICER391 (AY090559.1) int amplicons while a TauI site was not present in ICER997 (KY433363), ICESXT or ICEpMERPH (MH974755).
PCR, as described above, was carried out on selected isolates containing an amplified integrase gene, ICER391, ICER997 and ICESXT amplicons were used as controls. Additional controls consisted of the use of two different molecular weight markers (NEB 1 kb and NEB 100 kb), uncut PCR product and a control no enzyme "mock" digest. After the PCR, the restriction digest was carried out using 5 µl PCR product, 2 µl Tango Buffer, 0.5 µl of relevant restriction endonuclease and 12.5 µl Milli Q water, bringing the total volume to 20 µl. A cocktail of 15 µl was aliquoted into the PCR tube and 5 µl of the PCR product was then added before the digestion, the tubes were incubated at the appropriate temperatures, HpaII and ApoII at 37 °C and TauI at 65 °C. The tubes were incubated for 45 minutes in the thermocycler. To visualize the samples, 3 µl of 5X loading dye was added directly into the digest, 10 µl was loaded onto a 1.5% (w/v) agarose gel in 1X TAE Buffer. The gel was set to run for 1 hour at 100 V and photographed under UV light.
ice transfer via triparental mating. To determine if the identified ICE could undergo conjugation tri-parental mating was carried out using the isolated strain as a host strain, pBAM1 as a helper (pBAM1 reference) and E.coli J53 as a recipient. The idea was to mark the ICE with a mini Tn5 (Kanamycin resistance) via pBAM1 and then mobilise the marker ICE into E.coli J53, a Rif resistant strain. Conjugative mating was carried out following the method by Kristensen et al. with some modifications 18 . The donor, helper and recipient strains were grown overnight. Overnight cultures were then centrifuged for 5-10 minutes at 10,000 rpm to collect cells, and washed with fresh LB broth. The cultures were mixed in a proportion of 5:1:1 (donor population was increased) and 50 µl were applied onto 0.45 µm filter on LB Agar plates. The plates were incubated at 37 °C for 24 hours. The cells were then scraped off the surface of the filter and added to a 0.85% saline solution. The culture was diluted with antibiotics for counter selection of the donor and helper cells.
The dilutions were spread onto rifampicin (200 µg.ml −1 ) and kanamycin (50 µg.ml −1 ) supplemented plates and grown at 37 °C for 24 hours. Rifampicin had an MIC of 200 µg.ml −1 for the host strain as it was susceptible at this concentration while E. coli remained resistant. Controls for the donor, helper and recipient were also prepared. Transconjugants were then tested for the presence of the ICE as described above.
Genome sequencing & genome annotation. Full genome sequencing was carried out by MicrobesNG (University of Birmingham, UK) using 2 ×250 bp paired-end reads HiSeq Illumina Miseq technology giving 30X genome coverage. The genome sequences were analysed using the MicrobesNG's automated analysis pipeline, the closest available reference was found using Kraken software (Taxonomic Sequence Classification System) 19 and reads were mapped using BWA (Burrows-Wheeler Aligner) mem 20 software to assess the quality of the data. De novo assembly of the reads was carried out using SPAdes genome assembler software 21 and the reads were mapped back to the resulting contigs using BWA mem software. The genome was identified amongst the contigs by using the BLAST tool to investigate the presence of several different ICER391 (AY090559) and ICESXT (AY055428) core scaffold genes (int, jef, traLEKBVA, setCD). The sequence data was edited on notepad + + (https:// notepad-plus-plus.org/downloads/), submitted to RAST (Rapid Annotation using Subsystem Technology) for annotation (see below), exported as a Genbank and Fasta file and reannotated manually with predicted Orfs blasted via BLASTP to confirm ICE association. Molecular maps were created by inputting the ICE sequence files in GenBank (.gnk) or FASTA (.faa) format into RAST 22 (http://rast.nmpdr.org/). Gene Graphics was used to create gene representations of hotspot and variable regions by inputting the ICE sequence in GenBank (.gnk) format 23 . phylogenetic analysis. Phylogenetic analysis of fully sequenced SXT/R391 ICEs was performed based on comparison with the concatenated amino acid sequences of 48 to 52 SXT/R391 core ICE gene-encoded proteins on all previously sequenced whole SXT/R391 elements. An unrooted phylogenetic tree was constructed by the maximum-likelihood method based on the Poisson correction model using the MEGAX 24 . Bootstrap analysis with 1000 replications was performed to test the reliability of the tree.

Results & Discussion
Targeted isolation of SXT/R391 ICEs. Composite Table 2. List of primers used in this study with the gene amplified and the nucleotide sequence 12 .
Samples were plated on genera specific selective media as a significant number of SXT/R391 ICE MGEs have been detected in Proteus and Vibrio species 7 . After plates were incubated overnight, colonies of interest were swabbed and inoculated in LB Broth, the overnight cultures were then used for DNA extraction and PCR was performed to detect the presence or absence of an ICE specific integrase gene 12 indicating the presence of an SXT/ R391 ICE MGE containing host. Screening and re-plating (once PCR positive amplicons were detected) were carried out over several weeks.
Following multiple screening, and replica plating (see Fig. 1 in Supplementary Material) followed by repeated PCR analysis, the integrase gene was detected by PCR in three translucent blue extracts, ULP001, ULP004 and ULP014 identified on CLED selective media. Translucent blue colonies were screened for the integrase gene and, if detected, samples were streaked out to isolate a single pure colony. A potential SXT/R391-like ICE was detected following further screening from one Proteus species, selectively grown on CLED media. A PCR amplicon was detected at 1378 bp matching the positive control, AB1157ICER391, see Fig. 1. From the screened single colony, full sequence analysis of this isolate determined that the host was a Proteus species based on analysis of its 16S rDNA sequence. The novel ICE was named ICEPmiIre01 as per standard ICE nomenclature 25 . Proteus isolates isolated in the winter also showed positive amplification for the integrase gene, indicating that the presence of SXT/R391 ICE-like elements is apparently not season dependent. community analysis. RFLP was used to determine whether there might be more than one SXT/R391 ICE type present in the wastewater. The integrase specific PCR amplicon generated by IntFor1 and IntRev1 is normally 1378 bp in length. Alignments of numerous ICE integrase genes sequenced from SXT/R391 family members within the ICEberg database 8 revealed some heterogeneity in gene sequence. Analysis of this heterogeneity using NEBcutter restriction site analysis suggested that using the restriction enzymes HpaII, TauI and ApoII could reveal at least some of this RFLP heterogeneity. Therefore, a number of samples were analysed by amplifying putative int gene amplicons from wastewater isolates and then subjecting the amplicons to restriction analysis via HpaII, TauI and ApoII. We demonstrated that HpaII and TauI were able to cut amplicons generated from samples isolated from wastewater (but not ApoII). HpaII cleaved two types of environmental SXT/R391 ICE integrase genes amplified, one without an HpaII site and one with an HpaII site (each amplicon 1378 bp in length), TauI cleaved one SXT/R391 ICE (two amplicons equal to 1378 bp in length) as it contained a TauI site and ApoII did not show any cleavage of amplicons from isolates in our case.
This limited RFLP yield suggested that only a limited number of amplicon types were present, and indeed possibly only two different types (see Fig. 2, lane 1, where isolate ULP014 when amplified shows two ICE types each at 1378 bp in size, one that contains the HpaII site and one that does not. full sequence analysis. ICEPmiIre01 (was identified among 271 contigs with a mean fold coverage of 14.98 by searching for the integrase gene and blasting against the prototype ICER391 (AY090559.1) The ICE was found on node 1 (287508 bp in length), this node revealed a novel SXT/R391-like ICE element associated with Proteus mirabilis. 16 s rDNA analysis of the host-based on full genome analysis revealed a 100% match to Proteus mirabilis. Previously SXT/R391 ICE elements have been detected in Proteus mirabilis from China but isolated from clinical and animal samples generally associated with illness 26,27 . Proteus species are known to be human and animal opportunistic pathogens, and frequently isolated from urine, wounds and other clinical sites and is present in soil and water habitats possibly from spreading from clinical sources 28 .
This novel SXT/R391 ICE is of the Type 1 ICE variety based on comparative genomics 6 and is 81 kb in length with prediction of 75 orfs based on RAST annotation. A full genome map can be seen in Fig. 3 Comparative genomics revealed that ICEPmiIre01 contains 51 core genes, sharing the core molecular backbone and synteny of other SXT/R391 ICEs. A full list of orfs detected is presented in Supplementary Table 1 with % similarity to ICER391 and ICESXT illustrated. The core genes of ICEPmiIre01 are predicted to encode activities for basic ICE functionality such as integration, excision and conjugative transfer 29 . ICEPmiIre01 is one of the smallest SXT/ R391-like ICE MGEs to be discovered thus far, the smallest being ICEPmiChn3 (KY437727) at 57 kb in length also from a Proteus mirabilis, isolated from broiler carcasses in China 26 .
Structural analysis of ICEPmiIre01 revealed that it contains five hotspot regions and further insertions into one variable region. Some of the hotspot regions contain genes putatively encoding resistance to bacteriophage attack. This is not surprising as there is a continuous arms race between bacteria and phage and such resistance www.nature.com/scientificreports www.nature.com/scientificreports/ systems include restriction-modification systems and toxin-antitoxin systems 30 . Interestingly the element contained no antibiotic resistance or metal resistance determinants. HS1 contains two predicted genes, orf27 encoding a putative plasmid-related protein and orf28 encoding a hypothetical protein with 93% and 97% identity to hypothetical proteins found in ICER391 (AY090559.1). The nucleotide sequence of HS1is 99% similar to that of HS1 of ICER997 31 but it is unclear what the function of these HS1 predicted genes are.
HS2 contains two genes both putatively associated with the Type IV toxin-antitoxin 'innate immunity' bacteria abortive infection (Abi) system that protects bacteria from the spread of phage infection. The Abi system, sometimes denoted as phage exclusion, is characterized by a normal start of bacteriophage infection which is then followed by the interruption of phage development leading to the release of few or no progeny particles but can cause premature host cell death. This Abi system has a survival potential for the population rather than the individual by reducing phage bursts at the expense of individuals within the population 32 . orf35 encodes a putative AbiEi antitoxin, Type IV TA system with a 96% identity to an Ynd protein (AQT24150.1) found in ICER997 (KY433363.1). orf36 encodes a putative nucleotidyl transferase AbiEii/AbiGii toxin family protein with a 97% identity to an Ync protein (AQT24151.1) found in ICER997 (KY433363.1). Toxin-antitoxin genes are commonly found in HS2 of SXT/R391-like ICEs as revealed through comparative analysis. The nucleotide sequence of HS2 is 100% identical to that of Vibrio cholerae, ICESXT (AY055428).
HS3 contains two genes, none of which show any similarity to ICER391, ICER997 or ICESXT, orf62 encodes a putative restriction endonuclease, which is 100% identical to a restriction endonuclease (WP_004249392.1) found in Enterobacterial species while orf63 encodes a putative 5-methylcytosine restriction system component which is 100% identical to the mcrBC-5 methylcytosine restriction system of Gammaproteobacteria 33 . The nucleotide sequence of HS3 is 100% identical to that found in ICEPmiJpn1 (KT894734). These two putative enzymes in HS2 maybe a RMS, which recognizes a specific DNA sequence for cleavage and a cognate methyltransferase that protects from cleavage by methylation of adenine or cytosine bases within the same recognition sequence 34 and maybe involved in protection from invading phages 35 .
HS4 contains three predicted genes, orf44 encodes a hypothetical protein with 100% identity to a hypothetical protein (AQT24159.1) found in ICER997 (KY433363.1), orf45 encodes a putative Helicase HerA-like protein with 100% identity to a Bipolar DNA helicase HerA protein (AQT24160.1) present in ICER997 (KY433363.1). HerA is a DNA helicase able to utilize either 3′ or 5′ single-stranded DNA extensions for loading and subsequent DNA duplex unwinding 36 . Lastly, orf46 encodes a putative endonuclease protein with a 94% identity to an endonuclease  I precursor protein (AQT24161.1) present in ICER997 (KY433363.1). Endonucleases are thought to generate double-strand breaks in DNA. The biological role associated with this hotspot is unknown 37 .
HS5 contains eight predicted genes, six of which make up the novel putative BREX (bacteriophage exclusion) system found in ICEPmiIre01. This system has characterised homologs that allow for phage adsorption but blocks phage DNA replication. A similar BREX system was first identified in the genome of Bacillus subtilis H3081.97 (NZ_ABDL01000007.1) and there confers resistance to a broad range of phage both virulent and temperate 30 . The BREX six gene cassette in Bacillus subtilis includes a putative Lon-like protease, a putative alkaline phosphatase protein, a putative RNA-binding protein, a putative DNA methylase, a putative ATPase domain-containing protein and a protein of unknown function 30 . The six genes that make up the BREX system in ICEPmiIre01 are orf15, a putative inner membrane protein (BrxA) that is 100% identical to orf23, a hypothetical protein (AAM08040.1) found in ICER391. BrxA shares structural homology with the RNA-binding antitermination protein NusB 31 . The second gene orf16 is a putative DUF1788 domain-containing protein (BrxB) which is 100% identical to orf24, a hypothetical protein (AAM08046.1) contained in ICER391, the function of this gene is unknown 30 . The third gene, orf17 a putative BREX system P-loop protein (BrxC), is 100% identical to orf25 a putative ATPase (AAM07998.1) present in ICER391. BrxC contains a putative large ATP binding motif 30 . The fourth gene orf19 encodes a putative BREX-1 system adenine-specific DNA-methyltransferase (PglX), showing 100% identity to orf26, a hypothetical protein (AAM08017.1) found in ICER391. This gene homolog encodes a protein that has putative DNA modifying functions that enable BREX to differentiate between host and phage DNA 30 . The fifth gene is orf20 encoding a putative BREX-1 system phosphatase protein (PglZ, type A), showing 93% identity to orf30, a hypothetical annotated protein (AAM07999.1) present in ICER391. The final gene within the system orf21 encodes a putative a Lon-related BREX system protease (BrxL), showing 97% identity to orf31, a putative ATP-dependent Lon protease protein (AAM08002.1) found in ICER391. Lon proteases are ATP-dependent serine proteases that mediate the selective degradation of mutant and abnormal proteins, as well as certain short-lived regulatory proteins and is required for cellular homeostasis and for survival from DNA damage and development changes induced by stress 38 . Other genes present in the HS5 include orf18, encoding a hypothetical protein 100% identical to a hypothetical protein (EKA96589.1) found in Proteus mirabilis WGLW6 and orf22 a putative DNA repair protein showing 100% identity to orf32, a hypothetical protein (AAM08000.1) present in ICER391. A similar BREX system appears to be present in ICER391 (AY090559) 39 and ICEVpaCan1 (CP028481) 7 based on homology reported above but to our knowledge has not been annotated as a putative BREX system previously. Based on the observations above and based on comparative analysis with other SXT/R391 family members it is clear that ICEPmiIre01 is a mosaic of core ICE genes and variable genes located in hotspots and variable regions of the element. These genes are most likely accumulated by ICE encoded recombination systems which generated a mosaic-like genetic structure 39 . It is interesting to observe elements of this mosaic structure being found in disparate ICE-like elements isolated from different global locations as outlined above.
ICEPmiIre01 has an insertion in Variable Region 1 (VR1). Two predicted genes are present orf01, encoding a putative transcriptional regulator of the XRE family, which shows 100% identity to a hypothetical protein (AAM08068.1) present in ICER391 (AY090559.1) and orf02, encoding a putative Serine/threonine-protein kinase HipA, which is 100% identical to a hypothetical protein (AAM08010.1) also present in ICER391 (AY090559.1). The putative transcriptional regulator of the XRE family, orf01 is also present in ICEBs1 and is proposed to repress the transcription of genes required for excision and transfer and prevents the acquisition of additional ICE copies, therefore, conferring immunity 40 . HipA-like proteins are expressed by various bacterial species and are involved in high-frequency persistence to the lethal effects of inhibition of either DNA or peptidoglycan. When expressed in high amounts alone, HipA is toxic to bacterial cells 41 and may also be involved in multidrug tolerance 42 . These genes are commonly associated with VR1 of several other SXT/R391-family ICEs (see comparative genomics section).
ICEPmiIre01 does not have genes present in variable regions VRII, VRIII, VRIV and VRV as occurs in other SXT\R391 elements and this may be reflect its small size relative to other SXT/R391 ICEs. Accession number. The designated GenBank accession number for the nucleotide sequence ICEPmiIre01 is MN520463. transfer of icePmiIre01. Due to not detecting a selectable marker associated with the ICE it was not possible to demonstrate conjugation of ICEPmiIre01 in the normal way. Therefore a tri-parnetal mating was undertaken using the ICEPmiIre01 containing P. mirabilis host (susceptible to rifampicin), E.coli CC118λ pBAM1, containing a miniTn5 with kanR and E.coli J53 as described above. We isolated putative strains of E. coli J53 containing ICEPmiIre01 containing the transposed miniTn5. Transconjugants were then analysed to validate and confirm the conjugative transfer. Transfer frequency of the element can be seen in Table 3. phylogenetic analysis. A phylogenetic tree (Fig. 4) was constructed based on the concatenated amino acid sequences of all SXT-R391 core proteins for all published core genome sequences of these elements. ICEPmiIre01 is clustered with ICEPmiJpn1, ICER391, ICEIdBUSA1 and ICEPrSt33672. These results illustrate the wide geographic spread of Type 1 SXT/R391 ICE-like elements with these near neighbours being isolated in Japan, South Africa and the USA respectively. Information on SXT/R391 element in Fig. 4    www.nature.com/scientificreports www.nature.com/scientificreports/ BREX systems 30 . The putative BREX system located in HS5 of ICEPmiIre01 is related to a novel bacteriophage defence system first reported by Goldfarb et al. (2015) 30 . It consists of a six gene cassette which includes a putative Lon-like protease, an alkaline phosphatase domain protein, a putative RNA-binding protein, a DNA methylase, an ATPase-domain protein and a protein of unknown function. This system shows high levels of homology and synteny with the putative ICEPmiIre01 system.
By comparing each gene in the putative ICEPmiIre01 BREX system it was discovered that a near-identical unannotated system was also located in HS5 of ICER391 the prototype SXT/R391 ICE (showing between 93-100% homology) 39 . A recently discovered SXT/R391 ICE, ICEVpaCan1, was also found to contain this putative system 7 with homology levels between 68-99%. A comparison between the putative HS5 BREX system of ICEPmiIre01 demonstrated high (84-100%) homology to the BREX system found in Bacillus cereus H3081.9. This is most surprising as SXT/R391 ICEs are exclusively found in Gram-negative bacteria and finding a homologous system in a Gram-positive organism with such high levels of homology is unusual (Fig. 5).
The insertion in HS2 was also compared to ICER391 as it contains a putative T-A system and is commonly associated with this hotspot in other ICEs. T-A systems consist of a stable toxin and its labile antitoxin 43 that promote the maintenance of SXT/R391 ICEs, genes associated with this function are mosT and mosA (ynd, ync) 44 . This system is found in numerous SXT/R391 ICE elements and is generally also found inserted in HS2 15 . Examples of this system are displayed in Fig. 6 and compared with ICEPmiIre01 found in this study. conclusion Using a directed approach, a novel environmental SXT/R391 Type 1 family ICE, ICEPmiIre01 was isolated from a local municipal wastewater by using selective media for Proteus and Vibrio strains (common hosts for SXT/R391 ICEs) and monitoring the presence of ICEs using ICE specific int primers amplified by PCR. A Proteus mirabilis ICEPmiIre01 containing strain isolated in this way was subjected to genome sequencing and revealed an 81 kb element with 75 putative orfs. The element contained the 51 core ICE genes and contained insertions in some of its hotspots which indicated the presence of an encoded bacteriophage resistant mechanisms. A novel BREX system was identified in HS5, not previously identified or annotated in other SXT/R391 ICEs. Comparative analysis revealed that this system is also present in ICER391 but was not previously annotated there 40 and in ICEVpaCan1 again not previously annotated 7 . Unusually the nearest BREX homolog to ICEPmiIre01was reported in a Bacillus subtilis isolate which given the differences between Gram-positive and Gram-negative hosts is surprising. ICEPmiIre01 is a novel SXT/R391 ICE-like element with interestingly no antibiotic or metal resistance, the first to be identified in Ireland.

Figure 5.
Comparative analysis of the BREX systems present in ICEPmiIre01, ICER391 and ICEVpaCan1, located in HS5 of these ICE elements. Genes coloured moss are associated with brxA (predicted RNA-binding domain), genes coloured brown associated with brxB (unknown function), genes coloured light blue are associated with pglX (DNA methylation domain), genes coloured dark green are associated with pglZ (alkaline phosphatase domain) and brxL (lon-like protease domain), genes coloured yellow are associated with a putative DNA repair and genes coloured white of unknown function. Figure 6. Examples of the genetic organization of HS2 of SXT/R391 ICE elements containing a T-A system, including ICEPmiIre01, ICER997 31 , ICESXT 45 48 . Genes coloured orange are associated with ynd and genes coloured blue are associated with ync. Both genes encode a toxin-antitoxin system.