Analysis of pCl107 a large plasmid carried by an ST25 Acinetobacter baumannii strain reveals a complex evolutionary history and links to multiple antibiotic resistance and metabolic pathways

Abstract Acinetobacter baumannii has successfully spread during the last decades as one of the main critically important pathogens. However, many aspects including plasmids, are still under-investigated. Here, we report the complete sequence of an Acinetobacter baumannii strain, belonging to the ST25IP (Institut Pasteur) sequence type recovered in 2012 in Lebanon, using a combination of Illumina MiSeq and Oxford Nanopore sequencing and a hybrid assembly approach. This strain (Cl107) carries a 198 kb plasmid called pCl107 that encodes the MPFI conjugative transfer system. The plasmid carries the aacA1, aacC2, sul2, strAB, and tetA(B) antibiotic resistance genes. pCl107 region encompassing the sul2, strAB, tetA(B) is closely related to AbGRI1 chromosomal resistance islands, which are widespread in A. baumannii strains belonging to Global Clone 2. The resistance region found in pCl107 is one of the missing links in the evolutionary history of the AbGRI1 islands. pCl107 also contains a BREX Type 1 region and represents one of the two main evolution patterns observed in BREX clusters found in plasmids related to pCl107. pCl107 also harbours a ptx phosphonate metabolism module, which plays an ancestral structure compared to other large plasmids in ST25 strains. While the uric acid metabolic module found in pCl107 is incomplete, we identified possible ancestors from plasmids and chromosomes of Acinetobacter spp. Our analyses indicate a complex evolutionary history of plasmids related to pCl107 with many links to multiple antibiotic resistance and metabolic pathways.


Introduction
Acinetobacter baumannii is an important nosocomial pathogen that continues to attract worldwide attention due to its resistance to many antibiotics and its genomic plasticity, enabling it to acquire genetic material (Imperi et al. 2011, Peleg et al. 2012. Globally, most A. baumannii infections are due to the spread of a few clones known as global or international clones (GC) (Zarrilli et al. 2013, Hamidian and Nigro 2019, Koong et al. 2021. Besides the wellstudied ST1 (GC1) and ST2 (GC2), members of ST25 IP (IP: Institut Pasteur MLST scheme) have been reported as an epidemic, endemic and sporadic in many regions around the world (Sahl et al. 2015, Sennati et al. 2016, Nigro and Hall 2017, da Silva et al. 2018, Hamidian and Nigro 2019. ST25 was first described as a singleton (a sequence type without known singlelocus variants (SLV)) (Karah et al. 2012), but few SLV, e.g. ST402 and ST991, have then been reported hinting for further diversification of this well-established sequence type (Cerezales et al. 2020). In addition to their epidemiological importance, ST25 strains have gained special attention mainly due to increasing antimicrobial resistance, particularly to carbapenems (Zarrilli et al. 2013, Sahl et al. 2015, Cerezales et al. 2020, as well as their ability to produce extensive biofilm compared to other globally distributed clones (Sahl et al. 2015).
Plasmids can significantly impact the biology of A. baumannii by introducing additional genes encoding antibiotic resistance and metabolic functions (Fondi et al. 2010). Two plasmids (pD4 and pD46-4; GenBank acc. no. KT779035 and MF399199, respectively) characterized in two Australian ST25 strains are amongst those that carry a large number of genes involved in various functions Hall 2016, Nigro and. pD4 and pD46-4 belong to a broad and diverse family of conjugative plasmids that encode the MPF I conjugative transfer system and several important functions involved in many cellular pathways, e.g. Type 6 Secretion System (T6SS) (Weber et al.2015, Nigro and Hall 2017.
Understanding the biology and evolution of A. baumannii is vital in developing therapeutic agents and preventing infections. However, knowledge of A. baumannii strains is restricted due to the biased study of its different biological aspects. Genomic studies often look for antimicrobial resistance mechanisms and ignore other potentially novel gene clusters that remain buried within the genomic data, including metabolic genes. Over the last decade, genome sequencing has superseded traditional methods to track and characterize microbial pathogens. However, while A. baumannii genomes are being sequenced swiftly in many countries; the genomic data are still scarce from many regions, such as the Middle East Nigro 2019, Koong et al. 2021). Wars have served, not only at the Middle East but at the worldwide level, as a breeding ground (Rafei et al. 2014(Rafei et al. , 2015 for many pathogens as A. baumannii that take advantage of the weak health infrastructure and shattered infection control measures (Rafei et al. 2014, Ismail et al. 2018, Higgins et al. 2020. A. baumannii strains recovered in the Middle East region are important given that this region has been suggested as a potential source for the spread of some of the globally distributed GC1 lineages (Griffith et al. 2007, Calhoun, Murray and Manring 2008, Koong et al. 2021. In Lebanon, ST25 was one of the commonly identified sequence types in clinical settings, although at a lesser degree than the dominant sequence types (e.g. ST2 strains) (Rafei et al. 2015, Al Atrouni et al. 2016, Chamieh et al. 2019. Currently, no complete Middle Eastern ST25 genomes are available in GenBank (Table S1). Hence, we completed and analyzed the genome sequences of one ST25 isolate from Lebanon, strain Cl107. We showed that it contains a large 198 kb plasmid with several important modules, including two antibiotic resistance clusters, a ptx operon (involved in phosphate metabolism), a BREX (phage resistance) system, and a uric acid metabolic operon. These regions were analyzed in detail. We also explored publicly available complete plasmids to study related plasmids.

Bacterial strain and publicly available genome sequence data used in this study
A. baumannii CMUL Cl107 was recovered in 2012 from the urine of an 80-year-old male patient suffering from amyloidosis and anaemia in Tripoli, North Lebanon. Cl107 was previously reported as part of a study investigating the epidemiology of A. baumannii strains isolated from different hospitals in Lebanon (Rafei et al. 2015).

Whole-genome sequencing, assembly, and annotation
The DNA extraction was done with QIAGEN DNeasy genomic DNA extraction Kit following the manufacturer's instructions. Shortread library preparation was done using the Nextera (DNA Flex library preparation) library preparation kit and sequenced using Illumina MiSeq. MinION (Oxford Nanopore) long-read sequencing was done following library preparation using the Rapid Barcoding Sequencing Kit (SQK-RBK004) and sequenced on an FLO-MIN106D Flow cell according to the manufacturer's instructions (Oxford Nanopore Technologies, Inc., Oxford, UK). Illumina MiSeq generated 1102451 paired-end short reads with 50-fold coverage, an average length of 250 bp, and MinION produced a total of 52680 long reads with an N 50 length of 15.2 kbp and 25-fold coverage. FastQC (v.0.11.9) (https://bioinformatics.babraham.ac.uk/pr ojects/fastqc/) and Filtlong (v.0.2.0) (https://github.com/rrwick/Fi ltlong) were used to check the quality of Illumina and MinION reads, respectively. Filtlong was used to filter long reads with poor quality and those with <1 Kbp length. Reads were trimmed using the Porechop (v.0.2.3) programs using default parameters. The high-quality Illumina and MinION reads were assembled de novo using a hybrid assembly approach using the Unicycler program (v0.4.7) as previously described (Wick et al. 2017).

Phylogenetic analysis of the BREX operon
The entire DNA sequences of the brxA-brxL regions were extracted from genomes that showed homology with that of pCl107. Partial regions with incomplete or frameshifted BREX genes were dismissed from the phylogenetic analysis. In total, 90 entries were retained and aligned using the mafft (v7.457) tool (Katoh and Standley 2013). The length of the BREX operon varied in 90 sequences ranging from 13168 bp to 18839 bp with a mean of 14632 bp and a median of 14562 bp. FastTree (v2.1.11) (Price, Dehal, and Arkin 2009) was used to infer a maximum likelihood phylogenetic tree with GTR+CAT as models. FigTree software (v1.4.3) was used to visualize the final phylogenetic tree (http://tree.bio.ed.ac.uk/soft ware/figtree/).

Complete genome sequence of Cl107
The genome sequence of Cl107 was completed using a combination of Illumina MiSeq and Oxford Nanopore (MinION) data following a hybrid assembly approach. The genome assembly consists of a 4056235 bp chromosome and a 198716 bp plasmid named pCl107.
The Cl107 genome belongs to ST229 OX (Oxford scheme) and ST25 IP (Institut Pasteur scheme). It carries the KL 14 and OCL 6 capsular polysaccharide and lipooligosaccharide outer core, respectively. The genome contains a CRISPR locus belonging to CAS-TypeIFb (Cameranesi, Kurth and Repizo 2022) composed of 6 genes in the CAS cluster (bases 1008607-1017088 of GenBank acc. no. CP098521) and 48 spacers (bases 1017221-1020131 of GenBank acc. no. CP098521). Analysis of the spacers compared to the pool of spacers previously identified and catalogued by Karah et al. in 2015(Karah et al. 2015 revealed that Cl107 contains the CST24 CRISPR locus, which is very similar to CST23, except it lacks any duplicated spacer. According to Karah et al. (Karah et al. 2015), CST23-24 constitutes the main subclone of clonal complex 25 with a remarkable ability to be endemic, which has caused national outbreaks in the USA, UAE, Argentina, and Sweden (Karah et al. 2015). Cl107 has the same standard ST25 variants of the intrinsic bla OXA-Ab (encoding bla OXA-64 ) and bla AmpC (99% DNA identity to bla AmpC23 ), which are not preceded by an ISAba1 copy. Cl107 was originally reported to carry bla NDM-1 ; however, this gene appears to have been lost in its complete genome, likely during the storage or transport to Australia, explaining thus its susceptibility to imipenem (Rafei et al. 2015).
Quinolone resistance accounted for the mutations we found in the gyrA DNA gyrase (GyrA S81L) and parC topoisomerase IV genes (ParC S84L). These mutations are known to be associated with fluoroquinolones resistance in Gram-negative bacteria, including A. baumannii (Vila et al. 1995, 1997, Pérez-Varela et al. 2018. Several biofilm-related virulence determinants were identified in Cl107, including two pga loci (pgaABCD and pgaBCD), a csuA/BABCDE locus (chaperone-usher pili assembly system) regulated by bfmRS, an abaI gene encoding an autoinducer synthase part of the quorum-sensing system, ompA, a pglL, a blp2, a 5709 bp sized gene encoding BapA prefix-like domain-containing protein, as well as bap gene. The bap gene of 9315 bp encodes a BAP protein sharing 100% identity with the COOH extremity of type 3 (1169 aa), which is mainly observed in ST25 and differed from type 1 and 2 occurring in GC 1 and 2, respectively, according to De Gregorio et al. (De Gregorio et al. 2015).
Cl107 also carries a large plasmid (198716 bp) called pCl107 with a GC% content of 39.13% and 210 coding reading frames. pCl107 encodes several important modules, including the HigB/A toxin/anti-toxin system, the ParA/B and ParM/StbA partitioning protein systems, a DNA primase, a DNA helicase, a DNA topoisomerase, the UmuCD DNA damage repair system, and the BREX-1 phage resistance system. Similar to previously characterized plasmids, such as pD4 and pD46-4 (GenBank acc. no. KT779035 and MF399199, respectively), pCl107 contains several genes associated with the conjugative transfer, including traIJYWH, trbA, and trbN with 100% DNA identity making pCl107 a potentially conjugative plasmid. Sequence analysis showed that pCl107 is related to pA297-3, pD4, pD46-4, and pAB3 (GenBank acc. no. KU744946, KT779035, MF399199, and CP012005, respectively) (Hamidian and Hall 2016, Hamidian, Ambrose and Hall 2016, 2017, Nigro and Hall 2017). The complete sequences of 34 plasmids related to pA297-3, pD4, pD46-4, and pAB3 are available in GenBank nonredundant database (Table 1). Our analysis showed that pCl107 is closely related to pMC1.1 (GenBank acc. no. MK531536), which is a 184 kb plasmid recovered from an ST991 (an SLV of ST25 IP ) isolate from a catheter sample in Bolivia in 2015 (Cerezales et al. 2020). pCl107 carries four genes responsible for aminoglycoside resistance (strAB, aacA1, aacC2), sul2 for sulfamethoxazole resistance, and tetA(B) gene for tetracycline resistance. pCl107 also carries multiple complete insertion sequences, including three copies of each of ISAba1, ISAba34 and IS1008, two copies of each of IS1007 and ISAba12 and a single copy of each of ISAcsp1, ISAba11, ISAha2, IS17, and the common region CR2, along with many truncated insertion sequences, e.g. IS1006, ISAba22 and the common region elements CR1, and CR2.

pCl107 contains a resistance region that serves as a missing link in the evolution of AbGRI1 antibiotic resistance islands
The strAB genes, together with sul2 and tetAR(B), were found in pCl107 within a variant of Tn6172, which was previously described in pA297-3 (GenBank acc. no. KU744946) with an additional CR -tetAR(B) segment (compared to Tn6172) ( Fig. 1). We previously presented a detailed analysis of the resistance regions found in pA297-3 and pAB3 (GenBank acc. no. CP012005). We proposed that the ancestral structure of AbGRI1 island (AbGRI1-0), which is frequently found in the chromosomal comM genes of GC2 isolates, has evolved from a close relative of pAB3 with Tn6022 in place of Tn6021 and in which Tn6174 had already evolved to form Tn6172 (Hamidian and Hall 2017).
Here, analysis of the surrounding regions of the Tn6172-var in pCl107 showed the presence of a Tn6022 :: ISAba11 precisely where Tn6022 would be, relative to Tn6172-var, in AbGRI1 variants ( Fig. 1). In pCl107, the ISAba11 copy interrupts the tniC gene of the Tn6022 that is missing the tniD, and parts of the tniE and tniB genes. Notably, while the sequence of this Tn is identical to Tn6022, the ISAba11 copy is present precisely where an ISAba11 copy is also present in Tn6021 in pAB3 (Hamidian and Hall 2017). The Tn6022 :: ISAba11 structure found in pCl107 shares the IS-Aba11 position with pAB3 and the sequence of Tn6022 with Ab-GRI1 islands, indicating that the ISAba11 insertion is an earlier event. Tn6021 in pAB3 has further diverged from Tn6022 by the acquisition of a short DNA spanning the 3'-end of tniC and 5'-end of tniA from an unknown source via homologous recombination. It also indicates that pCl107 diverged differently by a deletion in the tniDE region before it moved to GC2s (prior to the acquisition of ISAba11). Putting together all this information, the resistance region found in pCl107 serves as an important missing link in the complex evolutionary history of AbGRI1 islands as it provides further support for the theory that this segment has transposed as a single unit into the comM gene of the ancestor of the GC2 clone.
In addition to the strAB-sul2-tetA(B) region, another resistance cluster region encompassing aacC2 and aacA1 genes was found in pCl107. Indeed, these genes are absent in pD4, pA297-3 and pD46-4 (GenBank acc. no. KT779035, KU744946, MF399199 respectively) but present in several other A. baumannii plasmids such as pMC1.1, p1AR_0088, pA1429c, pHWBA8_1, and p40288 (GenBank acc. no. MK531536, CP027531, CP046899, CP020596, CP077802, respectively) ( Table 1). The two aminoglycoside resistance genes are surrounded by several complete and truncated mobile elements, including two CR1 copies, an ISKpn11 , an IS1008, and an IS1006 . In the absence of an ancestral structure in GenBank, it was not possible to determine the evolution of this region. Hence,    Figure 1. Comparison of AbGRI1 and resistance regions found in pA297-3, pCl107, and pAB3. Red horizontal lines represent plasmid backbones and the chromosome of global clone 2 (GC2) is shown by a thick line with the comM gene present on either side of AbGRI1. Arrows represent the extent and orientation of genes or orfs, with their names shown below. Resistance genes are in red. The extents of Tn6021, Tn6022, Tn6172, and Tn6174 are shown below or above the genomic region. Vertical bars at the ends of the elements indicate their IRs. Different shades of grey join regions with significant DNA identities, with identities shown by red numbers. ISAba1, ISAba18 and CR2 are shown using a green box with a white arrow indicating the direction of the tnp gene. ISAba11 and GIsul2:: ISAba1 are shown above their location. The GenBank acc. no. for pA297-3, pCl107, and pAB3 are KU744946, CP098522, and CP012005, respectively. The scale bar is shown.
further sequence data will be required to show how this complex resistance region has evolved. pCl107 has an intact merDACPTR module,>99% DNA identical to that carried by pA297-3. This mercury resistance module is present in many plasmids, as described previously (Hamidian, Ambrose and Hall 2016) (Table 1). However, it is absent the pD4.

pCl107 encodes a type 1 BacteRiophage Exclusion (BREX) system
The pCl107 has a BREX (BacteRiophage Exclusion) system located at bases 125 913 to 139 090 (of GenBank acc. no. CP098522). BREX system has recently been described as a phage resistance system and added to the previously known mechanisms (CRISPR-Cas and restriction-modification systems) in the arsenal of bacterial defence against phages (Goldfarb et al. 2015, Hampton, Watson andFineran 2020). The BREX system includes at least six subtypes, each with a different gene component and operon organization (Goldfarb et al. 2015). Our analysis showed that the BREX system carried by pCl107 belongs to Type 1 as it includes all six genes (brxABC, pglXZ, and brxL) that are characteristic of this type (Fig. 2). This type shares two genes, pglX encoding an adeninespecific DNA methyltransferase and pglZ encoding an alkaline phosphatase, with the pgl system, which was first identified in Streptomyces coelicolor (Goldfarb et al. 2015). The other four genes encode a Lon-like protease (brxL), an-ATP binding protein (brxC), an unknown protein (brxB), and NusB-like RNA binding antitermination protein (brxA).
A previous study has shown that BREX and BREX-like systems are distributed in about 10% of microbial genomes (Goldfarb et al. 2015). Although BREX has been reported in plasmids of Acinetobacter bereziniae (Brovedan et al. 2019) and A. baumannii (Cerezales et al. 2020), no comparative analysis has been done. To identify BREX systems related to that found in pCl107, the GenBank nonredundant database was searched using the DNA sequence of this region (from pCl107) as a query. Several matches were found including 36 A. baumannii plasmids either belonging to pCl107 (26) or unrelated (10), fourteen A. baumannii chromosomes, and sev-eral other entries from Acinetobacter spp., listed in Table S3, were found. These BREX modules differ by the presence and absence of additional genes between pglZ and pglX (Fig. 2, Table S3), which can differentiate the three BREX gene clusters found in A. baumannii plasmids as well as those in A. baumannii chromosomes (Fig. 2).
The phylogenetic analysis of BREX regions revealed the presence of many clusters where BREX loci with similar arrangements and identities generally clustered together (Fig. 3). For instance, the BREX gene clusters found in A. baumannii include a DUF262 domain-containing HNH endonuclease family protein encoding gene between pglZ and pglX, which is clustered in three different subgroups (Fig. 3) where each cluster has different DUF262 amino acid identities: identity 65% between those of pA297-3 and pAB_C63_1 (GenBank acc. no. KU744946 and CP051867, respectively) and 59% between those of pA297-3 and DT0544C chromosome (GenBank acc. no. KU744946 and CP053215 respectively).
We found 24 plasmids related to pCl107 that include a complete BREX gene cluster. Two gene clustering patterns were generally identified-one group (15 plasmids including pA297-3) included the DUF262 domain-containing HNH endonuclease family protein-encoding gene, and the second group (with nine plasmids including pCl107) did not have this gene (Table 1, Fig. 3). The DNA identities between BREX pCl107 and pA297-3 ranges between 88% to 96%, indicating thus the presence of two different pathways for the evolution within the plasmids' BREX and perhaps different impact on immunity against phages (Fig. 3).
Many clusters delimited by the phylogenetic tree are diverse in species and genetic backgrounds (Fig. 3). For instance, the A. baumannii plasmid BREX Region (pAb-C63_1) clusters together with BREX regions from Acinetobacter johnsonii, A. indicus, and A. pittii. This indicates the extent to which horizontal gene transfer is possibly taken between Acinetobacter spp. and even between plasmid and chromosomal genomes to mitigate the phage pressure.

The ptxABCDE phosphonate metabolism module
Phosphorus is an essential component of cell molecules, metabolism, and structure, predominantly present in nature in Figure 2. Complete BREX structures, related to the pCl107 BREX module, found in A. baumannii plasmids and chromosomes. The BREX genes are shown using blue arrows. Other genes are shown using white arrows. duf262 encodes a DUF262 domain-containing HNH endonuclease family protein, dprA encodes a DNA-processing protein DprA, recQ codes for a RecQ family ATP-dependent DNA helicase, and hp encodes a hypothetical protein. The figure is drawn to scale.
its most oxidized state as inorganic phosphate (pi) and which is also directly incorporated into biological molecules (Bisson et al. 2017). For instance, in Escherichia coli there are more than 30 genes in a regulon called PHO to sense, acquire and process phosphate (Stasi, Neves and Spira 2019). Under phosphate limitations, bacteria can acquire phosphorus by metabolizing reduced phosphorus compounds. Cl107 does not include the ph-nCDEFGHIJKLMNOP operon, which is present in E. coli and involved in phosphonate transport and assimilation via C-P lyase complex, however, instead, it contains a putative cluster (phnXW-HpnW-ABC transporter and regulator encoding genes located at bases 2 485 529-2 493 621) for transporting and processing the 2aminoethylphosphonate (2-AEPn) which is one of the most widely occurring phosphonates in nature (Borisova et al. 2011, Zangelmi et al. 2021. Interestingly, an additional cluster potentially linked to phosphate metabolism (ptx module) was also found in pCl107. This ptx module consists of five genes, namely ptxABCDE, that are related to those found in Pseudomonas stutzeri (amino acid identities for PtxA: 59%, PtxB: 60%, PtxC: 64%, PtxD: 53%, PtxE: 38%). ptxABC encodes an ABC transporter system where PtxA (IPR012693 family with IPR003439 and IPR003593 domains) is an ATP-binding protein, PtxB (IPR005770 family with PF12974 domain) is a periplasmic-binding protein and PtxC (IPR005769 family with domain IPR000515) is a membrane-spanning protein. ptxD encodes a phosphonate (traditionally named phosphite) dehydrogenase (IPR006140 and IPR006139 domains) and ptxE encodes a transcriptional regulator (IPR005119 and IPR000847 domains) belonging to the LysR family. Additionally, the Cl107 PtxB has the same phosphite-specific cap residues as the PtxB family that are not found in other homologs as C-P lyase-linked phosphonate binding PhnD of E. coli and HtxB (part of HtxBCDE hypophosphite transporter), highlighting a potential role for the Cl107 PtxABC as a phosphonate (traditionally named as phosphite) transporter ( Fig. S1) (Bisson et al. 2017). GenBank was searched to find ptx modules closely related to that found in pCl107. Eighteen (n = 18) plasmids were found to carry either complete or partial segments of the ptx module found in pCl107. Seven were in A. baumannii plasmids related to pCl107 (Table 1), 5 in A. pittii, and 6 in Acinetobacter spp. plasmids (Table 2). Notably, all the ptx modules found in A. baumannii plasmids are incomplete (Table 1). Amongst others, the ptx module is found in two A. pittii plasmids (p1DUT-2 and p1Ac-14 GenBank acc. no. CP014652 and CP063770, respectively) are identical to that in pCl107 (Fig. 4). Given that most strains containing ptx modules were recovered in environmental samples suggests a potential environmental source for this ptx module and that it has also been introduced to clinical strains by plasmids related to pCl107 (Table 2).

pCl107 encodes a class B cytochrome P450
Additionally, in pCl107, a class B cytochrome P450 with three components: a FAD-containing flavoprotein 'NAD(P)H-dependent reductase' (domains: IPR028202, IPR023753), P450 (families: IPR002397, IPR001128), and an iron-sulphurs protein (family: IPR001055; domain: IPR001041) is present in the vicinity of the ptx module (Fig. 4). Although found in organisms across the tree of life, cytochrome P450 enzymes (CYP or P450s) have attracted particular interest primarily for their potential usages in bioremediation, where bacterial P450 often oxidize environmental chemicals to ease their catabolism (Behrendorff 2021). These enzymes are classified into several families and subfamilies (Nelson 2006). The class B cytochrome P450 encoded by pCl107 is a 463 aa iden- Figure 3. Maximum likelihood phylogenetic tree of BREX sequences drawn using the FastTree software. The phylogenetic tree was constructed using the DNA sequence of 90 BREX regions from brxA to brxL. The clusters encompassing A. baumannii are highlighted according to the nature of the intruding gene. The clusters with no intruding gene are highlighted in blue (enclosing the pCl107) or orange (without pCl107). The yellow and pink clusters contain one intruding gene encoding the DUF262 domain containing the HNH endonuclease family and hypothetical protein, respectively. The green cluster includes the BREX structure with three intruding genes. The first two or three letters of different taxa names include an abbreviation of genus and species of each entry followed by their plasmid/chromosome names (those start with letter p refere to plasmid names and other are chromosome/strain names) and GenBank accession numbers. The abbreviations of species names are: Ab for A. baumannii (red), Ap for A. pittii (Dark blue), An for A. nosocomialis (brown), Aj for A. johnsonii (pale brown), As for A. schindleri (orange), Ai for A. indicus (bitter lime), A. sp for Acinetobacter spp. (black), Ar for A. radioresistens (turquoise), Ah for A. haemolyticus (purple), Al for A. lwofii (medium purple), Apl for A. pseudolwofii (pink), Api for A. piscicola (grey), Aw for A. wuhouensis (light green), Av for A. variabilis (dark red), Abe for A. bereziniae (maroon), Ave for Acinetobacter venetianus (olive). Shimodaira-Hasegawa local support values ranged from 0 to 1.   Other genes with known functions are indicated by grey arrows. The genes likely involved in purine metabolism are coloured in light orange. The genes uacT for uric acid transporter encoding gene, hiuH for hydroxyisourate hydrolase encoding gene, uao for 2-oxo-4-hydroxy-4-carboxy-5-ureidoimidazoline decarboxylase encoding gene, puuE for allantoinase encoding gene, allA for ureidoglycolate lyase encoding gene, alC for allantoicase encoding gene.
tical to that of Moraxellaceae P450 (e.g. GenPept accession number WP_004 658 516), which belongs to the family 153 of the subfamily A, which is also 71% aa identical to CYP153A1 of Acinetobacter calcoaceticus (https://drnelson.uthsc.edu/bacteria/). In plasmids related to pCl107, all the P450 identified belong to the same family, with the majority 99% identical to the A. calcoaceticus P450 and only two almost identical to pCl107 P450 (Table 1). Indeed, CYP153 proteins are known to encode alkane hydroxylases able to degrade short-and medium-chain-length n-alkanes and are often retrieved in petroleum-contaminated soil, groundwater, and coastal seawater, adding more shreds of evidence on the potential environmental ancestor of pCl107 (Wang et al. 2010, Nie et al. 2014. Similarly, a recent study highlighted that the bacterial lifestyle in Gammaproteobacteria species shaped to such an
The uacT gene encodes a permease that belongs to the xanthin.uracile permease family proteins (IPR006042 and the UacTlike: IPR017588). The hiuH and uao genes encode an hydroxyisourate (HIU) hydrolase (family: IPR014306) and 2-oxo-4-hydroxy-4-carboxy-5-ureidoimidazoline (OHCU) decarboxylase (domain: IPR017595) respectively. The HiuH and Uao proteins are responsible for the second (hydrolysis of HIU to OHCU) and third (the decarboxylation of OHCU to (S)-allantoin) steps of uric acid processing, respectively (Lee et al. 2013). The puuE gene is an allantoinase encoding gene catalyzing the hydrolysis of (S)-allantoin to allantoate. While these three encoded enzymes can potentially take an HIU and transform it to allantoate, the urate oxidase (also known as HpxO in Acinetobacter (Michiel et al. 2012)) encoding gene responsible for the first step of uric acid catabolism is missing, likely due to an ISAha2-mediated deletion event in pCl107. Hence, the pCl107 uric acid module is likely unable to hydroxylate urate into HIU.
To find plasmids that can potentially catabolize uric acid, we searched 616 complete A. baumannii plasmids that were publicly available in the GenBank non-redundant database as of September 2021. Nine plasmids (related to pCl107 carried by A. baumannii strains) were found, including four plasmids with the hpxO gene and five plasmids without hpxO (Table 1 and S4). Among those plasmids with the hpxO gene, pA297-3 and pA1429c (GenBank acc. no. KU744946 and CP046899, respectively) include a related se-quence (DNA identity 99%) with a complete set of uric acid module genes except for the tsx, which is replaced by ISAha2 (Fig. 5).
Additionally, several partial uric acid modules related to that in pCl107 (puuE-uao-hiuH-uacT) were found in multiple Acinetobacter spp. chromosomes and plasmids (Table S5). Among Acinetobacter spp. plasmids, p1DUT-2 (GenBank acc. no. CP014652) shares a similar module and surroundings to pA297-3 with 99.9% DNA identities (Fig. 5). Amongst the entries found here, only pACI-3569 and p1FDAARGOS_1092 (GenBank acc. no. CP026416 and CP068202) have uric acid modules almost identical to that in pCl107 (99.9% DNA identity) (Fig. 5). In addition to the hpxO gene, those with complete modules also contain genes encoding allantoicase, which catalyzes the degradation of allantoate to urea and ureidoglycolate, and ureidoglycolate hydrolase that degrades ureidoglycolate to glyoxylate and another molecule of urea (Fig. 5). As well, it harbors a tsx gene located next to uacT gene and encoding a nucleoside-specific channel-forming protein Tsx. These uric acid clusters in those plasmids could represent a more ancestral clusters of p1DUT-2 and pA297-3, which have subsequently given rise to the pCl107 cluster (Fig. 5). It is also notable that the pCl107 uric acid module was found to be closely related to several Acinetobacter strains (Table S5); Acinetobacter lowfii H7 and Al-065 and Acinetobacter spp. ACNIH1 (GenBank acc. no. CP072549, CP078045 and, CP026420 respectively) have DNA identities ranging from 99% to 100%, suggesting a chromosomal origin.
Uric acid is a potent antioxidant in humans protecting cells from reactive oxygen species. It is released by necrotic mammalian cells as a signal to activate the immune system against microbes. It is well known that bacteria can exploit uric acid as a source of nitrogen and even carbon and energy. However, whether there is any relationship between the presence of uric acid and microbial virulence is poorly understood. In A. baumannii, genes associated with purine and pyrimidine biosynthesis have been identified within novel gene clusters required for surface-associated mobility (Blaschke, Skiebe and Wilharm 2021), raising potential roles of purine metabolism-associated and uric acid catabolism genes.

Conclusion
This study adds another complete genome sequence of an ST25 A. baumannii strain recovered in the MENA region in Lebanon and characterizes a large plasmid carried by this strain named pCl107. pCl107 carries a set of antimicrobial resistance genes (aacA1, aacC2, sul2, strAB, tetA(B)) distributed in two complex regions and many non-antimicrobial resistance clusters. We showed that the pCl107 region encompassing the sul2, strAB, tetA(B) is closely related to AbGRI1 islands, commonly found in the chromosomal comM gene of GC2 strains, with an addition of the insertion of IS-Aba11 as the only difference (Fig. 1). This ancestral form of Ab-GRI1 provides yet another piece of evidence for the transposition of this unit (before it acquired ISAba11) from a plasmid related to pCl107 into the chromosome of an ancestral GC2 strain. This is an important finding highlighting the role of plasmids related to pCl107 in the evolution and acquisition of large genomic islands that carry multiple antibiotic-resistance genes by the most widespread strains of A. baumannii, GC2.
pCl107 contains a Type 1 BREX system and constitutes one of the two main evolution patterns observed in BREX clusters found in its related plasmids. This could indicate an association with different impacts on bacterial immunity against phages. pCl107 also carries a ptx module for phosphonate metabolism and an incomplete uric acid catabolism module. Here, we untangled the relationships of these regions with several regions in many Acinetobacter spp. plasmids and chromosomal and identified possible ancestors for some of the clusters. The ptx module found in pCl107 is complete and is likely to be the ancestral structure compared to those found in other related plasmids in ST25 strains. The presence of those environmentally advantageous genetic clusters along with class B cytochrome P450 module in pCl107 and their relatedness with those from environmental isolates could pinpoint the potential environmental ancestor for pCl107.
Plasmids related to pCl107, pD4, and pA297-3 belong to an important class of Acinetobacter plasmids that play an important role in multidrug-resistant dissemination as well as regulation of virulence by encoding several regulatory functions interacting with other plasmids and the chromosome (Di Venanzio et al. 2019a, Weber et al. 2015, Hamidian, Ambrose and Hall 2016, Nigro and Hall 2017, 2019b, Benomar, Di Venanzio, and Feldman 2021, Sycz et al. 2021. These plasmids carry a range of genetic material and complex evolutionary history with many links to multiple antibiotic resistance and metabolic pathways.

Accession numbers
Complete genome sequence of Cl107 was deposited in GenBank under accession numbers CP098521 (chromosome) and CP098522 (pCl107). Illumina short reads and MinION long reads are also deposited under the Sequence Read Archive (SRA) accession numbers SRR20613520 and SRR20613519, respectively.