Complete Genome Sequencing of Four Arcobacter Species Reveals a Diverse Suite of Mobile Elements

Abstract

Arcobacter species are recovered from a wide variety of sources, including animals, food, and both fresh and marine waters. Several Arcobacter species have also been recovered from human clinical samples and are thus associated tentatively with food- and water-borne human illnesses. Genome sequencing of the poultry isolate Arcobacter cibarius H743 and the Arcobacter acticola, Arcobacter pacificus, and Arcobacter porcinus type strains identified a large number and variety of insertion sequences. This study presents an analysis of these A. acticola, A. cibarius, A. pacificus, and A. porcinus IS elements. The four genomes sequenced here contain 276 complete and degenerate IS elements, representing 13 of the current 29 prokaryotic IS element families. Expansion of the analysis to include 15 other previously sequenced Arcobacter spp. added 73 complete and degenerate IS elements. Several of these IS elements were identified in two or more Arcobacter species, suggesting movement by horizontal gene transfer between the arcobacters. These IS elements are putatively associated with intragenomic deletions and inversions, and tentative movement of antimicrobial resistance genes. The A. cibarius strain H743 megaplasmid contains multiple IS elements common to the chromosome and, unusually, a complete ribosomal RNA locus, indicating that larger scale genomic rearrangements, potentially resulting from IS element-mediated megaplasmid cointegration and resolution may be occurring within A. cibarius and possibly other arcobacters. The presence of such a large and varied suite of mobile elements could have profound effects on Arcobacter biology and evolution.

Introduction

The genus Arcobacter currently contains 27 named species (Whiteduck-Leveillee et al. 2016; Figueras et al. 2017; Ramees et al. 2017; Perez-Cataluna et al. 2018), which are recovered from a variety of animals, food, water, human clinical and environmental sources (Donachie et al. 2005; Collado et al. 2008; Collado and Figueras 2011; Levican et al. 2012; Ramees et al. 2017; Gilbert et al. 2019). The presence of these organisms in a multitude of hosts and environments makes them a worthy target for whole-genome sequencing, to analyze and characterize gene content relevant to, for example, host association or environmental adaptation. We sequenced the Arcobacter cibarius strain H743 (Levican et al. 2013) as part of a project to obtain complete genomes for all current Arcobacter species. Initial 454 sequencing with paired-end reads produced a highly fragmented assembly with 51 repeat contigs. Eighteen repeat contigs contained all or part of a mobile element and mobile elements were present at a minimum of 43 locations within the genome. Thus, the fragmented nature of the assembly was largely due to an abundance of insertion sequences in this organism.

Insertion sequences were also previously identified in the genomes of the Arcobacter species Arcobacterbutzleri (Merga et al. 2013), Arcobactercryaerophilus (Miller et al. 2018c), Arcobacterellisii (Miller et al. 2018a), and Arcobactertrophiarum (Miller and Yee 2018), with 1–16 intact IS elements observed per genome. Although insertion sequences are not unknown within the related genera Campylobacter and Helicobacter (Kersulyte et al. 1998, 2013; Arnold et al. 2011; van der Graaf-van Bloois et al. 2014; Miller et al. 2016), they are generally uncommon with approximately one to ten elements per genome, when present. Twenty-nine families have been defined for prokaryotic IS elements (Siguier et al. 2015). Nearly, all Campylobacter and Helicobacter IS elements are affiliated with the IS110, IS200/IS605, IS607, or IS1595 families.

Draft genome sequences were available for several Arcobacter genomes. Although draft genomes can be used to identify the presence of potential IS elements, the number of copies and locations for each element can often only be ascertained using complete genomes. Therefore, we sequenced the genomes of A. cibarius strain H743, Arcobacteracticola strain KCTC 52212^T (Park et al. 2016), Arcobacterpacificus strain LMG 26638^T (Zhang et al. 2016), and Arcobacterporcinus strain LMG 24487^T (Figueras et al. 2017) to completion and characterized their suite of IS elements. Most complete Arcobacter genomes possess only a few IS elements. However, a large number of IS elements were observed in these genomes. Insertion sequences from other Arcobacter genomes were also included to provide a comprehensive analysis and the potential role of these mobile elements in Arcobacter evolution and biology is discussed.

Materials and Methods

Growth Conditions and Chemicals

Arcobacter agar plates were prepared using Brain Heart Infusion (BHIA) or Marine (MA) broth (Thermo-Fisher Scientific, Waltham, MA) with 1.5% (w/v) agar and 5% laked horse blood. Arcobacter cibarius strain H743 and A. pacificus strain LMG 26638^T were grown aerobically at 30 °C for 48 h on either BHIA (A. cibarius) or MA (A. pacificus) plates. Arcobacter acticola strain KCTC 52212^T and A. porcinus strain LMG 24487^T were grown microaerobically (6% O₂, 7% H₂, 7% CO₂, and 80% N₂) at 30 °C for 48 h (A. porcinus) or 96 h (A. acticola) on BHIA plates. All chemicals were purchased from Sigma-Aldrich Chemicals (St. Louis, MO) or Thermo-Fisher Scientific. DNA sequencing reagents and consumables were purchased from Illumina Inc. (San Diego, CA) or Pacific Biosciences (Menlo Park, CA).

DNA Sequencing

Genomic DNAs (gDNA) were prepared using loopfuls (∼5 μl) of cells from the agar plates and the Promega Wizard Genomic DNA Purification Kit. Illumina HiSeq reads were obtained from SeqWright (Houston, TX) for A. cibarius strain H743. Illumina MiSeq libraries for the A. acticola, A. pacificus, and A. porcinus type strains were prepared and sequenced as described (Miller and Yee 2019). PacBio library construction and sequencing were performed as previously described (Miller et al. 2017; Miller and Yee 2019).

Genome Assembly and Annotation

SMRT sequencing resulted in single chromosomal contigs for all four strains, with an additional plasmid contig for A. cibarius strain H743. These contigs were circularized within Geneious Prime (ver. 2019.1.3; Biomatters Ltd, Auckland, New Zealand). For these five contigs, base calling was improved by mapping the HiSeq reads or quality-trimmed MiSeq reads onto the circularized PacBio contigs within Geneious Prime. Using the Geneious “Find Variations/SNPs” module, with a default minimum variation of 0.3, discrepancies between the Illumina and PacBio sequences were identified (usually 1-bp indels) and corrected to the Illumina consensus sequence. Final combined coverages ranged from 404 to 959×. Protein-, tRNA/transfer-messenger (tmRNA)-, and rRNA-encoding genes were identified using GeneMark (Besemer and Borodovsky 2005), ARAGORN (Laslett and Canback 2004), and RNAmmer (Lagesen et al. 2007), respectively. Open reading frames were annotated as described (Miller et al. 2018b).

Identification of IS Elements

Insertion sequences in the four Arcobacter genomes were identified and characterized as follows. First, the protein sequences of all the annotated transposases within the four genomes were compiled, aligned, and examined phylogenetically. For each phylogenetic cluster, the nucleotide sequences of the transposase genes, along with 200–250 bp of flanking sequences, were assembled within Lasergene SeqMan Pro (v. 15; DNAstar, Madison, WI). The ends of each element were determined through at least one of the following: 1) comparison of multiple copies of the same element; 2) analysis of genes inactivated by insertion of the element; 3) identification of any terminal imperfect inverted repeats (TIRs) and/or flanking direct repeats (DRs); and 4) comparison to IS element ends observed for other members of the same IS family (Siguier et al. 2015). Sequences representing each IS element were extracted; pseudosequences of elements inactivated by other IS elements were constructed through subtraction of the inactivating IS element sequences. IS element family/subgroup affiliations were determined through a BlastP query of the transposase sequence within ISfinder (https://www-is.biotoul.fr/index.php, last accessed February 4, 2020; Siguier et al. 2006). Each IS element was assigned a unique name/number designation (e.g., ISAact1, ISApac3) based on its species of origin and the position of the first identified copy in the chromosome (relative to dnaA) or plasmid. Within each species, IS elements with a ≥ 98% sequence identity were considered copies of the same element, with an additional letter designation appended (e.g., ISApac3A and ISApac3B). Names of interrupted, degenerate, or truncated IS elements were further appended with a terminal apostrophe (e.g., ISAcib6A’).

Results and Discussion

General Features of the Arcobacter Genomes

Genomic data for the four Arcobacter genomes are presented in table 1. The genome sizes ranged from ∼2 Mb (A. porcinus) to ∼3 Mb (A. acticola) with an average G + C content of 27.5%. A 237-kb megaplasmid (pACIB) was identified in the A. cibarius genome. This plasmid encodes two replication initiation proteins, an F-type type IV conjugative transfer system, and the tetracycline resistance gene tet(L); however, this plasmid is highly unusual, as it contains a complete ribosomal RNA locus, as well as several housekeeping genes typically located on the chromosome. These housekeeping genes encode proteins associated with amino acid biosynthesis (e.g., cysteine synthase, serine O-acetyltransferase, and threonine synthase) and transport of amino acids, sulfonate, sulfate (Sbp), molybdenum (ModA), and zinc (ZupT). Interestingly, although modB, modC, and modE are present on the A.cibarius chromosome, modA is present only on pACIB.

Table 1.

Genomic Data for the Four Arcobacter Species

	Value(s)a
	Arcobacter acticola	Arcobacter cibarius	Arcobacter pacificus	Arcobacter porcinus
Strain	KCTC 52212T	H743	LMG 26638T	LMG 24487T
Strain info/sequencing data
Source	Seawater	Poultry meat	Seawater	Piglet fetus
Location	Sea of Japan	Belgium	Pacific Ocean	Denmark
Coverage (×)(Illumina; PacBio; Total)	78; 326; 404	671; 228; 899	212; 342; 554	211; 748; 959
Accession #s	CP042652	CP043857 (chromosome)	CP035928	CP036246
		CP043858 (pACIB)
Genomic data
Generalb
Size (bp)	3,019,071	2,105,974 / 237,519	2,657,991	2,018,219
G+C content	27.57%	27.11% / 27.45%	28.03%	27.22%
CDS numbersc	2,845	1,984 / 189	2,493	1,989
Assigned function (% CDS)	985 (34.6)	859 (43.3) / 10 (5.3)	929 (37.2)	845 (42.5)
General function annotation (% CDS)	1,177 (41.4)	696 (35.1) / 101 (53.4)	924 (37.1)	625 (31.4)
Domain/family annotation only (% CDS)	222 (7.8)	129 (6.5) / 19 (10.1)	177 (7.1)	151 (7.6)
Hypothetical (% CDS)	461 (16.2)	300 (15.1) / 59 (31.2)	463 (18.6)	368 (18.5)
Pseudogenes	102	58 / 33	65	30
Genomic islands/CRISPR
Genetic islands	2	6; 0	4	7 [1]
Coding sequences in genetic islands	40 [2]	127 [12]; 0	79	229 [4]
CRISPR/Cas loci	0	0	Type II-C	0
IS elements/Mobile elements/Transposases	80 [6]	45 [20]; 21 [12]	66 [9]	28 [10]
Gene content/pathways
Signal transduction
Che proteins	9	9	9	8
Methyl-accepting chemotaxis proteins	9 [3]	12 [2]; 3 [3]	16 [5]	12
Response regulators	45 [2]	26 [1]; 2	44 [5]	19
Histidine kinases	56 [10]	31 [2]; 1 [2]	47 [9]	18[1]
Response regulator/histidine kinase fusions	8 [1]	0	5	1
Diguanylate cyclases	26 [4]	18 [1]; 0	22 [2]	6
Diguanylate phosphodiesterases	5 [1]	4 [1]; 0	7 [1]	3
Diguanylate cyclase/phosphodiesterases	10 [1]	10; 0	2 [1]	5
Other	13 [4]	7; 2	5 [2]	6
Restriction/modification
Type I systems (hsd)	2	1; 0	2	5
Type II systems	2	0	0	3
Type III systems	2	0	1	0
Transcription
Regulatory proteins	63 [2]	33 [3]; 4 [2]	41	41
Non-ECF σ factors	σ54, σ70	σ70	σ70	σ70
tRNAs, ribosomal loci	61, 6	51; 2, 5; 1	52, 5	48, 4
CO dehydrogenase (coxSLF)	Yes	No	Yes	No
Entner-Doudoroff pathway	Yes	No	Yes	No
Ethanolamine utilization (eutBCH)	Yes	No	No	No
Glycogen metabolism (glgABCP)	Yes	No	No	No
Nitrate reductase	Yes	Yes	No	Yes
Nitrogen fixation (nif)	Yes	No	No	No
Osmoprotection	proVWX	No	ectABC, proVWX	ectABC
Pyrroloquinoline quinone (PQQ) biosynthesis	Yes	No	Yes	No
Pyruvate → acetyl-CoA
Pyruvate dehydrogenase (E1/E2/E3)	No	Yes	Yes	Yes
Pyruvate: ferredoxin oxidoreductase	porABDG	No	porABDG	No
Urease	Yes	No	Yes	No

note.—CDS, coding sequence.

^aNumbers in square brackets indicate pseudogenes or fragments. Strain H743 genomic data values after the semicolon indicate plasmid-borne features.

^bStrain H743 values before the backslash are for the chromosome, whereas values after the backslash are for the pACIB plasmid.

^cNumbers do not include pseudogenes.

Analysis of the A. cibarius genome (chromosome + plasmid) identified 66 complete IS elements, 17 truncated, degenerate, or interrupted IS elements (termed hereafter as “degenerate”), and 15 transposase fragments (table 1), confirming the initial 454 sequencing results for this strain. A large number of complete IS elements were also identified in the A. acticola (80 elements), A. pacificus (66 elements), and A. porcinus (28 elements) genomes (table 1).

Characterization of the IS Elements in the Four Arcobacter Genomes

Two hundred seventy-six IS elements (240 complete, 36 degenerate) were identified in the four genomes characterized here; an additional 21 transposase fragments were also identified. Based on a 98% sequence identity cutoff, 77 unique elements were present in this set: 30 elements were present as a single copy, with the remainder present in 2 (e.g., ISApor5) to 28 (e.g., ISApac1) copies (supplementary table 1, Supplementary Material online). Of the 36 degenerate elements, 16 were inactivated through the insertion of another IS element; element ISAcib25’ was inactivated by two separate IS elements (ISAcib3K and ISAcib1C) and element ISAcib5D’ was inactivated by ISAcib1E, which was subsequently inactivated via insertion of ISAcib14B. Six degenerate elements were truncated, presumably due to the insertion of another IS element at or near one of the element ends with subsequent deletions. The remaining 14 degenerate elements possessed frameshift mutations within the transposase gene.

Phylogenetic and sequence analysis of these IS elements also indicated that they represent at least 13 (fig. 1 and supplementary table 1, Supplementary Material online) of the 29 current prokaryotic IS families (Siguier et al. 2015). One or two additional families may be represented by the phylogenetic clade containing elements ISAact16 and ISApac13 (fig. 1), a cluster with little similarity to any of the current families in ISfinder. Members of two families in this study can be sorted further into previously characterized subgroups (Siguier et al. 2015; fig. 1): members of the IS3 family are divided into three subgroups (i.e., IS3, IS51, and IS407) and IS256 members form two subgroups (IS256 and IS1249). In addition, phylogenetic analysis suggests that three additional families here (IS5, IS1380, and IS1634) are divided into two clades each, with the differences between the clades reflected in sequences at the element ends and/or element size. Most Arcobacter IS elements contain a single open reading frame; however, elements associated with four families, IS3, IS21, ISAs1, and Tn3, contain two or three open reading frames.

Fig. 1.

—Arcobacter insertion sequence families. Three hundred thirty-four Arcobacter IS element sequences were compiled and aligned using ClustalW. The dendrogram was constructed using the maximum likelihood method based on the Tamura–Nei model (Tamura and Nei 1993). Bootstrap values of ≥75%, generated from 500 replicates, are shown at the nodes. Initial tree(s) for the heuristic search were obtained by applying the neighbor-joining method to a matrix of pairwise distances estimated using the maximum composite likelihood approach. The tree is drawn to scale with branch lengths measured in the number of substitutions per site. All positions with <95% site coverage were eliminated. There were a total of 654 positions in the final data set. Evolutionary analyses were conducted in MEGA6 (Tamura et al. 2013). Some branches were collapsed based on sequence similarity to improve legibility. Labels were colored according to family/group affiliation. Degenerate IS elements are in parentheses.

With the exception of IS elements affiliated with the IS110 and ISAs1 families, the Arcobacter IS elements are bounded by TIRs, with many further bounded by DRs (supplementary tables 1 and 2, Supplementary Material online). These TIRs range from 11 to 37 bp, show strong sequence similarity with respect to family/subgroup/cluster affiliation, and possess conserved bases near the TIR ends that are consistent with those reported in other members of the same families (supplementary table 1, Supplementary Material online).

IS Elements in Other Arcobacter Taxa

To obtain a more comprehensive picture of the IS elements in the entire Arcobacter genus, we analyzed genomes representing an additional 15 Arcobacter spp. (supplementary table 3 and supplementary fig. 1, Supplementary Material online). Sixty-five complete and eight degenerate IS elements were identified in these genomes, for a total of 349 complete and degenerate IS elements across the 19 species. Four strains (Arcobacteranaerophilus DSM 2463^T, A. butzleri 7h1h, A. ellisii LMG 26155^T, and A. trophiarum LMG 25534^T) contained >5 IS elements, with the remainder of the genomes possessing zero to three elements each. These additional elements were members of seven IS families that formed a subset of the families identified above. Moreover, five of these additional elements were members of the novel phylogenetic clade described above (fig. 1).

Horizontal Gene Transfer among the Arcobacter IS Elements

Several IS elements from different taxa show strong sequence similarity (fig. 1), suggesting that some members of the Arcobacter IS element suite have moved from species to species via horizontal gene transfer. For example, the IS4 elements ISAact4 (A and B), ISAcib9B, and ISAell5 (A and B) are 1,486 bp, with identical TIRs and 97–100% sequence identity. Other examples are the IS1182 elements ISAact7 and ISApor3 (1,493 bp, identical TIRs, 99.9% sequence identity) and the IS3 elements ISAcib23 and ISAcryA1 (1,289 bp, similar TIRs, 98% sequence identity). The sequence similarities for these three examples are much higher than the average nucleotide identity between the A. acticola, A. cibarius, A. cryaerophilus, A. ellisii, and A. porcinus species (75–81%). Several IS elements are located within genomic islands and on three megaplasmids (pM830MA, pACRY43158, and pACIB), suggesting at least two possible mechanisms of horizontal gene transfer between Arcobacter strains.

Genomic and Phenotypic Implications of the Arcobacter IS Elements

The presence of a wide variety of IS elements in a single genome, with many elements in multiple copies, would be expected to impact gene content and genomic structure. Predictably, the three strains with the largest number of IS elements also contain the most pseudogenes (65–102; table 1), whereas those with <10 IS elements contain only 7–34 pseudogenes (supplementary table 3, Supplementary Material online). Chromosomal recombination between copies of an element oriented in the same direction would lead to deletions while recombination between oppositely oriented copies would result in an inversion. Park et al. (2016) described the A. acticola type strain as nonflagellated. However, the A. acticola KCTC 52212^T genome contains a flagellin gene and a full set of chemotaxis genes, in contrast to other nonmotile strains in which these genes are absent (Miller et al. 2015). Flagellar genes in the related species Arcobactervenerupis (supplementary fig. 1, Supplementary Material online) are arrayed in four clusters. In A. acticola, IS elements are present in three of the four flagellar gene clusters, with deletions in two of the four clusters. Thus, 26 of the 38 flagellar genes in these four clusters are either insertionally inactivated or deleted, potentially via IS element-mediated recombination. Evidence for IS element-mediated inversion can be observed by analyzing the ISApac1 set (supplementary table 2, Supplementary Material online). Copies of this element are defined by 7–9 bp DRs. However, the left DR of element ISApac1I is identical to the right DR of ISApac1AB’ and vice versa. As these two elements are oppositely oriented, it is likely that recombination occurred between them in the past, resulting in a chromosomal inversion and swapping of the DRs.

The Arcobacter IS elements can sometimes form more complex structures. Arcobacter thereius contains an aminoglycoside resistance gene flanked by two IS4 family elements (ISAth2A/ISAth2B). Arcobacter porcinus contains an identical triad (flanked by ISApor4B/ISApor4C); however, in this strain, another triad, containing an unrelated aminoglycoside resistance gene (18% identity), was inserted into ISApor4C. ISApor4C was subsequently inactivated via insertion of the IS3 element ISApor5B to create a composite mobile element containing two different drug resistance genes. The A. thereius and A. porcinus elements are both putatively present on different genomic islands.

The A. cibarius H743 megaplasmid contains a high density of mobile elements; insertion sequences or transposase gene fragments comprise 20% of pACIB. The presence of 11 identical elements on both the A. cibarius chromosome and the megaplasmid pACIB suggests that the A. cibarius megaplasmid could undergo cycles of cointegration and resolution. Evidence of this may be seen with the presence of ribosomal RNA and other housekeeping genes on the megaplasmid that almost surely have a chromosomal origin. Additionally, as pACIB contains a type IV conjugative transfer system, chromosomal genes could potentially be transferred into another strain either through conjugation or by an Hfr-type mechanism.

Conclusions

This study identified a diverse suite of insertion sequences in 19 Arcobacter species. Notably, the major IS families within Campylobacter and Helicobacter, that is, IS200/IS605 and IS607, were not identified in this study, suggesting that lateral transfer between those two genera and Arcobacter is limited. Additional A. cibarius and A. porcinus genomes were previously sequenced to draft level. Multiple and varied transposases were identified in these genomes, indicating that the IS elements and families described above are not confined to the set of strains characterized in this study. The Arcobacter IS elements will provide a rich source for future research.

Supplementary Material

Supplementary data are available at Genome Biology and Evolution online.

Data deposition: All genome sequencing data have been deposited at GenBank under accession numbers CP035928, CP036246, CP042652, CP043857, and CP043858.