Plasmid 50 (2003) 131–144 www.elsevier.com/locate/yplas
Three small, cryptic plasmids from Aeromonas salmonicida subsp. salmonicida A449 Jessica Boyd,* Jason Williams, Bruce Curtis, Catherine Kozera, Rama Singh, and Michael Reith National Research Council Institute for Marine Biosciences, 1411 Oxford Street, Halifax, NS, Canada B3H 3Z1 Received 21 February 2003, revised 2 June 2003
Abstract The nucleotide sequences of three small (5.2–5.6 kb) plasmids from Aeromonas salmonicida subsp. salmonicida A449 are described. Two of the plasmids (pAsa1 and pAsa3) use a ColE2-type replication mechanism while the third (pAsa2) is a ColE1-type replicon. Insertions in the Rep protein and oriV region of the ColE2-type plasmids provide subtle differences that allow them to be maintained compatibly. All three plasmids carry genes for mobilization (mobABCD), but transfer genes are absent and are presumably provided in trans. Two of the plasmids, pAsa1 and pAsa3, carry toxin–antitoxin gene pairs, most probably to ensure plasmid stability. One open reading frame (ORF), orf1, is conserved in all three plasmids, while other ORFs are plasmid-specific. A survey of A. salmonicida strains indicates that pAsa1 and pAsa2 are present in all 12 strains investigated, while pAsa3 is present in 11 and a fourth plasmid, pAsal1, is present in 7. Crown Copyright Ó 2003 Published by Elsevier Inc. All rights reserved. Keywords: Furunculosis; Plasmid addiction; ColE1-type replicon; ColE2-type replicon
1. Introduction Aeromonas salmonicida subsp. salmonicida is the aetiological agent of the salmonid disease furunculosis, a disease with both high mortality and morbidity (Smith, 1997). Furunculosis is a significant cause of economic loss in salmonid aquaculture throughout the world. Furunculosis is a complex disease and exists in different forms
* Corresponding author. Fax: +902-426-9413. E-mail address:
[email protected] (J. Boyd).
depending on the health, age, and species of fish and the conditions of their environment, particularly temperature. The acute form is characterized by rapid onset of general septicaemia, melanosis, inappetence, lethargy, and haemorrhage at the base of the fins. The chronic form is characterized by slow-onset with low mortality, affected animals often have raised skin lesions called furuncles which are considered pathognomic for the disease. Conversion from the chronic to the acute form can be caused by environmental stressors. Aeromonas salmonicida belongs to the family Aeromonadaceae of the c proteobacteria.
0147-619X/$ - see front matter. Crown Copyright Ó 2003 Published by Elsevier Inc. All rights reserved. doi:10.1016/S0147-619X(03)00058-1
132
J. Boyd et al. / Plasmid 50 (2003) 131–144
A. salmonicida is further subdivided into typical (A. salmonicida subsp. salmonicida) and atypical strains. Many groups have surveyed plasmid carriage in both typical and atypical A. salmonicida (Bast et al., 1988; Belland and Trust, 1989; Hanninen et al., 1995; Pedersen et al., 1996). A. salmonicida carries multiple plasmids of various sizes: small, multi-copy plasmids ranging from 1 to 6 kb, and larger low-copy plasmids from 11 to 150 kb. Most strains of A. salmonicida carry at least four plasmids, and some, as many as six. The variation in the number and size of plasmids is greater for atypical than for typical strains. Some of the large plasmids are known to carry antibiotic resistance genes, but other than this no correlation can be drawn between plasmid carriage and virulence (Brown et al., 1997). Most typical A. salmonicida strains carry a group of three small plasmids of 5.0, 5.2, and 5.4 kb. In the typical A. salmonicida strain A449, their copy number has been estimated to be very high, between 50 and 55 per cell (Belland and Trust, 1989). Further characterization of these plasmids demonstrated that they expressed very few genes and, despite great effort, it was not possible to cure the strain of any of these three plasmids (Belland and Trust, 1989). In addition to the three small plasmids, Belland and Trust (1989) identified a large, 145 kb plasmid in A. salmonicida A449. As part of the process of sequencing the entire genome of A. salmonicida A449, which is estimated to be 4.6 Mb (Umelo and Trust, 1998), we have
identified contigs corresponding to the three small plasmids. This paper describes the complete sequence and annotation of these plasmids from A. salmonicida A449.
2. Materials and methods 2.1. Bacterial strains Aeromonas salmonicida subsp. salmonicida strains used were kindly donated by Dr. Gilles Olivier of the Department of Fisheries and Oceans (DFO) in Moncton, N.B., Dr. Trevor Trust, Microtek International, Saanichton, B.C., and by Dr. Rafael Garduno at Dalhousie University, Halifax, N.S. They are described in Table 1. A. salmonicida strains were grown in Tryptic Soy broth (TSB, Difco) for 3 days at 17 °C with shaking. Plasmids were isolated from A. salmonicida using the Nucleobond Mini Prep kit (Clontech, Palo Alto, CA). Genomic DNA was isolated using the PureGene DNA isolation kit (Gentra Systems, Minneapolis, MN). 2.2. DNA sequencing Aeromonas salmonicida plasmid DNA was digested with BamHI and BamHI/EcoRI and cloned into pre-digested and dephosphorylated pTrueBlue vector (Genomics One, Laval, PQ) and transformed into Escherichia coli XL-1 Blue MRFÕ (Stratagene, La Jolla, CA). E. coli clones were
Table 1 Aeromonas salmonicida subsp. salmonicida strains used in this study Strain
Characteristics and origin
Source
A449 A450 A450-1 A450-3 80204 80204-1S SS70.1 NG10 97132 84222-S OAR MT004
Virulent, Brown trout, Eure, France Virulent, Brown trout, Tarn, France Lab-derived avirulent variant of A450 Lab-derived avirulent variant of A450 Virulent, Atlantic salmon, New Brunswick, Canada Lab-derived avirulent variant of 80204 Lab-derived avirulent, originally from Coho salmon, Oregon, USA Virulent, Atlantic salmon, New Brunswick, Canada Virulent, Atlantic salmon, New Brunswick, Canada Lab-derived avirulent, originally from Atlantic salmon, New Brunswick, Canada Virulent, oxolinic acid resistant, Atlantic salmon, New Brunswick, Canada Lab-derived avirulent, originally from Atlantic salmon, Scotland
T. Trust R. Garduno R. Garduno R. Garduno G. Olivier G. Olivier G. Olivier G. Olivier G. Olivier G. Olivier G. Olivier G. Olivier
J. Boyd et al. / Plasmid 50 (2003) 131–144
grown and plasmid DNA isolated using standard techniques. Clones containing the appropriatesized inserts were sequenced from both ends. These end-sequences were used as a tag to identify contigs from the main A. salmonicida genome assembly that had been generated by sequencing random clones and assembled using the Staden package (Bonfield et al., 1995). Primer-walking on the plasmid clones was done to fill any gaps in the respective contigs as well as to confirm the sequences. Total sequencing coverage on these plasmids is approximately 12-fold. These sequences have been deposited in GenBank under Accession Nos.: pAsa1, AY301063; pAsa2, AY301064; and pAsa3, AY301065. 2.3. PCR primers and methods PCR conditions to amplify specific regions of the plasmids were: 45 s at 94 °C; 30 cycles of 45 s at 94 °C, 45 s at 55 °C, and 90 s at 72 °C; followed by a 10 min extension at 72 °C. The reaction mix contained 100 ng genomic DNA, 0.3 mM each primer and 1.25 U of rTaq (Amersham–Pharmacia Biotech, Uppsala, Sweden). Primers used were: pAsa1F, 50 GGACGATTAACCTTCGCATC 30 ; pAsa1R, 50 GTATCGCCCAACTTCTTCCA 30 ; pAsa2F, 50 AAAAGAGCGTGCAACCCTAA 30 ; pAsa2R, 50 GCGATGCTACTTCATTCACC 30 ; pAsa3F, 50 TCATGGAGAATGTTCGCAAG 30 ; pAsa3R, 50 GCCCAATTATCACAGCAACA 30 ; pAsal1F, 50 TAACATGGGTGAGTCAGGA30 ; and pAsal1R, 50 TGCATGTTTGTAAAAAGTA GGTG 30 .
3. Results and discussion 3.1. General description of plasmids The restriction patterns of the completed plasmid sequences were compared to those generated by Belland and Trust (1989) who characterized but did not sequence three small plasmids from the same strain of A. salmonicida. The restriction maps of our plasmids were very similar to those previously reported, so we chose to use their nomenclature. The maps are not identical in size (pAsa1
133
and pAsa3) and restriction site pattern (pAsa2 and pAsa3) but this can be explained by the simultaneous mapping of all three plasmids by these authors, while we were able to sequence them individually. Plasmids pAsa1, pAsa2, and pAsa3 (Fig. 1) are 5424, 5247, and 5616 bp in length, respectively, with G + C contents of 57, 52, and 55%. The G + C content of the A. salmonicida chromosome is 58.3% (unpublished data). pAsa1 and pAsa3 appear to be ColE2-type replicons while pAsa2 is a ColE1-type plasmid. Both pAsa1 and pAsa3 carry toxin–antitoxin genes and all three plasmids encode genes for plasmid mobilization. Other than genes for plasmid replication, stability, and mobilization, each plasmid contains only one to three other open reading frames (ORFs)1 that encode proteins of unknown function. During the preparation of this manuscript, three A. salmonicida plasmid sequences were deposited in GenBank by D. Fehr, S.E. Burr, and J. Frey. One of these sequences, pAsal2 (Accession No: NC_004339.1), is identical to the pAsa1 sequence while pAsal3 (NC_004340.1) differs from the pAsa2 sequence at 4 positions and has 2 additional bases. Plasmid pAsal1 (NC_004338.1), another ColE2 plasmid, is not present in A. salmonicida A449, while pAsa3 was apparently not sequenced by Fehr et al. 3.2. Replication of plasmids pAsa1 and pAsa3 On the basis of similarity to other rep genes in the GenBank database, those of plasmids pAsa1 and pAsa3 are members of the ColE2 family (Table 2). Plasmids with ColE2-type replicons are usually small, have a high copy-number and replicate using the theta mechanism (del Solar et al., 1998; Espinosa et al., 2000). The minimum replicating unit consists of the rep gene; a short antisense RNA, RNAI, that is complementary to the 50 untranslated region of rep; and a cis acting origin, oriV, where the Rep protein binds. ColE2type Rep proteins are primases and thus they 1
Abbreviations used: ORF, open reading frame; DFO, Department of Fisheries and Oceans; TSB, Tryptic Soy broth; rm, restriction/modification system.
134
J. Boyd et al. / Plasmid 50 (2003) 131–144
Fig. 1. Maps of plasmids pAsa1, pAsa2, and pAsa3. Gene position and direction of transcription are indicated by arrows.
both bind to their cognate origin and synthesize a small primer RNA, ppApGpA, that is required for initiation of the leading-strand DNA synthe-
sis by the chromosomally encoded DNA polymerase I. The RNAI antisense RNA negatively regulates rep expression post-transcriptionally,
J. Boyd et al. / Plasmid 50 (2003) 131–144
135
Table 2 Percent similarity (lower left) and identity (upper right) of the Rep proteins
pAsa1 pAsa3 pAsal1 ColE2 ColE3 ColE5
pAsa1
pAsa3
pAsal1
ColE2
ColE3
ColE5
—
73
82 82 49 49 48
—
72 90
94 50 47 47
—
38 38 39
51 48 46
—
39 36 36 88
91 82
—
38 36 36 75 83
88
—
Similarity was calculated using the BLOSUM 62 matrix. GenBank Accession Nos. are as follows: pAsal1 (A. salmonicida), 23897236; ColE2-P9 (Shigella sp.), 808894; ColE3-CA38 (Escherichia coli), 808865; and ColE5-099 (Shigella sonnei), 809524.
although the exact mechanism of the regulation is unclear. ColE2-type plasmids are often mutually compatible, thus allowing some bacteria, including A. salmonicida, to carry more than one such plasmid. Incompatibility among ColE2-type plasmids is controlled by two factors: the RNAI molecule which specifically binds to its own sense transcript and controls copy number (Takechi et al., 1994) and the Rep protein which specifically binds to its cognate origin. Subtle variations in the sequence of the Rep protein and the origin of replication provide the possibility for mutual compatibility among ColE2-type plasmids (Hiraga et al., 1994; Shinohara and Itoh, 1996). Fig. 2 shows the predicted position of the antisense RNAI between the rep promoter and start codon in the A. salmonicida plasmids pAsa1, pAsa3, pAsal1, and three ColE2-type plasmids from E. coli and Shigella species. In the latter ColE2-type plasmids the RNAI promoters overlap the rep start codon on the opposite strand (grey boxes). The aligned sequences of the A. salmonicida plasmids are quite dissimilar from those of E. coli and Shigella and do not seem to carry RNAI promoter sequences in the same position. Instead putative )10 and )35 regions appear to reside about 30 bp upstream of the rep start codon (also shown with grey boxes). The difference in the position of the promoters reflects the different sizes of the A. salmonicida and ColE2 RNAI molecules. The RNAI molecules from the A. salmonicida plasmids are shorter and appear to form only a single stem–loop structure (solid arrows in Fig. 2), as does that of ColE5-099 and other ColE2 RNAI molecules (Hiraga et al., 1994). In contrast, the
RNAI molecules of ColE2-P9 and ColE3-CA38 form two stem–loop structures (solid and dashed arrows). The conserved stem–loop is very similar in all six plasmids, except that the loop sequences in the A. salmonicida plasmids are exactly complementary to those of the other ColE2 plasmids. A run of thymidines at the 50 end of the rep mRNA likely acts as a rho-independent transcriptional terminator for RNAI. In addition to the RNAI molecule, incompatibility among members of the ColE2 group has been shown to be controlled by three insertions in the C-terminal region of the Rep protein and corresponding insertions in the oriV sequence (Hiraga et al., 1994; Shinohara and Itoh, 1996). The three insertions in the Rep protein, termed A, B, and C are 9, 2, and 4–6 amino acids in length, respectively, and occur in the C-terminal 50 amino acids of the protein. Each is associated with a single nucleotide insertion in the origin region, a, b, and c; that occur at positions 5, 20, and 9 bp upstream of the ppApGpA primer, respectively. Chimeric rep and origin constructs (Shinohara and Itoh, 1996) demonstrate the specificity of these Rep/oriV insertions and their involvement in plasmid incompatibility. Alignment of the A. salmonicida pAsa1, pAsa3 and pAsal1 Rep, and oriV sequences with those of other ColE2 plasmids (Fig. 3A) reveals the presence of all three types of Rep/oriV insertions in the A. salmonicida plasmids. All three plasmids appear to have the B/b insertions, while pAsa1 additionally has the A/a insertions and pAsa3 has the C/c insertions. This variation in Rep/oriV type between the three A. salmonicida ColE2-type plasmids demonstrates how Rep/oriV specificity is
136
J. Boyd et al. / Plasmid 50 (2003) 131–144
Fig. 2. Alignment of the DNA sequences upstream of the rep genes of three sequenced A. salmonicida ColE2-type plasmids and three ColE2-type plasmids from E. coli and Shigella species. The rep-35 and rep-10 promoter region is indicated by boxes and is approximately 160 bases upstream from the rep start codon. Two sets of grey boxes indicate the most likely promoters for the antisense RNAI molecules. The promoters for the A. salmonicida RNAI molecules are not the same as those for the E. coli and Salmonella plasmids. Arrows indicate stem–loop structures formed in the RNAI molecule with the conserved loop region boxed. Dashed arrows indicate the second stem–loop structure found only in plasmids ColE2-P9 and ColE3-CA38. The RNAI terminator is likely the run of AÕs indicated by the solid line. Accession Nos. are: pAsal1, NC_004338.1; ColE2-P9, 487322; ColE3-CA38, 487267; and ColE5-099, 487324.
determined for each plasmid and indicates that all three ColE2-type plasmids would be compatible in a single strain. In addition, the A. salmonicida Rep proteins, while overall very similar to other ColE2type Rep proteins, have several conserved insertions throughout the length of the protein, as well as a single insertion unique to pAsa1 (Fig. 3A). The A. salmonicida Rep proteins thus appear to be a distinct subgroup of the ColE2-type Rep family.
3.3. Replication of pAsa2 On the basis of similarity to other plasmids, pAsa2 is a member of the ColE1-type group, and thus is a theta replicating, DNA polymerase dependent plasmid (Chan et al., 1985; del Solar et al., 1998; Espinosa et al., 2000). ColE1-type replication requires no plasmid-encoded proteins; instead it uses two RNA molecules. RNAII acts as the
J. Boyd et al. / Plasmid 50 (2003) 131–144
137
Fig. 3. (A) Alignment of the Rep proteins from the three sequenced A. salmonicida ColE2-type plasmids and three ColE2-type plasmids from E. coli and Shigella species. The conserved leucine zipper and helix–turn–helix motifs are boxed and shaded. The three insertions that determine compatibility are labeled A, B, and C. (B) Alignment of the origin of replication (oriV) of the three sequenced A. salmonicida ColE2-type plasmids and three ColE2-type plasmids from E. coli and Shigella species. The Rep stop codons are boxed and the ppApGpA primer site is indicated by a box and an arrow. The three insertions that determine compatibility are labeled a, b, and c. Rep protein Accession Nos. are: pAsal1, NP_710167.1;ColE2-P9, 808894; ColE3-CA38, 808865; and ColE5-099, 809524.
138
J. Boyd et al. / Plasmid 50 (2003) 131–144
primer for DNA synthesis while RNAI is a shorter, antisense RNA complementary to the 50 end of RNAII. RNAI is constitutively expressed, but is rapidly turned over, resulting in tight control of plasmid copy number. RNAI is also the main incompatibility determinant of ColE1-type plasmids (Tomizawa and Itoh, 1981). RNAI and RNAII molecules have both been predicted to form three stem–loop structures. RNAI forms structures I, II, and III, while RNAII forms the complementary structures of I and II, and a different larger structure, IV (Tamm and Polisky, 1983; Tomizawa, 1990). The initial interactions between RNAI and RNAII takes place at these loops. Alignment of the origin region of pAsa2 with those of other ColE1type plasmids shows a high degree of similarity (Fig. 4). Furthermore, as predicted by m-fold (Mathews et al., 1999; Zuker et al., 1999) the RNA molecules from pAsa2 also form similar stem–loop structures, (indicated by inverted arrows in Fig. 4). The stem structures are very highly conserved while the loop structures, which are responsible for the initial binding interactions and therefore incompatibility, are quite variable. Some ColE1-type plasmids encode a small protein, Rom or Rop that stabilizes the interaction between RNAI and RNAII; however pAsa2 does not appear to encode a Rom homologue. We have been able to show that pAsa2 can replicate stably in E. coli (data not shown). It is also interesting that cloning vectors based on E. coli ColE1 can replicate in A. salmonicida for a few generations only. Indeed we have successfully used pBluescript (Stratagene, La Jolla, CA) derivatives as suicide vectors in A. salmonicida. We have no information about whether the other two A. salmonicida plasmids can replicate in E. coli. 3.4. Mobilization and oriT of pAsa1, pAsa2, and pAsa3 All three A. salmonicida plasmids carry genes, mobABCD, encoding proteins similar to those of ColE1 that are involved in plasmid mobilization. MobA proteins are relaxases that nick the doublestranded plasmid DNA at a specific site, nic, in the origin of transfer (oriT). MobA becomes covalently attached to the plasmid DNA and the
MobA–DNA complex then moves through the mating bridge into the donor cell (Zechner et al., 2000). MobA and other conjugative relaxases are not highly similar (Table 3) but they do have three recognizable motifs in their N-termini. Motif I includes a conserved tyrosine residue that remains covalently attached to the DNA at the 50 end of nic, motif II includes a serine that is implicated in interacting with the 30 end of nic, motif III contains either three histidines, (HHH) or a histidine, a glutamate and an asparagine (HEN) (Varsaki et al., 2003). The MobA proteins of the A. salmonicida plasmids all have motifs I and II and the HEN sequence at motif III (data not shown). MobB, C, and D are accessory proteins that facilitate the action of MobA. As in ColE1, these genes are organized in an apparent operon with mobC upstream of mobA, while mobB and mobD are encoded on the same DNA segment as mobA, but in different reading frames (Fig. 1). The MobA, C, and D proteins of plasmid pAsa2 are very similar to those of ColE1 (Table 3), while those of pAsa1 and pAsa3 are similar to each other, but less similar to ColE1. The A. salmonicida MobB proteins are not very similar to each other or to that of ColE1. Putative oriT regions have been found upstream of the mobC genes in plasmids pAsa1 and pAsa3, and upstream of orf1 in pAsa2 (Fig. 1). These oriT regions are similar to those of ColE1-type plasmids, which reflects the similarity of the mobA genes to that of ColE1-type plasmids (Lanka and Wilkins, 1995; Zechner et al., 2000). Fig. 5 shows an alignment of the putative oriT regions from the A. salmonicida plasmids with those of ColE1-type plasmids. We have no evidence that the various MobA proteins can specifically identify their cognate oriTs, but the sequence diversity of both the oriTs and the MobA proteins suggests that this is the case. While the presence of the mob genes and oriT regions in these three plasmids suggests they are mobilizable, they are lacking genes encoding proteins required for transfer. To be mobilized, these plasmids must rely on transfer proteins encoded elsewhere. It is worth noting that the two large plasmids found in A. salmonicida A449 each carry transfer gene operons (unpublished results).
J. Boyd et al. / Plasmid 50 (2003) 131–144
139
Fig. 4. Alignment of the replication control region (oriV) of pAsa2 with that of other ColE1-type plasmids from E. coli. Stem structures are indicated by solid and dashed line arrows, loops are clear boxes. Structures I, II, and III are found in RNAI; structures I, II and IV are found in RNAII. Promoter elements are indicated by grey boxes. Open arrows indicate the site and direction of transcription and replication. Accession Nos. are: ColE1, J01566.1; ColA, 144670; and ColD, 144289.
J. Boyd et al. / Plasmid 50 (2003) 131–144
16 51 24
—
41 24
—
36
18
Fig. 5. Alignment of the origins of transfer (oriTs) of four A. salmonicida plasmids with that of ColE1 and ColA. The nick site (nic) is indicated by an arrow. Accession Nos. are: pAsal1, NP_710167.1; ColE1, J01566.1; and ColA, 144670.
35 60 —
3.5. Toxin–antitoxin systems in pAsa1 and pAsa3
35 57
Similarity was calculated using the BLOSUM 62 matrix. GenBank Accession No. of ColE1 is J01566.1.
39
38 58 —
—
—
34
—
34 70 34
27 76 41
—
23 39
36
—
39 88 40
—
—
31 57 29
—
26 44 24
—
pAsa3 pAsa2 pAsa1 pCo lE1
80 21
pAsa3 pAsa2
22 —
pAsa1 pColE1
26 22 18 54 9
pAsa3 pAsa2
15 —
pAsa1
22 44 21
pColE1
60 23
pAsa3 pAsa2
21
pAsa1
—
MobC MobB MobA
Table 3 Percent similarity (lower left) and identity (upper right) of the Mob proteins
pAsa1 pAsa2 pAsa3 pColE1
MobD
pColE1
140
Plasmids pAsa1 and pAsa3 both have proteic toxin–antitoxin genes that presumably act as plasmid stability mechanisms by post-segregational killing. These systems, also known as addiction modules, consist of two proteins, one a toxin and the other the specific antidote (Engelberg-Kulka and Glaser, 1999). The antidote is more labile than the toxin and if a cell stops making the antidote, as in the case of a plasmidless daughter cell, the toxin will be able to kill the cell. These genes are usually transcribed as an operon with the antitoxin gene upstream and overlapping the toxin gene. The systems on pAsa1 and pAsa3 also follow this general rule. Plasmids pAsa1 and pAsa3 carry genes homologous to the relBE system (Gronlund and Gerdes, 1999), in which RelB is the antitoxin and RelE is the toxin. RelE has recently been shown to cleave mRNAs bound to ribosomes in a codonspecific manner (Pedersen et al., 2003). This toxin– antitoxin system is widespread among prokaryotes and is found in Gram-negative and Gram-positive bacteria and archaea (Gerdes, 2000). Many of these RelBE systems are found on the chromosome where they act as a global regulator of translation (Christensen et al., 2001). The homologues most similar to the pAsa1 system are those of the E. coli and Salmonella enterica chromosomes, while those of pAsa3 are most similar to genes called pasAB in plasmids of Acidithiobacillus caldus (Gardner et al., 2001; Smith and Rawlings, 1997) and Pseudomonas fluorescens (Peters et al., 2001). While the RelE toxins
J. Boyd et al. / Plasmid 50 (2003) 131–144
from pAsa1 and pAsa3 do not show high identity (26%) or similarity (47%) to each other, the pasAB genes do belong to the relBE family (Gerdes, 2000), therefore an alignment of all these proteins is shown in Fig. 6. To our knowledge, the presence of two plasmidencoded, toxin–antitoxin systems in the same bacterial strain has not been detected. However, there are instances of strains carrying more than one plasmid with other types of plasmid stability systems, most notably restriction/modification (rm) systems. In this situation, two plasmids carrying rm systems will both be stably maintained as long as the target rm sites on the chromosome are different. If both methylases can protect the same site then one or the other restriction enzyme can be lost without cell death (Kusano et al., 1995). In the A. salmonicida case as long as each RelB antitoxin recognizes only its own toxin, neither plasmid can be lost without cell death. The low degree of se-
141
quence conservation among the antitoxin proteins (Fig. 6) may be indicative of their specificity for their cognate toxin. No other known form of plasmid stability system is found on these plasmids. Most notably no cer or xis sites are obvious. These are sites of sitespecific recombination that are used by the plasmid to resolve dimers created during replication. Dimer and higher multiple forms of all three plasmids are commonly seen in plasmid preparations (data not shown). 3.6. Other ORFs of unknown function There are several ORFs in these plasmids that share little or no sequence similarity with genes in the databases. The most notable is orf1, which is found on all three plasmids. orf1 shares limited identity with a putative gene of unknown function from a Ralstonia solanacearum plasmid, pJTPS1
Fig. 6. Alignment of the toxin and antitoxin proteins from plasmids pAsa1 and pAsa3 with their closest homologues. Accession Nos. are: RelE E. coli, 76201; RelE S. enterica, 16763250; RelB E. coli, 132283; RelB S. enterica, 16763249; PasB Pseudomonas fluorescens, 10834752; PasB Acidithiobacillus caldus, 14209913; PasA P. fluorescens, 10834752; and PasA A. caldus, 14209912.
142
J. Boyd et al. / Plasmid 50 (2003) 131–144
Table 4 Percent similarity (lower left) and identity (upper right) of the ORF1 proteins
pAsa1 pAsa2 pAsa3 pJTPS1
pAsa1
pAsa2
pAsa3
pJTPS1
—
36
50 86 29
—
77 33
50 24
—
20 15 21
30
—
Similarity was calculated using the BLOSUM 62 matrix. GenBank Accession No. of pJTPS1 OrfC1 (Ralstonia solanacearum) is NP_052314.
(Table 4). In addition to the sequence similarity, the orf1 genes are all positioned upstream of the mobC genes. In pAsa2, orf1 is positioned between mobC and oriT. This close proximity to the mob cluster and the fact that the mob and orf1 genes are the only genes shared among all the three plasmids suggests orf1 may be involved in mobilization of these plasmids. pAsa2 has two more ORFs of unknown function (orf2 and orf3) and pAsa3 has one (orf4). Each plasmid also has a large region (500–1000 bp) with no obvious features.
3.7. Presence of pAsa1, pAsa2, and pAsa3 in other strains of A. salmonicida Previous analyses of the plasmid profiles of A. salmonicida strains (Belland and Trust, 1989; Giles et al., 1995; Pedersen et al., 1996; Sorum et al., 2000) suggested that different strains often carried different sets of plasmids. To investigate the plasmid profiles of the strains present in our A. salmonicida strain collection, we designed primers to specific regions of the plasmids and used them to amplify total genomic DNA from the strains. For pAsa1 and pAsa3 the toxin–antitoxin regions were amplified, and for pAsa2, orf3 was chosen. We also amplified the aopP gene from plasmid pAsal1 that was sequenced by Fehr et al. (Accession No: NC_004338). Additionally, plasmid DNA was isolated from all the strains and analysed on an agarose gel following digestion with EcoRI. EcoRI cuts all plasmids only once, except pAsa3, in which case a diagnostic 4.8 kb fragment is generated. In all cases the results of the PCR and agarose gel were identical (Fig. 7). All of the
Fig. 7. Plasmids from A. salmonicida subsp. salmonicida strains. (A) Plasmids prepared from A. salmonicida were digested with EcoRI. EcoRI cuts once in pAsal1, pAsa1, pAsa2, and four times in pAsa3 creating one large 4800 bp fragment and three small ones (not visible). Plasmid names are indicated to the right and MW standards are on the left. (B) Presence (+) or absence ()) of PCR products amplified with primers specific to each of the four small plasmids using total genomic DNA as the template.
J. Boyd et al. / Plasmid 50 (2003) 131–144
strains in our collection carried pAsa1 and pAsa2. All but one carried pAsa3. Many of the strains, but not A449, also carried another larger plasmid of approximately 6.5 kb, which is similar to the size of pAsal1. These strains also gave positive results in PCR with primers to aopP, suggesting that the 6.5 kb plasmid in these strains is indeed pAsal1. As predicted by the different Rep/oriV insertions in the ColE2 plasmids, pAsa1, pAsa3, and pAsal1, all three can indeed be compatibly maintained in a single A. salmonicida strain. The virulence of the strains in our collection is quite varied, as is their plasmid profile. However, as others have shown (Bast et al., 1988; Brown et al., 1997), there does not appear to be any correlation between carriage of these small plasmids and virulence. We now know this is because, at least for the small plasmids, there are no known virulence factors except for the type III secretion system effector, AopP, encoded by pAsal1.
4. Summary The sequences of the three small plasmids from A. salmonicida subsp. salmonicida A449 demonstrate that one is a ColE1-type while the other two are ColE2-type plasmids. The genes encoded by these plasmids function primarily in replication, mobilization and plasmid stability. Differences in the Rep/oriV regions of the ColE2-type plasmids provide specificity for replication and the ability for multiple ColE2-type plasmids to be maintained in a single strain. The presence of two different plasmid addiction systems in a single bacterial strain supports the idea that these plasmids exist primarily to promote their own replication and spread.
Acknowledgments We thank the IMB DNA sequencing team for carrying out the DNA sequencing and Drs. S. Douglas and A. Patrzykat for comments on the manuscript. This work was supported by the NRC Genomics and Health Initiative. This is NRCC Publication No. 42381.
143
References Bast, L., Daly, J., DeGrandis, S.A., Stevenson, R.M.W., 1988. Evaluation of profiles of Aeromonas salmonicida as epidemiological markers of furunculosis infections in fish. J. Fish Dis. 11, 133–145. Belland, J., Trust, T.J., 1989. Aeromonas salmonicida plasmids: plasmid-directed synthesis of proteins in vitro and in Escherichia coli minicells. J. Gen. Microbiol. 135, 513–524. Bonfield, J.K., Smith, K., Staden, R., 1995. A new DNA sequence assembly program. Nucleic Acids Res. 23, 4992– 4999. Brown, R.L., Sanderson, K., Kirov, S.M., 1997. Plasmids and Aeromonas virulence. FEMS Immunol. Med. Microbiol. 17, 217–223. Chan, P.T., Ohmori, H., Tomizawa, J., Lebowitz, J., 1985. Nucleotide sequence and gene organization of ColE1 DNA. J. Biol. Chem. 260, 8925–8935. Christensen, S.K., Mikkelsen, M., Pedersen, K., Gerdes, K., 2001. RelE, a global inhibitor of translation, is activated during nutritional stress. Proc. Natl. Acad. Sci. USA 98, 14328–14333. del Solar, G., Giraldo, R., Ruiz-Echevarria, M.J., Espinosa, M., Diaz-Orejas, R., 1998. Replication and control of circular bacterial plasmids. Microbiol. Mol. Biol. Rev. 62, 434–464. Engelberg-Kulka, H., Glaser, G., 1999. Addiction modules and programmed cell death and antideath in bacterial cultures. Annu. Rev. Microbiol. 53, 43–70. Espinosa, M., Cohen, S., Couturier, M., del Solar, G., DiazOrejas, R., Giraldo, R., Janniere, L., Miller, C., Osborn, M., Thomas, C.M., 2000. Plasmid replication and copy number control. In: Thomas, C.M. (Ed.), The Horizontal Gene Pool: Bacterial Plasmids and Gene Spread. Harwood Academic Publishers, Amsterdam, pp. 1–47. Gardner, M.N., Deane, S.M., Rawlings, D.E., 2001. Isolation of a new broad-host-range IncQ-like plasmid, pTC-F14, from the acidophilic bacterium Acidithiobacillus caldus and analysis of the plasmid replicon. J. Bacteriol. 183, 3303– 3309. Gerdes, K., 2000. Toxin–antitoxin modules may regulate synthesis of macromolecules during nutritional stress. J. Bacteriol. 182, 561–572. Giles, J.S., Hariharan, H., Heaney, S.B., 1995. The plasmid profiles of fish pathogenic isolates of Aeromonas salmonicida, Vibrio anguillarum, and Vibrio ordalii from the Atlantic and Pacific coasts of Canada. Can. J. Microbiol. 41, 209–216. Gronlund, H., Gerdes, K., 1999. Toxin–antitoxin systems homologous with relBE of Escherichia coli plasmid P307 are ubiquitous in prokaryotes. J. Mol. Biol. 285, 1401–1415. Hanninen, M.L., Ridell, J., Hirvela-Koski, V., 1995. Phenotypic and molecular characteristics of Aeromonas salmonicida subsp. salmonicida isolated in southern and northern Finland. J. Appl. Bacteriol. 79, 12–21. Hiraga, S., Sugiyama, T., Itoh, T., 1994. Comparative analysis of the replicon regions of eleven ColE2-related plasmids. J. Bacteriol. 176, 7233–7243.
144
J. Boyd et al. / Plasmid 50 (2003) 131–144
Kusano, K., Naito, T., Handa, N., Kobayashi, I., 1995. Restriction–modification systems as genomic parasites in competition for specific sequences. Proc. Natl. Acad. Sci. USA 92, 11095–11099. Lanka, E., Wilkins, B.M., 1995. DNA processing reactions in bacterial conjugation. Annu. Rev. Biochem. 64, 141–169. Mathews, D.H., Sabina, J., Zuker, M., Turner, D.H., 1999. Expanded sequence dependence of thermodynamic parameters improves prediction of RNA secondary structure. J. Mol. Biol. 288, 911–940. Pedersen, K., Dalsgaard, I., Larsen, J.L., 1996. Characterization of atypical Aeromonas salmonicida isolates by ribotyping and plasmid profiling. J. Appl. Bacteriol. 80, 37–44. Pedersen, K., Zavialov, A.V., Pavlov, M.Y., Elf, J., Gerdes, K., Ehrenberg, M., 2003. The bacterial toxin RelE displays codon-specific cleavage of mRNAs in the ribosomal A site. Cell 112, 131–140. Peters, M., Jogi, E., Suitso, I., Punnisk, T., Nurk, A., 2001. Features of the replicon of plasmid pAM10.6 of Pseudomonas fluorescens. Plasmid 46, 25–36. Shinohara, M., Itoh, T., 1996. Specificity determinants in interaction of the initiator (Rep) proteins with the origins in the plasmids ColE2-P9 and ColE3-CA38 identified by chimera analysis. J. Mol. Biol. 257, 290–300. Smith, A.S., Rawlings, D.E., 1997. The poison–antidote stability system of the broad-host-range Thiobacillus ferrooxidans plasmid pTF-FC2. Mol. Microbiol. 26, 961–970. Smith, P., 1997. The epizootiology of furunculosis: the present state of our ignorance. In: Bernoth, E.-M., Ellis, A.E., Midtlyng, P.J., Olivier, G., Smith, P. (Eds.), Furunculosis Multidisciplinary Fish Disease Research. Academic Press, San Diego, USA, pp. 25–53. Sorum, H., Holstad, G., Lunder, T., Hastein, T., 2000. Grouping by plasmid profiles of atypical Aeromonas
salmonicida isolated from fish, with special reference to salmonid fish. Dis. Aquat. Organ 41, 159–171. Takechi, S., Yasueda, H., Itoh, T., 1994. Control of ColE2 plasmid replication: regulation of Rep expression by a plasmid-coded antisense RNA. Mol. Gen. Genet. 244, 49–56. Tamm, J., Polisky, B., 1983. Structural analysis of RNA molecules involved in plasmid copy number control. Nucleic Acids Res. 11, 6381–6397. Tomizawa, J., 1990. Control of ColE1 plasmid replication. Intermediates in the binding of RNA I and RNA II. J. Mol. Biol. 212, 683–694. Tomizawa, J., Itoh, T., 1981. Plasmid ColE1 incompatibility determined by interaction of RNA I with primer transcript. Proc. Natl. Acad. Sci. USA 78, 6096–6100. Umelo, E., Trust, T.J., 1998. Physical map of the chromosome of Aeromonas salmonicida and genomic comparisons between Aeromonas strains. Microbiology 144, 2141–2149. Varsaki, A., Lucas, M., Afendra, A.S., Drainas, C., De La Cruz, F., 2003. Genetic and biochemical characterization of MbeA, the relaxase involved in plasmid ColE1 conjugative mobilization. Mol. Microbiol. 48, 481–493. Zechner, E.L., de la Cruz, F., Eisenbrandt, R., Grahn, A.M., Koraimann, G., Lanka, E., Muth, G., Pansegrau, W., Thomas, C.M., Wilkins, B.M., Zatyka, M., 2000. Conjugative-DNA transfer processes. In: Thomas, C.M. (Ed.), The Horizontal Gene Pool: Bacterial Plasmids and Gene Spread. Harwood Academic Publishers, Amsterdam, pp. 87–174. Zuker, M., Mathews, D.H., Turner, D.H., 1999. Algorithms and thermodynamics for RNA secondary structure prediction: a practical guide. In: Barciszewski, J., Clark, B.F.C. (Eds.), RNA Biochemistry and Biotechnology. Kluwer Academic Publishers, Dordrecht, pp. 11–43. Communicated by I. Kobayashi
