Correspondence: Rolandas Meskys, Department of Molecular Microbiology and Biotechnology, Institute of Biochemistry, Vilnius University, Mokslininku 12, Vilnius LT-08662, Lithuania. Tel.: +370 5 2729149; fax: +370 5 2729196; e-mail: firstname.lastname@example.org
A cryptic plasmid from Arthrobacter rhombi PRH1, designated as pPRH, was sequenced and characterized. It was 5000 bp in length with a G+C content of 66 mol%. The plasmid pPRH was predicted to encode six putative open reading frames (ORFs), in which ORF2 and ORF3 formed the minimal replicon of plasmid pPRH and shared 55–61% and 60–69% homology, respectively, with the RepA and RepB proteins of reported rhodococcal plasmids. Sequence analysis revealed a typical ColE2-type ori located 45 bp upstream of the gene repA. Sequence and phylogenetic analysis led to the conclusion that pPRH is a representative of a novel group of pAL5000 subfamily of ColE2 family plasmids. Three shuttle vectors pRMU824, pRMU824Km and pRMU824Tc, encoding chloramphenicol resistance, were constructed. The latter two harboured additional antibiotic resistance genes kan and tet, respectively. All vectors successfully replicated in Escherichia coli, Arthrobacter and Rhodococcus spp. The vector pRMU824Km was employed for functional screening of 2-hydroxypyridine catabolism encoding genes from Arthrobacter sp. PY22. Sequence analysis of the cloned 6-kb DNA fragment revealed eight putative ORFs, among which hpyB gene encoded a putative monooxygenase.
Escherichia coli is the dominant screening host for functional metagenomics (Taupp et al., 2011). Although E. coli can support the expression of genes from numerous donor genomes, the main limitations in using E. coli cells for functional screens are recognition of promoters, protein maturation and cofactor requirements in heterologous genes and proteins. One of the opportunities to expand the diversity of the expression systems is to create new ones based on different microorganisms and intrinsic genetic elements such as phages, plasmids or transposons (Uchiyama & Miyazaki, 2009).
The bacteria of genus Arthrobacter are Gram-positive, nonmotile obligate aerobes that belong to class Actinobacteria (Zhi et al., 2009). Arthrobacter species are very common in soils and often constitute an important or even dominant culturable fraction of the microbial communities. A main feature of arthrobacters is their nutritional versatility coupled with the ability to grow in simple media utilizing a wide range of compounds as a source of carbon and nitrogen (Cacciari & Lippi, 1987; Jones & Keddie, 1992). Recently, these microorganisms have received considerable attention because of their potential use in detoxification of xenobiotics (Eaton, 2001; Brandsch, 2006; Shapir et al., 2007), while the genetic tools applicable for manipulation of cells of Arthrobacter spp. are not well developed. Different strains of the genus Arthrobacter harbour plasmids varying in size from 41 to 380 kb (Igloi & Brandsch, 2003; Mongodin et al., 2006; Jerke et al., 2008; Monnet et al., 2010). However, only two sequences of small plasmids from Arthrobacter species are deposited in GenBank database. The plasmid pA3 (AJ131246) is 2205 bp in length and harbours five hypothetical open reading frames (ORF). The second plasmid (pRE117-2, FQ311476) is 8528 bp in length, and 13 ORFs are predicted, two of them encode putative mobilization proteins (Monnet et al., 2010). Recently, Miteva et al. (2008) have described the cryptic plasmid p54 (1950 bp), which harbours seven ORFs, few of which sharing similarities with proteins of known function. However, the nucleotide sequence is not publicly available.
To date, a few vectors for the bacteria of Arthrobacter genus have been created. Two hybrid plasmids have been developed using the ori sequence of pCG100 from Corynebacterium glutamicum (Shaw & Hartley, 1988; Sandu et al., 2005) and pBL100 from Brevibacterium lactofermentum (Shaw and Hartley, 1988). One vector has been constructed on the basis of pULRS8 from Brevibacterium lactofermentum (Morikawa et al., 1994). The pART2 and pART3 vectors can be applied for both constitutive and nicotine-inducible gene expression as well as for promoter screening by GFP fusion (Sandu et al., 2005) or production of MalE-fused hybrid proteins (Kolkenbrock & Fetzner, 2010). All above-mentioned E. coli–Arthrobacter shuttle vectors are developed from cryptic plasmids of phylogenetically related species. Recently, the hybrid vector pSVJ21 has been constructed based on the cryptic plasmid p54 from Arthrobacter sp. (Miteva et al., 2008).
This paper reports on characterization of a small cryptic plasmid pPRH (5.0 kb) from Arthrobacter rhombi PRH1 strain and describes the pPRH-derived hybrid vectors, which replicates in both Arthrobacter and Rhodococcus species as well as in E. coli. One of the vectors has been successfully applied for functional screening of 2-hydroxypyridine catabolism encoding genes from Arthrobacter sp. PY22, using a nonconventional host.
Materials and methods
Bacterial strains and plasmids
The bacterial strains and plasmids are listed in Table 1. Arthrobacter and Rhodococcus spp. strains were cultivated at 30 °C on nutrient agar (NA) (Oxoid) plates or in nutrient broth (Oxoid) aerobically. When necessary, antibiotics were added to the media: ampicillin (50 μg mL−1), chloramphenicol (10–20 μg mL−1), kanamycin (40–60 μg mL−1 and tetracycline (10–40 μg mL−1).
1.9-kb fragment from pPRH cloned to 2.5-kb pART2 fragment, KmR
1.6-kb fragment from pPRH cloned to 2.5-kb pART2 fragment, KmR
Replication origins of pPRH and pACYC184, CmR, 3.2 kb
Replication origins of pPRH and pACYC184, LacZ, CmR, 4.0 kb
Replication origins of pPRH and pACYC184, LacZ, CmR, KmR, 4.9 kb
Replication origins of pPRH and pACYC184, LacZ, CmR, TcR, 5.3 kb
Cloning and DNA manipulations were performed as described by Maniatis et al. (1982). Plasmid DNA from Rhodococcus and Arthrobacter cells was isolated by alkaline lysis method following the incubation with lysozyme (10 mg mL−1) for 30 min. Escherichia coli and Arthrobacter (Rhodococcus) cells were prepared for electroporation by the method of Sharma & Schimke (1996) and Gartemann & Eichenlaub (2001), respectively.
Sequencing and analysis of the pPRH plasmid
A restriction analysis of the pPRH plasmid was carried out using single and double digestions. The DNA fragments were subcloned in pTZ57R. Potential protein-coding sequences were identified using blastx analysis (Altschul et al., 1997). The putative promoters were analysed using the online tools (http://www.fruitfly.org/seq_tools/promoter.html ). The predicted ORFs were further analysed by blastp and blastn. Phylogenetic trees were created using Mr. Bayes-3.1.2 (Huelsenbeck & Ronquist, 2001). Domain architectures in proteins were analysed using the online smart tool (http://smart.embl.de, Letunic et al., 2009).
To determine a minimal replicon plasmids, pAPrepAB4 and pAPrepA2 were created by replacing pCG100 origin of replication (2.1 kb BglII–SalI) in the pART2 plasmid with the appropriate DNA fragments from pPRH-containing ori sequence with repAB operon (1.9 kb BamHI–SalI) for pAPrepAB4 and ori sequence with repA gene (1.6 kb BamHI–XhoI) for pAPrepA2.
Construction of shuttle vectors
All PCRs were performed using T Personal Thermocycler (Biometra) and AccuPrime Pfx DNA polymerase (Invitrogen). The reaction mixtures (total volume 25 μL) contained 0.5 μL of template DNA (50–100 ng), 2.5 μL 10× AccuPrime Pfx reaction mix, 0.5 μL of each primer (Table 1, final concentration 2 μM) and 0.5 μL of AccuPrime Pfx DNA polymerase (1.25 units). The amplification conditions were as follows: 1 cycle of 95 °C for 5 min, 30 cycles of 95 °C for 30 s, 30 cycles of 52–62 °C for 30 s, 30 cycles of 72 °C for 1 min per kb and 1 cycle of 72 °C for 5 min.
The amplified fragments of plasmid pACYC184 (2120 bp) and plasmid pPRHHind4 (entire pPRH cloned into pTZ57R via HindIII site) (1223 bp) using DP1/RP1 and DP2/RP2 primer pairs, respectively, were ligated. The E. coli clones were selected for chloramphenicol resistance. The obtained plasmid pRMU8 and the amplified pTZ57R fragment (690 bp) using DP3 and RP3 primer pairs were double digested with BglII and XmaJI. After ligation and electroporation, the cells were spread on NA plates containing chloramphenicol, IPTG and X-Gal. Blue colonies were selected for the further work. The hybrid plasmid pRMU824 and the amplified pART2 (884 bp) or p34S-Tc (1300 bp) fragments using a pair of DP4/RP4 and DP5/RP5 primers, respectively, were hydrolysed with XmaJI. After ligation and electroporation, kanamycin- or tetracycline-resistant clones were selected. The plasmids were re-sequenced to confirm the structure and designated pRMU824Km and pRMU824Tc, respectively.
Plasmid stability test
The method described by Picardeau et al. (2000) was used to determine the segregational stability of the vectors.
Determination of the copy number of the shuttle vector pRMU824Km
Total DNA was isolated from the overnight cultures of Arthrobacter sp. 68b (negative control) and Arthrobacter sp. 68b harbouring plasmid pRMU824Km by the method described by Woo et al. (1992). DNA samples (50 μg mL−1) were diluted 100- and 1000-fold before analysis. Quantitative real-time PCR amplification was carried out using a Rotor-Gene Q 6plex instrument (Qiagen). qPCR was conducted in 0.1-mL tubes containing 15 μL of reaction mixture: 200 nM of each primer, 200 μM dNTP (Fermentas, Lithuania), 3 mM MgCl2 (Fermentas), 1.5 μM Syto9 (Invitrogen-Molecular Probes), 0.04 U μL−1 TrueStart™ Taq DNA Polymerase, TrueStart™ Taq buffer (Fermentas) and 1 μL of the DNA tested. The qPCR was initiated by 4 min of incubation at 95 °C, followed by 35 cycles of 95 °C for 20 s, 56 °C for 60 s and 72 °C for 60 s. Fluorescence data were recorded after the annealing steps. All experiments were carried out in triplicate. A genome target encoding the glycine oxidase (primers GlyOX68F and GlyOX68R) was used as a single-copy reference. The repAB genes (primers DP2 and RP2) were used as a plasmid target. The amplification efficiency for both targets was 1.12 and 1.06, respectively. The template-free negative control was used to estimate nonspecific binding. The copy number was calculated from the threshold cycle (CT). The CT values were calculated automatically according to the amplification plot (data not shown). The difference between the mean CT value of the single-copy reference and the mean CT value of the vector target was calculated.
DNA sequence data
DNA sequences have been deposited in GenBank and can be accessed via accession numbers: HQ624979 (pPRH), HQ624980 (pRMU824), HQ624981 (pRMU824Km), HQ624982 (pRMU824Tc) and FM202433 (2-hydroxypyridine catabolic genes from Arthrobacter sp. PY22).
Analysis of the pPRH plasmid from A. rhombi PRH1
Arthrobacter rhombi PRH1 was found to possess one small plasmid, designated as pPRH. The restriction and sequence analysis showed that pPRH was a circular DNA molecule, 5000 bp in length, with the G+C content of 66 mol%. It contained six putative ORFs and a putative promoter (859–899 nt) (Fig. 1a). The possible functions of the ORFs are presented in Table 2.
Table 2. Nearest orthologues and predicted functions of ORFs
A search against the GenBank protein database revealed that ORF2 and ORF3 encoded putative replication proteins RepA and RepB, respectively. The ORF2 shared 61%, 57% and 55% aa sequence similarity with the RepA protein from the Rhodococcus sp. plasmid pNC500 (Matsui et al., 2007), pREC2 (Sekine et al., 2006) and pNC903 (Matsui et al., 2006), respectively. The protein encoded by the ORF3 also shared significant homology with the Rhodococcus spp. proteins, and the similarity to the RepB of pNC903 (Matsui et al., 2006), pRC4 (Hirasawa et al., 2001), pREC2 (Sekine et al., 2006), pFAJ2600 (De Mot et al., 1997) and pKNR01 (Na et al., 2005) was 60%, 60%, 64%, 63% and 69%, respectively. Based on similarities mentioned, ORF2 and ORF3 were given functional annotation and designated as RepA and RepB, respectively. Phylogenetic analysis of RepA and RepB of pPRH showed that they formed a distinct cluster (Fig. 2a,b). Two conserved domains were detected in RepA protein. The N-terminal region (27–159 aa) was homologous to the replicase domain, which is usually found in DNA replication proteins of bacterial plasmids. The other domain (166–242 aa) shared structural features characteristic to the C terminal of primases. C-terminus of RepB (37–83 aa) was similar to a region 4 of sigma-70-like sigma factors.
The protein encoded by ORF6 was homologous to resolvases (Table 2). A resolvase from pPRH shared the same phylogenetic cluster with the homologous protein encoded by pRE117-2 from Arthrobacter arilaitensis (Fig. 2c); however, neither resolvases from Rhodococcus nor Corynebacterium spp. were related to the arthrobacterial counterparts.
A 23-nt site (1090–1067 bp) showing 60% similarity to ColE2 ori (Yagura et al., 2006) was found on the complementary DNA strand at 45 nt upstream of the repA gene (Fig. 1b).
Establishment of the minimal origin of replication
Specific combinations of genes were tested to determine the minimal region required for autonomous replication of pPRH. Plasmids pAPrepAB4 containing repAB genes and pAPrepA2 harbouring the repA gene only were constructed. pAPrepAB4 transformed Arthrobacter oxydans PY21, A. rhombi VP3, Arthrobacter sp. 68b and Rhodococcus sp. SQ1. By using a second derivative, pAPrepA2, no transformants were obtained in all Arthrobacter and Rhodococcus spp. strains tested.
Construction of shuttle vector
The Escherichia coli–Arthrobacter–Rhodococcus shuttle vector pRMU824 conferring resistance to chloramphenicol was constructed as described in ‘'Materials and methods'’ (Fig. 3). In addition, the tetracycline or kanamycin resistance gene was inserted into the plasmid pRMU824 to expand the applicability of the vector. Thus, two shuttle vectors pRMU824Km and pRMU824Tc were obtained (Fig. 3).
All shuttle vectors successfully replicated in Arthrobacter sp. 68b, 83, 85, A. oxydans PY21, Rhodococcus sp. SQ1 and E. coli. Approximately nine copies of the pRMU824Km vector per Arthrobacter sp. 68b cell were found. The analysis of plasmid loss, during cultivation in rich medium without antibiotic pressure, showed that segregational stability depended on the tested strain: 37 ± 3% of A. oxydans PY21 cells retained the plasmid after 40 generations, and under the same conditions, only 6 ± 0.7% of Arthrobacter sp. 68b cells contained the vector.
To analyse the compatibility of the developed vectors, A. oxydans PY21 harbouring the pRMU824Tc plasmid was additionally transformed with pART2gfp (Sandu et al., 2005). The clones simultaneously resistant to kanamycin and tetracycline and producing a green fluorescent protein were easily screened. Both recombinant plasmids were isolated from A. oxydans PY21 and used to transform E. coli in the presence of appropriate antibiotic. The restriction analysis of the isolated individual plasmids confirmed that the pRMU824Tc and pART2gfp plasmids were compatible with each other in the Arthrobacter spp. cells.
Application of the shuttle vector for functional screening in a nonconventional host
To test the applicability of the developed vectors for functional screening, the genes encoding the initial steps of 2-hydroxypyridine biodegradation in Arthrobacter sp. PY22 were chosen as a target. It was proposed that catabolism of 2-hydroxypyridine proceeds via formation of 2,3,6-trihydroxypyridine, which could spontaneously oxidize and dimerize to blue pigment, 4,5,4′,5′-tetrahydroxy-3,3′-diazadiphenoquinone-(2,2′) (for review, Kaiser et al., 1996). Total DNA from Arthrobacter sp. PY22, the strain degrading 2-hydroxypyridine and forming a blue pigment (Semėnaitė et al., 2003), was isolated, partially digested with KpnI and inserted into the KpnI site of pRMU824Km. Selection of clones was performed in E. coli DH5α, Arthrobacter sp. 68b and Rhodococcus sp. SQ1 bacteria, which could not utilize 2-hydroxypyridine. E. coli DH5α cells were transformed by ligation mixtures, and kanamycin-resistant clones were grown on NA plates supplemented with IPTG and 2-hydroxypyridine. As the visual inspection of plates did not reveal any coloured colonies, all clones were harvested from agar plates and pooled. The mixture of the recombinant plasmids was isolated and consequently used to transform Arthrobacter sp. 68b and Rhodococcus sp. SQ1 bacteria. A single clone (from ca 4000 clones) was selected on the NA medium supplemented with kanamycin and 2-hydroxypyridine by screening for pigment production. The pHYP1 plasmid containing a 6-kb DNA insert was isolated from the blue pigment producing clone of Rhodococcus sp. SQ1.
Sequence analysis of the cloned 6-kb DNA fragment from pHYP1 revealed eight putative ORFs (Fig. 4). Six of them shared the significant (71–87%) sequence homology with hypothetical proteins of the pSI-1 plasmid from Arthrobacter sp. AK-1 (Jerke et al., 2008). Moreover, an identical arrangement of ORFs in both plasmids was observed. The predicted functions of ORFs from pHYP1 are presented in Table 2.
New cryptic plasmid, not related to known arthrobacterial ones, was isolated from A. rhombi PRH1. A conserved sequence found at 45 bp upstream of the repA gene showed remarkable homology to the typical ColE2-type ori (Leret et al., 1998; Yagura et al., 2006). The predicted minimal replication operon of pPRH consisted of repAB genes, which is in accordance with the previous findings that both repA and repB are required for replication of pAL5000 (Stolt & Stoker, 1996a), pFAJ2600 (De Mot et al., 1997), pBLA8 (Leret et al., 1998) and pCASE1 (Tsuchida et al., 2009). All these results supported the conclusion that pPRH is a member of the pAL5000 subfamily of ColE2 family (Stolt & Stoker, 1996a), plasmids belonging to the theta replication C class (Bruand et al., 1993). Usually, repB is located downstream and overlaps with repA of these plasmids, suggesting that both of these genes form an operon. Correspondingly, the start codon of repB in pPRH overlaps with the stop codon of the repA by one nucleotide. A putative rep operon in pPRH may also include ORF4, which overlaps with repB and encodes a hypothetical protein. The function of ORF4 in the plasmid replication and/or maintenance remains unclear.
Phylogenetic analysis of RepA and RepB of pPRH showed that they formed a distinct branch on phylogenetic tree suggesting their evident divergence from homologous proteins (Fig. 2a,b). Two putative conserved domains related replicase and primase, respectively, were detected in RepA from pPRH, which is a common structural feature of other RepA proteins associated with theta replication (De Mot et al., 1997; Leret et al., 1998; Sekine et al., 2006; Matsui et al., 2007; Tsuchida et al., 2009). The analysis of RepB from pPRH revealed one conserved domain homologous to region 4 of sigma-70-like sigma factors, which is involved in binding of the −35 promoter element (Campbell et al., 2002). The RepB protein of pAL5000 was shown to bind to DNA near the ori site (Stolt & Stoker, 1996b). It could be proposed that the RepB encoded by pPRH has the same function.
According to the sequence analysis, ORF6 belongs to serine recombinase family, which includes resolvases, invertases, integrases and transposases (Smith & Thorpe, 2002), and might contribute to plasmid maintenance (Nordstrom & Austin, 1989). A putative resolvase of plasmid pPRH is phylogenetically most related to the enzyme from A. arilaitensis sharing the distinct branch (Fig. 2c). This demonstrates that, in contrast to both Rep proteins, the resolvase displays the independent patterns of evolution.
Escherichia coli–Arthrobacter–Rhodococcus shuttle vectors were built using the bottom-up approach, starting with the minimal requirement for the arthrobacterial replicon taken from the cryptic plasmid pPRH. The multiple cloning site of the lacZ′ cassette (Fig. 3) allowed using a common beta-galactosidase-based screening strategy in E. coli. The developed shuttle vectors were compatible with the pART vectors (Sandu et al., 2005). Hence, these plasmids might be used as original tools in genetic complementation studies as well as for a functional complementation-based screening in both Arthrobacter and Rhodococcus species.
The successful cloning of the genes encoding the initial steps of 2-hydroxypyridine biodegradation in Arthrobacter sp. PY22 showed a potential of the developed vectors for functional screening in the nonconventional host. The cloned genes or encoded proteins were inactive in E. coli cells; hence, screening based on enzyme activities was impossible in this host. However, the pHYP1 plasmid containing genes encoding 2-hydroxypyridine catabolism could be selected using Rhodococcus or Arthrobacter as a host. It is supposed that 2-hydroxypyridine biodegradation in Arthrobacter sp. PY22 bacteria proceeds via classical pathway by formation of 2,5-dihydroxypyridine and 2,3,6-trihydroxypyridine as intermediates (Semėnaitė et al., 2003). Implying that, the appropriate hydroxylases are expected. A sequence analysis of the cloned DNA fragment showed that hpyB gene encodes a putative flavin monooxygenase belonging to the family of flavin mononucleotide (FMN)-dependent bacterial luciferases and alkanesulphonate monooxygenases, enzymes that employ reduced flavin and usually act as two-component monooxygenases in concert with NAD(P)H-dependent FMN reductases (Ellis, 2010). The hpyD gene encoding a putative NAD(P)H-dependent FMN reductase is located in close proximity to the hpyB gene. Hence, a two-component flavin monooxygenase involved in the hydroxylation of 2-hydroxypyridine ring might be expected. While various flavin-dependent monooxygenases participate in catabolic pathways of xenobiotics or biosynthesis of secondary metabolites (van Berkel et al., 2006; Ellis, 2010), there are no reports on two-component monooxygenases involved in the biodegradation of N-heterocyclic compounds except for pyrrole-2-carboxylate monooxygenase (Hormann & Andreesen, 1994) or 2-methyl-3-hydroxypyridine-5-carboxylic acid oxygenase and 5-pyridoxic acid oxygenase, both catalysing a ring-cleavage reaction (Chaiyen, 2010). Clearly, additional studies are needed to show the gene functions at the protein level; however, the first genetic data related to catabolism of 2-hydroxypyridine shed some light on the putative enzymes involved in this pathway.
The authors thank Dr Laura Kaliniene for critical reading of the manuscript. This research was funded by a grant (No. MIP-076/2011) from the Research Council of Lithuania.