Introduction

Species of the Gram-positive genus Rhodococcus exhibit not only a wide range of catabolic activities (Hughes et al. 1998), but also some cell properties required for the technological application of microbial degraders (Lang and Philip 1998). Bacteria often suffer from growth inhibition at higher concentrations of various aromatic compounds during the course of their biodegradation (Prieto et al. 2002). To overcome such inhibition, cell immobilization, which enhances the ability of the cell to tolerate the cytotoxic effects of pollutants (Lazarova and Manem 1995), or genetically engineered microorganisms have been applied (Soda et al. 1998). Comparisons of biodegradation of aromatic compounds by suspended and attached bacteria proved that bacteria forming a biofilm were protected against cytotoxic effects and their ability to degrade phenol was enhanced (Prieto et al. 2002). In a preliminary cell attachment study, the strain Rhodococcus erythropolis CCM2595 was shown to form a phenol-resistant biofilm on inorganic surfaces and to degrade phenol efficiently. This strain was therefore chosen as a host for the development of a gene transfer system that would enable us to combine physiological and genetic improvements of this phenol degrader.

Genes coding for enzymes involved in degradation of various aromatic compounds have been found to be located on bacterial chromosomes or on plasmids (van der Meer et al. 1992). A number of large plasmids, of which only some were suitable for the construction of cloning vectors, has been isolated from Rhodococcus strains (Desomer et al. 1990; Denis-Larose et al. 1998). A few small plasmids, e.g., pFAJ2600 from R. erythropolis (De Mot et al. 1997), pNC903 from R. rhodochrous (Kalscheuer et al. 1999), pRC3 from R. rhodochrous (Hashimoto et al. 1992), and pMVS300 from Rhodococcus sp. (Singer and Finnerty 1988) have been used for the construction of shuttle vectors replicating in Escherichia coli and in various Rhodococcus strains. However, there is a general lack of information on the host range, classification into incompatibility groups, mechanism of replication and copy number control in the plasmids indigenous to Rhodococcus strains. As a basis for vectors replicating in Rhodococcus strains, some broad-host-range plasmids have also been used. The broad-host-range plasmid pNG2 from Corynebacterium diphtheriae also replicates in C. glutamicum and in E. coli (Serwold-Davis et al. 1987). Lessard et al. (1999) used the pNG2-based vector for transformation of Rhodococcus strains. The plasmid derived from the pNG2-related replicon pSR1 (Archer and Sinskey 1993) replicated in Rhodococcus sp. R312 (formerly Brevibacterium sp. R312) (Chion et al. 1991) and pJC1, also based on pSR1, replicated in Rhodococcus opacus (Kalscheuer et al. 1999). The plasmids pNG2, pSR1 and pGA1 (Nešvera et al. 1997), indigenous to corynebacteria, are members of a new family of plasmids replicating by a rolling-circle mechanism (Osborn et al. 2000). These plasmids might also be utilized for the construction of vectors replicating in other Rhodococcus strains. The present paper describes optimization of an electrotransformation procedure of the studied strain with the plasmid vector pFAJ2574, construction of vectors based on C. glutamicum plasmids pSR1 and pGA1, and use of the constructed vector for expression of the gene coding for green fluorescent protein (gfpuv), which represents a new useful marker for R. erythropolis.

Materials and methods

Bacterial strains, plasmids and growth conditions

The strains and plasmids used are listed in Table 1. E. coli was grown in LB medium at 37°C. R. erythropolis CCM2529 was grown in 2×YT medium (Sambrook et al. 1989) with 0.5% glucose and 1.5% Tween 80 (used against cell clumping) at 30°C. Selection media contained kanamycin (Km, 30 μg/ml) or chloramphenicol (Cm, 50 μg/ml) for E. coli and Km (200 μg/ml) or Cm (50 μg/ml) for R. erythropolis.

Table 1. Strains and plasmids. Km Kanamycin, Cm chloramphenicol
Full size table

DNA isolation and manipulation

Isolation of plasmid DNA from E. coli and DNA manipulations were carried out using standard techniques (Sambrook et al. 1989). Plasmid DNA from R. erythropolis was isolated by the same method as from E. coli with some modifications. Before lysis, cells were incubated for 2 h with lysozyme (5 mg/ml). For the lysis, a higher concentration of SDS (5%) was used. For PCR amplification of the promoter P-45, primers P45H (GCGGATCCTCATTCGCCGTGGC) and CM4 (GAAAATCTCGTCGAAGCTCG) and the plasmid pEKplCmP45 as a template (Pátek et al. 1996) were used. Primers KmTF1 (TCGGTACCTCATTCGCCGTGGCA) and KmTF2 (GCGGTACCTCTTGTTCAATCATGCGAAAC) and the plasmid pK45 as a template were used for PCR amplification of the regulatory sequences and the initial six codons of the Km resistance gene from pK45.

Electrotransformation

The optimized electrotransformation procedure was as follows: an overnight culture of R. erythropolis was used to inoculate 100 ml 2×YT medium in a 500-ml flask to OD600 =0.1. The medium was supplemented with 1.5% Tween 80 and 0.5% glucose. Cells were cultivated at 30°C with vigorous shaking (200 rpm). At OD600 =0.6 the cells were harvested by centrifugation, washed three times with 50 ml 10% glycerol and resuspended in 2 ml 10% glycerol. For electroporation, 200 μl cell suspension was mixed with 0.5 μg plasmid DNA, placed into an 0.2-cm cuvette and subjected to an electric pulse using a Gene Pulser (Bio-Rad, Richmond, Calif.) with the settings: 12.5 kV cm−1, 600 Ω and 25 μF. Time constants of 14–16 ms were observed. Medium (1 ml 2×YT with 0.5% glucose but without Tween 80) was added immediately and the cells were cultivated for 4 h at 30°C. They were plated on 2×YT with Km (200 μg/ml) or Cm (50 μg/ml). Colonies appeared after 3–4 days and transformants were verified by detection of the relevant plasmid.

Results

Optimization of R. erythropolis electrotransformation

For initial transformation of R. erythropolis CCM2595, plasmids pFAJ2574 and pNC9503, previously described to replicate in several Rhodococcus strains (De Mot et al. 1997) and in R. opacus (Kalscheuer et al. 1999), respectively, were used. Initially, the electrotransformation procedure for C. glutamicum (Liebl et al. 1989) was applied. In both cases transformants of R. erythropolis CCM2595 were obtained. To optimize the electrotransformation technique in the strain studied, the plasmid pFAJ2574 (10.4 kb, CmR) was utilized. Different growth phase, DNA concentration, field strength, external resistance and cultivation time before plating were tested. Effects of the experimental variables on the transformation efficiency are shown in Fig. 1. The transformation efficiency increased with the increasing final concentration of the cells in the electroporation cuvette. However, when the cells were concentrated to OD600 =100, the probability of electric arc increased. Cells were concentrated to OD600 =30 in all cases shown in Fig. 1A. The highest efficiency was achieved using cells from the early growth stage (OD600 =0.3–0.7). At late stationary phase (OD600 >15,) no transformants were obtained. The optimum DNA concentration was found to be 0.25–1.0 μg/ml (Fig. 1B). Highest transformation efficiency was observed at a field strength of 10–12.5 kV/cm and external resistance of 600 Ω. At these settings, time constants of 14–16 ms were reached, and a 4-h cultivation before plating was found to be most favorable for the transformation yield. The optimized procedure was used for electrotransformation of R. erythropolis with plasmids derived from Corynebacterium replicons.

Fig. 1.
figure 1

Dependence of electrotransformation efficiency on A optical density of the cell culture, B DNA concentration, C field strength and D external resistance. The data represents the average of three independent experiments

Full size image

Transformation of R. erythropolis with plasmid vectors based on Corynebacterium replicons

The vectors based on the C. glutamicum plasmid pSR1 and related C. diphtheriae plasmid pNG2 have been reported to replicate in some Rhodococcus strains (Chion et al. 1991; Lessard et al. 1999). We therefore tested the constructs pSRK1 and pKG43 (based on the C. glutamicum indigenous cryptic plasmids pSR1 and pGA1, respectively) using the optimized electrotransformation procedure. Plasmid pSRK1 was constructed by cloning pSR1 linearized with BclI into the BamHI site of pK19. Plasmid pKG43 (Nešvera et al. 1997) was created by cloning the large PstI fragment (4.3 kb) of pGA1 into pK19. In addition to these constructs, E. coli-C. glutamicum shuttle vectors pET2 and pEC6, derived from the C. glutamicum ssp. lactofermentum replicon pBL1, were used for the electrotransformation of R. erythropolis CCM2595. Transformants harboring pSRK1 and pKG43 were obtained, while no transformants were observed with either pET2 or pEC6. The plasmid pSR1 was used for the construction of further vectors. First, the strong promoter P-45 (originally randomly cloned from the C. glutamicum chromosome) (Pátek et al. 1996) on a 120-bp PCR-amplified BamHI fragment was inserted into the BclI site of pK19 upstream of the KmR gene, giving rise to pK45. This step allowed pSR1 linearized with BclI to be cloned into the BglII site of pK45. The resulting vector pSRK21 (5.8 kb, KmR; Fig. 2) provided the multiple cloning site of pK19 with six unique sites (PstI, SalI, XbaI, BamHI, SmaI, KpnI), α-complementation of β-galactosidase for cloning in E. coli, and a strong selectable marker. After transfer into R. erythropolis, the plasmid was found to be structurally stable in the transformants. The minimal inhibitory concentration of Km in this host was higher than 3,000 μg/ml. Segregational stability of the plasmids in R. erythropolis was determined. After 30 generations of cultivation under non-selective conditions, 100% R. erythropolis cells harbored pFAJ2574 and 90% cells harbored pSRK21. The plasmid pKG43 was maintained in R. erythropolis cells during growth in the presence of Km, while it was unstable under non-selective conditions and was lost from the cells after 20 generations.

Fig. 2.
figure 2

Restriction and genetic maps of the Escherichia coli-Rhodococcus erythropolis shuttle vectors pSRK21 and pSRKgfp. The part originating from pSR1 is shown as a thick black line. The empty box at the 5′-end of the genes depicts a segment containing the regulatory sequences (promoter and Shine-Dalgarno site) and initial six codons of the kanamycin resistance (Km R) gene used in translational fusion with gfpuv for efficient expression of the green fluorescent marker. The per gene of pSR1 is homologous to the per gene from pGA1, which positively influences the plasmid copy number and segregational stability (Nešvera et al. 1997)

Full size image

Expression of the gfpuv gene using vector pSRK21

The gfpuv gene, coding for green fluorescent protein, was used to demonstrate the usefulness of the vector pSRK21. The KpnI-EcoRI fragment (730 bp) carrying the gfpuv coding sequence from the plasmid pGFPuv was cloned into pSRK21. In the resulting construct, pSRK211, lacZ-gfpuv fusion ensured a strong fluorescence of colonies of the E. coli host on plates illuminated with UV radiation (360 nm). However, no fluorescence of R. erythropolis transformants was detected. Therefore, the fragment containing the strong promoter P-45, a Shine-Dalgarno sequence and the initial six codons of the Km resistance gene was inserted upstream of the gfpuv gene in pSRK211, resulting in another translational fusion with gfpuv. The resulting plasmid, pSRKgfp (Fig. 2), conferred a reliable green fluorescent marker to R. erythropolis cells.

Discussion

The strain R. erythropolis CCM2595, which has convenient properties for application as a phenol-degrading organism, was chosen for genetic studies. To facilitate genetic analysis and manipulation of this strain, we have optimized conditions for genetic transformation and constructed plasmid vectors for gene transfer.

Starting with the conditions for electrotransformation of C. glutamicum, a gene transfer system for R. erythropolis was optimized. It was necessary to add Tween 80 to the cultivation medium to prevent clumping of the cells. Moreover, it has been shown that addition of Tween 80 to the cultivation medium increases transformation efficiency in some corynebacteria (Haynes and Britz 1989) and rhodococci (Sekizaki et al. 1998). The highest transformation efficiency in R. erythropolis CCM2595 reached (7×104 transformants/μg DNA) is comparable to that reported in R. opacus (Kalscheuer et al. 1999) and R. fascians (Desomer et al. 1990) and will allow further genetic tools for gene manipulations in the strain studied to be developed.

R. erythropolis falls into a distinct group of rhodococci with R. globerulus, R. opacus and R. fascians (Goodfellow et al. 1998). Construction of a few vectors replicating in some of these species has already been reported (Desomer et al. 1990; De Mot et al. 1997; Kalscheuer et al. 1999). However, knowledge of the basic functions of the Rhodococcus replicons used is scarce. According to the homologies of 16S rRNA sequences, R. erythropolis is related to other mycolic acid-containing genera like Corynebacterium. We therefore tested several Corynebacterium plasmid replicons as the basis for plasmid vectors for gene transfer to R. erythropolis CCM2595. The vectors based on the plasmids pSR1 and pGA1, classified into a new fifth group of rolling-circle-replicating plasmids (Osborn et al. 2000), were found to replicate in R. erythropolis, while no transformants were obtained with the pBL1 replicon, belonging to the pC194 family of rolling-circle-replicating plasmids. Plasmid replication and regulation of copy number have been studied in the plasmids pSR1 and pGA1 from C. glutamicum (Archer and Sinskey 1993; Nešvera et al. 1997). The product of the per gene of the plasmid pGA1 was found to positively influence the copy number of this plasmid and, consequently, its segregational stability (Nešvera et al. 1997). The N-terminal part of the pSR1-encoded ORF1 gene product (Archer and Sinskey 1993), whose amino acid sequence is identical with nearly the whole sequence of the pGA1-encoded Per protein, appears to fulfill the same function (Nešvera et al. 1997). We have recently described negative control of pGA1 copy number by a small antisense RNA complementary to the rep gene leader region (Venkova-Canova et al. 2003). On the basis of sequence homology, we suppose that a similar mechanism functions also in pSR1. Since stable maintenance of small rolling-circle-replicating plasmids correlates with their copy number, vector molecules exhibiting higher segregational stability could be constructed by introducing sequence alterations alleviating the negative control of copy number of plasmids pSR1 and pGA1. These plasmids thus seem to represent a suitable basis for development of further tools for genetic analysis of R. erythropolis and for practical use in cloning the genes involved in degradation pathways. The gfpuv gene was efficiently expressed in R. erythropolis when cloned into the constructed vector pSRK21 using a suitable transcriptional and translational system. The gfpuv gene will be tested as a marker in a single copy after integration into the R. erythropolis chromosome and for construction of a promoter-probe vector for rhodococci.