Dam and Dcm methylations prevent gene transfer into Clostridium pasteurianum NRRL B-598: development of methods for electrotransformation, conjugation, and sonoporation

Kolek et al. Biotechnol Biofuels (2016) 9:14 DOI 10.1186/s13068-016-0436-y RESEARCH Dam andDcm methylations prevent gene transfer intoClostridium pasteurianum NRRL B-598: development ofmethods forelectrotransformation, conjugation, andsonoporation Jan Kolek1*, Karel Sedlar2, Ivo Provaznik2 and Petra Patakova1 Abstract Background: Butanol is currently one of the most discussed biofuels. Its use provides many benefits in comparison to bio-ethanol, but the price of its fermentative production is still high. Genetic improvements could help solve many problems associated with butanol production during ABE fermentation, such as its toxicity, low concentration achievable in the cultivation medium, the need for a relatively expensive substrate, and many more. Clostridium pasteurianum NRRL B-598 is non-type strain producing butanol, acetone, and a negligible amount of ethanol. Its main benefits are high oxygen tolerance, utilization of a wide range of carbon and nitrogen sources, and the availability of its whole genome sequence. However, there is no established method for the transfer of foreign DNA into this strain; this is the next step necessary for progress in its use for butanol production. Results: We have described functional protocols for conjugation and transformation of the bio-butanol producer C. pasteurianum NRRL B-598 by foreign plasmid DNA. We show that the use of unmethylated plasmid DNA is necessary for efficient transformation or successful conjugation. Genes encoding DNA methylation and those for restrictionmodification systems and antibiotic resistance were searched for in the whole genome sequence and their homologies with other clostridial bacteria were determined. Furthermore, activity of described novel type I restriction system was proved experimentally. The described electrotransformation protocol achieved an efficiency 1.2 × 102 cfu/ μg DNA after step-by-step optimization and an efficiency of 1.6 × 102 cfu/μg DNA was achieved by the sonoporation technique using a standard laboratory ultrasound bath. The highest transformation efficiency was achieved using a combination of these approaches; sono/electroporation led to an increase in transformation efficiency, to 5.3 × 102 cfu/μg DNA. Conclusions: Both Dam and Dcm methylations are detrimental for transformation of C. pasteurianum NRRL B-598. Methods for conjugation, electroporation, sonoporation, and a combined method for sono/electroporation were established for this strain. The methods described could be used for genetic improvement of this strain, which is suitable for bio-butanol production. Keywords: Clostridium, Butanol, Transformation, Conjugation, Methylation, Dam, Dcm, Electroporation, Sonoporation © 2016 Kolek et al. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/ publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Open Access Biotechnology for Biofuels *Correspondence: [email protected] 1 Department of Biotechnology, University of Chemistry and Technology Prague, Technicka 5, 166 28 Prague, Czech Republic Full list of author information is available at the end of the article Page 2 of 14 Kolek et al. Biotechnol Biofuels (2016) 9:14 Background Interest in biofuel production, which could represent a useful substitute for standard fuels derived from fossil resources, has increased significantly over the last decade [1]. Butanol formed during acetone-butanolethanol (ABE) fermentation by solventogenic clostridia represents an interesting option for biofuel production, especially taking into account its physico–chemical properties that better suit requirements of gasoline motors compared to ethanol. Although butanol production by ABE has been known for more than 100years [2], its industrial-scale production is hampered by a low final concentration, lower yield compared to ethanol, and in most species, an association of butanol production with sporulation. In addition, clostridia, including solventogenic species, are a polyphyletic group of bacteria, where transfer of knowledge gathered for one species, or even strain to another species/strain is difficult if not impossible. Most knowledge regarding the ABE process has been obtained from a single strain, Clostridium acetobutylicum ATCC 824, which differs in many features from other solventogenic clostridia [3]. Most other species, with the exception of C. beijerinckii NCIMB 8052 [4], have been described relatively poorly. These drawbacks have precluded the biotechnological production of bio-butanol on a larger scale [5]. Genetics and metabolic engineering represent new approaches with the possibility of significantly improving the ABE process. The existence of methods for genetic manipulation of industrial microorganisms is generally essential for improving their properties to be appropriate for biofuel production. However, these methods are also very important for better, quicker and more effective research that could lead to the acquisition of important information useful in industrial processes. The most commonly used method for introducing foreign DNA into bacterial cells is transformation (an exogenous molecule of DNA is introduced directly through the cell membrane), conjugation (mediated by tight contact between donorrecipient cells and pili formation), and transduction (mediated by virus particles). In most cases, transformation of Gram-positive bacteria is more difficult compared to Gram-negatives and the development of transformation protocols is demanding. Gram-positive bacteria possess a thick peptidoglycan layer that is further enveloped by a protein S-layer and these bacteria also have only one cytoplasmic membrane, whose distortion can lead to immediate disruption of cell homeostasis and often death. Transformation of gram-positive, strictly anaerobic bacteria of the genus Clostridium, is also usually accompanied by many drawbacks. For the introduction of foreign DNA into clostridial cells, several protocols have been described, based on conjugation with Escherichia coli [6, 7] or Enterococcus [8] donors, PEG-induced protoplast transformation [9, 10] and more recently, electroporation [11–14]. In addition, some less frequently used transformation approaches such as chemical treatment by Tris-PEG method [15] or sonoporation [16] have been tested. Here, we describe the development of methods for genetic modification of C. pasteurianum NRRL B-598—a solventogenic bacterium producing butanol, acetone, and ethanol [17]. This strain is unique in its exceptional oxygen resistance, which is much higher than the standard butanol-producing model strains such as C. pasteurianum ATCC 6013, C. beijerinckii NCIMB 8052 or C. acetobutylicum ATCC 824. Also the whole genomic sequence is available for this strain [18, 19]. Moreover, only one system for genetic manipulation of C. pasteurianum species (type strain ATCC 6013) has been published [12]. We found that the development of methods for introducing DNA into the non-type, and at first sight untransformable, strain C. pasteurianum NRRL B-598, was problematic and completely different from other clostridia. We believe that our contribution to this field will strengthen knowledge on bacterial (especially Clostridium) transformation methods and encourage those who tackle similar tasks, trying to apply protocols developed for different species/strains, to their particular microorganisms. Results Initial transformation attempts Initially, we conducted a series of pilot experiments based on previous descriptions of the transfer of foreign DNA to other clostridial species, as described in the literature [6, 8, 20, 21]. First, we tested various conditions for plasmid transfer by conjugation using various growth media (TYA, RCM, CBM, P2, YTG), time of conjugation (5–24h), donor:recipient ratios (from 1:10 to 10:1) and, when no transformants resulted, electroporation was tested using various growth states of cells (OD 0.4–1.2), electroporation buffers (SMP, PEG, glycerol), cuvettes (0.2 and 0.4 cm gap), and electrical parameters (field strength 2.5–15kVcm−1, time constant 5–20ms). We also used plasmids from the pMTL80000 series encoding different replicons and antibiotic resistance markers [21]; this was to minimize the possibility that the plasmids may encode unsuitable origins of replication or antibiotic resistance for our strain. Unfortunately, no conditions that we tested during these pilot experiments led to successful transformation. During pilot experiments, we discovered that strain C. pasteurianum NRRL B-598 was naturally resistant to chloramphenicol and thiamphenicol, therefore plasmids Page 3 of 14 Kolek et al. Biotechnol Biofuels (2016) 9:14 encoding thiamphenicol resistance, classically used as a selection marker for most clostridial strains, were not applicable. On the other hand, such a marker could be used for counter-selection during conjugation. We also verified that C. pasteurianum NRRL B-598 was not resistant to erythromycin or spectinomycin (20μg/μl, 700μg/ μl resp.) at concentrations previously reported in the literature [21], but when a lower concentration of antibiotic was used, or too many cells were seeded onto agar plates, a very strong background growth was observed. Similarly, almost normal growth of cells was observed after longer periods (2–3days) in TYA broth supplemented with appropriate concentrations of antibiotics. Bioinformatics analysis ofthe C. pasteurianum NRRL B‑598 genome Because all attempts at plasmid transformation of our strain failed, we decided to perform a more detailed bioinformatics analysis. The main purpose was to reveal genes encoding putative restriction-modification (R-M) systems that could present a problem during transformation of clostridia, and genes encoding putative DNA methyltransferases that could be connected with these R-M systems for protection of their own DNA [8, 12, 22, 23]. We took advantage of SMRT sequencing data used for the genome assembly [19] to study DNA methylation on a genome-wide scale. We analyzed all base modifications to determine modified sequence motifs. Out of the total, 2033 positions in the C. pasteurianum NRRL B-598 genome were detected as being methylated (m4C or m6A) with the majority being m6A methylations (1996 positions). Both detected motifs (GAAYNNNNNNNRTANYC, GAYNNNNNNCTAG) demonstrated novel recognition sequences that have not been described previously. Letters in bold denote methylated bases. Highlighted ‘T’ represents methylation of ‘A’ in the opposite strand. The data were deposited in the REBASE PacBio database (http://rebase.neb.com/cgi-bin/pacbiolist) [24] and were connected to the R-M system based on homology searching. The detected methylation motifs, both m6A types, are summarized in Table1, along with the corresponding methyl transferase (MT)-encoding genes. In addition to above-mentioned type I R-M systems, three more putative R-M systems were predicted, including two type II R-M systems and a single type IV R-M system. A summary of all five systems is found in Table2. BLAST results also showed that no genes homologous to E. coli Dam and Dcm were present in the C. pasteurianum NRRL B-598 genome. We also searched for antibiotic resistance genes. In total, 28 ORFs with antibiotic resistance functions, divided into nine resistance classes, were identified in the genome. All of these ORFs were assigned GenBank accession numbers for the relevant protein product (Fig.1). As expected, we verified the presence of a gene for chloramphenicol acetyltransferase (cat, [GenBank: ALB45592]) that encoded resistance to chloramphenicol and thiamphenicol, as observed during our experiments. Moreover, genes encoding erythromycin or spectinomycin resistance were not identified. A substantial part of the antibiotic resistance of C. pasteurianum NRRL B-598 is mediated by an antibiotic efflux system. Investigation ofpotential restriction barriers As described previously, nucleases can be located on the surface of cells and in some cases and degradation of DNA can already start after adding DNA to the cells [25]. In other cases, enzymes with nuclease activity are located in the cytoplasm. Hence we examined nuclease activities in both the protoplast crude lysate (without any parts of the cell envelope) as well as in the whole cell extract. We did not detect any restriction activity when pMTL83253 (plasmid does not contain motifs of predicted type I R-M systems) was incubated with crude extracts and whole cell lysate. In the case of pMTL82254 (contains one of each predicted motifs), plasmid DNA was nearly completely digested in broad spectrum of cultivation conditions. Restriction did not provide separate bands (DNA fragments) like in case of cultivation with crude extract from C. pasteurianum DSM 525, but led to one fuzzy smear (see Fig.2). The same restriction pattern was obtained at 30 and 37°C. Influence ofmethylation andestablishment ofan electroporation protocol As a next step, we wanted to test whether plasmid DNA without Dam and Dcm methylation could be used for transformation. We extracted plasmids from E. coli Table 1 Methylated motif detected forC. pasteurianum NRRL B-598 R‑M system type Motifs (±strand) No. in genome No. detected (±strand) % detected (±strand) Locus tag Nomenclature I GRNTAYNNNNNNNRTTC/ GAAYNNNNNNNRTANYC 406 385/380 94.83/93.60 X276_10630 M.Cpa598I I CTAGNNNNNNRTC/ GAYNNNNNNCTAG 606 573/560 94.55/92.41 X276_12360 M.Cpa598II Page 4 of 14 Kolek et al. Biotechnol Biofuels (2016) 9:14 JM110 (dam−/dcm−), a strain used for preparation of unmethylated DNA. After pilot electrotransformation experiments using unmethylated pMTL83253 (containing the pCB102 origin derived from C. butyricum) and conditions described previously for C. beijerinckii [25], a few erythromycin-resistant colonies (1–12CFU) were obtained after 48 h of growth on selective agar medium. Also other tested plasmids (pMTL83353pCB102 replicon and spectinomycin selection marker; pMTL82251-pBP1 replicon; pMTL84251-pCD6 replicon; pMTL85251-pIM13 replicon) were transformed successfully but the CFU yields were much lower (a maximum of 4CFU). Because of the best transformation efficiency achieved, as well as the fact that the pCB102 origin is the replicon that is used, for example, in standard pMTL007 plasmids (ClosTron system) [7] used for fast and specific knock-outs, we performed all following experiments with pMTL83253. The presence of pMTL83253 in erythromycin-resistant colonies was verified by its isolation and restriction digestion by PstI. Bands of the digested DNA were compared to bands of pMTL83253 isolated from E. coli and digested in the same way (Fig.3). The presence of pMTL83253 was confirmed in all erythromycin-resistant colonies that we tested. After achieving successful transformation, we aimed to improve transformation efficiency for unmethylated plasmid DNA because the twelve colonies observed (observed maximum) corresponded to a transformation efficiency of only 6cfu/μg DNA, which is very low and would not be compatible with the use of this method for genetic manipulations. Initially, we tested different voltages (2500– 15,000Vcm−1). A second parameter, investigated and optimized during the first experiments, was the growth state of the cells, represented by culture optical density. For this purpose, we prepared electrocompetent cells from cultures of different OD600 (0.6–0.8 and 1.2–1.4), representing the previously used states of culture for electrotransformation of clostridia. When cells at an OD600 of around 1.2–1.4 were used, transformation efficiency was significantly improved (Fig. 4). In the following electroporation experiments, time constant, as the main parameter of electroporation, was investigated using the best voltage and cell growth conditions (see Table 2 R-M systems inC. pasteurianum NRRL B-598 genome a R restriction endonuclease, M restriction endonuclease coupled methylation protein, S R-M specific protein b Predicted recognition site Type Name GeneaMeth. type Recognition Locus (X276_) Most similar (% identity) I Cpa598IP R m6A GAAYNNNNNNNRTANYC 10620 CspMORF4102P (95 %) M.Cpa598I M 10630 M.CbeG117ORFCP (97 %) S.Cpa598I S 10635 S.CspMORF4102P (49 %) I Cpa598IIP R m6A GAYNNNNNNCTAG 12355 Csc25775ORFJP (90 %) M.Cpa598II M 12360 M.Csc25775ORFJP (96 %) S.Cpa598II S 12365 S.Bme201ORFGP (56 %) II M1.Cpa598ORF20205 M m5C – – M1.CboKAPB3ORF12160P (87 %) M2.Cpa598ORF20205 M 01545 M2.CboKAPB3ORF12160P (84 %) R1.Cpa598ORF20205 R – R2.Cce743ORF4007P (46 %) R2.Cpa598ORF20205 R 01555 R1.Bce3081ORF2217P (31 %) II M.Cpa598ORF2410P M m6A GATCb20735 M.Cbe59BORF1284P (100 %) IV Cpa598ORF12465P R – – 12465 Cdi15410ORFAP (93 %) Fig. 1 Antibiotic resistance genes in C. pasteurianum NRRL B-598 genome. Overall resistance in the center, resistance classes in the middle, and individual resistance genes (and their NCBI accession numbers) on the outer ring Page 5 of 14 Kolek et al. Biotechnol Biofuels (2016) 9:14 above). We observed that shorter electric pulses (5ms) were significantly better for transformation efficiency compared to higher values. CFUs obtained using different time constants are shown in Fig.4. Square-wave pulse delivery was also tested, but transformation efficiencies were significantly lower than with exponential pulse mode (see Fig.4). We also tested a set of various electroporation buffers (30% PEG 8000 and SMP buffer at different pH values). However, no increase in transformation efficiency was obtained in any other buffers during these experiments. The addition of cell-wall weakening additives (different concentrations of glycine, ampicillin or Tween 80) or treatments with various concentrations of lysozyme prior to electroporation, which have been described previously [12, 26] as methods for significantly increasing transformation efficiency in Gram-positive bacteria, was not successful and no transformants or poor transformation efficiencies were observed (data not shown). Generally, very poor growth was observed in the presence of low concentrations of glycine (more than 0.25%), even with sucrose or PEG osmotic protection. Equally, addition of osmoprotective agents (various concentrations of sucrose, PEG or lactose) to the recovery medium always had detrimental effects on growth and transformation efficiency, and addition of sucrose to the growth medium at high concentrations (0.2M and more) led to a significant decrease in growth. Importantly, when culture degeneration [27] was observed (represented mainly by formation of very long, mycelium-like cells in log and late-log phase), transformation efficiency was reduced drastically and only a few colonies grew on the selective medium. After optimization of electrotransformation steps, we wanted to better understand the influence of Dam and Dcm methylation individually to resolve which one is detrimental or potentially helpful in transformation. We compared electroporation transformation Fig. 2 Testing the presence of potential restriction barriers. Cultivation of pMTL83253 (a) and pMTL82254 (b) with crude protoplast extract (PE) and whole cell lysate (WL) prepared from C. pasteurianum NRRL B-598 at 37 °C. Positive control (c): cultivation of pMTL83253 with PE prepared from C. pasteurianum DSM 525 by the same method Page 6 of 14 Kolek et al. Biotechnol Biofuels (2016) 9:14 efficiencies of experiments where plasmid DNA isolated from the following methylation-deficient E. coli strains were used: JM110 (dam−/dcm−), BL21 (dam+/dcm−) and GM33 (dam−/dcm+). DNA extracted from E. coli DH5α (dam+/dcm+) was also used for confirmation that Dam and Dcm methylations represent a real obstacle to transformation, even when the optimized electrotransformation protocol was performed. A few erythromycin-resistant colonies (a maximum 8 of CFU) containing pMTL83253 were sometimes obtained if DNA from DH5α (fully methylated) was transformed. Relatively consistent results were achieved by transformation of hemimethylated plasmid DNA. Both methylations led to a significant reduction in transformation efficiency. The influence of various methylations on electrotransformation efficiencies is summarized in Table3. Establishment ofconjugational transfer Conjugation was not observed when an E. coli strain supporting Dam or Dcm methylation was used as a donor for transmission of pMTL80000 series plasmids to our strain used in the pilot experiment (see above). Based on our experience from electrotransformation experiments, we constructed a new conjugation donor strain by transmission of RP4 helper plasmid to E. coli JM110 (dam−/dcm−) containing pMTL83253. With this donor ensuring transfer of unmethylated pMTL83253 between donor and recipient cells, we tested for conjugation. Conjugation using a methylation-deficient donor was successful and many erythromycin-resistant colonies were Fig. 3 Confirmation of pMTL83253 presence in C. pasteurianum NRRL B-598 erythromycin-resistant transformants. a pMTL83253 isolated from E. coli JM110 (a1) and C. pasteurianum NRRL B-598 transformants (a2). b pMTL83253 isolated from E. coli JM110 (b2) and C. pasteurianum NRRL B-598 transformants (b3) cleaved by PstI (resulting fragments 370 bp and 4413 bp) compared to the GeneRuler 1 kb DNA ladder - Thermo Scientific (b1) Fig. 4 Optimization of electrotransformation conditions. Influence of various voltages and cell growth state presented by OD600 (used conditions: 0.2 cm gap electroporation cuvette, time constant 11 ms) on transformation efficiency (a); influence of different time constants (conditions used: 0.2 cm gap electroporation cuvette, voltage 1 000 V) on transformation efficiency (b); influence of square-wave pulse delivery (c) Page 7 of 14 Kolek et al. Biotechnol Biofuels (2016) 9:14 observed after 48h. CFUs achieved after various conjugation times (6 or 24h) are summarized in Table4. Use ofsonoporation fortransmission ofplasmid DNA As described previously, ultrasound could also be a useful technique to use for transformation of Gram-positive bacteria. From a few sonoporation media tested (TYA broth, 0.5M CaCl2, sterile water, SMP and PEG), only 10 and 30% PEG 8000 were suitable for relatively high-efficiency transformation. No or only a few transformants were achieved when other sonoporation media were used. An adequate time of ultrasonic pulse was designed according to previous experiences with sonoporation of Gram-positive bacteria, where 20s was identified as a critical time for ultrasound-mediated plasmid DNA degradation but less time led to a reduction in transformation efficiency [16]. Sonoporation has been proven to be a very effective method of transformation that provides even higher transformation efficiencies than electrotransformation. Efficiencies of transformation achieved by sonoporation are summarized in Table4. Combined sono/electroporation forincreased transformation efficiency Because cell-wall weakening approaches were not successful, we compiled a combined method using both sonoand electroporation for improving transformation efficiencies. During the first set of sono/electroporation experiments, we observed that a square-wave pulse provided more consistent results and significantly higher efficiency than the previously used exponential pulse. Also, different amounts of DNA (0.25–2μg) were used for establishing the most efficient approach. Slightly higher voltage (1250 V) produced the most transformants in the square-wave mode and best transformation efficiency was achieved with 0.5μg of plasmid DNA (see Fig.5). By a combination of both techniques, we were able to reach a transformation efficiency of 5.3×102cfu/ μg DNA (see Table4). Discussion The development of methods for efficient genetic manipulation of clostridial bacteria is generally very challenging. Protocols for transmission of foreign DNA to many clostridial species have been developed [20], but these transformation procedures use very different conditions and their overall efficiencies vary by orders of magnitude from 100 to 106 transformants/μg of DNA. Furthermore, transformation conditions are often useful for only one strain and cannot readily be used for other species or even strains. At least, rational step-by-step optimization of the protocol is necessary in order to achieve consistent results. A unique approach to transformation must be developed when the strain expresses a specific restriction Table 3 Influence ofDNA methylation stage tothe electrotransformation efficiency a 2μg of DNA was used for transformation DNA amount [μg]/E. coli strain (designation) CFU (average count)aEfficiency (CFU perμg DNA) 2 μg/DH5α (dam +/dcm +) 3 1.5 2 μg/GM33 (dam −/dcm +) 27 13.5 2 μg/BL21 (dam +/dcm −) 28 14 2 μg/JM110 (dam −/dcm −) 236 118 Table 4 Summary ofpMTL83353 containing CFU yielded byconjugation, sonoporation, andcombined sono/electroporation approaches a 2μg of DNA was used for transformation b 0.5μg of DNA was used for transformation Method CFU (average count) Efficiency (CFU perμg DNA) Conjugation (E. coli JM110 containing RP4 and pMTL83253 donor) 6 h of conjugation 12 24 h of conjugation 37 Sonoporation 10 % PEG 8000 buffer, 20 s pulse 225a112.5 30 % PEG 8000 buffer, 20 s pulse 321a160.5 Sono/electroporation 30 % PEG 8000 buffer, 20 s ultrasound pulse, 5 ms square-wave pulse 5 ms (1250 V) 265b530 Page 8 of 14 Kolek et al. Biotechnol Biofuels (2016) 9:14 barrier that prevents effective transformation, or when conditions from previously published approaches are unsuccessful, as in our case. C. pasteurianum NRRL B-598 represents a non-type strain of solventogenic clostridium that could be a good candidate for production of organic solvents in an ABE process. This strain excels in very high oxygen resistance and overall robustness that could be helpful for a large-scale ABE process. Moreover, biosynthesis of some nonspecific proteases that allows the use of cheap nitrogen sources in its cultivation (e.g., waste whey products) has been described previously for this strain [28]. During our experiments, we showed that C. pasteurianum NRRL B-598 carries a cat gene encoding resistance to chloramphenicol and thiamphenicol, the antibiotics that is normally effective against many strains of clostridium bacteria. This finding is a little surprising because chloramphenicol and thiamphenicol resistances have only been observed in solventogenic species such as C. beijerinckii, but not C. pasteurianum. The action of various restriction-modification (R-M) systems represents a frequent obstacle in the transformation of clostridia, as well as other Gram-positive species. Type II R-M systems recognize a defined short sequence in the foreign DNA and promote its degradation after transmission to the cytoplasm, or even immediately on the cell surface [29]. R-M II systems were described as a reason preventing transformation of C. acetobutylicum ATCC 824 [22], C. pasteurianum ATCC 6013 [12] or C. cellulolyticum ATCC 35319 [8]. In these cases, special treatment by DNA-methyltransferase, which masks all recognition sequences, was necessary before transformation. Type I R-M systems could also be responsible for a decrease in transformation efficiency like in C. saccharobutylicum NCP 262 [23]. Specific protein inhibitors (such as TypeOne restriction inhibitor), protective methylation or heat inactivation could be approaches for overcoming these systems [29]. Equally, reduction of transformation efficiency could be caused by R-M III or IV, but these systems have, so far, been very poorly described in clostridia. Based on the analysis of PacBio SMRT data, we demonstrated the genomic existence of two type I R-M systems, Cpa598I and Cpa598II. Activity of these systems was also confirmed experimentally by cultivation of pMTL82254 which contained recognition sequences of both R-M systems. Restriction provides probably unspecific cleaving of DNA in the direction from the recognized motifs which is typical for type I R-M systems [30]. Both recognized motifs are included in sequence of pBP1 replication origin module of pMTL80000 plasmids system thus it is better to use other replicon for transformation of this strain. On the other hand, when unmethylated pBP1 replicon-based plasmid (pMTL82251) was transformed by electroporation, we were still able to obtain a few transformants. Both type II R-M systems are most certainly inactive because no methylated recognition sequence for Cpa598ORF2410 system was found and no m5C methylations assigned to Cpa598ORF20205 system were detected. We note that the kinetic signatures of m5C bases may not have been strong enough to study properly, but in a relatively high sequence coverage (79×) not a single m5C methylation was detected and also no active type II R-M system was obtained during experimental testing of their presence in the protoplast or whole cell lysates. Activity of the remaining type IV R-M system remains unclear, since these systems are poorly described and neither recognition sequence nor the type of methylation was assigned to this system. Nevertheless, because Cpa598ORF12465P is a methyl-directed restriction enzyme, its activity could also be the reason for decreased transformation efficiency. Further studies are required to verify these hypotheses. The C. pasteurianum NRRL B-598 genome contains a relatively large number of antibiotic efflux genes. Antibiotic resistance can be confirmed by almost normal Fig. 5 Optimization of sono/electroporation conditions. Influence of various voltages and exponential and square-wave pulse deliveries (a); influence of DNA amount on transformation efficiency (b) Page 9 of 14 Kolek et al. Biotechnol Biofuels (2016) 9:14 growth of cells in a medium containing various antibiotics over long periods of time. The addition of TypeOne restriction inhibitor, which has been described previously as a functional agent for overcoming R-M I systems in E. coli or Salmonella typhimurium [31], also did not lead to successful transformation. Based on these results, we assumed that a restriction barrier requiring methylation protection of plasmid DNA probably did not constitute a relevant obstacle during transformation of DNA extracted from E. coli or its conjugal transfer to C. pasteurianum NRRL B-598. Methylation of transmitted DNA can also clearly affect the efficiency of bacterial transformation. Significant reductions in transformation efficiencies when methylated DNA was used were described for many bacterial species such as Streptomyces or Lactobacillus. Methylspecific restriction systems probably play a major role in these observations [32, 33], but the fact that methylated ori sequences on a plasmid may not associate with a specific replication protein could also play an important role in transformation efficiency [34]. Fully methylated DNA isolated from Escherichia coli (dam+/dcm+) was, in most cases, referred to as the best template for clostridial transformation because Dam and Dcm methylation could protect DNA from degradation by nucleases and could increase clostridial transformation efficiencies. Reported cases of detrimental influences of E. coli methylation were observed in C. thermocellum DSM1313 and C. ljungdahlii DSM 13528, but eventually only Dcm methylation was identified as the origin of transformation problems in both experiments [13, 34]. Surprisingly, when unmethylated plasmid DNA was used for electrotransformation of C. pasteurianum NRRL B-598, we suddenly obtained a few transformants. For electrotransformation, a previously published protocol for C. beijerinckii NCIMB 8052 [25] was used and the maximum transformation efficiency, achieved with pMTL82353, was 6cfu/μg DNA. The transformation efficiency achieved was very low compared to other clostridia or Gram-positive bacteria and could not be used for effective genetic manipulations or research on this strain. Because a previously published protocol for other species was used without changes, we wanted to optimize it directly for C. pasteurianum NRRL B-598, hopefully leading to an improved transformation efficiency. The efficiency of electrotransformation may be affected by many parameters such as growth medium, cell growth phase, composition of electroporation buffer, voltage of electric pulse, or its length (influenced mainly by capacitance and resistance of the electroporator). For electrotransformation of clostridial species, cells in early-log to late-log growth phase, different electroporation buffers with low conductivity containing osmostabilizing agents (sucrose, PEG, etc.), and a relatively low electric field (around 5kVcm−1) are usually used [20]. We found that the best growth phase of C. pasteurianum NRRL B-598 for electrotransformation was between late logarithmic and early-stationary phase (OD600 1.2–1.4), which is not typical for most solventogenic strains. Similarly, the best transformation efficiency was obtained when electroporation was conducted in 10% PEG 8000 and decreased when the SMP electroporation buffer (at various pH values) was used. Through step-by-step optimization, we were able to achieve an average electrotransformation efficiency of 1.2×102cfu/μg DNA when unmethylated DNA was used. This was much lower than for the type strains C. acetobutylicum or C. beijerinckii, where the electrotransformation efficiencies reached 104–105 transformants per μg of DNA [22, 25]. Nevertheless, this efficiency is sufficient to use this method for some genetic improvements and basic research on this intractable strain. Achieved transformation efficiency showed clearly that with a decreasing number of any E. coli methylations, transformation efficiency significantly increased. Thus, both Dam and Dcm methylations were shown to be detrimental to transformation, a fact that has not been described previously in transformation of other clostridia. Previously, Pyne etal. [20] described similar effect of CpG methylation which presence led to obtain no transformants even though CpG provided good protection against digestion by described R-M system. If we take into account the number of Damand Dcmspecific methylation sites on pMTL83253 (10 and 18 resp.), we can postulate that Dam methylation could be a little more detrimental than Dcm, which is at variance with findings obtained previously [13, 35]. Decreased efficiency could be caused by a reduction in replication efficiency or some methyl-specific restriction system that may be present in cells as a protection against foreign DNA, e.g., bacteriophage, exhibiting a foreign methylation pattern. The best described similar systems are, for example, the DpnI system in Streptococcus pneumoniae [36] or model methylation-dependent systems McrA, McrBC, and Mrr as described in E. coli [30]. If some methyl-specific type IV restriction system occurs in our strain (see above), it would be quite interesting because no restrictions were obtained when we conducted an examination of restriction systems with Dam and Dcm methylated pMTL83253. However, we focused mainly on R-M I and II systems, so some putative R-M IV (methylspecific) systems may not be active under these invitro conditions. The influence of E. coli methylation was also verified in conjugation experiments, where pMTL82353 transmission was only successful in the methylation-deficient