Background

The development of new processes for sustainable energy and chemical production is in great demand due to the urge of reducing the carbon footprint, aiming to fix current environmental issues [1]. With raised climate ambitions to reduce at least 55% (from 1990 levels) of EU greenhouse gas (GHG) emissions by 2030 [2] and targets on climate and energy framework of at least 32% share for renewable energy [3], it is crucial to develop new technologies for generating renewable energy. While electricity seems to be the solution for some cases, such as road transportation, it is unlikely that electrical power alone will be able to replace fossil fuels as a platform for global chemical needs. Instead, for the chemical industry, heavy road transport, marine, and aviation sectors, bioprocesses are one of the few options to replace the current fossil feedstock with a renewable resource, thereby reducing the sectors’ GHG emissions [4, 5].

One promising emerging technology to address these challenges is to source low-carbon fuels and chemicals via fermentation of syngas (CO, CO2, H2), which can be obtained after gasification of renewable resources such as lignocellulosic biomass and also from industrial and municipal wastes [6], and as a by-product of industries like steel manufacturing [7]. Syngas fermentation is based on the ability of microorganisms to convert syngas components into alcohols and carboxylic acids [8, 9]. Most of these microorganisms are autotrophic acetogens that can fix CO/CO2 and H2 via the Wood–Ljungdahl pathway [10, 11]. Recent progress in genetic and metabolic engineering of acetogens [12,13,14] offers gateways to produce a wide variety of non-natural compounds through the introduction of exogenous pathways that are either borrowed from other organisms or synthetically designed. The autotrophic acetogens with developed genetic toolkits are mostly Clostridium ljungdahlii, Clostridium autoethanogenum, Clostridium carboxidivorans, Eubacterium limosum, Moorella thermoacetica, Thermoanaerobacterium kivui, and Acetobacterium woodii [12, 14]. Examples of production of non-native compounds in these organisms are butanol, butyrate, acetone, mevalonate, and isoprene production in C. ljungdahlii [15,16,17,18], or acetone, isopropanol, isobutanol, and butyrate in A. woodii [19,20,21,22]. C. autoethanogenum was also manipulated for the production of acetone and isopropanol from syngas, which could be achieved at industrial pilot scale [23]. Most used genetic parts for these organisms such as resistance genes and origins of replication are functional in multiple species [12] and thereby expanding the existing molecular tools to other acetogens should be possible. The currently studied bioprocesses involving gas fermentation usually face limitations concerning low growth rates, toxicity by gas streams, and low productivity [24]. By studying non-model organisms with different characteristics, these limitations may be overcome. In previous work, a novel acetogenic bacterium strain, Acetobacterium wieringae strain JM, was isolated by our research team [25]. This novel autotrophic acetogen is an interesting organism with different physiology when compared to its closest relative A. woodii. It is very versatile, having industrially relevant characteristics since it grows at high rates in different gas compositions and does not require additional carbon sources as supplementation [25]. Additionally, as shown in genomic analysis, non-model organisms within the Acetobacterium genus may provide varied metabolic capabilities and other abilities to survive in diverse environments and potentially be exploited for biotechnological applications [26]. A. wieringae JM produces mainly acetate and ethanol from gas fermentation [25]. With an interest in expanding the product frame of gas fermentation, A. wieringae JM was previously used in a co-cultivation strategy to produce propionate from carbon monoxide [27]. To further expand the product frame of A. wieringae JM, we aim to make this acetogen genetically accessible and implement heterologous pathways to produce more valuable compounds.

Here, we report the development of a robust, stable, and efficient electroporation-mediated transformation system for A. wieringae JM which can also be used in other Acetobacterium strains. We determine the impact of several parameters of the transformation procedure on the transformation efficiency, including different origins of replication. Furthermore, we demonstrate the expression of heterologous genes for acetone production.

Results

Screening of molecular tools for transformation of A. wieringae strain JM

To transform the novel acetogen A. wieringae strain JM, as commonly carried out for new potential hosts, we performed minimum inhibitory concentrations (MIC) assays to test the natural antibiotic resistance of A. wieringae JM to thiamphenicol, clarithromycin, spectinomycin, and tetracycline (Table 1). Strain JM did not show resistance to any of the tested antibiotics, so we decided to select the catP resistance marker to thiamphenicol at a working concentration of 15 μg/mL for the selection of A. wieringae JM transformants as the MIC was 3 μg/mL.

Table 1 Minimum inhibitory concentrations of four antibiotics in A. wieringae strain JM and typical working concentrations for the resistance markers of these antibiotics in A. woodii and C. ljungdahlii
Full size table

A series of plasmids from the pMTL80000 system of modular E. coli–Clostridium shuttle vectors [28] with proven efficiency in A. woodii [19] were used to transform A. wieringae JM. They differ in their Gram+ origins of replication: pMTL82151 (pBP1 from C. botulinum); pMTL83151 (pCB102 from C. butyricum); pMTL84151 (pCD6 from C. difficile); and pMTL85141 (pIM13 from Bacillus subtilis) [28]. Firstly, we attempted the conjugative plasmid transfer using the conjugal donor strain Escherichia coli CA434. Such a conjugation system works well for many organisms and is particularly useful when no electroporation protocol is available. In contrast with some Clostridium species where this method was applied [28], A. wieringae JM is not resistant to D-cycloserine at the working concentration of 250 μg/ml, which is typically used as counterselection against the E. coli donor strain. Therefore, we tried an alternative approach to select transconjugants, using 25 μg/ml of thiamphenicol to select for the plasmid and autotrophic conditions (carbon and energy source: 60% CO, 30% H2, 10% CO2; (v/v)) to select for A. wieringae JM, also omitting yeast extract from the medium. The resulting cultures of the conjugative mating process of E. coli CA434 (pMTL83151 and pMTL84151) with A. wieringae JM could be grown after four consecutive transfers on selective mineral media under autotrophic conditions, but transconjugants of A. wieringae JM could not be identified nor isolated as E. coli could grow in symbiosis with A. wieringae JM under autotrophic conditions.

Development of an electrotransformation procedure for A. wieringae strain JM

As the conjugation procedure failed to produce isolated transformants of A. wieringae JM, we developed a new protocol that can be used to transform A. wieringae JM, A. wieringaeT, and A. woodiiT, based on previously reported electrotransformation procedures for A. woodii. Table 2 lists the detailed differences between the published transformation protocols for A. woodii and the final version of our new procedure (protocol 7, Table 2). The first electrotransformation of A. woodii was published in 1994 by Strätz et al. [29] (protocol 1, Table 2). The transformed plasmids pMS3 and pMS4p carried the replicon derived from the plasmid pAMβ1. In that study, the resulting transformation efficiencies (ET), using tetracycline selection, were 4.5 × 103 CFU/μgDNA for both plasmids [29]. Publications using other procedures reported the successful usage of the following vectors although transformation efficiencies were not given: pJIR750 plasmids carrying the replicon pIP404 from C. perfringens [30] (protocols 3 and 4, Table 2); pMTL82151 (pBP1) (protocol 4, Table 2); pMTL83151 (pCB102) (protocols 2, 4 and 5, Table 2); and pMTL84151 (pCD6) (protocols 2 and 4, Table 2). In C. ljungdahlii, the utilization of protocol 2 (Table 2) with the plasmids pMTL82151 (pBP1) and pMTL83151 (pCB102) resulted in ET values of 3.8 × 103 and 3.1 × 103 CFU/μgDNA, respectively [31]. More recently, Baker et al. reported a new transformation protocol (protocol 6, Table 2) for A. woodii using the plasmid pMTL83141 (pCB102) with ET values of 4.0 × 105 CFU/μg [32].

Table 2 Comparison of published electrotransformation procedures used in Acetobacterium woodii and a new procedure for Acetobacterium strains
Full size table

The most important changes for higher efficiency and reproducibility were the use of fixed cell density of competent cells resulting in electroporation cell-plasmid ratios of 2 to 3 (OD600 × mL/µg), the use of electric field strength of 10 kV/cm, and the selection of transformants in serum anaerobic bottles with molten agar (Additional file 1 Fig. S1a). Furthermore, during competent cell preparation, cells are usually grown with cell wall-weakening agents, such as lysozyme, glycine, DL-threonine, or penicillin G [42, 43], to promote the passage of DNA through the thick Gram+ cell wall. Protocols 2 to 5 report the use of DL-threonine (20–40 mM) during cell growth before the preparation of competent cells (Table 2). In our hands, A. wieringae JM could not be transformed without the use of cell wall-weakening agents. The use of 40 mM D-threonine allowed the transformation of A. wieringae JM while the addition of 1.25% glycine (% w/v) caused growth inhibition. Protocols 1 and 3 do not allow storage of competent cells before electroporation, while protocols 2, 4, and 5 use DMSO, protocol 5 uses DMF in the SMP resuspension buffer as cryoprotectants to allow storage of competent cells. We used DMF as it is also utilized as a solvent during the preparation of the antibiotic thiamphenicol. In our protocol, the plasmid DNA was prepared from E. coli TOP10 (Invitrogen, MA, USA) cells that have the methyltransferases Dam and Dcm which methylate the DNA bases adenine and cytosine, respectively, allowing protection against endonucleases. No external, site-specific methylation was used, in contrast with protocol 3 where a methylase gene of C. ljungdahlii (CLJU_c03310) was used in addition to an E. coli Dam+/Dcm+ background. Without this, the transformation of A. woodii with pJIR750 would fail [39]. Protocols 1, 4, and 5 report the utilization of Dam+/Dcm+ methylation background with E. coli XL1-Blue (MRF’) strain (Stratagene, CA, USA), while protocol 2 utilizes an E. coli strain lacking Dcm (New England Biolabs, MA, USA) which yielded better transformation results in C. ljungdahlii [31].

To confirm the presence of the pMTL80000 plasmids in A. wieringae JM transformants, we isolated plasmid DNA from recombinant cells and transformed it into E. coli DH5α cells. After re-purification, the pMTL80000 vectors were verified by restriction analysis. Correct digestion profiles were obtained for each re-purified plasmid (Fig. 1), confirming the presence of the transformed plasmids in A. wieringae JM.

Fig. 1
figure 1

Plasmid restriction analysis of Acetobacterium wieringae strain JM transformants. Digestion of pMTL82151, pMTL83151, pMTL84151, and pMTL85151, using NcoI-HF and SpeI-HF, EcoRI-HF, NcoI-HF, NheI-HF, and NotI-HF, respectively, resolved on a 1% agarose gel. Digestion reactions contained 0.5 μg of plasmid DNA and 1 μL of restriction enzymes in a total volume of 50 μL. For comparison, plasmids are shown undigested and digested. Expected band sizes are: 3114 and 2140 bp, pMTL82151; 3475 and 1001 bp, pMTL83151; 3606 and 2697 bp, pMTL84151; 3453 and 276 bp, pMTL85151

Full size image

The development and optimization of our electrotransformation procedure for maximized efficiency in A. wieringae JM required the empirical study of parameters involved in the preparation of electrocompetent cells such as the density at which cells are harvested, number of washing steps, pH of resuspension buffer, and the study of key aspects of the electroporation process such as pulse voltage, plasmid amount, and recovery time. The plasmid pMTL83151 carrying the pCB102 replicon yielded the most colonies during the first tests, therefore it was used for the optimization experiments, results are shown in Fig. 2.

Fig. 2
figure 2

Optimization of electrotransformation procedure in Acetobacterium wieringae strain JM using the plasmid pMTL83151. a Effect of harvesting OD600 during the preparation of A. wieringae JM competent cells on transformation efficiency, ET. Growing cells were sampled during exponential phase (OD600 0.21, 0.36, 0.43, 0.51, 0.60, 0.86) for preparation of competent cells prior electroporation. After resuspension of competent cells in SMP 10% DMF, all conditions had an OD600 of 30. b Effect of the number of washing steps during preparation of A. wieringae JM competent cells on transformation efficiency. Cells were grown to OD600 of 0.36 and were sampled to be washed 1, 2, 3, 4, and 5 times with SMP buffer during the preparation of competent cells before electroporation. c Effect of pH of resuspension buffer on transformation efficiency. Cells were grown to OD600 of 0.40, washed in SMP buffer pH 6, and were sampled to be resuspended in SMP 10% DMF pH 5.5, 6.0, 6.5, 7.0, 7.5, and 8.0 during competent cell preparation before electroporation. d Effect of pulse voltage (field strength) on transformation efficiency. ET was measured using electric pulses of 1.0, 1.5, 1.8, 2.0, and 2.5 kV, corresponding to field strengths of 5.0, 7.5, 8.0, 10, and 12.5 kV/cm. e Effect of plasmid DNA amount on colony-forming units (CFU) and transformation efficiency. Separately, 0.5, 1.0, 2.0, 3.0, 4.0, and 5.0 μg of plasmid corresponding, respectively, to cell–plasmid ratios of 12.0, 6.0, 3.0, 2.0, 1.5, and 1.2, were added to competent cells of A. wieringae JM and electroporated. The total number of colony-forming units and transformation efficiency was quantified. f Effect of post-electroporation incubation time on transformation efficiency. Cells were electroporated, transferred to a 3 mL outgrowth medium, and incubated for 2, 4, 6, 8, and 20 h before selective plating

Full size image

When preparing competent cells, the cell density at which cells are harvested can impact the transformation efficiency [44]. The published transformation protocols for A. woodii harvested cells when OD600 reached 0.2–0.7 (protocols 1–6, Table 2). We harvested A. wieringae JM at six different optical densities (OD600 of 0.21, 0.36, 0.43, 0.51, 0.60, and 0.86). The cells harvested at OD600 of 0.43 produced the most colonies, resulting in transformation efficiency (ET) of 3.1 × 102 CFU/μgDNA. Cells harvested at OD600 of 0.36, 0.51, and 0.60 resulted in the relative reduction of ET by 0.41- to 0.51-fold, while harvesting cells early or late during the exponential phase of growth (OD600 of 0.21 and 0.86), reduced ET by 0.73- and 0.81-fold, respectively (Fig. 2a). Therefore, we defined the optimal harvesting OD600 range of 0.3–0.6 (protocol 7, Table 2).

All available transformation protocols report the washing of competent cells in two steps (protocols 1 – 5, Table 2). We attempted to increase the ET of A. wieringae JM by increasing the number of washing steps up to 5. However, maximal ET was obtained with only one washing step (5.0 × 102 CFU/μgDNA, Fig. 2b). Washing the cells two times resulted in a reduction of ET by 0.66-fold, while five washes had an even higher impact on ET, by 0.84-fold. As centrifugation of competent cells is performed inside the anaerobic chamber, the refrigeration efficiency decreases as the time of centrifugation increases. Therefore, increasing the number of washes compromises the ET of A. wieringae JM.

Cultivation media for A. wieringae JM and A. woodii usually have pH close to neutral (~ 6.8–7.0), but transformation procedures for A. woodii report the use of pH 5.8–6.0 on the washing and resuspension buffers (protocols 1–6, Table 2). We tested the effect of pH (range of 5.5–8.0) in the SMP 10% DMF resuspension buffer during the preparation of competent cells on ET of A. wieringae JM (Fig. 2c). Results show that the use of pH 6.0 and 6.5 produced the most colonies, resulting in ET values of 2.8 × 102 and 2.4 × 102 CFU/μgDNA, respectively. The more acidic pH of 5.5 reduced ET by 0.29-fold, while pH of 7.0, 7.5, and 8.0 caused reductions in ET of more than 0.56-fold. The pulse duration decreased with the increment of pH, without a significant impact on ET.

Key electroporation parameters, such as voltage, resistance, or field strength, which are dependent on cuvette gap width, influence the transformation efficiency. Other reported procedures for A. woodii use different combinations of cuvette gap width and pulse voltage, resulting in a pulse strength of 6.25 kV/cm with 600 Ω of resistance in protocols 2–5, and 50 kV/cm in protocol 1 with 400 Ω of resistance (Table 2). We investigated the effects of the electrical pulse concerning voltage (i.e., field strength), using cuvettes with 0.2 cm gap width, as the electroporation device used (E. coli Pulser™ Transformation Apparatus) only allows adjustment on pulse voltage. The electric resistance of this electroporation device is 20 Ω. Pulses of 1.0, 1.5, 1.8, 2.0, and 2.5 kV were administered, corresponding, respectively, to field strengths of 5.0, 7.5, 8.0, 10, and 12.5 kV/cm (Fig. 2d). Higher voltage resulted in higher transformation efficiency, the voltage of 2.5 kV was found to produce the greatest ET values (1.7 × 102 CFU/μgDNA) but sample ‘arcing’ occurred often, although pulse of 2.0 kV only slightly reduced ET by 0.06-fold without any ‘arcing’ occurrence. With the weakest tested pulse, 1.0 kV, it was still possible to obtain colonies but there was a 0.98-fold reduction in ET. 1.5 and 1.8 kV impacted ET by 0.68 and 0.45-fold, respectively. Pulse duration decreased by increasing pulse voltage.

Plasmid amount and cell–plasmid ratio (OD600 × mL/µg) can also influence the transformation efficiency. In reported transformation procedures for A. woodii, there is some variation in used plasmid amounts and the cell–plasmid ratios, between and within different protocols (Table 2). Plasmid amounts of 0.5, 1.0, 2.0, 3.0, 4.0, and 5.0 µg were used, corresponding, respectively, to cell-plasmid ratios of 12.0, 6.0, 3.0, 2.0, 1.5, and 1.2 (Fig. 2e). Although the total number of transformants was found to increase as expected between 0.5 and 5.0 μg of pMTL83141, the greatest ET occurred at the lowest quantity of DNA tested, 0.5 μg of plasmid DNA (2.8 × 102 CFU/μgDNA), but less than 100 colonies were observed in duplicates. By using 1–3 μg of plasmid, ~ 150–300 colonies were found which resulted in the average ET of 2.3 × 102 (± 7.9 × 101) CFU/μgDNA.

The recovery period after applying an electric field to the cells is performed in a liquid medium with no selective pressure and can also influence the transformation efficiency as longer periods may lead to plasmid loss and shorter periods may not allow the expression of the selection marker gene before plating. Recovery periods of 5 h up to 3 days have been reported for A. woodii which may also depend on the used recovery media and Gram+ replicon, which also differs between protocols (Table 2). For assessing outgrowth duration with pMTL83151 in A. wieringae JM, electroporated cells were incubated for 2, 4, 6, 8, and 20 h before plating on a selective medium. Transformants could be obtained in all conditions, 20 h of recovery time resulted in the greatest ET value of 2.3 × 102 CFU/μgDNA, but there was not a significant difference for the recovery periods of 2, 4, and 6 h, representing on average a 0.23 (± 0.03)-fold reduction, 8 h of recovery time had the highest impact on ET, with a 0.51-fold reduction (Fig. 2f).

Application of the electrotransformation protocol to other vectors and strains

The effect of the replicons pBP1, pCB102, pCD6, and pIM13 on the transformation efficiency of A. wieringae JM, A. wieringaeT, and A. woodiiT, was assessed using our optimized transformation protocol (protocol 7, Table 2). All Acetobacterium strains were harvested at OD600 of ~ 0.30, washed two times during the preparation of competent cells, and 3 μg of plasmid were used for electroporation. All vectors carrying the different replicons have the same backbone including the catP resistance gene but differ in size: 5254 bp, pMTL82151 (pBP1); 4476 bp, pMTL83151 (pCB102); 6297 bp, pMTL84151 (pCD6); 3729 bp, pMTL85151 (pIM13). Colonies could be obtained using the four different vectors in all tested Acetobacterium strains, and results show that ET differences between plasmids are not justified by plasmid size, but by the replication mechanism or stability of each replicon (Fig. 3). In A. wieringae JM, pCB102 produced the most colonies, with an ET of 1.6 × 102 CFU/μgDNA, while ET of pCD6, pBP1, and pIM13 were, respectively, 0.38-, 0.46-, and 0.90-fold lower (Fig. 3). In A. wieringaeT, the use of pCB102 also had the highest ET of 2.0 × 103 CFU/μgDNA, while ET of pCD6, pBP1, and pIM13 were, respectively, 0.20-, 0.93-, and 0.96-fold lower (Fig. 3). The transformation efficiency of A. wieringaeT with pCB102 was 12.3-fold higher than the ET of A. wieringae JM for the same replicon. Other replicons also resulted in higher ET with A. wieringaeT than with A. wieringae JM. Lastly, in A. woodiiT, transformation efficiencies were very similar for the replicons pBP1, pCB102, and pCD6, with an average ET of 4.8 × 102 (± 2.9 × 101) CFU/μgDNA, while ET of pIM13 was 0.97-fold lower (Fig. 3). For all tested plasmids, transformation efficiencies in A. woodii were equal (pIM13) or higher (pBP1, pCB102, and pCD6) than in A. wieringae JM, and lower than in A. wieringaeT.

Fig. 3
figure 3

Transformation efficiencies for different Acetobacterium strains and different origins of replication. A. wieringae strain JM, A. wieringae DSM 1911 T, and A. woodii DSM 1030 T were transformed with the plasmids pMTL82151 (pBP1), pMTL83151 (pCB102), pMTL84151 (pCD6), and pMTL85151 (pIM13)

Full size image

Validation of the vector pMTL83151 for plasmid-based expression in Acetobacterium strains

To test the heterologous production of chemicals in Acetobacterium strains via plasmid-based expression, as a proof of concept, we constructed the biosynthesis pathway for acetone production from acetyl-CoA in Acetobacterium by utilizing the shuttle vector pMTL83151. Acetone production in Acetobacterium, from acetyl-CoA is theoretically feasible with the presence of the acetone production operon (APO) from Clostridium acetobutylicum DSM 792, which consists of the genes thlA (encoding thiolase A), ctfA/ctfB (encoding CoA transferase), and adc (encoding acetoacetate decarboxylase). The APO was cloned into the pMTL83151 vector, under the control of the thlA native promoter, PthlA, and the transcriptional terminator from downstream of the CD0164 ORF of Clostridium difficile strain 630. The resulting plasmid, p83_APO was transformed in A. wieringae JM, A. wieringaeT, and A. woodiiT. Acetone production was detected via gas chromatography in the three recombinant Acetobacterium strains after 10 days of uncontrolled batch experiments under autotrophic (220 kPa 80% H2, 20% CO2) and heterotrophic (30 mM fructose) conditions (Fig. 4), confirming the expression of recombinant genes. From heterotrophic growth, A. wieringae JM [p83_APO] produced 1.5 ± 0.3 mM acetone, A. wieringae [p83_APO] 2.6 ± 0.6 mM, and A. woodii [p83_APO] 0.6 ± 0.2 mM, respectively. When grown on H2 + CO2, acetone production was lower (A. wieringae JM [p83_APO]: 0.26 ± 0.02 mM; A. wieringae [p83_APO]: 0.56 ± 0.03 mM; A. woodii [p83_APO]: 0.39 ± 0.01 mM). In the control experiments using wild-type strains and mutant strains carrying the empty vector pMTL83151, acetone could not be detected.

Fig. 4
figure 4

Acetone production in recombinant Acetobacterium derived from different host strains using fructose or H2/CO2 as substrate. Acetobacterium [p83_APO] strains were cultivated in serum bottles under autotrophic and heterotrophic conditions. Data represent the mean with SDs of three biological replicates

Full size image

Discussion

This study provides a reliable and reproducible electrotransformation protocol for use in the new acetogen Acetobacterium wieringae strain JM, but also applicable to the type strains of A. wieringae and A. woodii, and potentially to other acetogenic strains. We achieved robust ET values from ~ 1.0 × 102 up to 2.0 × 103 CFU/μgDNA (Fig. 2) for plasmids with the replicons pBP1, pCB102, and pCD6 in all three Acetobacterium strains, which are within the ET range of the same plasmids in other acetogenic species (Table 3). This reinforces the transferability of Clostridial molecular tools to other acetogenic bacteria and uncovers multiple replicons for the manipulation of A. wieringae, which allows the diversification of applicable genetic engineering strategies. For instance, the use of two plasmids with different compatible replicons is common practice for the simultaneous replication of genes in a host, leading to increased size and number of exogenous genes expressed at once [45, 46]. The high transformation efficiencies achieved here also enable library-based approaches, which require high frequencies, in contrast to the conventional transfer of individual plasmids.

Table 3 Summary of electrotransformation efficiencies for different acetogens using the replicons pBP1, pCB102, pCD6 and pIM13 with the plasmid vectors pMTL82151, pMTL83151, pMTL83141, pMTL84151, and pMTL85151
Full size table

Different replicons may replicate at different levels which can later influence copy number and gene expression. For example, in A. woodii it was shown that out of the four replicons pIP404, pBP1, pCB102, and pCD6, a negative impact on acetone production was observed with pCB102 [19]. However, also in A. woodii, the use of pCB102 led to higher lactate productivity than earlier constructions using pIP404 [47]. In Clostridium tyrobutyricum, high transformation efficiency was associated with high segregational stability for the replicons pBP1 and pCB102 but the same relation was not observed for the replicon pIM13 [48]. Generally, replicon stability can vary within species, influencing transformation efficiency, copy number, and segregational stability. Identifying the replicon stability is essential in guiding the replicon selection for downstream applications. While high replicon stability is essential for plasmid-based expression of recombinant proteins, replicon instability leads to easier plasmid loss from growth without selective pressure, which is central to allele-coupled exchange [49] and CRISPR–Cas9-mediated genome editing [50]. Future studies will assess the replicon stability in A. wieringae. As higher transformation efficiencies in A. wieringae JM and A. wieringaeT were obtained with the replicon pCB102, we selected the plasmid pMTL83151 for demonstration of heterologous production of acetone in A. wieringae JM, A. wieringaeT, and A. woodiiT via plasmid-based expression of the acetone biosynthetic pathway from C. acetobutylicum. Acetone production was detected for all three recombinant Acetobacterium strains.

The investigation of various electrotransformation parameters enabled us to understand the impact of each parameter which gives flexibility in adapting the protocol to diverse plasmids and conditions, securing the success of future transformations with the different plasmids. Two factors that greatly influence the success of transformation experiments are the plating procedure and the colony-picking method. For plating, we used 50 mL of media for 1.5 mL of inoculum in 200 mL anaerobic serum bottles, using a pour-plating method described in the Materials and methods section. Most reported protocols use Petri dishes for the plating procedure, either by mixing the inoculum with molten agar [31, 53] or by spreading the inoculum on the agar [21, 29], with further incubation inside the anaerobic chamber (Protocols 2, 4–6, Table 1). In our hands, these approaches led to lower transformation efficiencies and low reproducibility when using anaerobic jars (Oxoid™ AnaeroJar™) for the incubation of Petri dishes, as we could not incubate the cells inside the anaerobic chamber. Other studies report the use of liquid media for the selection of transformants [19, 39], but even though this method led to successful transformation experiments in our hands, it required additional subculturing steps and a final plating step to obtain a genetically homogeneous population of transformants.

Even though the electrotransformation protocol was optimized for A. wieringae JM, lower transformation efficiencies were obtained for this strain compared to the type strains A. wieringae and A. woodii. This may be explained by single-nucleotide polymorphisms (SNPs) which may have occurred during the enrichment and isolation of A. wieringae JM [25]. For instance, SNPs were identified in a lab-adapted strain of A. woodii which were associated with lower transformation efficiencies when compared to the type strain A. woodii DSM 1030 [32]. In addition, different restriction-modification (RM) systems in Acetobacterium strains may exhibit different protection against foreign DNA. All three Acetobacterium strains have RM systems that target and cleave unmethylated adenosine such as the HsdS-HsdM-HsdR system [51]: A. woodiiT (H6LBT5, H6LHR7, H6LHR4, UniProtKB); A. wieringaeT (A0A1F2PIJ0, A0A1F2PL45, A0A1F2PJC8, UniProtKB); A. wieringae JM (A0A5D0WQE5, A0A5D0WKG3, UniProtKB). However, only in A. woodiiT and A. wieringaeT RM systems targeting and cleaving unmethylated cytosine could be found, such as the Dcm, ydiP, and aplIM systems (H6LJP1, H6LIV2, A0A1F2PM82, A0A1F2PKZ9, UniProtKB). Additionally, a type II restriction endonuclease from the NgoFVII family, which recognizes the double-stranded sequence ‘GCSGC’ and cleaves after G-4 could be found in A. wieringae JM (A0A5D0WXZ3, UniProtKB), but not in A. woodiiT and A. wieringaeT. With the use of the E. coli TOP10 strain for plasmid preparation, which provides Dam and Dcm methylation, protection against the RM systems described above should be secured in the type strains of A. woodii and A. wieringae. Since A. wieringae JM does not methylate DNA on the cytosine, the use of an E. coli strain without Dcm for plasmid preparation may improve transformation efficiencies in A. wieringae JM, as observed in C. ljungdahlii [31] and C. autoethanogenum [52].

Conclusions

In this work, the optimization of competent cell preparation, electroporation parameters, and plating procedure enabled the electrotransformation of A. wieringae strain JM, A. wieringae DSM 1911 T, and A. woodii DSM 1030 T at efficiencies of up to 2.0 × 103 CFU/μgDNA. The work here also demonstrates the establishment of a molecular toolkit in A. wieringae, currently comprising the thiamphenicol selection marker catP and four Gram + origins of replication (pBP1 from C. botulinum, pCB102 from C. butyricum, pCD6 from C. difficile, and pIM13 from B. subtilis). This is the first report of a genetic transformation procedure for A. wieringae and represents a key advancement for this industrially C1-gas fermenting relevant species with important applications for renewable chemical and biofuel production.

Materials and methods

Bacterial strains and growth conditions

Bacterial strains and plasmids used in this study are listed in Table 4. E. coli was routinely grown at 37 °C or 30 °C in LB broth or on LB agar. E. coli TOP10 was used for plasmid propagation and E. coli CA434 as a conjugative donor strain. C. acetobutylicum was grown at 37 °C under strictly anaerobic conditions using the DSMZ medium 104b. Acetobacterium strains were grown anaerobically at 30 °C in basal medium prepared as described previously [56], with the addition of yeast extract (0.5 g/L), d-fructose (5 g/L), and phosphate buffer, pH 7.0 (K2HPO4/KH2PO4, 10 mM). All basal media were extensively flushed with N2 and residual oxygen was removed with ~ 0.8 mmol L−1 sodium sulfide (Na2S.7-9H2O) as a reducing agent [56]. Media were supplemented with antibiotics, when necessary, in the following concentrations: for E. coli, chloramphenicol (25 μg/mL) and kanamycin (50 μg/mL); for Acetobacterium strains, thiamphenicol (15 μg/mL).

Table 4 Bacterial strains and plasmids used in this work
Full size table

For the growth of Acetobacterium strains on solid media, a pour-plating method was used. In detail, basal media supplemented with 1.5 g/L yeast extract, was prepared with 1.3% agar and aliquoted by 45 mL in 200 mL serum bottles (Glasgerätebau Ochs, Bovenden, Germany) for anaerobic cultivation. Bottles were then flushed with 100% N2, pressurized to 100 kPa, and sterilized. After this, media were cooled down to 40–50 °C and then mixed with vitamins, d-fructose (5 g/L), reducing agent, antibiotics, and 1.5 mL of liquid bacterial culture. After solidification, bottles were filled up with 20% CO2 and 80% N2 to a final pressure of 170 kPa. The bottles were then incubated at 30 °C until single colonies were visible (3–10 days). Colonies were picked inside the anaerobic working station with long wooden toothpicks.

Recombinant E. coli stocks were stored at − 80 °C in 25% glycerol, and Acetobacterium wild type and recombinant stocks were prepared by concentrating 10 × cultures in fresh basal media (OD600 of 0.30–0.80) and stored at − 80 °C with 10% DMSO.

In the case of batch experiments to detect acetone production in recombinant Acetobacterium strains, cells were cultivated in 200 mL serum bottles with 50 mL of basal media, as described above, using thiamphenicol with a final concentration of 25 μg/mL. For heterotrophic growth fructose was added at a final concentration of 30 mM under the headspace of 80% N2 and 20% CO2 (v/v) at 220 kPa. For autotrophic growth, the headspace was set at 220 kPa with 80% H2 and 20% CO2 (v/v). Recombinant strains were previously adapted autotrophy by subculturing each clone three times in fresh medium before the growth experiment.

Conjugative plasmid transfer

Conjugative transformation of A. wieringae strain JM was performed as described previously [28], with a few modifications. Donor cells of E. coli CA434 carrying the plasmids pMTL82151, pMTL83151, pMTL84151, and pMTL85151, previously transformed by heat-shock transformation, were collected by centrifugation of 1 mL of antibiotic-supplemented overnight culture at 4000 × g for 2 min. Donor cells were washed once with PBS, then the pellet was gently resuspended in 1 mL of 2 times concentrated overnight culture of the A. wieringae JM recipient. This mating mixture was spotted onto a non-selective plate of basal media and incubated anaerobically for 24 h at 30 °C without inverting the plates. The mating mixture was resuspended from the surface of the plate in 2 mL of PBS using a spreader and transferred into 50 mL of selective (25 μg/mL thiamphenicol) liquid basal medium without fructose nor yeast extract, and grown autotrophically on syngas (60% CO, 30% H2, 10% CO2; (v/v)) till growth was obtained. Cultures were then sub-cultured four times under the same conditions. To identify the presence of E. coli, 0.2 mL of the latest culture was spread on LB agar and grown aerobically.

DNA manipulation

E. coli DH5α (NZYTech, Lisbon, Portugal) cells were used for plasmid construction and storage. E. coli TOP10 cells (Invitrogen, MA, USA) were used for plasmid preparation before electrotransformation in Acetobacterium. Plasmids were then isolated from E. coli using the GenElute™ plasmid miniprep kit (Sigma-Aldrich, MO, USA). For Acetobacterium transformants, plasmid DNA was extracted and purified using also the GenElute™ plasmid miniprep kit (Sigma-Aldrich, MO, USA) after the following pre-treatment: 10 mL of late-exponential phase cells were collected by centrifugation (4000 × g, 15 min, 4 °C), resuspended in 1 mL of 10 mM Tris–NaCl, pH 8, with the addition of 10 mg of lysozyme and incubated at 37 °C for 1 h. For genomic DNA extraction from C. acetobutylicum, twenty milliliters of culture were collected by centrifugation (4000 × g, 15 min, 4 °C) and used in the FastDNA SPIN kit for soil (MP Biomedicals, Solon, OH, USA). Restriction enzymes for DNA digestion, T4 DNA ligase for DNA ligation, and Phusion High-Fidelity DNA Polymerase for amplification of DNA fragments were purchased from New England Biolabs GmbH (Main, Germany), Promega Corporation (WI, USA), and Thermo Fisher Scientific (Waltham, MA, USA), respectively. The QIAquick PCR Purification Kit (Qiagen, Valencia, CA, USA) was used to purify the PCR products and DNA fragments obtained by treatment with restriction enzymes. Gel extraction was performed using the QIAquick Gel Extraction Kit (Qiagen, Valencia, CA, USA). All primers were synthesized by Eurofins Genomics (Ebersberg, Germany).

All primers used to construct plasmids for acetone production are listed in Table 5. The acetoacetate decarboxylase gene (adc; CA_P0165), the acetoacetyl-CoA:acetate/butyrate CoA transferase genes (ctfA/ctfB; CA_P0163/CA_P0164), and the thiolase gene (thlA; CA_C2873) with its promoter region (PthlA) were obtained from C. acetobutylicum via PCR amplification using each respective primer pair. Amplified genes were sequentially cloned into the MCS of vector pMTL83151 using the restriction enzymes NotI-HF, KpnI-HF, BamHI-HF, and SalI-HF to form the plasmid p83_APO.

Table 5 Primers used in this study
Full size table

Preparation of electrocompetent cells and electrotransformation

For the preparation of electrocompetent cells of Acetobacterium strains using the optimized protocol, 0.5 L of basal media supplemented with 40 mM d-threonine were inoculated with − 80 °C culture stocks to an OD600 of 0.05 to 0.1 and incubated at 30 °C overnight. At an OD600 range of 0.3–0.6, cultures were placed on ice for 20 min and brought into an anaerobic workstation (Coylab, MI, USA) under a gas atmosphere of about 5% H2 and 95% N2 (v/v). The bottles of the cultures were then opened and stepwise harvested in pre-deoxygenated 50 mL tubes by centrifugation (4000 × g, 15 min, 4 °C), and cell pellets were washed once or twice with 50 mL ice-cold SMP buffer pH 6 (270 mM sucrose, 1 mM MgCl2, 7 mM NaH2PO4). Subsequently, the washed cells were resuspended in SMP buffer 10% DMF pH 6, to a final OD600 of 30, and aliquot by 2 mL in anaerobic vials and stored at − 80 °C.

For electroporation, competent cells were thawed on ice and brought into an anaerobic workstation. Electroporation cuvettes with a gap width of 0.2 cm (Bio-Rad, CA, USA) were placed on ice, and 1–3 μg of plasmid DNA were added, following 0.2 mL of competent cells. Electroporation was performed at the following conditions: 2.0 kV, 20 Ω, and 25 μF (E. coli Pulser™ Transformation Apparatus, Bio-Rad, CA, USA). Directly after electroporation, cells were transferred into 3 mL of basal media and incubated for 20 h at 30 °C. After this non-selective outgrowth step, 1.5 mL of culture were plated using the pour-plating method described above. The transformants were further confirmed by plasmid digestion analysis or PCR amplification. For colony picking, long wooden toothpicks produced the best results, while the use of the pointy end of plastic inoculation loops, long pipette tips, or metallic long needles resulted in lower ratios of grown/picked colonies. The porous surface of wooden toothpicks and their thin pointy ends allow better accuracy and a sticking effect when picking colonies submerged in agar plates. During plating, appropriate dilutions should be done to obtain plates in fewer colonies (Additional file 1: Fig. S1c), and therefore facilitate colony picking, reducing contamination risks and promoting the homogeneity of clones.

Acetone measurement

The concentration of acetone was analyzed by use of a gas chromatograph (Varian 4000, Agilent Technologies, CA, USA) equipped with a flame ionization detector (FID), an autosampler (CP-8400, Bruker, MA, USA), and a Teknokroma Meta.WAX 30 m × 0.25 mm × 0.25 mm TR-810232 capillary column. For sample preparation for GC analysis, 1-butanol was added as an internal standard to cell-free supernatant, in a working concentration of 5 mM. For detection, 10 μL of the sample were injected at a split ratio of 10. The injector temperature was 120 °C, and the column oven temperature was held at 35 °C for 5 min and then increased to 120 °C at a rate of 15 °C/min. The detector temperature was set at 160 °C. The carrier gas was N2 at a flow rate of 1 mL/min.