H. Wernérus, Department of Biotechnology, Royal Institute of Technology (KTH), AlbaNova University Center, Roslagstullsbacken 21, SE-106 91 Stockholm, Sweden. E-mail: email@example.com
Aims: The study was conducted with an aim to optimize the transformation efficiency of the Gram-positive bacterium Staphylococcus carnosus to a level that would enable the creation of cell surface displayed combinatorial protein libraries.
Methods and Results: We have thoroughly investigated a number of different parameters for: (i) the preparation of electrocompetent cells; (ii) the treatment of cells before electroporation; (iii) the electroporation step itself; and (iv) improved recovery of transformed cells. Furthermore, a method for heat-induced inactivation of the host cell restriction system was devised to allow efficient transformation of the staphylococci with DNA prepared from other species, such as Escherichia coli. Previously described protocols for S. carnosus, giving transformation frequencies of approximately 102 transformants per transformation could be improved to reproducible procedures giving around 106 transformants for a single electroporation event, using plasmid DNA prepared from either S. carnosus or E. coli. The transformed staphylococcal cells were analysed using flow cytometry to verify that the entire cell population retained the introduced plasmid DNA and expressed the recombinant protein in a functional form on the cell surface at the same level as the positive control population.
Conclusions: The results demonstrate that the transformation frequency for S. carnosus could be dramatically increased through optimization of the entire electroporation process, and that the restriction barrier for interspecies DNA transfer, could be inactivated by heat treatment of the cells prior to electroporation.
Significance and Impact of the Study: The generation of large combinatorial protein libraries, displayed on the surface of S. carnosus can be envisioned in the near future, thus dramatically improving the selection compared with the traditional biopanning procedure used in phage display.
Gram-positive bacteria are in general more resistant to plasmid DNA transformation than Gram-negative bacteria. This is mainly attributed to the structural differences in the cell surface barrier between the two cell types. The cell wall of Gram-negative bacteria consists of a dual membrane structure separated by a thin peptidoglycan layer. The Gram-positive cell wall is much thicker in comparison, with a single membrane and a massive exterior peptidoglycan layer, acting as a physical barrier for the DNA, and making efficient transformation of Gram-positive bacteria far more challenging.
Some applications demand a very high efficiency in the transformation step, one of which is combinatorial biochemistry, where large libraries of randomized DNA are introduced into cells in order to express the encoded proteins on the cell surface for subsequent screening for a specific function (Georgiou et al. 1997; Wittrup 2001). The probability to find a clone with the desired properties increase with the library size, and consequently the transformation step is an important parameter for successful generation of cell-displayed combinatorial protein libraries.
The Gram-positive bacterium Staphylococcus carnosus has been used for the display of proteins and peptides on the cell surface in a number of different applications (Cano et al. 2000; Lehtiöet al. 2001; Wernérus et al. 2001; Wernérus and Ståhl 2004), and at present, we are investigating its potential also for combinatorial engineering approaches (Wernérus and Ståhl 2002; Wernérus et al. 2003; Löfblom et al. 2005). The traditional method for in vitro combinatorial engineering of proteins and peptides is phage display (Smith 1985; Smith and Petrenko 1997), where large libraries of randomized peptides or proteins are displayed on the surface of phages. It is used in biopanning procedures for the selection of novel variants with desired properties. In recent years, alternative strategies have emerged, such as display on cells like bacteria (Georgiou et al. 1997; Daugherty et al. 1998; Bessette et al. 2004; Harvey et al. 2004) and yeast (Wittrup 2001; van den Beucken et al. 2003; Feldhaus et al. 2003), having the significant advantage of employing flow cytometry in the selection step, thus dramatically improving the selection, compared with the traditional biopanning procedure used in phage display. Gram-negative systems have hitherto been the bacterial approach of choice, primarily owing to the reasonably high transformation frequencies obtained on a routine basis, enabling the construction of large libraries containing 109–1010 library members (Vaughan et al. 1996). Still, thousands of transformations might be needed to create such libraries. Nevertheless, it would be interesting to create surface displayed libraries also for Gram-positive bacteria, because the thick cell wall makes them ideal for the harsh environment generated in high-speed flow cytometers. However, significantly improved transformation strategies are needed to make Gram-positive bacteria a competitive alternative. In addition, as our long-term objective is to develop a cell display selection system for affibody molecules (Nord et al. 1997), which are based on an engineered Staphylococcal aureus protein A domain, a staphylococcal display system would be a rational choice, as it is likely that a vast majority of the library members would be functionally expressed with a native fold on the cell surface.
To function in library display applications, the staphylococcal surface display system used here has been improved in two previous studies; firstly, by engineering of the display vector for increased genetic stability and improved growth properties (Wernérus and Ståhl 2002), and secondly, by the development of a flow cytometric surface expression-level normalization strategy to avoid biases, generated by cell-to-cell expression level variations, in the selection step (Löfblom et al. 2005). The capacity of the selection principle has so far been demonstrated using model libraries with no need for high-frequency transformation (Wernérus et al. 2003; Löfblom et al. 2005), and the positive results have encouraged this endeavor to optimize the transformation frequency to allow the generation of full-size protein libraries for true de novo selections.
The expression vector used for surface display on S. carnosus is a shuttle vector with one origin of replication for E. coli and one for replication in staphylococci, enabling more efficient library construction and cloning in E. coli followed by transformation of the E. coli-prepared DNA library to the host strain. However, transforming bacteria with DNA prepared from another strain is often problematic, owing to the differences in the methylation pattern resulting in the recognition of the introduced DNA as foreign material and its subsequent digestion by the host cell restriction machinery. Different strategies have been reported to overcome this restriction barrier in other bacteria, such as passing the DNA via a mutant strain defective in the restriction system (Yoshimoto et al. 1997); treating the DNA with cell extracts for in vitro methylation (Kwak et al. 2002); or temporarily knocking out the restriction enzymes in the host strain (Edwards et al. 1999).
Electroporation (Shigekawa and Dower 1988), in which a strong electric field pulse is applied to the cells in order to create temporary pores in the plasma membrane and thereby enable the DNA to cross the lipophilic layer via aqueous pathways, is the most widely used technique when high-frequency transformation of bacteria is required. The procedure can roughly be divided into four separate procedures: (i) preparation of electrocompetent cells; (ii) treatment of cells before electroporation; (iii) the electroporation step; and (iv) the recovery of transformed cells after electroporation. In this study, we have investigated parameters in all four stages to optimize the entire process for S. carnosus, and thereby obtain a transformation frequency suitable for the generation of cell-displayed combinatorial libraries. In addition, a method has been investigated to temporarily inactivate the host cell restriction machinery to enable efficient transformation of S. carnosus with DNA prepared from other species.
Materials and methods
Bacterial strains, plasmids and growth media
Staphylococcus carnosus strain TM300 (Augustin and Götz 1990) was used as bacterial host in the electroporation experiments, and for the preparation of plasmid DNA. Escherichia coli strain RR1ΔM15 (Rüther 1982) was used for the preparation of plasmid DNA. The plasmid vector pSCX:Zwt (Wernérus et al. 2003), derived from the surface display, shuttle vector pSCXm (Wernérus and Ståhl 2002), and containing an IgG-binding domain Zwt (Nilsson et al. 1987), was used in all electroporation experiments. Staphylococcus carnosus cells were grown in B2 medium (Augustin and Götz 1990; Schenk and Laddaga 1992) consisting of 1·0% casein hydrolysate, 2·5% yeast extract, 0·5% glucose, 2·5% NaCl and 0·1% K2HPO4 (pH 7·5). For recovery of transformed S. carnosus cells after electroporation, B2 medium or SMMP medium [7·5 parts of SMM (1-mol l−1 sucrose, 0·04-mol l−1 maleic acid and 0·04-mol l−1 MgCl2) and 2·5 parts of 7% antibiotic medium 3 (Difco, Detroit, MI, USA)] was used. Escherichia coli cells were grown in tryptic soy broth, supplemented with yeast extract (TSB + Y; Merck, Darmstadt, Germany).
Plasmid DNA prepared from E. coli was purified using the Jetstar Plasmid Maxiprep Kit (Genomed, Bad Oeynhausen, Germany). Plasmid DNA prepared from S. carnosus was purified using the QIAprep® Spin Miniprep Kit (Qiagen GmbH, Hilden, Germany), including an additional lysostaphin incubation step to degrade the peptidoglycan cell wall. The DNA was eluted or dissolved in sterile water to reduce the conductivity and minimize the risk of arcing caused by salt contaminants in the electroporation step.
Preparation of electrocompetent cells
The optimized protocol for the preparation of electrocompetent cells is as follows: an overnight culture of S. carnosus cells was diluted into 500-ml B2 medium to an optical density at 578 nm (OD578) of 0·5 and grown at 37°C and 110 rev min−1 to an OD578 of 0·6. The cells were incubated on ice for 15 min to stop growth, harvested by centrifugation (3000 g, 10 min and 4°C), and washed with 500, 250 and 80 ml of ice-cold sterile water. Subsequent washes were performed with 10 and 5-ml ice-cold 10% glycerol. The cell pellet was resuspended in 2 ml of ice-cold 10% glycerol to a final concentration of around 4 × 1010 cells ml−1. Finally, 60-μl aliquots of cells were immediately transferred for storage to a −80°C freezer.
The optimized electroporation protocol is as follows: electrocompetent S. carnosus cells were thawed on ice for 5 min, and incubated at room temperature for 30 min. The pSCX:Zwt plasmid, prepared from E. coli, was thawed and 1 μl (4 μg μl−1) was added to the cells, and the mixture was incubated at room temperature for 10 min prior to electroporation. After incubation with DNA, 50 μl of the cells were transferred to a 1-mm gap electroporation cuvette and electrotransformed with a Bio-Rad MicroPulser (Bio-Rad Laboratories, Hercules, CA, USA), using a field strength of 21 kV cm−1 and a pulse width of 1·1 ms. The complete electroporation procedure was performed at room temperature. Immediately after electroporation, 1-ml B2 medium was added to the cuvette, and the cells were transferred to a 14-ml tube and incubated at 37°C and 130 rev min−1 for 2 h. Finally, the cells were diluted 1/200 in B2 medium, spread on tryptose blood agar base (TBAB; Oxoid Ltd., Basingstoke, UK) plates with 10-μg ml−1 chloramphenicol, and incubated at 37°C for 48 h.
Heat-induced inactivation of the host restriction system
Plasmid DNA was prepared both from S. carnosus and E. coli, and electrotransformed into S. carnosus cells, as described earlier, under the Electroporation procedure section, to evaluate potential differences in transformation efficiency. Additionally, S. carnosus cells were heat-treated and electrotransformed by the same method as non-heat-treated cells to evaluate the effects of heat treatment on the transformation frequency. In order to determine the number of cells that survived during heat treatment and optimal time intervals for each temperature, a viability study was performed at 47, 50, 53 and 56°C. Electrocompetent S. carnosus was incubated at different temperatures, and the samples were removed from the heat block every minute during 9 min. Immediately after removal from the heat block, the samples were diluted in B2 medium, spread on TBAB plates without antibiotics, and incubated at 37°C for 24 h. The optimal temperature and time combination for transformation of electrocompetent S. carnosus was determined experimentally by heating the cells at 47, 50, 53 and 56°C for different time intervals with subsequent electrotransformation according to the electroporation protocol.
The optimized heat treatment protocol is as follows: after thawing of the electrocompetent cells, according to the electroporation protocol, the cells were collected by a quick centrifugation and incubated in a heat block at 56°C for 2 min. Immediately after heating, 500-μl electroporation buffer consisting of 0·5-mol l−1 sucrose and 10% glycerol was added to the cells. The cells were pelleted by centrifugation at 5000 g for 15 min and resuspended in 40-μl electroporation buffer. Subsequent procedures were performed according to the optimized electroporation protocol described earlier.
Flow cytometric analysis of electrotransformed cells
Two different cultures of S. carnosus containing the pSCX:Zwt plasmid (hereafter denoted Sc:Zwt) – one inoculated with a single colony of earlier verified Sc:Zwt and the other inoculated with all colonies from an electrotransformation as described earlier, were grown overnight in B2 medium with 10-μg ml−1 chloramphenicol, after which 1 ml from each culture was pelleted, washed with 1-ml phosphate buffered saline with 0·1% Pluronic® F108 NF surfactant (PBSP; BASF Corporation, Mount Olive, NJ, USA) and diluted to 107 cells ml−1. Approximately 106 cells were transferred to a new 1·5-ml Eppendorf tube. The cells were pelleted by centrifugation (3500 g, 4°C, 6 min) and resuspended in 1-ml PBSP containing 3·5-nmol l−1 biotinylated IgG, and incubated at room temperature with gentle mixing for 1 h. The cells were then washed with 1 ml of ice-cold PBSP, followed by incubation in 1-ml PBSP containing 1·25-μg ml−1 Streptavidin, R-phycoerythrin conjugate (Molecular Probes, Eugene, OR, USA) and 225-nmol l−1 Alexa Fluor® 488, human serum albumin (HSA) conjugate for 40 min on ice. After a final washing step in 1 ml of ice-cold PBSP, the cells were resuspended in 400 μl of ice-cold PBSP before analysis using a FACSVantage SE flow cytometer (BD Biosciences, San Jose, CA, USA).
In order to reach transformation frequency levels that are required for the construction of large cell surface displayed protein libraries, the various parameters in the transformation process of S. carnosus were investigated and evaluated. In this study, the transformation frequency is presented as transformants per transformation event and not per amount of DNA (CFU μg−1) added, as the former way is more relevant to the application than the latter expression, because minute amounts of transformed DNA typically give high (but not always scalable) transformation frequencies if expressed as transformants per microgram of DNA. The effects of all parameters investigated in the optimization experiments, both from the preparation of electrocompetent cells and the electroporation procedure, are presented in Table 1 together with short comments and references to the literature.
Table 1. Results from the optimization of parameters in the transformation procedure for Staphylococcus carnosus
PEG, polyethylene glycol; SMMP, medium comprising of sucrose, maleic acid and MgCl2.
*Effect on transformation frequency (CFU transformation−1): +, positive effect; −, negative effect; 0, no significant effect.
The growth, washing and storage conditions of the S. carnosus cells in the preparation process were potential targets in the optimization, and we investigated several different parameters at each stage (Table 1).
1Glycine-rich growth medium for weakening of the cell wall. There are a number of different methods reported to increase the permeability of competent cells in order to improve the transformation frequency, and the addition of glycine to the growth medium has been demonstrated to have a positive effect for a variety of Gram-positive species (Holo and Nes 1989; Dunny et al. 1991; Aukrust and Blom 1992; Buckley et al. 1999). The effect and the optimal concentration of glycine as a cell wall weakening agent are highly specific for different bacterial species. Glycine was added to the B2 medium at concentrations of 2%, 3%, 4%, 5%, 6%, 8% and 10%. Overnight cultures were diluted in glycine-rich medium to an OD578 of 0·5 and grown to an OD578 of 0·6 before harvesting. The results were negative, with a decrease in transformation frequency correlating with increasing glycine concentration (data not shown).
2Growth phase. In order to determine the effects of growth phase on transformation frequency, different cell concentrations were tested both at the dilution of overnight cultures step and at the harvesting step. Overnight cultures, at an OD578 of around 12, were diluted to an OD578 of 0·1, 0·5 or 0·6, and the cells were grown to an OD578 of 0·6 before harvesting. For the cell culture with a start OD578 of 0·6, cells were either harvested immediately after dilution or incubated for 40 min at 37°C, without any detectable cell growth. The results demonstrated an approximately twofold increase in transformation frequency, when the cells were grown from an OD578 of 0·5–0·6 or when they were diluted to an OD578 of 0·6, with 40 min of incubation, compared with the previously used OD578 of 0·1 (data not shown). On the other hand, harvesting the culture immediately after dilution to an OD578 of 0·6 resulted in an approximately 50-fold decrease in transformation frequency. The results indicate that the cell concentration at initiation of growth is not a critical factor for the transformation competence of the cells. However, it was found that the cells must be incubated for at least 40 min after dilution to allow a phenotypic change from the stationary phase in the overnight culture. In addition to an increased transformation frequency, starting the culture at a higher cell concentration resulted in a considerably faster preparation protocol for the competent cells. Growing the cells to an OD578 of 0·5, 0·7, 0·9 and 1·1 at harvest, gave similar transformation frequencies as for OD578 of 0·6, used in the initial protocol. However, the higher cell concentrations for OD578 of 0·7, 0·9 and 1·1, made the cell pellet more difficult to handle in the electroporation procedures and an OD578 of 0·6 at harvest was therefore considered to be the optimal concentration.
3Washing temperature for preparation of electrocompetent cells. Although previous studies with Staphylococcus aureus have demonstrated an increased transformation frequency for cells prepared at 20°C compared with those at 4°C (Augustin and Götz 1990; Schenk and Laddaga 1992), our experiments resulted in a different outcome. The washing procedure, including centrifugation steps and washing buffers, was performed either at 4°C or at room temperature to determine the effect on the transformation frequency. No difference in transformation frequency could be observed for washings performed at 4°C or at room temperature (data not shown). However, preparation at 4°C was preferred in the final protocol, because centrifugation at 4°C gave a more compact cell pellet.
4Washing solution. The effects of different washing buffers on the electrocompetence were evaluated by the addition of 0·5-mol l−1 sucrose to the 10% glycerol buffer to increase osmotic stability (Holo and Nes 1989), and by the addition of CaCl2 (5, 25, 50 and 100 mmol l−1) to the 10% glycerol buffer in order to weaken the cell wall (Suga and Hatakeyama 2003). A buffer of 10% glycerol was used for the final resuspension of the cells before freezing in the experiments with CaCl2. Furthermore, the freezing procedure was investigated by comparing a quick freezing method, using liquid nitrogen before storage at −80°C, to no-freezing step before storage. The addition of 0·5 -mol l−1 sucrose as an osmotic stabilizer to the 10% glycerol washing solution (Holo and Nes 1989) resulted in an approximately tenfold decrease in transformation frequency (data not shown). One possible explanation is that the sucrose decreases the cryoprotection of the competent cells during storage at −80°C compared with those stored in 10% glycerol. Addition of different amounts of CaCl2 to the washing solution in order to increase the permeability of the cells (Suga and Hatakeyama 2003) resulted in very few transformants and was not further investigated (data not shown). Storage at −80°C without freezing the cells in liquid nitrogen (Xue et al. 1999) gave a threefold higher electrocompetence, and in addition, the use and handling of liquid nitrogen was avoided.
The optimized preparation protocol has resulted in significantly higher electrocompetence of the S. carnosus cells – a faster protocol and also in a better reproducibility. Nonetheless, there are still batch-to-batch variations in the preparation process, making it difficult to evaluate the importance of parameters that give rise to positive, but small effects on the transformation frequency. The variations also have an impact on the optimization of the electroporation procedure, and comparisons between different batches of electrocompetent cells should be avoided, which makes it difficult to perform a true multifactorial analysis.
Several different parameters – both in the pretreatment, the actual electroporation step and the recovery of transformed cells, were independently tested and evaluated to study the effects on transformation frequency (Table 1).
1Different amounts of plasmid DNA – 0·1–6 μg, were added to the electroporation mixture, containing around 4 × 1010 cells ml−1, to determine the optimal DNA concentration for electroporation. The results in Fig. 1 show that the number of transformants increased with the amount of plasmid DNA that was added. Even though a higher transformation frequency was obtained with 6 μg of DNA (Fig. 1), 4 μg was preferred in the final optimized protocol, owing to the extra amount of time and work that was required to produce higher DNA amounts.
2Preincubation of cells with DNA. Preincubation of the cells with plasmid DNA has been reported to increase the transformation frequency in S. carnosus (Augustin and Götz 1990). This would, according to the previous study, indicate that binding of the DNA to the cell surface is necessary for efficient transformation (Augustin and Götz 1990). However, those results could not be verified in this study. Cells were incubated with DNA for 0, 10, 20 or 30 min, prior to electroporation without any change in transformation frequency (data not shown). This is an important aspect for scale-up attempts of the electroporation procedure for large library constructions, in which numerous electrotransformations are performed in parallel resulting in different incubation times for each sample. The obtained results demonstrate that variations in the incubation time within the range of 30 min can be allowed without any decrease in the transformation frequency (data not shown).
3Temperature during the electroporation process. The complete electroporation procedure was performed on ice or at room temperature to determine the effects of temperature on the transformation frequency. For electroporation on ice, all cells, cuvettes and solutions were kept on ice during the entire procedure. Thawing of the electrocompetent S. carnosus cells on ice for 5 min, followed by incubation for 30 min at room temperature, and subsequent handling and electroporation at room temperature resulted in an approximately fivefold increase in transformation frequency compared with cells treated on ice with precooled electroporation cuvettes (data not shown). This is consistent with results from Augustin and Götz (1990), where electroporation experiments with Staphylococcus epidermidis using cells and cuvettes at 20°C gave higher transformation frequencies than at 4°C. Similar observations have also been made for S. aureus by Schenk and Laddaga (1992).
4Electroporation buffer. Several different compounds were added to the original 10% glycerol electroporation buffer to determine the effects on the transformation frequency (Vehmaanperä 1989; Augustin and Götz 1990; Dunny et al. 1991; Aukrust and Blom 1992; Xue et al. 1999; Serror et al. 2002). The additives that were investigated and the results on the transformation frequency are shown in Table 1. Electroporation buffers with 0·5-mol l−1 sucrose and 10% glycerol proved to increase the transformation frequency approximately fivefold compared with the initial electroporation buffer, consisting of only 10% glycerol (data not shown). Similar results were obtained for 0·5- mol l−1 mannitol and 10% glycerol, whereas all other additives resulted in decreased transformation frequency (data not shown).
5Electrical parameters. The outcome of an electrotransformation process is to a large extent determined by the electrical parameters that are used in the actual electroporation step. The two main parameters are the field strength – which is the applied voltage divided by the gap size of the electroporation cuvette, and the pulse width – which is the duration of the electrical pulse. These two parameters must be optimized for different species and even strains, and for the chosen electroporation equipment. The Bio-Rad MicroPulser, used in this study, offers the ability to truncate the exponential electrical pulse after a predefined time. Hence, the duration of the pulse is denoted here as pulse width, in contrast to the time constant, which is defined as the time required for the voltage of the untruncated pulse to decline to 1 e−1 (∼37%) of the peak amplitude (Shigekawa and Dower 1988). The effects on transformation frequency for pulse widths between 1·0 and 2·0 ms and field strengths of 19, 21 and 23 kV cm−1 were evaluated. In all the cases, the maximal transformation frequency was obtained with a pulse width of 1·1 ms (Fig. 2a–c). Note that the experiments were carried out with different electrocompetent cell batches for each field strength, and therefore comparisons between Fig. 2a–c should be avoided. Pulse widths longer than 2·0 and 1·7 ms at 21 and 23 kV cm−1, respectively, caused frequent arcing, and were not further investigated. To compare the results between different field strengths, an electroporation experiment with the pulse widths for each field strength that gave the highest transformation frequency (1·1 ms; Fig. 2a–c) was performed, using competent S. carnosus cells from the same batch (Fig. 2d). A field strength of 21 or 23 kV cm−1, together with a pulse width of 1·1 ms gave the highest transformation frequencies (Fig. 2d). The field strength of 21 kV cm−1 was however chosen, owing to the lower risk of arcing, and the transformation frequency increased approximately fivefold compared with the initial parameters of 20 kV cm−1 and 2·5 ms.
6Recovery medium. Resuspension of cells in SMMP medium for recovery of transformed cells after electroporation has been suggested to increase the transformation frequency for S. aureus. In contrast, Schenk and Laddaga (1992) have reported that no increase in transformation frequency was obtained for S. aureus, when SMMP was used as recovery medium instead of B2 medium. In this study, B2 and SMMP media were both used in electroporation experiments to determine the effect on the transformation frequency of S. carnosus. The results showed an approximately fivefold increase in transformation frequency, when B2 medium was used for the recovery of the transformed cells instead of SMMP (data not shown).
Heat-induced inactivation of the host cell restriction system
Transformations with plasmid DNA that is prepared from other strains or species often result in significantly decreased transformation frequencies, because the restriction machinery recognizes the DNA as foreign, owing to the differences in the methylation pattern, and degrades it. Previous studies have reported that the problem can be avoided by the manipulation of the host cell restriction enzymes and Edwards et al. (1999) obtained an approximately 1000-fold increase in transformation frequency for interspecies DNA transfer between E. coli and Salmonella typhimurium/enteritidis, when cells were heat treated prior to electroporation. Edwards et al. (1999) suggested that a temporary heat-induced inactivation of the host cell restriction enzyme machinery was the mechanism behind this increase. As we are using a shuttle vector system and DNA prepared from E. coli, this method and hypothesis were tested and optimized in a number of experiments. Initial electroporation of electrocompetent S. carnosus cells with plasmid DNA prepared from E. coli or S. carnosus, respectively resulted in approximately 20-fold lower transformation frequency for the interspecies transfer (Fig. 3a). Interestingly, for heat-treated cells, the transformation frequency increased when using DNA prepared from E. coli to a level comparable with that obtained with the DNA prepared from S. carnosus (Fig. 3a). The results suggest that a heat-induced temporary inactivation of the restriction system is responsible for this increase, because a much smaller effect was observed for transformations with DNA prepared from S. carnosus.
Furthermore, a viability study was performed after heat treatment to investigate the temperature and time intervals at which a heat-treatment experiment for increased transformation frequency would be relevant. Electrocompetent S. carnosus cells were heat treated at 47, 50, 53 and 56°C for different time intervals, and the results demonstrated that a majority of the cells were dead after heat treatment at 56°C for more than 2 min, but on lowering the temperature, the viability was increased (Fig. 3b). Optimization of the heat treatment was performed to identify the time and temperature combination, giving the highest transformation frequency. Staphylococcus carnosus cells were heat treated at 47, 50, 53 and 56°C for different time intervals, and washed according to the earlier described heat-treatment protocol (Fig. 3c). To compare between different temperatures and to avoid batch-to-batch variations, an electroporation experiment using the incubation times giving the highest transformation frequency for each temperature (Fig. 3c) was performed (Fig. 3d). This comparative study showed that heating the electrocompetent cells at 56°C for 2 min resulted in the highest transformation frequency (Fig. 3d). A washing step after the heat treatment was introduced, and proved to be necessary in order to reduce the conductivity and prevent arcing in the electroporation step.
The novel heat-treatment protocol, together with the optimized protocols for the preparation of electrocompetent cells and the electroporation procedure yielded a transformation frequency of approximately 106 transformants per transformation when S. carnosus cells were electrotransformed with pSCX:Zwt plasmid DNA prepared from E. coli. As a summary, the most important modifications of the initial protocol are presented in Table 2.
Table 2. Initial and optimized procedures for transformation of Staphylococcus carnosus
SMMP, medium comprising of sucrose, maleic acid and MgCl2.
Preparation of electrocompetent cells
Growth from OD578 0·1–0·6
Growth from OD578 0·5–0·6
Thawing of cells
10 min on ice
5 min on ice, 30 min at room temperature
56°C for 2 min followed by a washing step
0·5-mol l−1 sucrose and 10% glycerol
Amount of plasmid DNA
Cells, cuvettes and medium on ice
Cells, cuvettes and medium at room temperature
20 kV for 2·5 ms (untruncated)
21 kV for 1·1 ms (truncated)
0·5 ml SMMP medium, 37°C, 48 h
1-ml B2 medium, 37°C, 48 h
Freezing of cells
Liquid nitrogen, −80°C freezer
Transformation frequency (CFU transformation−1)
Flow cytometric analysis of electrotransformed cells
To verify that all transformants obtained from the electroporation process expressed the Zwt and the albumin-binding protein (ABP) in a surface accessible and functional form, a flow cytometric assay was performed. Briefly, all colonies from an electroporation procedure were collected and used to inoculate an overnight culture containing antibiotics, and in parallel, a single colony of a previously verified S. carnosus strain transformed with the plasmid pSCX:Zwt was used to inoculate a second culture used as a positive control. Cells from the overnight cultures were washed and incubated with biotinylated IgG and thereafter labelled with streptavidin, R-phycoerythrin conjugate and Alexa Fluor® 488, HSA conjugate (Löfblom et al. 2005) to verify if the cells were expressing the full-length recombinant protein on the surface, and if both the IgG-binding and the HSA-binding parts present in the chimeric protein were functional (Fig. 4a). The labelled cell populations were then analysed and compared using flow cytometry. The results showed almost identical populations for the transformed cells and the control strain, when the fluorescence intensity corresponding to the IgG binding and the HSA binding was analysed (Fig. 4b,c). This demonstrated that a vast majority of all cells from the electroporation retained the plasmid, and expressed the recombinant protein in a functional and surface-accessible form.
In this study, we have optimized the electroporation protocol for the Gram-positive bacterium S. carnosus to yield approximately 106 transformants per transformation, which is sufficient for the construction of a large cell-displayed combinatorial protein library. All parts of the process – from the preparation of competent cells to the recovery of cells after transformation, have been investigated in order to achieve the improvements needed. For comparison, the optimized protocol and the initial protocol are shown in Table 2. To solve the problem of interspecies DNA transfer, a previously described method, where a short heat treatment of the competent cells is performed, in order to temporarily inactivate the host restriction system, was optimized and successfully employed to enable crossing of the species restriction barrier and obtain high frequency transformations of S. carnosus with DNA prepared from E. coli. The results demonstrated that it indeed is the manipulation of the restriction system that is the major contributor to the increased transformation frequency, as no significant increase could be seen when heat-treated competent S. carnosus cells were transformed with plasmid DNA prepared from S. carnosus. To verify if the transformed cells retained the plasmid and the surface expression of the complete and functional chimeric protein, a flow cytometric assay was performed using two proteins, labelled with fluorophores, directed at two different parts of the recombinant surface protein. The flow cytometric assay demonstrated that the transformed cells retained the surface expression of functionally recombinant proteins at a level almost identical to the previously verified positive control. The results in this study demonstrate that the transformation efficiency can be dramatically increased also for Gram-positive bacterial species when thorough optimizations are performed in the entire electroporation process.
To comment on whether 106 transformants per transformation are enough to generate functional combinatorial libraries, it is relevant to consider that the initial phage libraries used for functional selection of S. aureus protein A-derived affibody ligands had a size of only 4 × 107 variants (Nord et al. 1997). Today, larger affibody phage libraries are often in the order of 109 members (Wikman et al. 2004), and these are typically created by hundreds to thousands of transformation events. Such libraries could very well be envisioned for S. carnosus, as the entire transformation procedure would readily lend itself for automation (except for the preparation of electrocompetent cells, which could be done in bulk using scaled-up procedures). Finally, although results from electroporation optimizations obtained for a specific strain, generally, are difficult to directly apply to different species, the optimized process and the investigated parameters, should be of value in transformation optimization efforts of bacteria in general and Gram-positive species in particular.
The authors would like to thank Drs Bengt Guss, Lars Frykberg, Karin Jacobsson and late Prof Martin Lindberg for valuable advice. This work was financed with a grant (No. 621-2003-2876) from the Swedish Research Council (VR).