Effects of Sr^2 + on the preparation of Escherchia coli DH5α competent cells and plasmid transformation

Introduction

Escherichia coli is extensively applied in genetic engineering as a desired microorganism for various cloning experiments. In 1970, Mandel & Higa (1970) firstly found E. coli K12 strains treated with calcium chloride solution was easy to be infected by λ phage DNA by a short-time heat shock. Besides, it was proven that the plasmid DNA could also enter the bacteria by the same method by Cohen & Chang (1972). Thereafter, the chemical transformation methods (Nørgard, Keem & Monahan, 1978; Dagert & Ehrlich, 1979; Zhang, Xu & Xu, 2004) have continuously been modified with various combinations of chemical solutions including different cations (Mg²⁺, Ca²⁺, Sr²⁺, Ba²⁺, Mn²⁺, K⁺, Na⁺, Rb⁺), PEG, DMSO (Chung & Miller, 1988; Chan et al., 2013) and glycerin Shanehbandi et al. (2013). Considering the practicality and convenience of different artificial methods, more efficient exogenous gene transfer systems also have been developed by physical or chemical methods, including chemical procedures (Inoue, Nojima & Okayama, 1990; Sarkar, Choudhuri & Basu, 2002a; Sarkar, Choudhuri & Basu, 2002b; Song et al., 2007), high-voltage electroporation (Dower, Miller & Ragsdale, 1988; Dunny, Lee & LeBlanc, 1991; Sheng, Mancino & Birren, 1995), a biolistic propulsion system (Shark et al., 1991), liposome-mediated DNA transfer (Kawata, Yano & Kojima, 2003), microwaves and ultrasounds assisted DNA transfer (Zarnitsyn & Prausnitz, 2004; Fregel, Rodriguez & Cabrera, 2008; Tripp, Maza & Young, 2013; Deeks et al., 2014) and chemical-physical (Rb⁺, sepiolite and nanomaterials) induced transformation (Ren et al., 2017; Ren, Na & Yoo, 2018; Ren et al., 2019), etc. The transformation efficiency (10⁴∼10⁸ CFU/µg DNA) of different strains can obviously be obtained from these methods described.

Each of the methods mentioned above has their advantages in transformation efficiency, the corresponding experimental complexity and cost-effective limitations (Chen, Christie & Dubnau, 2005; Ren et al., 2017). However, the higher transformation efficiency in each chemical-mediate method all depended on the function of CaCl₂. This implied Ca²⁺ ions played the key role in the preparation and transformation process of the competent cells by plasmid, that can be affected by the growth of the bacterial strain, plasmid DNA concentration, heat shock temperature and duration, cold incubation duration of CaCl₂ treated cells (Singh et al., 2010). The various strains showed the different transformation efficiency with the same treatment method (Ren et al., 2017). In our previous study (Wang et al., 2013; Liu et al., 2014), the transformation efficiency with pUC19 plasmid DNA was almost equal between the cells prepared by CaCl₂ and SrCl₂ treatment, respectively. Whereas, how much concentration of Sr²⁺ resulted in maximum transformation efficiency compared with Ca²⁺ has not yet been reported up to now, and the exact mechanism of SrCl₂-mediated artificial transformation process is still largely obscure.

Chemical transformation was referred to as one of the most important techniques in genetic manipulation. The transformation efficiency depended upon several factors, including cation type, cation concentration, treatment time, thermal shock, and incubation time (Huff et al., 1990; Cosloy & Oishi, 1973; Chan et al., 2002; Broetto et al., 2006; Singh et al., 2010). Bacterial transformation is a fundamental technology to deliver engineered plasmids into bacterial cells, which is essential in industrial protein production, chemical production, etc. Hence, a novelty statistical method could be applied in the experimental design for higher transformation efficiency. Response surface methodology (RSM) consists of a group of mathematical and statistical methods, initiates from the design of experiment (DOE) that is generally considered an effective statistical technique for optimizing complex processes, due to its obvious advantages, e.g., the reduced number of experimental trials needed to evaluate multiple parameters and their interactions, less laborious and lesser time-consuming (Khuri & Mukhopadhyay, 2010; Bezerra et al., 2008). An optimal process with the mathematical equations is obtained by determining the significant factors affecting an experiment and reducing the number of experimental runs while maximizing output through the data generated. As one of the RSM, Box-Behnken Design (BBD) (Ferreira et al., 2007) is introduced to optimize the preparation of E. coli DH5 α prepared with SrCl₂ andthe transformation with pUC19. The optimal conditions are verified by different E. coli strains with different size of plasmid DNA, including pUC19, pET32a and pPIC9k, respectively. In addition, the transformation mechanism is preliminarily explored by SEM, EDS, AAS and FT-IR, respectively.

Material and Methods

Results

Effects of different divalent metal ions on transformation efficiency

Five strains of E. coli were treated by five kinds of divalent metal ions (100 mM) under the same condition and then transformed with 10.0 ng pUC19. The results shown in Fig. 1 displayed that the transformation efficiency of strains treated by Mg²⁺ or Mn²⁺ was lower than those prepared by Ca²⁺, Sr ²⁺ or Ba²⁺. The transformation efficiencies of E. coli BL21, DH5α, HB101 and JM109 induced by CaCl₂ were the highest. However, E. coli TG1 and TOP10 treatment with SrCl₂ obtained the highest transformation efficiencies. Differences between these strains maybe could take account for this phenomenon. Some receptor proteins existed in the surface of E. coli cells, such as outer membrane proteins (OmpA, OmpC and so on), responsing the signal of divalent metal ions, and then the receptor signalling pathways were triggered for the absorption of exogenous DNA (Aich et al., 2012; Finkel & Kolter, 2001). Moreover, the previous reports displayed RNase, DNase, β-galactosidase and alkaline phosphatase could be released on the cell surface or extracellular by osmotic shock (Nossal & Heppel, 1966). The different strains also showed different changes (permeability, enzyme activities) with the chemical treatment. The different enzymatic reactions in the cells may be activated or inhibited by the metal ions. For example, nuclease can hydrolyze DNA with the assistant of Mg²⁺ or Mn²⁺ to reduce the transformation efficiencies for different plasmids.

Effects of Sr²⁺ concentration on transformation efficiency

As we all know, divalent metal ions and their concentration play an important influence on the formation of competent cells and transformation efficiency for exogenous DNA. In this experiment, the effects of different Sr²⁺ concentrations on transformation efficiency were examined, while Heat-shock time and the amount of pUC19 plasmid DNA was fixed at 45 s and 10.0 ng. Plots of transformation efficiency against Sr²⁺ concentrations were shown in Fig. 2A, the transformation efficiency was significantly enhanced when the concentration of Sr²⁺ increased from 20 mM to 80 mM, and the value reached at a maximum at the concentration of 80 mM Sr²⁺. However, transformation efficiency kept stable between the Sr²⁺ treatment of 80 mM and 100 mM, and then decreased gradually when the amount of Sr²⁺ exceeded 100 mM.

Optimization of the parameters on transformation efficiency

According to the single-factor experiment results, three independent variables (SrCl₂ concentrations, heat-shock time, the amount of pUC19 plasmid DNA) with three levels (Listed in Table 2) were performed for developing the model. The transformation efficiency was regressed to a second-order polynomial function shown in the following equation: ${\displaystyle Y=-4.41\times 10^7+3.13\times 10^5{X}_1+4.92\times 10^5{X}_2+2.41\times 10^6{X}_3-355.17X_1{X}_2+1091.14X_1{X}_3+986.49X_2{X}_3-1737.21X_1^2-2661.05X_2^2-1.43\times 10^5{X}_3^2}$ where Y denoted transformation efficiency, X₁, X₂ and X₃ represented SrCl₂ concentrations (mM), heat-shock time (s) and the amount of pUC19 plasmid DNA (ng) in coded values, respectively. Table 3 showed the analysis of variance (ANOVA) for the fitted quadratic polynomial model. Based on this analysis, F-value of 71.59 and P-value < 0.01 indicated that the response surface quadratic model was significant difference. The ANOVA of the quadratic regression model showed that experimental data for transformation efficiency had a correlation coefficient (R²) of 0.9893 for the calculated model, which implied that the experimental results were agreement with the theoretical values predicted by the polynomial model. The model was suitable for application in the competence cells preparation and plasmid transformation. The Lack of Fit (F-value of 0.13) was not significant compared with pure error, which further indicated this model could accurately predict the preparation of competent cells and plasmid transformation. The data in Table 3 showed that all the linear coefficients (X₁, X₂) and quadratic term coefficients (X${}_1^2$, X${}_2^2$, X${}_3^2$) significantly affected the transformation efficiency (P <0.05). To further understand the relationship between the three independent variables in transformation efficiency, the three-dimensional response surface were plotted in Fig. 3. The interactive effects of SrCl₂ concentration and heat-shock time, DNA concentration and heat-shock time, and SrCl₂ concentration and DNA concentration on the transformation efficiency of E. coli DH5α with pUC19 plasmid were displayed in Figs. 3A–3C, respectively. It was observed that the transformation efficiency enhanced gradually up to a threshold level with the increase of Sr²⁺ concentration, as well as heat-shock time, the amount of pUC19 plasmid. Then, the transformation efficiency slightly decreased beyond this level. The maximum transformation efficiency in E. coli DH5α predicted by the model was 1.63 ×10⁶, when Sr²⁺ concentration, heat shock time, the amount of pUC19 plasmid DNA were set at 83.87 mM, 88.47 s, and 9.05 ng, respectively. Considering the operability in actual production, the optimal conditions could be modified as follows: Sr²⁺ concentration of 90.0 mM, heat-shock time of 90 s, the amount of pUC19 plasmid DNA of 9 ng. The average transformation efficiency obtained under these conditions was (1.76 ± 0.21) ×10⁶ (n = 3). These results confirmed that the response model was sufficient to optimize the preparation of E. coli DH5α competent cells and transformation for pUC19 plasmid. Thereafter, other 5 strains of E. coli were chosen to verify the optimal conditions with different size of plasmid including pUC19 (2.68 kb), pET32a (5.9 kb) and pPIC9k (9.27 kb). The results were shown in Fig. 4, the strain (E. coli DH5α and JM109) with a higher transformation efficiency was obtained, and the transformation efficiency decreased with the increase of molecular weight of the plasmid. A similar result was described by Chan et al. (2002).

Table 3:

Regression coefficients of the predicted quadratic polynomial model.

Sources	SSa	DFb	MSc	F value	Significance
Model	6.16 ×1012	9	6.85 ×1011	71.58562	<0.0001**
X1	5.31 ×1011	1	5.31 ×1011	55.45105	0.0001**
X2	8.27 ×1010	1	8.27 ×1010	8.642465	0.0217*
X3	3.43 ×109	1	3.43 ×109	0.358188	0.5684
X1X2	4.54 ×1010	1	4.54 ×1010	4.746211	0.0658
X1X3	7.62 ×109	1	7.62 ×109	0.796378	0.4018
X2X3	3.5 ×109	1	3.5 ×109	0.366159	0.5642
X${}_1^2$	2.03 ×1012	1	2.03 ×1012	212.4936	<0.0001**
X${}_2^2$	1.51 ×1012	1	1.51 ×1012	157.757	<0.0001**
X${}_3^2$	1.38 ×1012	1	1.38 ×1012	143.8017	<0.0001**
Residue	6.7 ×1010	7	9.57 ×109
Lack of fit	6.06 ×109	3	2.02 ×109	0.132688	0.9356
Pure error	6.09 ×1010	4	1.52 ×1010
Cor total	6.23 ×1012	16

DOI: 10.7717/peerj.9480/table-3

Notes:

Note: The data with “**” represented extremely significant difference (P <0.01); the data with“*” represented significant difference (0.01 <P <0.05), respectively. SS^a, sum of squares. DF^b, degree of freedom. MS^c, mean square.

Surface observation with SEM and determination of content of Sr²⁺ in the cell surface

SEM is a useful method for checking the higher resolution of cell surfaces (Li et al., 2004; Panja et al., 2008). The surface structure of E. coli DH5α competent cells treated with 90 mM SrCl₂ and E. coli DH5α cells were observed by SEM at a magnification of 10,000 and 50,000, as shown in Fig. 5. E. coli DH5α cells showed the typical rod-shaped bacteria with the length of 1.63–2.54 µm, a width of 0.63–0.9 µm (Fig. 5A), and the cells treated by Sr²⁺ kept rod-shaped characteristics with a typical length of 1.63–2.12 µm, a width of 0.43–0.81 µm (Fig. 5B). The cells treated with Sr²⁺ seemed to gather together, which attributed to the cell surface charge is changed. The cell’s shape and size almost remained unchanged between the two groups. However, the surface of the treatment cells appeared rough and wrinkled morphology (Fig. 5D) compared with that of normal cells (Fig. 5C). To check whether Sr²⁺ only attached to the cell surface, the contents of Sr²⁺ on the surface between the normal cells and competent cells were determined using an SEM-EDS as shown in Fig. 6. The contents of Sr²⁺ on the surface of competent cells was significantly higher than that of normal cells, which implied that a part of Sr²⁺ remained on the surface of the competent cells. The content of Sr²⁺ in the supernatant of E. coli DH5α treated with 90 mM of Sr²⁺ was measured by AAS. The content of Sr²⁺ concentration was measured 82.15 mM according to the calibration curve of SrCl₂ solution, which illustrated that a small part of Sr²⁺ may be entered into the cells or adsorbed on the surface of the cells. Based on the analysis results of EDS, we deduced Sr²⁺ indeed was adsorbed on the cell surface, and whether Sr²⁺ could enter into cells needs further research.

FT-IR analysis

Fourier transform infrared spectrum (FT-IR) could not only provide molecular groups characteristics of vibration absorption spectrum band but also detect the change of the molecular groups and the strain identification (Hellwig, Stolpe & Friedrich, 2004; Carlos et al., 2011). FT-IR of E. coli DH5α cells and competent cells were all shown in Fig. 7. A detailed analysis of FT-IR absorbance spectra of the strains was presented as follow: the peak at 1,060 cm⁻¹ was attributed to a sugar absorbance, which was mainly caused by the C-OH stretch or the P-O-C stretch form polysaccharide skeleton vibration. A spectral feature between 1,220 cm⁻¹ and 1,384 cm⁻¹ was assigned to the symmetric stretching vibration ν_s (PO${}_2^{-}$) and symmetric stretching vibration ν_as (PO${}_2^{-}$) of the phosphodiester group of nucleic acid molecules (Muntean et al., 2014). An obvious absorption peak at 1,456.82 cm⁻¹ was assigned to a stretching vibration of asymmetric bending vibration δ_s (CH₃) of methyl in a protein molecule. The peaks at 1,537.01 cm⁻¹ and 1,614 cm⁻¹ were regarded as the stretching vibration of C-N, C=O and the bending vibration of N-H from protein amide (Goormaghtigh, Cabiaux & Ruysschaert, 1994). The peaks at 2,919.46 cm ⁻¹ ν_as (CH₂) and 2,937.17 cm⁻¹ ν_as (CH₃) were the typical stretching vibration of C-H of aliphatic carbon chain, which reflected the essential feature of E. coli cell wall including the information of fatty acids, membrane proteins and other amphiphilic molecules (Muntean et al., 2014). A broad peak at near 3417 cm⁻¹ has been assigned to the stretching vibration of O-H, N-H from polysaccharides, fatty acids and proteins (Goormaghtigh, Cabiaux & Ruysschaert, 1994; Muntean et al., 2014). According to the above description, the cell displayed slightly different after the Sr²⁺ treatment (Fig. 7), which was consistent with the SEM observation. The differences of the relative intensity and the peak area mainly occurred at 1,537.01 cm ⁻¹, 1,384.66 cm⁻¹, 1,220.74 cm⁻¹ and 1,060.67 cm⁻¹ among three cells, which implied that the cell surface structure including phospholipids and membrane proteins. The lipopolysaccharide could generate some changes by the combination with Sr²⁺. The results were similar to those described in SEM-EDS and AAS analysis.

Discussions

The development of competent cells is one of the key techniques for introducing foreign DNA into cells. Many literature have reported the development of chemo-transformation, physical-assisted transformation, electrotransformation and other new nanocarrier-mediated transformation (Chan et al., 2013; Choi et al., 2013; Brito et al., 2017; Roychoudhury, Basu & Sengupta, 2009; Ren et al., 2019). Among these methods, the chemical artificial transformation was regarded as the simple and efficient bacterial transformation technique. E. coli X1776 was optimally transformed by pBR322 DNA and a higher efficiency was obtained under the following conditions: 100 mM C MgCl2aCl₂ in 5 mM Tris buffer (250 mM KCl and 5 mM MgCl₂, pH 7.6) (Nørgard, Keem & Monahan, 1978). Ren et al. (2017), described a chloride (RbCl)-based chemical-physical method that the best transformation efficiency for E.coli DH5α was obtained 4.3 ×10⁶ CFU/µg of pUC19 plasmid, in which glycerol (2.6%), MnCl₂ (2.5 M), potassium acetate (0.1 M) and CaCl₂ (100 mM) was used as transformation solution. Choi et al. (2013) also exhibited a new bacterial transformation method by magnesium and calcium aminoclays, and a transformation efficiency of up to 2 ×10⁵ CFU/µg pBBR122 in E. coli XL-1 and Streptococcus mutans (KCTC 3065). According to the abovementioned description, these divalent cations (Mg²⁺, Mn²⁺, Ba²⁺, Sr²⁺), especially Ca²⁺ ions played the most important roles in higher transformation efficiency and displayed different transformation efficiency for various strains. The important factors of competent cells preparation and transformation included the growth state of bacteria, kinds of the metal ions, the concentration of the metal ions, duration of ice bath, incubation time at 42 °C and concentration of plasmid DNA, which had significant effect on the transformation efficiency.

Hence, more efficient methods should be established and employed for multiple microorganisms. In this study, the preparation of E. coli DH5α competent cells treated with SrCl₂ and the transformation by heat-shock with pUC19 plasmid were optimized by RSM. An equation of the regression model and the optimized value for generation of competent cells and transformation was obtained and was verified by the other five E. coli strains with different size of plasmids. The transformants approximately reached to 10⁶ CFU/µg DNA, which was sufficient for the gene clone and expression. The ideal parameters to obtain maximum transformation efficiency in E. coli DH5α strain with pUC19 plasmid DNA was Sr²⁺ of 90 mM, heat-shock time of 90 s, the amount of pUC19 plasmid DNA of 9 ng. However, the precise details of transformation mechanisms remain vague (Aich et al., 2012; Claverys & Martin, 2003; Wang, Sun & Ma, 2016; Ren et al., 2019). Some researchers considered that cell membranes fluidity was weakened by using cations (e.g., Na⁺, Mg²⁺, Ca²⁺, Ba²⁺) and low temperature (ice-water bath). The phosphate groups and other negatively charged groups (hydroxyl and the carboxyl group) on the bacterial cell membrane could bind with these cations, that played a bridge to introduce the exogenous DNA to permeabilize the membrane to absorb plasmid DNA. Panja et al. summarized the cations form stable coordination compounds to promote DNA into the cells by eliminating electrostatic repulsion between DNA and lipo-polysaccharide (LPS) molecules on the outer membrane (Panja et al., 2006; Panja et al., 2008). As an amphiphile, DMSO or glycerin could further improve the efficiency, which might be due to the fact that amphiphiles promotes a better bond between the cation and the outer membrane, thereby facilitate DNA-membrane contact (Hanahan, 1983). While some other reports indicated the cations and cold-hot alternation in the process of transformation causes the cell membrane to transform into a more regular packing structure with a less fluid-like behavior. This solidified membrane may increase the permeability to DNA, which may be due to the complex forming of PHB/Ca²⁺/PPi leading to small pores in the membrane structure (Addison, Chu & Reusch, 2004; Huang & Reusch, 1995; Pavlov et al., 2005). Meanwhile, the pores could be enlarged by adding of other bioorganic compounds, such as 10% ethanol (Sarkar, Choudhuri & Basu, 2002a; Sarkar, Choudhuri & Basu, 2002b; Sharma, Singh & Gill, 2007) and cyclodextrins (Aachmann & Aune, 2009). It was recommended to incorporate more unsaturated lipids into the membrane at lower temperatures to maintain membrane fluidity (Aachmann & Aune, 2009). Since the membrane is more flexible, this seems to relieve DNA uptake during the shock and recovery phases. In this paper, we preliminarily explored the transformation mechanism by SrCl₂ method. The results showed that the surface between E. coli DH5α and competent cells treated with 90 mM SrCl₂ appeared rough and wrinkled and shrank by SEM analysis. A part of Sr²⁺ was adsorbed on the cell surface, which implied that Sr²⁺ played a role in foreign DNA entering into the competent cells by the interaction between Sr²⁺ andthe cell surface. Also, FT-IR differences of the relative intensity and the area of the peak between E. coli DH5α and competent cells mainly occurred at 1,537.01 cm⁻¹, 1,384.66 cm⁻¹, 1,220.74 cm⁻¹ and 1,060.67 cm⁻¹, respectively. These wavenumbers were related to the phosphate group, hydroxyl and carboxyl group involved in phospholipids, membrane proteins, and lipopolysaccharide (LPS) on the cell surface. The results further were verified by SEM-EDS and AAS, some Sr²⁺ ions were adsorbed on the cell surface, and others may be entered into the cells. However, whether Sr²⁺ could enter into the cell, and why the higher transformation efficiency must be achieved by divalent metal ions such as Mg²⁺, Ca²⁺, Sr²⁺, Ba²⁺ and Mn²⁺, rather than K⁺, Na⁺, Co²⁺, Zn²⁺ and Al³⁺, these questions need to be further explored. Although many studies have shown that membrane proteins and environmental stress factors played important roles in the formation of competent cells and transformation, the mechanisms are still not well known until now. Based on all of these phenomena and conclusions, we also speculate that the transformation is not only a kind of microbial physiological phenomenon but also a physicochemical and biochemical process (Yoshida & Sato, 2009; Aune & Aachmann, 2010), in which complex regulatory mechanisms maybe existed (Finkel & Kolter, 2001; Claverys & Martin, 2003; Aich et al., 2011).

Conclusion

In this study, E. coli competent cells were prepared by the SrCl²⁺ treatment, and the method was optimized by Response Surface Methodology (RSM). The differences between the normal cells and competent cells were analyzed by multispectral techniques, that revealed the minor changes occurred and strict regulatory mechanisms existed in E. coli cells.

Supplemental Information