High efficiency transformation of E.coli by high voltage electroporation

Volume 16 Number 13 1988 Nucleic Acids Research High efficiency transformation of E.coli by high voltage electroporation William J.Dower*, Jeff F.Miller2 and Charles W.Ragsdale' Molecular Biology Group and 1Electrophoresis Engineering Group, Bio-Rad Laboratories, 1414 Harbour Way So., Richmond, CA 94804 and 2Department of Medical Microbiology, Stanford Medical Center, Stanford, CA 94305, USA Received April 13, 1988; Revised and Accepted May 31, 1988 ABSTRACT E. coil can be transformed to extremely high efficiencies by subjecting a mixture of cells and DNA to brief but intense electrical fields of exponential decay waveform (electroporation). We have obtained 109 to 1010 transformants/jg with strains LE392 and DH5ox, and plasmids pUC18 and pBR329. The process is highly dependent on two characterstics of the electrical pulse: the electric field strength and the pulse length (RC time constant). The frequency of transformation is a linear function of the DNA concentration over at least six orders of magnitude; and the efficiency of transformation is a function of the cell concentration. Most of the surviving cells are competent with up to 80% transformed at high DNA concentration. The mechanism does not appear to include binding of the DNA to the cells prior to entry. Possible mechanisms are discussed and a simple procedure for the practical use of this technique is presented. INTRODUCTION The efficient introduction of DNA into bacteria is a phenomenon of great practical importance in molecular biology. Methods of chemical treatment yielding 108 to 109 transformants/4g of DNA with some strains of E. co#l have been described; but the preparation of such highly competent cells is affected by many factors and the level of competence obtained can vary considerably from batch to batch (1,2). Using a different approach, we have obtained transformation efficiencies exceeding those available with the best chemical methods. We have transformed several strains of E. colito efficiencies consistently in the range of 109 to 1010 transformants/4g of pBR and pUC plasmids by subjecting concentrated suspensions of cells and DNA to electrical fields of very high amplitude. Rendering eukaryotic cells permeable to nucleic acids by exposure to electrical fields is now a commonly used technique and is referred to as electroporation (3,4,5). Recently, intact bacteria of several species, both gram negative and gram positive, have been transformed to reasonable efficiencies with this technique (6, 7,8,9,10). Although a theoretical framework exists to explain this process (11) the actual mechanism of electrical field-induced DNA uptake is not understood. © I R L Press Limited, Oxford, England. Nucleic Acids ResearchVolume 16 Number 13 1988 6127 Nucleic Acids Research Our results with E. coi were obtained with single pulses of exponential decay waveform generated by a commercially available device. The procedure for preparing the cells and applying the pulse is simple and reproducible. We believe that electroporation provides a significant advance over chemical means for transforming many strains of E. coli and a variety of other bacterial species, and we propose the term "electro-transformation" for this technique. MATERIALS AND METHODS E. coli strains LE392 [F-, hsd R514 (rk -, mk+), sup E44, sup F58, lac Yi, gal k2, gal T22, met Bi trp R55, N-] and DH5o [080d lac ZA Mi5, end Al, rec Al, hsd Ri 7 (rk-,mk+), sup E44, thi-1, N -, gyr A, rel Al, F-, A(lac ZYA- arg F), Ul 69] were grown in L-broth (10 gm Bacto tryptone, 5 gm Bacto yeast extract, 5 gm NaCI per liter) with vigorous shaking at 370 to an ABS600 of 0.5 to 1 (taking cells still in mid-log growth phase). The cells were harvested by chilling the flasks briefly on ice and centrifuging at 4000 X gmax for 15 min at 40. Electroporation at the high voltages used in this study requires a cell suspension of very low conductivity. To achieve this, the ionic strength of the suspension was reduced by extensive washing as follows: The cells from a 1 liter culture were resuspended in 1 liter of either cold 1 mM HEPES (pH7), or water, centrifuged as above, resuspended in 0.5 liter of cold 1 mM HEPES, or water, centrifuged, resuspended in 20 ml of 10% glycerol, centrifuged, and finally resuspended in from 2 to 20 ml of 10% glycerol (a 50 to 500-fold concentration from the culture, depending on the experiment. See table 2 for high efficiency procedure). This concentrated suspension was distributed in small aliquots, frozen on dry ice, and stored at -700. DNJA pUC18 and pBR329 were prepared by alkaline lysis and double banded in CsCI. These preparations were about 90% supercoiled and 10% relaxed circles. The plasmids were resuspended in TE (10 mM tris-CI pH 8.0, 1 mM EDTA) and stored in aliquots frozen at -200. Because estimates of transformation efficiency are dependent on accurate DNA quantitation, the DNA concentration of these stocks was measured in two ways. First, we measured the ABS260 and calculated the concentration assuming a molar extinction coefficient of 1.3 x 104 L mol-1 cm-1 (1 ABS260 unit = 50 gg/ml). Agarose gel electrophoresis showed no contaminating, ethidium bromide-stained material that might contribute to the absorbance readings. Second, we linearized the plasmid DNA by restriction digestion, and compared the band intensity on an ethidium bromide-stained agarose gel with that of several other linear DNA species of known mass. These two methods were in agreement. 6128 Nucleic Acids Research For comparison, competent DH50x cells ("Hanahan cells") were obtained from BRL and transformed with various DNA samples exactly according to the instructions of the supplier. Electronics and electrodes The exponential decay pulses were generated by a Gene PulserTM apparatus (Bio-Rad Laboratories, Richmond, CA) set at 3 or 25 jF and from 0.2 to 2.5 kV. The output of the pulse generator was directed through a Pulse Controller unit (Bio-Rad) containing a high power, 20 Q resistor in series with the sample, and a selection of resistors of 100 to 1000 Q in parallel with the sample. The effective resistance placed in parallel with the electrodes is much lower than that of the sample, and determines the time constant of the pulse (for example, 200 Q with the 25 gF capacitor gives a 5 msec time constant). Electrode gaps of either 0.15 cm with a special "mini-electrode", or 0.2 cm with the small gap, Potter-type cuvette (Bio-Rad) were used. These electrode configurations provided field strengths of up to 16.7 kV/cm and 12.5 kV/cm. The circuit and electrodes are illustrated and described in more detail in the appendix. Transformation protocol The concentrated cells were thawed at room temperature, and placed on ice. 40 gl of cells were transferred to a cold, 1.5 ml polypropylene tube; 1 to 2 il of DNA solution (in a low ionic strength medium such as TE) was added to give a final concentration of from 10 pg/ml to 7.5 gg/ml, and the suspension was mixed vigorously by flicking the tube. The cell/DNA mixture was placed between the chilled electrodes, the electrode assembly or cuvette placed in the safety chamber, and the appropriate pulse applied. Following the pulse, the cells were immediately removed from the electrodes and mixed into 25 to 50 volumes of outgrowth medium (SOC: 2% Bacto tryptone, 0.5% Bacto yeast extract, 1OmM NaCI, 2.5mM KCI, 1OmM MgCI2 10 mM MgSO4 20 mM glucose) in a 17 x 100 mm polypropylene tube. The samples were incubated, with shaking at 225 RPM, for 1 hour at 370. At the end of this expression period, the cells were diluted appropriately in SOC and plated on L-agar containing either ampicillin (100 gg/ml), tetracycline (10 gg/ml), or chloramphenicol (34 jg/mI) to screen for transformants. Transformation efficiency was calculated as CFU/jig of plasmid DNA added. Dilutions were plated on non-selective L-agar to assess cell survival. Transformation frequency was calculated as transformants/survivors. RESULTS Our initial attempts to transform E. coli using field strengths of up to 6 kV/cm were modestly successful, producing 105 to 106 transformants/jg. With improved 6129 Nucleic Acids Research equipment we have evaluated some of the variables of bacterial electroporation; this has resulted in some understanding of the mechanism, and improved protocols yielding very high levels of transformation. Electrical variables The pulses we used were of exponential decay waveform generated by the discharge of a capacitor (see Appendix). These pulses are described by the peak voltage, V0, and the RC time constant, -r. V0 is the amplitude of the pulse, and T is a convenient expression of the pulse length (T (seconds) = R (ohms) x C (farads) = the time for V0 to decline to V0/e]. The voltage drop experienced by each cell depends on the size of the cell and the field strength, Eo = Vo/d, where d is the distance between the electrodes. As we change V, R, and C, we also change current, power, charge, and energy applied to the sample. These, in turn may lead to changes in other effects such as heating and hydrolysis. The effect of field strength and pulse length on transformation Our preliminary experiments showed that both the viability and transformability of the cells is very sensitive to the initial electric field strength of the pulses. Therefore, we compared the effectiveness of pulses with a wide range of field strengths and two time constants. The results are shown in Fig. 1: With pulses of 20 msec (panel A), maximum transformation occurs with a field of about 7 kV/cm; with shorter pulses of 5 msec (panel B and C), fields greater than 11 kV/cm are required to obtain maximum transformation. It is noteworthy that pulses of either 7 kV/cm and 20 msec or 11 kV/cm and 5 msec produce about the same level of transformation. Cell survival declines steadily with increasing field strength; and in each case shown, the maximum transformation efficiency is reached when 30 to 40% of the cells survive the pulse. The data in Fig. 1 also shows an effect of pulse length on transformation; we examined the role of pulse length in more detail. We varied the time constant from 0.4 to 18 msec by changing the size of the capacitor and the resistor in parallel with the sample. (When this resistor is much smaller than the resistance of the sample, it is the prmary determinant of the pulse length and provides a convenient means for varying the time constant obtained with any given capacitor.) We did this experiment at three field strengths and the results are shown in Fig.2. In the strongest field (16.7 kV/cm, panel A), pulses as short as 2.3 msec produced high levels of transformation. In the weaker fields (12.5 and 7.0 kV/cm, panels B and C) pulses considerably longer were required to produce similar levels of transformation. At each field strength, increasing the pulse length caused transformation to rise and cell survival to decline. In each case, maximum transformation is reached when about 50 to 75 % 6130 Nucleic Acids Research 2.0 . 100 A 1.5 75 1.0 50 0.5 25 Bx B 1.5 75 1.0 / 50 E 0.5 25 00 10 5 1 150 Field strength (kV/cm) Figure 1. Efct of field strength on transformation 4 pg of pBR329 DNA was added to 40 gl of LE392 cell suspension (2.5 x 1010 cells/ml) and placed between cold electrodes of (A and B) 0.2 cm or (C) 0.15 cm gap, and pulsed at field strengths of 1 to 15 kV/cm with time constants of (A) 20 msec or (B and C) 5 msec. The transformation efficiency (* ), and the percent of cells surviving the pulse (O) are displayed. The electrical conditions were, 200 to 2500 vofts with a 25 gF capacitor, 20 LI in series with the sample, and either (A) 1000 Q, or (B and C) 200 Ql in parallel with the sample. Transformants were selected on ampicillin. The limit of detection in this experiment was 10o6 transformants/li. 6131 Nucleic Acids Research 3.0 100 A 75 2.0 50 1.0 25 o . , I x5. B 0) ~~~~~~~752.0 co CD) 50 0 . _igure2.Efetfpuslntho tasfrato25n C~~~~~~~ 75 2.0- 50 1.0 25 0 5 1 0 1 5 2 0 RC Time Constant (msec) Figujre. Effect of pulse length on transformation 4 pg pBR329 DNA was added to 40 gl of a suspension of LE392 cells (2.5 x 1010 cells/ml) and placed between cold electrodes of (A) 0.15 cm or (B and C) 0.2 cm gap and pulsed with time constants of 0.4 to 18 msec at field strengths of (A) 16.7 kV/cm, (B) 12.5 kV/cm, or (C) 7.0 kV/cm. The transformation efficiency (U), and the percentage of cells surviving the pulse (0) are displayed. The electrical conditions were (A and B) 2500 volts, or (C) 1400 volts, with (A) 3 RF, or (B and C) 25 gF capacitor, 20 Q in series and 100 to 1000 Ql in parallel with the sample. Transformants were selected on ampicillin. The limit of detection in this experment was 106 transformants/jg. 6132 Nucleic Acids Research of the cells are killed. Further increases in pulse length caused still more cell death resulting in a decline in the recovery of transformants (Fig.2A,B). These experments demonstrate a compensatory relationship between the pulse amplitude and duration. Decreasing the field increased the length of the pulse required to maximally transform the cells; decreasing the pulse length increased the amplitude of the field required to maximally transform the cells, and each of these optimal combinations of field strength and pulse length produced similar efficiencies of transformation (2 to 3 x 1 09/jig) and cell death (50 to 7.h%). Our ability to compensate for lower field strength by increasing the pulse length is quite limited, however. Under the conditions described above, but with fields of only 2.0 kV/cm and very long pulses (time constants up to 900 msec), we have been unable to detect transformation above 106/gg. The effect of the concentration of DNA and cells on the recovery of transformants We examined the effect of DNA concentration on electro-transformation of E. coli. Table 1 shows the result of pulsing the same volume of cell suspension (40 g1) with quantities of pBR329 DNA of 0.4 pg to 0.3 jg (10 pg/mI to 7.5 jg/mI). The recovery of transformants increased linearly with DNA input (and DNA concentration) over this very wide range. (The transformation efficiencies for these points were constant, within experimental variation, with a mean (± SD) of 2.9 ± 1.2 x 1 09/jg.) In contrast, varying the DNA concentration by pulsing the same mass of DNA with different volumes of cell suspension gave the same yield of transformants in a manner independent of DNA concentration. In each case. however, the transformation frequency (proportion of cells transformed) was related to the DNA concentration. To illustrate this relationship we have replotted some of the data from Table 1 and this is .shown in Fig. 3. This indicates that under a given set of conditions, the DNA concentration determines the probability of any cell becoming transformed. In this case, the yield of transformants should increase with the number of cells present when the DNA concentration is held constant. To test this, we electroporated cells at concentrations of from 1.5 x 109/ml to 2.8 x 1010/ml in the presence of a fixed DNA concentration. The data (Fig. 4) show a steady increase in transformants recovered over this range of cell concentration. The effect of pre- and post-shock incubation of cells with DNA If binding of DNA to the cells is required for electro-transformation, increased time of incubation of DNA with the cells prior to pulsing might increase the level of transformation; conversely, increased incubation time might be detrimental for transformation if, for example, nucleass .are present in the cell suspension. We tested the effect of pre-shock incubation time and found very little difference over the range of 0.5 to 3'. min (Fig. 5a), indicating that a binding step may not be necessary. 6133 Nucleic Acids Research Table 1: The effect of the addition of DNA and cells on the recovery of transformants. DNA Vol [DNA] Transformants Efficiency Frequency (pg) (gil) (pg/mi) (trans/jg xi 0-9) (trans/survivor*) A 0.4 40 10 9.7 x 102 2.6 1.6 x 10-6 1.0 40 25 3.8 x 103 4.3 6.2 x 10-6 4.0 40 100 1.3x 104 3.6 2.1 x 10-5 10 40 250 2.4x 104 2.5 3.9 x 10-5 40 40 1.0 x 103 9.0 x 104 2.3 1.5 x 10-4 100 40 2.5 x 103 2.8 x 105 2.8 4.6 x 10-4 1 x 103 40 2.5 x 104 5.4 x 106 5.4 9.0 x 10-3 1 x104 40 2.5x105 3.6x107 3.6 6.0x10-2 1 x105 40 2.5x106 1.1 x108 1.2 1.8x10-1 3 x 105 40 7.5 x 106 4.8 x 108 1.6 7.8 x 10-1 B 4.0 40 100 2.8 x 103 0.7 9.3 x 10-6 4.0 100 40 4.2 x 103 1.1 3.6 x 10-6 4.0 200 20 3.4 x 103 0.8 1.8 x 10-6 C 40 40 1000 5.5x 104 1.4 1.8x 10-4 40 100 400 5.1 x 104 1.3 6.7x 10-5 40 200 200 5.4x 104 1.4 3.6x 10-5 See legend to Fig. 3 for experimental details. A. 0.4 pg to 0.3 gg in 40 gi of cell suspension. B. 4.0 pg of DNA in 40 to 200 p1 of cell suspension. C. 40 pg of DNA in 40 to 200 p1 of cell suspension. * Survival rate for all points was about 30 to 40 % The effect of the time of incubation following the pulse (but preceeding the expression period) was also examined. The data in Fig. 5b shows that the post-shock incubation time had a profound effect on transformation efficiency, which decreased about 3-fold in the first minute of post-pulse incubation, and continued to decline by more than 20-fold in 30 min. We now transfer the sample directly from the electrodes to the SOC expression medium as quickly as possible following the pulse. Survivor data shows that a delay in transfer to the SOC medium reduces cell 6134 Nucleic Acids Research z 0 0 U 0o -2 E 0 0 c -4 0 I- 0 2 4 6 8 log [DNA] (pg/ml) Figur Effec of DNA concentration on transformation frequency (U) From 0.4 pg to 0.3 9g of pBR329 DNA was added to 40 gl of LE392 cell suspension (3.6 x 1010 cells/ml); (0) 4 pg of pBR329 DNA was added to from 40 to 200 gl of an LE392 cell suspension (2.8 x 1010 cells/ml); and ( 0 ) 40 pg of pBR329 was added to from 40 to 200 p1 of LE392 cell suspension (2.8 x 1010 cells/ml). One pulse of 12.5 kV/cm, 5 msec was applied to each sample and the cells were suspended in SOC and incubated as usual. Transformants were selected on ampicillin and surviving cells were estimated by plating on L-agar. Transformation frequency is calculated as transformants/survivors. The electrical conditions were 25gF, 2.5 kV, 200 Q in parallel, 0.2 cm ciuvette. viability (data not shown). In general, our outgrowth protocol has been adapted from that used for chemically treated competent cells, and many of the same factors such as the use of SOC medium and shaking in large tubes also served to increase the efficiency of electro-transformation. High efficiency electro-transformation protocol: comparison of strains and plasmids By combining various improvements to the method, we have developed an optimized protocol for electro-transformation of E. coil (see Table 2). With this protocol we have compared strains LE392 and DH5o for the uptake of several plasmids. The results are shown in Table 3. The most efficient transformations were obtained with strain LE392, which yields up to 1010 transformants/gg of pUC18 with this protocol. The somewhat larger plasmid pBR329 consistently transforms LE392 to 2 to 4 x 109 transformants/JLg, depending on the antibiotic selection regimen. Strain DH5o also transforms to extremely high efficiencies with these plasmids, though usually to about half that of LE392. This difference may be due to the reduced viability of rec A cells (DH5o<). In these experiments the lower limit of detection was about 106 transformants/gg. At this level of sensitivity we saw no 6135 Nucleic Acids Research 3 a F/ I-l x 2 E 1 0) 0 co " 00 1 2 3 Cells/ml (x10 ', FigureA. Effect of cell concentration on transformation efficiency 4 pg pBR329 DNA was added to 40 gl of a suspension of LE392 cells at a concentration of 1.5 x 109 to 2.8 x 1010 cells/ml in a cold 0.2 cm cuvette. The cells were pulsed at 12.5 kV/cm with a time co