Gene, 96 (1990) 23-28

Elsevier

23

GENE 03769

High efficiency transformation of E s c h e r i c h i a coli with plasmids

(Competent cell; strain DH5; pBR322 vector; simple and efficient method (SEM); cDNA library; electroporation; frozenstock; stability)

Hiroaki Inoue', Hiroshi Nojima b and Hiroto 0kayama b

a Research Department, Turuga Enzyme Plant, Toyob~ Co, Ltd., Turuga City, Fukui-ken 914 (Japan) Tel. (0770)22 7643, and b Department
of Molecular Genetics, Research Institute for Microbial .Oisease~, Osaka University, Yamadaoka 3-!, Suita City, Osaka 565 (Japan)
Received by H. Yoshikawa: 26 April 1990
Revised: 4 June 1990
Accepted: 15 June 1990

SUMMARY

We have re-evaluated the conditions for prepmmg competent Escherichia coil cells and established a simple and efficient
method (SEM)for plasmid transfection. Cells (DH5, JMI09 and HB101) prepared by SEM are extremely competent for
transformation (1-3 x 109 efu//~g of pBR322 DNA), and can be stored in liquid nitrogen for at least 40 days without loss
of competence. Unlike electroporation, transformation using these competent cells is affected minimally by salts in DNA
preparation. These competent cells are particularly useful for construction of high-complexity cDNA libraries with a minimum
expenditure of mRNA.

INTRODUCTION

Recent advances in recombinant DNA techniques have

made it possible to isolate a eDNA copy of a gene by
complementation of a particular defect in cells without
information of its gene product. The efficacy ofthis strategy,
which is called 'expression cloning' or 'complementation
Correspondenceto: Dr. H. Okayama, Department of Molecular Genetics,
Research Institute for Microbial Diseases, Osaka University,
Yamadaoka 3-1, Suita City, Osaka 565 (Japan) Te!.(06)8753980;
Fax (06) 875 5192.
Abbreviations: Ap, ampicillin; BES, N,N-bis(2-hydroxyethyl)-2-aminoethanesulfonic acid; bp; base pair(s); cfu, colony-forming unit(s);
DMSO, dimethyl sulfoxide; FTB, freeze-thaw buffer (see section c and
Fig. 1 legend); Hepes, N-(2-hydroxymethyl)piperazine-N'-2-ethanesulfonic acid; kb, kilobase(s) or 1000 bp; LB, Luria-Bertani (medium);
MOPS, 3-(N-morpholino)propanesulfonic acid; nt, nucleotide(s); Pipes,
piperazine-N,N'-bis(2-ethanesulfonic acid); SEM, simple and efficient
method for transformation; SOB; 2% Bacto tryptone/0.5% yeast
extract/10 mM NaCI/2.5 mM KCI/10 mM MgCI2/10 mM MgSO4; SOC,
SOB with 20 mM glucose; TB, transformation buffer (see section a); TE,
10 mM Tris. HCI (pH7.5)/1 mM EDTA.
0378-1119/90/$03.50 1990ElsevierScience Publishers B.V.(BiomedicalDivision)

cloning', largely depends on the quality of the eDNA library

used for the experiments. We have hitherto developed a
vector system that permits construction of full-length
eDNA libraries and subsequent screening based on phenotypic expression of the eDNA inserts in mammalian cells
(Okayama and Berg, 1982; 1983; Okayama et al., 1987;
Chen and Okayama, 1987). To screen for very rare eDNA
clones, a high complexity eDNA library consisting of at
least 5-10 x 106 independent clones often needs to be ~tmstructed with a limited amount of mRNA. This task could
not be achieved if highly competent E. coli cells were not
available. We have used the protocol described by Hanahan
(1983) with a slight modification (Okayama et al., 1987).
This Hanahan protocol is one of the best chemical methods
presently available) whichwas developed to attain the maximum transformation effieiencies after a great deal of studies
(Taketo, 1972; Taketo and Kuno, 1974; Lederberg, 1974;
Kretschmer et al., 1975; Enea et al., 1975: Dagert and
Ehrlich, 1979; Morrison, 1979; Jones etal., 1981;
Bergmans et al., 1981) since the first demonstration of
C a 2 + -dependent DNA transfer into E. coli by Mandel and
Higa (1970). However, preparation of cells with a con-

24
sistently high transformation efficiency is difficult in general.
Moreover, frozen competent cells often rapidly lose their
competence during storage, particularly when the final ligation product in the process of eDNA-library construction
is used for transfection. To circumvent these problems, we
systematically optimized each condition and parameter for
the prepacation of E. coli competent cells.
The aim of the present study was to develop SEM for the
preparation of such cells with an extremely high transforming frequency (1-3 x 10 9 cfu//2g of plasmid DNA).

RESULT AND DISCUSSION

(a) Standard protocol for the preparation of competent cells

by SEM
Media (SOB, SOC and LB) were prepared as described
by Hanahan (1983). For selection of transformed E. coli,
LB plates containing 50/~g Ap/ml were used. To make TB
(10 mM Pipes/55 mM MnCI2/15 mM CaC12/250mM
KCI), all the components except for MnCI 2 were mixed and
the pH was adjusted to 6.7 with KOH. Then, MnCI 2 was
dissolved, the solution was sterilized by filtration through a
preripsed 0.45 pm filter unit and stored at 4C. All salts
were added as solids.
Frozen stock (in LB/7~o DMSO) DH5 cells were
thawed, streaked on an LB agar plate, and cultured overnight at 37C. About ten to twelve large (2-3 mm in
diameter) colonies were isolated with a plastic loop, inoculated to 250 ml of SOB medium in a 2-liter flask, and grown
to an A6o o of 0.6 at 18C, with vigorous shaking
(200-250 rpm). The flask was removed from the incubator
and placed on ice for 10 min. The culture was transferred
to a 500ml centrifuge bottle and spun at 2500 x g
3~00 rpm in Beckmann J-6B centrifuge) for 10 min at 4C.
The pellet was resuspended in 80ml of ice-cold TB,
incubated in an ice bath for 10 min, and spun down as
above. The cell pellet was gently resuspended in 20 ml of
TB, and DMSO was added with gentle swirling to a final
concentration of 7~o. After incubating in an ice bath for
10 min, the cell suspension was dispensed by 1-2 ml into
tissue-culture cell-freezing tubes and immediately chilled by
immersion in liquid nitrogen. The frozen competent cells
were stored in liquid nitrogen for at least one month without
a detectable loss of competence.
(b) Standard protocol for SEM transformation
A tube of competent cells was thawed at room temperature, dispensed by 200/~1 into 15 ml polypropylene tubes
(Greiner, F.R.G.) and placed in an ice bath. Glass tubes
should not be used since they largely decrease (tenfold less)
tlae competence. Usually, 1-5/~1 of the plasmid zeMtion
was added to each tube, and the cells were incubated in an

ice bath for 30 min. They were then heat-pulsed without

agitation at 42C for 30 s and transferred to an ice bath.
After 0.8 ml of SOC was added, the tubes were placed in
a 37 C incubator and shaken vigorously for 1 h. A desired
portion of the mixture was tranferred into a 5 ml polypropylene or glass tube, mixed with 3 ml of melted LB soft
agar preincubated at 47C, and poured on LB plates
containing 50 #g Ap/ml. Colonies were counted after overnight incubation at 37 oC.

(c) Optimization of transformation buffer

To be well suited for construction of high-complexity
libraries, competent cells must meet two requirements:
(i)high-frequency transformation with ligated DNA;
(ii) ability to be stored frozen in liquid nitrogen without loss
of competence. To find the best conditions for preparing
competent cells, we began by optimizing FTB (Hanahan,
1983). To simplify our strategy, we first restricted the plasmid DNA and the E. coli strain to pBR322 and DH5,
respectively. The E. coli DH5 strain (supE44 hsdRl7
(r~ ,m~ ) recAI endA 1 gyrA96 thi-I relA I) appears to be
suitable as a host for generating cDNA libraries, since
plasmids containing homologous sequences are stable in
this strain owing to its recA 1 genotype. Moreover, this
strain is efficiently transformed with large plasmids because
it is defective in endonuclease I. These properties are
important when full-length cDNA libraries containing more
than 6-7 kb inserts are to be constructed.
First, we examined the concentration of the components
and pH values of FTB without significant improvements.
Second, we examined the amount ofglycerol added to FTB
and found that, in our hands, glycerol (a glass-distilled
grade from BRL) in FTB was highly inhibitory to competence of DH5 (Fig. la). In the course of this investigation,
we found that all salts, e~xcept KCI, CaCI2, MnCI2 and
K.acetate, could be removed from FTB. Salts such as
hexamine-Co chloride, COSO4, LiCI and RbCl were found
to be inhibitory rather than stimulatory to transformation.
Therefore, they were removed from the transformation
buffer. The pH value of the transformation buffer which has
little influence on transformation efficiency in a range of
5.6-7.0, could not be raised more than 7.0 because of
sedimentation of manganic salts at this a~d higher pH
values. Hence, we chose pH 6.7.
Fig. lb shows that among buffers tested, Pipes yielded
the highest competence. Therefore, K. acetate was replaced
by Pipes or Hepes. After optimizing the pH (Fig. 2a) and
salt concentrations (Fig. 2b-d) of this solution, we decided
to use TB (see section a) as our standard transformation
buffer. The optimum concentration of these omponents
remained the same when 10 mM Pipes (pH 6.7) was
replaced by 10 mM Hepes (pH 6.7). Competent cells prepareG ".y the standard transfection protocol (see section a;

25
1010

P-

u.

lO 9

v-

g~

07

F-

106
0

10

20

Glycerol

(%)

Fig. I. Effect of glycerol and buffers on transformation frequency (cfu/

/Jg of pBR322 DNA). (a)Effect of glycerol included in the FTB (Hanahan, 1983; 10 mM K.acetate/45 mM MnCl2/10 mM CAC12/3 mM
[Co(NH~)2]CI3 pH6.4). Competent cells were stored for I h in liquid
nitrogen. (h) Difference in transformation frequency when 10raM
K. acetate (bar 1) in the modified FTB (see section e) was replaced by
10 mM Hepes (bar 2), 10 mM BES (bar 3), 10 mM MOPS (bar 4) or
10 mM Pipes (bar 5). In both experiments, cells (DH5) were grown at
18"C and harvested at an A6oo of 0.37. Each value for competence was
an average of duplicate measurements.

200/tl of competent cells transfected with 10 pg pBR322

DNA), yielded a transformation frequency of 1-3 x
109 cfu//zg pBR322. This value corresponds to about one
transformed cell per 70-200 plasmid DNA molecules.
(d) Optimization of cell growth

After testing various concentrations of the components

of the medium for E. coil growth, we found that SOB and

SOC media (Hanahan, 1983) were suitable for obtaining

optimal efficiency for our method. The optimum temperature used for cell growth, however, was very different from
Hanahan's (1983). As shown in Fig. 3, incubation at 18C,
although it might take two days to reach a proper cell
concentration, gave high competence in a wide range of cell
concentrations at harvest. The optimum cell density was an
A6o o of 0.75. On the other hand, incubation at 37C as
described by Hanahan (1983) resulted in both a decrease in
transformation frequency and a sharp dependence of competence on cell density. The optimum cell density at this
temperature was an A6oo of 0.45 (Hanahan, 1983). Incubation at room temperature (about 25 C) also gave results
similar to incubation at 18 C, but with a slight decrease in
competence (data not shown). Incubation at a temperature
lower than 15C is impractical since growth is too slow.
Use ofFTB (Hanahan, 1983), however, resulted in only low
competences (2 x 107-2 x 10s cfu/#g of pBR322) at any
concentration of cells even after incubation at 18C.
(e) Optimization of transfection steps

In our standard transformation protocol, frozen competent cells were thawed at room temperature and transferred to an ice bath as soon as thawing was completed.
Competent cells (0.2 ml) have been dispensed into polypropylene of polystyrene tubes (Greiner, F.R.G.). The
pBR322 DNA (10 pg each in 1 #1 of TE) was then added
to the competent cells and the mixture incubated in an ice
bath for 30 min. The volume of DNA solution (up to 20 #1)
had little influence on transformation frequency (data not
shown). The number of transformants increased in proportion to the amount of DNA up to 10 ng. Afterwards,
10 l o

l.,.f-l-lr,.,,.,I

mm~m~ m

I:
S~IO 9

~,

fm- ".,~I,~i
%mm

~ "m\

g[ ~a lO9

lf~i

.m,

~m

"_/ A

-o~

l=Q-

qa

,)

Q.

h.

. ~

100

.m
l--

1o7

s.o

6.0
pH

7.0

lO

"

20

PIPES(mM)

10 7

20 4o eo
MnCI2(mM)

o 12oo3~4ooo
K C I (raM)

lo 20 ~o 4o
CaCl=(mM)

o.o

o.s

1.~
A soo

Fig. 2. Influences ofpH (a) and concentrations of TB components (b-e) on transformation frequency(cfu/#gpBR322 DNA). One of the parameters in
TB (see section a) was varied. The cells DH5 were grown at 18C and harvested at an A6ooof 0A2. Transformation frequencyat each point was obtained
by duplicate measurements.
Fig. 3. Effectof cell concentration at harvest and growth temperature on transformation frequency(cfu/#g of pBR322 DNA). Cells were grown at 37C
(D) or 18C (m), treated with standard TB and transfected following the standard protocol (see section a). Each point denotes an average value of
duplicate measurements.

26
however, the ratio of transformants to the amount of used
D N A decreased rapidly (data not shown).
To optimize heat-shock conditions, transfected cell suspensions were subjected to a heat pulse at 40 , 42 and
44C for 0-90 s. As shown in Fig. 4a, the highest number
of transformants was yielded at 30 s. The temperature for
the heat pulse was not critical, and even after heat shock at
an extremely high temperature (52C) for 30 s, a similar
transformation efficiency was obtained (Fig. 4b).

1010

>,
0e.
~'~

109

e- Q.

'3~

'

'

le

I-

(f) Freezing competent cells and their stability

The benefit of freezing competent cells is twofold.
Freezing increases the competence, and storage of frozen
competent cells permits knowing their competence prior to
use for precious transformation such as library construetion. D M S O is most widely used as a stabilizer of frozen
cells and worked well for our competent cells. The optimum
concentration of D M S O was 7~o for the cold shock and
storage of the competent cells (Fig. 5a). Freezing in liquid
nitrogen (cold shock) enhanced the transformation efficiency four- to fivefold. Frozen competent cells in the presence of 7% DMSO were stable. As shown in Fig. 5b, cells
retained their original conipetence even after a 40 days'
storage in liquid nitrogen. Oxidation products of D M S O
did not seem to be critical to our competent cells. We used
D M S O from a bottle that had been opened many times and
stored at room temperatu,e without a detectable decrease
in effectiveness. D M S O from commercial sources other
than MCB (optical grade) gave similar good results.

10 7

~ I

10

"

15

10

DMSO (%)

20

30

40

Days

Fig. 5. EffectsofDMSO and stabilityofcompetence. (a) Transfor~la~ion

efficiency as a function of DMSO concentration. Transformation frequency was measured after thawing competent cells prepared by the
standard transformation protocol except that the DMSO concentration
was varied (see section b). (b) Stabilityof competent cells after long-term
storage in liquid nitrogen. Cells were grown at 18C and harvested at an
A6oo of 0.60. Each point indicates the average value of duplicate experiments.
(g) Effect of plasmid size
The influence of plasmid D N A size on the transformation frequency of our competent cells was assessed by
tra~sfecting plasmid DNA whose size ranged from
4.4-17.6 kb. As shown in Fig. 6, the transformation frequency is almost e,,ual for plasmids 4.4-7.3 kb in size on
a molar basis. When plasmids with sizes of 10.3 kb, 13.2 kb
and 17.6 kb were used, relative transformation frequency
decreased to 46~o, 26% and 15% of that of pBR322
(4.4 kb). Thus, plasmids up to 10 kb long are able to trans-

10 10

(%)

'!
c

Z~'A"Z'-,,,)[

0o

L
C

60

4o

a,

ii
o~

"#"'J'-

lOO
J
o

10 9

11,

\.

.m
I-.

20

10 8

20

"

40

"

61~

"

"

80

Time (second)

30

4~)

50

60

Temperature (c)

Fig. 4. Effectsof time (a) and temperature (b) of heat shock on transformation frequency.Cells(DHS) were grown at 18C, harvested at an A60o
of 0.6I and treated with standard transformation buffer. Effect of time
was assessed at three different temperatures; 40C (1), 42C (1"1)and
44C (O). Average values of duplicate measurements were plotted.

5.0

10.0

15.0

20.0

Plasmid size (kb)

Fig. 6. Effectof plasmid size on transformation frequency. The plasmid
DNAs with the size of 4.4 kb (pBR322), 6.0 kb, 7.3 kb, 10.3kb, 13.2kb
and 17.6kb were transfected to the same lot of competent cell. The
ordinate denotes the relative value of transformation frequency (per
molecule of each plasmid DNA) when the value ofpBR322 was taken as
100%. Average values of duplicate measurements were plotted.

27
form this competent cell within a twofold difference in
frequency. Competent cells with this size limitation are
generally considered sufficient for constructing eDNA
libraries. Since the majority of mRNA ranges from 1-6 kb,
plasmid vectors of up to 4 kb such as the Okayama-Berg
vector can accomodate 6 k b cDNA inse~*.s without
jeopardizing transformation frequency. Hence, use of competent cells prepared by this method is expected to yield a
cDNA library representing the original population of
mRNA

(h) Comparison with electro-transformation

Transformation of E. coli with high-voltage electroporation has been described by Dower et al. (1988). They
claimed that it yielded 109-101 transformants/#g plasmid
DNA. We tested this method for comparison. It gave a
transformation frequency of 1-2 x 109 efu/#g of pBR322
(DNA in TE), a value comparable to the frequency of our
chemical method. The major disadvantage of the electrotransformation method, however, is that transformation
frequency is greatly influenced by salts and other chemicals
generally contained in a ligated DNA solution (Table I).
Therefore, their removal prior to electroporation is essential
for obtaining the best results. This step is cumbersome, and
tends to lose DNA during removal of the interfering chemicals, making eleetroporation less desirable for such purposes as construction of high-complexity eDNA libraries.
(i) Competence of other Escherichia coil strains
The applicability of this method to other E. coli strains
was assessed by testing HBI01 (Boyer and RoulandTABLE I

Effect of salt concentrations on transformation frequency of competent

cells prepared by our chemical method (SEM) and electroporation

TE a
TE/3 e
TE/9 f

SEM a
Transformation
frequency

Electroporation b
(cfu/#g of pBR322 DNA) c

1.6 109
1.5 109
1.4 x 109

1.1 109
<1 x 10TM
-h

a Described in sections a and b and Fig. 1.

b Dower et al. (1988).
Obtained in duplicate measurements.
d pBR322 DNA (1 pg) in 1 #1 of TE was added to 200 #1 of competent
cells.
e TE + 3 #1 of the reaction cocktail used in the cyclization procedure
(Okayama et al., 1987; 30 mM Tris. HCI pH 7.5,'100 mM NaCI/100 mM
KCI/10 mM (NH4)2SO4/4 mM MgCIa/I mM EDTA/5% (w/v) bovine
serum albumin) was added to 200 #1 of competent cells.
f TE + 9 #1 of the same reaction cocktail was added to 200 #1 of competent cells.
g Very low value and h no transformation because of sparking during the
application of high voltage to the prewashed cells (Dower et al., 1988).

Dussoix, 1969) and JMI09 (Messing et al., 1981) strains.

The competences of these E. cob strains were (1.84 +
0.85) x 109 cfu (average of five preparations) for DH5,
(1.44 _+ 0.65) x 109 cfu (average of four preparations) for
JM109, and (1.00 + 0.18) x 10 9 cfu (average of three
preparations) for HB101 per #g of pBR322 DNA. The
maximum competence we have hitherto obtained was
2.8 x 109 cfu/#g of pBR322 DNA for DH5.

(j) Conclusions
We have described a protocol for preparing competent
E. coli cells which are extremely efficient for plasmid transfection (1-3 x 109 cfu/#g pBR322 DNA) and can be stored
in liquid nitrogen for more than a month. These competent
cells are best suited for the construction of eDNA libraries.
Using these competent cells, we have constructed a eDNA
library with more than 107 independent colonies with as
little as 1 #g of mRNA. Our method is simple, reliable and
highly reproducible. It works for at least three E. coli
strains (DHS, JMI09 and HB101), and possibly for any
other E. coil strains.
REFERENCES
Bergmans, H.E.N., Van Die, I.M. and Hoekstra, W.P.M.: Transformation in Escherichia coil: stages in the process. J. Bacteriol. 146
(1981) 564-570.
Boyer, H.W. and Rouland-Dussoix, D.: A complementation analysis of
the restriction and modification of DNA in Escherichia coll. J. Mol.
Biol. 41 (1969) 459-472.
Chert, C. and Okayama; H.: High-efficiency transformation of mammalian cells by plasmid DNA. Mol. Cell. Biol. 7 (1987) 2745-2752.
Dagert, M. and Ehrlich, S.D.: Prolonged incubation in calcium chloride
improves the competence of Escherichia coil cells. Gene 6 (1979)
23-28.
Dower, W.J., Miller, J.F. and Ragsdale, C.W.: High efficiency oftransformation of Escherichia coil by high voltage electroporation. Nucleic
Acids Res. 16 (1988) 6127-6145.
Enea~, V., Volvis, G.F. and Zinder, N.D.: Genetic studies with heteroduplex DNA of bacteriophage fl. Asymmetric segregation, base correction and implications for the mechanism of genetic recombination.
J. Mol. Biol. 96 (1975) 495-509.
Hanahan, D.: Studies on transformation ofEscheriehia coli with plasmid.
J. Moi. Biol. 166 (1983) 557-580.
Jones, I.M., Primrose, S.B. and Ellwood, D.C.: Effect ofg~owth rate and
nutrient,limitation on the transformability of Escherichia coli with
plasmid deoxyribonucleic acid. J. Bacteriol. 146 (1981) 841-846.
Kretschmer, P.J., Chang, A.C.Y. and Cohen, S.N.: Indirect selection of
bacterial plasmids lacking identifiable phenotypic properties. J. Bacteriol. 124 (1975) 225-231.
Lederberg, E.M. and Cohen, S.N" Transformation of Salmonella
typhimurium by plasmid deoxyribonucleic acid. J. Bacteriol. !19
(1974) 1072-1074.
Mandel, M. and Higa, A.:Calcium-dependent bacteriophage DNA infection. J. Mol. Biol. 53 (1970) 159-162.
Mes~mg, J., Crea, R. and Seeburg, P.H.: A system for shotgun DNA
sequencing. Nucleic Acids Res. 9 (1981) 309-321.
Morrison, D.A.:Transformation and preservation of competent bacterial
cells by freezing. Methods Enzymoi. 68 (1979) 326-331.

28
Okayama, H. and Berg, P.: High-efficiency cloning of full-length eDNA.
Mol. Cell. Biol. 2 (1982) 161-170.
Okayama, H. and Berg, P.: A cDNA cloning vector that permits expression of eDNA inserts in mammalian cells. Mol. Cell. Biol. 3 (1983)
280-289.
Okayama, H., Kawaichi, M., Brownstein, M., Lee, F., Yokota, T. and
Arai, K.: High-efficiency cloning of full-length eDNA; construction

and screening of eDNA expression libraries for mammalian cells.

Methods Enzymol. 154 (1987) 3-28.
Taketo, A.: Sensitivity of Escherichia coli to viral nucleic acid, V. Competence of calcium-treated cells. J. Biochem. 72 (1972) 973-979.
Taketo, A. and Kuno, S.: Sensitivity of Escherichia coli to viral nucleic
acid, VI. Further studies on Ca: + -induced competence. J. B::~ochem.
75 (1974) 59-67.