Eukaryotic Cells to Antibiotic Selection ... At 39 to 40°C, selection of antibiotic-resistant transfected mammalian cell linesor .... of that synergy is probably different.
ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Aug. 1992, p. 1782-1784
Vol. 36, No. 8
0066-4804/92/081782-03$02.00/0 Copyright © 1992, American Society for Microbiology
Increased Susceptibility of Transfected Prokaryotic and Eukaryotic Cells to Antibiotic Selection FRED A. M. ASSELBERGS* AND PETRI VAN SOMEREN Department of Biotechnology, K681. 4. 42, Ciba-Geigy Ltd., CH-4002 Basel, Switzerland Received 16 September 1991/Accepted 26 May 1992
At 39 to 40°C, selection of antibiotic-resistant transfected mammalian cell lines or Escherichia coli required lower aminoglycoside antibiotic concentrations than at 37°C. The thermosensitivity of antibiotic susceptibility was much more manifest during genetic selection experiments than in conventional growth inhibition assays.
Compared with the dramatic differences in selection efficiency between 37 and 40°C seen after DNA transfection (Fig. 1), the effect of mild hyperthermia appeared small in standard short-term cell cultures (Fig. 2A and D). If a correction is made for the slower growth of cells at 40°C, a leftward shift of the resistance curve is observed at intermediate drug concentrations; this shift amounts to at most a twofold difference in drug susceptibility (Fig. 2C and D). One obvious difference between the two assays is certainly that 10 to 15 cell divisions are needed to form visible colonies of 1,000 to 5,000 stably transfected cells, whereas growth inhibition during one cell passage spans only 5 or 6 cell divisions. Similar results of transfection and growth inhibition assays were obtained with Bowes melanoma cells (3), except that this human cell line required 30% lower antibiotic concentrations (unpublished data). An inhibitor affecting drug transport across the cell membrane was used to investigate whether acceleration of the intracellular accumulation of aminoglycosides would have an effect similar to that of hyperthermia. The ionophore nigericin facilitates uptake of aminoglycoside antibiotics (1, 12). Although with nigericin 1/10 as many colonies were obtained after transfection of pSV2911neo and pCGA25cNAT, the same concentration of geneticin or hygromycin B was needed for selection as in the absence of nigericin (unpublished data). The aminoglycoside concentrations used for bacteria (0.025 to 0.050 mg/ml) are much lower than the lowest concentrations used for mammalian cells (0.2 to 0.5 mg/ml). In E. coli at 37°C, the number of plasmid-transformed colonies did not decrease until 20 times the minimally effective selective dose was reached (Fig. 3A and B). At low concentrations, hyperthermia did not affect transformation efficiency, but the cloning efficiency at 40°C relative to the efficiency at 37°C decreased linearly with progressively higher antibiotic concentrations (Fig. 3C and D). Addition of 1 ,uM nigericin to neutral or acidic medium (8, 12) increased the susceptibility of nonresistant E. coli to aminoglycosides two- to fivefold but did not affect transformation efficiency with pSV2911neo or pCGA25cNAT significantly (unpublished data). Presumably, mild hyperthermia does not influence antibiotic susceptibility through a single mechanism. Heat stress and inhibition of protein synthesis probably affect many cellular processes, but the net effects in E. coli and mammalian cells are very similar. One might be able to identify the factors involved genetically in E. coli. Clinically, severe hyperthermia (10 min to 1 h at 43 to 45°C) is used to enhance the toxicity of tumoricidal agents (5, 9), but the mechanism
Aminoglycoside antibiotics of the neomycin and hygromycin B groups are widely used for eukaryotic and prokaryotic cells (11). However, the concentrations required in tissue culture are so high that hygromycin B has even been referred to as "not toxic to mammalian cells" (6). In this article, we demonstrate that mild hyperthermia enhances antibiotic susceptibility such that lower concentrations are needed to select transfected mammalian cells and bacteria. Two shuttle plasmids were used for transfection: pSV2911neo (4), which confers resistance to neomycin-type antibiotics (kanamycin in bacteria and geneticin in mammalian cells), and pCGA25cNAT (2), which confers resistance to hygromycin B. In both plasmids, the simian virus 40 early promoter is used to express resistance in mammalian cells, and in Escherichia coli, the promoter of the antibiotic resistance genes of transposon Tn5 is used. The lowest concentration of geneticin that could be used to select transfected CHO cells at 37°C was 0.4 to 0.5 mg/ml (Fig. 1A). Cell lines isolated at high geneticin concentrations were also found to be more resistant in subsequent growth inhibition assays than those selected at lower concentrations. At 40°C, good yields of stably transfected clones with complete elimination of the untransfected cells were obtained at concentrations as low as 0.15 mg/ml (Fig. 1A), but no colonies appeared at concentrations above 1.5 mg/ml (Fig. 1A). Expressed as a percentage of the transfection efficiency at 37°C, the transfection efficiency at 40°C declined linearly from 15% at 0.4 mg/ml to less than 0.25% at 2 mg/ml (Fig. 1C). No difference in susceptibility to geneticin at either 37 or 40°C between pooled cell lines selected at 40°C with 0.2 mg/ml and cells selected at 37°C with 1.0 mg/ml was observed. With hygromycin B, cloning efficiencies were also similar at 40°C to those at 37°C but required two- to threefold lower drug concentrations (Fig. 1B and D). Selection at 40°C also eliminated the background problem observed with CHO and other cell lines at 37°C in that, with low concentrations of hygromycin B, clumps of nontransfected cells survived for more than 6 weeks. No apparent differences in plasmid copy number were found in a Southern dot blot comparison of DNA from pools of resistant clones obtained with an equally selective concentration of either hygromycin B or geneticin at 40 or 37°C (data not shown). Preliminary experiments also failed to reveal a significant difference in the status of the chromosomally integrated plasmids in individual cell lines obtained at 40°C. *
Corresponding author. 1782
CHO colonies/microgram 300 250 200150 100 50O 0 A.
DNA
CHO colonies/microgram DNA
A:nn, buv
10
E. coli colonies/dish x 1000 10_ = =
E. coli colonies/dish x 1000
_
400 300200
_
_ 0.0 1 F7=-
L
U-
0.5 1 1.5 mg/ml Geneticin
2
0
B.
0.2 0.4 0.6 0.8 mg/ml Hygromycin-B
'40oC
37 oC 80
1
0
A.
'37oC
--
__
K.
_:
0.4 0.6 0.8 mg/ml Kanamycin 40 oC ° 37 oC
0.2
1
of control
100 -e
0
B.
of control
%
1010O
v
-o
2
D.
0.25
o.0
0.1 0.15 0.2 mg/ml Hygromycin-B
of that synergy is probably different. Perhaps the effect of mild hyperthermia described here also has some medical importance, for example, during treatment of febrile patients with antibiotics or tumoricidal agents. Alternatively, it may be of practical interest in the laboratory. Hyperthermia millions of cells
1; 0.8 0.6 0.4 0.2
0.8 0.6 0.4 0.2 0
0
0.2
0.4
0.8
0.6
0 B.
mg/ml Geneticin
40 oC % of control
37 oC
0.05
0.1 0.15 0.2 mg/ml Hygromycin-B + 40 oC + 37 oC
0.25
% of control
120, 100-
1
_-
80
00
_ ,,
20
v
0.2
0.4
0.6
0.8
mg/ml Geneticin + 40 oC
,, __
80 6040 20
60 40
-
37 oC
1
0 D.
i
_
u-
0.05
0.1
0.05
0.1
0.15
0.2
mg/ml Hygromycin-B
+ 40 oC of control
37 oC
--
-~
61 41
21
0.2
0.4
0.6
0.8
mg/ml Kanamycin
1 1
0
D.
0.15 0.05 0.1 mg/ml Hygromycin-B
0.2
FIG. 3. Transformation of E. coli at 37 and 40'C. E. coli DH5a (Stratagene) was transformed (11) with plasmid pSVneo2911 (A and C) or pCGA25cNAT (B and D) and spread on plates containing Luria-Bertani agar, which were incubated at 37 or 40'C. Colonies were counted after 18 h. In panels C and D, the number of antibiotic-resistant colonies obtained at 40'C is expressed as a percentage of those obtained at 37°C. Like geneticin, kanamycin is inactivated by the neomycin phosphotransferase encoded by pSV2911neo, but kanamycin is not toxic for mammalian cells.
increases the stringency of antibiotic selection. This may allow the use of resistance genes in cells or microorganisms ordinarily resistant to antibiotics. Another possible application is to force amplification of resistance genes, which at least with geneticin is not possible or highly impractical in mammalian cells because of the high concentrations needed (10). Finally, the altered antibiotic susceptibility at elevated temperatures may require adjustments in the antibiotic concentration in experiments in which thermosensitive genes and antibiotic resistance markers are coexpressed. The present work would have been impossible without the excellent technical assistance of Sonja Schneider. We thank L. Chasin for DUKXB1 cells and N. Hardman for critical reading of the manuscript.
1.2
1.4 1.4 1.2
0
C.
FIG. 1. Selection of transfected tissue culture cells. CHO cells of line DUKXB1 (deficient in dihydrofolate reductase) were transfected as described previously (4) with plasmid pSVneo2911 (A and C) or pCGA25cNAT (B and D). The cells were trypsinized and subcultured 24 h after the osmotic (glycerol) shock at the end of the transfection procedure. At that time, a part of each culture was transferred to 40°C. The next day, the culture medium was replaced by medium containing the selective antibiotic. Subsequently, the culture medium was replaced every third or fourth day. Colonies obtained at 37 or 40'C were scored on day 17. In panels C and D, the number of drug-resistant colonies obtained at 40'C is expressed as a percentage of those obtained at 37'C.
1.6millions of cells
I
81!0-
v
1 1.5 mg/ml Geneticin
= =
