- Email: [email protected]
A highly sensitive, mixed-phase assay for chloramphenicol acetyltransferase activity in transfected cells
179,1%?3
ANALYTICALBIOCHEMISTRY
(1989)
A Highly Sensitive, Mixed-Phase Assay for Chloramphenicol Acetyltransferase Activity in Transfected Cells’ David
A. Nielsen,
Department
Received
Tsu-Chung
of Biochemistry,
August
Chang,2 and David
University
of Illinois,
J. Shapiro
1209 West California
Quantitation of promoter activity in vertebrate cells is usually achieved by covalently joining the DNA fragments to be tested to the easily assayable bacterial reporter gene encoding chloramphenicol acetyltransferase
i This research was supported by Grant DCB 8804109 from the National Science foundation, by an American Cancer Society postdoctoral fellowship (PF-2486) (to D.A.N.), and by a fellowship from the Chung Shan Institute of Science and Technology Taiwan, Republic of China (to T.C.C). * On leave from the National Defense Medical Center, Taipei, Taiwan, Republic of China.
Copyright All rights
$3.00
0 1989 by Academic of reproduction
Illinois
61801
5,1988
We describe a simple, rapid yet extremely sensitive assay for chloramphenicol acetyltransferase (CAT) activity in extracts from transfected eukaryotic cells. Using our modified reaction conditions and the mixedphase assay, less than 0.000010 unit of CAT activity in transfected cells can be reliably detected. The mixedphase assay is based on the inability of the polar [‘I-I]acetyl-&enzyme A (CoA) substrate to partition out of a urea containing aqueous phase into the nonpolar scintillation fluor, while the [%I]chloramphenicol reaction products partition into the toluene scintillation fluor and are quantitated by scintillation counting. The increased sensitivity of this assay is due to the optimization of the acetyl-CoA concentration, to a urea-containing aqueous phase which lowers the assay background, and to the use of extract blanks. The mixed-phase assay is simpler, is quantitative, uses less costly substrates, and is far more sensitive than the most widely used CAT assays, which require solvent extraction followed by thin-layer chromatography to separate the unreacted substrate from product. @ 1999 Academic POW, I~~.
0003-2697/89
Street, Urbana,
Press, in any form
Inc. reserved.
(CAT).3 The most widely used CAT assay entails incubating soluble extracts from transfected cells with [14C]chloramphenicol and an excess of unlabeled acetyl-CoA, extracting with organic solvents, separating the substrate and products by thin-layer chromatography, and quantitating CAT activity by autoradiography or by scraping the spots off the thin-layer plates and counting them (1). Because this assay is labor intensive and relatively slow, several alternative methods for quantitating CAT activity have been described. These include assays using a [3H]acetyl-CoA generating system (2), an assay based on the acetylation of chloramphenicol with [ 14C] acetyl-CoA (3), and an assay in which the reaction products are continuously extracted into scintillation fluor (4,5). Several groups have recently described transfection assays based on the use of luciferase as a reporter gene (6). Although luciferase assays are probably the most sensitive transfection assays in current use, they employ a costly luminometer to assay the enzyme and require the recloning of a large number of constructions made using CAT as a reporter gene. In this paper we describe a CAT assay which combines simplicity, speed, and low cost to achieve a level of sensitivity which approaches that obtained with luciferase assays. The assay is based on the principle of the mixedphase assay for 3-hydroxy-3-methylglutaryl-Coenzyme A reductase we developed several years ago (7). In the method presented here the polar 3H-labeled substrate remains dissolved in the urea-containing aqueous phase, while the nonpolar reaction product partitions into the toluene scintillation fluor. This mixed-phase assay has ’ Abbreviations used: CAT, chloramphenicol acetyltransferase; CoA, Coenzyme A; XLF, Xenopuc Levis fibroblast; HSV TK, herpes simplex virus thymidine kinase; TFIIIa, transcription factor IIIa; VIT-CAT, vitellogenin promoter CAT; TKXER, thymidine kinase promoter X. lueuia estrogen receptor; HMG, 3-hydroxy-3-methylglutaryl.
19
20
NIELSEN,
CHANG,
AND
SHAPIRO
been successfully used in our laboratory for over a year to assay transfections of mammalian and amphibian cells. It differs in several important respects (see below) from a continuous two-phase CAT assay independently described by Neumann et al. (4). MATERIALS
AND
METHODS
Materials. [3H]Acetyl-Coenzyme A (16 Ci/mmol) was obtained from ICN Radiochemicals and [14C]chloramphenicol (60 mCi/mmol) from New England Nuclear. Chloramphenicol, acetyl-CoA, and bovine serum albumin (crystallized) were obtained from Sigma. Cell lines and transfections. The Xenopus laevis fiCAT ACTIVITY (E.U.1 broblast (XLF) and liver (AllO) cell lines were isolated in our laboratory and will be described elsewhere (M. FIG. 1. Standard curve. Purified CAT was diluted in Buffer I containing 80 pg bovine serum albumin/100 ~1 reaction and assayed by Barton and D. Shapiro, manuscript in preparation). assay described under Materials and Methods. HepG2 cells, a human hepatocyte line, was kindly pro- the mixed-phase vided by Dr. Barbara Knowles (8). MCF-7, human breast cancer cells (9), were obtained from the Michigan hour incubations were chosen because the mixed-phase Cancer Foundation. Transfections were performed with assay was linear over incubation times of at least 2 h various plasmid DNA constructions containing the CAT (data not shown). The complete reaction contained 0.4 structural gene preceded by a variety of promoters (HSV PCi [3H]acetyl-CoA, 30 PM acetyl-CoA, 1 mM chloramTK, Xenopus vitellogenin and TFIIIa [gift of Dr. Wilphenicol, 4 mM EDTA, and 200 mM Tris, pH 7.8. The liam Taylor, Vanderbilt University] followed by the 3’ reaction was terminated by mixing 90 ~1 of the reaction regions of either the SV40 early gene or the vitellogenin Bl gene). Twenty micrograms of supercoiled DNA (2X mixture with 1 ml 7 M urea (in water) in a 20-ml scintillaCsCl gradient purified) was transfected into 2-4 X lo6 tion vial. Ten milliliters of 0.8% PPO (2,5diphenyloxacells by a modification of the calcium phosphate copre- zole) in toluene was added and the vial shaken for 10 cipitation method (10) incorporating a glycerol shock s. After allowing 15 min for the phases to separate, the (11) or by the DEAE-dextran method with omission of cocktail was counted for 10 min. Samples are stable in the two-phase system for at least several hours. the chloroquine step (12). Preparation of cell extract. After 48 or 72 h the transAND DISCUSSION fected cells were harvested by scraping, pelleted at 700g RESULTS for 3 min, and resuspended in 100 ~1 Buffer I (250 mM Most CAT assays employ labeled chloramphenicol Tris, pH 7.8, 5 mM EDTA) (13). Following sonication and a large excess of unlabeled acetyl-CoA. Since we for 3 s, the extract was spun at 12,OOOgfor 15 min. Super- were using labeled acetyl-CoA, we determined that the natants from liver cell lines were then incubated for 7 Km for acetyl-CoA in the CAT reaction is 80 PM (data min at 60°C and spun at 12,000g for 15 min and the re- not shown). This was somewhat higher than the 50 pM sultant supernatant was quick-frozen in liquid N2 and Km reported by Shaw et al. (14) or the 40 PM K,,, deterstored at -70°C. Fibroblast extracts were frozen and as- mined by Neumann et al. (4). This difference may reflect sayed without preheating. Protein was determined by differences in assay conditions. When the enzyme assay the Coomassie blue protein assay (Bio-Rad). was performed with a substrate concentration of 100 PM acetyl-CoA, the sensitivity was 0.0010 unit of CAT, and Mixed-phase CAT Assay. For maximum sensitivity, the mixed-phase assay was linear over three orders of samples were assayed as a complete reaction with chlormagnitude (from 0.0010 to 1.0 unit of CAT) [data not amphenicol and an extract blank was assayed without shown]). Decreasing the acetyl-CoA concentration to 30 chloramphenicol. Specific CAT activity was determined by subtracting the extract blank from the complete reac- PM dramatically increased the sensitivity to 0.000010 tion. Sixty microliters of cell extract containing 20-50 unit CAT. The standard curve in Fig. 1 was obtained pg of protein plus Buffer I was mixed with either 20 ~1 of from mixed-phase assays of samples using 30 pM acetyl5 mM chloramphenicol in Buffer I (complete reaction) or CoA and increasing amounts of purified bacterial CAT. The response was essentially linear over a range of CAT 20 ~1 of Buffer I (extract blank). After a 5-min incubaactivities spanning more than three orders of magnitude tion at 37”C, 20 ~1 of 20 &i/ml [3H]acetyl-CoA, 150 PM acetyl-CoA, in 75 PM HCl, was added to both samples of CAT activity (from 0.000010 to 0.010 unit of CAT). As long as less than approximately 30% of the acetyland the incubation was continued for 2 h at 37°C. TWO-
A MIXED-PHASE
CHLORAMPHENICOL TLC
BAnPW (a011
line,
transfmztion
ACETYLTRANSFERASE MIXED
ASSAY
vector)
TK-CAT
MCF-7,
B.
XL
Fibroblast,
XL
Liver
TK-CAT
(AllO),
VIT-CAT
PHABE
COMPLETE
EXTRACT
REACTIOl
BLAUA
cpn/!N
21
ASSAY
cwh
9
ABSAY CAT ACTIVITY
cPWq
808
18
790
102
11
93
K.U./uq
82
x lo+
7.6
~10-~
+ TK-XER
Exp.
1
97
26
71
2.4
lo+
Exp.
2
46
25
21
0.74
x 10-e
Exp.
3
40
25
16
0.56
x
x
lO-6
FIG. 2. Comparison of CAT assays of transfected cell extracts performed by extraction followed by thin-layer chromatography and by the mixed-phase assay. The indicated cell lines of human and amphibian origin were transfected by calcium phosphate coprecipitation as described under Materials and Methods. Aliquots of the transfected cell extracts were assayed for formation of [W]acetyl-chloramphenicol by thin-layer chromatography (1) and for formation of [3H]acetyl-chloramphenicol by the mixed-phase assay. The thin-layer chromatograph origin is to the left. The promoters used in the transfections are shown next to the names of the cell lines. The thin-layer and mixed-phase assays shown in panel A were performed with 50 pg of cell extract protein. The assays shown in panel B were carried out on 20 fig of cell extract protein. The autoradiogram shown in panel A was exposed for 10 h, while the autoradiogram shown in panel B was exposed for 2 days. The spots from the TLC were quantitated by scintillation counting of the scraped TLC. In panel A, 35,326, 6615, 175,910, and 16,002 cpm of acetylated products were formed by the HepG2, XL liver, MCF-7, and XL fibroblast extracts, respectively, from 342,000 cpm of substrate. Mock-transfected HepG2 cell extract formed 1234 cpm of acetylated product (TLC not shown). In panel B, 3837,1732, and 1773 cpm of acetylated products were formed by the XL liver extracts in Experiments 1,2, and 3, respectively, from 274,000 cpm of substrate. Data from the mixed-phase assay are presented as counts per minute per microgram extract protein. CAT activity in enzyme units per microgram extract was calculated from total counts per minute in the 50 and 20 pg reactions and the standard curve in Fig. 1.
CoA in the mixed-phase assay is converted to product, the reaction displays excellent linearity. The increased sensitivity of the assay and its linearity over a broad range of enzyme activities led us to employ 30 pM acetylCoA in all of the remaining assays. A lower concentration of acetyl-CoA allows the use of a higher specific radioactivity [3H]acetyl-CoA substrate, while keeping the cost of the labeled substrate well below the cost of [l*C]chloramphenicol. The mixed-phase assay is considerably less expensive than our modified [‘“Clchloramphenicol thin-layer assay (see below), which uses only one-fifth the level of costly [14C]chloramphenicol specified by Gorman et al. (1). The accuracy and sensitivity of the mixed-phase assay were determined by transfecting a variety of vertebrate cell lines with promoters linked to CAT and assaying
the resulting extracts in parallel by solvent extraction followed by thin-layer chromatography and by the mixed-phase assay. The data presented in Fig. 2 demonstrate a good correlation between results obtained with the two assays. The sensitivity of the [‘“Clchloramphenicol thin-layer chromatography assay has been optimized by using a 3-h incubation and multiple additions of acetyl-CoA while using only 0.2 &i of [‘“Clchloramphenicol per reaction. The sensitivity of the mixed-phase assay is illustrated by the facile, rapid, and quantitative detection of low levels of CAT activity, which gave weak signals by the thin-layer assay even after exposure of the autoradiogram for 2 days (Fig. 2, panel B). Although the cell lines used in this study are transfected with reasonable efficiency, many widely used mammalian cells are much more efficiently transfected.
22
NIELSEN, TABLE CAT
Assay
Complete reaction& (cpmlai3 protein)
Samples Xenopus fibroblasts (XLF) A. VIT-CAT, TKXER, +estrogen’ Experiment Experiment Experiment Experiment Experiment
1 2 3 4 5
Average of 16 total B. SVOCAT* Experiment 1 Experiment 2 Experiment 3 HepG2 cells’ Experiment Experiment
Background
1 2
CHANG,
1
in the Mixed-Phase Extract blank (-cbloramphenicol) km/pi3 protein)
Assay
CAT activity (cpm/wg protein)
436 202 265 596 211
16 20 14 16 16
416 162 271 566 195
452 f 67
16.6 f 0.6
435 f 67
25 24 37
15 17 16
10 7 21
145 210
21 20
126 190
“XLF cells were cotransfected with VIT-CAT and TKXER and grown in medium plus 100 PM estradiol-17@, and 20-50 fig of cell extract was assayed as described under Materials and Methods. The results of the first 5 transfection experiments as well as the average (&SE) of 16 transfection experiments are shown. * XLF cells were transfected with the promoterless vector SVOCAT in three experiments and 30 fig of cell extract protein was assayed. ’ HepG2 cells were transfected with pRCV12 (a rous sarcoma virus promoter-CAT plasmid) in two experiments and 50 pg of cell extract protein was assayed. d Results are presented as per microgram of extract protein.
These cell lines were chosen to provide a variety of cell lines and to illustrate the application of the mixed-phase assay to relatively low activity samples. The mixedphase assay was developed to allow quantitative determination of CAT activity in cells transfected with promoters of low transcriptional activity. Detection of very low levels of CAT activities by the mixed-phase assay depends on the ability to discriminate between radioactivity derived from CAT enzyme activity and background radioactivity from the extracts. To attain a precise determination of CAT activity in an extract, we performed parallel incubations of the complete reaction and an extract blank (which contains [3H]acetyl-CoA but lacks chloramphenicol) and subtracted the extract blank from the radioactivity obtained with the complete reaction. Our approach assumes that the extract blank, which is performed on one aliquot of the extract, is the same as the extract blank would be in the aliquot used to assay the complete reaction. TO evaluate the reproducibility of the extract blanks, we have summarized the data from transfections with sev-
AND
SHAPIRO
eral sets of vectors. Although extracts from each cell line give slightly different reaction blanks (see the table in Fig. 2, panel A), the blanks within a cell line are remarkably reproducible. The data summarized in Table 1 (part A) show the results of the first five separate transfection experiments for the expression of the vitellogenin promoter under the control of the X. Levis estrogen receptor and estradiol-17& In a total of 16 assays carried out over many months the reaction blank averaged 16.6 f 0.5 (SE) cpm/pg protein. The level of uncertainty which surrounds the extract blank to be subtracted is therefore extremely low, less than 1 cpm/pg protein. This allows even the SVOCAT vector, which lacks a promoter and exhibits extremely low background transcription in amphibian cells, to be quantitatively assayed (Table 1, part B). Because of the exceptionally high level of reproducibility of the extract blanks, transfections which employ promoters of average activity can be assayed without the use of extract blanks. The major factors contributing to the reaction blank appear to be the partition of a small quantity of the [3H]acetyl-CoA into the toluene scintillation fluor and the amount of protein in the assay samples. Samples containing extremely high levels of CAT activity (outside the linear range of this assay) exhibit an elevated extract blank, which nevertheless represents only a tiny fraction of their CAT activity. This suggests that even in the absence of chloramphenicol, CAT may be capable of converting minute amounts of acetyl-CoA to a nonpolar product. Our experience with an earlier mixed-phase assay for HMG-CoA reductase (7) led us to a number of procedures which increased the sensitivity of this assay and which differ from the recently described two-phase CAT assay based on continuous extraction of the products into scintillation fluor (43). The urea in the aqueous phase both terminates the reaction and lowers the reaction blank. Determining the optimum concentration of [3H]acetyl-CoA substrate effected a large increase in sensitivity without decreasing linearity. Comparison of the signal-to-noise ratios of the mixed-phase assay reported here and the data reported for the continuous extraction procedure of Neumann et al. (4) indicates that the mixed-phase assay we present is approximately 14 times more sensitive.* It is difficult to compare the sensitivity of the mixedphase assay with that of other CAT assays because in 4 The signal-to-noise ratios of the two assays were compared as follows: The two-phase assay of Neumann et al. (Ref. 4, Fig. 1, panel B) yields approximately 146,000 cpm of activity when 0.010 unit of CAT enzyme was assayed over a background of 9050 cpm (with no CAT), for a signal-to-noise ratio of 16:l. The mixed-phase assay yields a signal of 87,638 cpm over a background of 395 cpm for the same assay of 0.010 unit of CAT enzyme. This gives a signal-to-noise ratio of 222:1, indicating that our mixed-phase assay is 14-fold more sensitive than the continuous two-phase assay described by Neumann et al. (4)
A MIXED-PHASE
CHLORAMPHENICOL
many cases either comparison to the original thin-layer assay or quantitative expression of the data in CAT enzyme units is lacking. The sensitivity of the assay and the range of activities assayable by the mixed-phase assay are illustrated by assay of the promoterless SVOCAT vector (Table 1). The transcriptional activity of this vector in Xenopus cells is ~0.5% the activity of a vector containing a strong promoter, such as the adenovirus major late promoter CAT (data not shown). The sensitivity of the mixed-phase assay is comparable to that of the original luciferase assay. The mixedphase assay is able to detect 0.000010 unit of CAT or 1.5 X lo6 molecules of CAT (l&16). In fact, the extract of Xenopus fibroblasts transfected with SVOCAT assayed in Table 1 (part B, Experiment 2) contained only 0.000020 unit or 3 X lo6 molecules of CAT. This is equivalent to the limit of sensitivity of 3 X lo6 molecules reported for the original luciferase assay (6). However, improved luminometers effect a considerable increase in the sensitivity of subsequent luciferase assays. The mixed-phase assay presented here is an extremely sensitive assay for the determination of CAT activity in transfected cell extracts. It is extremely simple and rapid to perform, yielding reliable CAT activity determinations of transfected cell extracts in less than 3 h. Note added in proof. [3H]Acetyl-Coenzyme A from ICN Radiochemicale and from New England Nuclear has been used successfully. The
ACETYLTRANSFERASE
23
ASSAY
single lot of [3H]acetyl-Coenzyme an approximately ten fold higher
A from Ameraham assay blank.
we tested
yielded
REFERENCES 1. Gorman, C. M., Moffat, Biol. 2,1044-1051.
L. F., and Howard,
B. H. (1982)
2. Nordeen, S. K., Green, P. P., and Fowlkes, D. M. 173-178. 3. Sleigh, M. J. (1986) Anal. Biochem. 156,251-256.
(1987)
Mol.
Cell.
DNA
6,
4. Neumann, J. R., Morency, C. A., and Russian, K. 0. (1987) BioTechniques 6,444-447. 5. Eastman, A. (1987) BioTechniques 6(8), 730-732. 6. De Wet, J. R., Wood, K. V., DeLuca, M., Helinski, D. R., and Subramani, S. (1987) Mol. Cell. Biol. ‘7(2), 725-737. 7. Phillipp, B. W., and Shapiro, D. J. (1979) J. Lipid Res. 20, 588593. 8. Knowles, B. B., Howe, C. C., and Aden, D. P. (1980) Science 209, 497-499. 9. Soule, H. D., Vazquez, J., Long, A., Albert, S., and Brennan, M. (1973) J. N&l. Cancer Inst. 61(5), 1409-1416. 10. Graham,
F. L., and Van der Eb, A. J. (1973)
Virology
11. Parker, B. A., and Stark, G. R. (1979) J. Virol. 31(2), 12. Luthman, H., and Magnusson, G. (1983) Nucleic ll(ll), 1295-130s. 13. Crab, D. W., and Dixon, J. E. (1987) Anal. 14. Shaw, W. V., Bently, D. W., and Sands, 104(3), 1095-1105.
Biochem. L. (1970)
62,456-467. 360-369. Acids Res. 163,88-92. J. Bacterial.
15. Shaw, W. V., Packman, L. C., Burleigh, B. D., Dell, A., Morris, H. R., and Hartley, B. S. (1979) Nature (London) 282.870-872. 16. Shaw, W. V., andBrodsky, R. F. (1968) J. Bacterial. 96.28-36.
ANALYTICALBIOCHEMISTRY
(1989)
A Highly Sensitive, Mixed-Phase Assay for Chloramphenicol Acetyltransferase Activity in Transfected Cells’ David
A. Nielsen,
Department
Received
Tsu-Chung
of Biochemistry,
August
Chang,2 and David
University
of Illinois,
J. Shapiro
1209 West California
Quantitation of promoter activity in vertebrate cells is usually achieved by covalently joining the DNA fragments to be tested to the easily assayable bacterial reporter gene encoding chloramphenicol acetyltransferase
i This research was supported by Grant DCB 8804109 from the National Science foundation, by an American Cancer Society postdoctoral fellowship (PF-2486) (to D.A.N.), and by a fellowship from the Chung Shan Institute of Science and Technology Taiwan, Republic of China (to T.C.C). * On leave from the National Defense Medical Center, Taipei, Taiwan, Republic of China.
Copyright All rights
$3.00
0 1989 by Academic of reproduction
Illinois
61801
5,1988
We describe a simple, rapid yet extremely sensitive assay for chloramphenicol acetyltransferase (CAT) activity in extracts from transfected eukaryotic cells. Using our modified reaction conditions and the mixedphase assay, less than 0.000010 unit of CAT activity in transfected cells can be reliably detected. The mixedphase assay is based on the inability of the polar [‘I-I]acetyl-&enzyme A (CoA) substrate to partition out of a urea containing aqueous phase into the nonpolar scintillation fluor, while the [%I]chloramphenicol reaction products partition into the toluene scintillation fluor and are quantitated by scintillation counting. The increased sensitivity of this assay is due to the optimization of the acetyl-CoA concentration, to a urea-containing aqueous phase which lowers the assay background, and to the use of extract blanks. The mixed-phase assay is simpler, is quantitative, uses less costly substrates, and is far more sensitive than the most widely used CAT assays, which require solvent extraction followed by thin-layer chromatography to separate the unreacted substrate from product. @ 1999 Academic POW, I~~.
0003-2697/89
Street, Urbana,
Press, in any form
Inc. reserved.
(CAT).3 The most widely used CAT assay entails incubating soluble extracts from transfected cells with [14C]chloramphenicol and an excess of unlabeled acetyl-CoA, extracting with organic solvents, separating the substrate and products by thin-layer chromatography, and quantitating CAT activity by autoradiography or by scraping the spots off the thin-layer plates and counting them (1). Because this assay is labor intensive and relatively slow, several alternative methods for quantitating CAT activity have been described. These include assays using a [3H]acetyl-CoA generating system (2), an assay based on the acetylation of chloramphenicol with [ 14C] acetyl-CoA (3), and an assay in which the reaction products are continuously extracted into scintillation fluor (4,5). Several groups have recently described transfection assays based on the use of luciferase as a reporter gene (6). Although luciferase assays are probably the most sensitive transfection assays in current use, they employ a costly luminometer to assay the enzyme and require the recloning of a large number of constructions made using CAT as a reporter gene. In this paper we describe a CAT assay which combines simplicity, speed, and low cost to achieve a level of sensitivity which approaches that obtained with luciferase assays. The assay is based on the principle of the mixedphase assay for 3-hydroxy-3-methylglutaryl-Coenzyme A reductase we developed several years ago (7). In the method presented here the polar 3H-labeled substrate remains dissolved in the urea-containing aqueous phase, while the nonpolar reaction product partitions into the toluene scintillation fluor. This mixed-phase assay has ’ Abbreviations used: CAT, chloramphenicol acetyltransferase; CoA, Coenzyme A; XLF, Xenopuc Levis fibroblast; HSV TK, herpes simplex virus thymidine kinase; TFIIIa, transcription factor IIIa; VIT-CAT, vitellogenin promoter CAT; TKXER, thymidine kinase promoter X. lueuia estrogen receptor; HMG, 3-hydroxy-3-methylglutaryl.
19
20
NIELSEN,
CHANG,
AND
SHAPIRO
been successfully used in our laboratory for over a year to assay transfections of mammalian and amphibian cells. It differs in several important respects (see below) from a continuous two-phase CAT assay independently described by Neumann et al. (4). MATERIALS
AND
METHODS
Materials. [3H]Acetyl-Coenzyme A (16 Ci/mmol) was obtained from ICN Radiochemicals and [14C]chloramphenicol (60 mCi/mmol) from New England Nuclear. Chloramphenicol, acetyl-CoA, and bovine serum albumin (crystallized) were obtained from Sigma. Cell lines and transfections. The Xenopus laevis fiCAT ACTIVITY (E.U.1 broblast (XLF) and liver (AllO) cell lines were isolated in our laboratory and will be described elsewhere (M. FIG. 1. Standard curve. Purified CAT was diluted in Buffer I containing 80 pg bovine serum albumin/100 ~1 reaction and assayed by Barton and D. Shapiro, manuscript in preparation). assay described under Materials and Methods. HepG2 cells, a human hepatocyte line, was kindly pro- the mixed-phase vided by Dr. Barbara Knowles (8). MCF-7, human breast cancer cells (9), were obtained from the Michigan hour incubations were chosen because the mixed-phase Cancer Foundation. Transfections were performed with assay was linear over incubation times of at least 2 h various plasmid DNA constructions containing the CAT (data not shown). The complete reaction contained 0.4 structural gene preceded by a variety of promoters (HSV PCi [3H]acetyl-CoA, 30 PM acetyl-CoA, 1 mM chloramTK, Xenopus vitellogenin and TFIIIa [gift of Dr. Wilphenicol, 4 mM EDTA, and 200 mM Tris, pH 7.8. The liam Taylor, Vanderbilt University] followed by the 3’ reaction was terminated by mixing 90 ~1 of the reaction regions of either the SV40 early gene or the vitellogenin Bl gene). Twenty micrograms of supercoiled DNA (2X mixture with 1 ml 7 M urea (in water) in a 20-ml scintillaCsCl gradient purified) was transfected into 2-4 X lo6 tion vial. Ten milliliters of 0.8% PPO (2,5diphenyloxacells by a modification of the calcium phosphate copre- zole) in toluene was added and the vial shaken for 10 cipitation method (10) incorporating a glycerol shock s. After allowing 15 min for the phases to separate, the (11) or by the DEAE-dextran method with omission of cocktail was counted for 10 min. Samples are stable in the two-phase system for at least several hours. the chloroquine step (12). Preparation of cell extract. After 48 or 72 h the transAND DISCUSSION fected cells were harvested by scraping, pelleted at 700g RESULTS for 3 min, and resuspended in 100 ~1 Buffer I (250 mM Most CAT assays employ labeled chloramphenicol Tris, pH 7.8, 5 mM EDTA) (13). Following sonication and a large excess of unlabeled acetyl-CoA. Since we for 3 s, the extract was spun at 12,OOOgfor 15 min. Super- were using labeled acetyl-CoA, we determined that the natants from liver cell lines were then incubated for 7 Km for acetyl-CoA in the CAT reaction is 80 PM (data min at 60°C and spun at 12,000g for 15 min and the re- not shown). This was somewhat higher than the 50 pM sultant supernatant was quick-frozen in liquid N2 and Km reported by Shaw et al. (14) or the 40 PM K,,, deterstored at -70°C. Fibroblast extracts were frozen and as- mined by Neumann et al. (4). This difference may reflect sayed without preheating. Protein was determined by differences in assay conditions. When the enzyme assay the Coomassie blue protein assay (Bio-Rad). was performed with a substrate concentration of 100 PM acetyl-CoA, the sensitivity was 0.0010 unit of CAT, and Mixed-phase CAT Assay. For maximum sensitivity, the mixed-phase assay was linear over three orders of samples were assayed as a complete reaction with chlormagnitude (from 0.0010 to 1.0 unit of CAT) [data not amphenicol and an extract blank was assayed without shown]). Decreasing the acetyl-CoA concentration to 30 chloramphenicol. Specific CAT activity was determined by subtracting the extract blank from the complete reac- PM dramatically increased the sensitivity to 0.000010 tion. Sixty microliters of cell extract containing 20-50 unit CAT. The standard curve in Fig. 1 was obtained pg of protein plus Buffer I was mixed with either 20 ~1 of from mixed-phase assays of samples using 30 pM acetyl5 mM chloramphenicol in Buffer I (complete reaction) or CoA and increasing amounts of purified bacterial CAT. The response was essentially linear over a range of CAT 20 ~1 of Buffer I (extract blank). After a 5-min incubaactivities spanning more than three orders of magnitude tion at 37”C, 20 ~1 of 20 &i/ml [3H]acetyl-CoA, 150 PM acetyl-CoA, in 75 PM HCl, was added to both samples of CAT activity (from 0.000010 to 0.010 unit of CAT). As long as less than approximately 30% of the acetyland the incubation was continued for 2 h at 37°C. TWO-
A MIXED-PHASE
CHLORAMPHENICOL TLC
BAnPW (a011
line,
transfmztion
ACETYLTRANSFERASE MIXED
ASSAY
vector)
TK-CAT
MCF-7,
B.
XL
Fibroblast,
XL
Liver
TK-CAT
(AllO),
VIT-CAT
PHABE
COMPLETE
EXTRACT
REACTIOl
BLAUA
cpn/!N
21
ASSAY
cwh
9
ABSAY CAT ACTIVITY
cPWq
808
18
790
102
11
93
K.U./uq
82
x lo+
7.6
~10-~
+ TK-XER
Exp.
1
97
26
71
2.4
lo+
Exp.
2
46
25
21
0.74
x 10-e
Exp.
3
40
25
16
0.56
x
x
lO-6
FIG. 2. Comparison of CAT assays of transfected cell extracts performed by extraction followed by thin-layer chromatography and by the mixed-phase assay. The indicated cell lines of human and amphibian origin were transfected by calcium phosphate coprecipitation as described under Materials and Methods. Aliquots of the transfected cell extracts were assayed for formation of [W]acetyl-chloramphenicol by thin-layer chromatography (1) and for formation of [3H]acetyl-chloramphenicol by the mixed-phase assay. The thin-layer chromatograph origin is to the left. The promoters used in the transfections are shown next to the names of the cell lines. The thin-layer and mixed-phase assays shown in panel A were performed with 50 pg of cell extract protein. The assays shown in panel B were carried out on 20 fig of cell extract protein. The autoradiogram shown in panel A was exposed for 10 h, while the autoradiogram shown in panel B was exposed for 2 days. The spots from the TLC were quantitated by scintillation counting of the scraped TLC. In panel A, 35,326, 6615, 175,910, and 16,002 cpm of acetylated products were formed by the HepG2, XL liver, MCF-7, and XL fibroblast extracts, respectively, from 342,000 cpm of substrate. Mock-transfected HepG2 cell extract formed 1234 cpm of acetylated product (TLC not shown). In panel B, 3837,1732, and 1773 cpm of acetylated products were formed by the XL liver extracts in Experiments 1,2, and 3, respectively, from 274,000 cpm of substrate. Data from the mixed-phase assay are presented as counts per minute per microgram extract protein. CAT activity in enzyme units per microgram extract was calculated from total counts per minute in the 50 and 20 pg reactions and the standard curve in Fig. 1.
CoA in the mixed-phase assay is converted to product, the reaction displays excellent linearity. The increased sensitivity of the assay and its linearity over a broad range of enzyme activities led us to employ 30 pM acetylCoA in all of the remaining assays. A lower concentration of acetyl-CoA allows the use of a higher specific radioactivity [3H]acetyl-CoA substrate, while keeping the cost of the labeled substrate well below the cost of [l*C]chloramphenicol. The mixed-phase assay is considerably less expensive than our modified [‘“Clchloramphenicol thin-layer assay (see below), which uses only one-fifth the level of costly [14C]chloramphenicol specified by Gorman et al. (1). The accuracy and sensitivity of the mixed-phase assay were determined by transfecting a variety of vertebrate cell lines with promoters linked to CAT and assaying
the resulting extracts in parallel by solvent extraction followed by thin-layer chromatography and by the mixed-phase assay. The data presented in Fig. 2 demonstrate a good correlation between results obtained with the two assays. The sensitivity of the [‘“Clchloramphenicol thin-layer chromatography assay has been optimized by using a 3-h incubation and multiple additions of acetyl-CoA while using only 0.2 &i of [‘“Clchloramphenicol per reaction. The sensitivity of the mixed-phase assay is illustrated by the facile, rapid, and quantitative detection of low levels of CAT activity, which gave weak signals by the thin-layer assay even after exposure of the autoradiogram for 2 days (Fig. 2, panel B). Although the cell lines used in this study are transfected with reasonable efficiency, many widely used mammalian cells are much more efficiently transfected.
22
NIELSEN, TABLE CAT
Assay
Complete reaction& (cpmlai3 protein)
Samples Xenopus fibroblasts (XLF) A. VIT-CAT, TKXER, +estrogen’ Experiment Experiment Experiment Experiment Experiment
1 2 3 4 5
Average of 16 total B. SVOCAT* Experiment 1 Experiment 2 Experiment 3 HepG2 cells’ Experiment Experiment
Background
1 2
CHANG,
1
in the Mixed-Phase Extract blank (-cbloramphenicol) km/pi3 protein)
Assay
CAT activity (cpm/wg protein)
436 202 265 596 211
16 20 14 16 16
416 162 271 566 195
452 f 67
16.6 f 0.6
435 f 67
25 24 37
15 17 16
10 7 21
145 210
21 20
126 190
“XLF cells were cotransfected with VIT-CAT and TKXER and grown in medium plus 100 PM estradiol-17@, and 20-50 fig of cell extract was assayed as described under Materials and Methods. The results of the first 5 transfection experiments as well as the average (&SE) of 16 transfection experiments are shown. * XLF cells were transfected with the promoterless vector SVOCAT in three experiments and 30 fig of cell extract protein was assayed. ’ HepG2 cells were transfected with pRCV12 (a rous sarcoma virus promoter-CAT plasmid) in two experiments and 50 pg of cell extract protein was assayed. d Results are presented as per microgram of extract protein.
These cell lines were chosen to provide a variety of cell lines and to illustrate the application of the mixed-phase assay to relatively low activity samples. The mixedphase assay was developed to allow quantitative determination of CAT activity in cells transfected with promoters of low transcriptional activity. Detection of very low levels of CAT activities by the mixed-phase assay depends on the ability to discriminate between radioactivity derived from CAT enzyme activity and background radioactivity from the extracts. To attain a precise determination of CAT activity in an extract, we performed parallel incubations of the complete reaction and an extract blank (which contains [3H]acetyl-CoA but lacks chloramphenicol) and subtracted the extract blank from the radioactivity obtained with the complete reaction. Our approach assumes that the extract blank, which is performed on one aliquot of the extract, is the same as the extract blank would be in the aliquot used to assay the complete reaction. TO evaluate the reproducibility of the extract blanks, we have summarized the data from transfections with sev-
AND
SHAPIRO
eral sets of vectors. Although extracts from each cell line give slightly different reaction blanks (see the table in Fig. 2, panel A), the blanks within a cell line are remarkably reproducible. The data summarized in Table 1 (part A) show the results of the first five separate transfection experiments for the expression of the vitellogenin promoter under the control of the X. Levis estrogen receptor and estradiol-17& In a total of 16 assays carried out over many months the reaction blank averaged 16.6 f 0.5 (SE) cpm/pg protein. The level of uncertainty which surrounds the extract blank to be subtracted is therefore extremely low, less than 1 cpm/pg protein. This allows even the SVOCAT vector, which lacks a promoter and exhibits extremely low background transcription in amphibian cells, to be quantitatively assayed (Table 1, part B). Because of the exceptionally high level of reproducibility of the extract blanks, transfections which employ promoters of average activity can be assayed without the use of extract blanks. The major factors contributing to the reaction blank appear to be the partition of a small quantity of the [3H]acetyl-CoA into the toluene scintillation fluor and the amount of protein in the assay samples. Samples containing extremely high levels of CAT activity (outside the linear range of this assay) exhibit an elevated extract blank, which nevertheless represents only a tiny fraction of their CAT activity. This suggests that even in the absence of chloramphenicol, CAT may be capable of converting minute amounts of acetyl-CoA to a nonpolar product. Our experience with an earlier mixed-phase assay for HMG-CoA reductase (7) led us to a number of procedures which increased the sensitivity of this assay and which differ from the recently described two-phase CAT assay based on continuous extraction of the products into scintillation fluor (43). The urea in the aqueous phase both terminates the reaction and lowers the reaction blank. Determining the optimum concentration of [3H]acetyl-CoA substrate effected a large increase in sensitivity without decreasing linearity. Comparison of the signal-to-noise ratios of the mixed-phase assay reported here and the data reported for the continuous extraction procedure of Neumann et al. (4) indicates that the mixed-phase assay we present is approximately 14 times more sensitive.* It is difficult to compare the sensitivity of the mixedphase assay with that of other CAT assays because in 4 The signal-to-noise ratios of the two assays were compared as follows: The two-phase assay of Neumann et al. (Ref. 4, Fig. 1, panel B) yields approximately 146,000 cpm of activity when 0.010 unit of CAT enzyme was assayed over a background of 9050 cpm (with no CAT), for a signal-to-noise ratio of 16:l. The mixed-phase assay yields a signal of 87,638 cpm over a background of 395 cpm for the same assay of 0.010 unit of CAT enzyme. This gives a signal-to-noise ratio of 222:1, indicating that our mixed-phase assay is 14-fold more sensitive than the continuous two-phase assay described by Neumann et al. (4)
A MIXED-PHASE
CHLORAMPHENICOL
many cases either comparison to the original thin-layer assay or quantitative expression of the data in CAT enzyme units is lacking. The sensitivity of the assay and the range of activities assayable by the mixed-phase assay are illustrated by assay of the promoterless SVOCAT vector (Table 1). The transcriptional activity of this vector in Xenopus cells is ~0.5% the activity of a vector containing a strong promoter, such as the adenovirus major late promoter CAT (data not shown). The sensitivity of the mixed-phase assay is comparable to that of the original luciferase assay. The mixedphase assay is able to detect 0.000010 unit of CAT or 1.5 X lo6 molecules of CAT (l&16). In fact, the extract of Xenopus fibroblasts transfected with SVOCAT assayed in Table 1 (part B, Experiment 2) contained only 0.000020 unit or 3 X lo6 molecules of CAT. This is equivalent to the limit of sensitivity of 3 X lo6 molecules reported for the original luciferase assay (6). However, improved luminometers effect a considerable increase in the sensitivity of subsequent luciferase assays. The mixed-phase assay presented here is an extremely sensitive assay for the determination of CAT activity in transfected cell extracts. It is extremely simple and rapid to perform, yielding reliable CAT activity determinations of transfected cell extracts in less than 3 h. Note added in proof. [3H]Acetyl-Coenzyme A from ICN Radiochemicale and from New England Nuclear has been used successfully. The
ACETYLTRANSFERASE
23
ASSAY
single lot of [3H]acetyl-Coenzyme an approximately ten fold higher
A from Ameraham assay blank.
we tested
yielded
REFERENCES 1. Gorman, C. M., Moffat, Biol. 2,1044-1051.
L. F., and Howard,
B. H. (1982)
2. Nordeen, S. K., Green, P. P., and Fowlkes, D. M. 173-178. 3. Sleigh, M. J. (1986) Anal. Biochem. 156,251-256.
(1987)
Mol.
Cell.
DNA
6,
4. Neumann, J. R., Morency, C. A., and Russian, K. 0. (1987) BioTechniques 6,444-447. 5. Eastman, A. (1987) BioTechniques 6(8), 730-732. 6. De Wet, J. R., Wood, K. V., DeLuca, M., Helinski, D. R., and Subramani, S. (1987) Mol. Cell. Biol. ‘7(2), 725-737. 7. Phillipp, B. W., and Shapiro, D. J. (1979) J. Lipid Res. 20, 588593. 8. Knowles, B. B., Howe, C. C., and Aden, D. P. (1980) Science 209, 497-499. 9. Soule, H. D., Vazquez, J., Long, A., Albert, S., and Brennan, M. (1973) J. N&l. Cancer Inst. 61(5), 1409-1416. 10. Graham,
F. L., and Van der Eb, A. J. (1973)
Virology
11. Parker, B. A., and Stark, G. R. (1979) J. Virol. 31(2), 12. Luthman, H., and Magnusson, G. (1983) Nucleic ll(ll), 1295-130s. 13. Crab, D. W., and Dixon, J. E. (1987) Anal. 14. Shaw, W. V., Bently, D. W., and Sands, 104(3), 1095-1105.
Biochem. L. (1970)
62,456-467. 360-369. Acids Res. 163,88-92. J. Bacterial.
15. Shaw, W. V., Packman, L. C., Burleigh, B. D., Dell, A., Morris, H. R., and Hartley, B. S. (1979) Nature (London) 282.870-872. 16. Shaw, W. V., andBrodsky, R. F. (1968) J. Bacterial. 96.28-36.