- Email: [email protected]
The Development of β-Lactamase as a Highly Versatile Genetic Reporter for Eukaryotic Cells
ANALYTICAL BIOCHEMISTRY ARTICLE NO.
247, 203– 209 (1997)
AB972092
The Development of b-Lactamase as a Highly Versatile Genetic Reporter for Eukaryotic Cells John T. Moore, Stephen T. Davis, and Inderjit K. Dev Glaxo Wellcome, Division of Biological Sciences, 5 Moore Drive Research Triangle Park, North Carolina 27709
Received November 19, 1996
We describe in this report that TEM-1 b-lactamase has several desirable characteristics as a genetic reporter. First, it has no endogenous counterpart in eukaryotic cells and therefore provides a backgroundfree measure of gene expression. Second, because of the uniqueness of the substrate cleavage reaction, a wide variety of substrates which are efficiently cleaved can be synthesized for b-lactamase. Third, since the assays involve no more than addition of substrate to media, it is possible to continuously monitor a culture without destruction of the cells. Fourth, the enzyme is extremely versatile in that it can be fused to other proteins and retain activity. To demonstrate the versatility of b-lactamase, we created three forms of the enzyme including secreted, intracellular, and membrane-bound forms of the enzyme, each form having distinct advantages as a reporter system. We also showed that levels of secreted b-lactamase were proportional to both the levels of transfected DNA, b-lactamase mRNA, as well as activity of the chloramphenicol acetyl transferase gene controlled by the same promoter, validating the reliability of this reporter. bLactamase thus represents a novel and highly versatile genetic reporter. q 1997 Academic Press
With the advent of molecular cloning, certain genes which code for proteins with enzymatic activity that is easily distinguishable from the endogenous proteins have been manipulated for use in monitoring biological events. Currently, the genetic reporters most widely employed in animal cells are the b-galactosidase (1), chloramphenicol acetyltransferase (CAT)1 (2), luciferase (3), and alkaline phosphatase (4) genes, the human growth hormone gene (5), and a modified human heat1 Abbreviations used: CAT, chloramphenicol acetyltransferase; PADAC, 3-(2,4-dinitrostyryl)-(6R,7R)-7-(2-thienylacetamido)-ceph-3em-4-carboxylic acid; PBS, phosphate-buffered saline; PMSF, phenylmethylsulfonyl fluoride.
stable placental alkaline phosphatase gene (6). Another reporter enzyme which is gaining in popularity is the green fluorescent protein, which can autocatalytically generate a fluorophore without addition of exogenous substrate (7, 8). Each reporter gene has certain advantages and disadvantages (9). At present, such reporter genes are used to monitor biochemical events preceding transcription to events following transcription. Such events include DNA methylation, effects of cis-acting DNA elements, RNA processing, and protein translation as well as a wide range of other processes. However, the development of novel high-throughput assay formats, such as cellcoated bead formats, lawn formats, and flow-cytometry-based cell sorting procedures has placed new demands on the desired characteristics of a genetic reporter in these systems. Reporter gene systems which are more versatile are needed to meet these demands. The enzyme TEM b-lactamase (penicillin amido-b-lactamhydrolase, EC 3.5.2.6) is an extremely versatile enzyme in that it can be fused to other proteins and retain activity (10). The enzyme is efficient, and has no mammalian endogenous counterpart activity. Two sensitive chromogenic substrates, PADAC and nitrocefin, are currently commercially available and thus assay for b-lactamase in tissue culture media requires no more than addition of these substrates and monitoring their rate of decomposition. The number of additional substrates which can be synthesized for b-lactamase is very large, thus extending the potential of this enzyme as a reporter (Fig. 1A). This paper describes three different forms of b-lactamase including secreted, intracellular, and membrane-anchored. Each form is a sensitive reporter gene with distinct advantages and can be used for a wide variety of specialized and general assays. MATERIALS AND METHODS
b-Lactamase reporter construction. In order to express b-lactamase in mammalian cells for use as a re203
0003-2697/97 $25.00 Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
AB 2092
/
6m2f$$$601
03-28-97 22:21:53
aba
204
MOORE, DAVIS, AND DEV
FIG. 1. Detection and expression of b-lactamase in mammalian cells. (A) Cephalosporin-based reporter substrate where R represents the wide variety of potential reporter molecules released upon cephalosporin ring cleavage. (B) Schematic representations of pCMV-BL, pCMVdBL, and pCMV-BLIgM. (C) b-Lactamase activity measurements were carried out on human A549 cells transfected with pCMV-BL (secretory), pCMV-dBL (intracellular), or pCMV-BL-IgM (membrane). b-Lactamase activity was measured using the chromogenic substrate PADAC. Relative expression (%) represents the percentage of activity represented in each fraction relative to the total amount of b-lactamase observed in all three cellular fractions.
porter gene, the coding region from RTEM1 b-lactamase (11) was utilized. In Escherichia coli, the gene product is secreted into the periplasm, and the unmodified coding region of b-lactamase contains a signal peptide sequence (12). To potentially enhance expression, a consensus sequence which confers optimal translation efficiency in eukaryotes (13) was added immediately adjacent to the initiator codon via PCR. The modified gene was inserted into an expression vector (pRc-CMV, InVitrogen Corp.) between the human intermediate early cytomegalovirus promoter and a downstream bovine growth hormone polyadenylation sequence. The vector containing this sequence is referred to as pCMVBL. The sequence of the insert in pCMV-BL was confirmed by dideoxy sequencing. To create a DNA construct for expression of intracellular b-lactamase, modifications to the 5* end of the blactamase gene in pCMV-BL were carried out using PCR. The signal sequence was removed and replaced by an initiator methionine. The forward primer contained a HindIII restriction site for cloning, a consensus site for optimal translation efficiency (GCCACC) in vertebrates (13), and an ATG initiator codon immediately adjacent to the sequence representing the mature
AID
AB 2092
/
6m2f$$$602
03-28-97 22:21:53
amino terminus of TEM b-lactamase (11). The b-lactamase coding region containing this modification was cloned into pRc-CMV and designated pCMV-dBL (Fig. 1B). The sequence of the insert in pCMV-dBL was confirmed by dideoxy sequencing. To create a DNA construct for expression of a membrane form of b-lactamase, the membrane-spanning domain of human IgM was appended to the coding region of secretory b-lactamase. In the first step of this modification, the termination codon in the b-lactamase sequence contained in pCMV-BL was deleted during by PCR amplification of the pCMV-BL and replaced with an XbaI site. The hexameric XbaI sequence is inframe with the coding region of b-lactamase and represents a Ser-Arg amino acid sequence. This PCR product was cloned into the HindIII and XbaI sites of pRc-CMV and designated pCMV-MEM1. To attach a carboxy-terminal membrane spanning domain, a 300-bp sequence from pIgM/TM/PCRII (kindly provided by C.-A. Ohmstede of this institution) (14) which contains the amino terminus of the IgM transmembrane domain beginning at nucleotide 489 to the carboxy terminus of the transmembrane domain and ending at nucleotide 815 (GenBank Accession No. X14939). These oligos were used
aba
b-LACTAMASE AS A EUKARYOTIC GENETIC REPORTER
to carry out PCR as described above and the approximately 300-base-pair product was restriction digested, gel-purified, and cloned into the XbaI and ApaI sites of pCMV-MEM1. The schematic diagram of the construct (designated pCMV-BLIgM) is shown in Fig. 1B. The sequence of the insert in pCMV-BLIgM was confirmed by dideoxy sequencing.
b-Lactamase assays. In in vitro assays, the chromogenic substrate PADAC (Calbiochem) was added to a final concentration of 20 mM (15) and the decrease in absorbence at OD570nm was monitored (a rate of 01 OD/ min was found to correspond to a concentration of approximately 1 mg/ml). In tissue extraction experiments, tumor lines transfected with pCMV-BL were harvested from animals; the excised cells were rinsed and then minced into single cells by a tissue mincer. The cells were then counted using a hemocytometer to normalize each sample. An equal number of cells were placed into PBS / 20 mM PADAC at 377C. In the nitrocefin assays, lungs from Balb-c mice treated by intrabrochial administration of a pCMV-BL/liposome mixture. After two days, lungs were removed, rinsed in PBS, and weighed. Tissues were then homogenized in Dounce homogenizers. Tissue extracts (50 ml) were assayed at 377C in 0.5 mM nitrocefin (Becton Dickinson, Inc.), 0.5 mM oleic acid (Sigma Corp.) in a total volume of 1 ml. Activity was measured as an increase in absorbance at 488 nm. In some samples, an inhibitor of b-lactamase, clavulanic acid (Smith/Kline/Beecham, Inc.), was added at a concentration of 25 mM (16, 17). Cell fractionation. Cells transfected with pCMVBL, pCMV-dBL, or pCMV-BLIgM were fractionated as follows. The secreted fraction consisted of the spun media from the transfected cells. Remaining cells were harvested and resuspended in 50 mM Tris –Cl (pH 7.4), 0.1 mM EDTA containing PMSF and leupeptin, swollen on ice for 10 min, then lysed using a Dounce homogenizer. The cytosolic fraction consisted of the supernatant after centrifugation at 800g for 6 min. The membrane fraction consisted of the pellet from the above as well as from recentrifugation of the supernatant at 30 psi for 20 min in a Beckman Airfuge. RNA analysis. RNA was prepared as described in Current Protocols in Molecular Biology (Section 4.2.4). Ten micrograms of purified RNA was slot blotted onto Hybond-N filters (Amersham, Arlington Heights, IL), cross-linked by UV irradiation, and probed with 32 Plabeled DNA representing pCMV-BL. Plasmid DNA was labeled by a random primers DNA labeling system (Life Technologies, Gaithersburg, MD) following manufacturer’s instructions. The relative amounts of RNA determined by hybridization of the 32P-labeled probe was quantitated using a phosphorimager (Molecular Dynamics) using ImageQuant software. RNA levels are presented in arbitrary units which represent inte-
AID
AB 2092
/
6m2f$$$602
03-28-97 22:21:53
205
grated volumes. Before slot-blot analysis, Northern blot analysis was carried out to show that a single RNA band of the appropriate size is detected in total mRNA from A549 transfected with pCMV-BL using a b-lactamase probe (data not shown). CAT assays. The CAT vector was obtained commercially (pCAT-control, Promega Corp.). CAT activity was measured in cell extracts using a mixed-phase assay (18). Transient transfections and clonal line production. A549 cells were transfected with pCMV-BL using lipofectamine (Life Technologies) according to manufacturer’s instructions and selection in G418 (Life Technologies) was carried out at 0.5 mg/ml. To screen for blactamase-producing cells, transiently transfected cells were selected for 2 weeks with G418 and subsequently plated at approximately 1 cell/well on 96-well plates. After cells had amplified to 100–200 cells/well, 100 ml was removed from each well to a new well containing 100 ml PBS and 20 mM PADAC. RESULTS AND DISCUSSION
b-Lactamase (penicillin amido-b-lactamhydrolase) is a relatively small (27,000 Da) soluble, monomeric enzyme which is located extrachromosomally on a variety of plasmids (19). The enzyme has a high catalytic efficiency (20) (kcat for PADAC substrate is 135 s01 and the difference in the extinction coefficient between cleaved and uncleaved product at 570 nm is 4 1 104 m01cm01), and has no mammalian endogenous counterpart. The assay for b-lactamase in tissue culture media requires only the addition of substrate and monitoring of the rate of hydrolysis. Two sensitive chromogenic substrates, PADAC and nitrocefin, are commercially available. Also, additional substrates with increased sensitivity or for specific assays (for example, histochemical assays) could be synthesized. The potential for multiple substrates for b-lactamase stems from the fact that any molecule which can be chemically attached to the 3 * substituent of a cephalosporin is subsequently released upon cleavage of the b-lactam ring by b-lactamase (21). Thus, reporter substrates with a wide range of properties could be used, such as fluorescent dyes or intracellular stains (Fig. 1A). The coding region from E. coli, RTEM1 b-lactamase (11) was used to create genes for secretory, intracellular, and membrane reporter enzymes (Fig. 1B). The native RTEM b-lactamase contains a signal peptide sequence which directs the gene product to the periplasm in E. coli (11, 12). The expression vector for secretory b-lactamase, pCMV-BL, was created by retaining the signal peptide. To generate an intracellular form of b-lactamase, a construct was created which was identical to pCMV-BL except that the signal peptide sequence was deleted and replaced with a methionine
aba
206
MOORE, DAVIS, AND DEV
FIG. 2. Correlation of secretory b-lactamase activity with b-lactamase mRNA levels and with CAT reporter activity. Three cell lines (A549, human lung adenocarcinoma cell line; WiDr, human colon carcinoma; and H441, human lung Clara-like cell line) were cotransfected with equal amounts of plasmids pCMV-BL and pCAT-control and subsequently assayed for b-lactamase activity, CAT activity, and blactamase RNA levels. Rows 1 –3 of the graph represent control transfections. Other samples are identified as follows: Rows 4, 6, and 7— WiDr, A549, H441, 0.5 mg total plasmid; Rows 5, 8, and 9—WiDr, A549, H441, 2.0 mg plasmid total.
codon (pCMV-dBL). To create a membrane-bound form of b-lactamase, a membrane-spanning domain was appended to the region representing the carboxy terminus of the secretory b-lactamase coding region contained in pCMV-BL. The membrane sequence was derived from the human C m IgM heavy chain gene (14). This was carried out by fusing a 300-bp sequence representing the human IgM membrane-spanning domain (from plasmid IgM/TM/PCRII which contains exons M1 and M2 separated by a single intervening sequence) in-frame to the carboxy terminus of the secretory b-lactamase gene. To test whether b-lactamase is enzymatically active when produced in human cells and to determine the cellular or extracellular location of the gene product, in vitro transient transfections using pCMV-BL, pCMV-
TABLE 1
Comparison of b-Lactamase and CAT Activity as Measured by Standard Assays in H460 Cells Cotransfected with pBLControl and pCAT-Control Plasmid Vectors Cotransfection DNA (mg)
b-Lactamase (pmol/min/ml medium)
CAT (pmol/min/mg protein)
0 0.5 2.0
10 38 173
N.D. N.D. 0.2
Note. Constructs containing b-lactamase (pCMV-BL) and CAT (pCAT-Control, Promega Corp.) were utilized in this experiment. The coding regions were contained in identical vector contexts, both under transcriptional control of the SV40 promoter. Various amounts of the plasmids were cotransfected into H460 lung cell line as shown. In some cases, CAT activity was nondetectable (ND).
AID
AB 2092
/
6m2f$$$603
03-28-97 22:21:53
dBL, and pCMV-BLIgM were carried out in A549 human lung adenocarcinoma cells. b-Lactamase activity was assessed in secreted, cytosolic, and membrane cell fractions using the chromogenic substrate PADAC (Fig. 1C). When cells were transfected with pCMV-BL, active b-lactamase was produced and greater than 90% was located extracellularly. In contrast, when the clone containing the deleted signal peptide (pCMV-dBL) was tested, greater than 90% of the activity was detected in the intracellular fraction. Finally, when the membrane form was tested, greater than 90% was located on the cell membrane. The total activity in intracellular fractions in transfection using pCMV-dBL was approximately 10-fold less than the total activity seen in media from pCMV-BL transfected cells and the activity in the pCMV-BLIgM sample was 7-fold less. In order to further characterize the polarity of the active membrane form of b-lactamase, cellular assays were carried out. Transient transfections of human lung adenocarcinoma with pCMV-BLIgM or pCMVdBL were carried out. b-Lactamase activity was then measured either in media removed from the cells or directly in media still in contact with the transfected cells on the plate. In pCMV-dBL transfections, b-lactamase activity was not detected in the media removed from the cells, indicating that little or no enzyme was secreted from the cells into the media. b-Lactamase activity was also not detected in media which remained in contact with the transfected cells when pCMV-dBL DNA was used in the transfections, indicating that the PADAC substrate does not penetrate cells. b-Lactamase activity was detected when pCMVBLIgM DNA was utilized in the transfections, but only when the
aba
b-LACTAMASE AS A EUKARYOTIC GENETIC REPORTER
207
FIG. 3. b-Lactamase assays in tissue samples. (A) Mouse lungs were transfected with pCMV-BL via liposome-mediated gene transfer. bLactamase assays were subsequently carried out in mouse lung tissue samples using the chromogenic substrate nitrocefin (27). Activity is measured as a increase in absorbance at 488 nm. To avoid the inherent breakdown of b-lactamase substrates in tissue samples (15), oleic acid was used to stabilize the substrate (27). The amount of background nitrocefin breakdown was assessed through addition of the blactamase-specific inhibitor, clavulanic acid. (B) Subcutaneous tumors containing cells pretransfected with pCMV-BL were assayed using PADAC. After harvesting tumors, de novo production of b-lactamase by single cell suspensions of the tumor cells was measured.
assay medium was in contact with the cells, indicating that the enzyme is membrane-bound and is located on the exterior face of the membrane. These experiments show that b-lactamase is amenable to targeted expression in different cellular compartments while retaining activity. All of the described forms of b-lactamase have potential utility as reporters. It is necessary that a reporter gene is expressed in a wide variety of cell types. b-Lactamase is expressed in all cell lines we have tested using pCMV-BL. Overall, we have detected expression of b-lactamase in 18 cell lines, including multiple human lung cell lines. Cell lines tested included the human established cell lines A549, lung adenocarcinoma; WiDr, colon carcinoma; NCI-H551, NCI-H209, NCI-H128, small cell carcinoma; CEM, lymphoblastoid; MOLT-4, lymphoblastoid; HCT-8 colon carcinoma; H441, Clara-like cell carcinoma; H460, large cell carcinoma; H520, squamous cell carcinoma; SK-Lu 1, lung adenocarcinoma; Detroit 562, pharynx carcinoma) and human primary cell cultures (human bone marrow stem cells), a monkey cell line (CV-1, kidney fibroblast), mouse cell lines (AtT-20, pituitary; MOPC-315, mouse myeloma), and a rat cell line (A-10, rat smooth muscle). Another important attribute of a reporter gene is that the steady-state level of secreted enzymatic activity is representative of steady-state mRNA levels. This has indeed been demonstrated for several reporter genes including CAT and alkaline phosphatase (22, 23). In contrast, the secretion of hGH has not been a reliable estimate of mRNA levels in all cell lines (24, 25). The reliability of secretory b-lactamase as a reporter was
AID
AB 2092
/
6m2f$$$603
03-28-97 22:21:53
established by determining a relationship between extracellular b-lactamase activity and b-lactamase mRNA steady-state levels in three cell lines (A549, human lung adenocarcinoma cell line; WiDr, human colon carcinoma; and H441, human lung Clara-like cell line). Relative RNA levels were determined by dot-blot hybridization followed by densitometry. We found that increases in b-lactamase activity in the growth medium in general correlated with increases in the steady-state b-lactamase mRNA levels, thus changes in extracellular b-lactamase activity reflect changes in the relative abundance of b-lactamase mRNA (Fig. 2). Additionally, extracellular b-lactamase activity in these cell lines also increased in parallel with CAT (r2 Å 0.97), an independent established reporter gene (Fig. 2). The b-lactamase gene is a very sensitive reporter. When CAT and b-lactamase activities were compared in cotransfection experiments in human H460 cells, blactamase activity was detected under conditions in which CAT activity was undetectable (Table 1). In this experiment, both reporters were cloned downstream from the SV40 promoter to compare them under equal promoter strengths. Using purified bacterial b-lactamase (Sigma P3553), the detection limit in growth medium (RPMI 1640/ 10% fetal bovine serum) was 108 –109 molecules or approximately 0.02 ng using the PADAC assay. Assuming that the mammalian-expressed enzyme has the same specific activity, the limit of detection of b-lactamase using the PADAC assay is similar to the standard hGH and CAT assays (26) and may exceed the sensitivity of these reporters under certain circumstances.
aba
208
MOORE, DAVIS, AND DEV
FIG. 4. Utilization of b-lactamase reporter. A549 lung adenocarcinoma cells were transfected with plasmid pCMV-BL which contains the secretory b-lactamase reporter gene as well as the neomycin-resistance gene. Clonal lines of transfected cells were screened for b-lactamase activity as described under Materials and Methods. The photograph shows the visual impression of the plate after 20 min incubation at room temperature in the presence of PADAC. Yellow wells identify media from cells producing significant levels of b-lactamase.
In addition to cultured cells, we found that b-lactamase can be used as a reporter in cells after either (A) in vivo transfection of pCMV-BL DNA into tissue or (B) implantation and growth of cells pretransfected with pCMV-BL. b-Lactamase activity was measured in pCMV-BL-transfected tissues using two different methods. In the first, mouse lungs were transfected with pCMV-BL using liposome-mediated gene transfer (Fig. 3A). After transfection, b-lactamase activity was measured in lung tissue homogenates using the substrate nitrocefin (activity monitored by increase in A488nm). The b-lactamase-specific inhibitor, clavulanic acid, was used to determine the background level of substrate hydrolysis. In a second test, a stable clonal lung adenocarcinoma cell line expressing secretory blactamase was utilized. Cells were harvested after subcutaneous growth in mice, disaggregated to give rise to a single cell suspension, and washed to remove serum, and de novo b-lactamase production was assayed using PADAC as a substrate (activity monitored by decrease in A570nm) (Fig. 3B). Thus, using two different
AID
AB 2092
/
6m2f$$$603
03-28-97 22:21:53
approaches, we have shown that b-lactamase is useful as a reporter for gene activity. To date at our institution, the b-lactamase reporter has been used to evaluate the strength and specificity of DNA control elements, to determine the relative efficiencies of novel gene transfer reagents, to determine the effects of compounds on DNA repair and for use in the development of high-throughput screens for antiviral compounds, and to screen for stably transformed lines (an example of the use of the b-lactamase assay to screen transformed lines in a 96-well plate format is shown in Fig. 4). From the animal data, it is clear that b-lactamase could also be a very useful reporter in in vivo experiments in which gene transfer must be monitored, as in gene therapy experiments. The coding region of b-lactamase is very amenable to modification for customized reporter usage. For example, the intracellular form of b-lactamase may be optimal for use with a histochemical stain to determine cell-specific expression. We have also shown that membrane and secreted chimeras fused to peptide hormones as well as
aba
b-LACTAMASE AS A EUKARYOTIC GENETIC REPORTER
another protein (leptin) at the carboxy terminus retain enzyme activity (unpublished data). These results demonstrate the utility of the b-lactamase gene as a sensitive and reliable secreted reporter, while also demonstrating the flexibility of the gene in individualized reporter assays due to the high degree of choice of both enzyme and substrate forms. REFERENCES 1. Hidaka, G. K., and Siminovitch, L. (1982) Mol. Cell. Biol. 2, 1628 – 1632. 2. Gorman, C. M., Moffat, L. F., and Howard, B. H. (1982) Mol. Cell. Biol. 2, 1044 – 1049. 3. Gould, S. J., and Subramani, S. (1988) Anal. Biochem. 175, 5– 11. 4. Yoon, K., Thiede, M. A., and Rodan, G. A. (1988) Gene 66, 11– 17. 5. Berger, J., Hauber, J., Hauber, R., Geiger, R., and Cullen, B. R. (1988) Gene 66, 1–7. 6. Cullen, B. R., and Malim, M. H. (1992) in Recombinant DNA (Wu, R., Ed.), Vol. 216, Part G, Academic Press, San Diego. 7. Chalfie, M., Tu, Y., Euskirchen, G., Ward, W., and Prasher, D. (1994) Science 263, 802– 805. 8. Wood, K. V. (1995) Curr. Opin. Biotechnol. 6, 50–58. 9. Alam, J., and Cook, J. L. (1990) Anal. Biochem. 188, 245 –254. 10. Broome-Smith, J. K., Taddayyon, M., and Zhang, Y. (1990) Mol. Microbiol. 4, 1637 –1644. 11. Sutcliffe, J. G. (1978) Proc. Natl. Acad. Sci. USA 75, 3737 –3741.
AID
AB 2092
/
6m2f$$$604
03-28-97 22:21:53
209
12. Ambler, R. P., and Scott, G. K. (1978) Proc. Natl. Acad. Sci. USA 75, 3732 –3736. 13. Kozak, M. (1991) J. Cell Biol. 115, 887– 903. 14. Dorai, H. (1989) Nucleic Acids Res. 17, 6412 –6414. 15. O’Callaghan, C. H., Morris, A., Kirby, S. M., and Shingler, A. H. (1972) Antimicrob. Agents Chemother. 1, 283– 288. 16. Thompson, J. S., and Malamy, M. H. (1990) J. Bacteriol. 172, 2584 –2593. 17. Brown, A. G., Butterworth, D., Cole, M., Hanscombe, G., Hood, J. D., Reading, C., and Rolinson, G. N. (1976) J. Antibiot. 29, 668–669. 18. Nielsen, D., Chuang, T.-C., and Shapiro, D. J. (1989) Anal. Biochem. 179, 19–23. 19. Sykes, R. B., and Matthew, M. (1976) J. Antimicrob. Chemother. 2, 115–157. 20. Meyer, D. L., Jungheim, L. N., Mikolajczyk, S. D., Shepherd, T. A., and Starling, J. J. (1992) Bioconjugate Chem. 3, 42– 48. 21. Albrecht, H. A., et al. (1991) J. Med. Chem. 34, 669– 675. 22. Gorman, C. M., Moffat, L. F., and Howard, B. H. (1982) Mol. Cell. Biol. 2, 1044 –1049. 23. Berger, J., Hauber, J., Hauber, R., Geiger, R., and Cullen, B. R. (1988) Gene 66, 1– 7. 24. Chu, G., Hayakawa, H., and Berg, P. (1987) Nucleic Acids Res. 15, 1311 –1326. 25. Pavlakis, G. N., and Hamer, D. H. (1983) Proc. Natl. Acad. Sci. USA 80, 397– 401. 26. Yoon, K., Thiede, M. A., and Rodan, G. A. (1988) Gene 66, 11– 17. 27. Nerli, B., and Pico, G. (1994) Biochem. Mol. Biol. Int. 32, 789– 795.
aba
247, 203– 209 (1997)
AB972092
The Development of b-Lactamase as a Highly Versatile Genetic Reporter for Eukaryotic Cells John T. Moore, Stephen T. Davis, and Inderjit K. Dev Glaxo Wellcome, Division of Biological Sciences, 5 Moore Drive Research Triangle Park, North Carolina 27709
Received November 19, 1996
We describe in this report that TEM-1 b-lactamase has several desirable characteristics as a genetic reporter. First, it has no endogenous counterpart in eukaryotic cells and therefore provides a backgroundfree measure of gene expression. Second, because of the uniqueness of the substrate cleavage reaction, a wide variety of substrates which are efficiently cleaved can be synthesized for b-lactamase. Third, since the assays involve no more than addition of substrate to media, it is possible to continuously monitor a culture without destruction of the cells. Fourth, the enzyme is extremely versatile in that it can be fused to other proteins and retain activity. To demonstrate the versatility of b-lactamase, we created three forms of the enzyme including secreted, intracellular, and membrane-bound forms of the enzyme, each form having distinct advantages as a reporter system. We also showed that levels of secreted b-lactamase were proportional to both the levels of transfected DNA, b-lactamase mRNA, as well as activity of the chloramphenicol acetyl transferase gene controlled by the same promoter, validating the reliability of this reporter. bLactamase thus represents a novel and highly versatile genetic reporter. q 1997 Academic Press
With the advent of molecular cloning, certain genes which code for proteins with enzymatic activity that is easily distinguishable from the endogenous proteins have been manipulated for use in monitoring biological events. Currently, the genetic reporters most widely employed in animal cells are the b-galactosidase (1), chloramphenicol acetyltransferase (CAT)1 (2), luciferase (3), and alkaline phosphatase (4) genes, the human growth hormone gene (5), and a modified human heat1 Abbreviations used: CAT, chloramphenicol acetyltransferase; PADAC, 3-(2,4-dinitrostyryl)-(6R,7R)-7-(2-thienylacetamido)-ceph-3em-4-carboxylic acid; PBS, phosphate-buffered saline; PMSF, phenylmethylsulfonyl fluoride.
stable placental alkaline phosphatase gene (6). Another reporter enzyme which is gaining in popularity is the green fluorescent protein, which can autocatalytically generate a fluorophore without addition of exogenous substrate (7, 8). Each reporter gene has certain advantages and disadvantages (9). At present, such reporter genes are used to monitor biochemical events preceding transcription to events following transcription. Such events include DNA methylation, effects of cis-acting DNA elements, RNA processing, and protein translation as well as a wide range of other processes. However, the development of novel high-throughput assay formats, such as cellcoated bead formats, lawn formats, and flow-cytometry-based cell sorting procedures has placed new demands on the desired characteristics of a genetic reporter in these systems. Reporter gene systems which are more versatile are needed to meet these demands. The enzyme TEM b-lactamase (penicillin amido-b-lactamhydrolase, EC 3.5.2.6) is an extremely versatile enzyme in that it can be fused to other proteins and retain activity (10). The enzyme is efficient, and has no mammalian endogenous counterpart activity. Two sensitive chromogenic substrates, PADAC and nitrocefin, are currently commercially available and thus assay for b-lactamase in tissue culture media requires no more than addition of these substrates and monitoring their rate of decomposition. The number of additional substrates which can be synthesized for b-lactamase is very large, thus extending the potential of this enzyme as a reporter (Fig. 1A). This paper describes three different forms of b-lactamase including secreted, intracellular, and membrane-anchored. Each form is a sensitive reporter gene with distinct advantages and can be used for a wide variety of specialized and general assays. MATERIALS AND METHODS
b-Lactamase reporter construction. In order to express b-lactamase in mammalian cells for use as a re203
0003-2697/97 $25.00 Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
AB 2092
/
6m2f$$$601
03-28-97 22:21:53
aba
204
MOORE, DAVIS, AND DEV
FIG. 1. Detection and expression of b-lactamase in mammalian cells. (A) Cephalosporin-based reporter substrate where R represents the wide variety of potential reporter molecules released upon cephalosporin ring cleavage. (B) Schematic representations of pCMV-BL, pCMVdBL, and pCMV-BLIgM. (C) b-Lactamase activity measurements were carried out on human A549 cells transfected with pCMV-BL (secretory), pCMV-dBL (intracellular), or pCMV-BL-IgM (membrane). b-Lactamase activity was measured using the chromogenic substrate PADAC. Relative expression (%) represents the percentage of activity represented in each fraction relative to the total amount of b-lactamase observed in all three cellular fractions.
porter gene, the coding region from RTEM1 b-lactamase (11) was utilized. In Escherichia coli, the gene product is secreted into the periplasm, and the unmodified coding region of b-lactamase contains a signal peptide sequence (12). To potentially enhance expression, a consensus sequence which confers optimal translation efficiency in eukaryotes (13) was added immediately adjacent to the initiator codon via PCR. The modified gene was inserted into an expression vector (pRc-CMV, InVitrogen Corp.) between the human intermediate early cytomegalovirus promoter and a downstream bovine growth hormone polyadenylation sequence. The vector containing this sequence is referred to as pCMVBL. The sequence of the insert in pCMV-BL was confirmed by dideoxy sequencing. To create a DNA construct for expression of intracellular b-lactamase, modifications to the 5* end of the blactamase gene in pCMV-BL were carried out using PCR. The signal sequence was removed and replaced by an initiator methionine. The forward primer contained a HindIII restriction site for cloning, a consensus site for optimal translation efficiency (GCCACC) in vertebrates (13), and an ATG initiator codon immediately adjacent to the sequence representing the mature
AID
AB 2092
/
6m2f$$$602
03-28-97 22:21:53
amino terminus of TEM b-lactamase (11). The b-lactamase coding region containing this modification was cloned into pRc-CMV and designated pCMV-dBL (Fig. 1B). The sequence of the insert in pCMV-dBL was confirmed by dideoxy sequencing. To create a DNA construct for expression of a membrane form of b-lactamase, the membrane-spanning domain of human IgM was appended to the coding region of secretory b-lactamase. In the first step of this modification, the termination codon in the b-lactamase sequence contained in pCMV-BL was deleted during by PCR amplification of the pCMV-BL and replaced with an XbaI site. The hexameric XbaI sequence is inframe with the coding region of b-lactamase and represents a Ser-Arg amino acid sequence. This PCR product was cloned into the HindIII and XbaI sites of pRc-CMV and designated pCMV-MEM1. To attach a carboxy-terminal membrane spanning domain, a 300-bp sequence from pIgM/TM/PCRII (kindly provided by C.-A. Ohmstede of this institution) (14) which contains the amino terminus of the IgM transmembrane domain beginning at nucleotide 489 to the carboxy terminus of the transmembrane domain and ending at nucleotide 815 (GenBank Accession No. X14939). These oligos were used
aba
b-LACTAMASE AS A EUKARYOTIC GENETIC REPORTER
to carry out PCR as described above and the approximately 300-base-pair product was restriction digested, gel-purified, and cloned into the XbaI and ApaI sites of pCMV-MEM1. The schematic diagram of the construct (designated pCMV-BLIgM) is shown in Fig. 1B. The sequence of the insert in pCMV-BLIgM was confirmed by dideoxy sequencing.
b-Lactamase assays. In in vitro assays, the chromogenic substrate PADAC (Calbiochem) was added to a final concentration of 20 mM (15) and the decrease in absorbence at OD570nm was monitored (a rate of 01 OD/ min was found to correspond to a concentration of approximately 1 mg/ml). In tissue extraction experiments, tumor lines transfected with pCMV-BL were harvested from animals; the excised cells were rinsed and then minced into single cells by a tissue mincer. The cells were then counted using a hemocytometer to normalize each sample. An equal number of cells were placed into PBS / 20 mM PADAC at 377C. In the nitrocefin assays, lungs from Balb-c mice treated by intrabrochial administration of a pCMV-BL/liposome mixture. After two days, lungs were removed, rinsed in PBS, and weighed. Tissues were then homogenized in Dounce homogenizers. Tissue extracts (50 ml) were assayed at 377C in 0.5 mM nitrocefin (Becton Dickinson, Inc.), 0.5 mM oleic acid (Sigma Corp.) in a total volume of 1 ml. Activity was measured as an increase in absorbance at 488 nm. In some samples, an inhibitor of b-lactamase, clavulanic acid (Smith/Kline/Beecham, Inc.), was added at a concentration of 25 mM (16, 17). Cell fractionation. Cells transfected with pCMVBL, pCMV-dBL, or pCMV-BLIgM were fractionated as follows. The secreted fraction consisted of the spun media from the transfected cells. Remaining cells were harvested and resuspended in 50 mM Tris –Cl (pH 7.4), 0.1 mM EDTA containing PMSF and leupeptin, swollen on ice for 10 min, then lysed using a Dounce homogenizer. The cytosolic fraction consisted of the supernatant after centrifugation at 800g for 6 min. The membrane fraction consisted of the pellet from the above as well as from recentrifugation of the supernatant at 30 psi for 20 min in a Beckman Airfuge. RNA analysis. RNA was prepared as described in Current Protocols in Molecular Biology (Section 4.2.4). Ten micrograms of purified RNA was slot blotted onto Hybond-N filters (Amersham, Arlington Heights, IL), cross-linked by UV irradiation, and probed with 32 Plabeled DNA representing pCMV-BL. Plasmid DNA was labeled by a random primers DNA labeling system (Life Technologies, Gaithersburg, MD) following manufacturer’s instructions. The relative amounts of RNA determined by hybridization of the 32P-labeled probe was quantitated using a phosphorimager (Molecular Dynamics) using ImageQuant software. RNA levels are presented in arbitrary units which represent inte-
AID
AB 2092
/
6m2f$$$602
03-28-97 22:21:53
205
grated volumes. Before slot-blot analysis, Northern blot analysis was carried out to show that a single RNA band of the appropriate size is detected in total mRNA from A549 transfected with pCMV-BL using a b-lactamase probe (data not shown). CAT assays. The CAT vector was obtained commercially (pCAT-control, Promega Corp.). CAT activity was measured in cell extracts using a mixed-phase assay (18). Transient transfections and clonal line production. A549 cells were transfected with pCMV-BL using lipofectamine (Life Technologies) according to manufacturer’s instructions and selection in G418 (Life Technologies) was carried out at 0.5 mg/ml. To screen for blactamase-producing cells, transiently transfected cells were selected for 2 weeks with G418 and subsequently plated at approximately 1 cell/well on 96-well plates. After cells had amplified to 100–200 cells/well, 100 ml was removed from each well to a new well containing 100 ml PBS and 20 mM PADAC. RESULTS AND DISCUSSION
b-Lactamase (penicillin amido-b-lactamhydrolase) is a relatively small (27,000 Da) soluble, monomeric enzyme which is located extrachromosomally on a variety of plasmids (19). The enzyme has a high catalytic efficiency (20) (kcat for PADAC substrate is 135 s01 and the difference in the extinction coefficient between cleaved and uncleaved product at 570 nm is 4 1 104 m01cm01), and has no mammalian endogenous counterpart. The assay for b-lactamase in tissue culture media requires only the addition of substrate and monitoring of the rate of hydrolysis. Two sensitive chromogenic substrates, PADAC and nitrocefin, are commercially available. Also, additional substrates with increased sensitivity or for specific assays (for example, histochemical assays) could be synthesized. The potential for multiple substrates for b-lactamase stems from the fact that any molecule which can be chemically attached to the 3 * substituent of a cephalosporin is subsequently released upon cleavage of the b-lactam ring by b-lactamase (21). Thus, reporter substrates with a wide range of properties could be used, such as fluorescent dyes or intracellular stains (Fig. 1A). The coding region from E. coli, RTEM1 b-lactamase (11) was used to create genes for secretory, intracellular, and membrane reporter enzymes (Fig. 1B). The native RTEM b-lactamase contains a signal peptide sequence which directs the gene product to the periplasm in E. coli (11, 12). The expression vector for secretory b-lactamase, pCMV-BL, was created by retaining the signal peptide. To generate an intracellular form of b-lactamase, a construct was created which was identical to pCMV-BL except that the signal peptide sequence was deleted and replaced with a methionine
aba
206
MOORE, DAVIS, AND DEV
FIG. 2. Correlation of secretory b-lactamase activity with b-lactamase mRNA levels and with CAT reporter activity. Three cell lines (A549, human lung adenocarcinoma cell line; WiDr, human colon carcinoma; and H441, human lung Clara-like cell line) were cotransfected with equal amounts of plasmids pCMV-BL and pCAT-control and subsequently assayed for b-lactamase activity, CAT activity, and blactamase RNA levels. Rows 1 –3 of the graph represent control transfections. Other samples are identified as follows: Rows 4, 6, and 7— WiDr, A549, H441, 0.5 mg total plasmid; Rows 5, 8, and 9—WiDr, A549, H441, 2.0 mg plasmid total.
codon (pCMV-dBL). To create a membrane-bound form of b-lactamase, a membrane-spanning domain was appended to the region representing the carboxy terminus of the secretory b-lactamase coding region contained in pCMV-BL. The membrane sequence was derived from the human C m IgM heavy chain gene (14). This was carried out by fusing a 300-bp sequence representing the human IgM membrane-spanning domain (from plasmid IgM/TM/PCRII which contains exons M1 and M2 separated by a single intervening sequence) in-frame to the carboxy terminus of the secretory b-lactamase gene. To test whether b-lactamase is enzymatically active when produced in human cells and to determine the cellular or extracellular location of the gene product, in vitro transient transfections using pCMV-BL, pCMV-
TABLE 1
Comparison of b-Lactamase and CAT Activity as Measured by Standard Assays in H460 Cells Cotransfected with pBLControl and pCAT-Control Plasmid Vectors Cotransfection DNA (mg)
b-Lactamase (pmol/min/ml medium)
CAT (pmol/min/mg protein)
0 0.5 2.0
10 38 173
N.D. N.D. 0.2
Note. Constructs containing b-lactamase (pCMV-BL) and CAT (pCAT-Control, Promega Corp.) were utilized in this experiment. The coding regions were contained in identical vector contexts, both under transcriptional control of the SV40 promoter. Various amounts of the plasmids were cotransfected into H460 lung cell line as shown. In some cases, CAT activity was nondetectable (ND).
AID
AB 2092
/
6m2f$$$603
03-28-97 22:21:53
dBL, and pCMV-BLIgM were carried out in A549 human lung adenocarcinoma cells. b-Lactamase activity was assessed in secreted, cytosolic, and membrane cell fractions using the chromogenic substrate PADAC (Fig. 1C). When cells were transfected with pCMV-BL, active b-lactamase was produced and greater than 90% was located extracellularly. In contrast, when the clone containing the deleted signal peptide (pCMV-dBL) was tested, greater than 90% of the activity was detected in the intracellular fraction. Finally, when the membrane form was tested, greater than 90% was located on the cell membrane. The total activity in intracellular fractions in transfection using pCMV-dBL was approximately 10-fold less than the total activity seen in media from pCMV-BL transfected cells and the activity in the pCMV-BLIgM sample was 7-fold less. In order to further characterize the polarity of the active membrane form of b-lactamase, cellular assays were carried out. Transient transfections of human lung adenocarcinoma with pCMV-BLIgM or pCMVdBL were carried out. b-Lactamase activity was then measured either in media removed from the cells or directly in media still in contact with the transfected cells on the plate. In pCMV-dBL transfections, b-lactamase activity was not detected in the media removed from the cells, indicating that little or no enzyme was secreted from the cells into the media. b-Lactamase activity was also not detected in media which remained in contact with the transfected cells when pCMV-dBL DNA was used in the transfections, indicating that the PADAC substrate does not penetrate cells. b-Lactamase activity was detected when pCMVBLIgM DNA was utilized in the transfections, but only when the
aba
b-LACTAMASE AS A EUKARYOTIC GENETIC REPORTER
207
FIG. 3. b-Lactamase assays in tissue samples. (A) Mouse lungs were transfected with pCMV-BL via liposome-mediated gene transfer. bLactamase assays were subsequently carried out in mouse lung tissue samples using the chromogenic substrate nitrocefin (27). Activity is measured as a increase in absorbance at 488 nm. To avoid the inherent breakdown of b-lactamase substrates in tissue samples (15), oleic acid was used to stabilize the substrate (27). The amount of background nitrocefin breakdown was assessed through addition of the blactamase-specific inhibitor, clavulanic acid. (B) Subcutaneous tumors containing cells pretransfected with pCMV-BL were assayed using PADAC. After harvesting tumors, de novo production of b-lactamase by single cell suspensions of the tumor cells was measured.
assay medium was in contact with the cells, indicating that the enzyme is membrane-bound and is located on the exterior face of the membrane. These experiments show that b-lactamase is amenable to targeted expression in different cellular compartments while retaining activity. All of the described forms of b-lactamase have potential utility as reporters. It is necessary that a reporter gene is expressed in a wide variety of cell types. b-Lactamase is expressed in all cell lines we have tested using pCMV-BL. Overall, we have detected expression of b-lactamase in 18 cell lines, including multiple human lung cell lines. Cell lines tested included the human established cell lines A549, lung adenocarcinoma; WiDr, colon carcinoma; NCI-H551, NCI-H209, NCI-H128, small cell carcinoma; CEM, lymphoblastoid; MOLT-4, lymphoblastoid; HCT-8 colon carcinoma; H441, Clara-like cell carcinoma; H460, large cell carcinoma; H520, squamous cell carcinoma; SK-Lu 1, lung adenocarcinoma; Detroit 562, pharynx carcinoma) and human primary cell cultures (human bone marrow stem cells), a monkey cell line (CV-1, kidney fibroblast), mouse cell lines (AtT-20, pituitary; MOPC-315, mouse myeloma), and a rat cell line (A-10, rat smooth muscle). Another important attribute of a reporter gene is that the steady-state level of secreted enzymatic activity is representative of steady-state mRNA levels. This has indeed been demonstrated for several reporter genes including CAT and alkaline phosphatase (22, 23). In contrast, the secretion of hGH has not been a reliable estimate of mRNA levels in all cell lines (24, 25). The reliability of secretory b-lactamase as a reporter was
AID
AB 2092
/
6m2f$$$603
03-28-97 22:21:53
established by determining a relationship between extracellular b-lactamase activity and b-lactamase mRNA steady-state levels in three cell lines (A549, human lung adenocarcinoma cell line; WiDr, human colon carcinoma; and H441, human lung Clara-like cell line). Relative RNA levels were determined by dot-blot hybridization followed by densitometry. We found that increases in b-lactamase activity in the growth medium in general correlated with increases in the steady-state b-lactamase mRNA levels, thus changes in extracellular b-lactamase activity reflect changes in the relative abundance of b-lactamase mRNA (Fig. 2). Additionally, extracellular b-lactamase activity in these cell lines also increased in parallel with CAT (r2 Å 0.97), an independent established reporter gene (Fig. 2). The b-lactamase gene is a very sensitive reporter. When CAT and b-lactamase activities were compared in cotransfection experiments in human H460 cells, blactamase activity was detected under conditions in which CAT activity was undetectable (Table 1). In this experiment, both reporters were cloned downstream from the SV40 promoter to compare them under equal promoter strengths. Using purified bacterial b-lactamase (Sigma P3553), the detection limit in growth medium (RPMI 1640/ 10% fetal bovine serum) was 108 –109 molecules or approximately 0.02 ng using the PADAC assay. Assuming that the mammalian-expressed enzyme has the same specific activity, the limit of detection of b-lactamase using the PADAC assay is similar to the standard hGH and CAT assays (26) and may exceed the sensitivity of these reporters under certain circumstances.
aba
208
MOORE, DAVIS, AND DEV
FIG. 4. Utilization of b-lactamase reporter. A549 lung adenocarcinoma cells were transfected with plasmid pCMV-BL which contains the secretory b-lactamase reporter gene as well as the neomycin-resistance gene. Clonal lines of transfected cells were screened for b-lactamase activity as described under Materials and Methods. The photograph shows the visual impression of the plate after 20 min incubation at room temperature in the presence of PADAC. Yellow wells identify media from cells producing significant levels of b-lactamase.
In addition to cultured cells, we found that b-lactamase can be used as a reporter in cells after either (A) in vivo transfection of pCMV-BL DNA into tissue or (B) implantation and growth of cells pretransfected with pCMV-BL. b-Lactamase activity was measured in pCMV-BL-transfected tissues using two different methods. In the first, mouse lungs were transfected with pCMV-BL using liposome-mediated gene transfer (Fig. 3A). After transfection, b-lactamase activity was measured in lung tissue homogenates using the substrate nitrocefin (activity monitored by increase in A488nm). The b-lactamase-specific inhibitor, clavulanic acid, was used to determine the background level of substrate hydrolysis. In a second test, a stable clonal lung adenocarcinoma cell line expressing secretory blactamase was utilized. Cells were harvested after subcutaneous growth in mice, disaggregated to give rise to a single cell suspension, and washed to remove serum, and de novo b-lactamase production was assayed using PADAC as a substrate (activity monitored by decrease in A570nm) (Fig. 3B). Thus, using two different
AID
AB 2092
/
6m2f$$$603
03-28-97 22:21:53
approaches, we have shown that b-lactamase is useful as a reporter for gene activity. To date at our institution, the b-lactamase reporter has been used to evaluate the strength and specificity of DNA control elements, to determine the relative efficiencies of novel gene transfer reagents, to determine the effects of compounds on DNA repair and for use in the development of high-throughput screens for antiviral compounds, and to screen for stably transformed lines (an example of the use of the b-lactamase assay to screen transformed lines in a 96-well plate format is shown in Fig. 4). From the animal data, it is clear that b-lactamase could also be a very useful reporter in in vivo experiments in which gene transfer must be monitored, as in gene therapy experiments. The coding region of b-lactamase is very amenable to modification for customized reporter usage. For example, the intracellular form of b-lactamase may be optimal for use with a histochemical stain to determine cell-specific expression. We have also shown that membrane and secreted chimeras fused to peptide hormones as well as
aba
b-LACTAMASE AS A EUKARYOTIC GENETIC REPORTER
another protein (leptin) at the carboxy terminus retain enzyme activity (unpublished data). These results demonstrate the utility of the b-lactamase gene as a sensitive and reliable secreted reporter, while also demonstrating the flexibility of the gene in individualized reporter assays due to the high degree of choice of both enzyme and substrate forms. REFERENCES 1. Hidaka, G. K., and Siminovitch, L. (1982) Mol. Cell. Biol. 2, 1628 – 1632. 2. Gorman, C. M., Moffat, L. F., and Howard, B. H. (1982) Mol. Cell. Biol. 2, 1044 – 1049. 3. Gould, S. J., and Subramani, S. (1988) Anal. Biochem. 175, 5– 11. 4. Yoon, K., Thiede, M. A., and Rodan, G. A. (1988) Gene 66, 11– 17. 5. Berger, J., Hauber, J., Hauber, R., Geiger, R., and Cullen, B. R. (1988) Gene 66, 1–7. 6. Cullen, B. R., and Malim, M. H. (1992) in Recombinant DNA (Wu, R., Ed.), Vol. 216, Part G, Academic Press, San Diego. 7. Chalfie, M., Tu, Y., Euskirchen, G., Ward, W., and Prasher, D. (1994) Science 263, 802– 805. 8. Wood, K. V. (1995) Curr. Opin. Biotechnol. 6, 50–58. 9. Alam, J., and Cook, J. L. (1990) Anal. Biochem. 188, 245 –254. 10. Broome-Smith, J. K., Taddayyon, M., and Zhang, Y. (1990) Mol. Microbiol. 4, 1637 –1644. 11. Sutcliffe, J. G. (1978) Proc. Natl. Acad. Sci. USA 75, 3737 –3741.
AID
AB 2092
/
6m2f$$$604
03-28-97 22:21:53
209
12. Ambler, R. P., and Scott, G. K. (1978) Proc. Natl. Acad. Sci. USA 75, 3732 –3736. 13. Kozak, M. (1991) J. Cell Biol. 115, 887– 903. 14. Dorai, H. (1989) Nucleic Acids Res. 17, 6412 –6414. 15. O’Callaghan, C. H., Morris, A., Kirby, S. M., and Shingler, A. H. (1972) Antimicrob. Agents Chemother. 1, 283– 288. 16. Thompson, J. S., and Malamy, M. H. (1990) J. Bacteriol. 172, 2584 –2593. 17. Brown, A. G., Butterworth, D., Cole, M., Hanscombe, G., Hood, J. D., Reading, C., and Rolinson, G. N. (1976) J. Antibiot. 29, 668–669. 18. Nielsen, D., Chuang, T.-C., and Shapiro, D. J. (1989) Anal. Biochem. 179, 19–23. 19. Sykes, R. B., and Matthew, M. (1976) J. Antimicrob. Chemother. 2, 115–157. 20. Meyer, D. L., Jungheim, L. N., Mikolajczyk, S. D., Shepherd, T. A., and Starling, J. J. (1992) Bioconjugate Chem. 3, 42– 48. 21. Albrecht, H. A., et al. (1991) J. Med. Chem. 34, 669– 675. 22. Gorman, C. M., Moffat, L. F., and Howard, B. H. (1982) Mol. Cell. Biol. 2, 1044 –1049. 23. Berger, J., Hauber, J., Hauber, R., Geiger, R., and Cullen, B. R. (1988) Gene 66, 1– 7. 24. Chu, G., Hayakawa, H., and Berg, P. (1987) Nucleic Acids Res. 15, 1311 –1326. 25. Pavlakis, G. N., and Hamer, D. H. (1983) Proc. Natl. Acad. Sci. USA 80, 397– 401. 26. Yoon, K., Thiede, M. A., and Rodan, G. A. (1988) Gene 66, 11– 17. 27. Nerli, B., and Pico, G. (1994) Biochem. Mol. Biol. Int. 32, 789– 795.
aba