How would you define a quantitative and qualitative assessment of reporter gene constructs?

1. Why was β-gal used in this study?

2. How would you define a quantitative and qualitative assessment of reporter gene constructs?

3. How is the consensus sequence necessary for gene expression to occur?

4. Please explain figure 1 in your own words. Are these results qualitative or quantitative? How do you know?

5. In Figure 2, what was the best promoter to use? In your explanation, please provide evidence through your interpretation of this figure.

transformed cell differentiation. Proc. Natl. Acad. Sci. USA 93:5705-5708.

11.Simanis, V. and P. Nurse. 1989. Characteri- zation of the fission yeast cdc10+ protein that is required for commitment to the cell cycle. J. Cell Sci. 92:51-56.

The authors would like to thank Victoria M. Richon and Paul A. Marks at the Memo- rial Sloan-Kettering Cancer Center, NY, NY, for kindly providing the SAHA and CBHA hybrid polar compounds. Address corre- spondence to Arthur B. Pardee, Division of Cancer Biology, Dana-Farber Cancer Insti- tute, 44 Binney Street, Boston, MA 02115, USA. Internet:

Received 14 June 1999; accepted 13 August 1999.

H. Melichar, I. Bosch, G.M. Molnar, L. Huang and A.B. Pardee Dana-Farber Cancer Institute Harvard Medical School Boston, MA, USA

Green Fluorescent Protein as a Quantitative Reporter of Relative Promoter Activity in E. coli BioTechniques 28:82-89 (January 2000)


Green fluorescent protein (GFP) has be- come a valuable tool for the detection of gene expression in prokaryotes and eukary- otes. To evaluate its potential for quantita- tion of relative promoter activity in E. coli, we have compared GFP with the commonly used reporter gene lacZ, encoding β-galac- tosidase. We cloned a series of previously characterized synthetic E. coli promoters into GFP and β-galactosidase reporter vec- tors. Qualitative and quantitative assess- ments of these constructs show that (a) both reporters display similar sensitivities in cells grown on solid or liquid media and (b) GFP is especially well suited for quantita- tion of promoter activity in cells grown on agar. Thus, GFP provides a simple, rapid and sensitive tool for measuring relative promoter activity in intact E. coli cells.


Green fluorescent protein (GFP) has enjoyed widespread use as a qualitative reporter of in vivo gene expression in prokaryotes and eukaryotes (9,13). Sur- prisingly few studies have reported the use of GFP as a quantitative reporter of promoter activity in prokaryotes (5,7,11, 14). In contrast, β-galactosidase (β-gal) has been used extensively as a reporter for quantifying gene expression in Es- cherichia coli. The ease of use of GFP makes it an attractive candidate as a re- porter of promoter strength in E. coli.

Numerous variants of GFP are avail- able (13). We chose to work with GFPuv because it offers several advan- tages over wild-type GFP: it is more soluble at high expression levels, less toxic and yields more intense fluores- cence (3). (From this point, we will re- fer to GFPuv simply as GFP.)

To test the utility of GFP as a quanti-

tative reporter, we constructed a pro- moterless GFP vector into which we inserted a series of previously studied synthetic promoters having varying ac- tivities (Table 1). Because β-gal has been so well studied as a quantitative reporter, we also tested the same pro- moters in a promoterless β-gal vector (6) to directly compare β-gal and GFP as reporters of promoter activity in E. coli. Here, we present the use of GFP as a rapid, simple and quantitative reporter of relative promoter activity in intact E. coli cells.



We obtained DNA polymerase I (Klenow fragment), T4 DNA ligase, T4 polynucleotide kinase, Taq DNA poly- merase, competent JM109 and a Wiz- ard® Midi Prep Kit from Promega (Madison, WI, USA), restriction en- donucleases, λ BstEII marker and calf intestinal alkaline phosphatase from New England Biolabs (Beverly, MA, USA) and oligonucleotides from Life Technologies (Gaithersburg, MD, USA).

Cell Growth Conditions

For the plasmid copy number assay and for the GFP and β-gal liquid cul- ture assays, cells were grown at 37°C in M9 minimal medium (8) supplemented with 1% glucose, 0.2% casamino acids, 1 µg/mL thiamine and 100 µg/mL am- picillin. (From this point, we will refer to this supplemented medium as M9 Glu+Caa.) LB broth was unsuitable for GFP liquid culture assays because of its high and variable background fluores- cence (data not shown). For GFP plate assays, single colonies were streaked on LB agar plates supplemented with 100 µg/mL ampicillin, and plates were incubated at 37°C for 20 h followed by storage at 4°C.


We obtained the pGFPuv plasmid from Clontech Laboratories (Palo Alto, CA, USA) and excised the lac promot- er using SapI and HindIII followed by

Short Technical Reports

Vol. 28, No. 1 (2000)

filling in the overhangs with DNA polymerase I (Klenow fragment) and religation. This promoterless vector was named pJC1.

To limit transcriptional readthrough, the bacteriophage T7 early transcrip- tional terminator (T7Te) (10) was di- rectionally cloned downstream of the GFP coding sequence. The following complementary oligonucleotides were annealed and cloned between the EcoRI and SpeI sites in pJC1: EcoGFP, 5′-AATTCAAGCTTAAATGTAATCA- CACTGGCTCACCTTCGGGTGGG- CCTTTCTGCGA-3′ and SpeGFP, 5′- CTAGTCGCAGAAAGGCCCACCCG AAGGTGAGCCAGTGTGATTACA- TTTAAGCTTG-3′. The resulting plas- mid, which contains a promoterless GFP coding sequence with a T7Te ter- minator, was designated pJC2.

The pQF50 β-gal reporter vector (6) was used without modification.

Promoter Cloning

The synthetic promoters used in this study (Table 1) had previously been cloned into a different reporter vector (12). The promoters were amplified from this vector by PCR using primers that contained restriction sites for cloning the amplified promoters be- tween the SphI and KpnI sites in pJC2 and pQF50. The pQF50 transformants were identified by the production of blue pigment on medium spread with 50 µL of X-gal (20 mg/mL in dimethyl- formamide) and/or by restriction analy- sis. The pJC2 transformants were iden- tified by green fluorescence under longwave UV illumination and/or by restriction analysis. Plasmid DNA for each promoter clone was isolated using the Wizard Midi Prep Kit, and the se- quence of each promoter region was determined by automated DNA se- quence analysis.

Estimate of Relative Amounts of Plasmid in Each Strain

Cells were grown overnight in M9 Glu+Caa, diluted 25-fold in fresh M9 Glu+Caa and grown at 37°C to A600 = 0.8±0.05. Triplicate 1.5 mL aliquots were removed from each culture, plas- mid DNA was isolated by alkaline lysis miniprep and an aliquot of each prep

was run on a 1% agarose gel containing 0.5 µg/mL ethidium bromide. Each gel also included a lane with 250 ng of λ BstEII molecular weight standard for es-

timation of plasmid mass. Gels were an- alyzed with 1D Gel Analysis software (Eastman Kodak, Rochester, NY, USA). Multiple plasmid isoforms were present

Short Technical Reports

84 BioTechniques Vol. 28, No. 1 (2000)

Figure 1. Expression of GFP and ββ-gal under the control of synthetic promoters. Strains containing plasmids that express GFP and β-gal driven by the indicated promoters were streaked on LB-ampicillin plates that contained X-gal and were grown at 37°C. Vector strains are promoterless. The upper portion of the figure shows GFP fluorescence produced by longwave UV illumination. The lower portion of the figure shows β-gal expression detected by the blue pigment produced by hydrolysis of X-gal.

Promotera -35 region -10 region





6/6,6/6 TTGACA TATAAT aPromoter names reflect the number of bp in the -35 and -10 regions that match the E. coli consensus sequences for each region. For example, the 6/6,5/6 pro- moter matches the consensus in 6/6 bp in the -35 region and 5/6 bp in the -10 region. The sequence between the -35 and -10 regions, shown for 4/6,4/6, is the same in all constructs except for a single bp substitution in the 6/6,6/6 GFP con- struct (see Results).

Table 1 Promoters Used in This Study

in each lane, so the sum of the mass of DNA present in each lane was deter- mined. The average plasmid DNA mass produced by each strain was then calcu- lated and used to estimate the relative amounts of plasmid DNA in each strain.

Assay Conditions

β-gal assays were carried out essen- tially as described using chloroform and SDS to permeabilize the cells (1,8). Briefly, overnight cultures of cells grown in M9 Glu+Caa were diluted 25- fold into fresh M9 Glu+Caa and were grown at 37°C to A600 = 0.7 ± 0.1. Cells were harvested and the assay was conducted on 0.1, 0.2 and 0.3 mL of each culture using the pQF50 strain as the control. The optical density of the o-nitrophenol produced was measured at 420 nm in a Spectronic® 21D spec- trophotometer (Milton Roy, Rochester, NY, USA) and was used to calculate

units of β-gal (A420/[min mL A600]). To assay GFP in cells grown on LB-

ampicillin plates (see above for the cell growth conditions), plates were stored at 4°C for 24–28 h after growth at 37°C because this treatment produced a sig- nificant increase in fluorescence inten- sity (data not shown). Two plate-assay methods, (A) and (B), were used, each using pJC2 to determine background fluorescence.

(A) Plate scrape assay: Colonies were scraped from plates with an inoc- ulating loop and resuspended in M9 Glu+Caa that lacked thiamine and con- tained chloramphenicol (CAM) (to pre- vent cell growth) to A600 = 0.20 ± 0.01. Approximately 20–40 colonies were used for each measurement. Duplicate fluorescence measurements of each sample were made at room temperature with an LS-3B Fluorescence Spectrom- eter (PE Biosystems, Foster City, CA, USA) with excitation at 395 nm and

emission at 509 nm using the promoter- less pJC2 strain as the blank. The re- sponse of the fluorometer was linear over approximately 3 orders of magni- tude (data not shown). Parallel experi- ments were also conducted in which cells were scraped from the densest part of the streak.

(B) Charged-coupled device (CCD) camera assay: Images of single plates were obtained with a GDS-8000 cooled CCD camera system (UVP, Upland, CA, USA) using 365 nm epi-illumina- tion and a green filter. Exposure time was 1/16 s, and the aperture was opened to maximum. Individual colonies were identified, and the fluo- rescence intensity was determined us- ing the colony counting function in the GDS-8000 LabWorks® software.

To assay GFP in liquid culture, overnight cultures of cells grown in M9 Glu+Caa were diluted 50-fold into fresh M9 Glu+Caa and were shaken at 37°C until they reached the target OD (A600 = 0.50 ± 0.03). At this point, CAM was added to a final concentra- tion of 200 µg/mL to stop protein syn- thesis and to allow complete folding of GFP, which folds slowly at 37°C (13). All strains reached the target ab- sorbance within 4–5 h except for 5/6,5/6 that took 6.0–6.5 h. The cul- tures were started at staggered times so that all cultures reached the target OD within the same 60–90 min period. In- creases in cell density and GFP fluores- cence ceased within 1 h of the addition of CAM (data not shown). Cultures were kept shaking at 37°C for 1 h after CAM addition. The final A600 was de- termined (range 0.52–0.65), and cul- tures were diluted with M9 Glu+Caa containing CAM to A600 = 0.50 ± 0.02. GFP fluorescence was determined in a fluorimeter as described above. GFP fluorescence was stable for at least 2 h at room temperature.


The GFP reporter vector we devel- oped, pJC2, encodes GFPuv, which is easily detected by longwave UV illumi- nation (3). pJC2 was derived from the pGFPuv vector so it has several unique restriction sites (SphI, PstI, XbaI, XmaI, SmaI, KpnI, Asp718I and AgeI) to facil-

Short Technical Reports

Vol. 28, No. 1 (2000)

itate promoter cloning. We chose pQF50 as the β-gal reporter vector, which also has an extensive multiple cloning site (6).

We selected a previously character- ized set of synthetic constitutive pro- moters that vary by a factor of >100 in terms of promoter strength (Table 1) and cloned each promoter into both vectors. DNA sequence analysis of these clones confirmed the expected -35 and -10 sequences. In addition, the 6/6,6/6 promoter cloned into pJC2 had acquired a T to A substitution at posi- tion -25, which would not be expected to affect promoter strength (2). Both pJC2 and pQF50 clones carrying the 4/6,4/6 promoter had unusual substitu- tions in which the CG dinucleotide at the -3 and -2 positions had been re- placed by a single A. Wild-type 4/6,4/6 promoters have very low activity (12), so we did not use the 4/6,4/6 constructs in our quantitative assays.

Because plasmid copy number can be sensitive to the strength of a cloned promoter, we estimated the relative amounts of plasmid produced by each strain. The difference between high and low values for plasmid copy number for promoter clones in pJC2 was ap- proximately 20%, while for pQF50 clones, the difference was approxi- mately 35%. Because these differences were small, we did not consider them in evaluating the results of the assays.

All promoters except for 4/6,4/6 consistently produced visible levels of expression in strains streaked on agar plates (Figure 1) and in individual colonies (data not shown). It was possi- ble to visually distinguish among the GFP strains, in terms of fluorescence, as follows: vector = 4/6,4/6 < 5/6,4/6 < 5/6,5/6 < 6/6,5/6 = 6/6,6/6. In contrast, for β-gal strains, the relative intensities of blue color were as follows: vector = 4/6,4/6 < 5/6,4/6 < 5/6,5/6 = 6/6,5/6 = 6/6,6/6. Therefore, on plates containing X-gal, GFP displayed a greater dynam- ic range than β-gal.

Figure 2 shows the results of quanti- tative analysis of relative promoter ac- tivity using GFP and β-gal reporter genes. GFP plate and β-gal liquid cul- ture assays were able to distinguish among weak [5/6,4/6], moderate [5/6, 5/6] and strong [6/6,5/6 and 6/6,6/6] promoters. Therefore, GFP and β-gal

are comparable in terms of sensitivity and dynamic range as quantitative re- porters of relative promoter strength. Both methods used to quantify GFP ex- pression on plates (i.e., plate scrape and CCD camera) are rapid, simple and in excellent agreement. In the plate scrape assay, essentially identical relative flu- orescence values were obtained using either colonies or cells scraped from the densest part of the streak (data not shown). Interestingly, the absolute val- ue of fluorescence in cells scraped from the densest part of the streak was 3–4- fold higher than the signal from colonies (data not shown).

Results obtained using a GFP liquid culture assay were inconsistent with re- sults from GFP plate and β-gal liquid culture assays (Figure 2). Specifically, activity of the 5/6,4/6 promoter was un- detectable in liquid culture, while the activity of the 5/6,5/6 promoter exceed-

ed that of 6/6,5/6. In addition, we en- countered difficulties with the growth of GFP-expressing strains in liquid cul- ture including extremely poor growth of the 6/6,6/6 strain and variable growth rates (all strains reached the tar- get absorbance within 4–5 h except for 5/6,5/6, which took 6–6.5 h). Curious- ly, all GFP strains seemed to grow well on LB-ampicillin plates at 37°C and all β-gal expressing strains grew well in liquid and on solid media.


GFP displays a sensitivity and dy- namic range equivalent to β-gal as a re- porter of relative promoter activity in both qualitative (Figure 1) and quanti- tative assays (Figure 2). This result is somewhat surprising because the lack of signal amplification inherent in GFP

Short Technical Reports

88 BioTechniques Vol. 28, No. 1 (2000)

Figure 2. Quantitative analysis of relative promoter activity using GFP and ββ-gal reporters. GFP fluorescence and β-gal activity were quantified as described in Materials and Methods. For each assay, the data were obtained from three independent experiments and normalized to the activity of the 6/6, 5/6 promoter. For GFP-colonies/suspension and GFP-colonies/CCD camera, n=6 for each promoter; for GFP liquid culture, n=8 for 5/6, 4/6, n=15 for 5/6,5/6, and n=8 for 6/6, 5/6; and for β-gal, n=3 for each pro- moter. Error bars indicate standard deviation. *Not determined.

assays has been thought to be a short- coming in comparison with enzymatic assays (9,13).

The two different GFP plate assays that we developed, plate scrape and CCD camera, are rapid, simple and sensitive. They allow ready quantita- tion of relative promoter strength over an activity range comparable to that for β-gal. In fact, GFP offers advantages over β-gal for assessing promoter activ- ity. On plates, detection of GFP does not require the use of special media (e.g., media that contain X-gal). While a variety of sophisticated colorimetric plate assays for β-gal have been devel- oped (8), all require the inclusion of one or more reagents in the growth medium. In addition, our GFP plate as- says are much more rapid than the β- gal liquid culture assay because re- porter activity is detected directly on plates and does not require a period of growth in liquid to a particular cell den- sity. Weak and moderate strength pro- moters take no more time to assay than strong promoters in the GFP plate as- says because the amount of reporter produced is measured directly. In con- trast, weak and moderate strength pro- moters take longer to assay with β-gal because of the longer time required to produce measurable amounts of the product of β-gal activity, o-nitrophenol.

The GFP reporter vector that we have developed has a number of possi- ble uses in addition to quantifying the activity of known promoters. For exam- ple, it should be useful as a vector for shotgun cloning and detection of pro- moters from a variety of eubacteria. In particular, the dynamic response of this vector in cells grown on agar should al- low screening for promoters with a wide range of activities. The CCD cam- era assay should be ideal for such screens in which digitized images of colonies on plates could be subjected to image analysis to allow automated identification of colonies that express various levels of GFP. Similarly, the CCD camera assay should allow genet- ic screens for altered GFP expression levels to identify mutations affecting the regulation of a particular promoter.

Based on the data presented here, GFP may not be as good a reporter of promoter activity in a liquid culture as- say, at least for constitutive promoters.

We observed variable and sometimes poor growth of strains that express GFP in liquid culture, an effect not seen when these strains were grown on agar. The shortcomings of the liquid culture assay may stem from the apparent toxi- city of GFP caused by the postulated release of H2O2 during chromophore formation (4). Interestingly, others have reported the apparently successful use of GFP as a quantitative reporter in liq- uid culture using inducible promoters (11,14). It would be interesting to see if engineering cells to express high levels of catalase (13) would reduce the pos- tulated toxicity. Alternatively, GFP variants that are claimed to be non-tox- ic (e.g., blue fluorescent protein) may be better reporters for cells grown in liquid culture.


1.Allen, T. 1986. M.S. thesis, Case Western Re- serve University School of Medicine.

2.Auble, D.T., T.L. Allen and P.L. deHaseth. 1986. Promoter recognition by Escherichia coli RNA polymerase: effects of substitutions in the spacer DNA separating the -10 and -35 regions. J. Biol. Chem. 261:11202-11206.

3.Crameri, A., E.A. Whitehorn, E. Tate and W.P.C. Stemmer. 1996. Improved green fluo- rescent protein by molecular evolution using DNA shuffling. Nat. Biotechnol. 14:315-319.

4.Cubitt, A.B, R. Heim, S.R. Adams, A.E. Boyd, L.A. Gross and R.Y. Tsien. 1995. Un- derstanding, improving and using green fluo- rescent proteins. Trends Biochem. Sci. 20:448-455.

5.Dhandayuthapani, S., L.E. Via, C.A. Thomas, P.M. Horowitz, D. Deretic and V. Deretic. 1995. Green fluorescent protein as a marker for gene expression and cell biology of mycobacterial interactions with macrophages. Mol. Microbiol. 17:901-912.

6.Farinha, M.A. and A.M. Kropinski. 1990. Construction of broad-host-range plasmid vectors for easy visible selection and analysis of promoters. J. Bacteriol. 172:3496-3499.

7.Kremer, L., A. Baulard, J. Estaquier, O. Poulain-Godefroy and C. Locht. 1995. Green fluorescent protein as a new expression marker in mycobacteria. Mol. Microbiol. 17:913-922.

8.Miller, J.H. 1972. Experiments in Molecular Genetics. CSH Laboratory Press, Cold Spring Harbor, NY.

9.Misteli, T. and D.L. Spector. 1997. Applica- tions of the green fluorescent protein in cell biology and biotechnology. Nat. Biotechnol. 15:961-964.

10.Reynolds, R., R.M. Bermudez-Cruz and M.J. Chamberlin. 1992. Parameters affecting transcription termination by Escherichia coli RNA polymerase. I. Analysis of 13 Rho-inde- pendent terminators. J. Mol. Biol. 224:31-51.

11.Siegele, D.A. and J.C. Hu. 1997. Gene ex- pression from plasmids containing the araBAD promoter at subsaturating inducer concentrations represents mixed populations. Proc. Natl. Acad. Sci. USA 94:8168-8172.

12.Szoke, P.A., T.L. Allen and P.L. deHaseth. 1987. Promoter recognition by Escherichia coli RNA polymerase: effects of base substitu- tions in the -10 and -35 regions. Biochemistry 26:6188-6194.

13.Tsien, R.Y. 1998. The green fluorescent pro- tein. Annu. Rev. Biochem. 67:509-544.

14.Zhao, H., R.B. Thompson, V. Lockatell, D.E. Johnson and H.L.T. Mobley. 1998. Use of green fluorescent protein to assess urease gene expression by uropathogenic Proteus mirabilis during experimental ascending uri- nary tract infection. Infect. Immun. 66:330- 335.

We thank Drs. P. de Boer, R. Maurer and P. Rather for critically reviewing this manu- script, Dr. R. Bohinski for instruction in fluorimeter operation, J. Pieri of UVP for advice with the GDS-8000 software, G. Fe- doriw for technical assistance and three anonymous reviewers for many helpful comments. This work was supported by in- ternal funds from John Carroll University, National Institutes of Health Grant No. GM31808 to P.L.dH and USPHS Grant No. P30CA43703 to the CWRU Molecular Biol- ogy Core Laboratory. Address correspon- dence to James L. Lissemore, Department of Biology, John Carroll University, Univer- sity Heights, OH 44118, USA.

Received 8 March 1999; accepted 12 August 1999.

J.L. Lissemore, J.T. Jankowski, C.B. Thomas, D.P. Mascotti and P.L. deHaseth1 John Carroll University University Heights, OH 1Case Western

Reserve University Cleveland, OH, USA

Vol. 28, No. 1 (2000) BioTechniques 89

“Get 15% discount on your first 3 orders with us”
Use the following coupon

Order Now