Localization of Upstream Silencer Elements Involved in the Expression of Cone Transducin a-Subunit (GNAT2) T. Allen Morris*^ Wei-Bao Fong,* Melissa J. Ward,* Hu Hu,% and Shao-LingFong*% Purpose. To localize as-acting elements involved in the expression of the cone-specific G- protein, cone transducin a-subunit (GNAT2). Methods. In this study, the authors used a genomic clone, HGLG3, to sequence 3139 base pairs of the upstream region of the GNAT2 gene and to localize as-acting elements involved in the expression of GNAT2. Upstream elements were localized functionally by transfection of chloramphenicol acetyltransferase gene constructs containing nested deletions of this up- stream region into WERI-Rbl cells. Cell specificity of the localized elements was determined by transfection of the HeLa cells. Transacting factor-binding sites to functional as-acting elements were determined by DNasel footprinting. Cell specificity of protein interaction with footprinted regions was tested by electrophoretic mobility shifts with nuclear extracts from WERI-Rbl and HeLa cells. Results. Transfection of WERI-Rbl and HeLa cells revealed the presence of a strong, noncell- specific silencer region between —1130 and —23, a weak, cell-specific promoter between — 151 and —10, and a stronger, noncell-specific element between +143 and +167. DNasel footprinting showed three major footprints (SI, S2, and S3) between —807 and —176, indicat- ing the binding sites for putative negative trans-acting factors. Individual footprinted sequences had similar electrophoretic mobility shifts when they were incubated with nuclear extracts from either WERI-Rbl or HeLa cells, suggesting that these cells express the same negative factors. Conclusions. The expression of the GNAT2 gene is controlled by a strong silencer region, a weak upstream cell-specific promoter, and a strong downstream element. The silencer region interacts with similar proteins from retina- and nonretina-derived cell lines. Invest Ophthalmol Vis Sci. 1997;38:196-206. J. ransducin, a G protein, plays an important role as units from cGMP-PDE. The activated cGMP-PDE a liaison in the visual transduction cascade in photore- rapidly lowers the concentration of cGMP in the outer ceptors. During interaction with a light-activated pho- segment of the photoreceptor, causing the closure of topigment, the a-subunit of transducin is stimulated cGMP-dependent cation channels and hyperpolariza- to exchange bound guanosine diphosphate for guano- tion of the plasma membrane.1'2 sine triphosphate and is released from its /?y-subunits. For the most part, the biochemistry for the verte- The guanosine triphosphate-bound a-subunit then ac- brate visual cascade has been elucidated using pro- tivates cyclic guanosine monophosphate-phosphodi- teins purified from rod photoreceptors.1 Cone photo- esterase (cGMP-PDE) by removing inhibitory y-sub- receptors appear to use a similar phototransduction mechanism.3'4 For example, cones and rods each have their own homologous visual pigments,5'6 transduc- From the Departments of * Ophthalmology and %Biochemistry and Molecular Biology, s Indiana University, Indianapolis, and the ^Department of Biological Sciences, ins, and cGMP-PDEs. Although many of the Purdue University, West Lafayette, Indiana. Supported by National Institutes of Health grant EY08562 and supported in part genes encoding these proteins have been character- by an unrestricted grant from Research to Prevent Blindness (New York, Neiu York) ized, there is much to be learned about the regulatory and by the Department of Ophthalmology, Indiana University School of Medicine. Received for publication March 21, 1996; revised August 16, 1996; accepted mechanisms involved in their expression. Recently, we August 16, 1996. began using transient transfections of a human retino- h-oprietary interest category: N. Reprint requests: Shao-Ling Fong, Department of Ophthalmology, 702 Rotary blastoma cell line, WERI-Rbl, as a model for studying Circle, Indiana University, Indianapolis, IN 46202. photoreceptor-specific transcriptional regulation.11 Investigative Ophthalmology & Visual Science, January 1997, Vol. 38, No. 1 196 Copyright © Association for Research in Vision and Ophthalmology Downloaded from iovs.arvojournals.org on 09/26/2021 Upstream Elements for the Human GNAT2 Gene 197 We have characterized the genes for human rod out as previously described11 using a cationic lipo- (GNAT112) and cone (GNAT213) transducin a-sub- some, Lipofectin (GIBCO BRL), with some modifica- units and have demonstrated that GNAT2 is expressed tions. For each WERI-Rbl transfection, 3.3 X 10° in WERI-RB1 cells. Studies of as-acting elements and WERI-Rbl cells were pelleted and then mixed with 18 transacting factors for these two homologous genes fig of GNAT2-CAT plasmid, 1 //g of a cytomegalovirus will provide knowledge about the similarities and dif- promoter-driven luciferase internal control plasmid, ferences in transcriptional regulation between rods CMV.LUC, and 10 /il of Lipofectin in 1 ml of Opti- and cones. In this article, we report the initial localiza- Meml medium (GIBCO BRL) and were plated imme- tion and characterization of ds-elements within the diately in a 35-mm well. After 16 hours, 1 ml of culture upstream region of the human GNAT2 gene. medium with 20% fetal bovine serum was added. Twenty-four hours later, cells were replated in 1.7 ml of fresh culture medium and were grown for 2 more MATERIALS AND METHODS days before they were harvested. DNA Sequencing and Comparison Adherent HeLa cells (ATCC no. CCL 2) were grown in Dulbecco's modified Eagle's medium The upstream sequence of GNAT2 gene was charac- (GIBCO BRL) with high glucose and the same supple- terized using a Pstl fragment of a previously character- 13 ments used for WERI-Rbl cells. HeLa cells were trans- ized human genomic clone, HGLG3. The complete fected at approximately 80% confluency in 35-mm sequence for both strands of this fragment was ob- wells by the immediate addition of a mixture of 10 //g tained using Sanger dideoxy-chain termination14 and 15 CAT construct DNA, 0.5 //g CMV.LUC DNA (used as the method of Dale et al. Overlapping sequences an internal control), and 10 /xl Lipofectin in 1 ml were aligned using Beckman (Palo Alto, CA) Micro- Opti-Meml. After 6 hours, the medium was replaced genie software. Functional regulatory sequences were with 1.6 ml of fresh, supplemented Dulbecco's modi- compared to a exacting element data base, tfsites, fied Eagle's medium and was incubated for 36 addi- by GCG Package (Genetics Computing Group, Inc, tional hours before it was harvested. Madison, WI). Harvested WERI-Rbl and HeLa cells were resus- pended in 0.1 M KPO4 (pH 7.8) and 0.1 mM dithio- Plasmid Constructions threitol (70 fi\ for WERI-Rbl and 60 fu.\ for HeLa Because no convenient restriction sites were available cells). One third of the WERI-Rbl and HeLa cytosols between the transcription ( + 1 to +31) and translation 13 were used for CAT assays as described by Gorman et (+214) start sites, an insert-containing sequence al.17 [14C]-labeled chloramphenicol (Amersham, Ar- from —3139 to +167 in M13mpl9 was generated ac- 15 lington Heights, IL) and its derivatives were separated cording to the Dale method and was used for plas- by thin-layer chromatography and quantitated by a mid constructions. This fragment was removed by Pstl two-dimensional /^-scanner (Ambis, San Diego, CA). and EcoKL digestion, subcloned into the Pstl site of Luciferase assays were carried out using a Luciferase pUC19, and directionally ligated into the Sphl and Pstl Assay System (Promega). sites immediately upstream of the chloramphenicol acetyltransferase (CAT) reporter gene in pCAT.basic Nuclear Extract Preparation (Promega, Madison, WI). Nested deletions from both Nuclear extracts for DNasel footprinting and mobility- the upstream and the downstream ends of this insert shift assays were prepared from WERI-Rbl and HeLa in pCAT.basic were created using the Erase-a-Base sys- 18 16 cells by the method of Dignam et al and were stored tem (Promega) and available restriction endonucle- in liquid nitrogen. Protein concentrations were deter- ase sites within the insert. Plasmid preparations were mined by the Bradford assay19 (BioRad, Hercules, banded twice by CsCl-ethidium bromide density gra- CA). dient centrifugation. WERI-Rbl and HeLa cells were transfected with two independent preparations of DNasel Footprinting Analysis each construct. DNasel footprinting was performed by the method of Dynan20 with some modifications. Briefly, two geno- Cell Transfection and CAT Assays mic fragments containing upstream sequences from The retinoblastoma cell line, WERI-Rbl (ATCC no. -807 to -553 and -546 to -176 were end-labeled HTB 169), was grown in suspension in a culture me- with [y-32P]ATP (7000 Ci/mmol; ICN, Costa Mesa, dium made up of RPMI 1640 supplemented with 10% CA) and T4 polynucleotide kinase (GIBCO BRL) on fetal bovine serum, 2 mg/ml sodium bicarbonate, 0.29 the 5'-end of either the coding or the noncoding mg/ml L-glutamine, 100 U/ml penicillin, and 100 strand and were purified by polyacrylamide gel elec- mg/ml streptomycin (GIBCO BRL, Gaithersburg, trophoresis followed by DE52 ion exchange chroma- MD) at 37°C and 5% CO2. Transfections were carried tography. A 50-//1 mixture containing 10 mM Tris (pH Downloaded from iovs.arvojournals.org on 09/26/2021 198 Investigative Ophthalmology 8c Visual Science, January 1997, Vol. 38, No. 1 8), 5 mM MgCl2, 5 mM CaCl2, 50 mM KC1, 500 nM of exon 1. The region used for plasmid construction dithiothreitol, 0.05 mg/ml bovine serum albumin, is shown in Figure 1. This sequence was compared and 40.5 ng///l poly(dldC) (Boehringer Mannheim, to the as-acting element database, tfsites, using the Indianapolis, IN) was incubated on ice for 15 minutes Genetics Computing Groups program. As expected in the presence or absence of 75 /ig WERI-Rbl nu- for such a large sequence, many matching sequences clear extract.