.l .9»... .. ’11; .ul .litovls . '0 :3! (I. ...\.l:. u for; o ) . {‘5}? la... . . It... {.0031 3):;fv...’ I .- .xr .Mnnfiun. o ,1 v!l III) 5}" 21.. 1 a! 15 5427, I 5}. :36! .. (.4: I} 13. .! 35'031) , o.‘ .. 5.7: .b \r. . . 2!:- r. 311.!!! I. 3.5 f {:1 ’3‘ . 5.») 1:1 1... f :I..; v: 7:... t. . :w L... .fl. ...v 7.....Il. . .v.. n.» p .u . ‘ ‘ , ‘ I. >2» “Kai-mean. , . 53.... . at... gigs, .. 1 ‘ ”(Ht-.848 mill/:7 ”ill/lull“? ill/ill?! This is to certify that the thesis entitled STUDIES OF CGMP PHOSPHODIESTERASES AND PHOSPHOINOSITIDE-SPECIFIC PHOSPHOLIPASES C IN BOVINE RETINAL ROD OUTER SEGMENTS presented by Jeffrey Robert Leipprandt has been accepted towards fulfillment of the requirements for M.S. degreeinB1ochem15try lam-WM Major professor Date__LA?).C9:L_ 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution LIBRARY Mlchlgan State Unlversity PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE W 2't “fig—7 ——_l 4;; fluf— F7 MSU Is An Affirmative Action/Equal Opportunity Institution ammunmn ———————-—— ___________.___ _,_... STUDIES O! 06!? PEOSPHODIEBTERABEB AND PEOBPEOINOBITIDE-SPECIPIC PHOBPEOLIPABES C IN BOVINE RETINRL ROD OUTER SEGMENTB BY Jeffrey Robert Leipprandt A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree MASTER OF SCIENCE Department of Biochemistry 1993 AEETBLQI STUDIES 01‘ cm PEOSPEODIESTERASES AND PEOSPEOINOSITIDE-SPECIFIC PEOSPHOLIPASES C IN BOVINE RETINLL ROD OUTER SEGMENTS BY Jeffrey Robert Leipprandt In vertebrate photoreceptors cGMP phosphodiesterase (PDE) is involved in mediating and amplifying the visual signal. Phosphoinositide-specific phospholipases C (PLCs) , are implicated in signal transduction in many cell-types including invertebrate photoreceptor transduction, and may play a role in light adaptation in vertebrates as well. Evidence suggests both P015: and PLC have multiple isoforms in the bovine rod outer segment (ROS). This work was undertaken to identify these isoforms. Peptide sequences obtained from isolated PD]: 0: and 6 subunits revealed distinct differences from published sequences, suggesting multiple isoforms. Work was initiated to identify cDNA clones that code for the corresponding PDE isoforms. Five bovine retinal PLC isoforms were identified from polymerase chain reaction products and cDNA clones. One of these is likely to have an important role in vertebrate vision because it bears striking amino acid sequence resemblance to norpA, a mutant B-PLC gene that results in blindness in Drosophila. I would like to thank some of the many people who have helped me get to this significant point in my life: the members of my committee, past and present, Drs. Ron Davis, Jerry Dodgson, John Linz, Bill Smith, and John Wilson, for their advice and guidance; Joe Leykum, David Dewitt, Annette Thelen, and Bill Helferich for special technical assistance and‘ suggestions; my laboratory co-workers and aides, past and present, Rich Pinke, Carol Mindock, Barry Gehm, Donna Schultz, Frank Wilkinson, Linda Doorenbos, David Pepperl, John Su, Denise Messing, Stephanie Wardwell, and Trish Palmer for their help and camaraderie; Dr. David McConnell, for his support, friendship and mentorship; my parents, Robert and Diana Leipprandt, whose love and support gave me an enviable start in life; and especially my partner in life, Ann Leipprandt, for her understanding, caring, and help, but mostly for her love. ii LIST OE TABLES LIST OE FIGURES ABBREVIATIONS INTRODUCTIOE CHAPTER I: 06!? PEOSPEODIESTERASES LITERATURE REVIEW MATERIALS AND METHODS Materials. ROS Isolation. cGMP PDE Isolation. Gel Filtration HPLC. SOS-PAGE. Electroelution. Carboxymethylation. Cyanogen Bromide Cleavage. Tryptic Digestion. HPLC Purification of Tryptic Peptides. Peptide Sequencing. Oligonucleotide Synthesis and Purification. RNA Isolation. Poly AI RNA Isolation. iii 10 11 11 12 15 15 15 16 16 17 17 18 CHAPTER II: cDNA Synthesis. Polymerase Chain Reaction. Agarose Gel Electrophoresis. Recovery of DNA from Agarose Gels. DNA Cloning. DNA Sequencing. Agtlo cDNA Library Plating. Library Replicas. Probe Labeling. Hybridization with Oligonucleotide Probes. Southern Blot Analysis. RESULTS Peptide Sequences. Polymerase Chain Reaction. Oligonucleotide ”Guessmer" Probe. DISCUSSION PDE Subunit Separation. Peptide Sequences. Polymerase Chain Reaction. A Focus for Future Research. LITERATURE REVIEW The Role and Control of Calcium in the ROS. Background of Phospholipases C. MATERIALS AND METHODS iv PEOSPEOIEOSITIDE-SPECIEIC PBOSPEOLIPASES C 18 18 19 19 20 20 20 20 21 21 22 23 23 26 34 36 36 36 38 41 46 46 46 47 50 Materials. Oligonucleotide Synthesis and Purification. RNA Isolation. cDNA Synthesis. Polymerase Chain Reaction. Agarose Gel Electrophoresis. Recovery of DNA from Agarose Gels. Cloning PCR Products. Oligonucleotide End-labeling. DNA Sequencing. Probe Labeling. Agtlo Library Plating and Replicas. Southern and Northern Blot Analyses. Hybridization with Oligonucleotide Probes. Hybridization with Double-stranded Probes. Computer Analysis of Sequences. RESULTS Polymerase Chain Reaction. Direct Sequencing. Cloning PCR Products. Probe Labeling. Test of Cloned PCR Products as Probes. Agtlo Library Screening. Characterization of Agth Clones. DISCUSSION Design of Primers. 50 50 50 51 51 51 51 51 51 52 52 52 52 52 53 53 54 54 59 62 68 68 69 76 9O 9O Polymerase Chain Reaction. Cloning of PCR Products. norpA-like PLCs in Bovine Photoreceptors. Proposals for Future Research. APPENDIX BIBLIOGRAPHY vi 91 92 92 94 97 99 CHAPTER I Table 1. Peptide Sequence Similarity. CHAPTER II Table 1. Comparisons of Cloned PCR Products. Table 2. Agth Clones. Table 3. Comparisons of Cloned Sequences. APPENDIX Table A1. Bovine Rod Cell Codon Utilization Table 25 67 80 88 98 INTRODUCTION Figure CHAPTER I Figure Figure Figure Figure Figure Figure CHAPTER II Figure Figure Figure Figure Figure Figure Figure Figure 1. 5. 6. 7. A Rod Cell. Separation of Pa and P3. PCR Primers. Initial PCR Products. PCR Using Additional Primers. Oligonucleotide ”Guessmer” Probe. Sequence Comparisons of Pa to P, and P0,. PCR Primers. PCR Products Using Initial Primers. PCR Products Using Secondary Primers. Southern Analyses. Northern Analyses. Cross-hybridization of Cloned PCR Products Comparison of Primers with Clones Overlap of Clones with Published PLCs viii 14 28 3O 3—. 35 4O 56 58 61 71 73 75 83 86 BHT: butylated hydroxytoluene BME: B-mercaptoethanol cDNA: complementary DNA cGMP: cyclic guanosine monophosphate DAG: diacylglycerol EDTA: ethylenediaminetetraacetic acid ‘: transducin GDP: guanosine diphosphate GTP: guanosine triphosphate HPLC: high pressure liquid chromatography 1P5: inositol-l,4,5-trisphosphate MW: molecular weight IQ: phosphodiesterase a-subunit lg: phosphodiesterase B-subunit I’: phosphodiesterase 7-subunit PAGE: polyacrylamide gel electrophoresis PDE: phosphodiesterase pfu: plaque forming unit PIP2: phosphotidylinositol-4 , S-bisphosphate PLC: phospholipase C PMSF: phenylmethylsulfonylfluoride ROS: retinal rod outer segment ix SDS: sodium dodecyl sulfate SM: phage suspension medium sscDNA: single-stranded cDNA TEA: tetraethylamine TFA: trifluoroacetic acid Tris: tris(hydroxyethyl)aminomethane INTBQDHQTIQH In the retina specialized photoreceptor cells absorb light and convert it to an electrical signal by a process termed visual phototransduction. ‘Vertebrate retinas have two morphologically distinct types of photoreceptor cells, cones and rods. Cones operate in bright light and are of three types characterized by their spectral sensitivity. Cones are responsible for color vision. Rods are of a single type and operate at low light levels. Visual biochemistry has been studied most extensively in rods because they: 1) are larger and outnumber cones in.most retinas; 2) are easier to isolate and, 3) do not have the complication of multiple types. The rods are long cells that can be divided into four defined regions (see Figure 1). The outer segment contains numerous stacked membrane disks and is the site of photo- transduction. It is attached to the inner segment by a slender ciliary connection. The inner segment contains typical cell structures such as mitochondria and endoplasmic reticulum. Next is the nucleus and finally, the synaptic terminus, which connects the rod to other neural cells of the retina [Rodieck, 1973]. Phototransduction is initiated in the disks of the outer W]. This schematic diagram illustrates the major components of a rod cell. The cell is elongated and has four major divisions. The outer segment, the site of phototransduction, contains numerous stacked membrane disks which are not continuous with the plasma membrane. The photon receptor, rhodopsin, lies in these disk membranes. In the dark Ca++ and Na* enter the outer segment via the cGMP-gated cation channels. Upon photon absorption, rhodopsin initiates a cascade that results in depletion of cGMP and subsequent closure of cGMP-gated cation channels. This leads to hyperpolarization of the plasma membrane and initiation of neural signaling. It also~causes a decrease in cytosolic Ca++ which continually exits through a light-insensitive Na‘lCa+*,K+ exchanger. The outer segment is attached to the inner segment by a constricted ciliary connection. The inner segment contains many of the typical cellular structures and is the site of most of the biosynthetic pathways. Additional cellular constrictions separate the inner segment from the nucleus and the nucleus from the synaptic terminal. (2:222:22) °Gm°‘°‘t°d E Cation Channel CZZIIIZIZD Na+ (“H-fly: Outer Segment ' (2222:2222) ca++l K+ : 4 Na+ (2222:2222) Na+/Ca++,x+ Exchanger C111"— Mitochondria Nucleus Synaptic Terminal 4 segment when light is absorbed by the visual pigment, rhodopsin. Rhodopsin consists of an integral membrane protein, opsin, and a covalently attached chromophore, 11-cis- retinal. When a photon is absorbed by 11-cis-retinal, it isomerizes to all-trans-retinal, activating the rhodopsin molecule by causing conformational changes [Dratz & Hargrave, 1983]. In its active form rhodopsin activates a GTP-binding protein known as transducin or G. This enzyme is a periph- eral membrane protein composed of three subunits, Gh(40 kDa) , 6,,(35 kDa), and G"(8 kDa). It is a member of a well characterized family of GTP-binding proteins including the G, and Gi regulatory proteins of the hormone-sensitive adenylate cyclase. Rhodopsin activates G, by mediating the exchange of GDP for GTP on G... and dissociating it from the GW complex. A single activated rhodopsin can activate as many as 500 Gt molecules [Stryer, 1986]. The active Gh-GTP complex in turn activates the cGMP phosphodiesterase (PDE) . This enzyme is also a peripheral membrane protein that contains three different types of subunits, Pa(88 kDa), P,(85 kDa), and P7(12 kDa) [Kohnken et a1. 1981] . The PDE is activated when the Gh-GTP complex removes the inhibitory P, [Stryer, 1986] . The active PDE hydrolyzes the 3'-5’ phosphodiester bond of cGMP to form 5'-GMP. A single PDE is capable of hydro- lyzing over 1000 cGMP molecules per second. Thus in this 5 entire cascade the signal can be amplified as much as 106 [Baehr & Applebury, 1986]. In phototransduction cGMP has the role of a soluble second messenger between the disk membranes and the plasma membrane. In the dark cGMP levels are high and the molecules act as allosteric effectors keeping open cGMP- gated cation channels in the plasma membrane. In the light the active PDE depletes the cGMP thus causing the closure of the cation channel. This closure in turn results in the hyperpolarization of the rod cell membrane and the initiation of the neural relay to the brain [Lamb, 1986]. CHAPTER I CGHP PHOSPHODIESTERASES W The light activated cGMP PDE from the ROS plays a central role in the visual cascade [Pober and Bitensky, 1979]. It consists of three different subunits denoted P“, P,, and P7, which are 88, 85, and 12 kDa respectively [Baehr et a1. 1979; Kohnken et al. 1981a]. In darkness there is very little PDE activity in the ROS but when light impinges on the photo- ‘ receptor, PDE activity is dramatically increased. This activation has been demonstrated to be caused by removal of the inhibitory P1, mediated by the active Gu-GTP complex [Hurley and Stryer, 1982]. Though the stoichiometry of the isolated PDE has been determined to be 1:1:2, Pa:P,:P,, [Deterre et a1. 1988] , this may not strictly be the case for the enzyme in viva. Hurwitz et a1. [1985] presented evidence that P0,, or P“, complexed with Gb-GTP, retained PDE activity. While this report did not indicate the subunit makeup of the PDE holo- enzyme, it did suggest that P, and P,3 may be subunits of two distinct PDE isoforms, or at the very least are able to act independently. Complementary DNA sequences have been published for all three of the PDE subunits, P [Ovchinnikov et a1. 1987; Pittler et a1. 1990], P, [Lipkin et a1. 1990], and P7 7 [Ovchinnikov et a1. 1986]. The cDNAs for P, and P, code for quite similar proteins (72% sequence identity); however, the evidence that the cDNAs actually code for their designated subunits is not completely convincing. First, both cDNAs predict subunits of significantly higher MW than either P, or P, as revealed by SDS-PAGE, 99 vs. 88 kDa for P,., and 98 vs. 85 kDa for P,. Second, neither sequence is consistently corroborated by peptide sequences. At first glance the amino acid sequences deduced from the cDNAs appear to be well supported by numerous direct peptide sequences, 38 for P, [Ovchinnikov et a1. 1987] and 27 for P, [Lipkin et a1. 1990]. However, careful reading of the texts reveal that the vast majority of those peptides were derived from undifferentiated Pd. In fact only four peptides were derived from purified subunits, of which only two were actually sequenced, both from P,. One of these was almost identical with its Pa counterpart. The other two peptides, from P,, in too low a yield for sequencing merely had compositions from amino acid analysis that matched deduced peptides from the cDNA. However, these also were almost identical with their’rg counterparts. Thus confirmation that these cDNAs code for P, and P, hinges almost exclusively on the fact that a single sequenced P, peptide could not be found in a P, digest from which no peptides were sequenced, Nor' do ‘the authors ‘mention if any' peptide sequences were obtained that did not match either the P, or P, deduced amino acid sequences. 8 The authenticity of the cDNAs is further cast in doubt when considering the evidence for multiple PDE isoforms in the ROS. Hurwitz et a1. [1985] first indicated that multiple isoforms exist on the basis of HPLC fractionated PDE activity and immunological crossreactivity. Their existence is also implied by two cDNA clones reported by Pittler [Ph.D. Thesis, 1989; Pittler et a1. 1990]. These clones, which are not full length, are almost identical to the Ovchinnikov/Pittler P, sequence, but each codes for a different C-terminal region diverging at amino acid 841, approximately the same place the published sequences diverge. I The research reported on the following pages was performed.at the time that.much.of this information was coming to light. .As originally'proposed, part.of this.project was to be in conjunction with the work done by Steve Pittler, providing peptide sequence information from purified P, and P, by which cDNA clones could be identified. However, when differences between peptide and cDNA sequences arose that were not easily rectified the two projects were separated. Continuation of this project proved to be quite challenging but insightful, leading to proposals for research that may reconcile sequence discrepancies and.provide an assured basis for subsequent research. W W Radiolabeled compounds were purchased from New England Nuclear. DE-52 cellulose was purchased from Whatman. HPLC grade acetonitrile was purchased first from Aldrich and later from Burdick and Jackson. TFA and TEA were purchased from Pierce. Oligo d(T) cellulose was purchased from Stratagene. Tag polymerase was purchased from Cetus. DNA modifying enzymes were purchased from either New England I Biolabs, Bethesda Research Labs, or Boehringer Mannheim. Reagents for media preparation were purchased from Difco. All other chemicals were purchased from Sigma. W Bovine eyes were obtained from Murco, Inc. of Plainwell, MI. They were removed within minutes of slaughter, packed on ice, and kept dark. All subsequent operations were performed under dim red light and on ice. Within 24 hours the retinas were dissected from the eyes. ROS were prepared by the method of Kohnken et al. [1981a]. Briefly, isolated retinas were vigorously shaken in buffered 1.32 M sucrose, 150 mM KCl, shearing the ROS at their ciliary connection, to the inner segments. After centrifugation at 1100 x g the supernatant was diluted three-fold with buffered 150 mM KCl and centrifuged at 2100 x g. The pellet was 10 resuspended in buffered sucrose-KCl, p=1.10, and applied to a p=1.12 to 1.16 sucrose density step gradient. After centri- fugation at 90,000 x g, ROS were collected from bands at the interface of gradient steps p81.10 to 1.12 and p=1.12 to 1.14 and stored in liquid nitrogen until use. g§n£_£Q£_1§91agign,, The PDE was purified by the methods of Kohnken et al. [1981a] with modifications inspired by Kfihn [1982]. ROS was diluted with 10 volumes of 250 mM sucrose, 150 mM KCl, 10 mM Tris-Cl, pH 8.0, 10 mM DRE, 1 mM EDTA, 0.1 mM PMSF, 0.01% BET (solution A, isotonic) and centrifuged at 15,000 x g. The pellets were resuspended in.a minimal amount of 10 mM Tris-Cl, pH 8.0, 1 mM MgCl,, 1 mM DTT (solution M, hypotonic) , mixed, diluted to 500 ml with solution A, and centrifuged at 15,000 x g. The pellets were resuspended in minimal solution M, homogenized, bleached with bright light, diluted to 300 ml with solution M, and centrifuged at 90,000 x g. The supernatants were collected and the pellets were resuspended, homogenized, diluted, and centrifuged twice more pooling the supernatants. A Whatman DE-52 cellulose column (1 x 8 cm) was equilibrated with solution M. The supernatants were applied and then the proteins were eluted with a 200 ml linear NaCl gradient (0-500 mM) in solution M. Fractions (5 ml) were collected and assayed for absorbance at 280 nm. Peak fractions containing the PDE were pooled and concentrated to 2 ml with an Amicon concentrator and YM-S membrane. 11 A Sephadex G-100 column (1.2 x 115 cm) was equilibrated with solution M. Then the concentrated DE-52 pool was applied to the column and eluted with solution M. Fractions (3 ml) were collected and assayed for absorbance at 280 nm. Am peak fractions with elution times the characteristic of the PDE were pooled and concentrated to 1.5 ml. The fractions were analyzed by analytical SDS-PAGE and protein content determined by the method of Lowry et a1. [1951]. W was conducted using a Waters HPLC system employing a Zorbax Bio-Series GF-250 gel filtration column (DuPont). Material applied to the column was eluted. with 10 mM Tris-Cl, pH 7.4, 1 mM DTT at a flow rate of 1 ml/min. The eluants were monitored for absorbance at 280 nm. Freshly prepared, soluble samples of PDE applied to the column eluted in the column's void volume indicating an aggregate greater than 500 kDa was present. There was essentially no material that eluted at a retention time indicative of the separate P, or P,. Neither was there a peak that might suggest a few subunits binding together as a holoenzyme. There was, however, a sharp peak with the retention time of a 10 kDa protein, most likely representing P,. Considering that only fairly harsh denaturing conditions could solubilize the PDE, preparative SDS-PAGE seemed like the best technique for Pfi separation. §Q§;£A§£, Some early analytical SDS-PAGE was performed in 10% acrylamide/0.267% bis-acrylamide as described by 12 Kohnken et al. [1981a]. The proteins were visualized with Coomassie brilliant blue [Gordon, 1975] or by silver staining [Wray et a1. 1981]. To improve the separation of Pa and P,, preparative and subsequent analytical SDS-PAGE were performed in 15% acrylamide/ 0.08% bis-acrylamide as described by Baehr et a1. [1979] (see Figure 1A). Preparative gels were stained by gentle shaking in 0.25 M KCl at 0° for 10 to 30 min. Viewing was aided by a dark background and strong fluorescent side lighting. Bands were cut from the gel with a scalpel or sharp scissors. W The device used was an Elutrap electro-i separation chamber (Schleicher & Schuell). It was set up so that when gel strips were placed in one of the Elutrap's chambers and voltage was applied, proteins migrated from the gel into a second, low volume chamber (the trap). The trap was bounded on the far end by a BT-l membrane which allowed only small ions, driven by the voltage drop, to pass. It was bounded on the near end by a BT-2 membrane which allowed any charged, soluble ions, including proteins, to pass, under an applied voltage. The Elutrap was placed in a horizontal gel electro- phoresis chamber filled with buffer (192 mM glycine, 25 mM Tris base). The Elutrap’ s chambers were also filled with buffer and gel slices containing individual KCl-stained bands were placed in the larger of the two chambers. .A.potential of 200 V was applied for about six hours with the positive pole 13 W A. This photograph depicts a SDS-polyacrylamide gel with a typical separation of preparative amounts of P, and P, which were identified by their characteristic mobilities. Between 0.5 and 1.0 mg of purified bovine ROS cGMP PDE was applied to a preparative 15% polyacrylamide, 0.08% bis-acrylamide gel (15 x 15 cm, 1.5 mm thick). The optimum amount of P,,”, applied to the gel was 500-1000 pg. More material overloaded the gel and hindered the separation and less material resulted in poor visualization by KCl-staining. It was run at 150V until P, and P, had migrated more than halfway down the gel, ~24 hr. The KCl-stained bands were viewed against a dark background. The bands were cut apart with a scalpel, trimmed away from the rest of the gel, and eluted as described in the Methods section. B. The eluted P, and P, were analyzed for the completeness of their separation from each other by subjecting samples to analytical SDS-PAGE using the same gel composition as above. The bands on this gel, however, are visualized using the much more sensitive silver staining techniques as described in Methods. Lanes 1 and 7 are material from a DE-52 fraction of the PDE isolation protocol which contains both the PDE and G,, used here as MW indicators. Lanes 2 and 6 contain well- separated P, and P, fractions, respectively. Lanes 3 and 4 show P, fractions that were not well separated from P, and lane 5 shows a P, fraction that was not well separated from P,. Only fractions that were very pure, like those in lanes 2 and 6 were used in subsequent analyses. 14 15 toward the trap and of the Elutrap. The contents of the trap were then put directly into dialysis tubing (10-12 kDa cutoff) and dialyzed against distilled water. The yield of the elution was never very good, 50% at best and sometimes as low as 10%. Eluted P, and P, were analyzed for the quality of separation from each other by SDS-PAGE (see Figure 13). mm The separated P, and P, were carboxymethylated or ’H-carboxymethylated as described by Glazer et a1. (1975). gyan9ggn_flzgm1§g_§1eayage,, The protocol was essentially. that of Erhard Gross [1967]. The carboxymethylated protein was dissolved in 0.1 N HCl. CNBr was added in a 100-fold excess of the estimated methionine content of the sample and the reaction was allowed to proceed for 24 hr. at room temperature in the dark. The solution was then frozen and lyophilized. For P, and P,, this resulted in a yellow, largely insoluble material in both cases. These were intractable tars that yielded no detectable protein upon amino acid analysis and could not be resolved by 2-dimensional electrophoresis. I;yp3;j,g;_]2.;i.gg§33152;;L Carboxymethylated proteins in 10 mM NILECO, were incubated at 37°. Bovine trypsin was dissolved in 0.4 M NI-IJICO3 at a concentration of 100 ug/ml. An amount of trypsin, equal to 1% of the weight of the carboxymethylated protein, was added and the mixture was incubated at IT’fOr 3 hr. An equal amount of trypsin was added again and the 37° 16 incubation was continued overnight. “Wages. Initial fraction- ations of the P. and P, tryptic digests were performed using a Waters uBondapak C" column with an 80 min. linear 1%/min gradient of acetonitrile/0.1% TFA in H20/0.1% TFA. Fractions were collected by hand, changing Eppendorf tubes between Au, peaks. IBoth. digests yielded similar' but not identical chromatographic profiles with numerous peaks. Individual tryptic peptides were purified in one or more additional chromatographic steps using the Waters uBondapak C“ column or a narrow bore Vydac Protein and Peptide C18 column. Fractions? were rechromatographed in one of two different organic gradients: 1)a 1%/min. gradient of acetonitrile/0.1% TEA in H,O/0.1% TEA, or 2)a 1%/min. gradient of 65:20:15 aceto- nitrile:2-propanol:H,O in FRO/0.1% TFA. Some peptides were further purified using one or more of the above solvent systems. When chromatograms revealed sharp isolated peaks without shoulders, peptides were considered pure. Fractions were dried on a Speed Vac concentrator and submitted for sequencing. 2gp§19g_§ggggnging&, Some of the peptides were submitted to the Macromolecular Sequence, Synthesis, and Structure Facility (MSSSF), Department of Biochemistry, Michigan State University, where they were sequenced on a Beckman System 890 M protein sequencer. Other peptides were submitted to the Protein Structure Lab, University of California, Davis, where 17 they were sequenced on an Applied Biosystems 470 gas phase protein sequencer. Qlig2nnsle9tide_5xnthesis_and_znrifisationi Degenerate oligonucleotides based on peptide sequences were synthesized by the MSSSF, Department of Biochemistry, Michigan .State University on an Applied Biosystems 3808 DNA synthesizer. The crude synthetic oligonucleotides were purified by HPLC using a Zorbax Bioseries Oligo column (DuPont). The column was equilibrated with 80% 50 mM NaPO,, pH 7.0, 20% acetonitrile, 0.1 M NaCl. The oligonucleotides were applied and then eluted with a linear 0.1 to 0.8 M NaCl gradient in the same solvent. in 70 min. Peak Am fractions were pooled and desalted using C" Sep Paks (Waters) as described by Atkinson and Smith [1934]. W RNA was isolated by the method of Chirgwin et a1. [1979] with modifications by Dr. Bill Helferich [personal communication, 1989]. Bovine eyes were obtained from Murco, Inc of Plainwell, MI. Within 3 min. of slaughter the retinas were removed from the eyes and dropped into liquid nitrogen. Frozen retinas were crushed in a supercooled tissue grinder and homogenized in a solution of 4 M guanidine isothiocyanate, 25 mM sodium citrate, pH 7.0, 100 mM BME, 0.5% N-lauryl sarcosine, and then centrifuged at 5,000 x g. The supernatant was put into ultracentrifuge tubes layered on top of pads of 5.7 M CsCl, 100 mM EDTA and then centrifuged in a SW 55 TI rotor at 35,000 rpm for 20 hr. The 18 pellets were resuspended in 7 M guanidine HCl, 20 mM sodium acetate, pH 7.0, 1mM DTT, 10 mM iodoacetic acid, 1 mM EDTA, and then ethanol precipitated. Pellets were washed with 3 M sodium acetate, 10 mM iodoacetamide, pH 5.0; 66% ethanol, 33 mM sodium acetate, pH 5.0, and then 100% ethanol. Absorbance spectra, 320 nm to 220 nm, were taken from samples of the RNA to assess purity. Ana/Am ratios were routinely >1.8. The integrity of the RNA was assessed by formaldehyde-1% agarose gel electrophoresis as described by Fourney et a1. [1988]. W Poly A” RNA was isolated from total RNA by oligOsd(T) cellulose chromatography as described. by Sambrook et a1. [1989]. W Single stranded cDNA (sscDNA) was initially synthesized using bovine retinal poly AI RNA, oligo d(T)m, (Pharmacia) , and M-MLV reverse transcriptase (BRL) applying protocols provided by BRL. Subsequently sscDNA.was synthesized using specific antisense primers as described by Lee and Caskey [1990]. MW Thermal cycling was performed with a DNA Thermal Cycler (Perkin Elmer Cetus). PCR was carried out by protocols developed by Perkin Elmer Cetus with empirically determined adjustment made for particular ampli- fications. Typical reaction conditions were 10 mM Tris-Cl, pH 8.3, 50 mM KCl, 1.5 mM MgC12, 0.01% gelatin, 200 uM each dNTP, 1-uM sense and antisense primers, 40 units/ml Taq polymerase, and 1-200 ng sscDNA. The reaction mixtures were layered with 19 mineral oil and put through a heating regimen that typically included a 3-5 min. 94° initial denaturation step, followed by 25-30 thermal cycles containing a 1 min. 94° denaturation step, a 2 min. 37° primer annealing step, and a 3 min. 72° elongation step. Wig was performed essentially as described by Sambrook et a1. [1989] except that agarose concentration as high as 5% was routinely used to provide resolution of DNA as small as 50 bp. Such gels were made as lower percentage gels but required careful heating in a microwave oven. Gels and chamber buffers were 1 x TAE, 0.5- ug/ml ethidium bromide. W was accomplished by the band interception technique originally developed by Winberg and Hammarskjold [1980] using NA-45 DEAE membranes (Schleicher & Schuell) . After DNA was sufficiently separated on an agarose gel, the gel was out just ahead of the band of interest so that a pre-equilibrated NA-45 membrane could be inserted. Then the gel was returned to the buffer chamber and electrophoresed until the DNA ran into the membrane. The membrane was briefly washed in 0.15 M NaCl, 0.1 mM EDTA, 20 mM Tris-Cl, pH 8.0. and then the DNA was eluted by incubating the membrane at 60° for 30 min. in two washes of 1.0 M NaCl, 0.1 mM EDTA, 20 mM Tris-Cl. The solution was extracted with three volumes of water-saturated n-butanol to remove residual ethidium bromide and precipitated by adding 2.5 volumes 20 ethanol followed by immersion of the tube in crushed dry ice for 10 min. W DNA was digested with restriction enzymes as described by Sambrook et a1. [1989] or as recommended by vendors. Purified PCR products were ligated, with ends either blunt or 5' overlapped, into pUC19 using protocols from Ausubel et a1. [1987]. Recombinant DNA was used to transform DHSa E. coli and DNA was prepared from the resulting transformants by methods of Sambrook et a1. [1989] W The nucleotide sequence of cloned DNA was determined using Sequenase version 2.0 modified. bacteriophage T7 DNA polymerase and its accompanying reagents and protocols with [”SJ-dATPaS and M13 -40 primer. Reaction products were separated on a 60 cm, 7 M urea, 8% poly- acrylamide gel. W A bovine retinal Agtlo cDNA library (Clontech) was diluted with SM to 2.5 x 105 pfu/ml. 100 pl aliquots were mixed with 300 pl SM and 600 ul of an overnight culture of Y1090 E. coli grown in LB broth plus 0.2% maltose. The mixture was incubated at 37° for 20 min. Then 10 ml aliquots of molten (45°) LB soft top agarose were added, mixed, and poured onto dry, 37°, 150 mm LB agar plates. After the top agarose hardened the plates were inverted and incubated at 37° until the plaques were 0.1-0.2 mm in diameter (typically 4-5 hr.) . Then the plates were stored at 4°. W BA85 (Schleicher & Schuell) nitro- 21 cellulose filters (132 mm) were wetted in 5x SSC, blotted on filter paper, and placed for 5 min on the surface of plates at 4° containing plaques. Second and third lifts were left for 7 and 9 min, respectively. Filters were removed and placed plaque-side-up on 2 sheets of 3MM paper saturated with 0.5 N NaOH, 1.5 M NaCl for 5 min. Filters were transferred to 2 sheets of 3MM paper saturated with 0.5 M Tris-Cl,pH 8.0, 1.5 M NaCl for 5 min, then to 2 sheets of 3M paper saturated with 2x SSC for 5 min. ‘Filters were blotted and air dried on 3MM paper, then dried in a vacuum oven at 80° for 1 hr. Finally, the filters were stored at room temperature in a vacuum. desiccator. Engg_L§bgling;_ Oligonucleotides were end-labeled using T4 polynucleotide kinase and [7-32P] ATP by the method of Ausubel et a1. [1987]. WWW; Nitro- cellulose filters containing plaque replicas were wetted and washed 3 times for 10 min each in 3x SSC, 0.1% SDS at room temperature, and once in the same solution at 65° for 1 hr. Southern blots were simply wetted in.the same solution at room temperature. All filters, regardless of type, were next prehybridized for 1 hr at 37miwith gentle shaking in 6x SSC, 5x Denhardt's solution, 0.05% Na pyrophosphate, 100 ug/ml denatured herring sperm DNA, and 0.5% SDS. Filters were removed from the prehybridization solution and put into heat- sealable bags with 6x SSC, 1x Denhardt's solution, 100 ug/ml 22 yeast tRNA, 0.05% Na pyrophosphate, and approximately 1 ng/ml end-labeled oligonucleotide at >5 x 10’ cpm/ml. Filters were hybridized 24-48 hr at 50% Then they were washed 3 times at room temperature in 6x SSC, 0.05% Na pyrophosphate for 10 min each and 1 time at 65° in the same solution for 30 min. Damp filters were resealed in heat-sealable bags, mounted on filter paper, and fluorographed. This protocol was adapted from Ausubel et a1. [1987]. W221: was performed by the methods of Sambrook et a1. [1989]. W As the P, and P, tryptic peptides were purified they were submitted to either the Macromolecular Sequence, Synthesis, and Structure Facility, Department of Biochemistry, Michigan State University or to Alan J. Smith, Protein Structure Lab, University of California, Davis, for N- terminal sequence analysis. The first three peptides submitted, a-43-6, a-29-9, and a-42-16, did not yield any. sequence, each appearing to be obscured by a contaminant. The source of the contaminant turned out to be a bad batch of "BPLC grade” acetonitrile. After switching vendors there was considerably more success at sequencing peptides. There were an additional nine a and nine 3 peptides submitted. Of those, 12 yielded peptide sequence, one yielded two independent peptide sequences, and one yielded a very unsure three residues (see Table 1). Of the other four, three yielded no detectable sequence and one was lost from the sequencer during a power failure at MSU. The ’1! labeled peaks from the chromatograms were easily located but were significantly smaller than other peaks. When sequencing was attempted on some of these the amounts proved to be too small. This seemed to indicate that labeling by 23 24 W This table reveals the sequences of numerous P, and P, tryptic peptides and their similarity with published P,, P,, and Pa. deduced amino acid sequences. The peptides are identified in the two left-hand columns. The a- and B- prefixes indicate that the peptides were derived from P; and P, respectively, which were physically separated on denaturing gels, stained with KCl, and cut apart with a razor blade. After electroeluting from the gel slices the subunits were carboxymethylated and then digested with trypsin. Numerous tryptic peptides were separated by C" reverse phase HPLC and several of those were sequenced by Edman degradation. Residues that could not be determined are designated by a dash (-) . Residues that could not be determined with complete confidence are bracketed. In the three right hand columns the peptides are compared with the published deduced amino acid sequences for P,,, P,, and cone cell P,,. from Ovchinnikov et a1. [1987], Lipkin et a1. [1990], and Li et a1. [1990] respectively. Comparisons without significant similarity between published and peptide sequences were marked with two dashes (--). * Tentative partial match with P, residues 483-502. At residue 5 a determination between proline or leucine could not be made . Peptide_ln a-24-3-6 a-26-6-7 a-28-16-9 a-32-6-5-3* a-36-12-4 a-44-1,2-5-2 0-44-6-8 B-23-4 8-23-5 B-23-7 B-30-9a fi-ao-eb B-41-1 B-41-2R B-46-12 Segnense DFQDNLQAEK EINFYK VIDYILHGK VFEQ[P,L]NMILQGDAIV[TS-QS] FLDQNPGFADQY VALLVAAEF-E [VIV] SQNPLAK AAELPK V[T]FK LFNVHK AFLNCDR VAAEFWEQGDLER FHEEI-PM VLAD[T]ELWI 25 Rod .2._ 100% 100% 100% 30% 67% 100% 100% 83% 100% 100% 100% 100% Rod .19.. 40% 83% 100% 100% 100% 100% 100% 50% 67% 100% 100% Sequence Similarity C0118 .1Lt. 66% 89% 50% 90% 57% 66% 33% 57% 92% 86% 26 carboxymethylation did not occur as expected. In addition one of the unlabeled peaks, 8-30-9-5, yielded a sequence with an unlabeled carboxymethylcysteine. The comparisons in Table 1 reveal that while some of these peptides corresponded to published sequences of either the P, and P,, some, including the longest Pa peptide and the second longest P, peptide, corresponded to neither. These nonmatching peptides were isolated as major HPLC peaks. They cannot be dismissed as peptides from contaminating proteins, nor do they match P7 [Ovchinnikov et a1. 1986] , cone cell Pa. [Li et a1. 1990], or bovine trypsin [Keil, 1971]. A In addition, two of the P, peptides matched the published P sequence but only partially matched the published P, sequence. The converse was true for one of the P, peptides. These results indicated that while the parent subunits of these peptides were closely related to the published sequences, neither one was actually encoded by them. Even though the published Pa and P, cDNAs almost certainly code for authentic cGMP PDE subunits, other isoforms most likely exist in significant quantities. WW Three PCR primers were synthesized corresponding to the peptide, a-32-6-5-3 (see Figure 2A), and used in PCRs as described in Methods. This resulted in a product of the expected size, 60 bp, seen as a minor band on an agarose gel (see Figure 3). Interestingly, the two major products, 120 and 180 bp, were two and three 27 Iisurg_zi_293_£rinsrai A. The initial PCR primers were designed corresponding to opposite ends of peptide.a-32-6-5-3, VFEQ[PorL]NMILQGDAIV[TS- QS]. It was chosen because it was the longest peptide and it had only limited similarity to the published P, deduced amino acid sequence [Ovchinnikov et a1. 1987]. The primers were designed with restriction sites to aid in subsequent cloning of the products and 14 bases of degenerate sequence corresponding to amino acid sequence from the peptide as was done by Lee et a1. [1988]. To reduce the amount of degeneracy in these mixed oligonucleotide primers a bovine rod cell codon utilization table (2,511 codons from 8 rod cell cDNA sequences) was created (see Table A1 in the Appendix, p.98) and used to eliminate some of the least often-used codons. This table used only rod cell sequences to eliminate the effects of tissue-specific differences in codon usage [Newgard et a1. 1986]. The sense primers corresponded to the five N-terminal amino acids of the peptide. Because of the ambiguity of the fifth amino acid 2 different sense primers were made. On the 5' end of each there was a HindIII restriction site and both had a 16-fold degeneracy. The antisense primer was made corresponding to amino acids 11 though 15 of the peptide, the last 5 certain amino acids. On the 5' end there was an ECoRI restriction site. It had a 24 fold degeneracy. All the primers are lined up with the corresponding amino acid sequence above them. Nucleotides below the primer line designate the sites of oligonucleotide degeneracy. B. From several.Ig peptides additional PCR primers were designed. In an effort to make these primers more specific for their target sequences the restriction sites on the 5' ends were omitted, leaving each oligomer with 20 bases that corresponded to the peptides. It was thought these 5’ noncomplementary portions of the primers might hinder initial annealing to the target templates thus favoring nonspecific interaction resulting in more artifacts. Sense and antisense primers were again produced corresponding to a-32-6-5-3, but this time, both corresponded to the internal region. One primer, Sense*”'“'”, was produced corresponding to two peptides, a-26-6-7 and a-28-16-9, that were found to be next to each other in the published sequence [Ovchinnikov et a1. 1987]. 28 W . Val Phe Glu Gln Pro Sense Primer #1 5' CAAQQIIG GTC TTC GAA CAA CC 3' Mind III G T G G Val Phe Glu Gln Leu Sense Primer #2 5' CAAGQIIG GTC TTC GAA CAA CT 3’ Hind III G T G G Val Ile Ala Asp Gly Antisense Primer 5' GQAAITQC AC AAT AGC ATC CCC 3' Eco RI G G G G T W Asn Met Ile Leu Gln Gly Asp Ala 1) Sense°*m” 5' AAC ATG ATC c'rc CAG sec GAC cc 3 . c G Ala Asp Gly Gln Leu Ile Met Asn 2) Antisense‘rm” 5' so are ccc CTG CAG GAT CAT cm 3' G c Gln Asn Pro Gly Phe Ala Asp Gln Tyr 3) Sense°*”‘“ 5' CAG AAT ccc soc m ccc GAC CAG TA 3' G G Asp Phe Gln Asp Asn Leu Gln Ala 4) Sense*“4‘ 5’ GAC TTT CAG GAC AAC CTG CAG GC 3' Asn Phe Tyr Lys Val Ile Asp Tyr 5) Sense’”""'” 5' AAC TTT TAC AAG GTC ATT GAC TA 3' Glu Trp Phe Glu Ala Ala Val Leu Leu 6) Antisenseflm’” 5' TC CCA GAA 'r'rc AGC AGC AAC CAG CAG 3' 29 MW Three PCR primers, two sense and one antisense, were created corresponding to peptide a-32-6-5-3 and used with oligo d(T)-primed sscDNA in PCR. The reaction mixture was extracted with phenol/CECE, ethanol precipitated, dissolved iJIIgO, and fractionated on a 5% agarose gel in 1x TAE, 0.5 ug/ml ethidium bromide in lane 4. Molecular weight markers (1 pg pBR322 HaeIII digest) were run in lanes 2 and 6. Visualization revealed a faint band at 60 bp, the size of the expected product, and predominant bands at 120 and 180 bp. 30 31 times the length of the expected product. There also was a smear of background products and some other minor products. Three bands, 60, 120, and 180 bp, were intercepted with, and then eluted from NA245 membranes (Schleicher and Schuell). This material was cloned into pUC19 using both cohesive end and blunt end ligations. Several different clones were found with approximately 60 bp inserts but upon sequencing none was found to code for the peptide sequence of interest" No clones were recovered containing the 120 bp insert and only one clone from the 180 bp ligations was found, ZHowever its sequence did not code for the peptide nor did it have any ORF coding for a1 PDE-like sequence. Six additional primers were synthesized corresponding to various P; peptides (see Figure 28). Three different bovine retinal sscDNA templates were used in these PCRs, one primed with oligo d(T) and the others primed with one of the two antisense primers. Many different combinations of primers and templates were tried, each yielding a variety of products, most under 500 bp long (see Figure 4). Those PCRs that utilized primers matching the published sequences did not yield products of the expected sizes (2190, 2034, and 1293 bp in Figure 4, lanes 3, 5, and 7 respectively). Still, several bands from other PCRs were isolated from agarose gels. Direct sequencing attempts of the bands were frustrated by the presence of multiple templates. It was concluded that this approach could not succeed unless PCR products were 32 W11: The PCR products fractionated on a 1% agarose gel illustrate some of the numerous combinations of primers. In this case the PCR primers listed in Figure 2B were used in the relevant combinations with sscDNA primed with the same antisense primer used in the PCR. Lanefl contains 500 ng A DNA-EcoRI-HindIII digest. The following lanes contain PCR products created using primers identified in Figure 28: Lane#2- Sense"*3’"*5‘3 and Antisensewm'” Lane#3- Sense““““ and Antisensewm'” Lanef4- Sense°'3‘*"“ and Antisense°‘”“'3 Lane#5- Sense“"*“ and Antisense"""‘v2‘5'2 Lanef6- Sense*“4‘iand Antisense"'32'°'5'3 Lane#7- Sense°‘”“’” and Antisensewm'“ Lanef8- Sense°"°""'” and Antisense°*32“'3. 33 Eisnre_11 12345678 564 bp-- 34 individually trapped by cloning, as was successfully accomplished later on with the phospholipase C probes. W Without immediate success at PCR, a longer, less degenerate oligonucleotide ”guessmer" probe, also based on peptide a-32-6-5-3, was synthesized. Such probes have been touted as a better alternative to using short highly degenerate mixed probes [Ausubel et a1. 1987; Wood, 1987]. The best codons were determined using the bovine rod cell codon utilization table (see Table A1, p.98). This resulted in a synthetic 41 base oligonucleotide that was degenerate only at the position where’ the amino acid was uncertain (see Figure 5). This probe was used on Southern blots of genomic DNA and even on lifts of a bovine retinal cDNA library without.gaining any reliable results. However, background hybridization appeared unusually high, and this led to the discovery that the probe bound very well to A-DNA. Considering that the libraries on hand were both in A vectors this probe was not going to be immediately useful. 35 Leu Phe Glu Gln Pro Asn Met Ile Leu Gln Gly Asp Ala Ile Val 5’ TTT GAG CAG CCC.AAC ATG ATC CTG CAG GGC GAT GCC ATT GT 3' TO A 41 base oligonucleotide was designed corresponding to the a-32-6-5-3 peptide following the recommendations of Ausubel et a1. [1987] and Wood [1987]. For each amino acid the single best codon was determined from a bovine rod cell codon utilization table (see Table A1 in the Appendix, p.98). 21m 29E_§gbgni§_§gparatign‘ The long-range purpose of this research was to better understand the function and regulation of the cGMP PDE of vertebrate photoreceptors. As initially proposed, my role was to isolate the large PDE subunits, Pa and P,, and to use these purified subunits to: 1)obtain peptide sequence information that would in turn be used by Steve Pittler to find the corresponding cDNAs, and then 2)study their functional roles in the visual cascade through reconstitution experiments. Because reconstitutions were planned it was initially hoped that the Pa and P, could be separated under nondenaturing conditions. However, soon after isolation from the ROS, the PDE invariably started to aggregate yielding a white precipitate. This phenomenon was alsoreported earlier by Kohnken et a1. [1981] . This material could only be solubilized again in the presence of strong denaturants like SDS, urea, or guanidine hydrochloride. Even then not all of the material could be dissolved. As a result of this aggregation, HPLC gel filtration using nondenaturing conditions was found to be unsuitable for separation of P,,I and P,, leading to the successful separation by SDS-PAGE. W Amino acid sequences could not be 36 37 determined directly from the intact P, and P, because their N- terminal amino groups were blocked, preventing Edman degra- dation [Ovchinnikov et a1. 1986]. To obtain sequencable peptides it was decided to trypsinize the subunits. At the time preparations for trypsinization of P, and P, were being made, Steve Pittler requested that cysteine-containing peptides specifically be searched for in order to verify'which of his cDNA clones actually coded for a PDE subunit. He had recently identified three cDNA clones that presumably coded for a PDE subunit but were not full length. The three clones were identical except that they diverged at the 3 ' end of their ORFs. The clone that Steve suspected represented a ”major” PDE isoform had in its deduced peptide sequence a C- terminal CAAX motif commonly associated with isoprenylation of peripheral membrane proteins [Hancock et a1 . 1989] . The predicted C-terminal tryptic peptide from this clone contained two cysteine residues. To aid in the search for cysteine-containing peptides, P, and P, were 3H-carboxymethylated using tritiated iodoacetic acid. However the resulting tryptic digest yielded only very minor labeled peaks that proved to contain too little material to sequence. It also yielded at least one unlabeled peptide that contained a carboxymethylated peptide. One possible explanation for these results is that the labeled iodoacetic acid may have been degraded, such as by exposure to light. During the time the 14 sequenceable peptides in Table 1 38 were being generated, Ovchinnikov's group published it's Pa cDNA and deduced amino acid sequence [1987] . Other bovine retinal PDE sequences followed: another P, from Pittler et a1. [1989], P, from Lipkin et a1. [1990], and the cone cell P,. from Li et a1. [1990]. To find cDNA coding for the missing isoforms it was decided to use the polymerase chain reaction. Comparison of the deduced amino acid sequences of P,, P,, and the cone P,. reveals that the emerging family of photoreceptor cGMP PDEs has extended consensus regions and only very narrow hypervariable (HV) regions (see Figure 6). It is anticipated that the source of Table 1's non-matching peptides is/are one or more HV regions of predominant PDE subunit isoforms other than those_ published. In fact the longest peptide, a-32-6-5-3, does exhibit limited sequence similarity (30%) to the Pa deduced amino acid sequence at residues 483-502 (see Table 1 and Figure 6). This is a variable region where P, and Pa. also have limited sequence similarities with PM of 45% and 40%, respectively. Another peptide, 6-46-12, has no apparent similarity to published P, or P, sequences. It pre-sumably comes from either the N- or C-terminal ends of its parent protein where the different isoforms diverge almost entirely. WW After noting the discrep- ancies between the peptide sequences and the deduced amino acid sequence from published P, cDNA [Ovchinnikov et a1. 1987, the only one published at the time], a DNA probe was designed to preferentially identify, from a cDNA library, P, clones 39 MW The amino acid sequences deduced from the published cDNAs for P, [Ovchinnikov et a1. 1987], P, [Lipkin et a1. 1990], and P, [Li et a1. 1990] were lined up and gaps were introduced to reveal their greatest sequence overlap. Dashes (-) in the P, and P,. sequences represent identity with the P, residue directly above it. Substantial hypervariable regions are seen near the C- and N-termini (approximately residues 25-80 and 830-850) but only in limited areas in the internal regions. 40 limb. NGEVTAEEVEKFLDSIVSFAKOYYILRYRAKVISOLLGPREAAVDFSNVHALISVEESEIIFDLLRDFODILOAEKCVFN - SLSEGfl-flR---0-PG--D--FGRKLSPEDVAIACEDGCPEGCT-FIE -C0----AAL-E-V0-l-E-VIH-RV--K ---ISO-T'--Y-EA-P0---E-F-RKLOVE-P-GGAOAPAS-SFPGRTL-EEAALYL-LLEV-'EEASSVEL-ANRALO VIKKLCFLLQADRMSLFIVIARIGIAELAIILFIVHKDAVLEECLVIPDSEIVFPLDISVVGHVALSKKIVWVPITEEDE lLRl--Sl-fl---C ------ e---v -------- S-OP-S---D---P ---------- r ------- or--u---eovn-cp n-Ae------c-u-Lc ----- TP-V-SK-LD-IPISKF-DI--V-~l-A-----V-I--H--uT--TF---DVKKNS HFCUFVUILIEYUTKIILASPIH'GKDVVIIINVV'KVDGPHFTENDEEILLKYLIFANLIHKVFHLSYLHNCEIRRGOI .-ss-‘-E.-D-V.R----T.---------v--A---L---c--“--DvF ...... Gr-uL-lY .............. v "S'-I'KQ'G'V'I'L"T"VI"E'L'VF‘A ..... ASE'S‘O"'VFS"'S'VSI‘L'LH'TI"Y'I'S"S'- LLUSGSKVFEELIDlEIOFNKALYIVIAFLICDRYSVGLLDMTKOKEFFDVUPVLMGEAPPYAGPRTPDGREIUFYKVID ----Au ---------------- r ----- v --------------- E -------------- eA-s ---------- L ------ -u--Au -------- v ------------ rv---£---x-------£---v-e---xL--ve--x--x ------ v1---r-- v1Luexzorxv:PuPPPouHALvseLPrvvneueLIcuruIApseorrArexepLoeseunrKIVLsuprvuxxzeivev ------------- S--A-----A--------ES-F--------ADEM-l--EG---D---lV------------------ ------- e-----r--u---r-t--------£-~r---uL---Ao£v-r---e-v--r--v------L-------o---- ArrvunxnexproenosrLMESLrerLGHSVLIPDIVELnuxieuRKn1raonvxvnvxcoueerorrixrne vvexep -------------- 0--V---------------T---DK----------A----L---R--R----L--P--- RL---- -------------- Y-~fllA-1--------L--1----K'---------A-E-Lll-T-ATPD--KS---FK-KLNIDVI HECEEEELAEILOGELPDADKYEINKFHFSOLPLTELELVKCGIOMYYELKVVDKFHlPOEALVRFMYSLSKGYRRITYH M"’D"K"KEV"WA'FD'YE ...... EC ................ 6--R-.°----v----LF-v .......... ED---K0-VT--KED---PRTADLYE-l-Rfl--l--H--l---LRLFF-ll--E--KV-V-V-T-H--TVR----AV--- unerwmrnmwrexinmeuumercmwuacruuumxsoupunuessrLeauuiermu - ------- A----r--n ----- s-v ----- r ----- or --------------------------------------- r-- ------------- r--u--n--K-v-----r--LA--------------------rs---a-------------vs---- RDESLIIFQNLNRRONEHAIHIIDIAIIATDLALYFKKRTHFOKIVDOSKTYETOOEHTOYHHLDQTRKEIVHAIHHTAC 35.1.-.1 .......... V"L ......................... E"I-'DRKS'VE'LS'ET .............. g ........... K"Y'TV"LFEV ...................... ACEKI--EE-AIK-VTI-P'K'--I ........ DLSAIrKPuevoSKVALLVAAErueeeotearVLaouPiPunnnuonELPKLovertorvcrrvvxrrsarnEEIIPuL ---------------------------------- 0-0----------A----------------------------L--F ------------- e---~--u---------------e-—-------x--------------------------x------ DC!TIIRKEUKALADEYEIKIKGLEEEKOKOOAANOAAAGSONGGKOPGGGPASKSCCVO -RLO -------------- A-V-A---DOK-E rr-xxverercue-PA-n-sr-nrL l-L0---V---S ----- o£---vr--n- --eee-rr suaveoseee-ooxx--r-Lni 160 157 240 237 320 318 317 398 397 479 477 477 559 557 557 639 637 719 717 717 797 797 059 853 855 41 coding for the alternative sequences. It‘was to be created by PCR techniques adapted from Lee et a1. [1988] using primers corresponding to both ends of peptide a-32-6-5-3 (see Table 1 and Figure 2A) . It was chosen first because it was the longest peptide sequence obtained and it did not have signif- icant sequence similarity with the published P,. When it was found that identifying PCR products corre- sponding to peptide a-32-6-5-3 from a mixture including many artifacts was going to be exceedingly difficult, additional PCR primers were prepared. These corresponded to several of the shorter isolated P, peptides (see Figure 28) . However, with these primers the subsequent PCRs, like the earlier ones, yielded complex mixtures of products. Long products that were expected in some of the PCRs did not show up, in retrospect because reactions with degenerate primers and complex template mixtures favor shorter products. Still, the number of PCR products obtained gave promise that with sufficient care and persistence, the underlying variant Pa sequences, probably multiple, are very likely accessible by PCR amplification. A_Eggu§_1gz_£utuzg_3g§eazgn; Future research should be directed at identifying the various isoforms of the photo- receptor PDE, verifying their localization in the photo- receptor cell, and determining their relative expression levels. With those items accomplished, the functional role of each isoform can be examined. An important first step for any future research is to obtain considerable additional peptide 42 sequence from separated P, and P, to further verify the alternative isoforms. This will yield.more information about the predominant isoform(s) and allow the production of probes for cDNA libraries. When producing peptides, trypsin could be used again which may be useful in verifying already identified peptides. However the purified Pa and P, subunits could also be cleaved with other proteases that will yield fewer and longer sequenceable peptides. Enzymes are commercially available that cut exclusively at arginine or at lysine rather than at both as trypsin does. using published cDNAs the following ‘ table was created giving the number of peptides theoretically produced, followed by the average peptide length: Rod Rod Cone PDE-a PDE-B PDE-a ’ Trypsin 97/ 8.9 98/ 8.7 108/ 7.9 Endoproteinase Lys-C‘ 63/13.6 55/15.5 73/11.7 Submaxillaris Proteaseb 35/24.5 44/19.4 37/23.1 ' Boehringer Mannheim Biochemicals ° Pierce. From such alternative digests it will potentially be easier to isolate purified peptides of which many should yield more sequence information than those from a tryptic digest. There are nonenzymatic methods that could also be used to fragment the subunits. Peptides were successfully isolated from cyanogen bromide digests of undifferentiated Pd, 43 [Ovchinnikov et al, 1987, Lipkin et al, 1990]. Early on in these studies I attempted cyanogen bromide cleavage of purified P, and P, but ended up in both instances with mostly intractable tars. Possibly with greater diligence we can successfully employ this cleavage method. A second chemical method that could be utilized is treatment with 70% formic acid. This method cleaves the relatively uncommon aspartate-proline peptide bond as was done with G” [McConnell et a1. 1984]. No such sites exist in the deduced amino acid sequences for P, and P, but two are found in the closely related cone P,. This method may be able to ’ reveal very quickly the existence of multiple isoforms within an apparently homogeneous PDE-subunit. If after treatment with 70% formic acid only a portion of the material remains uncleaved as revealed by SDS-PAGE then that argues that more than one sequence is present. It is also a potential way to purify a formic acid insensitive isoform away from a sensitive one. Formic acid fragments may be directly sequenced or digested further by other means to obtain manageable peptides. The P, variants identified by Pittler suggest that the C- terminal region of the subunits could be important in identifying the predominant isoforms. Not one of the published cDNAs, nor Pittler's two variants, have yet had these C-terminal regions confirmed with any direct peptide sequence. A technique for quickly purifying C-terminal tryptic peptides has recently been offered commercially 44 (Immobilized.Anhydrotrypsin Kit, Pierce). The anhydrotrypsin binds tryptic peptides with C-terminal arginine or lysine residues while C-terminal peptides, lacking these residues, elute through the column. These peptides can then be resolved and sequenced confirming which, if any, of the C-terminal regions represents the predominant isoform. With additional peptide sequence in hand we will be in a better position to develop successful probes for screening bovine retinal cDNA libraries for clones that code for various PDE isoforms. One set of probes could be produced by PCR using primers corresponding to some of the more complete ' peptide information. Cloning and screening PCR products has been made much easier with the introduction of the TA Cloning kit [Invitrogen] as was later demonstrated with PCR products coding for PLCs (Chapter II). Individual clones can be sequenced to verify that they code for the peptides in question and to identify likely variants amplified from the cDNA of different isoforms. Another potential set of probes could be antibodies raised against synthetic peptides based on the generated amino acid sequences. The specificity of the antibodies can be tested on Western blots of total ROS protein and purified PDE and then used to screen an expression library. One strategy for identifying'multiple PDE isoforms would be to screen a bovine retinal Agt11 expression library (Clontech) with a set of PCR produced probes corresponding to 45 the conserved catalytic domain of PDEs [Li et al. 1990]. The numerous clones that are likely to be identified by this screening can then be further probed with oligonucleotide or antibody probes which correspond to unique peptides coming from hypervariable regions. After the variant PDE subunit isoforms have been identified and their sequences revealed by cDNA clones, the expression level of each isoform can be examined. A.first approximation of the transcript levels could come from Northern blot analysis of total RNA and poly A? RNA [Sambrook et a1. 1989] using probes that correspond to unique regions of * each isoform. Next, a more accurate measurement might be made by ribonuclease protection assays [Ausubel et a1. 1987; Sonnenburg et 81. 1991]. Labeled antisense RNA probes, also corresponding to unique regions of each isoform, are hybrid- ized to the target RNA and then treated with RNases which digest single-stranded RNA. Transcript levels are then calculated from the amounts of incorporated label remaining. If the transcript level of any isoform proves to be too low to be detected accurately by Northern blot analysis or ribo- nuclease protection assays, it can be determined after amplification by PCR using the methods of McConnell and Linz [unpublished studies]. Knowing the expression levels of the variant isoforms will aid in subsequent research examining their localization and functional roles in vision, and in other forms of cellular signaling. CHAPTER II PHOSPHOINOSITIDE-SPECIPIC PHOSPHOLIPASES C W W In the vertebrate ROS, modulation of [Ca“] is believed to mediate light-adaptation [Matthews et al. 1988]; however its mode of action has not been established. It is known that in the dark Ca++ enters the ROS along with Na+ through the cGMP-gated cation channel [Koch, 1992] (see Figure 1, p. 3). Simultaneously CaM is extruded via a Na“/Ca+”,K+ exchanger, ' harnessed to an outward K+ gradient [Cervetto et al. 1989] . Upon illumination the cGMP is depleted and the cGMP-gated cation channel shuts stopping the influx of Ca++ and Na+ while Ca“ efflux via the exchanger continues [Koch, 1992]. One possible explanation of CaH-mediated light adaptation is the activation of guanylate cyclase at low [Ca”]. It has been shown that a 23 kDa protein named recoverin activated ROS guanylate cyclase in vitro when free [CaH] was lowered from 450 nM to 20 nM [Dizhoor et al. 1991]. However, this has been recently challenged by a report that purified bovine recoverin, as well as homologs from chicken and Gecko lizard retina, added to intact Gecko ROS were shown to delay recovery after illumination [Gray-Keller et al. 1993]. Another possible role for Ca++ is the activation of cGMP 46 47 PDE. A 26 kDa protein from frog ROS reportedly binds to ROS disks and stimulates cGMP hydrolysis at 1 uM Ca++ level [Kawamura and Murakami, 1991]. The mechanisms for the modulation of [Ca“] have not been established either. Certainly calcium flux via the cGMP-gated cation channel and the Na“/Ca“+,K+ exchanger plays an important role [Koch, 1992] . But so too must the calcium stores associated with ROS disks where most of the ROS calcium is found. These stores are sensitive to cGMP and calcium ionophore A23187 [Devenny and Clack, 1991] . It is also likely that these stores are sensitive to inositol 1,4,5-tris- ‘ phosphate (IP,) as are Ca“ stores in most cell types [Berridge 8 Irvine, 1989]. WWW Phoephatidyl- inositol 4,5-bisphosphate (PIP2) phospholipases C (PLCs) are a class of enzymes involved in various signal transductions. They are activated by transmembrane receptors, either directly or via a G-protein or a kinase [Majerus et al. 1990] . They in turn cleave the PIP, phosphodiester linkage to yield two second messengers: diacylglycerol (DAG) which activates protein kinase C, and IP3 which has the ability to release Ca2+ stores in many different cell types [Berridge and Irvine, 1989] . Considering that both the activators and the substrate are membrane-associated it follows that PLCs need to bind to a membrane to be functional. However, PLC activity has been demonstrated in both particulate and cytosolic fractions [Rhee 48 et a1. 1989; Gehm and McConnell, 1990b] and at least one isoform, a bovine brain B-PLC, has been isolated from both [Lee et al. 1987]. Certainly one attractive model for PLC activation is translocation from the cytosol to the membrane. In the ROS, PLCs likely play a role in calcium modulation via IP3 production. Rapid PIP2 synthesis and multiple PLC isoforms have been detected in the bovine ROS [Gehm and McConnell, 1990a&b] although light activation has not yet been established. A number of PLC cDNAs have been cloned and sequenced by S. G. Rhee's laboratory at NIH and by others. They have been ‘ categorized based on immunological cross-reactivity and extended sequence similarities into four different families denoted a, B, 7, and 6 [Rhee et al. 1989]. With the exception of the a-PLCs, all the families have consensus regions denoted X and Y. Outside of these regions the PLC families have little sequence similarity among them. Additionally, within the 7-PLC family are highly conserved regions between X and Y. These have been designated SH2 and SH3 (src homology 2 and 3) and have significant sequence similarity with the src-encoded family of tyrosine kinases. One or both of these sequences have also been found in crk oncogene from avian sarcoma virus CT10, the GTPase activating protein (GAP) which activates the res gene product p21, and the a-subunit of spectrin [Rhee et al. 1989]. The work presented in this chapter was initiated to 49 complement the continuing ROS PLC purification and character- ization studies in our laboratory. It has led to some provocative insights and has provided some powerful tools, not before available, with which to further analyze these enzymes. HATEEIAL§_AND_HETEQD§ W Commercial suppliers were the same as in Chapter I. c s ' c . Oligo- nucleotides were synthesized by the Macromolecular Structure, Synthesis, and Sequence Facility, Department of Biochemistry, Michigan State University on an Applied Biosystems 3808 DNA synthesizer. The crude synthetic oligonucleotides were - purified by HPLC using a Zorbax Bioseries Oligo column (DuPont). The column was equilibrated with 80% 50 mM NaPO“ pH 7.0, 20% acetonitrile, 0.1 M NaCl. The oligonucleotides were applied and then eluted with a linear 0.1 to 0.8 M NaCl gradient in the same solvent in 70 min. Peak Am, fractions were pooled and desalted using C1, Sep Paks (Waters) as described by Atkinson and Smith [1984]. BNA_1§91gtign; Bovine eyes were obtained from Murco, Inc. of Plainwell, MI. Within 3 min. of slaughter fresh retinas were removed from the eyes and dropped into liquid nitrogen. Poly A+ RNA was isolated from the frozen retinas using the Fast Track kit (Invitrogen). This kit yielded poly A+ RNA with Ago/Am ratios consistently >1.8 in significantly less time than the traditional protocols used earlier. It 50 51 also had less contaminating non-poly A+ RNA as judged by the relatively low intensity of rRNA bands seen on formaldehyde-1% agarose gels [Fourney et al. 1988]. gpflg_§yn§hg§1§ was performed as described in Chapter I. W Thermal cycling was performed with a DNA Thermal Cycler or a Gene Amp PCR System 9600 (both Perkin Elmer Cetus). The reactions were carried out as described in Chapter 1. W was performed as described in Chapter I. W was Performed as ' described in Chapter I. Cloning_£g3_zrggggt§‘ The TA Cloning kit (Invitrogen) was utilized. It came with a precut vector, pCR 1000, which has 5' single T overhangs. This allowed direct ligation of the PCR products without having to resort to restriction cutting and subsequent cleanup. The PCR products, which inherently have 5' single A overhangs as catalyzed by Taq polymerase, were ligated to the vector with T4 ligase. The recombinant DNAs were used to transform competent INV 1aF' E. call which were provided with the kit. Transformants were then screened by their ability to catabolize the chromogenic substrate, x-gal (blue/white screening). W was performed by the methods of Ausubel et a1. [1987] using T4 polynucleotide kinase and [7-32P] ATP. 52 W was performed using two different methods. PCR products were sequenced both directly and after being cloned using Sequenase version 2.0 modified bacteriophageT7 DNA.polymerase (USB) and its accompanying reagents and proto- cols with [”SJ-dATPaS and with either the M13 -40 forward primer or a PCR primer. The PCR-amplified inserts of Agt10 clones were sequenced by cycle sequencing using the dsDNA Cycle Sequencing Kit (BRL) and 32P-end labeled primers (see above). Probe_Lgbeling4, Double-stranded DNA probes were labeled in three different ways: by the method of random primers [Sambrook et al. 1989]; by T4 PNK end-labeling [Ausubel et a1. 1987], and by PCR techniques. To label by PCR the reactions were set up and run the same as standard PCR except that the normal 200 uM dGTP was replaced with [a-“PJ dGTP diluted with an equal amount.of unlabeled.dGTP for a final concentration of approximately 1 uM [a-nP] dGTP. The quality of the reactions was monitored from autoradiograms of products on agarose gels and the total amount and specific activity of the products were calculated on the basis of the total amount of incorporated label. WWW were performed as described in Chapter I. Sggthern and Northern Blot Analyses were performed by the methods of Sambrook et al. [1989]. Hybridizatign with Oligonucleotide Probes was performed 53 as described in Chapter I. W Nitro- cellulose filters containing plaque replicas or Southern- blotted DNA were wetted in 6x SSC, 0.5% SDS, then inserted into heat-sealable bags and incubated at 42° for 1 hr with 6x SSC, 5x Denhardt’s solution, 0.5% SDS, 100 ug/ml denatured herring sperm DNA, and 50% formamide. Double-stranded, random primer- or PCR-labeled DNA probes were denatured by boiling for 5 min, chilled on ice, and injected into the bags. After resealing, the bags were incubated at 42° with gentle shaking for 2-3 times the Cotm. Label concentrations were 2 x 10‘ to ' 1 x 10‘ cpm/ml with a probe specific activity >5 x 107 cpm/ug. After incubation, filters were washed with four times in 2x SSC, 0.1% SDS for 10 min at rm temp, then once in 1x SSC for 1 hr at 68% Filters were either dried or resealed still damp in heat sealable bags, and then fluorographed. This protocol was adapted from Sambrook et al. [1989]. W was done on IBM-compatible PC equipment in our laboratory, using software obtained from other laboratories and modified to our own needs. WW... Initially three PCR primer pairs were synthesized corresponding to the X-region of three published PLC cDNA sequences: one from a bovine brain 7-PLC [Stahl et al. 1988], and two from.rat brain sources, 3- and 6- PLCs [Suh et a1. 1988] (see Figure 1A). These were used in PCRs as described in methods but had.only limited success“ To start, the individual primer pairs did not yield any product at all. Only when all six primers were used together were PCR products formed. When they were used in PCRs with bovine retinal oligo d(T)-primed sscDNA as the template no discern- ible bands were seen, only a smear of material between 0.2 and 4.0 kbp. However when the sscDNA had been primed using the antisense PCR primers the reactions yielded several distinct bands including one (or more) of approximately the expected size, 450 bp, though the yields were low and there were numerous artifacts of various sizes (see Figure 2). The PCR products were blotted onto nitrocellulose and probed with an internal oligonucleotide probe. The 450 bp product hybridized to the internal probe but nothing in the smear of DNA did (not shown). This suggested that at least some small amount of a PLC x-region had been amplified. In an attempt to improve the 54 55 Marina A. The initial three PCR.primer pairs were complementary to 1)the bovine brain 7-PLC X-region cDNA [Stahl et a1. 1988], 2)the rat brain B-PLC X-region cDNA [Suh et al. 1988], and 3) the rat brain tumor 6-PLC X-region cDNA [Suh et al. 1988]. They were constructed similar to those used by Lee et al. [1988] . Each one was 23 bases long, had a complementary region 15 bases long and an 8 base region at the 5' end that contained a 6 base restriction site and a 2 base ”clamp" to aid in subsequent cloning of PCR products [Scharf, 1990]. 8. The second set of primers (1 sense and 2 antisense) were mixed oligonucleotides that were complementary to both the bovine brain 7-PLC X-region cDNA [Stahl et al. 1988] and the bovine adrenal cortex S-PLC X-region cDNA [Suh et a1. 1988] . They were 20 bases long and complementary to the published sequences for their full length. The mixed nucleotide sites are denoted with an additional base below the main line. 56 Iisnra_1A1_Initial.£§8_£riuer_zaira 1) Bovine Brain y-PLC X-Region: Sense primer corresponding to bases 973-987 XbaI P L S H Y 5’ TGIQIAQA CCC CTG TCC CAC TAC 3’ Antisense primer corresponding to bases 1,372-1,385 BamHI K I L I K 5’ TTQQAIQQ CTT GAT AAG GAT CTT 3’ 2) Rat Brain B-PLC x-Region: Sense primer corresponding to bases 961-975 XbaI P L S H Y 5' TGIQIAQA CCC CTG TCT CAC TAT 3' Antisense primer corresponding to bases 1,381-1,395 BamHI K V L I K 5’ TTQQAIQQ TTT CAC CAA GAT TTT 3' 3) Rat Brain Tumor 6-PLC X-Region: Sense primercorresponding to bases 901-915 XbaI P L S H Y 5' TGIQIAQA CCA CTG AGT CAC TAC 3’ Antisense primer corresponding to bases 1,300-1,314 BamHI K L L I K 5’ TTQQAIQQ TTT CAA CAG GAT CTT 3’ zisnra_181.5esondarx_298_2riners Sense, 7-PLC Bases 994-1,014, J-PLC Bases 922-942 S S S H N T Y 5’ TCC TCC TCG CAC AAC ACC TA 3' G G Antisense, 7-PLC 1,093-1113, s-ch 1,021-1,041 G D w c D L E 5' cc ATC CCA GCA GTC CAA CTC 3' G G G Antisense2, 7-PLC 1,219-1,239, 6-PLC 1,147-1,167 I S L I V P Y 5’ AA GGA CAG GAT GAC CGG GTA 3’ T G 57 limb—2.9mm All six of the initial PCR.primers were combined in PCRs using either sscDNA primed with oligo d(T) (lane 4) or sscDNA primed with all three antisense primers together (lane 6). They were fractionated on a 1% agarose gel in 1 x TAE, 0.5 ug/ml ethidium bromide and compared with MW markers (lanes 3 and 7, 1 ug A-HindIII digest). A product of the expected (~450 bp) size was seen in lane 6. The DNA was blotted onto Nytran and hybridized with a labeled internal probe (the secondary sense PCR.primer, Figure 18). Fluorograms revealed a faint signal at the ~450 bp level of lane 6 and no signal elsewhere (not shown). 58 Figure 2 . Lane# 34567 —' 564 bp 59 product yield the PCR conditions were manipulated. This led to the discovery that an approximately 450 bp product could be formed with just antisense primers present, suggesting that purification of the desired product(s) from the artifacts would be difficult. A second set of primers was more successful. The internal probe mentioned was also used as one of three new PCR primers. These primers, one sense and two different anti- sense, each corresponded to sequences between the initial PCR primer pairs. They were mixed oligonucleotides complementary to the bovine brain 7-PLC and the bovine adrenal cortex 6-PLC . (see Figure 18) . These had greater specificity and when used in PCRs yielded fewer artifacts and larger amounts of the appropriate products, 120 or 250 bp depending on the antisense primer used (see Figure 3) . Products were formed equally well whether oligo d(T) -primed or antisense-primed cDNA was used as the template, arguing that primer specificity was the reason the initial PCR primers were less successful. W The 120 and 250 bp PCR products were gel-purified and samples were directly sequenced using dideoxynucleotides and Sequenase (USB) . In separate reactions the sense and appropriate antisense PCR primers for each product were used as sequencing primers. The resulting sequences had ambiguities and were not full length but did reveal two different PLC X-region-like sequences (data not shown). There also were underlying sequences, possibly 60 W The secondary PCR primers (Figure 18) were combined in PCRs using, in this case, oligo d(T)-primed sscDNA. The products were fractionated on a 3% agarose gel and compared with MW markers (lanes 1 and 6, 500 ng p8R322-HaeIII digest). The PCRs used either the sense and antisenseil primers (lanes 2 and 3) yielding a predominant band of the expected 120 bp size, or used the sense and antisensefz primers (lanes 4 and 5) yielding a predominant band of the expected 250 bp size. These products were isolated by the band interception technique and subsequently cloned and sequenced. 61 123456 62 related, consistent with the proposal that PCR would amplify multiple PLC cDNAs if the primers matched closely enough to anneal. The predominant 120 bp sequence had the most sequence similarity with the bovine brain y-PLC [Stahl et al. 1988]. Where the 120 bp sequences were unambiguous they had 89% identity with the 7-PLC. The predominant 250 bp sequence was most like the bovine adrenal cortex S-PLC [Suh et al. 1988]. Likewise they were 60% similar. One result that was not immediately understood was that the 250 bp product yielded two different predominant sequences depending on which primer was used for sequencing: one, the S-PLC-like product using the- sense primer, and the other, an unrelated sequence using the antisense primer. Only after cloning and sequencing multiple PCR products was it determined that there was a major PCR artifact of approximately the same size that had antisense primers at both ends. (glgning_£gg_zzgggg§§‘ The 120 and 250 bp PCR products were cloned using the TA Cloning Kit (Invitrogen) . This system ‘takes advantage of the fact ‘that, Taq jpolymerase catalyzes the addition of a single A to the 3' end of the newly amplified cDNA yielding a single base overhang. The supplied vector, pCR 1000, comes precut and has a comple- mentary 3' single T overhang allowing direct ligation of PCR products using T4 ligase. The INV 1aF' E. coli transformed with these constructs theoretically should yield white colonies when grown on LB agar plus 50 ug/ml kanamycin and 63 spread with 1 mg X-gal, whereas those transformants that have self-ligated vectors should yield blue colonies denoting a- complementation. Some transformants with recombinant DNA were initially identified by this blue/white screening method. However, additional transformants with recombinant plasmids were discovered in subsequent screenings that still had either functional or partially functional B-galactosidases that in the presence of X-gal yielded deep blue and light blue colonies. Streaks of 63 transformants from the 120-mer ligation and 261 from the 250-mer ligation (these included all of the white . transformants as well as numerous blue transformants) were grown on duplicate sets of nitrocellulose filters. One set was probed with the labeled 250 bp PCR products to identify all the transformants containing a 250 bp PCR product, or 120 bp-containing transformants with the same or closely related sequences. The other set was probed with a labeled PCR oligo- nucleotide antisense primer to identify transformants with a 120 bp PCR product, and 250 bp-containing transformants that had significant sequence similarity to the primer. In tandem these two probes filtered out PCR artifacts unrelated to the two PLC-derived PCR products. However, they were not expected to, and in fact did not identify all PLC types present. The 250 bp PCR products hybridized to 244 out of 261 clones from the 250 bp transformation and to 15 out of 63 clones from the 120 bp transformation. The PCR primer hybridized to 26 out of 64 63 clones from the 120 bp transformation and 8 out of 261 clones from the 250 bp transformation. Thus this double screening identified 15 120-mer clones and 8 250-mer clones that possibly contained PLC sequences. These 23 transformants and selected others that had been identified by just one hybridization were grown up in culture and their plasmids were purified by methods of Sambrook et a1 . [1989] . The desire at this point was to check the insert size by restriction cutting and then sequence many potential PLC- containing clones so as to obtain the maximum number of different PLC-like sequences. The plasmids were analyzed, - first by restriction cutting, to determine insert size. However not all of them yielded the expected inserts. Those plasmids which showed the correct-size inserts sequenced well and many revealed PLC-like sequences. A few plasmids had PLC-like inserts plus unknown foreign DNAs that were of neither PCR nor vector origin. Some other clones had PLC-like inserts in an aberrant vector which was missing over 300 bp including half of the polylinker region. This was puzzling at first since no insert could be removed from them using restriction enzymes but it became clear after sequencing that they did indeed have PLC inserts. Still others could not be sequenced at all using the M13 forward (-40) primer leading to suspicion that this part of the vector was missing here as well. Although these clones were not pursued further they could potentially be sequenced using the M13 reverse primer or 65 a PCR primer. Of the 23 clones identified by the double screening, 17 contained PLC-like sequences of 3 different types, 3 were unsequenceable, and 3 had unidentified DNA inserts of non-PCR origin. There were another 12 clones from the 120 bp ligation that were identified as PCR products but did not hybridize with the 250-mer PCR product. Five of them contained PLC-like sequences of two additional types, 3 were unsequenceable, and 4 were unidentified DNAs. Of the 236 250-mer clones that were identified as PCR products but which did not hybridize with the internal oligonucleotide probe, 4 were sequenced and 8 others were analyzed by restriction cutting and they all appeared to be the same 246 bp artifact. This artifact was a PCR product of close to the expected length but had antisense primers at each end and did not resemble the PLC X-region cDNAs. This indicates that a majority of the 250 bp PCR products were non-PLC artifacts, vindicating individual cloning of PCR products to sort them. The 5 different PLC types discovered among the sequenced clones were designated PLCfl-S. They were compared with each other and with those of published PLC sequences for the levels of sequence similarity at both the cDNA level and the deduced amino acid level (see Tables 1A and 18) . From the comparisons it seemed likely that PLC#2 represented a 'y-PLC x-region cDNA, perhaps the same clone that was purified from bovine brain cDNA library by Stahl et al. [1988] . PLC#3 likely represented 66 WW These tables show the percent similarity of both the cDNA and the deduced amino acid sequences amongst the five identified bovine retinal PLC X-regions (Table 1A) and with those of published PLCs [A,C, and D from Suh et al. (1988), B from Stahl et a1. (1988)] (Table 18). All of the comparisons were made excluding the PCR primers. 67 Tahls_1A1_Q2mparison_of_§lones_witb_8asb_9tber PLCfl PLCfZ PLCf3 PLCf4 PLCfl PLCf2 PLC#3 PLCI4 PLCIS PLCfZ PLCf3 EDNA__AA_ EDNA__AA_ 47%, 50% 48%, 50% 52%, 46% .u., _., . ., s A. B. B-PLC from y-PLC from Rat Brain Bovine Brain QDHA__AA_ QDHA__AA_ 57%, 51% 59%, 59% 49%, 62% 100%,100% 90%,100% 52%, 46% 38%, 77% 51%, 54% 64%, 47% 66%, 53% PLCf4 PLC#5 EDNA__AA_ EDNA__AA_ 44%, 54% 53%, 50% 51%, 54% 60%, 54% 67%, 77% 61%, 50% 57%, 54% ’i. :!'0 Bl ’ -_ C. D. G-PLC from 6-PLC from Bovine Adrenal Rat Brain Cortex Tumor EDNA__AA_ EDNA__AA_ 73%, 59% 68%, 59% 57%, 54% 57%, 54% 49%, 50% 48%, 50% 39%, 46% 38%, 46% 63%, 57% 62%, 57% 68 a B-PLC X-region and possibly the bovine analog of’a rat brain fi-PLC characterized by Suh et a1. [1988] with.which it shared 89.9% cDNA and 100% amino acid sequence similarity. It was not as clear into which category the others fell comparing these limited sequences. Non-X-region sequences will probably be required to differentiate them. Eggp§_Labgling,, When the cloned PCR.products were first to be tested as probes on blots they were cut out of the vector with restriction enzymes and labeled by random primer extension. However subsequent attempts at labeling by this method were inconsistent and unsatisfactory. An alternative- labeling method was needed that could consistently attain the desired levels of total counts and specific activity. End- labeling with T4 polynucleotide kinase was attempted but proved no better than random primer-labeling. Calculations showed that even using 6,000 Ci/mmole [7-”P] ATP and getting 100% end-labeling the probes would be below the recommended 10’ dpm/ug [Sambrook et al. 1989] simply because they were too long. Committee member Jerry Dodgson agreed that the length of my probes made them less than ideal for the established methods of random primer and end-labeling, and suggested that labeling by PCR might be the easiest way to bring these probes to a sufficiently high specific activity. This turned out to be sound advice. PCR labeling has consistently yielded probes of a very high specific activity (>2x109 dpm/ug) . W The 5 Cloned PCR 69 products were trimmed from their vectors, labeled by either random primer (initially) or PCR (later) methods, and tested as probes. On Southern blots of EcoRI-cut bovine genomic DNA, each probe hybridized to one or more bands, none of them the same size from probe to probe (see Figure 4). Similar results were seen with probe hybridization to Northern blots of bovine retinal poly A+ RNA (see Figure 5) . The five probes were also tested for cross-hybridization among each other. Slot blots of the unlabeled probes were hybridized to the individual labeled probes. The results showed that there was little or no cross-hybridization between pairs of different probes (see. Figure 6). This implied that in a library screening all of the probes would be required to identify clones representing each PLC even though all of them code for some part of the PLC X-consensus regions. Lg§19_Libzgry_§gz§gning&_.After some initial attempts at library screening failed (in retrospect most likely due to using random primer-labeled probes) I wanted to test the screening protocols to determine if they were being correctly executed and to develop a positive control probe for compar- ison to the PLC probes. The probe chosen was a cloned PCR product developed by Carol Mindock and Cindy Wheeler, corres- ponding to the first exon of bovine rhodopsin [Nathans and Hogness, 1983]. Its attributes for this work were that: 1) it was 260 bp, close in length to some of the PLC probes and could be labeled similarly; and 2) it was expected to identify 70 Iisurg_1i_flouthgrn_Analxasa Bovine genomic DNA was digested with EcoRI and then fractionated on a 1% agarose gel with 7.5 pg DNA/lane. It was transferred to a Nytran filter by standard blotting techniques. The cloned PCR products, PLC#1-5, were individually labeled by PCR and hybridized to filter strips, each containing two lanes of DNA. After washing, the filter strips were fluorographed, revealing the following bands (in bp): PLEtl ELQtz ELEtl ELQtl 2LQ£§ 5,850 >21,000* 3,850 >21,ooo* >21,ooo* 4,450** 6,300** 7,400** 3,850** 1,420**. * Although the size of these bands could not be measured accurately from this fractionation they were clearly not identical bands. ** These bands were faint and may be the result of cross- hybridization of closely related isoforms. 71 ch#1 ch#2 ch§3 ch#4 PLC#5 1.. is 1"“ —21,226bp . ‘1 . — 5,148 hp — 2,027 bp 72 zigur9_§i_ngrthern_aaalzsga Bovine retinal poly A? RNA was fractionated on a formaldehyde 1% agarose gel with 5 pg/lane. It was transferred to a Nytran filter by standard blotting techniques. The cloned PCR products, PLCf1-3 and #5 were individually by the method of random primers. (PLCf4 was not included in this analysis due to difficulties associated with its aberrant vector.) The others were hybridized to filter strips each containing a lane of RNA. After washing, the strips were fluorographed, revealing the following bands (in bases): BLQfl. H.512. £18513. M 5,800 6,100 7,250 8,030 4,390 2,620 6,460 3,630 5,200 3,130. 73 um chh PLCSZ PLCS3 PLCSS D U o O was — 4, 712 Bases ‘3‘ — 2, 904 Bases — 1, 869 Bases — 1, 541 Bases 74 The cloned PCR products, PLC#1-5, were each slot-blotted five times onto a Nytran filter in equimolar amounts (0.6 pmoles). Also blotted were herring sperm DNA and A DNA in 6 pg aliquots. Additional amounts of PLC#1-5 were individually labeled by the method of random primers and hybridized to strips of the filter containing one blot each of the DNAs. After washing, the filter strips were autoradiographed. As expected, each of the probes hybridized most strongly to itself. PLCfl and #5 cross-hybridized to each other at ~20% of the level that they self-hybridized as determined by densitometric scans of the autoradiogram. Other pairs did not significantly cross-hybridize. 75 Probes chn 91.053 ch#5 pm“ pm“ Fish Sperm DNA ‘ A-DNA i plan ' n... 3133“ chis PLC#1 PLCS4 76 the most abundant clone type in the library, as rhodopsin is the most abundant protein in the retina. Triplicate plaque lifts of 10’ pfus were hybridized and washed by standard methods [Sambrook et a1 . 1989] . Fluorograms revealed 10 plaques that hybridized to the rhodopsin probe at varying intensities. Next the library was screened with the five cloned PLC probes applied simultaneously. Triplicate plaque lifts of 7.5x10’ pfus were hybridized with the probes and the resulting fluorograms revealed 10 plaques that hybridized to the PLC probes to varying degrees. The plaques were subjected to“ standard plaque purification protocols (see Methods) where eight of the original ten hybridized again to the five pooled probes. For the other two, upon purification attempts no hybridizing plaques could be found and they were presumed to be artifacts of the first hybridization and were not pursued any further. WW Thus far the clones have been characterized by which probe they hybridize to, determining the size of their inserts, and sequencing. When the clones were first plaque-purified, samples of the purified stocks were applied to separate sectors of a single plate with a Y1090 E. coli lawn and incubated until clear areas were visible. Then five plaque lifts were taken from the plate and separately hybridized with one of the five PLC probes. The fluorograms indicated that PLCf4 identified six of the eight 77 clones and PLCfl and PLCf3 each identified a single clone. These results suggest that the six PLC#4-hybridizing clones come from transcripts of a single gene or multiple closely related genes and that it/they are likely transcribed at a higher level than those genes represented by the other clones and/or’ probes. The fact that neither PLC#2 nor PLC#5 hybridized to any plaques in this particular library screening indicates only that they may be transcribed at a lower level than those of other PLC. Because there were PCR products generated.it.can be assumed that they were transcribed at some level. Presumably clones that hybridize to these probes could . be identified by a more extensive screening of the Agth or an alternative library. To determine the size of the inserts of the clones the initial plan was to produce large phage stocks, isolate the phage DNA, cut the DNA with EcoRI, and separate it on an agarose gel. When this was carried out using four of the clones, 1-1, 3-2, 4-1, and 4-2 it was found that only 3-2 and 4-2 actually yielded any insert fragments, of approximately 900 and 1,350 bp respectively. The other two clones had high MW A-DNA but no discernable EcoRI insert fragments. This problem had been encountered previously by Steve Pittler when he was isolating PDE clones from this same library. At that time Ron Davis offered that when the library was constructed there may have been some BooRI* activity which gave rise to some erroneous EcoRI restriction sites in the recombinant DNA. 78 Such sites were confirmed later upon sequencing (see below). In order to rescue these clones a plan was devised to use PCR to amplify the insert cDNA, with primers specific for the Agth vector which flank the EcoRI insertion site. It was quickly discovered that this was a much quicker and easier way to obtain large quantities of very pure insert DNA. Hence it was employed for all the clones. The resulting PCR products were sized by agarose gel electrophoresis taking into account that they contained the inserts plus 81 bp of vector DNA, The clones are identified in Table 2. Six clones hybridized to PLCf4. Three of these had inserts of 1000 bp, two had inserts. of 1900 bp, and one had an insert of 1400 bp. It was suspected that the inserts of identical length were identical clones. This was confirmed later by sequencing (see below). Comparing the length of published PLC cDNAs and.the inserts it was considered unlikely that any of the clones represented a full length open reading frame. The PCR-amplified clone inserts were also used for DNA sequencing. This DNA was purer and yielded more readable sequence than the phage DNA isolated by traditional methods. To sequence the DNA a relatively new technique called cycle sequencing was employed (dsDNA Cycle Sequencing Kit, BRL) . It uses multiple cycles of labeled primer annealing and elongation with Taq polymerase in the presence of dideoxy- nucleotides to yield a readable sequence. There are some advantages of cycle sequencing over traditional dideoxy- . . my 79 We: Eight clones were identified by the screening of a bovine retinal Agt10 cDNA library provided to us by Jeremy Nathans [Nathans & Hogness, 1983] with PCR-labeled PLCfl-S together. Then each of the plaque lifts of the purified.phage was tested with each of the probes independently to determine which probe hybridized to it. 'The inserts of the clones were amplified by PCR using Agt10 forward and reverse primers and then sized on a 1% agarose gel, These PCR.products included 81 bp of vector DNA that was also amplified using these primers. 4-1 4-2 7-1 11-1 80 1,900 1,000 800 1,400 1,000 1,900 1,800 1,000 Probe that Win—hp). Wit PLCf4 PLCf4 PLCfl PLCf4 PLCf4 PLCf4 PLCf3 PLCI4. - “an" 81 sequencing. One is that template secondary structure is less of a problem because of the elevated temperatures at which Tag polymerase operates. Another is that less template is needed overall because the cycling yields a linear amplification of dideoxy-termination products (as compared to the logarithmic amplification of PCR). Using this technique sequences over 500 bp have been routinely read. To date all 8 of the amplified clone inserts have been partially or fully sequenced. All of these clones were sequenced using the Agt10 forward and reverse primers originally used to amplify the inserts. However, clones 4-1. and 13-1 could also be sequenced using the PLC sense and antisense primers originally used to create the PCR probes. The other 6 clones, which all hybridized to PLC#4 and are closely related, could not be sequenced with either of those 2 PCR primers. Clone 3-2, which was fully sequenced using the Agt10 forward and reverse primers, revealed a lower sequence similarity to the PLC primers than clones 4-1 and 13-1 (see Figure 7). As suspected, clones 3-2, 7-1, and 15-1, which all hybridized to PLCf4 and had approximately 1000 bp inserts (see Table 2), were identical. Clones 1-1 and 11-1 have 1900 bp inserts and may also be identical although they have not yet been fully sequenced. It was confirmed that all six of the clones that hybridized to PLCf4 coded for very similar PLCs. In fact through the regions that have been sequenced and 82 WW Two of the PCR.primers, originally used to create PLC#1- 5, are shown at the top (shown is the inverse complement of the antisense primer, see Figure 1B) . In those original PCRs the primers were able to anneal to the corresponding sequences below them and amplify the intervening sequences. When the same primers were tried for thermal cycle sequencing clones 4- 1 and 13-1 could be read, however, neither primer yielded detectable sequence for the identical clones 3-2, 7-1, or 15- 1, most likely due to the lower degree of sequence similarity. Clones 1-1, 4-2, and 11-1, which are closely related to clones 3-2, 7-1, and 15-1, also could not be sequenced with the PCR primers. 83 11m W BMW: 6 c c c a m TCGTCCTCGCACAACACGTA GAGTTGGACTGCTGGGACGG 3-2, 7-1, 15-1 --AGGTATCTG---A ----- --A--T-----T-----T-- 4-1 -------- A ---------------------------- c-- 84 overlap, they are identical except that clones 3-2, 7-1, and 15-1 (1000 bp) contain a 36 bp insert missing from.the 1400 bp and two 1900 bp clones“ This suggests that these 6 clones may be members of a subfamily of PLCs. To simplify the subsequent comparisons these 6 sequences were put together in a prelim- inary composite. The deduced amino acid sequences of the composite and of the other two clones, 4-1 and 13-1 were compared with the deduced amino acid sequences of selected published PLCs that contain X-regions and with each other. To follow the compar- isons, refer to Figure 8 and Table 3 which follows it. Clone 4-1 was the only one that did.not have a 100% amino acid sequence match with the probe that hybridized to it (97% with PLCf1). It also did not match predominantly with any of the PLC families, ranging between 38 and 54 overall sequence identity. This clone was, however, relatively short and sequence from overlapping clones may firmly place it in one of the established classes. Alternatively it may represent a PLC class that has not yet been characterized. Clone 13-1, like the PCR probe (PLCf3) that hybridized to it, was almost identical tera rat brain B-PLC cDNA isolated by Suh et al. [1988] with 97% identity of the deduced amino acid sequences. This clone potentially represents the bovine analog of that enzyme that might be found in both the brain and the retina or a closely related brain-like PLC expressed only in the retina. It had much less sequence similarity to 85 WW Schematic linear representations of the deduced amino acid sequences of the Agt10 clones and four published PLCs are presented. Dashed lines represent regions that have not yet been sequenced. Heavy boxes represent the PLC X and Y consensus regions. Larger vertical lines represent the ends of the ORFs or the clone inserts. Smaller vertical lines within the clones represent the limits of known sequence or the presumed X and Y regions (here denoted x and y). The line marked "Composite" is the compilation of six overlapping clone sequences. Where they overlap they are identical except for a 12-base insert in three of the clones, marked by an arrow. All of the sequences are aligned relative to their greatest sequence overlap within the PLC X consensus region. For the sake of clarity gaps needed to maximize the overlap in the actual sequences are not shown. At the bottom are shown the approximate extent of the designated regions compared in Table 3. 86 .51 m _ as u _ H _ " panama—00 33.30% %H .r x n “ $913.32, HT... . 1 1 Jami-melts s x 1. one «33 629m one“ A a I x T “ Quench—.50 " N u? n x c.l I." “ Him." 23.3" N + 1. a llllll I. are 688 u 87 MW The deduced.amino acid.sequences of the Agt10 clones were compared with those of selected published PLCs and with each other. To better illustrate PLC family resemblance, the sequences were divided into numbered regions defined by the limits of the X and Y consensus regions and identified in the bottom line of Figure 8. Percentages in the table represent sequence identity. Comparisons without any significant sequence identity were denoted with two dashes (--) . The published sequences used are Drosophila norpA [Bloomquist et al. 1988], rat brain B-PLC and J-PLC [Suh et al. 1988], and bovine brain 7-PLC [Stahl et a1. 1988]. 88 12212.11 8egi9ns_Q9mpared_uith_§9mposite_9f_ - 3- - 7- - a — Seguenselfiesion§11,1 2_lz) 3 4 (Y) Overall norpA 46 72 37 62 61 fi-ch 3o 58 -- 42 35 y-ch -- 45 -- 28 20 s-ch -- 44 -- 15 17 Clone 13-1 49 58 -- 42 43 Clone 4-1 -- 47 NA NA 32 Begi9na_compared_with_slone_13:1 B-ch 97 97 100 92 97 norpA 44 58 -- 4o 42 s-ch 25 50 -- 25 32 7-PLC 13 50 -- 35 29 Clone 4-1 19 48 NA NA 36 8esi9ns_Qompared_witb_§lone_4:1 6-PLC 41: 62 NA NA 54 B-ch 28 52 NA NA 42 norpA 24 50 NA NA 40 y-ch 12 55 NA NA 38 89 norpA, a fl-PLC from Drosophila [Bloomquist et al. 1988], and even less to mammalian 7- and S-PLCs (see Table 3). The six related clones that hybridized to PLCf4, however, were the most intriguing. Their greatest deduced amino acid sequence similarity was to the norpA B-PLC, 61% overall. Comparisons with published mammalian PLCs revealed signifi- cantly lower sequence identities (see Table 3). Although the match is not as close as that between clone 13-1 and the rat brain B-PLC, the match is provocative considering the evolu- tionary distance between cattle and Drosophila. ‘ 1" mm W The X and Y regions of four pub- lished PLC cDNA sequences were compared to locate consensus regions to which PCR primers could be fashioned. It was anticipated that properly chosen primers would be able to amplify multiple retinal PLC sequences. Among the three X and Y containing classes 8, 7, and 6, the published X-regions had the most extensive sequence similarity. The initial three PCR. primer pairs were complementary to 1) the bovine brain y-PLC X-region cDNA [Stahl et al. 1988], 2) the rat brain B-PLC X-region cDNA [Suh et al. 1988], and 3) the rat brain tumor 6-PLC X-region cDNA [Suh et al. 1988]. They corresponded to homologous positions in the three PLC sequences and they bracketed the majority of the X-region of those sequences. The design of the primers was similar to those of Lee et a1. [1988]. They were each 23 bases long with 15 bases at the 3’ end that were complementary to the published cDNA sequences (Figure. 1A) . They also had an additional 8 bases including 6- base restriction sites to facilitate cloning of the subsequent PCR products and 2 bases at the 5' end of each primer which were added as a ”clamp" for the restriction sites because their respective enzymes cut inefficiently when the 90 91 recognition sequences are at the end of the DNA [Scharf, 1990]. In addition, those 2 bases were complementary to the cDNAs in hopes of increasing the specificity of the primers. The second set of PCR primers was designed differently for several reasons. First they corresponded to regions between the initial primer pairs so they could function as internal probes of the initial PCR.products. Second, because the first set of primers had only limited success in PCR, the new primers were complementary to areas of greater sequence similarity between.the published cDNAs, hopefully to increase their specificity for the template cDNAs. Also to increase- specificity, the primers were complementary for their full length, 20 bases (see Figure 18). The restriction sites were eliminated because of the possibility that the noncomple- mentary regions destabilized the initial primer-cDNA hybrid. In addition, a new product, the TA Cloning Kit (Invitrogen), was available for direct cloning of unmodified PCR products (see below) eliminating the need for restriction sites in the primers. 2g1ymg;age_§haig_3eagtign, As described in the results section the initial set of three primer pairs had only limited success, low expected product yield and numerous artifacts, whereas the second set of primers yielded much more satisfying results. These results could have resulted from a low specificity of the initial primers for the cDNA templates. Having only a 15-base complementary region and then a 92 noncomplementary restriction site could limit the formation of the primer-cDNA hybrid during the PCR annealing step. Having 20 bases of complementary DNA without noncomplementary regions increased the specificity enough to enhance the PCRs. W The 120 and 250 bp PCR products were cloned as reported in Results. The system utilized (TA Cloning kit, Invitrogen) employs the established technique of blue/white screening where blue colonies denote the presence of and white colonies denote the lack of a-complementation of the lacz gene. Theoretically white colonies contain an insert in the lacz gene and blue colonies do not. This, however, was- not always the case as some dark blue transformants were found to have recombinant DNA. In addition some of the white colonies when transferred and grown in streaks displayed varying shades of blue. It was anticipated that these results might occur. A likely explanation for this is that some inserts do not significantly disrupt the lacz gene. This could happen if the insert did not cause a shift in the reading frame and did not affect the overall structure of the resulting B-galactosidase. W In the invertebrate photoreceptor a G protein-linked PLC is light activated and plays a central role in the visual cascade [Rayer et al. 1991] . The norpA mutant Drosophila has a defective photoreceptor PLC (the product of the norpA gene) and as a result is blind [Bloomquist et al. 1988]. The fact I:llu..'l".1'ua-L. , 93 that.a bovine retinal PLC is more closely related to the.norpA PLC than it is to other mammalian PLCs is quite suggestive. Considering that this particular PLC has been conserved between Drosophila and cattle it follows that some of its functions and regulatory mechanisms may have been conserved as well. Almost certainly, distinct differences will be revealed upon detailed examination, but the similarities between invertebrate and vertebrate visual phototransduction may be greater than previously recognized, and a unified theory may in fact center around the role of PLC in both systems. There are two important implications suggested by the' relationship of this bovine retinal PLC to norpA. First is that the bovine retinal norpA-like PLC is very likely of photoreceptor origin as opposed to coming from other retinal cells. Drosophila norpA has been localized to the fly's photoreceptor cells [Bloomquist et al. 1988]. Bovine retinal cDNA libraries were expected to be biased toward photoreceptor clones due to the higher levels of gene expression than in other cell types [Hall at al. 1969]. This means that most, if not all the clones identified were expected to be represented in photoreceptors, and those that are represented by a greater number of clones are more likely expressed in photoreceptors. The second implication is that like norpA this enzyme is quite possibly light-activated via a G protein. There are numerous reports of light-activation of various amphibian photoreceptor PLCs [Ghalayini and Anderson, 1984; Hayashi and 94 Amakawa, 1985; Waloga and Anderson, 1985; Brown et al. 1987]. However, in bovine ROS such activation has not yet been esta- blished likely because of the numerous PLC isoforms present [Gehm.and.MCConnell, 1990b]. The norpA-like enzyme in bovine photoreceptors may be the PLC that is analogous to the lower vertebrate PLCs. These implications assure that in future research the norpA-like bovine retinal PLC will be pursued most aggressively. 2Igpggalg_fig;_£n§gzg_3gggargh, The future of this line of research lies in expression of PLCs from full length cDNAs- and the verification of their comparability with native photoreceptor PLCs. It is very much intertwined with the current work of Rich Pinke, purifying multiple ROS PLC isoforms. Ultimately the aims are to verify the number, primary structure, subcellular location, and expression levels of native PLCs in bovine photoreceptors; to determine the regulatory mechanisms of each PLC in vitro, and its possible role in vision. Since none of the present cDNA clones contains a full length ORF the Agt10 library will be rescreened and a bovine retinal Agt11 cDNA library (Clonetech) will be screened with labeled restriction fragments from these clones. DNA from positive plaques will be subjected to restriction analysis to establish apparent overlap with the present clones, and then sequenced to confirm the overlap. 95 If clones containing full length ORFs cannot be found in the libraries then perhaps composites can be constructed by PCR using retinal sscDNA or by ligation of restriction fragments of shorter clones. Another alternative may be in constructing an expression library using cDNA derived from photoreceptor-enriched poly A+ RNA. This is obtainable by isolating RNA not from whole retinas but from the first supernatant of ROS preparation (see PDE Methods). This material has been shown to be enriched in polysomes from the inner segment of rod cells [McConnell et a1. 1969]. Full length ORFs will be linked to strong promoters and- put into an expression system. It is likely that no single expression system will work equally well with all ORFs and several may have to be tried. But with persistence one that yields the proper post-translational modifications and/or protein folding will be identified for each ORF. The expressed products will be assayed for PLC activity directly in the crude homogenates of cultured cells and again after purification [Gehm and McConnell, 1990b]. They will be assayed more specifically to determine if they share char- acteristics of native ROS PLCs such as activation by W-7 [Gehm et a1. 1991] and arrestin [Ghalayini and Anderson, 1992]. Also limited peptide sequencing will be done to insure that the proper cDNA is expressed, and apparent MW of expression products, determined by denaturing gel electrophoresis, will be compared with MW of their presumed native counterparts. 96 Next polyclonal antibodies will be raised against the recombinant protein and/or unique peptides from the protein. These will first be applied to dot blots and Western blots of native and recombinant ROS PLCs to determine their cross- reactivity. Those that appear to be specific for a single species can then be used to localize the enzyme by immuno- cytochemistry. When correspondence between the recombinant PLCs and their native counterparts has been firmly established, regulatory mechanisms of each PLC can be explored by in vitro reconstitution experiments. Such experiments will initially. involve, co-reconstituting the PLCs in phospholipid vesicles with a variety of potential regulatory/regulated proteins such as rhodopsin, G,or other G proteins, arrestin, cGMP PDE, and PKC. Ultimately, understanding of regulatory mechanisms will demand site-directed mutagenesis studies, which are probably several years away in the case of bovine retinal PLCs. APPENDIX 97 The codon usage for 8 genes specifically expressed in the bovine rod photoreceptor cells were compiled from published cDNA sequences. The number of times each codon was seen in the 8 sequences was totaled with a grand total of 2,511 codons. The frequency, expressed as a percentage, of each codon in representing its corresponding amino acid was then calculated. The 8 cDNA sequences were: 1) Ga [Yatsunami et al. 1985a]; 2) Ge [Sugimoto et al. 1985]; 3) G,y [Yatsunami et al. 1985b]; 4) P, [Ovchinnikov et al. 1987]; 5) P,, [Ovchinnikov et al. 1986]; 6) Rhodopsin [Nathans and Hogness, 1983]; 7) Arrestin [Wistow et al. 1986]; and 8) Peripherin [Connell and Molday, 1990]. 98 m AA____§odon_;Totalt___iEregi_ __AA____§odon__Totalt___1Eresr Arg CGT 11 10.19 Ile ATT 34 23.78 CGC 32 29.63 ATC 103 72.03 CGA 7 6.48 ATA 6 4.20 coo 24 22.22 AGA 12 11.11 Lys AAA 51 31.68 A66 22 20.37 AAG 110 68.32 Leu CTT 20 8.89 Asn AAT 31 27.19 CTC 56 24.89 AAC 83 72.81 CTA 9 4.00 CTG 116 51.56 Gln CAA 18 18.00 TTA 5 2.22 CAG 82 82.00 TTG 19 8.44 His CAT 13 22.81 Ser TCT 22 16.30 CAC 44 77.19 TCC 37 27.41 TCA 15 11.11 Glu GAA 59 31.55 TCG 12 8.89 GAG 128 68.45’ AGT 13 9.63 AGC 36 26.67 Asp GAT 46 31.72 GAC 99 68.28 Thr ACT 22 14.67 ACC 71 47.33 Tyr TAT 21 23.86 ACA 27 18.00 TAC 67 76.14 ACG 30 20.00 Cys TGT 19 31.67 pro CCT 19 18.81 TGC 41 68.33 CCC 45 44.55 CCA 17 16.83 Phe TTT 40 31.75 CCG 20 19.80 TTC 86 68.25 Ala GCT 33 19.64 Met ATG 91 100.00 GCC 90 53.57 GCA 21 12.50 Trp TGG 34 100.00 GCG 24 14.29 Gly GGT 18 12.24 GGC 55 37.41 GGA 27 18.37 GGG 47 31.97 Val GTT 11 6.43 GTC 59 34.50 GTA 6 3.51 GTG 95 55.56 BIBLIOGRAPHY Atkinson, T. and Smith, M. (1984) in Oligonucleotide synthesis: a practical approach (Gait, M. J ., ed) pp. 35- 81, IRL Press, Ltd., Oxford, England. Ausubel F. M., Brent.R., Kingston R. E., Moore D. D., Smith J. A., Seidman J. G. and Struhl K. (1987) carrent Protocols in Moleculapriology, Wiley, NY. 8aehr,‘W. and Applebury, M. L. (1986) Trends NeuroSci. 9, 198- 203. Baehr, W., Devlin, M. J. and.Applebury, M. L. (1979) J; Biol. Chem. 254, 11669-11677. ‘ Berridge, M. J. and Irvine, R. F. (1989) Nature 341, 197-205. Bloomquist, B. T., Shortridge, R. D., Schneuwly, S., Perdew, M., Montell, C., Steller, H., Rubin, G. and Pak, W. L. (1988) Call 54, 723-733. Brown, J. E., Blazynski, C. and Cohen, A. I. (1987) Biochem. Diaphys. Res. Comm. 146, 1392-1396. Cervetto, L., lagnado, L., Perry, R. J., Robinson, D. W. and McNaughton, P. A. (1989 Nature 337, 740-743. Chirgwin, J. M., Przybyla, A, E., MacDonald, R, J. and.Rutter, W. J. (1979) Biochemistry 18, 5294-5299. Connell, G. J. and Molday, R. S. (1990) Biochemistry 29, 4691-4698. Devenny, J. J. and Clack, J. W. (1991) Biochem. Biophys. Res. Cbmm. 176, 70-78. Deterre, P., Bigay, J., Forquet, F., Robert, M., and Chabre, M. (1988) Proc. Natl. Acad. Sci. USA 85, 2424-2428. Dizhoor, A. M., Ray, S., Kumar, S., Niemi, G., Spencer, M., Brolley, D., Walsh, K., Philipov, P. P., Hurley, J. 8. and Stryer, L. (1991) Science 251, 915. 99 100 Dratz, E. A. and Hargrave, P. A. (1983) Trends Biochem. Sci. 0, 128-131. Fourney, R. M., Mayakoshi, J., Day, R. S. III, Paterson, M. C. (1988) Fbcus 10, 5-7. Gehm, 8. D. and McConnell, D. G. (1990a) Biochemistry 29, 5442-5446. Gehm, B. D. and McConnell, D. G. (1990b) Biochemistry 29, 5447-5452. Gehm, 8. D., Pinke, R. M., Lauqerre, S., Chafouleas, J. G., Schultz, D. A., Pepperl, D. A. and McConnell, D. G. (1991) Biochemistry 30, 11303-11306. Ghalayini, A. J. and Anderson R. E. (1984) Biochem. Biophys. Res. Cbmm. 124, 503-506. Ghalayini, A. J. and Anderson R. E. (1992) Biophys. J. 61, Abstr. 2117. - Glazer, A. N., DeLange, R. J. and Sigman, D. S. (1975) "Chemical Modification of Proteins: Selected Methods and Analytical Procedures" in Laboratory Techniques in Biochemistry and Molecular.8iology, eds. Work, T. S. and Work, E., North-Holland Publishing Co. Amsterdam. Gordon, A. H. (1975) "Electrophoresis of Proteins in Polyacrylamide and Starch.Gels” in Laboratory Techniques in Biochemistry and Malecular Biology, eds. Work, T. S. and Work, E., North-Holland Publishing Co. Amsterdam. Gray-Keller, M., Polans, A., Palczewski, K. and Detwiler, P. (1993) Biophys. J. 64, Abstr. M-PM-L2 (p. 19). Gross, E. (1967) Methods in Ehzymology 11, 238-255. Hall, M. O., Bok, D. and Bacharach, A. D. E. (1969) J. Mol. Biol. 45, 397-406. Hancock, J. F., Magee, T., Childs, J. E., Marshall, C. J. (1989) C811 57, 1167-1177. Hayashi, F. and Amakawa, T. (1985) Biochem. Biophys. Res. 00mm. 128, 954-959. Hurley, J. 8. and Stryer, L. (1982) J3 Biol. Chem. 257, 11094- 11099. Hurwitz, R. L., Bunt-Milam, A. H., Chang, M. L. and Beavo, J. A. (1985) J. Biol. Chem. 260, 568-573. 101 Kawamura, S. and Murakami, M. (1991) Nature 349, 420-423. Keil, P. (1971) The Enzymes 3, ed. Boyer, P. D., Academic Press, New York. Koch, K. W. (1992) Trends Biochem. Sci. 17, 307-311. Kohnken, R. E., Eadie, D. M., Revzin, A. and.McConnell, D. G. (19818) J. Biol. Chem. 256, 12502-12509. Kohnken, R. E., Eadie, D. M. and McConnell, (1981b) J. Biol. Chem. 256, 12510-12516. Kfihn, H. (1982) Methods in Ehzymology 81, 556-564. Lamb, T. D. (1986) Trends NeuroSci. 9, 224-228. Lee, C. C. and Caskey, C. T. (1990) in.PCR.Protocols: A Guide to Methods and Applications (Innis, M. A., Gelfand, D. H., Sninsky, J. J. and White, T. J., eds.) pp. 84-91, Academic Press, Inc., San Diego. - Lee, C. C., Xiangwei, W., Gibbs, R. A., Cook, R, G., Muzny, D. M. and Caskey, C. T. (1988) Science 239, 1288-1291. Lee, K. Y., Ryu, S. H., Suh, P. G., Choi, W. C. and Rhee, S. G. (1987) PTOC. Notl. Acad. Sci. USA 84, 5540-5554. Li, T., Volpp, K. and Applebury, M. L. (1990) Proc. Natl. Acad. Sci. USA 87, 293-297. Lipkin, V. M., Kharmtsov, N. V., Vasilevskaya, I. A., Atabekova, N. V., Muradov, K. G., Gubanov, V. V., Li, T., Johnston, J. P., Volpp, K. J. and Applebury, M. L. (1990) J. Biol. Chem. 265, 12955-12959. Lowry, O. H., Rosebrough, N. J., Farr, A. L. and Randall, R. J. (1951) J. Biol. Chem. 193, 265-275. Majerus, P. W., Ross, T. S., Cunningham, T. W., Caldwell, K. K., Jefferson, A. B. and Bansal V. S. (1990) Cell 63, 459-465. Matthews, H. R., Murphy, R. L.‘W., Fain, G. L. and Lamb, T. D. (1988) Nature 334, 67-69. Nathans, J. and Hogness, D. S. (1983) cell 34, 807-814. Newgard, C. 8., Nakano, K., Hwang, P. K. and Fletterick, R. J. (1986) Proc. Natl. Acad. Sci. USA 83, 8132-8136. 102 Ovchinnikov, Yu. .A., Gubanov, V. ‘V., Khramtsov, N. ‘V., Ischenko, K. A., Zagranichny, V. E., Muradov, K. G., Shuvaeva, T. M. and Lipkin, V. M. (1987) FEBS.Lett. 223, 169-173. Ovchinnikov, Yu. A., Lipkin, V. M., Kumarev, V. P., Gubanov, V. V., Khramtsov, N. V., Akhmedov, N. V., Zagranichny, V. E. and Muradov, K. G. (1986) FEBS Lett. 204, 288-292. Pittler, S. J., Baehr, W., Wasmuth, J. J., McConnell, D. G., Champagne, M. S., vanTuinen, P., Ledbetter, D. and Davis, R. L. (1990) Genomics 6, 272-283. Pober, J. S. and Bitensky, M. W. (1979) Adv. Cyclic Nucleotide Res. 11, 265-301. Rayer, 8., Naynert, M. and Stieve, H. (1991) J. Photochem. Photobiol. 7, 107. Rhee S. G., Suh, P. G., Ryu, S. H. and Lee, S. Y. (1989) Science 244, 546-550. - Rodieck, R. W. (1973) The vertebrate Retina: Principles of Structure and FUnction Freeman 6 Co., San Francisco. Sambrook, J. Fritsch, E. and Maniatis, T. (1989) Molecular Cloning, A Laboratory Manual (2d ed.) , Cold Spring Harbor Laboratory Press. Scharf, S. J. (1990) in.PCR.Protocols: A.Guide to.Methods and Applications (Innis, M. A., Gelfand, D. H., Sninsky, J. J. and White, T. J., eds.) pp. 84-91, Academic Press, Inc., San Diego. Sonnenburg, W. K., Mullaney, P. J. and Beavo, J. A. (1991) J. Biol. Chem. 266, 17655-17661. Stahl, M. L., Ferenz, C. R., Kelleher, K. L., Kriz, R. W. and Knopf, J. L. (1988) Nature 332, 269-272. Stryer, L. (1986) Ann. Rev. Neurosci. 9, 87-119. Sugimoto, K., Nukada, T., Tanabe, T., Takahashi, H., Noda, M., Minamino, N., Kangawa, K., Matsuo, H., Hirose, T., Inayama, S. and Numa, S. (1985) FEBS.Lett. 191, 235-240. Suh, P. G., Ryu, S. H., Moon, R. H., Suh, H. W. and Rhee, S. G. (1988) Cell 54, 161-169. Waloga, G. and Anderson, R. E. (1985) Biochem. Biophys. Res. comm. 126, 59-62. 103 Winberg, G. and Hammarskjold, M. L. (1980) Nucleic Acid Res. 8, 295-264. Wistow, G. S., Katial, A., Craft, C. and Shinohara, T. (1986) FEBS Lett. 196, 23-28. Wood, W. I. (1987) Methods in Ehzymology 152, 443-447. Wray, W., Boulikas, T., Wray, V. P. and Hancock, R. (1981) Anal. Biochem. 118, 197-203. Yatsunami, K. and Khorana, H. G. (1985a) Proc. Natl. Acad. Sci. USA 82, 4316-4320. Yatsunami, K., Pandya, B. V., Oprian, D. D. and Khorana, H. G. (1985b) Proc. Natl. Acad. Sci. USA 82, 1936-1940. MICHIGAN STATE UNIV. LIBRARIES ll”WWIWWWWWIWWW“Will 31293008977609