Journal of Chemical Neuroanatomy, Vol. 5:289-298 (1992)

Antipeptide Antibodies against the 5-HTIAReceptor Efrain C. Azmitia*t, llje Y u t , Homayoon M . Akbar# ~, Nancy Kheck*, Patricia M . Whitaker-Azmitia~ and Daniel R. M a r s h a k t *Department of Biology, New York University, New York, NY 100003, USA tCold Spring Harbor Laboratory, Cold Spring Harbor, NY 11794, USA ~Department of Psychiatry, State University N-Y, Stony Brook, NY 11794, USA ABSTRACT The availability of the primary amino acid sequence for a large numbers of molecules provides a fruitful opportunity for their cellular localization by utilizing the procedure of antipeptide antibody formation. This procedure permits a synthetic peptide sequence to be attached to a carrier molecule for the purpose of inoculating an animal to raise specific antibodies against the selected protein sequence. In this report we describe a number of steps that can be taken to increase the likelihood that the selected peptide sequence will be specific and antigenic. In addition, we describe how the peptides are synthesized, purified and coupled to keyhold limpet hemocyanin. The preparation of the antibody and its characterization are also presented in this method report. The immunocytochemical staining at both the light and ultrastructural level with serotonin (5-HTIA) receptor antipeptide antibodies is discussed. The advantages and disadvantages of this procedure are summarized. KEYWORDS; Serotonin Immunocytochemistry Ultrastructure INTRODUCTION The serotonin (5-HTIA) receptor gene from the rat has been cloned and shown to have 1266 nucleotides corresponding to 422 amino acids (Albert et al., 1990). It has seven transmembrane domains, a large third cytoplasmic loop and is 89% homologous with the human gene. Its mRNA tissue distribution showed high levels in septum, hippocampus, thalamus, amygdala olfactory bulb, mesencephalon, medulla and hypothalamus. Detectable levels were seen in cortex and basal ganglia but not in pineal and pituitary (Albert et al., 1990). These results are in general agreement with the [3H]8-OH-DPAT binding studies in the rat which showed widespread distribution of receptor labeling except in extrapyramidal areas (substantia nigra, caudate nucleus and the globus pallidus), cerebellum and habenula, where levels were undetectable (Pazos and Palacios, 1985; Verge et al., 1986). Antibodies directed or raised against synthetic peptides have been used to study many previously uncharacterized proteins (Sutcliffe et al., 1980; Walter et al., 1980; Yu and Marshak, 1991). This approach has the advantages of being applicable as soon as the cDNA sequence is known and that the antipeptide antibody can be directed against a specific short region of the molecule. In general, the short sequence is bound to a larger carrier protein to Address correspondenceto: Dr Efrain Azmitia, Department of Biology, 104)9 Main Building, 100 Washington Square East, New York, NY 10003,USA. 0891-0618/92/040289-10 $10.00 © 1992 by John Wiley and Sons Ltd

increase its antigenicity (Hollow and Lane, 1988). This approach has been applied to the 5-HT~A receptor against a region of the third cytoplasmic loop. Lefkowitz and Caron (Raymond et al., 1989) raised an antibody (JWR21) against the sequence 242-267 and attached it to keyhole limpet hemocyanin (KLH) using the method of Avrameas and Ternynck (1969), and El Mestikawy and Hamon (El Mestikawy et al., 1990) raised their antipeptide antibodies against a very similar sequence 243-268 and bound it to bovine serum albumin by 1-ethyl-3-(3dimethylaminopropyl) carbodiimide. Anatomical immunoautoradiographic study with this latter antibody at 1/1000 dilution showed'a 'strikingly similar' distribution to that seen with [3H]8-OHDPAT binding. In an immunocytochemical study, labeling of 5-HT perikarya and dendrite membranes was seen in the midbrain raphe nuclei (Sotelo et al., 1990). We present in this report a method for selecting two new sites for antibody recognition against the 5HT~A receptor: S1A-170 (aa 170-186) and S1A-258 (aa 258-274) (Fig. 1A). These antibodies labeled a protein band of approximate molecular weight of 49 000. They showed excellent staining in rat neonatal and adult brain and in adult monkey brain at dilutions as high as 1/10 000. Similarities and differences with the distribution of [3H]8-OH-DPAT binding were seen. Most importantly, clear laminar labeling was seen in those areas known to have high 5-HTIA binding such as the hippocampus and cortex. In addition, selective cells were labeled in areas

290

E.C. Azmitia et al.

thought to have little or no 5-HT1A receptors, such as the cerebellum and striatum. Light cellular labeling was apparent in many areas. In the hippocampus, polymorphic interneurons were stained (Fig. I B). There was less intense labeling in the granule cell layer. The neurons in the rostral dorsal raphe nucleus were clearly stained in the monkey. In the cerebellum, labeling was very high in the neonate but mainly confined to glial cells in the adult rat. High label was seen in epithelial cells lining the brain and the ventricular system. Tanocytes were labeled in the third ventricle near the median eminence. Staining was seen in astroglial cells in many brain areas and in astroglial cultures. Ultrastructural studies revealed heavy staining in processes in the hippocampus and midbrain. The label was associated with the smooth endoplasmic reticulum and in patches along the outer plasma membrane (Fig. 2).

SELECTION OF PEPTIDE The selection of the peptide domain of the 5-HT1A receptor against which the antibodies were raised was made on the basis of several criteria.

Functional domain of the receptor The receptor is homologous to the beta-adrenergic receptor family and many of the functions of the various segments of the 5-HT1A receptor can be inferred from extensive work with the [3-adrenergic receptor. The agonist binding site consists of at least two aspartate (asp) molecules within the 2nd and 3rd transmembrane regions (Dohlman et al., 1987). Asp residues (numbers 82 and 116) exist in a similar site in the 5-HTIA receptor (Fig. 1A). A histidine (His) in the 3rd transmembrane site (number 126) would provide the needed positive charge for 5-HT binding. The third cytoplasmic loop is believed to be the site of interaction with the G-proteins in the cytoplasm for regulation of the second messenger systems (Kobilka et a l , 1988). Therefore, peptide sequences can be selected to indicate the anatomical location of various segments of the full molecule and determine the cellular distribution of particular functional regions. We selected domains in the 2nd external loop (S1A-170) and in the 3rd cytoplasmic loop (SlA-258) (Fig. 2).

Hydropathicity regions The antigenic sites on a peptide can be approximated based on the hydropathicity score which assumes that the greater the hydrophilic characteristics, the more antigenic the sequence (Hopp and Wood, 1981). This measure assigns a numerical value to the various amino acids; for example K, R, D and E have a value of + 3.00, W has a value of - 3.4, and G and P have a value of 0. The calculated window

average at a residue is calculated across six residues. The hydropathicity score for S1A170 is shown in Table 1.

Two-dimensional protein structure There are three states in which a sequence can exist in a secondary structure of a protein molecule. These are beta sheets, alpha helix and turns (Chou and Fasman, 1974, 1978). In the latter state, it is assumed that the amino acids are most exposed. The selected peptides both have a significant number of predicted turns (Table 1 shows results for S 1A 170).

Charge balance of the protein The net charge of a peptide sequence should be near neutrality. If the molecule is too highly charged it will present problems during the purification procedure after the peptide is synthesized. If the net charge is highly basic or acidic, a cation or anion exchange resin can be used. Bio-rad AG-50 resin has been successfully used for very basic peptides. However, strong deviations from neutrality is also a problem during the attachment to KLH which should proceed at neutral pH (see below). S 1A-170 has six charged residues and a net - 2 charge while S1A-258 has five charged residues and a net + 1 charge.

Amino acid length The sequence for an ideal peptide for antibody formation should have 15-20 amino acids. A strand of seven amino acids is the lower limit for a recognition site. More than 20 presents some additional problems with the synthesis and structural considerations. Both SIA-170 and S1A-258 have 17 residues.

Phosphorylation and glycosylation sites A protein molecule has many possible phosphorylation and glycosylation sites. These sites should be avoided in choosing a sequence unless a particular confirmation is sought. Antibodies have been raised against phosphorylated sequences but these antibodies have altered affinity for the unphosphorylated site. Furthermore, a phosphorylated segment of the molecule often confers allosteric changes in the protein structure which may reduce the affinity for the peptide segment artificially produced. It can be appreciated that sites adjacent to modified sites may be less desirable for the same reasons. The glycosylation sites are located on the asparagine (ASN, N) residues at positions 10, 11 and 24. Three potential protein kinase C phosphorylation sites are located at 147-152, 227-232 and 341-345 and one additional phosphorylation site at 251-253 (El Mestikawy et al., 1991). Neither S1A-170 nor

Antipeptide Antibodies against the 5-HT1A Receptor

291

S1A-258 have phosphorylation or glycosylation sites,

have little consequence to the antigenicity of the sequence.

Position of cysteine Cysteine (Cys, C) residues are commonly involved in disulfide bridges. For this reason it is advisable to avoid a Cys residue in the middle of a peptide sequence. There are 15 cysteine residues in the 5-HT~A receptor, six in the transmembrane regions, three in the 3rd cytoplasmic loop, four in the three extracellular loops and two in the C-terminal cytoplasmic tail. The cysteines in extracellular loops 1 and 2 have been proposed to form a disulfide link in the 132-adrenergic receptor (Dohlman et al., 1987). Similar cysteines exist in the 5-HT~A receptor (Fig. 1A). In designing the sequence, it is advisable to have a terminal cysteine residue in order to bind to KLH protein (Hollow and Lane, 1988). This can be either at the C- or N-terminus of the peptide, depending on how the peptide is predicted to be exposed in the molecule. For instance, if the desired sequence is at the N-terminal end of the protein, then the cysteine should be placed at the C-terminal end of the peptide. In this way, the N-terminal end will be exposed after its attachment to the carrier protein. Both peptides have an N-terminal Cys but S1A-258 also contains a Cys in the center of the peptide at position 266.

SYNTHETIC PEPTIDES

Homology with known proteins In selecting a region of the 5-HTIA receptor protein, comparisons with the 5HT2, 5 HTlc, a 2 and B2adrenergic, muscarinic M1, and the D E receptor were performed (Julius et al., 1990; Bunzow et al., 1988). Once a segment has been selected it should be compared to all known protein sequences. The sequence data bank searching program (Intelligenetics, Inc., Mt View, CA-415-962-7300) we used is based on the algorithm of Welbur and Lipman. Sequence homology can be searched against the Protein Identification Resource data bank and the Protein Sequence data bank, which contains the translated European Molecular Biology Library. This search will identify those known structures that could react with the antibody raised. This is especially relevant when the protein in question has been identified in the same species chosen for study, and conversely, significant homology to an invertebrate protein is not necessarily a problem. High homology for the selected sequence of the same protein in different species is advantageous. Neither selected sequence has any significant homology with any other published receptor or mammalian protein. Interestingly, S1A-170 has a 94% homology with the human 5-HTIA receptor, while S1A-258 has only 44% homology with the human 5-HTjA. In the case of SIAI70, the methionine ( - 1.3 hydropathicity) at position 172 is replaced by an isoleucine ( - 1 . 8 hydropathicity) which should

The peptides were made at Cold Spring Harbor Laboratory in the Protein Chemistry Core Facility; this has been described in detail previously (Yu and Marshak, 1991). Briefly, the two sequences selected, CSH 228 (S1A-170-86: H-Pro-Pro-Met-Leu-GlyTrp-Arg-Thr-Pro-Glu-Asp-Arg-Ser-Asp-Pro-AspAla-Cys-NH2) and CSH 229 (S 1A-258-274: H-ProGly- Ser-Gly-Asp-Trp-Arg-Arg-Cys-Ala-Glu-AsnArg-Ala-Val-Gly-Cys-NH2), were synthesized by the solid-phase methods (Barany and Merrifield, 1979) on p-methyl-benzylhydrylamine polystyrene resin using pre-formed, symmetric anhydrides and hydroxybenzotriazole-activated esters of N-a-Bocprotected amino acids on an Applied Biosystems Inc. Model 430A automated peptide synthesizer. A modified small-scale (0.1 mmol) rapid cycle was used. Couplings were done in dimethylformamide and dichloromethane as solvents, and unreacted peptide was capped with acetic anhydride. The side chain-protected amino acids were: Arg (Mts); His (BOM); Thr (Bzl); Cys (4-CH3-Bzl); Trp (CHO); Ser (Bzl); Glu (OBzl); and Asp (OcHex). Double coupling was necessary for several amino acids such as Trp, Leu, Thr, Glu, Ser, Cys and Val. The peptides were deprotected and cleaved from the resin with liquid HF at - 1 0 ° C for 2 h in the presence of 5% (v/v) anisole and 5 % (v/v) dimethyl sulfide. The peptide was precipitated with ethyl ether, and solubilized in 6M-guanidine HCI at 10 mg/nl (41 ml). The Trp was formyl chased after cleavage to remove the -CHO protecting group. The sample was cooled to 0°C in a salt ice bath in a round-bottom flask with stirring. Ethanolamine was added at a final concentration of 1 M (2.5 ml) and stirred for 4 h at 0°C. The temperature was critical with the pH > 8 since a low temperature prevents the cyclization of glutamic acid and aspartic acid. The reaction was quenched by reducing the pH to less than 7.0 with HCL (Baker, HPLC grade). The sample was 0.45 Ixm filtered prior to HPLC purification. The solution was subjected to HPLC using a Waters Delta Prep 3000 instrument on a column (4.9 x 30 cm) of 300 A, C18 silica (Waters) and eluted with 0.1% (w/v) trifluoroacetic acid with a linear gradient to 60% aceto-nitrile (Burdick and Jackson). Both peptides (CSH228 and 229) eluted at approximately 23min at approximately 49% acetonitrile. The peptide can be further purified if necessary by HPLC on a column (2.2 x 25 cm) of silica using C18-bonded , 300 /~ pore size silica, 10 ~tm in diameter (Vydac, The Separations Group, Hesperia, CA). We did not find a second pass necessary over the HPLC. The mass spectrometry data showed a perfect [M+H] + ion with no other

0

2:

N

t~

Antipeptide Antibodies against the 5-HT1A Receptor adducts or delection products. Final products were 50 mg of 228 and 26 mg of 229. The structure o f the final peptide can be verified by amino acid analysis, automated sequence, analysis, plasma desorption mass spectrometry, and analytical microbore H P L C as described previously (Marshak and Carroll, 1991).

ANTIBODY PREPARATION The peptides, CSH 228 and 229, were coupled to K L H (Sigma) via maleimidobenzoyl-N-hydroxysuccinimide (Pierce Chemical Co., Rockford, IL) as described (Hollow and Lane, 1988). Briefly, the K L H was dissolved in phosphate-buffered saline (PBS); 10 mg/ml). M-Maleimidobenzoyl-Nhydroxysuccinimide ester (MBS; Pierce, N22312) was dissolved in dimethylformamide (Burdick and Jackson, Baxter, NJ). The MBS solution was very slowly added to the K L H solution, one drop at a time, while stirring, and the mixture allowed to mix for 30 min at room temperature. The unbound MBS was removed by filtration on Sephadex G25 in PBS (Pharmacia). The peptides were dissolved in PBS at 10mg/ml and added slowly to the M B S / K L H solution. The p H was adjusted to 7.2 and the mixture stirred for 3 h at room temperature. The solution was extensively dialyzed against PBS in the cold room overnight. The protein concentration in the dialysate was determined and the solution adjusted to I mg/ml. Antisera against the complexes were raised as follows. The complexes (0.5 mg) in 0.5 ml of PBS were mixed with 0.5ml of Freund's complete adjuvant and 300 ~tl injected subcutaneously in four spots on the back of adult female New Zealand white rabbits weighing 8.5 and 8 lb that were certified specific pathogen free (Hare Marland, Hewitt, N J). Booster injections were given with incomplete adjuvant at 2-week intervals with 200 Ixl complex and 200 ~tl incomplete adjuvant at two spots on the back. Additional booster injections were given at 2-week intervals until maximum serum titer was reached. The first bleed of 6 ml was made from the central artery o f the ear 3 days after the second boost. Once the maximum titer was reached (fourth boost), full bleeds of 15-20 ml were made. The blood was kept at 5°C and spun once at low speed for 10min and the serum spun a second time in Eppendorf tubes for 10 min before being aliquoted and frozen.

293

Serum antibody titer was determined by radioimmunoassay. Wells of 96-well poly-vinylchloride microtiter plates (Falcon Microtest III) were coated with 50 Ixl of the appropriate peptide (1 mg/10 ml) for at least 3 h at room temperature. The plates were washed three times with PBS and unbound sites were saturated with 200 lxl of 3% (w/v) bovine serum albumin. Dilutions of immune and preimmune serum were added to wells at concentrations from 1/50 to 1/50000 for at least 1 h at room temperature. The plates were washed three times with PBS; 50 000cpm of ~25I-labeled goat anti-rabbit IgG F(ab) 2 (NEX167, NEN) were added per well and allowed to mix at room temperature for 1 h. After three washes, the specific activity of each serum sample was determined by counting the radioactivity for 1 min in a gamma radiation counter (Beckman, G a m m a 5500 B). The results for S1A170 and S1A258 are shown in Fig. 3. GEL ELECTROPHORESIS AND IMMUNOBLOTTING Electrophoresis was performed on I mm 12.5% (w/v) polyacrylamide gels in the presence of sodium dodecyl sulfate using the buffer system of Laemmli (1970). 20 and 40 I~g of tissue from the hippocampus were run in each well along with 5 lal of standard (Bio-Rad biotinylated SDS-PAGE standards, low range, catalogue no. 161-0306). The hippocampus was removed from a young Long-Evans female rat (100-150g; Charles Rivers, Kingston, NY) and immediately frozen in liquid nitrogen. A single hippocampus was transferred to lysis buffer (see below) containing 2% SDS. The tissue was immediately homogenized by hand in an Eppendorf tube (pellet pestle with disposable tube, no. 749520, Kontes, N J) and allowed to sit for 15 min on ice before spinning in a microfuge (Eppendorf microcentrifuge 5414) at 5°C for 15 min. The supernatant was collected and assayed for protein using a BioRad protein microassay. Approximately 3 I~1 were added to 1 ml of solution and read on a spectrophotometer at wavelength 595 nm. The solution was adjusted to a final concentration of 2 mg/ml. The supernatant was mixed with an equal amount of sample buffer (Laemmli), 13-mercaptoethanol added to a final concentration of 10% (v/v) and the sample boiled for 5 min. The gel was run for approximately 5 h at 100 V on a vertical gel electrophoresis apparatus. Proteins were transferred electrophoretically to nitrocellulose using a Bio-Rad

Fig. 1. (A) A schematicrepresentationof the 5-HT~Areceptorprimarystructure(modifiedfrom Dohlmanet al., 1987).The locationof the two peptides synthesizedare numbered as 170-187 and 258-273. A proposed binding site for 5-HTis shown between the 2nd and 3rd transmembraneregions.Four putativephosphorylationsitesare designatedas phosphokinase-C(PKC)or merelya generalphosphokinase (PK) target (see E1 Mestikawyet al. (1991)).(B) A photograph of the 5-HT~Aimmunocytochemicalstaining of the hilus of the dentate gyrus of the monkey.Note the intensestaining of interneuronin the hilus (largearrows). Largedendritesextendingfrom thesecellshave spines that are labeled (small arrows). Numerous small cells can also be seen to be labeled and these have been shown to be astrocytes (small arrow with curvedtail). Scalebar is 100 lam.

ii~ ii!¸

,!i!~!~~¸~?!i~

~'

~,~i~!~~

~ i~ ¸~

Antipeptide Antibodies against the 5-HT~A Receptor Trans Blot Cell overnight at 50 V in the cold room (Towbin et al., 1979; Towbin and Kurstak, 1984). The nitrocellulose was incubated with antiserum at dilutions from 1/100 to 1/10000 in 0.1% (v/v) Tween-20 (Sigma) in Tris-buffered (0" 1 M, pH 7.4) saline (0.9%) solution (TTBs). In our experience the best results were obtained with our antibodies at dilutions of 1/1000 to 1/5000. The antibody at lower dilutions (1/250) gave increased background and less sensitivity. The avidin-biotin peroxidase procedure was used to identify the protein band as described by Vector Laboratories. Briefly, the nitrocellulose sheets were cut to include a standard and the appropriate rows and washed in small 75 mm disposable petri dishes. The nitro-cellulose strips were incubated with the antisera for at least 2 h at 40°C (the strips could be left with antisera for several days at room temperature in the cold room). The strips were rinsed three times in TBS for a total of 10 min on a shaker between biotinylated secondary (30-min incubation), the ABC reagents (30-min incubation) and the DAB peroxidase reaction. The strips were first incubated with freshly filtered diaminobenzidine (5 mg/20ml TBS; Sigma) and 0.2% nickel ammonium sulfate for 5 min at room temperature and then H202 was added at a final concentration of 0.003% (v/v). Boehringer-Mannheim produces biotinylated reagents that are comparable with Vector Stains products. One major band (49.5kDa) and two minor bands 42.0 and 37.0kDa) were stained with S1AI70 at a dilution of 1/1000. Lysis buffer (Draeta and Beach, 1988) was made from highest grade chemicals from Sigma (St Louis) and Boehringer-Mannheim: 50 mM-Tris, pH 7.4; 150mM-NaC1; 1% NP-40; 10mM-EDTA; l mMMgC12; 1 mM-CaCI2; 10% glycerol; 400 I~M-sodium orthovanadate; 50 mM-NaFluoride; 50 mg/l phenylmethane sulfonylfluoride from 10mg/ml isopropanol; 1 mg/1 leupeptin; 10 mg/l soybean trypsin inhibitor; 1 mg/1 aprotinin and 10 mg/1 L- 1-chloro-3[4-tosylamido]-4-phenyl-2-butanone from 3 mg/ml ethanol. IMMUNOCYTOCHEMISTRY AT THE L I G H T A N D U L T R A S T R U C T U R A L LEVEL Neonatal (1-2 weeks) and adult female rats were perfused with a variety of fixatives and prepared for immunocytochemistry according to our published procedures (Azmitia and Gannon, 1983). Rats (Sprague-Dawley, female, Taconic Breeders, 220 g)

295

and monkeys (Macacafascicularis, female, Charles River Breeding Laboratory, 3.3 kg) were perfused through the ascending aorta with 4% paraformaldehyde, 2% glutaraldehyde, 4% formaldehyde plus 0.1% glutaraldehyde or 3.25% acrolein with 2% paraformaldehyde at 20°C and 0.1% MgSO 4 in 0.1 M-phosphate buffer (pH7.4) at 20°C. The glutaraldehyde and acrolin fixatives were continued after 10 min with the same solution with only paraformaldehyde (total perfusion volume was 100 ml for neonates, 250 ml for adult rats and 1500 ml in monkeys) for an additional 20 min. The brains were postfixed at 5°C for at least 4 h before being processed for immunocytochemistry. Thirtymicrometer sections of the hippocampus and brainstem were cut on a Vibratome (Oxford). The primary antiserum and the secondary sera were diluted in 0.1 M-TBS (pH 7.4, 0.85%) containing 1% normal sheep serum and 0.1% Triton X100. The sections were incubated for 18-72 h at 5°C followed by 2 h at room temperature in antipeptide antibody serum at a dilution of 1/1000-1/10000. The sections were then processed with the elite Vector stain ABC-kit as directed by the manufacturer. The reaction was run for 2 min at room temperature in 0.05% 3,3-diaminobenzidine containing 0.2% nickel ammonium sulfate in 0.1 MTBS (pH 7.4) (Azmitia and Gannon, 1983) followed for 5-10min at room temperature in the same solution with the addition of 0.003% hydrogen peroxide. The sections for electron microscopy were viewed after postfixing for 1 h at 20°C in 2% osmium tetroxide containing 1.5 % potassium ferricyanide in 0.1 M-phosphate buffer (pH 7.2) and then block stained in 0.5% uranyl acetate at 5°C for 30min. Ultrathin sections were taken from the surface of Epon/Araldite-embedded tissue slices and viewed on the electron microscoI~e without further heavy metal staining. ADVANTAGES Specificity of the antibody is assured because a specific region of the molecule is targeted. Regions with high homologies with other molecules can be avoided. A functional region (3rd cytoplasmic loop) or a structural portion (second extracellular loop) can be selected for study (Fig. 1A). The antibody peptide can be produced as soon as the primary cDNA structure has been demonstrated. High titers with the antipeptide antibody can be obtained because the peptide is bound to a carrier

Fig. 2. An ultrastructuralmicrographof the processesimmunocytochemicallylabeledwith the antibodyraised against peptide 170-187in the midbrain raphe area. (A) The micrograph shows a single tangential process that is heavilylabeled. The label can be seen largely confined to smooth endoplasmic reticulum(SER) occurringboth within the process (arrow head) or in the eisterna (large arrow). In severalareaswithin the processthe labelappears to be localizedto smallsphericalorganelles(smallarrow). Magnification= 8200 ×. (B) A highermagnificationmicrographof a processshowinga singlelabeledprocess.The labelcan be seenwithinthe SER. Labelis also found in patches near the externalplasma membrane(largearrow). Note that small sphericallabeled organellescan be seen within the d cisterna (small arrows). Mitochondriaare seen(m). Magnification= 15500 x.

296

E.C. Azmitia et al. Table 1. The hydropathicity score according to H o p p and W o o d (1981) and the secondary structure prediction by the algorithm o f C h o u and F a s m a n (1974) are shown for the peptide S I A l 7 0 (Intelligenetics program). Notice that a turn structure is predicted for the region showing a high positive hydropathicity value. Position

Structure

170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187

1.6E4,

1.4E4

010\/

- 2

--Beta Beta Beta Beta Beta Turn Turn Turn Turn Turn Turn Turn Turn Turn Turn Turn

- 1

0 O..O_o

1.2E4. 1.0E4. •

4000.0. 2000.0.

,-, P-S1A258

+ 1

+2

**********p **********p ************M ***********L ********G *******W R T******* P****************** ************************* ************************** ***************************** ***************************** ************************** P**************** D********** A***** *C

O~

EoE 8000.0 0 6000.0.

0

o $1A17~) \

, , -'W"O--r•.2m 0.0 ~-mli~ 10.0 100.0 1000.0 1.0E4 ANTIBODY DILUTIONS

1.0E5

Fig, 3. The r~sults of a serum dilution titer curve are shown, with the a m o u n t of counts of ~25I-labeled goat anti-rabbit IgG F(ab)2 on the ordinate and the dilution of the primary antibody (S1A 170 and S1A258) or the pre-immune (P) sera on the abscissa, The peak of binding for both antibodies was around 1/1000 but significant binding was seen at dilutions as low as 1/50 000.

not suggested by earlier autoradiographic studies using ligand binding. The great advantage of a receptor antibody lies in its adaption to ultrastructural immunocytochemical localization (see Fig. 2) and to identify specific cell labelling in brain regions (Fig. 1B). Isolation of the receptor by immunoaffinity purification has been reported previously with an antipeptide antibody to the 5-HT1A receptor (Raymond et al., 1989). This provides a rapid and sensitive method to isolate a single protein fraction. The receptor can also be concentrated by the method of immunoprecipitation (El Mestikawy et al., 1990; see also Hollow and Lane, 1988). Western analysis of the amount of the 5 - H T I A receptor can be performed with both the SIAl70 and S1A258 antibodies that we have raised. DISADVANTAGES

protein that provides highly immunogenic sites for T-cell receptor binding. We were able to obtain radioimmunoassay binding of the native peptide over 100-fold that seen in the preimmune serum at a dilution of 1/10 000 (see Fig. 3). Anatomical and cellular localization of the receptor can be performed in neonatal and adult rats as well as in adult monkeys. All fixation solutions produced good results. The best staining of the central nervous system was in neonates and this confirms the ligand binding studies which have shown higher values for the 5-HT~A receptor during early prenatal periods (Bar-Peled et al., 1991; Daval et al., 1987; Whitaker-Azmitia et al., 1987). The most interesting observation was the specific cellular staining observed in neurons, astrocytes, tanocytes and epithelial cells. This complex distribution was

Characterization of the antipeptide antibody must establish that the native protein is selectively recognized. This requires demonstrating a specific cellular staining pattern in immunocytochemical studies. However, as discussed by E1 Mestikawy et al. (1991), the distribution of receptor staining with antibodies may not be completely consistent with the radioautographic distribution seen with ligand binding. This is explained by the observation that [3H]8-OH-DPAT, for instance, does not recognize the receptor if it is not linked to a G-protein. The antibody, on the other hand, can recognize all the receptor molecules, even those in transit from the cell body to the distal portion of the process via smooth endoplasmic reticulum (see Fig. 2).

Antipeptide Antibodies against the 5-HT~A Receptor Immunostaining of a single band in Western analysis is usually considered an indication of the general specificity of an antibody. However, this requires careful manipulation o f the antibody dilution used and preparation of the tissue sample. Protein fragments or aggregation of the molecule can result in several bands on a Western even if the antibody only recognizes a single protein. For this reason, we used a special lysis buffer and treated the tissue with reducing and denaturing conditions (such as 13-mercaptoethanol and boiling). Immunoprecipitation and immunoaffinity purification of a single protein that functions as an active receptor with the appropriate characteristics is the best criteria that the native protein is labeled by the antibody. In studies with transfected COS-7 cells, the antibody JWR21 (242-267) was shown to precipitate the [~25I]N:NAPS photoaffinity-labeled receptor (Raymond et al., 1989). The preimmune serum, the antigenic peptide-blocked JWR21 antibody or an antibody from a non-overlapping region (268-293) were all unable to precipitate the receptor. In the study by El Mestikawy et al. (1990), the 5-HTLA receptor antibody against 243-268 precipitated the binding sites of [3H]8-OH-DPAT when protein A-Sepharose CL-4B was added. No influence of the antiserum alone was seen in the binding. Costs of peptide sequencing can be very high and commercial laboratories charge several thousand dollars for mg quantities. At the Cold Spring Harbor Laboratory, the synthesis was performed in house with a 430A ABI sequencer. The 431A is a less versatile model and significantly less expensive. The H F apparatus is also a serious consideration. The apparatus alone costs approximately $5000 and requires a dedicated hood with a flow of > 150 cfm. The gas is dangerous and very difficult to obtain with current D O T laws. The alternative is T F M S A cleavage, which can be done at the bench.

ACKNOWLEDGEMENTS We would like to thank G. Binns and M. Meneilly for the preparation of the peptides at Cold Spring Harbor Laboratory Protein ChemistryCore Facility,and Lisa Bianco for preparing the antibody at Cold SpringHarbor LaboratoryAnimal Facility. Help was provided by Y. Chert in light microscopyand by P. Gannon in ultrastructural immunocytochemical analysis of the antibody. We thank Dr R. Kobayashi for his biochemical expertise and Dr J. Traber for his sponsorship. The work was supported by the NSF DevelopmentalNeuroscienceSectionand a generousgift from Miles Inc.

REFERENCES Albert, P. R., Zhou, Q.-Y., Van Toi, H. H. M., Bunzow, J. R. and Civelli, O. (1990). Cloning, functional expression, and mRNA tissue distribution of the rat 5hydroxytryptamine~A receptor gene. J. Biol. Chem. 265, 5828-5832.

297

Avrameas, S. and Ternynck, T. (1969). The cross-linking of proteins with glutaraldehyde and its use in the preparation of immunoabsorbents, lmmunochemistry 6, 53-66. Azmitia, E. C. and Gannon, P. J. (1983). The ultrastructural localization of serotonin immunoreactivity in myelinated and unmyelinated axons within the medial forebrain bundle of rat and monkey. J. Neurosci. 3, 2083-2090. Bar-Peled, O., Gross-Isseroff, R., Ben-Hur, H., Hoskins, I., Groner, Y. and Biegon, A. (1991). Fetal human brain exhibits a prenatal peak in the density of serotonin 5-HTtA receptors. Neurosci. Lett. 127, 173-176. Barany, G. and Merrifield, R. B. (1979). The Peptides: Analysis, Synthesis, Biology, Vol. 2 (eds Gross, E. and Meienhofer, J.), pp. 1-284. Academic Press, NY. Bunzow, J. R., Van Tol, H. H. M., Grandy, D. K., Albert, P., Salon, J., MacDonald, C., Machida, C. A., Neve, K. A. and Civelli, O. (1988). Cloning and expression of a rat D 2dopamine receptor cDNA. Nature 336, 783-787. Chou, P. and Fasman, G. D. (1974). Prediction of protein conformation. Biochemistry 13, 222-245. Chou, P. and Fasman, G. D. (1978). Prediction of the secondary structure of proteins from their amino acid sequence. Adv. Enzymology 47, 45-147. Davai, G., Verge, D., Becerrd, A., Gozlan, H., Spampinato, U. and Hamon, M. (1987). Transient expression of 5-HT~A receptor binding sites in some areas of the rat CNS during postnatal development. Intern. J. Neurosci. 5, 171-180. Dohlman, H. G., Caron, M. G. and Lafkowitz, R. J. (1987). Family of receptors coupled to guanine nucleotide regulatory proteins. Biochemistry 26, 2657-2663. Draeta, G. and Beach, D. (1988). Activation of CDC-II protein kinase during mitosis in human cells: cell cycle dependent phosphorylation and subunit rearrangement. Cell 54, 17-26. E1Mestikawy, S., Fargin, A., Raymond, J. R., Gozlan, H. and Hnatowich, M. (1991). The 5-HTIA receptor: an overview of recent advances. Neurochem. Res. 16, 1-10. E1 Mestikawy, S., Riad, M., Laporte, A. M., Verge, D., Daval, G., Gozlan, H. and Hamon, M. (1990). Production of specific anti-rat 5-HT l^ receptor antibodies in rabbits injected with a synthetic peptide. Neurosci. Lett. 118, 189-192. Hollow, E. and Lane, D. (1988). Antibodies: A Laboratory Manual. Cold Spring Harbor Laboratory Press, New York. Hopp, T. P. and Woods, K. R. (1981). Prediction of protein antigenic determinants from amino acid sequences. Proc. Natl. Acad. Sci. USA 8, 3824-3828. Julius, D., Huang, K. N., Livelli, T. J., Axel, R. and Jessell, T. M. (1990). The 5HT2 receptor defines a family of structurally distinct but functionally conserved serotonin receptors. Proc. Natl. Acad. Sci., USA 87, 928-932. Kobilka, B. K., Frielle, T., Collins, S., Yang-Feng, T., Kobilka, T. S., Francke, U., Lefkowitz, R. J. and Caron, M. G. (1987). An intronless gene encoding a potential member of the family of receptors coupled to guanine nucleotide regulatory proteins. Nature 329, 75-79. Laemmli, E. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-685. Marshak, D. R. and Caroll, D. (1991). Synthetic peptide substrates for casein kinase II. Methods in Enzymology, in press.

298

E.C. Azmitia et al.

Pazos, A. and Palacios, J. M. (1985). Quantitative autoradiographic mapping of serotonin receptors in the rat brain. Brain Res. 346, 205-230. Raymond, J. R., Fargin, A., Lohse, M. J., Regan, J. W., Senogles, S. E., Lefkowtiz, R. J. and Caron, M. G. (1989). Identification of the ligand-binding subunit of the human 5-hydroxytryptaminelA receptor with N(p-azido-[~25]iodophenethyl)spiperone, a high affinity radioiodinated photoaffinity probe. Molecular Pharamacology 36, 15-21.

Sotelo, C., Cholley, B., El Mestikawy, S., Gonzlan, H. and Hamon, M. (1990). Direct immunocytochemical evidence for the existence of 5-HTIA autoreceptor on serotonergic neurons in the midbrain raphe nuceli. Eur. J. Neurosci. 2, 1144-1154. Sutcliffe, J. C. et al. (1980). Chemical synthesis of a polypeptide predicted from nucleotide sequence allows detection of a new retroviral gene product. Nature 287, 801-805. Towbin, H., Staehlin, T. and Gordon, J. (1979). Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications, proc. Natl. ,4cad. Sci. USA 76, 4350-4354.

Towbin, H. and Kurstak, E. (1984). Immunoblotting and dot immunoblotting-current status and outlook. J. Immunol. Methods 72, 313-340. Verge, D., Daval, G., Marcinkiewicz, M., Patey, A., Mestikawy, E., Gozlan, S. and Hamon, M. (1986). Quantitiative autoradiography of multiple 5-HTj receptor subtypes in the brain of control or 5,7dihydroxytryptamine-treated rats. J. Neurosci. 6, 3474-3482. Whitaker-Azmitia, P. M., Lauder, J. M., Shemmer, A. and Azmitia, E. C., (1987). Postnatal changes in 5-HT-1 receptors following prenatal alterations in serotonin levels: further evidence for functional fetal 5-HT-1 receptors. Develop. Brain Res. 33, 285-295. Walter, G. et al. (1980). Antibodies specific for the carboxy- and amino-terminal regions of simian virus 40 large tumor antigen. Proc. Natl. Acad. Sci. USA 77, 5 i 97-5200. Yu, I. J. and Marshak, D. (199). Immunocytochemical localization of casein kinase II during interphase and mitosis. J. Cell Biology 114, 1217-1232.