Odorant-binding protein: Localization to nasal glands and secretions

Proc. Nail. Acad. Sci. USAVol. 83, pp. 4942-4946, July 1986Neurobiology

Odorant-binding protein: Localization to nasal glandsand secretions

(olfaction/mucus/immunohistochemistry/pyrazines)

JONATHAN PEVSNER, PAMELA B. SKLAR, AND SOLOMON H. SNYDER*Departments of Neuroscience, Pharmacology, and Experimental Therapeutics, Psychiatry, and Behavioral Sciences, Johns Hopkins University School ofMedicine, 725 North Wolfe Street, Baltimore, MD 21205

Contributed by Solomon H. Snyder, January 6, 1986

ABSTRACT An odorant-binding protein (OBP) was iso-lated from bovine olfactory and respiratory mucosa. We haveproduced polyclonal antisera to this protein and report itsimmunohistochemical localization to mucus-secreting glands ofthe olfactory and respiratory mucosa. Although OBP wasoriginally isolated as a pyrazine binding protein, both rat andbovine OBP also bind the odorants [3H]methyldihydrojasmon-ate and 3,7-dimethyl-octan-1-ol as well as 2-isobutyl-3-[3HImethoxypyrazine. We detect substantial odorant-binding ac-tivity attributable to OBP in secreted rat nasal mucus and tearsbut not in saliva, suggesting a role for OBP in transporting orconcentrating odorants.

In an effort to clarify molecular mechanisms of olfaction,several groups have examined the binding of radioactiveodorants to nasal mucosa (1-6). Recently, we purified tohomogeneity an odorant-binding protein based on its inter-actions with the potent odorant 2-isobutyl-3-[3H]methoxy-pyrazine ([3H]IBMP) (7), a finding obtained independently byBignetti et al. (8). Odorant-binding protein (OBP) is a solubledimeric protein with subunits of =19 kDa. OBP may have aselective function in olfaction, since it occurs in nasal mucosaand not in other tissues, and the relative potencies of ahomologous series of pyrazine derivatives in competing forthe binding sites parallels their potencies as odorants.We have produced antisera to bovine OBP and now report

the immunohistochemical localization ofOBP to bovine nasalglands, which secrete mucus. Odorant-binding studies dem-onstrate high levels of OBP in rat nasal mucus and tears,suggesting a role for OBP in concentrating odorants from theair.

MATERIALS AND METHODS

Materials. [3H]IBMP (43.8 Ci/mmol; 1 Ci = 37 GBq),2-[3H]methoxypyrazine (63.6 Ci/mmol), and [pentyl-2,3-3H]methyldihydrojasmonate ([3H]MDHJ; 65.5 Ci/mmol)were prepared by New England Nuclear Dupont. 3,7-Di-methyl[6,7(N)-3H]octan-1-ol ([3H]DMO; 57 Ci/mmol),[amyl-3H]isoamyl acetate (49 Ci/mmol), and [3H]isovalericacid (4-[2,3(n)3H]methylbutanoic acid; 52 Ci/mmol) wereprepared by Amersham. Unlabeled odorants were fromInternational Flavors and Fragrances (Union Beach, NJ) orPyrazine Specialties (Atlanta, GA). Protein molecular sizemarkers were obtained from Bio-Rad. All other reagentswere from commercial sources.

Preparation of Antibodies to OBP and Radioimmunoassays.Bovine OBP was purified as described (7), and antibodies tothis protein were raised by standard techniques (9). BovineOBP was iodinated with lodo-Beads (Pierce) according to themethod of Markwell (10). Radioimmunoassays were per-

formed at an antiserum dilution of 1:8100 in 0.1 M Tris HCl,pH 8.0/0.5% Triton X-100, in a final vol of 50 A.l. Incubationswere carried out at 370C for 90 min with 35,000 cpm of'251-labeled OBP per tube. Immunoprecipitation was accom-plished by using 25 1LI of 5% Staphylococcus aureus cells(Calbiochem) in 0.1 M Tris HCl (pH 8.0) at 370C for 30 min.Bound 1251I-labeled OBP was separated from free OBP byfiltration over glass fiber filters (No. 32, Schleicher &Schuell) pretreated with 10% fetal bovine serum, using aBrandel cell harvester (Brandel, Gaithersburg, MD). Intypical experiments, maximal and nonspecific binding were30% and 2% of added radioactivity, respectively.Immunoblots. Whole bovine nasal epithelia were homog-

enized in buffer A (50 mM Tris HCl, pH 7.6/1 mM EDTA)and filtered over cheesecloth. Samples of purified OBP,bovine serum albumin, and nasal homogenates prepared inbuffer A were electrophoresed into 14% NaDodSO4/poly-acrylamide gels. Transfer of proteins to nitrocellulose wasaccomplished in 12 hr at 60 V (11). The nitrocellulose wasincubated for 12 hr in buffer A supplemented with 0.1%gelatin and 0.1% Triton X-100 to decrease nonspecific staini-ing. Immunoblots were incubated with a 1:1000 dilution ofantisera for 2 hr as indicated in Fig. 2. Antisera were pre-adsorbed for 24 hr at 40C with either bovine serum albuminor purified bovine OBP that had been further purified byHPLC (see below). Immunoblots were developed by usingthe avidin/biotin/peroxidase technique (Vector Laborato-ries, Burlingame, CA) with 4-chloro-1-naphthol and hydro-gen peroxide as substrates.

Immunohistochemistry. Whole bovine nasal epithelia wereobtained immediately after slaughter and were fixed in i%glutaraldehyde in 0.15 M sodium phosphate buffer (pH 7.4)for 2 hr. Tissue was embedded in a mixture of 50% brainpaste/50% Tissue-Tek (Miles Scientific, Naperville, IL) andrapidly frozen. Cryostat tissue sections (8 ,um) were cut witha microtome and immunohistochemically stained with theavidin/biotin/peroxidase complex technique (Vector Labo-ratories) using diaminobenzidine and hydrogen peroxide assubstrates (12-14). Tissue sections were incubated with a1:20,000 dilution of antisera against either bovine OBP,bovine serum albumin (Sigma), or normal rabbit serum for 48hr at 4°C. Additional sections were counterstained withtoluidine blue.

Purification and Characterization of Bovine OBP and RatMucus OBP. Bovine OBP was purified from bovine olfactoryand respiratory epithelium as described (7) by sequentialcentrifugation; ammonium sulfate fractionation; and DEAE-cellulose, hydroxylapatite, and gel filtration chromatogra-

Abbreviations: IBMP, 2-isobutyl-3-methoxypyrazine; MDHJ,methyldihydrojasmonate (3-oxo-2-pentyl-cyclopentane-acetic acidmethyl ester); DMO, 3,7-dimethyl-octan-1-ol; OBP, odorant-bindingprotein.*To whom reprint requests should be addressed.

4942

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Proc. Natl. Acad. Sci. USA 83 (1986) 4943

phy. A single protein was purified from rat olfactory andrespiratory epithelium by using the same purification proce-dure. Purified bovine OBP and proteins present in crude ratmucus were separated by HPLC on a Vydac C4 proteincolumn (4.6 x 200 mm; 5-gim particle size) (15). The columnwas equilibrated with 0.1% (vol/vol) aqueous trifluoroaceticacid, and the samples were loaded in the same solvent. Aftera 5-min wash, the proteins were eluted with a linear gradientof 0-100% organic solvent over 40 min [0.1% (vol/vol)trifluoroacetic acid in acetonitrile/n-propanol (2:1)]. Proteinswere analyzed by NaDodSO4/PAGE (16) on 14% polyacryl-amide gels in the presence of 2-mercaptoethanol and werestained with Coomassie brilliant blue. Protein was deter-mined according to the method of Bradford (17) using bovineserum albumin as a standard.Odorant Binding to Secretions from Rat. Male Sprague-

Dawley rats (6-10 weeks old) were anesthetized with sodiumpentobarbital (60 mg/kg, i.p.) and injected with isoproterenol(30 mg/kg, i.p.) to induce secretions (18-20). After 5 min,secreted nasal mucus, tears, and saliva were collected with5-,41 or 100-gl glass micropipettes. Mucus was collected fromthe external nares and tears were collected from the puncta.Typically, 10 Al of nasal mucus, 20 Al of tears, and 200 A.l ofsaliva were obtained from a single rat in 1 hr.

Binding assays were performed by filtration over polyeth-ylenimine-coated filters as described (7). Assay mixturestypically contain secretions (2-20 jig of protein), 3-25 nMradioactive odorant, and unlabeled odorants in a final vol of100 ,ul, and they were incubated at 4°C for 1 hr.In Vivo Binding of [3H]DMO. Individual male Sprague-

Dawley rats were anesthetized and injected with isoproter-enol as described above. Two minutes after the isoproterenolinjection, a polypropylene test tube (1.5 ml capacity) con-taining 50 ,Ci in 50 ,ul of [3H]DMO was positioned around theanimal's nose, eyes, or mouth for 3 min. Secretions werecollected and assayed for protein content. Labeled odorantpresent in the mucus or saliva was measured by liquidscintillation spectrometry. Radioactivity bound to proteinwas measured by filtration (as described above) within 1 minof collection of the secretions.

RESULTSOdorant Binding to Purified Bovine OBP. OBP was first

detected on the basis of its pyrazine-binding properties (7, 8).Homogeneous bovine OBP also binds three nonpyrazineodorants-[3H]DMO, [3H]MDHJ, and [3H]amyl acetate (Ta-ble 1). Bovine OBP was purified as described (7). After

Table 1. Binding constants for odorants to purified bovine OBPand rat mucus and tears

Bovine OBP Rat mucus Rat tears

Odorant Kd,,4M Bm, Kd,,LM Blx Kd, juM Bmx[3H]Amyl acetate 68 22 >1000 ND >1000 ND[3H]DMO 0.3 30 44 1.6 100 15.7[3H]IBMP 3 27 20 2.3 11 1.7[3H]MDHJ 8 18 35 0.3 25 0.32-[3H]Methoxy-

pyrazine >1000 ND >1000 ND >1000 ND

Binding assays were performed as described (7) using bovine OBP(1-2 ,j.g of protein per tube), crude rat mucus (30-60 jig ofprotein pertube), crude rat tears (30-60 Ag of protein per tube), and 5-15 nM ofeach tritiated odorant. B.. values are nmol bound per mg of protein.No specific binding was detected for the binding of 10 nM[3H]isovaleric acid to bovine OBP or rat mucus. Less than 5 fmol permg of protein of [3H]IBMP or [3H]DMO binding to crude rat salivawas detected (n = 2). ND, not determined. Results are the mean of3-6 determinations.

o 1:

x

E

CL0

a

0

.0

.0

Cu

Ict

EC-

._--0

0-

Fraction

FIG. 1. Partial purification of OBP from bovine and rat nasalepithelium by DEAE-cellulose chromatography. After ammoniumsulfate fractionation, bovine (A) or rat (B) protein was dialyzed andapplied to a DEAE-cellulose column with a linear NaCl gradient(0-400 mM NaCl; bovine OBP elutes at 330 mM NaCl; rat bindingprotein elutes at 270 mM NaCl). All fractions were assayed for thebinding activities of [3H]DMO (8 nM), [3H]IBMP (10 nM), and[3H]MDHJ (6 nM). All three odorants bind with a single major peakof binding activity. Binding of [3H]amyl acetate and 2-[3H]methoxy-pyrazine also reveal single peaks of binding activity (not shown).Data are from a single experiment.

DEAE-cellulose chromatography, column fractions wereassayed for the binding of four tritiated odorants. A singlemajor peak of activity was found (Fig. 1A). Using the sametechniques, we have purified an OBP from rat olfactory andrespiratory epithelium. DEAE-cellulose chromatography re-veals a single peak of binding activity (Fig. 1B) and thepurified protein has an apparent subunit molecular size of 21kDa by NaDodSO4/PAGE (data not shown).We wondered if four chemically distinct 3H-labeled odor-

ants, which bind to a single protein, interact at the same ordifferent sites on bovine OBP. In competitive binding studies,each of the odorants displays similar potency in competingfor the four different 3H-labeled odorants (Table 2). Theseresults suggest that all four odorants bind to the same site onOBP. However, one cannot rule out the possibility that theybind to different sites that display allosteric interactions.

Table 2. Inhibition of 3H-labeled odorant binding to purifiedbovine OBP

IC50, JIMAmyl

Odorant acetate DMO IBMP MDHJ

[3H]Amyl acetate 0.5 18 30[3H]DMO 30 13 23[3H]IBMP 70 0.5 - 13[3H]MDHJ 65 2 2

IC50 values (concentration of unlabeled odorant that inhibitsresponse by 50%o) were calculated from displacement curves by usingbovine OBP (1-2 ,ug of protein per tube), 30-45 nM [3H]amyl acetate,3-10 nM [3H]DMO, 3-10 nM [3H]IBMP, and 3-10 nM [3H]MDHJ.Results are the mean of at least two determinations.

Neurobiology: Pevsner et al.

4944 Neurobiology: Pevsner et al.

kDa 1 2 3 4 5 6 7 8 966.2:-

45 iw-

FIG. 2. Immunoblot analysisof antisera to bovine OBP. Puri-fied OBP (20 ,g; lanes 1-3), crude

31 nasal olfactory epithelium homog-enates (40 /Lg; lanes 4-6), crudenasalrespiratoryepitheliumhomog-enates (40 ,ug; lanes 7-9). Anti-OBPantiserum at a dilution of 1:1000(lanes 1, 4, and 7); anti-OBP antise-

21 .5 I .nrum (1:1000), preadsorbed for 24 hr

144with HPLC-purified bovine OBP

14.4 3W-(50 pHg/ml) (lanes 2, 5, and 8); nor-mal rabbit serum (1:1000) (lanes 3,6, and 9).

Characterization of the Antiserum to Bovine OBP. The titerof the polyclonal antiserum and its affinity for OBP wereevaluated by radioimmunoassay. Half-maximal binding ofOBP occurs at an antibody dilution of 1:8100. To determinethe affinity ofthe antibody for OBP, we examined the bindingof 115I-labeled OBP to the antiserum in the presence ofincreasing concentrations of unlabeled OBP. Fifty percent ofmaximal binding is apparent at 0.6 nM OBP. In the radio-immunoassay, the lowest detectable level of OBP is 30 pM.

Based on the binding of [3H]IBMP, the levels of OBP in ratnasal mucosa are about the same as in bovine nasal mucosa(data not shown). However, radioimmunoassay for OBP inrat olfactory mucosal homogenates detects levels only 0.1%of those found in bovine olfactory mucosa, indicating verylittle antiserum cross-reactivity between the species.To ensure that the antiserum is selective for OBP, we

performed immunoblot analysis of bovine olfactory andrespiratory mucosa (Fig. 2). Coomassie blue staining ofpurified OBP detects the major band of protein at 19 kDa anda minor contaminant at 66 kDa (see Fig. 4 Inset, lane 2). Inimmunoblots of purified bovine OBP, antisera against OBPrecognize a band of 19 kDa and a minor contaminant at 63kDa (Fig. 2, lane 1). The 66-kDa contaminant observed withCoomassie blue staining is not recognized by the antiserumto OBP. In crude olfactory and respiratory epitheliumhomogenates, only the 19-kDa protein is recognized, con-firming that the immunohistochemical stain images onlyauthentic 19-kDa protein (Fig. 2, lanes 4 and 7). Preadsorp-tion of antiserum with HPLC-purified bovine OBP abolishesstaining of the 19-kDa protein (lanes 2, 5, and 8). The stainingand adsorption patterns are the same for both olfactory andrespiratory mucosa, although the olfactory mucosa hassomewhat less OBP, confirming earlier results based on[3H]IBMP binding (7).Immunohistochemical Loclizaon of Bovine OBP. Immu-

nohistochemical staining of bovine nasal epithelium revealsOBP-like immunoreactivity in the glands of the lamina

-s.6-

OE-4

A.^ 4b. tS

RBC J

OEAI~~~~~~

aD,'~~~~~~J

KI ; - A 'a'!WIF-~ws *s - -FS;;_ *8X ** 8S.:q ,,RE D

. @ @ . o *'f *.

- --,' u

* 'l- ~ F *;A ,

RBC

FIG. 3. Immunohistochemical localization of bovineOBP. Bovine olfactory (A, C, and E) and respiratory (B andD) epithelium are shown. Normal histology was demonstrat-ed by toluidine blue staining (A and B). The antiserumproduced against purified bovine OBP as described in Ma-terials and Methods was used at a dilution of 1:20,000 toimmunohistochemically stain bovine olfactory (C) and res-piratory (D) epithelium. (E) Staining of bovine olfactoryepithelium by normal rabbit serum (1:20,000). OE, olfactoryepithelium; RE, respiratory epithelium; RBC, erythrocyte;G, glands; LP, lamina propria; A, arteriole.

Proc. Natl. Acad. Sci. USA 83 (1986)

Proc. Natl. Acad. Sci. USA 83 (1986) 4945

120

C.,

0

x(n

C

0)

CQ

0

co

6

60

30

20

10

0

10 3020

Time, min

FIG. 4. Reverse-phase HPLC analysis of bovine OBP and ratmucus. HPLC analysis was performed after fractionation on aDEAE-cellulose column as described in Materials and Methods. (A)Bovine OBP (50 jig; retention time, 20.58 min); (B) partially purifiedrat mucus (40 ,g; retention time, 20.88 min). (Inset) NaDodSO4/PAGE analysis of bovine and putative rat OBP was performed asdescribed. Lanes: 1, molecular size markers; 2, bovine OBP (15 jg)purified as described (7); 3, the major HPLC peak ofabsorbance fromA; 4, the major HPLC peak of absorbance from B.

propria and along the surface of the epithelium (Fig. 3 C andD). Anatomically, mammalian nasal glands are differentiatedinto Bowman's glands in the olfactory mucosa and respira-tory glands, which underlie the respiratory mucosa (21).Specific staining for OBP occurs in both types of glands.Preadsorption of the antiserum with HPLC-purified OBPeliminates glandular staining (data not shown). Preadsorptionofthe antiserum with bovine serum albumin does not alter thestaining pattern. A band ofimmunoreactivity along the ciliary(outer) surface of the olfactory and respiratory epithelium isalso seen in the normal rabbit serum control and thus is notspecific for OBP. Reaction product is deposited in erythro-cytes when either OBP antiserum or normal rabbit serum isused and presumably reflects endogeneous peroxidase activ-ity.

Demonstration of Odorant Binding in Rat Nasal Mucus andTears. A major function of the nasal glands is the secretion ofmucus into the nasal cavity (22). Based on the localization ofOBP to bovine mucus-secreting glands, we investigated thebinding of odorants to secreted mucus, tears, and saliva inrats whose secretions were stimulated with isoproterenol

(Table 1). [3H]DMO, [3H]IBMP, and [3H]MDHJ bind spe-cifically and saturably to secreted rat mucus and tears, butnot to saliva. We also evaluated [3H]amyl acetate, 2-[3H]me-thoxypyrazine and [3H]isovaleric acid, all of which havedistinctive odors. Only very low levels of [3H]amyl acetateand 2-[3H]methoxypyrazine binding are detectable, and bind-ing constants cannot be obtained by filtration binding assays.We have failed to detect any binding of [3H]isovaleric acid atneutral pH to nasal mucus or OBP.Using a two-step procedure consisting of DEAE-cellulose

and HPLC chromatographic steps, we have purified OBP toapparent homogeneity from rat nasal mucus. A single peak ofbinding activity is observed after DEAE-cellulose chroma-tography (data not shown). This peak (40 ,ug; 8% of the totalprotein) is further purified by reverse-phase HPLC (Fig. 4B).A single peak is observed, closely corresponding to theelution volume of bovine OBP (Fig. 4A). This HPLC peakcontains 21 pmol of [3H]DMO bound per mg of protein([3H]DMO concentration in the assay, 41 nM). NaDodSO4/PAGE of the peak reveals a single band of 21 kDa (Fig. 4Inset) corresponding to that observed for purified rat OBPderived from olfactory epithelium (data not shown).

Concentration of [3H]DMO by OBP in Rat Mucus. Afunction ofOBP might be to interact with certain odorants invivo. To examine this possibility, rats were anesthetized andinjected with isoproterenol, and then either the nose, eyes, ormouth was exposed to [3H]DMO in ambient air as describedin Materials and Methods.

Total [3H]DMO levels in the nasal mucus are -'60 nM,while filter-bound radioactivity is "14 nM (Table 3). Slightlylower concentrations of tritiated odorant accumulate insecreted tears. No bound radioactivity is detected in salivaexposed to [3H]DMO in vivo, and saliva contains only 1% ofthe concentration of [3H]DMO measured in nasal mucus.

DISCUSSIONThe major findings of the present study are that OBP,originally isolated as a pyrazine-binding protein, binds sev-eral structurally unrelated odorants, that OBP is localized inbovine nasal glands, and that OBP is secreted into the nasalmucus of rats, where it comprises 2% of total protein.The binding of several structurally unrelated odorants to

OBP and the lack of binding of non-odorants (7) indicates thatOBP is involved in olfaction. Of numerous nasal proteins(22-27), only OBP binds odorants selectively. Moreover, themicromolar Kd of OBP for certain odorants is in the range ofconcentrations relevant to olfactory transduction bothelectrophysiologically (28) and biochemically (29). Our evi-dence that OBP interacts with odorants in the ambient airsuggests a specific role in olfaction. Odorants subjectivelydetected when present at apparent nanomolar concentrationsin the air are then concentrated to the micromolar levels inmucus required to influence olfactory neuronal firing (28) andadenylate cyclase (29). The parallel between sensory poten-cies of pyrazine odorants and their affinities for OBP sup-ports such a role.

Table 3. Concentration of [3H]DMO in rat secretions after odorant exposure in vivo

Total [3H]DMO Bound [3H]DMO

Secretion cpm/,ul Concentration, nM cpm/Al Concentration, nM % total

Nasal mucus 2280 572 (8) 60 15 539 ± 109 (8) 14 ± 3 24 ± 6Tears 1942 821(6) 51 ±22 304 ± 95 (7) 8 ± 3 16 ± 8Saliva 16± 14 (6) 0.4 0.4 0 ± 4 (6) 0 ± 0.1 0 ± 1

Individual rats were anesthetized and injected with isoproterenol, and the nose, eyes, or mouth were

exposed to odorant (17.5 AM [3H]DMO; 7.3 x 101 cpm/Al) as described in Materials and Methods.Three to 10 A.l ofmucus or tears (25-30 mg of protein per ml) and 50-100 A.l of saliva (20-30 mg of proteinper ml) were collected and assayed as described. Data are expressed as cpm/pl ± SEM of (n)determinations. Scintillation counter background (12 cpm) was subtracted from each data point.

A

20 BkDa 1 2 3 492

I ~~~~~~~~~~~66.2-45 -

I- ~~~~~31--

21.5- .w_

_ 14.4

_ I

Neurobiology: Pevsner et al.

4946 Neurobiology: Pevsner et al.

Immunohistochemical localization of OBP to bovine nasalglands indicates that this protein is not a neuronal odorantreceptor. However, the olfactory epithelium is bathed in alayer of mucus secreted by the underlying glands, whichodorants must traverse prior to their interaction withneuronal receptors (21). OBP is secreted into the nasal mucusafter isoproterenol stimulation. This finding is consistent withthe reduction in the secretory granule content of acinar cellsof the salamander olfactory glands after stimulation withisoproterenol (20) or with micromolar concentrations ofIBMP (30). OBP is presumably stored in the granules of theolfactory glands and released upon appropriate stimulation.The binding of various odorants to OBP may be relevant to

models proposed for transporting odorants (31, 32) prior totheir interaction with olfactory receptors or eliminating themafter signal transduction (21). The time required for anodorant to traverse the olfactory mucosa from the surface toreceptor sites is in part a function of the diffusion coefficientof the molecule and the viscosity of the medium (33, 34). Inaddition to concentrating odorants in the mucosa relative tothe air, OBP could decrease the diffusion delay of odorantstraversing the mucosa by acting as a selective carrier.[3H]DMO in ambient air accumulates in the mucus but not

in tears or saliva. However, the observed levels of odorant inthese secretions may not be directly comparable becausetritiated odorant is actively inhaled into the nose, but itdiffuses passively into the tears and saliva. The binding ofinhaled [3H]DMO suggests that OBP can interact with odor-ants in vivo.

It is unclear whether odorant binding in rat tears involvessecretions from the lacrimal glands or from the nose withsubsequent transport to the tears via the nasolacrimal duct(35).

We thank Drs. Karen Braas and Randall Reed for valuable advice,and Jeffrey Nye for helpful discussions. This work was supported bya grant from International Flavors and Fragrances, Incorporated.J.P., P.B.S., and S.H.S. are respective recipients of NationalInstitutes of Health Training Grant GM-07626, Training GrantMH-15330, and RSA Award DA-00074.

1. Pelosi, P., Pisanelli, A. M., Baldaccini, N. E. & Gagliardo, A.(1981) Chem. Senses 6, 77-85.

2. Pelosi, P., Baldaccini, N. E. & Pisanelli, A. M. (1982)Biochem. J. 201, 245-248.

3. Topazzini, A., Pelosi, P., Pasqualetto, P. L. & Baldaccini,N. E. (1985) Chem. Senses 10, 45-49.

4. Gennings, J. N., Gower, D. B. & Bannister, L. H. (1977)Biochim. Biophys. Acta 496, 547-556.

5. Fesenko, E. E., Novoselov, V. I. & Krapivinskaya, L. D.(1979) Biochim. Biophys. Acta 587, 424-433.

6. Rhein, L. D. & Cagan, R. H. (1980) Proc. Natl. Acad. Sci.

USA 77, 4412-4416.7. Pevsner, J., Trifiletti, R. R., Strittmatter, S. M. & Snyder,

S. H. (1985) Proc. Natl. Acad. Sci. USA 82, 3050-3054.8. Bignetti, E., Cavaggioni, A., Pelosi, P., Persaud, K. C., Sorbi,

R. T. & Tirindelli, R. (1985) Eur. J. Biochem. 149, 227-231.9. Eipper, B. A. & Mains, R. E. (1978) J. Supramol. Struct. 8,

247-262.10. Markwell, M. A. K. (1982) Anal. Biochem. 125, 427-432.11. Towbin, H., Staehelin, T. & Gordon, J. (1979) Proc. Natl.

Acad. Sci. USA 76, 4350-4354.12. Childs, G. & Unabia, G. (1982) J. Histochem. Cytochem. 30,

713-716.13. Hsu, S. M., Raine, L. & Fauger, H. (1981) J. Histochem.

Cytochem. 29, 577-580.14. Braas, K. M., Newby, A. G., Wilson, V. S. & Snyder, S. H.

J. Neurosci., in press.15. Tarr, G. E. & Crabb, J. W. (1983) Anal. Biochem. 131, 99-107.16. Laemmli, U. K. (1970) Nature (London) 227, 680-685.17. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254.18. Burton, L. E., Wilson, W. H. & Shooter, E. M. (1978) J. Biol.

Chem. 21, 7807-7812.19. Wallace, L. J. & Partlow, L. M. (1976) Proc. Natl. Acad. Sci.

USA 73, 4210-4214.20. Getchell, M. L. & Getchell, T. V. (1984) J. Comp. Physiol.

155, 435-443.21. Getchell, T. V., Margolis, F. L. & Getchell, M. L. (1984)

Prog. Neurobiol. 23, 317-345.22. Widdicombe, J. G. & Wells, U. M. (1982) in The Nose: Upper

Airway Physiology and the Atmospheric Environment, eds.Proctor, D. F. & Andersen, I. B. (Elsevier, New York), pp.215-244.

23. Rosen, R. D., Alford, R. H., Butler, W. J. & Vannier, W. E.(1966) J. Immunol. 97, 369-378.

24. Josephson, A. S. & Weiner, R. S. (1968) J. Immunol. 100,1080-1092.

25. Creeth, J. M. (1978) Brit. Med. Bull. 34, 17-24.26. Boat, T. F. & Cheng, P. W. (1980) Fed. Proc. Fed. Am. Soc.

Exp. Biol. 39, 3067-3074.27. Gower, D. B., Hancock, M. R. & Bannister, L. H. (1981) in

Biochemistry of Taste and Olfaction, eds. Cagan, R. H. &Kare, M. R. (Academic, New York), pp. 7-31.

28. Kashiwayanagi, M. & Kurihara, K. (1984) Brain Res. 359,97-103.

29. Pace, U., Hanski, E., Salomon, Y. & Lancet, D. (1985) Nature(London) 316, 255-258.

30. Getchell, T. V., Zielinski, B. & Getchell, M. L. (1985) Chem.Senses 10, 398-399.

31. Mozell, M. M. (1970) J. Gen. Physiol. 56, 46-63.32. Hornung, D. E. & Mozell, M. M. (1981) in Biochemistry of

Taste and Olfaction, eds. Cagan, R. H. & Kare, M. R. (Aca-demic, New York), pp. 33-45.

33. Getchell, T. V., Heck, G. L., DeSimone, J. A. & Price, S.(1980) Biophys. J. 29, 397-411.

34. Getchell, T. V. & Getchell, M. L. (1977) Chem. Senses Flavor2, 313-326.

35. Bojsen-Moller, F. (1964) Anat. Rec. 150, 11-24.

Proc. Natl. Acad. Sci. USA 83 (1986)