THE J O ~ A L OF BIOLQGICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 269, No. 11, Issue of March 18, pp. 8314-8318, 1994 Printed in U.S.A.

Mutational Analysis of Disulfide Bridges in the Type C Atrial Natriuretic Peptide Receptor*

(Received for publication, November 12, 1993, and in revised form, December 13, 1993)

Makoto Itakura, Masatora Iwashina, Takeshi Mizuno, Teizo Ito, Hiromi Hagiwara, and Shigehisa HiroseS From the Department of Biological Sciences, Tokyo Institute of Technology, 4259 Nagatsuta, Midoriku, Yokohama 227, Japan

The type C receptor (ANP-C or NPR-C) for the natri- uretic peptides was demonstrated, by site-directed mu- tagenesis, to have an immunoglobulin-like disulfide bonding pattern that is very similar to that of the cyto- kine receptor superfamily. The mature form of ANP-C has a disulfide-linked homodimeric structure and con- tains 5 conserved cysteine residues per subunit, all in the extracellular domain. To identify the cysteine resi- due involved in the dimerization and further to deter- mine the intramolecular disulfide bridges and their functional roles, cysteine to serine mutations of the 5 cysteine residues were constructed. An analysis of the mutant receptors expressed in COS-1 cells by lzaI-ANP binding assay and by measuring difference in their elec- trophoretic mobilities on sodium dodecyl sulfate-poly- acrylamide gels indicated that 1) the first 4 cysteine resi- dues are joined sequentially, forming the C y s ’ ” W y ~ ~ ~ ~ and Cys209-Cys267 loops of 29 and 49 residues, respec- tively; 2) the two disulfide-linked loops are essential for the ligand binding activity; 3) the 5th cysteine residue C Y S ~ ‘ ~ is used in the formation of covalently linked dimers; and 4) the covalent association of the subunit through the disulfide bond involving CysqSs has no ap- parent influence on ligand-receptor interactions. The intramolecular disulfide bond was also confirmed by direct protein sequencing of tryptic frag- ments of purified ANP-C receptor. The secondary struc- tural features revealed here will be useful in under- standing the structure and function relationships of not only the dimeric ANP-C receptor, which has only a short cytoplasmic tail, but also the ANP-A (GC-A) and ANP-B (GC-B) receptor subtypes, which have a guanylate cy- clase domain in their long cytoplasmic tail and have recently been shown to possess an oligomeric structure, since they have similarly spaced cysteine residues in their extracellular domains.

Natriuretic peptides are a family of peptides that are thought to be intimately involved in the regulation of body fluid and electrolyte balance and arterial pressure (1-3). Three members have been identified and termed ANP,I BNP, and CNP, which

* This work was supported by grants-in-aid for scientific research from the Ministry of Education, Science, and Culture of Japan and by a research grant for cardiovascular diseases from the Ministry of Health and Welfare of Japan. The costs.of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

(ext. 2281); Fax: 81-45-921-7790. f To whom correspondence should be addressed. Tel.: 81-45-922-1111

triuretic peptide A , BNP, natriuretic peptide B; CNP, natriuretic peptide The abbreviations used are: ANP, atrial natriuretic peptide or na-

C; ANP-A or GC-A, type Anatriuretic peptide receptor; ANP-B or GC-B, type B natriuretic peptide receptor; ANP-C, type C natriuretic peptide

share a common structural motif consisting of a 17-amino acid loop formed by an intramolecular disulfide bond. The receptor family (NPR) is also diverse, and currently three types of re- ceptors have been identified (for a recent review see Ref. 4): NPR-A (51, NPR-B (6-81, and NPR-C (9). These are also called ANP-A, ANP-B, and ANP-C, respectively. NPR-A and NPR-B form one subclass that belongs to the expanding family of re- ceptors that have an intracellular guanylate cyclase domain (10). In contrast, NPR-C has only a short cytoplasmic tail and lacks the guanylate cyclase activity, forming another subclass. Furthermore, NPR-B (11) and NPR-C (12) have been demon- strated to have variant forms produced by alternative splicing. Considerable sequence homologies are present in the extracel- lular domains between the guanylate cyclase-coupled and non- coupled receptors; especially, conservation of similarly spaced exoplasmic cysteine residues is remarkable. Another structural feature of the natriuretic peptide receptors is their oligomeric structures; for example, NPR-C or the type C atrial natriuretic peptide receptor (ANP-C) has been demonstrated to be a di- sulfide-linked homodimer (13, 14), and ANP-A has been sug- gested to be a n oligomer (15-17).

To advance our knowledge about the structure-function re- lationship of NPR, we carried out site-directed mutagenesis analysis of bovine NPR-C. Here we report the results of experi- ments designed to identify the cysteine residue(s) involved in the intersubunit disulfide linkage and those important in de- fining the receptor conformation. Previous cloning of the bovine (9, 12) and human (18, 19) type C receptors (NPR-C or ANP-C) indicated that the subtype consists of two products of splice variations, namely ANP-C5 (or NPR-C5) and ANP-C6 (or NPR- C6) (12). Mature forms of ANP-C5 and ANP-C6 have 5 and 6 cysteine residues, respectively, in their extracellular domains (Fig. 1). Our mutational analysis indicated that the type C receptors have a secondary structure with two loops involving the first 4 cysteines, which are paired sequentially; the 5th cysteine is important in the dimerization of the receptor sub- unit of about 65 kDa.

EXPERIMENTAL PROCEDURES

obtained from Toyobo, Osaka, Japan. DNA ligation kits and Mutan-K Muterzals-Restriction enzymes and T4 polynucleotide kinase were

site-directed mutagenesis system were from Takara, Kyoto, Japan. 32P- Labeled nucleotides, lZ6I-ANP (74 TBq/mmol), and [14Cliodoacetamide were from DuPont NEN. Rat ANP was from Peptide Institute, Osaka, Japan. Fetal bovine serum was from Filton, Victoria, Australia. Alka- line phosphatase-conjugated goat anti-rabbit IgG antiserum was from Tag0 Inc., Burlingame, CA. Oligonucleotides were synthesized with a MilligeniBiosearch Cyclone Plus DNA synthesizer.

Detection ofFree Sulfhydryl Groups ofANP-CS Receptor-The type C ANP receptor was purified from bovine lung as described previously (13,

ANP-C receptor with 6 cysteine residues; NPR, natriuretic peptide re- receptor; ANP-C5, ANP-C receptor with 5 cysteine residues; ANP-C6,

ceptods); PAGE, polyacrylamide gel electrophoresis; HPLC, high per- formance liquid chromatography.

8314

Disulfide Bridges of ANP-C Receptor 8315

3 P

Prepro MSD

FIG. 1. Schematic representation of the primary structures of bovine and human ANP-C receptors. Cysteine residues are indi- cated by horizontal linrs and italic nurnhrrs. Prepro sequences are shadowed; the presence of pre and prosequences has been suggested by sequence comparison of mature receptor (13) and its precursor (9). The complete amino acid sequences of bovine ANP-C5 (9 ) and ANP-C6 (12), and human ANP-CS (19) and ANP-Cf (181 have been reported in the indicated references; to simplify the figure. human ANP-Cf is omitted. M S D (hlack hoxrs), membrane-spanning domain.

20). Purified ANP receptor (10 pmol of subunit or 7 pg) was denatured by treating for 1 h in 6 u urea, 0.5 M 'his-HCI, pH 8.2. containing 2 mu EDTA a t 50 "C. The denatured ANP receptor was reduced with dithio- threitol (2.5 nmol; 2.5 mw) for 2 h a t 50 "C and carboxymethylated with [ITliodoacetamide (4.7 nmol; 2 kRq) for 20 min a t room temperature. The samples were electrophoresed on 7.5% polyacrylamide gels in the presence of 0.1% SDS, and the radioactivity incorporated in the receptor bands was quantitated by radioluminography using a Fuji Film bioim- age analyzer (BAS 3000).

Mutanf Construction and Expression-The cDNA clones encoding bovine ANP-CS and ANP-C6 receptors were subcloned into the expres- sion vector pcDLSRn296 (21) as described previously (12). Mutant cDNAs were created by site-specific mutagenesis according to the method of Kunkel et al. (22). The cysteine codons were modified to those of serine using the following oligonucleotides: Cys" ~ * Ser, 5"TTCTC- CGCGAGCGTCCTGCTC-3'; Cysl:I2 - * Ser, 5'"XXCGGTGTCCGAG- TACGCC-8'; Cys2"" + Ser. ~ ' - A G C G G A A C T C C T T C T T C A C - T C ~ S ~ ~ ~ -* Ser. 5'-TGATCATGTCTGCGAGTAG-3'.xd Cys""" -* Ser, 5'- GCTCTCCTTCCAAAGCTC-3'. COS-1 cells were maintained in Dul- becco's modifirdEagle's medium (Life Technologies. Inc.) containing 10 mM Hepes, 10% fetal bovine serum, 50 unitdml penicillin, and 50 pg/ml streptomycin in a controlled atmosphere of 5% CO, a t 37 "C. Transfec- tion was carried out by electroporation a s described previously ( 12). The transiently expressed receptor levels were estimated by I2"I-ANP bind- ing assay as described previously (12) and were found to vary from 140 to 680 fmoVmg of protein reflecting variations in transfection efliciency.

Drtermination of Disulfidr-linked Peptides of Purified ANP-C Receptor-Purified receptor (300 pg in 3 ml of 0.1 u sodium acetate, pH 5.0; Ref. 20) was dissolved in 4 M urea, adjusted to pH 7.5 with solid ammonium carbonate, and digested with trypsin (4 pg) which was added in two portions immediately and after 24 h of incubation a t 37 "C. Total incubation time was 48 h. The digested receptor was divided into two equal aliquots and incubated in the absence or presence of 1 mu dithiothreitol a t 50 ̂ C for 12 h. The reduced and nonreduced samples were separated by reverse phase HPLC using a pBondasphere 5-pm C18-300Acolumn (3.9 x 150 mm). Elution was performed with a linear gradient of 10-50% acetonitrile in the presence of 0.05% trifluoroacetic acid over 3 h a t a flow rate of 1 mumin. The HPLC-isolated fragments were then analyzed for amino acid sequence. Amino acid sequence was determined by automated Edman degradation in an Applied Riosys- tems 470A protein sequenator.

Other Methods-Radioligand binding assay, aflinity labeling, and immunoblot analysis (23, 24) of mutant receptors were carried out according to the published methods (12).

RESULTS

No Free Sulfhydryl Groups i n Dimeric ANP-C5-To deter- m i n e w h e t h e r ANP-C5 has unpaired cysteine residues, we iso- la ted ANP receptors f rom bovine lung by aff ini ty chromatogra- phy (20) and measured sulfhydryl content in the receptor

Silver stain Fluorograrn "

D7T "1 1 2 3 4

FIG, 2. Incorporation of ["Cliodoacetamide into reduced hut not into native ANP-CS receptor. Typr (' ANI' rrcrptor. which hns been demonstrated to br a disulfidr-linkrd hornodlmrr 1 I.?. 141 . was purified from bovine lung and rractrd wlth I"(:Ilntfoacrtarn~dr, a sulf-

2 and 4 1 conditions. The reaction products wrrr srpnratrd hy SDS- hydryl reagent. under nonrrducinp llanrs I and .7 I and rrduclnp rlnnp.5

PAGE and visualized by silvrr stainlng I1nne.s I and 2 I and fluorogra- phy (lanes 3 and 4 I. For fluorography. thr stainrd prl was rxposcd t o a

I'TIiodoacetamidr lahrling in lanr 3. D7". dlthiothrritnl. Fuji Film imaging plate for RAS 3000 fnr 2 days. Sotr thr ahsrncr of

preparat ion, whose major const i tuent was ANP-C5 (1.7. 201. using the radiolabeled sulfhydryl reagent [ "'C liodoacetamide. No significant labeling ofANP-C5 with I ' "Cliodoacetnmide was observed when i t was in a dimeric form (Fig. 2, l a n r 3 I . Upon reduct ion of the sample, however , the 65-kDa monomeric form ofANP-C5 was labeled with 1 ' "Cliodoacetamide (Fig. 2. l a n r 4 ).

These resul ts indicate that all of t h e 5 cysteine residues in ANP-C.5 are l inked by intra- and interchain disulfidc. bridges. and homodimeric ANP-C5 conta ins no free su l lhydry l poups .

C ~ S ~ ' ' ~ Is Involved in Intrrchain Disulfide Bond--Ry a n a l o m wi th the disulfide bonding pattern of the cytokine receptor superfamily (2.5) whose members ( including the granuloc-yte colony-stimulating factor receptor, prolactin receptor. erythro- poietin receptor, and growth hormone receptor) have a dimeric structure a n d are supposed to have immunoglohulin-like di- su l f ide l oops , we a s sumed t ha t t he 1st and 2nd cys te ine resi- dues ( C y s I o 4 a n d C ~ S ' : ' ~ ) of bovine ANP-C5 form a disulfide bond as do the 3rd a n d 4 t h f Cys2n'' a n d Cys2" I , l eav ing the 5 th Cys4"' unpai red and tha t th i s unpai red cys te ine res idue is responsible for dimerization of ANP-C5 by an interchain disul- fide bond (see Figs. 1 a n d 7 ) . This assumption was supported hy the following observations.

We first mutagenized Cys"", a candidate for the intersuh- uni t d isulf ide bond, expressed the mutant in COS-1 cel ls , and ana lyzed i t s subuni t s t ruc ture and '2"I-ANP hinding activity. Cysteine to ser ine subs t i tu t ion at position 469 ~ANF"C5"~"""~ did not affect the receptor activity including the aFfinity and ligand specificity (data not shown); however, it resulted in a complete loss of disulfide-linked dimerization capacity. produc- ing on ly t he 65-kDa monomeric form of the receptor (Fig. 3. l ane 2 ). T h e term monomeric stands for rnonornrric whrn ana- lyzed by nonreducing SDS-PAGE, and it is highly possible that the m u t a n t ANP-C5"4'""s exists as a noncovalently associated homodimer in the absence of SDS because d r te rgent ex t rac ts of the COS-1 cel ls expressing the wild type and mutant receptors exhibited, on gel filtration, ""I-ANP hinding activities in 140- kDa or higher molecular mass f ract ions. This is consistent with our previous observation, using a purified ANP-C receptor prepara t ion , that the receptor subuni t is active and has a st rong tendency to form a noncovalently associated dimer even after par t ia l reduct ion and carboxymethylat ion (201.

In the case ofANP-CG which has an extra cysteinr rcs idur a t

8316 Disulfide Bridges of ANP-C Receptor

200 - 116-

66 - 42 - 30 -

"ohheled 1 + 1 - ] + I - ] + I - 1.1-1 ANP

1 2 3 4 5 6 7 8

FIG. 3. Afflnity labeling of the 5th-cysteine ( C y ~ ' ~ 9 mutants of ANP-C5 and ANP-CI receptors demonstrating the involvement of C y P " in disulfide-mediated covalent dimerization. Mem- branes of COS-1 crlls rxpressinfi the control ilanrs 3 . 4 , 7. and 8 ) or mutant (lanes 1.2, .5, and 6 ) rcccptors were labeled with 1251-ANP using

a n excess of unlabeled ANP and analyzed by SDS-PAGE under nonre- 1,5-difluoro-2,4-dinitrobenzene (Pierce) in the presence and absence of

ducing conditions (12).

position 472 (12) (Fig. 1). a position very close to the 5th Cys"' and in close proximity to the membrane spanning domain, the cysteine mutation ( C ~ S " ~ Ser) did not inhibit the dimeriza- tion (Fig. 3, lane 6 ) , indicating that the 6th cysteine (Cys"") in ANP-C6 can also be involved in the formation of intersubunit disulfide bond. As expected, the ANP-C6'472s mutant, whose 5th cysteine (Cys"") is intact, behaved like the wild type ANP-c5 (12).

Cysteine Pairs Forming Intramolecular Disulfide Uonds-To determine disulfide bonding pattern of the first 4 cysteine resi- dues in ANP-C5, we prepared the mutants shown in Fig. 4 by replacing, individually or in combination, the cysteine residues CysrO", CYS':'~, Cys""!', and CYS'"~ with serine. To prevent thiol- catalyzed disulfide shuming and to make the interpretation of results simple, the 5th cysteine Cys4"", which is not involved in intramolecular disulfide bonding, was also changed to serine in all mutants. The presence of intrachain disulfide bonds and their disruption by mutagenesis were detected by the differ- ence in their migration on SDS-PAGE. Fig. 4 shows a typical example of such analyses. The 3rd-cysteine mutant ANP- C5""""s. '"" exhibited a considerable difference in the elec- trophoretic mobility (Fig. 4, lane 3 ) compared with the control (ANP-C5""6fi!'S; lane 2 ), whereas the 2nd-cysteine mutant ANP- C5"'"2s. C4fi'S showed only a small, almost inappreciable, dif- ference (lane 4 ) . This result indicates that 1) t h e 2 n d ( C ~ S " ~ " ) a n d 3 r d ( C ~ S " " ~ ) cysteines are not joined by a disulfide bond; and 2) the 2nd and 3rd cysteines are involved, respectively, in the formation of a short and a relatively long loop. The remain- ing possible combinations are: CysIo4 (lst)-Cys'"" (2nd) and

(2nd)-CyP7 (4th) . The latter set of the combinations appear to be unlikely because they generate two large disulfide-linked loops of more than 100 amino acids. These considerations leave us with the sequential pairing of the first 4 cysteine residues, namely C y ~ " ~ - C y s ' ~ ~ a n d C y ~ ~ ~ " - C y s " " . T h i s secondary struc- ture was confirmed by the following analysis using double cys- teine mutants of the control ANP-C5':"""". ANP-C5"rZz2"~ c209s. C460R in which the 2nd and 3rd cysteines, as well as the 5th one, are converted to serine showed an electrophoretic mobility

CYS""" (3 rd ) -Cy~""~ (4 th ) or Cys"" (1st)-Cyszo" ( 3 r d ) a n d C Y S ~ ~ "

@ * ' ?

+ Dtmc,

"

I

Mutants

1 2 3 4 5 6

Fw. 4. SDS-PAGE analysis of the intrachain disulfide bonding patterns of ANP-CS and its cysteine mutants hased on the dif- ferences in their electrophoretic mohilitirs. Xlc-mhranr prrpnr:l- tlons from COS-1 cells rxprc-ssing the rndlc:ltcd wild typr Inatlvrj and mutant receptors were suhjrcted to SDS-l'A(;E on IO? grls undrr non- reducing conditions. and the rrceptor bands wcrr visuallzrd by LVrstrm blotting. This analysis relies on the fact that disruption of Intracharn disulfidr bonds of proteins usually results in rrtardation in thrir SDS- PAGE mobilities. The sc.hrmnfir cfinprnrn in t h r l o u w port shows thr structures of native and mutants rrcrptors wlth intact and dlsruptrd disulfide loops.

identical to that of fully reduced receptor, indicating disruption of both disulfide loops by the mutation (Innc. 5 J; and ATP- C5"2"11S."2R~S~("~fioS (lanr 6), in which the 3rd and 4th cysteines as well as the 5th one are mutagenized to serine, behaved. on SDS-PAGE, in exactly the same way as the 3rd-cysteine mu- tant (ANP-C5"20"s, c"";ns) (lane 3 ), suggesting that the 3rd and 4th cysteines are joined together.

Confirmation of the Intramolrcular Disulfide Bonding Pat- tern by Direct Protein Srqcwncing-Although the above assign- ment of pairing of cysteine residues appeared to be unambigu- ous, we carried out the following experiment for confirmation because the assignment was based solely on the changes of cysteine mutants in their electrophoretic mobilities, ANP-C re- ceptor was purified from bovine lung and digested with trypsin in the presence of 4 urea. The tryptic digests were chromato- graphed on a reverse-phase HPLC column in the absence (Fig. 5A) and presence (Fig. 5 R ) of dithiothreitol. Among many peaks, there were one major and several minor peaks that disappeared on reduction, indicating that such peaks contained the disulfide-linked peptides of interest. The major peak uniquely present in the nonreduced sample (Fig. SA; indicated by shaded bar) was collected and sequenced. As expected, each sequencing cycle gave the phenylthiocarbamoyl derivatives of 2 amino acids (Fig. 5 , inset 1 ): cycle 1 (Phe, Glv), cycle 2 (Cln. Ala), cycle 3 (Val, Lys), cycle 4 (Ala, Pro), cycle 5 ' 5 r . Asp). and so on. This combination is consistent with the above assign- ment of disulfide bridge between the 1st and 2nd cysteine resi- dues ( C y ~ ' ~ ~ - - C y s ~ ~ ~ ) , which yields the following disulfide- linked peptide when the receptor molecule is cleaved with trypsin as schematically shown in Fig. 5 . inwt 2.

Phe"'-GIn-VaI-AIa-~-Glu-A.sp-Ser-Asp-Cvs'"'-GIy-AIa-Arp

Gly1Z1-AIa-Lys-Pro-ksp-Leu-Ilr~I~u-Gly-Pro~Val-Cys"'-Clu-~r-Ala- Ala-Ala-Pro-Val-Ala-Arg

I

PErT11)E I

Disulfide Bridges of ANP-C Receptor 8317

0 d l Cycle 1 7 3 4 5 *R94

F Q V A Y G A K P D

A- Non-reduced

(1 30 kDa) Dimer -

I

85 90 95 100 105 110 115 .1

Elutlon Tlme fmln)

FIG. 5. Isolation of disulfide-linked peptides from tryptic di- gests of bovine ANP-C receptor by HPLC and their sequences. Purified ANP-C receptor was denatured with urra and digested with trypsin. The digest was incubated in the absence (A ) or presence ( H ) of the reducing agent dithiothreitol fD7'7'). A, elution profile of the non- reduced tryptic digest from a C18 HPLC column. H . elution profile of the reduced tryptic digest. Acomparison ofA and B reveals a relatively high peak at a position of 96 min which is unique to A. Insrt 1 , amino acid sequences of the 96-min peak indicated by shaded bar. Only the first five cycles are shown. Inset 2, schematic representation of generation of disulfide-linked peptides from the first disulfide loop by trypsin diges- tion of ANP-C receptor. Arrou&mfs represent the cleavage sites corre- sponding to the NH, termini, and open arrotuheads indicate the car- boxyl-terminal cleavage sites; for detailed sequences of the resulting peptides see "Results."

The amounts of peptides contained in the minor peaks, which may include the one corresponding to the link between the 3rd and 4th cysteines, were not sufficient for sequencing. This low yield may be because of partial digestion of the receptor prepa- ration; such incomplete digestion is often encountered even in the presence of urea when proteins are digested without reduc- tion of their disulfide bonds.

Intrachain Disulfide Bridges Are Indispensable for Binding Activity-Binding assay using the mutants prepared above in- dicated that although the 5th-cysteine mutation had no effect on receptor affinity, the cysteine to serine mutation of any of the first 4 cysteine residues resulted in a complete loss of IZ5I-

ANP binding activity (data not shown); to ensure that the mu- tant proteins were expressed at similar levels, the mutant re- ceptor levels in membranes from each transfection were compared by Western blotting (for example, Fig. 4). These re- sults indicate that the two disulfide bonds Cys'"4"cys1~''2 and Cy~'"~-Cys'"~ are essential for proper folding of the polypeptide chain and bringing the discontinuous determinants of the ANP binding site into spatial proximity.

Characterization of Intermediate Molecular Mass Form of ANP-C-Our previous analysis of the subunit structures of ANP-C5 and ANP-C6 by Western blotting revealed the pres- ence of intermediate molecular mass forms of them (12) (Fig. 4, lane 1, and Fig. 6). For example, under nonreducing conditions, an antiserum against the bovine type C ANP receptor stained three bands: a 140-kDa dense band, which corresponds to the disulfide-linked homodimer; a weak band of 70-80 kDa, which is termed here the "intermediate molecular mass form"; and a weak band of about 65 kDa, which represents the nondisulfide- linked subunit. These three bands have also been clearly dem- onstrated by Porter et al. (26) by affinity labeling of ANP-C5 expressed on cultured bovine aortic smooth muscle cells and that expressed in mouse fibroblast L-M cells using a vaccinia virus expression vector. We first considered that the interme- diate form may be a n unprocessed biosynthetic precursor of the receptor. This possibility was, however, eliminated by the fol- lowing observations. The 70-80-kDa band of the intermediate molecular mass form was abolished 1) by the cysteine to serine

Monomer + ( 65 kDa)

1 2

Flc;. 6. Elimination of involvement of the signal sequence in the formation of intermediate molecular mass form of ANP-CS. A mutant of ANI'-C5 was constructed, In which thc. vystcslnc. r l w d u r t C y s ' ? ~ in the signal scqurncr IS rrplacrd with srnnr. rxprrswd In COS-1 cells, and its molrcular forms wrrr rxnmlnrd hy Sl)S-i'A(;E under nonreducing conditions and hy aflinlty Iahrllnp.

mutation of the 5th cysteine (ANP-C.i''"""s; Fig. 4, lane 2 I and, furthermore, 2) simply by reduction with 2-mrrcaptorthanol (data not shown). These results indicate that the intermcdiatr form is a disulfide-linked heterodimer consisting of thr 65-kDa subunit of ANP-C and an as yet unidentified componrnt of about 5-15 kDa. Since the presequences of thr human (18. 19) and bovine (9, 12) ANP-C receptors contain a cysteine rrsidur in a conserved position, Cys". we examined the possihility that the presequences might be the component of the intermediate molecular mass species by site-directed mutagenesis. Contrary to expectations, replacement of the cystcine with serinr flLUP- c5'"2sJ had no effect on the appearance of thr intermrdiatc form (Fig. 6, lane 2) . The fact that the mutant receptor ANP- C5(''2s was processed to the mature active receptor, howrvrr. indicates that the Cys" is dispensahle for the function and processing of the presequence.

DISCUSSION

We previously demonstrated, bv using purifird hovinr ANP-C receptor, that reduction of it with 5 m>c dithiothrritol followed by carhoxymethylation disrupted almost complctely the disulfide cross-linking between the two homologous suh- units without affecting its ligand binding activity ( 2 0 ) . This result of chemical modification, if taken together with that of the above mutational modification, indicates that thr two in- tramolecular disulfide bonds that are indispensablr for ligand binding are quite resistant to reducing agents, whereas the interchain disulfide bond is relatively readily accessible to thr reagents. The two intramolecular disulfide bridges, which have been demonstrated here to be important in defining the proper geometry of the ligand binding site, therrfnre, appear to br buried inside the molecule.

The proposed model for the pairing of cysteine rrsidurs (Fig. 7 ) is very similar to those of the members of the cytokinr rr- ceptor superfamily which share an immunoglobulin-likr disul- fide-linked homo- or heterodimeric structure. On this struc- tural basis, we can expect that the ANP-C rrceptor. l ik r th r cytokine and growth factor receptors. might he involved in t h r control of cell growth and differentiation. In this context. the report by Levin and Frank (27) seems to he very intrrcsting in that ANP and BNP decreased the incorporation of ["H Ithymi- dine into cultured rat astroglia and inhibited their prolifrration

Disulfide Bridges of ANP-C Receptor 8318

A ANP-C5

0

ANP-AIANP-B

Kinase-like

GC

FIG. 7. Disulfide bonding patterns of ANP receptors. Panel A , a model of ANP-C5 receptor illustrating the intra- and interchain disul- tide bonds determined in the present study by site-directed mutagen- esis. The first 4 cysteine residues form two disulfide loops: a short and a relatively long one. Numbers indicate the positions of cysteine resi-

model for ANP-A and ANP-B receptors proposed based on the conser- dues in the bovine ANP-C5 receptor. Panel B , a hypothetical structural

vation of cysteine residues; the 2 cysteine residues (arrows) near the transmembrane domain are probably involved in oligomerization of the guanylate cyclase (GC)-coupled receptors (15).

and that such a strong antiproliferative effect was also ob- served with the ring-deleted analog C-ANF (de~[Glnl~-Ser '~- Gly20-Leu21-Gly221ANP9-23). As the authors claimed, these ef- fects appears to be mediated by the ANP-C receptor because the analog C-ANF has been demonstrated to be very specific to the type C receptor (28). Cahill and Hassid (29,301 have also dem- onstrated a close relationship of antimitogenic effectiveness of natriuretic peptides to expression of ANP-C receptors in rat cultured aortic smooth muscle cells. Growth-regulatory prop- erties of ANP have been reviewed by Appel (31).

Examination by alignment of the NPR sequences of different subtypes and species (5 ,6 ,8 ,9 , 12, 18, 19,32,33) revealed that there is significant sequence conservation in their extracellular domains; most notably conserved is the pattern of cysteine residues (15) except the numbers of cysteine residues in the juxtamembrane segment: ANP-C5 has 1 (CYS~~ ' ) in this region, but ANP-A and ANP-B have 2 such cysteine residues (Fig. 7). This similarity suggests that 1) the two-disulfide-linked-loop structure determined here for the ANP-C receptor is also pos- sessed by the guanylate cyclase-coupled ANP-A (GC-A) and ANP-B (GC-B) receptors, and 2) in all probability, the 2 cys- teine residues located in the juxtamembrane regions of ANP-A and ANP-B (one corresponds to the 5th cysteine of ANP-C and the other is unique to the types A and B receptors) are involved in the formation of the recently identified disulfide-linked oli- gomers, probably tetramers (15, 161, since each subunit can be

covalently linked with two other subunits through the 2 cys- teine residues (Fig. 7).

Acknowledgments-We thank Dr. Yutaka Takebe (National Institute of Health, Japan) for providing the pcDL-SRa296 expression vector and Dr. Sadami Shibabe (Japan Atomic Energy Research Institute) for help in radioluminography. We also thank Setsuko Satoh and Kazuko Tanaka for secretarial and technical assistance.

2 1

3 4 5

6

7

8.

9.

10. 11.

12.

13.

14.

15.

16. 17. 18.

19.

20.

21.

22.

23.

24.

25. 26.

27. 28.

29.

30. 31. 32.

33.

REFERENCES de Bold, A. J. (1985) Science 230, 767-770 Brenner, B. M., Ballermann, B. J., Gunning, M. E., and Zeidel, M. L. (1990)

Rosenzweig, A., and Seidman, C. E. (1991)Annu. Rev. Biochem. 60,229-255 Koller, K. J., and Goeddel, D. V. (1992) Circulation 86, 1081-1088 Chinkers, M., Garbers, D. L., Chang, M.-S., Lowe, D. G., Chin, H., Goeddel. D.

V., and Schulz, S. (1989) Nature 338, 78-83 Chang, M.-S., Lowe, D. G., Lewis, M., Hellmiss, R., Chen, E., and Goeddel, D.

V. (1989) Nature 341, 68-72 Koller, K. J., Lowe, D. G., Bennett, G. L., Minamino, N., Kangawa, K., Matsuo,

H., and Goeddel, D. V. (1991) Science 252, 120-123 Schulz, S., Singh, S., Bellet, R. A., Singh, G., Tubb, D. J., Chin, H., and

Garbers, D. L. (1989) Cell 58, 1155-1162 Fuller, F., Porter, J. G., Arfsten, A. E., Miller, J., Schilling, J. W., Scarborough,

R. M., Lewicki, J. A., and Schenk, D. B. (1988) J. Biol. Chem. 263, 9395- 9401

Physiol. Rev. 70, 665499

Chinkers, M., and Garbers, D. L. (1991) Annu. Reu. Biochem. 60, 553-575 Ohyama, Y., Miyamoto, K., Saito, Y., Minamino, N., Kangawa, K., and Matsuo,

H. (1992) Biochem. Biophys. Res. Commun. 163,743-749 Mizuno, T., Iwashina, M., Itakura, M., Hagiwara, H., and Hirose, S . (199315.

Biol. Chem. 268,5162-5167 Shimonaka, M., Saheki, T., Hagiwara, H., Ishido, M., Nogi, A,, Fujita, T.,

Wakita, K., Inada, Y., Kondo, J., and Hirose, S . (1987) J. Biol. Chem. 262,

Schenk, D. B., Phelps, M. N., Porter, J. G., Fuller, F., Cordell, B., and Lewicki, 5510-5514

Iwata, T., Uchida-Mizuno, K., Katafuchi, T., Ito, T., Hagiwara, H., and Hirose, J. A. (1987) Proc. Nut!. Acad. Sci. U. S . A. 84, 1521-1525

Lowe, D. G. (1992) Biochemistry 31, 10421-10425 S. (1991) J. Biochem. 110,35-39

Chinkers, M., and Wilson, E. M. (1992) J . Biol. Chem. 267, 18589-18597 Lowe, D. G., Camerato, T. R., and Goeddel, D. V. (1990) Nucleic Acids Res. 18,

Porter, J. G., Arfsten, A., Fuller, F., Miller, J. A,, Gregory, L. C., and Lewicki, J.

Uchida, K.. Mizuno, T., Shimonaka, M., Sugiura, N., Nara, K., Ling, N., Hagi-

Takebe, Y., Seiki, M., Fujisawa, J. , Hoy, P., Yokota, K., Arai, K., Yoshida, M.,

Kunkel, T. A,, Roberts, J. D., and Zakour, R. A. (1987) Methods Enzymol. 154,

Shimonaka, M., Saheki, T., Hagiwara, H., Hagiwara, Y., Sono, H., and Hirose,

Kawaguchi, S., Uchida, K., Ito, T., Kozuka, M., Shimonaka, M., Mizuno, T., and

Kaczmarski, R. S., and Mufti, G. J. (1991) Blood Rev. 5, 193-203 Porter, J. G., Scarborough, R. M., Wang, Y., Schenk, D., McEnroe, G. A., Kang,

Levin, E. R., and Frank, H. J. L. (1991)Am. J. Physiol. 261, R453-R457 Maack, T., Suzuki, M., Almeida, A., Nussenzveig, D., Scarborough, R. M.,

Cahill, P. A,, and Hassid, A. (1991) Biochem. Biophys. Res. Commun. 179,

Cahill, P. A., and Hassid, A. (1993) J. Cell. Physiol. 154, 28-38 Appel, R. G. (1992) Am. J. Physiol. 262, F911-F918 Lowe, D. G., Chang, M.-S., Hellmiss, R., Chen, E., Singh, S., Garbers, D. L.,

Pandey, K. N., and Singh, S. (1990) J. Biol. Chem. 265, 12342-12348 and Goeddel, D. V. (1989) EMBO J. 8, 1377-1384

3412

A. (1990) Biochem. Biophys. Res. Commun. 171, 796-803

wara, H., and Hirose, S. (1989) Biochem. J. 263, 671478

and Arai, N. (1988) Mol. Cell. Biol. 8, 466-472

367-382

S. (1987) Am. J. Physiol. 253, F1058-F1062

Hirose, S. (1989) J. Histochem. Cytochem. 37, 1739-1742

L.-L., and Lewicki, J. A. (1989) J. Biol. Chem. 264, 14179-14184

McEnroe, G. A,, and Lewicki, J. A. (1987) Science 238, 675-678

160G-1613