Eur. J. Biochem. 201,607-614 (1991) 0 FEBS 1991

Structurelactivity relationship of eel calcitonin A study using a newly devised method for designing analogs

Akira INOUE, Mayumi SHIKANO, Yasuhiko KOMATSU, Junko OBATA, Junko OCHIAI, Hiroko NISHIDE, Noriko ITO, Hiromasa NAGAO, Kiyoshi KONDO, Daiei TUNEMOTO, Hiromichi HEMMl and Naganori NUMAO Protein Engineering Section, Sagami Chemical Research Center, Sagamihara, Kanagawa, Japan

(Received April 2, 1991) - EJB 91 0426

A series of analogs of eel calcitonin (eCT) was synthesized according to a newly devised scheme, ‘the insertion- inactivation method’, to clarify the structure/activity relationship of a given peptide. This method consists of two steps: the deletion of a residue of the peptide is first chosen and then a series of analogs with the residue reinserted into serial positions is synthesized and biological activities are assessed in each step.

An analog lacking Lysl8 (dK), selected as a deleted analog for the first step, showed marked loss of activities determined by inhibition of L251-eCT binding, growth inhibition, and cAMP production in a porcine kidney cell line LLC-PKI. Activities of a set of 20 analogs with the reinserted lysine residue at serial positions from 12 to 32 (K12 -K32) were then evaluated. The results showed the following three patterns of the expression of activities according to the position of the reinsertion: (a) analogs K12-Kl6 (positions 12-16) and K25 (position 25) showed lower activities than eCT in all assays; (b) K17-K24 (positions 17-24) showed slightly lower activities than eCT in the receptor binding and the growth inhibition and similar level in cAMP production; (c) K26 - K32 (positions 26- 32) showed considerably lower activities in the former two assays and slightly lower activity in cAMP production. Further, analogs considerably less active than eCT showed unchanged a-helix contents and destroyed amphiphilicity by the insertion of a lysine residue, indicating that amphiphilicity is one of important factors for expressing the activity.

The results obtained here lead to a conclusion on the significance of each region of eCT molecule as follows: (a) the presence of Lysl8 is necessary for the complete expression of biological activity; (b) the length of amphiphilic a-helix to be required for the activity is at most 10 residues ranging from position 8 to position 17; (c) the receptor binding region is located within 9 residues ranging from position 24 to position 32.

Calcitonin (CT) is a calcium-regulating peptide hormone found in many vertebrate species. The common structure of CT is 32 amino acid residues with a seven-residue cyclic loop formed by a disulfide bond between cysteines at positions 1 and 7 and prolinamide at the carboxyl terminal. CTs are categorized into three groups by both sequence and biological potency: ultimobranchial CTs [salmon CT (sCT), eel CT (eCT), and chicken CT]; artiodactyl CTs [porcine CT (pCT), bovine CT, and ovine CT]; and primate or rodent CTs (human CT and rat CT) [l]. Among them, CTs of ultimobranchial origin are the most potent in hypocalcaemic activity deter- mined in human and rat [l -41. eCT and sCT are also known to increase the cAMP production in mammalian primary and established cells [4- 61 and to inhibit the growth of established cell lines [6, 71. The correlation between in vivo hypocalcaemic activity and the in vitro activity such as the cAMP production and the growth inhibition is not clear.

Many studies on the structure/activity relationship in ul- timobranchial CTs such as sCT and eCT have been carried out. sCT [8] and pCT [9] are known to form a-helical structure

Correspondence to A. Inoue, Sagami Chemical Research Center, 4-4-1 Nishi-Ohnuma, Sagamihara, Kanagawa, Japan 229

Ahhreviutions. CT. calcitonin; sCT, salmon calcitonin: eCT, eel calcitonin; pCT, porcine calcitonin; dK, analog of eCT with Lysl8 deleted; K12-K32, analogs of dK with the Lys residue reinserted at positions 12 - 32 (see Table 1); iBuMeXan, isobutylmethylxanthine.

in the presence of amphiphiles such as trifluoroethanol and certain phospholipids. Kaiser and his coworkers have dra- matically demonstrated that residues in amphiphilic a-helix- forming regions of sCT at the positions between 8 and 22 can be replaced with unrelated sequences without marked loss of activity [lo- 121. Other factors such as conformational flexibility [13, 141 and long-range interactions [14] also modu- late hormonal potency. CT exerts its biological manifestations via specific receptors on the membrane of target cells [15, 161. Thus, CT is assumed to have the region on the molecule to be recognized by the receptors. Yamamoto et al. showed that a C-terminal fragment of eCT had weak but essential activity for the receptor binding assay [17]. However, the sites of receptor binding and/or other activities encoded in the CT molecules have not been determined into a small region. More- over, there is a limitation in the study using fragment peptides of CTs because the fragment(s) can express activity much less than that of the native form [17, 181.

In this study, to clarify the structure/activity relationship of eCT, we carried out a newly developed method ‘the inser- tion-inactivation method’. This comprises two steps. First, an analog lacking residue Lysl8, whose importance has been pointed out by D’Santos et al. [19], was synthesized. Then, a set of analogs with the lysine residue reinserted at a serial position of eCT was synthesized. It is assumed that the inser- tion of a lysine residue into important regions of the peptide

608

should result in loss of activity. To determine the biological activity of these analogs, we employed three in vitro assay procedures in a porcine kidney cell line, LLC-PK1 [20], i.e. receptor binding, growth inhibition, and CAMP production. We confined the length of the regions of eCT responsible for the expression of its in vitro biological activity. We also discuss the effectiveness of the newly devised method for designing analogs.

MATERIALS AND METHODS

Peptides and other reagents

sCT was purchased from Novabiochem AG (Laufelfingen, Switzerland). Arginine vasopressin [Arg8]vasopressin and bovine insulin were purchased from Sigma Chemical Co. (St. Louis, MO). Human parathyroid hormone 1 - 34, human atrial natriuretic peptide and adrenocorticotropic hormone 1 - 24 were purchased from Peptide Institute, Inc. (Minoh, Japan). Glucagon was purchased from Peninsula Laborato- ries, Inc. (Belmont, CA). Isobutylmethylxanthine (iBuMeXan) was purchased from Sigma. Fatty-acid-free bov- ine serum albumin was a product of Chiba Chikusan Kogyo (Chiba, Japan). Other reagents used are available in commer- cial with a reagent grade.

eCT and its analogs (sequences shown in Table 1) were synthesized by solid-phase methods [21] using a benzhydryl- amine-substituted polystyrene resin on a peptide synthesizer (model 430A or 431A, Applied Biosystems, Inc., Foster City, CA) by using the reagents supplied. Deprotection and cleav- age from the resin were performed with the mixed solvent of trifluoroacetic acid and trifluoromethanesulfonic acid in the presence of cation scavengers.

To remove the low-molecular-mass materials, crude peptide solution was gel-filtered on a Bio-Gel P-2 column

Table 1. Amino acid sequences of eCT and its analogs Amino acid residues are expressed in single-letter code

(Bio-Rad Laboratories, Richmond, CA; 25 mm internal diameter x 35 cm) in 1O0/o (by vol.) acetic acid. Formation of the intra-molecular disulfide bond occurred during the overnight air-oxidation at room temperature in 6 M guanidine . HC1 solution (pH 7.5) and was confirmed by titration with 5’,5’-dithiobis (2-nitrobenzoic acid) using cysteine as the stan- dard. Peptides were further purified by three steps of HPLC. The first chromatography was carried out on a preparative CI8 reverse-phase (RP) column (TSKgel ODS ~OTM, 21.5 mm internal diameter x 30 cm, Tosoh Co., Ltd, Tokyo, Japan) with the solvent system of a linear gradient of 20-50% CH3CN in 0.1 YO trifluoroacetic acid with a flow rate of 10 ml/ min. The second chromatography was done on a cation-ex- change column (ES-502C, 7.6 mm internal diameter x 10 cm, Asahi Chemical Industry Co., Ltd, Kawasaki, Japan) with the solvent system of a linear gradient of 0 - 500 mM NaCl in 50 mM ammonium formate pH 4.5, with 20% (by vol.) CH3CN with a flow rate of 2 ml/min. The final chromatography was done under the same conditions used in the first RP column chromatography.

Amino acid analysis was carried out on an amino acid analyzer (A-8700, Irica, Kyoto, Japan) after hydrolysis of the peptides in 6 M HCl at 3 10 “C for 20 h. Sequence analysis was carried out on a protein sequencer (model 477A4, Applied Biosystems) employing the Edman degradation method. Pu- rity and the retention time of each peptide in RP-HPLC were checked on an analytical CI8 column (TSKgel ODS ~OTM, 4.6 mm internal diameter x 15 cm, Tosoh Co., Ltd) with the solvent system of a linear gradient of 20-50% CH,CN in 0.1% trifluoroacetic acid with a flow rate of 1 ml/min for 30 min.

Circular dichroism ( C D )

Peptides were dissolved to a final concentration of 10 kg/ ml in 20 mM Pipes and 150 mM NaC1, pH 7.4, without or

~~

Peptide Position of the reinserted Sequence lysine residue

~

eCT dK K12 K13 K14 K15 K16 K17 K19 K20 K21 K22 K23 K24 K25 K26 K27 K28 K29 K30 K31 K32

18

12 13 14 15 16 17 19 20 21 22 23 24 25 26 21 28 29 30 31 32

- CSNLSTCVLGKLSQELHKLQTY PRTDVGAGTP-NHZ CSNLSTCVLGKLSQELH-LQTYPRTDVGAGTP-NH, CSNLSTCVLGKKLSQELHLQTYPRTDVGAGTP-NH, CSNLSTCVLGKLKSQELHLQTYPRTDVGAGTP-NHZ CSNLSTCVLGKLSKQELHLQTYPRTDVGAGTP-NH2 CSNLSTCVLGKLSQKELHLQTY PRTDVGAGTP-NH 2 CSNLSTCVLGKLSQEKLHLQTYPRTDVGAGTP-NHZ CSNLSTCVLGKLSQELKHLQTYPRTDVGAGTP-NH, CSNLSTCVLGKLSQELHLKQTYPRTDVGAGTP-NH2 CSNLSTCVLGKLSQELHLQKTY PRTDVGAGTP-NH2 CSNLSTCVLGKLSQELHLQTKY PRTDVGAGTP-NHZ CSNLSTCVLGKLSQELHLQTYKPRTDVG AGTP-NH2 CSNLSTCVLGKLSQELHLQTY PKRTDVGAGTP-NHZ CSNLSTCVLGKLSQELHLQTYPRKTDVGAGTP-NH2 CSNLSTCVLGKLSQELHLQTYPRTKDVGAGTP-NH2 CSNLSTCVLGKLSQELHLQTYPRTDKVGAGTP-NHZ CSNLSTCVLGKLSQELHLQTYPRTDVKGAGTP-NH, CSNLSTCVLGKLSQELHLQTYPRTDVGKAGTP-NH2 CSNLSTCVLGKLSQELHLQTYPRTDVGAKGTP-NH 2

CSNLSTCVLGKLSQELHLQTYPRTDVGAGKTP-NH2 CSNLSTCVLGKLSQELHLQTY PRTDVGAGTKP-NH 2

CSNLSTCVLGKLSQELHLQTYPRTDVGAGTPK-NH,

609

with 50% trifluoroethanol. CD spectra were obtained with a Jasco model J600 spectropolarimeter (Japan Spectronic Co., Tokyo, Japan). The CD was measured from 240 nm to 195 nm in a 1-mm sample cell that was maintained at 25°C with a thermostatted cell holder. The results were expressed as the mean residue ellipticity (0). The contents of a-helix was calcu- lated by using a calculating program model SSE302 (Japan Spectronic Co.).

Celh and cell culture

LLC-PK, cells were obtained from the Japanese Cancer Research Resources Bank (JCRB, Tokyo, Japan). Cells were maintained in a polystyrene tissue culture flask containing Dulbecco’s modified Eagle’s medium supplemented with 10% (by vol.) heat-inactivated fetal bovine serum at 37°C in a humidified atmosphere of 5% C 0 2 in air. For subculture, cell monolayers were washed once with NaCl/P, consisting of 137 mM NaCl, 2.7 mM KC1,g.l mM Na2HP04, and 1.5 mM KH2P04, pH 7.2, and digested with 0.05% trypsin and 0.53 mM EDTA in NaC1/Pi for 2 rnin at 37°C.

Preprution of iodinated eCT

Radioactive eCT was prepared by the chloramine T method [22]. Purified eCT (25 pg) was dissolved in 200 pl 0.5 M sodium phosphate pH 7.5 and mixed with 200 pl of the buffer containing 74 MBq NaI2’I (New England Nuclear, Boston, MA). Then, 200 pl of 1 .O mg chloramine T/ml buffer was added and mixed continuously for 20 s. To stop the reac- tion, 200 pl of 2.5 mg sodium metabisulfite/ml buffer was added and left at room temperature for 15 s followed by the addition of 200 p15 mM KI in NaCl/P, containing 10% bovine serum albumin (albumin/NaC1/Pi). The mixture was applied on Bio-Gel P-2 column equilibrated with albumin/NaCl/P,. The radioactive eCT fractions were collected and the concen- tration was determined by the growth inhibition assay (see below). The specific activity of I2’I-eCT used in the binding experiments was 2.9 - 4 MBq/nmol.

To confirm the equipotency of eCT and iodinated eCTs, non-radioactive iodinated eCT was prepared by the same method mentioned above in the use of NaI instead of Na12,1. After the P-2 column chromatography, mono-iodo-eCT (I- eCT) and di-iodo-eCT (I2-eCT) were purified with RP-HPLC in the solvent system described above. The chromatogram by HPLC using an analytical column showed three major peaks. The molecular masses of the contents of these peaks were 3415.1, 3540.2, and 3666.1 Da in the order of elution, which was determined with a system of RP-HPLC combined with a mass spectrometer (LC/MASS; JMS-SX102, Jeol, Akishima, Japan). These values corresponded to eCT, I-eCT, and 12- eCT, respectively. In addition, a negatively charged ion of 127 Da, i.e. the iodic ion, was detected in two slower-eluted peaks as expected. The biological activities of purified I-eCT and I,-eCT determined by the growth inhibition assay and the cAMP production assay were identical to those of eCT.

Receptor binding assay Confluently grown LLC-PK, cells were digested with

0.05% (massivol.) trypsin and 0.53 mM EDTA in NaCl/P, for 1.5 rnin at 37°C. Cells were washed twice with fresh medium and once with the assay buffer (15 mM Tris/HCl, 120 mM NaCl, 4 mM KCl, 1.6 mM MgSO,, 2 mM NaH2P04, 10 mM glucose, and 0.1 YO mass/vol. bovine serum albumin, pH 7.4)

[23]. Cell pellets were resuspended with the assay buffer at the cell density of 1.6 x 107/ml. Cell suspension (100 pl) was mixed with an equal volume of I2’I-eCT (approximately 100000 cpm) and incubated at 0°C for 60 min. The mixture was layered on 1 ml 1 M sucrose and centrifuged at 12000 g for 2 min. The supernatant was removed and the radioactivity of cell pellets was counted with a y-counter (1282 Compugamma CS, LKB, Turku, Finland). Specific binding was calculated according to the formula: radioactivity of specific binding = radioactivity of the 12’I-eCT bound - radioactivity of the 12’1-eCT bound in the presence of 1 pM unlabeled eCT. Specificity of the binding was confirmed by the failure of displacement of ”’I-eCT binding with 1 pM unrelated peptides such as [Arg8]vasopressin, bovine insulin, human parathyroid hormone 1 - 34, human atrial natriuretic peptide, adrenocorticotropic hormone 1 - 24 and glucagon. Under this experimental condition, the dissociation constant of receptor binding (Kd) and the number of receptors/cell were 2 . 4 4 ~ M and 3.76 x lo4, respectively. Receptor binding ability of eCT analogs was determined by the inhibition of 12’I-eCT binding in duplicates and expressed as 50% inhi- bition dose (IDSO), the concentration inhibiting 50% specific binding.

Growth inhibition assay

Cell growth of LLC-PK1 was monitored by a dye-staining method as described previously [24] with small modifications [7]. LLC-PK1 cells (3 x lo3) in 200 p1 DMEM with 10% (by vol.) fetal bovine serum were added to each well of a 96-well flat-bottomed microplate. After incubation at 37°C for 6 h, medium was replaced by the same volume of a peptide solution serially diluted with medium in quadruplicate and the culture was continued for a further 72 h. After incubation, the me- dium was decanted and the cells were fixed and stained with a staining solution consisting of 4.2% (by vol.) neutralized formaldehyde, 10% (by vol.) ethanol, and 0.05% (mass/vol.) crystal violet. The stained plates were washed with running water and dried in air. The dye was extracted with 100 p1 of 0.05 M NaH2P04 in 50% ethanol. The absorbance at 550 nm of each well was measured by a microplate photometer (MTP- 32, Corona Electric Co., Ltd, Hitachi, Japan). An A s s 0 of 1 for the control culture was equivalent to approximately 5.7 x 1 O4 cells determined with a hemocytometer. The absorbance at 550 nm of the control culture was defined as 100% growth. The activity of each peptide was expressed as ID,,, the concen- tration retarding 50% growth.

cAMP production assay

Cells (5-8 x 105/well) grown in a 24-well plate for 48 h were washed once with NaC1/Pi. The samples diluted with 0.5 ml NaC1/Pi containing 1 mg bovine serum albumin/ml, 4 mM MgC12, and 2 mM iBuMeXan were added in triplicate and incubated at 37°C for 20 min. Reaction was stopped by placing the plates on boiling water for 5 min. The contents of the wells were frozen and thawed three times. cAMP concen- tration was measured by the radioimmunoassay method with a Yamasa CAMP assay kit (Yamasa Shoyu, Co., Choshi, Japan) and calculated from a standard curve. The cAMP amount produced by 0.1 pM eCT was defined as 100%. The activity of each peptide was expressed as 50% effective dose (EDSO), the concentration evoking 50% production of CAMP.

610

10.' 10" 10.' lo6 10" 10.' 10.' 10" lo-li lo'io

[Ligand] (M)

Fig. 1. Dose/response curve of biological activities of eCT and dK measured in LLC-PK1 cells. The activities of eCT (0 ) and dK (0) in receptor binding (A), growth inhibition (B), and cAMP production (C) were determined by the methods described under Materials and Methods. The values defined as 100% were 8000 cpm for A, ,4550 = 1.49 for B, and 2200 pmol/106 cells for C . Bars are SE

RESULTS

Biological activities of des-Lysl8 analog (dK) and a series of 20 analogs of eCT with the reinserted lysine residue

The in vitro biological activity of dK in LLC-PK1 was first determined by three assay methods: (a) inhibition of 1251-eCT binding; (b) growth inhibition and (c) cAMP production. dK showed inhibition of lz5I-eCT binding, growth inhibition, and stimulation of CAMP production in a dose-dependent manner (Fig. 1). ID50 of the receptor binding and growth inhibition, and ED50 of the cAMP production of dK were 7.1 x M, 1.32 x l o p 6 M, and 1.11 x M, respectively, while those of eCT were 2 . 2 ~ 1 0 - ~ M, 4.8xlO-"M, and 6.81 x lo-'' M, respectively (Table 2). dK was approximately 320-, 2708-, and 16.1-fold less active in abilities for the binding, growth inhibition, and the production of CAMP, respectively, than eCT. Thus, deletion of Lysl8 induced an extreme de- crease in the eCT activity.

Since the importance of Lysl8 for the biological activity was suggested, we synthesized a set of analogs (K12-K32) with a reinserted lysine residue. These analogs also showed biological activities in a dose-dependent manner (data not shown). Activities of all analogs expressing ID50 and ED50 determined by a dose/response curve of each analog are shown in Table 2. None of the analogs showed higher activity than eCT detected in either the binding or the growth inhibition assay. On the other hand, in the cAMP production some analogs such as K19, K23, and K27 showed activity similar to or higher than eCT. The relationship between the position of the reinserted lysine residue and the biological activities of analogs is shown in Fig. 2. Three categories were recognized according to the pattern of activities. K17 - K24, belonging to the first group, showed slightly lower activities than eCT in the receptor binding and the growth inhibition and similar activity in the ability of the cAMP production. The second group, K26 - K32, showed considerably lower activities both in the receptor binding and growth inhibition and slightly lower activities in the cAMP production. K12 - K16 and K25, the third group, showed the lowest activities. These results indicate that both regions of positions from 12 to 16 or 17

Table 2. The ID5o and ED50 values of peptides in three assay method

Peptide IDSO ED50

binding growth cAMP inhibition inhibition production

M

eCT 2.2 x 10-8 4.8xIO-'O 6 . 8 ~ 1 0 - ' ~ K12 2.1 x 10-6 2.9 x 5.5 x 10-8 K13 9.1 x 10-6 > 1 . 0 ~ I O - ~ 4.1 10-7 K14 5.2x 10-6 > 1.ox 10-5 3.ox 10-7

K16 > 1.0 x 10-5 1.0 x 10-6 5.2 x 10-9 K17 3.8 x lo-' 1.3 x lo-' 1.2 10-9

3.7 x 10-9 1.2 10-9 K21 3.2 x 10-7 3.2 x 1.8 x 10-9 K22 3.0 x 1 0 - ~ 2.2x 1 0 - 8 1.2 10-9

K24 4.2 x 10-7 1 . 5 ~ 1 0 - ~ 1.5 x 10-9

K26 > 1 . 0 ~ 10-5

K29 4.1 x 1.7 x lo-' 1.1 x 10-9 K30 > 1.0 x 10-5 3.9 x 10-6 8.7 x 10-9 K31 > 1.0 x 10-5 1.2 x 10-6 8.5 x 10-9

K15 > i .ox 10-5 > i . 0 ~ 1 0 - 5 8.8 x lo-'

K19 2.7 10-7 3.1 x 1 0 - ~ 5.1 x 10-10 K20 8.3 x

K23 4.1 10-7 5.4 x 10-8 8.1 x lo-''

K25 > 1.oX10-5 2.ox10-6 > 1.ox10-5 2.4 x lo-'

K27 1.7 x 8.9 x 8.2 x 10- lo

1.0 x 10-8 K28 > i .ox 10-5 1.3 x

7.7 x 10-6

K32 > i .ox 10-5 2 . 7 ~ 3.3 x 10-8

and from 24 or 25 to 32 of eCT, at least, are required for expressing the activity, especially in receptor binding and growth inhibition.

To obtain the relationship among the three assays, corre- lations between the relevant two assays were examined using the values obtained from the results shown in Table 2. The correlation coefficient of activities between the binding and the growth inhibition was 0.919 (Fig. 3A), indicating that there is a good correlation between the receptor binding and the growth inhibition of LLC-PK1. Those between the cAMP

61 1

0-

B b -7.0 - Q L

k0.663 0 E

-8.0 -

-5.0

r, h

8 Q

-6.0 Y

0) 0 - s a 2 P) -7.0 c m C .- .- m

-8.0

Peptide Fig. 2. Relationship between theposition of the reinserted lysine residue and biological activities. The values, shown in Table 2, of 50% inhibi- tory or effective doses for receptor binding (O) , growth inhibition (0) and cAMP production (U) were plotted

1

A ' V -5.0 -

0) -0 -7.0 -

r=O.919 -10.0 ,

-5.0

-6.0

-7.0

-8.0

k0.752

-10.0 -8.0 -7.0 -6.0 -5.0 -10.0 -9.0 -8.0 -7.0 -6.0 -5.0 -10.0 -9.0 -8.0 -7.0 -6.0 -5.0

cAMP production, log (EDw)/M Receptor binding, log (ID5o)IM cAMP production, log (ED5o)/M

Fig. 3. Correlation between 50% inhibitory or effective doses of eCT and its analogs. Values shown in Table 2 were plotted: (A) receptor binding vs growth inhibition; (B) receptor binding vs cAMP production and (C) growth inhibition vs CAMP production. The correlation coefficients (r) are shown in each figure

production and either the growth inhibition or the receptor binding were 0.752 and 0.663, respectively (Fig. 3B, C), in- dicating low correlations.

Table 3. CD of eCT and its analogs Peptides (10 pg/ml) were measured in the absence or presence of 50% trifluoroethanol (CF3CH20H) at 25°C in 20 mM Pipes and 150 mM NaCI, pH 7.4. n. t. = not tested

Physicochemical properties of peptides

Ellipticities of eCT and selected analogs obtained from CD spectra were determined without or with 50% trifluoroethanol (Table 3). All the peptides, including eCT, showed increased ellipticity in the presence of trifluoroethanol compared to without. Calculated contents of a-helix of eCT, K13, K17, K19, and K25 were 51.1%, 6l.2%, 52.7%, 74.2% and 37.2%, respectively, indicating that all the tested analogs except K25 had enough a-helix content. Since it has been suggested that a non-amphiphilic peptide shows shorter retention time than an amphiphilic peptide in the analytical RP-HPLC if they have the same amino acid composition and length [25], we next measured the retention times of eCT and its analogs. K12 - K16, which were relatively less active (Table 2), showed shorter retention times (12.5- 13.5 min) than eCT (19.8 min; Fig. 4). The retention times of other analogs were similar to eCT. These results indicate that the reduced extent of the

Peptide Position of the -[0]222 reinserted lysine residue -CF3CH20H + CFJCH20H

deg cm2 dmol- '

eCT 18 3942 15 897 K13 13 n. t. 19463 K17 17 3819 17719 K19 19 663 19 724 K25 25 1119 11 222

amphiphilicity leads to the decreases in activity. To analyze amphiphilicity further, Edmundson helical wheel plot analysis in various lengths of residues was performed on eCT and all the analogs. Only when residues from position 8 to 17 were

61 2

plotted, the good correlation between the amphiphilicity and the activity was recognized in eCT and all analogs. The wheel plots of residues from positions 8 to 17 and 8 to 22 of selected peptides such as eCT, K13, K17, and K22 are shown in Fig. 5 as examples. Namely, the amphiphilicity of eCT and analogs with relatively high activity such as K17 and K22 was well maintained and that of an analog (K13) having low activity was destroyed (Fig. 5A).

DISCUSSION

We demonstrated that at least two regions (positions 8 - 17 and 24-32) of eCT are required for the in vitro biological activity in the porcine kidney cell line LLC-PK1. This was concluded by the study for the structure/activity relationship of eCT and its analogs using a newly developed method for designing analogs. It is the basis of this conclusion that the

8 0 c ._ w .

~ m b ~ ~ h ~ ~ o - ~ m t ~ w h r n r n o - ~ Y Y Y Y Y Y * Y Y Y Y Y Y Y Y Y Y Y Y Y Y

Peptide

- - - - - - * - N N N N N N N N ~ N " " "

Fig. 4. Relationship between the position of the reinserted lysine residue and extent of amphiphilicity of eCT and its analogs. The extent of amphiphilicity was determined by the retention time of analogs in analytical RP-HPLC

A a 8

11 16 11 16

9 9

10 17 10 17

insertion of a lysine residue into some important regions of the peptide should result in the loss of activity. The conclusion made is summarized and illustrated in Fig. 6.

The importance of basic residues of sCT had been pointed out by D'Santos et al. [19]. They employed substitutions of Arg at positions 11 and 18 and lysine at position 14. In agree- ment with their conclusion, the Lysl8 of eCT is one of the essential residues for the expression of the activity, because deleting this residue lead to marked loss of activity determined by three assay procedures (Fig. 1). Thus, we synthesized a set of lysine-reinserted analogs (K12 - K32). Two major regions, positions 12 - 16 and 25 - 32, were suggested to be fundamen- tal for the activity, since analogs with lysine reinserted at the positions from 12 to either 16 or 17 and 25 to 32 showed marked loss of activity (Table 2). On the other hand, the region from positions 17 to 24 does not bear any sequence specificity because the insertion of the lysine residue into this region did not result in any severe loss of activity.

The first region, including the positions from 8 to 16, may contribute to keep the amphiphilic a-helix structure, which is essential for preserving the CT activity [8, 10-121. This speculation was supported in part by evidence that retention times of analogs K12-Kl6 on an analytical RP-HPLC, a parameter of content of the amphiphilicity [25,26], were much shorter than others except K25 (Fig. 4). However, the a-helical content of K13 was sufficient compared to that of eCT (Ta- ble 3). These data are readily expected because lysine is a strong producer of a-helix [27]. Thus, it is likely that analogs K12 - K16 contain sufficient a-helix structure. These indicate that the attenuation of eCT activity in analogs K12 -K16 was due to the break of their amphiphilicity. The amphiphilicity revealed by the wheel plot of the region at the position from 8 to 22 of eCT and a set of analogs is not coincident with in vitro activity determined by the receptor binding and the growth inhibition assays (Fig. 5B, Table 2, unpublished re- sults), whereas the wheel plot of the regions at the position from 8 to 17, i.e. decapeptide portion, is completely coincident

B R i q 8 19

11 16 11 16

9 18 9 18

20 20

a a 8 19 8 19

I1 11

18

10 17 10 17

Fig. 5 . Wheelplots ofportions of eCT and typical analogs. Hydrophobic and hydrophilic residues are expressed as closed and shadowed circles, respectively. Fragments between (A) 8 and 17 and (B) 8 and 22 were plotted

613

- Amphiphilic a-helix

iF%/ i n ---/ 15

30

Receptor binding site Fig. 6. Schematic illustration o j eCT showing two regions associated with biological activity. In the amphiphilic cc-helix structure, hydro- phobic and hydrophilic amino acids are represented as closed and shadowed circles, respectively

with the activity pattern (Fig. 5A, Table 2, unpublished re- sults). Any set of wheel plots of longer regions than the deca- peptide is not coincident with the activity pattern. Therefore, the length of amphiphilic a-helix required for the expression of activity must be at most the decapeptide portion (between positions 8 and 17; Fig. 6).

The receptor binding region of eCT has been demonstrated to be located on the fragment 22-32 by studying various fragment peptides of eCT [17]. Further, calcitonin-gene-re- lated peptide (CGRP), a peptide alternatively processed from the calcitonin gene, is known to bind the calcitonin receptor of LLC-PK1 [16]. CGRP and ultimobranchial CTs must share a common sequence as the receptor-binding region. The highly similar sequence of CGRP with eCT and sCT is Thr30-Asn31- Va132-Gly33-Ser34, of which counterparts in eCT and sCT are Thr25-Asp26-Val27-Gly28-Ala29 and Thr25-Asn26- Thr27-Gly28-Ser29, respectively. Thus, this pentapeptide por- tion may be the determinant for the specificity of ligand- receptor binding in eCT, sCT, and CGRP. In our study, serial analogs with the reinserted lysine at the position from 25 to 32 had weaker activities than eCT. Thus, we initially specu- lated that the receptor binding region must be located at the position from 25 to 32. However, the possibility cannot be excluded that the receptor binding region begins from Arg24. The insertion of lysine into the position 24 in the analog K24 might result in the similar sequence with the native one, i.e. Arg24-Thr25-Asp26-Va127 in eCT and Lys24-Thr25-Asp26- Val27 in K24, because both arginine and lysine are basic amino acids. Thus, we could not determine the precise length of the receptor binding region by the present method. However, we recently synthesized several deleted analogs of eCT to define the role of the middle region of eCT in LLC-PK1. An analog with residues deleted at the positions from 18 to 23 showed similar ability to dK in binding and growth inhibition and an analog with residues deleted from 18 to 24 showed much lower activity, indicating that Arg24 is required for the activity, especially binding to LLC-PK1 [28]. Therefore, we concluded that the receptor binding region of eCT is residues from 24 to 32.

Analogs with lysine-reinserted at positions 17 and 19 to 24 (K17, K19-K24) had similar ability to eCT in cAMP production and slightly decreased activities in the receptor binding and the growth inhibition. Further, lysine insertion into this region did not affect the conformation of the whole molecule determined by CD spectra and retention times in HPLC. The region from 18 to 23 seems to bear neither sequen-

tial specificity for the biological activities nor contribution to the maintenance of the conformation of eCT. The basic or hydrophilic property of the lysine residue in this region may contribute to the activity, though it is unclear whether this is due to the presence of a lysine residue itself or the peptide length. However, this spacing region might be essential for expressing full biological activity, because deletion of this region caused a significant decrease in the eCT activity as low as dK [28].

Three assay procedures of receptor binding, growth inhi- bition, and cAMP production were employed to evaluate the eCT activity. There is good correlation between ID50 values of the receptor binding and the growth inhibition ( r = 0.919), indicating that growth inhibition is the result of direct events involving receptor binding. On the other hand, in the case of cAMP production, there was less correlation to the others. Some analogs, such as K17, showed the dissociated activities among these assays. K17 had a relatively high capability for cAMP production, but was low in the others. These results indicate that regions on the molecule required for each activity must be different and cAMP is not mainly involved in the growth inhibition. It supports conclusion in our previous study that the major signalling pathway for growth inhibition of LLC-PK1 by ultimobranchial CTs is other than adenylate cyclase system [7].

We employed a newly developed method for designing analogs to clarify the structure/activity relationship in a given peptide. In the case of peptides bearing amphiphilic a-helix structure such as calcitonin, this method is especially effective because only amphiphilicity can be destroyed without changing amino-acid composition, peptide length, and the whole conformation. This method may be useful not only for studying the structure/activity relationship but also for designing a novel analog of many other peptides.

The authors are very grateful to Drs T. Suda and N. Takahashi (School of Dentistry, Showa University), Dr S. Ishii (Faculty of Edu- cation, Waseda University) and Dr S. Kidokoro, for helpful dis- cussions. We also thank K. Matsuura and K. Kijima of the Jeol and S. Mohara for the technical assistance in HPLC/mass spectrometry of iodinated eCT.

REFERENCES 1. Potts, J. T. Jr & Aurbach, G. D. (1976) Handb. Physiol. Sect. 7 :

Endocrinol. 7,423 -430. 2. Gennari, C., Chierichetti, S. M., Vibelli, C., Francini, G., Maioli,

E. & Gonnelli, S. (1981) Curr. Ther. Res. 30, 1024-1032. 3. Homma, T., Watanabe, M., Hirose, S., Kanai, A,, Kangawa,

K. & Matsuo, H. (1986) J . Biochem. (Tokyo) 100,459-467. 4. Otani, M., Noda, T., Ydmauchi, H., Watanabe, S., Matsuda,

T., Orimo, H. & Narita, K. (1974) in Proceedings of the 5th Parathyroid Conference (Talmage, R. V., Owen, M. & Parson, J. A., eds) p. 11 1, Oxford. Excerpta Medica, Amsterdam.

5 . Goldring, S. R., Dayer, J.-M., Ausiello, D. A. & Krane, S. M. (1978) Biochem. Biophys. Res. Commun. 83,434-440.

6. Jam, D. A,, Gajdas, E. L., Dierks-Ventling, C., Hemmings, B. A. & Fahrenholz, F. (1987) Biochim. Biophys. Acta 930, 392- 400.

7. Inoue, A., Komatsu, Y., Ochiai, J., Itagaki, S., Nishide, H., Shikano, M., Hemmi, H. & Numao, N. (1990) Cell Biol. Int. Rep. 14, 887 - 896.

8. Epand, R. M., Epand, R. F., Orlowski, R. C., Schlueter, R. J., Boni, L. T. & Hui, S. W. (1983) Biochemistry 22, 5074- 5084.

9. Brewer, H. B. & Edelhoch, H. (1970) J . Bid. Chem. 245, 2402- 2408.

614

10. Moe, G. R., Miller, R. J. & Kaiser, E. T. (1983) J . Am. Chem.

11. Moe, G. R. & Kaiser, E. T. (1985) Biochemistry 24, 1971 -1976. 12. Green 111, F. R., Lynch, B. & Kaiser, E. T. (1987) Proc. Natl

Acad. Sci. USA 84, 8340- 8344. 13. Epand, R. M., Seyler, J. K. & Orlowski, R. C. (1986) Eur. J .

Biochem. 159, 125-127. 14. Epand, R. M., Epand, R. F., Orlowski, R. C., Seyler, J . K. &

Colescott, R. L. (1986) Biochemistry 25, 1964- 1968. 15. Nicholson, G. C., Moseley, J. M., Sexton, P. M., Mendelsohn,

F. A. 0. & Martin, T. J. (1986) J . Clin. Invest. 78, 355-360. 16. Wohlwend, A,, Malmstrom, K., Henke, H., Murer, H., Vassalli,

J.-D. & Fischer, J. A. (1985) Biochem. Biophys. Res. Commun.

17. Yamamoto, I., Morita, R., Fukunaga, M., Dokoh, S., Shigeno, C., Torizuka, K. & Noda, T. (1981) Endocrinology 108, 698- 702.

18. Epand, R. M., Stahl, G. L. & Orlowski, R. C. (1986) Int. J . Peptide Protein Rex 27, 501 - 507.

19. D’Santos, C. S., Nicholson, G. C., Moseley, J . M., Evans, T., Martin, T. J. & Kemp, B. E. (1988) Endocrinology 123, 1483- 1488.

SOC. 105,4100-4102.

131,537-542.

20. Hull, R. N., Cherry, W. R. &Weaver, G. W. (1976) In vitro 12,

21. Barany, G. & Merrifield, R. B. (1979) in The peptides (Gross, E. & Meienhofer, J., eds) vol. 2A, pp. 1 -284, Academic Press, New York.

22. Hunter, W. M. & Greenwood, F. C. (1962) Nature 194, 495- 496.

23. Findlay, D. M., Michelangeli, V. P., Eisman, J . A., Frdmpton, R. J., Moseley, J. M., MacIntyre, I., Whitehead, R. & Martin, T. J. (1980) Cancer Res. 40,4764-4767.

24. Hemmi, H., Nakamura, T., Tamura, K., Shimizu, Y., Kato, S., Miki, T., Takahashi, N., Muramatsu, M., Numao, N. & Sugamura, K . (1987) J . Immunol. 138,664-666.

25. Zhou, N. E., Mant, C. T. & Hodges, R. S. (1990) Peptide Res. 3,

26. Heinitz, L. M., Flamigan, E., Orlowski, R. C. & Regnier, F. E.

27. Chou, P. Y. & Fasman, G. D. (1978) Adv. Enzymol. 47,45-148. 28. Shikano, M., Inoue, A., Komatsu, Y., Hemmi, H., Tunemoto.

D. & Numao, N. (1991) in Peptide chemistry 1990 (Shimonishi, Y., ed.) pp. 363 - 366, Protein Research Foundation, Osaka.

670 - 677.

8-19.

(1988) J . Chromatogr. 443,229-245.