The insulin A and B chains contain structural information for the formation of the native molecule

Transcription

1 Biohem. J. (199) 268, (Printed in Great Britain) The insulin A and B hains ontain strutural information for the formation of the native moleule Studies with protein disulphide-isomerase Jian-Guo TANG and Chen-Lu TSOU* nstitute of Biophysis, Aademia Sinia, Beijing 18, China 429 t has been shown previously r[ang, Wang & Tsou (1988) Biohem. J. 255, ] that, under appropriate onditions, native insulin an be obtained from srambled insulin or the S-sulphonates of the hains with a yield of 25-3 %, together with reation produts ontaining the separated A and B hains. The native hormone is by far the predominant produt among the isomers ontaining both hains. t is now shown that the presene of added C peptide has no appreiable effet on the yield of native insulin. At higher temperatures the ontent of the native hormone dereases whereas those of the separated hains inrease, and in no ase was srambled insulin ontaining both hains the predominant produt in the absene of denaturants. Both the srambling and the unsrambling reations give similar h.p.l.. profiles for the produts. Under similar onditions ross-linked insulin with native disulphide linkages an be obtained from the srambled moleule or from the S-sulphonate derivative with yields of 5 % and 75 % respetively at 4 C, and with a dilute solution of the hexa-s-sulphonate yields better than 9% an be obtained. The regenerated produt is shown to have the native disulphide bridges by treatment with CNBr to give insulin and by the identity of the h.p.l.. profile of its pepti hydrolysate with that for ross-linked insulin. t appears that the insulin A and B hains ontain suffiient information for the formation of the native moleule and that the role of the onneting C peptide is to bring and to keep the two hains together. NTRODUCTON t was suggested some years ago that, among all the possible isomeri strutures, the native insulin struture is the most stable, on the basis of a reasonably good yield of the hormone when the redued A and B hains were reoxidized together (Du et al., 1961, 1965). The suessful resynthesis has sine not only been onfirmed in different laboratories (Zahn et al., 1966; Katsoyannis et al., 1967) but has also been applied to the prodution of human insulin from the hains obtained separately by a reombinant DNA tehnique (Chane et al., 1981). However, both the fat that insulin an be obtained from its hains in reasonably good yield and the related suggestion that the insulin A and B hains ontain suffiient information for the formation of the native struture have not met with general reognition (Anfinsen & Sheraga, 1975; Brandenburg et al., 1977; Hillson et al., 1984; Lehninger, 1984; Alberts et al., 1989). This is probably partly due to the fats that previous workers failed to obtain native hormone from 'srambled' insulin with randomly joined disulphide bonds through thiol-disulphide exhange reations atalysed by protein disulphide-isomerase (PD) and that treatment of insulin itself by this enzyme led to the destrution of native insulin struture (Givol et al., 1965; Varandani & Nafz, 197). Varandani & Nafz (197) also reported that some insulin was obtained from srambled moleules in the presene of added C peptide, suggesting that strutural information was onferred on the whole moleule by this peptide. n a previous paper we showed that, in the presene of PD and dithiothreitol (DTT) at ph 7.5 and 4 C, native insulin an indeed be formed from srambled insulin in 25-3 % yield (Tang et al., 1988). t is now shown that the yield dereases with inreasing temperature and that the addition of C peptide has no appreiable effet. The derease in insulin yield is due to the formation of produts ontaining only one of the hains, and in no ase were produts with isomeri strutures ontaining both hains the predominating produts for both the srambling and unsrambling reations. With N2A1NeB29-ross-linked insulin, moleules ontaining native disulphide bonds an be obtained in better than 9 % yield from the hexa-s-sulphonate derivative. t is suggested that the sequenes of the A and B hains provide the neessary information for the formation of the native insulin struture. EXPERMENTAL Bovine insulin was from Sigma Chemial Co. Srambled insulin, the S-sulphonates of the hains and PD were prepared as desribed previously (Tang et al., 1988) exept that srambled insulin was obtained by oxidation of the redued hains in 6 M- guanidinium hloride and had a residual ativity 1-2 % of that of native insulin. C peptide was a syntheti produt kindly provided by Professor Niu of the Shanghai nstitute of Biohemistry, Aademia Sinia, Shanghai, China (Niu et al., 1981). CNBr and pepsin with a speifi ativity of 1 munits/mg were from Merk. Carbonylbis-(L-methionine p-nitrophenyl ester) was a produt of Piere Chemial Co. Other reagents were loal produts of analytial grade used without further purifiation. CBM-insulin, ross-linked at the a- and e-amino groups of Gly-Al and Lys-B29 residues respetively, was prepared in aordane with Busse & Carpenter (1976) exept the olumn hromatography on DEAE-ellulose was arried out twie. t had a reeptor-binding ativity 1.8 % of that of native insulin, and an amino aid analysis showed the presene of the expeted amount of methionine. Srambled CBM-insulin was prepared by dissolving CBM-insulin (15 mg; final onentration 5.2 mm) in Abbreviations used: CBM-insulin, [N AlN6B29-arbonylbismethionyl-insulin; DTT, dithiothreitol; PD, protein disulphide-isomerase. * To whom orrespondene should be addressed.

2 43.5 ml of.1 M-Tris/HCl buffer, ph 8. 1, ontaining 2 mm-edta and 6 M-guanidinium hloride and redued with 32 mm-dtt for 4 min at 37 'C. The solution was dialysed against 5 mm- NH4HCO3 to remove exess DTT and then diluted to 3 ml with.1 M-glyine/NaOH buffer, ph 1.8, ontaining 6 M- guanidinium hloride followed by air oxidation at 15 'C for 24 h. The produt was then dialysed against 5 mm-aeti aid and freeze-dried. Srambled CBM-insulin thus prepared ontained 5 % CBM-insulin, whih is also approximately the amount expeted from random joining of the thiol groups. The S- sulphonate of CBM-insulin was prepared by sulphitolysis in 1 ml of.1 M-Tris/HC buffer, ph 7.5, ontaining 1 mm-edta and 6 M-guanidinium hloride for 1 h at 37 'C. The reation mixture ontained 2.6 mm-cbm-insulin, 178 mm-na2so3 and 56 mm- Na2S46. t was then dialysed against 5 mm-nh4hco3 and freeze-dried. Conditions for the PD-atalysed reations were as desribed previously (Tang et al., 1988). Unless otherwise speified, the PD-atalysed reations were arried out with 13 jm-insulin or its derivatives, 6.4 4uM-PD and the speified onentration of DTT in.1 M-Tris/HC buffer, ph 7.5, ontaining 1 mm-edta at 4 'C for 24 h. The partial inativation of insulin and destrution of CBM-insulin struture in the presene of PD is referred to as the srambling reation, and the regeneration of native insulin from either the srambled preparation or the S-sulphonates of the hains and the formation of CBM-insulin with the native disulphide linkages from a srambled preparation or its S-sulphonated derivative by PD are simply desribed as the unsrambling reation. Digestion of insulin and derivatives with pepsin was arried out by dissolving the protein at a onentration of 1 mg/ml in.1 M-HC and hydrolysing with.2 mg of pepsin/ml for 24 h at 37 'C, at the end of whih the initial material had ompletely disappeared from the h.p.l.. profile. CNBr treatment of CBM-insulin was arried out in aordane with Busse & Carpenter (1976) by treatment of.2 mg of protein in 5,ul of 7 % (v/v) formi aid with 1 mg of CNBr at 25 'C for 2 h. The ontent of regenerated insulin in the produt was analysed by a reeptor-binding assay. The reverse-phase h.p.l.. analysis of insulin and derivatives was arried out as desribed previously (Wang & Tsou, 1986; Tang et al., 1988) with a Waters Assoiates h.p.l.. system. The insulin or CBM-insulin ontents of the reation produts were determined by integration of the respetive h.p.l.. peaks. For the h.p.l.. analysis of the pepti hydrolysate, the same reverse-phase olumn was used but with a solvent of aetonitrile in aq..1 % (v/v) trifluoroaeti aid and a gradient (urve 7 on the Waters instrument) from to 7% (v/v) aetonitrile in 35 min. The eluate was monitored at 22 nm with a Hitahi tunable-wavelength effluent monitor. nsulin ontents of the reation produts were also determined by ompetition of reeptor binding with 1251-insulin (Mori & Wond, 1984) and by the stimulation of lipogenesis (Moody et al., 1974). RESULTS Amount of DTT required for the unsrambling reation t was shown in a previous paper (Tang et al., 1988) that the presene of DTT is neessary for the unsrambling of insulin. Figs. 1 and 2 show the effet of DTT onentration on the yield of native insulin for srambled insulin and from the S-sulphonates of the hains respetively. The insulin ontents of the reation produts were determined from the h.p.l.. profile and by biologial assay. t is to be noted that for the unsrambling of srambled insulin a atalyti amount of DTT is enough for the rearrangement of the disulphide bonds, but for the S-sulphonates of the hains lose to the stoihiometri amount of DTT for the C., -3 n 3 2 _ J.-G. Tang and C.-L. Tsou DTT/disulphide ratio Fig. 1. Effet of DTT onentration on the yield of native insulin from the srambled preparation For experimental onditions see the text. following reation (Frank et al., 1981) is required to produe the optimal yield: R(SH)2 + R'(SSO3-)2-+ R(S)2 + R'(S)2 + 2 HSO3- Effet of temperature on the unsrambling reation The h.p.l.. profiles of the produts for the unsrambling of srambled insulin at different temperatures are shown in Fig. 3, showing the derease in peak height for native insulin with inreasing temperature. The unsrambling of an equimolar mixture of the A hain and B hain S-sulphonates at a thiol/ssulphonate ratio of 1.25 and a final DTT onentration of 375 j#m produed h.p.l.. profiles losely similar to those shown in Fig. 3 at the respetive temperatures. The identities of the respetive peaks have been previously haraterized (Tang et al., 3 a, o :: lo _- ~~ l l l l Thiol/S-sulphonate ratio Fig. 2. Effet of DT onentration on the yield of native insulin from a mixture of the S-sulphonates of A and B hains The molar ratio of A and B hains was 1:1 and the final protein onentration was.6 mg/ml. For other experimental onditions see the text. 199

3 A and B hains have strutural information for formation of insulin 431 (a) 2 \ (b) ()J )(d))j Time (min) Fig. 3. H.p.l.. profiles at different temperatures for the unsrambling of srambled insulin As identified in a previous paper (Tang et al., 1988), peaks 1-5 indiate respetively the positions of oxidized A hain, inative produts ontaining both hains joined presumably with inorret disulphide linkages, native insulin, oxidized B hain and PD. The reation mixture ontained 17 /SM- DTT and srambled insulin with a DTT/disulphide ratio of.55. The h.p.l.. profiles of unsrambling produts obtained at 4, 15, 26 and 37 C are shown in (a), (b), () and (d) respetively. 1988). t is to be noted that with the derease in the peak for native insulin at higher temperatures the peaks for the separate hains inrease, but in no ase is the peak orresponding to produts ontaining both hains higher than the insulin peak. The derease in yield of insulin is therefore not due to the inrease in the formation of A.B. isomers with inorret disulphide bridges, whih should be eluted at peak 2, immediately preeding peak 3. Further, lowering the temperature to -5 C in 2% (v/v) ethylene glyol did not inrease the yield of the unsrambling reation, nor had altering solvent polarity with.5 M-KC any effet. Treatment of insulin with PD The h.p.l.. profiles of the produts of the srambling of insulin at different temperatures are losely similar to those of the unsrambling reation shown in Fig. 3. The yield of insulin at 37 C is in agreement to that reported by Givol et al. (1965) and by Varandani & Nafz (197). However, like the unsrambling reation, there is no signifiant formation of produts ontaining both hains. With inreasing temperature the gradual derease in the insulin peak was aompanied by inreases in the peaks for the separate hains, but not the peak at the position expeted for the AP, isomers of insulin (Tang et al., 1988). Fig. 4 ompares the ontents of native insulin at different temperatures for both the srambling and the unsrambling reations. t is also lear from a omparison of the respetive h.p.l.. profiles (not shown) that, no matter whether one started from insulin, srambled insulin or the S-sulphonates of the hains, similar relative amounts of the final produts were obtained in the presene of DTT and PD. Effet of C peptide Varandani & Nafz (197) reported that no native insulin ould be obtained from srambled insulin unless the C peptide was added. The impliation is that the C peptide ontained strutural information that it somehow onferred on the A and B hains so as to promote the formation of the native hormone. However, in our hands the addition of an equivalent amount of C peptide (.35 mg/ml, 14 ftm) to srambled insulin (.6 mg/ml, 13 um), as used by Varandani & Nafz (197), had no appreiable effet on the yield of native insulin in the presene of PD and DTT at either 37 C or 4 C (results not shown). Unsrambling and srambling reations of CBM-insulin and its S-sulphonate Under appropriate onditions both srambled CBM-insulin and its hexa-s-sulphonate an be unsrambled with PD and DTT to give CBM-insulin with the orret disulphide bridges. The h.p.l.. profiles of the produts of the unsrambling reation from srambled CBM-insulin and the hexa-s-sulphonate are shown in Fig. 5. Addition of DTT alone generates 8.3 % and 4)._ ) 3 ~-- 2 _ 1 _ Temperature (OC) Fig. 4. Contents of native insulin at different temperatures for the srambling and unsrambling reations The experimental onditions for srambled insulin were as given in Fig. 3 legend. Conditions for the other reations were as desribed in the text. Curves represented by A-A, O-O and *-* are for reations starting from native insulin, srambled insulin and an equimolar mixture of the S-sulphonates of the A and B hains respetively.

4 432 J.-G. Tang and C.-L. Tsou (a) (b) () (d) (e) '. ' k- -- ' Time (min) Fig. 5. H.p.l.. profiles of the unsrambling produts from srambled CBM-insulin and CBM-insulin hexa-s-sulphonate The arrow indiates the positions of CBM-insulin with native disulphide bonds. Srambled produts with inorret disulphide linkages are eluted after CBM-insulin and are presumably more hydrophobi. (a) H.p.l.. profile of srambled CBM-insulin. (b) After reation of the srambling produts with DTT alone at a DTT/disulphide ratio of.16. () As (b), but after reation with both DTT and PD. (d) Reation produts of the S-sulphonate of CBM-insulin with 36,uM-DTT at a thiol/s-sulphonate ratio of 1.2. (e) As (d), but reation produts obtained in the presene of both DTT and PD. The seond peak in () and (e) is the peak for PD as a sharp peak at the end of the elution. 11.6% of CBM-insulin from srambled CBM-insulin and the hexa-s-sulphonate respetively, whereas in the presene of both DTT and PD yields of 5 % and 75 % were obtained respetively, as shown by the appearane of a peak at the position of authenti CBM-insulin, the identity of whih is demonstrated below. At high DTT onentrations the speies with native disulphide bonds dereases, probably as a result of the formation of partially redued produts. For the unsrambling from the hexa-ssulphonate derivative, the thiol/s-sulphonate ratio is optimal around 1.2, lose to the value required by the stoihiometry of the reation. The unsrambling of CBM-insulin is a faster reation than that for insulin and approahes ompletion within about.5 h at 4 'C. n ontrast with insulin, the unsrambling of CBM-insulin is favoured in dilute solutions, so as to avoid the formation of oligomers. n fat, for the dimeri moleule, the total number of possible isomers will be 5243, as alulated by the equation given by Wang et al. (1987). ndeed, with CBMinsulin hexa-s-sulphonate, when the protein onentration is lowered to.1 mg/ml yields better than 9 % an be obtained at 4 C. The addition of DTT and PD to CBM-insulin leads to a partial srambling of the moleule, resulting in a derease in the speies ontaining the orret disulphide bridges. The h.p.l.. profile of the produts is similar to that obtained in the unsrambling reation. Effet of temperature on the srambling and unsrambling reations of CBM-insulin The h.p.l.. profiles of the produts for both the unsrambling and the srambling reations of CBM-insulin, as well as the unsrambling reation of CBM-insulin hexa-s-sulphonate, at different temperatures were ompared. n all ases the yields of the produt orresponding to the speies ontaining the native disulphide bridges derease at higher temperature, as shown in Fig. 6. There was also an inrease of produts eluted at higher onentrations of methanol, these presumably being more hydrophobi in nature than CBM-insulin (results not shown). Comparison of the temperature effets on the srambling and unsrambling reations of insulin and ross-linked insulin learly shows (Figs. 4 and 6) that the native struture of CBM-insulin is less sensitive to temperature than is the orresponding struture of insulin. Fig. 6 also shows that CBM-insulin dereases at a, U Temperature ( C) Fig. 6. Effet of temperature on the srambling and unsrambling reations for CBM-insulin The srambling of CBM-insulin at DTT/disulphide ratios of.16 and.55 are shown in urves A-A and A- -A respetively. Unsrambling reations are shown by urves - and --O for the unsrambling of srambled CBM-insulin at DTT/disulphide ratios of.16 and.55 respetively. Curve - shows the unsrambling of CBM-insulin hexa-s-sulphonate with a thiol/ssulphonate ratio of 1.2. For other experimental onditions see the text. 199

5 A and B hains have strutural information for formation of insulin 433 (a) 7 9 (b) 9i () 1 (d) 5 1J Time (min) Fig. 7. H.p.l.. profiles of the produts of pepti digestion The h.p.l.. proffiles of the produts of pepti digestion of (a) insulin, (b) CBM-insulin, () srambled CBM-insulin and (d) regenerated CBM-insulin are shown. Regenerated CBM-insulin was obtained by olleting the h.p.l.. eluate at the position expeted for CBM-insulin from the reation produt of PD-treated CBM-insulin hexa-s-sulphonate. t was freeze-dried after removal of methanol by evaporation, redissolved and then digested by pepsin. For other experimental details and a disussion of the respetive peaks in the profile see the text. higher DTT/disulphide ratio for both the srambling and the unsrambling reations, probably as a result of the formation of partially redued produts. dentifiation of the unsrambling produt As CBM-insulin is muh less biologially ative than insulin (Busse & Carpenter, 1976; Brandenburg et al., 1977), ontamination by a small amount of insulin in the CBM-insulin preparation ould seriously interfere with determination of the ativity of the unsrambling produt. A fingerprinting tehnique and leavage by CNBr to regenerate insulin were used for the identifiation of the produt obtained from the unsrambling reations. The peak eluted at the expeted position of CBMinsulin during h.p.l.. separation was olleted and hydrolysed with pepsin, and the h.p.l.. profile of the produts was ompared with those for the hydrolysates of insulin, CBM-insulin and srambled CBM-insulin, as shown in Fig. 7. t is evident from a omparison of Figs. 7(a) and 7(b) that peaks 1 and 7 from the insulin hydrolysate disappeared and were replaed with a new peak, 5, in the profile for the pepti hydrolysate of CBM-insulin. This is as expeted, sine the C- and N-terminal peptides of B and A hains respetively would be joined by the ross-linking reagent. From the expeted sites of pepti leavage of insulin hains, peak 1 is very probably the relatively hydrophili peptide B-(26-3), peak 7 the hydrophobi peptide A-(1-4) and peak 5 the two peptide fragments onneted by the ross-linking reagent. The h.p.l.. profile of the peptide fragments of the srambled produt (Fig. 7) is onsiderably different from that obtained from CBM-insulin by virtue of the derease in the heights of peaks 8 and 9, whih are no doubt the peptides ontaining the native disulphide linkages, and the appearane of a number of new peaks, 6, 1, 11 and 12, whih are very probably peptides ontaining disulphide linkages formed during the srambling reation. t is possible that the leavage sites of the srambled moleule are not ompletely idential with those of CBM-insulin, beause of onformational hanges brought about the formation of disulphide linkages not present in CBM-insulin. t is to be noted that peaks 1, 11 and 12 are more hydrophobi than the peptides ontaining native disulphide linkages, as represented by peaks 8 and 9. That the prinipal unsrambling produt is indeed CBM-insulin is shown by the omplete identity of the h.p.l.. profiles of their pepti hydrolysates (Figs. 7b and 7d). The identity of the regenerated CBM-insulin has further been shown by the formation of insulin after treatment with CNBr as desribed by Busse & Carpenter (1976), with a yield similar to that obtained by these authors. DSCUSSON n spite of reports from different laboratories that good yields of insulin an be obtained when the redued A and B hains were reoxidized together (Du et al., 1961, 1965; Zahn et al., 1966; Katsoyannis et al., 1967; Chane et al., 1981), it has not been generally reognized that insulin an be obtained from its hains with a yield better than expeted by random formation of the disulphide bridges, and onsequently the suggestion has been made that the A and B hains ontain suffiient information for the formation of the native struture (Anfinsen & Sheraga, 1975; Brandenburg et al., 1977; Hillson et al., 1984; Lehninger, 1984; Alberts et al., 1989). Apparently, the alulation on the yield of random formation of the disulphide linkages made by some authors was based on the onsideration of the formation of the AAB1 isomers only (Pruitt et al., 1966). n fat, had the formation of the disulphide linkages been ompletely random, monomeri as well as oligomeri produts ontaining one or both of the hains ould be formed, and the total possible number of produts would be immense indeed, as has been alulated by Kauzmann (1959) and Chane et al. (1981) and more reently orreted for the number of A2 isomers by Wang et al. (1987). Completely random joining of the thiol groups to form disulphide linkages would result in a negligible yield of the native moleule, as has been shown by the oxidation of the redued hains in the presene of either 6 M-guanidinium hloride or.1 % SDS (Qian & Tsou, 1987). Givol et al. (1965) and Varandani & Nafz (197) reported that when insulin was treated with PD at 37 C the native struture was ompletely destroyed, and that, although it was possible to inrease the ontent of proinsulin in a preparation of the srambled moleule by treatment with PD from 12% to 37%, very little insulin was obtained from a srambled insulin prep-

6 434 aration ontaining about 2 % insulin under similar onditions. We have now shown that, under suitable onditions, 25-3 % insulin ould be obtained for both the srambling and unsrambling reations. The main differene between our experimental onditions and those of previous workers is a differene in temperature. As for the resynthesis of insulin from the redued hains, a temperature of 4 C was used, whereas both Givol et al. (1965) and Varandani & Nafz (197) used 37 'C. ndeed, we an onfirm the results obtained by these authors in.that we have shown that on inreasing the temperature to 37 'C the final yields of insulin derease markedly for both the srambling and the unsrambling reations. However, h.p.l.. analysis of the produts shows that, with inreasing temperature, the derease in native insulin among the reation produts is aompanied by an inrease, not of produts ontaining both hains, but of produts ontaining only one of the hains. n all ases the peak orresponding to produts ontaining both hains also dereases slightly with inreasing temperature. n fat Givol et al. (1965) did produe evidene suggesting that the preipitate formed during the srambling reation of insulin with PD ontained produts with only one of the hains. The addition of the onneting C peptide has been reported (Varandani & Nafz, 197) to inrease the yield of the unsrambling reation of srambled insulin by PD to about 12.8 % at 37 'C. However, in the present study the addition of a similar amount of C peptide was found to have no appreiable effet on the yield at either 4 'C or 37 C. Nevertheless with the two hains ross-linked with arbonylbismethionyl, a yield of 5% ould be obtained from the srambled moleule at 4 'C. Under arefully ontrolled onditions and with a dilute solution of CBM-insulin hexa-s-sulphonate, yields over 9% an be obtained. t is also to be realled that oxidation of the NAlNeBs_ ross-linked redued hains gives the native disulphide linkages in good yields (Brandenburg & Wollmer, 1973; Busse & Carpenter, 1976), as obtained for native proinsulin from the oxidation of redued proinsulin reported by Steiner & Clark (1968) and by Bullesbah et al. (198) or from the S-sulphonate derivative as reported by Frank et al. (1981). As our results and those reported by different authors were obtained with different hemial ross-linking reagents, it is most unlikely that the rosslinking moiety ontains any strutural information required to assist the orret pairing of the hains. The above results suggest strongly that the role of the onneting C peptide in proinsulin is limited to bringing and to keeping the A and B hains together during both the oxidation and the unsrambling reations. The neessary strutural information is provided entirely by the A and B hains themselves. The formation of the native disulphide bridges is favoured by low temperatures, as shown previously during the oxidation of the redued hains (Du et al., 1961, 1965; Katzen & Tietze, 1966). t was suggested in a previous paper (Tang et al., 1988) that, for the separate hains, low temperatures would derease the mobility of the hains in solution and onsequently inrease the opportunity for the hains to be orretly paired. This is likewise essential not only for the separately oxidized hains but also for the srambled moleules ontaining both hains, even though the hains ould be onneted together all the time during the rearrangements of the disulphide bridges. The latter onsideration applies espeially to the unsrambling of rosslinked insulin and its S-sulphonate derivative. The formation of the native disulphide linkages in the ross-linked moleules is less sensitive to an inrease in temperature. The fat that they are still sensitive to inreasing temperature suggests that some other fators are to be onsidered. The insulin hains are known to have the tendeny to polymerize (Lu & Tsao, 1962; Lu & Cui, 1981), espeially the B hain, and this dereases the formation 1 ~ ~ ~ ~ ~ ~ ~ ~ ~ J.-G. Tang and C.-L. Tsou of the A,B1 isomers. The tendeny towards self-assoiation of the individual hains would be dereased but probably not ompletely prevented by ross-linking the two hains, with the formation of oligomeri CBM-insulin, espeially at high onentrations, as evidened by the inrease in yield of the unsrambling reation in dilute solutions. t is not understood why low temperature would favour the pairing of A and B hains at the expense of oligomerization of the separate hains, unless different fores with different temperature oeffiients are responsible for the above two reations. We express our sinere thanks to Professor C.-. Niu of the Shanghai nstitute of Biohemistry, Aademia Sinia, for his gene)'ous gift of C peptide, Dr. C.-C. Wang for many helpful suggestions, Dr. Y.-M. Wang for the use of the Hitahi tunable-wavelength monitor and Dr. D. Saunders for a gift of the pepsin used. This work was supported in part by Researh Grant RO DK 3435 from the U.S. National nstitutes of Health. REFERENCES Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K. & Watson, J. D. (1989) Moleular Biology of the Cell, 2nd edn., p. 122, Garland Publishing, New York Anfinsen, C. B. & Sheraga, H. A. (1975) Adv. Protein Chem. 29, 25-3 Brandenburg, D. & Wollmer, A. (1973) Hoppe-Seyler's Z. Physiol. Chem. 354, Brandenburg, D., Gattner, H.-G., Shermutzki, W., Shuttler, A., Ushkoreit, J., Weimann, J. & Wollmer, A. (1977) in Protein Crosslinking, part A (Friedman, M., ed.), pp , Plenum Press, New York Bullesbah, E. E., Danho, W., Helbig, H.-J. & Zahn, H. (198) Hoppe- Seyler's Z. Physiol. Chem. 361, Busse, W. D. & Carpenter, F. H. (1976) Biohemistry 15, Chane, R. E., Hoffmann, J. A., Kroeff, E. P., Johnson, M. G., Shirmer, E. W., Bromer, W. W., Ross, M. J. & Wetzed, R. (1981) in Peptides, Synthesis, Struture and Funtion (Pro. Am. Pept. Symp. 7th) (Rih, D. H. & Gross, E., eds.), pp , Piere Chemial Co., Rokford Du, Y.-C., Zhang, Y.-S., Lu, Z.-X. & Tsou, C.-L. (1961) Si. Sin. (Engl. Ed.) 1, Du, Y.-C., Jiang, R.-Q. & Tsou, C.-L. (1965) Si. Sin. (Engl. Ed.) 14, Frank, B. H., Pettee, J. M., Zimmerman, R. E. & Burk, P. J. (1981) in Peptides, Synthesis, Struture and Funtion (Pro. Am. Pept. Symp. 7th) (Rih, D. H. & Gross, E., eds.), pp , Piere Chemial Co., Rokford Givol, D., De Lorenzo, F., Goldberger, R. F. & Anfinsen, C. B. (1965) Pro. Natl. Aad. Si. U.S.A. 53, Hillson, D. A., Lambert, N. & Freedman, R. B. (1984) Methods Enzymol. 17, Katsoyannis, P. G., Takatellis, A. C., Johnson, S., Zalut, C. & Shwartz, G. (1967) Biohemistry 6, Katzen, H. M. & Tietze, F. (1966) J. Biol. Chem. 241, Kauzmann, W. (1959) in Sulfur in Proteins (Benesh, R., ed.), pp. 9-18, Aademi Press, New York Lehninger, A. L. (1984) Priniples of Biohemistry, 3rd edn., pp , Worth Publishers, New York Lu, Z.-X. & Cui, T. (1981) Shengwu Huaxue Yu Shengwu Wuli Xuebao 13, Lu, Z.-X. & Tsao, T.-C. (1962) Shengwu Huaxue Yu Shengwu Wuli Xuebao 2, Moody, A. J., Stan, M. A., Stan, M. & Gliemann, J. (1974) Horm. Metab. Res. 6, Mori, A. F. & Wond, R. J. (1984) J. Biol. Stand. 12, Niu, C.-., Yang, S.-Z., Ma, A.-Q., Chen, Y.-N., Jiang, R.-Q. & Huang, W.-T. (1981) Biopolymers 2, Pruitt, K. M., Robison, B. S. & Gibbs, J. H. (1966) Biopolymers 4, Qian, Y.-Q. & Tsou, C.-L. (1987) Biohem. Biophys. Res. Commun. 146,

7 A and B hains have strutural information for formation of insulin 435 Steiner, D. F. & Clark, J. L. (1968) Pro. Natl. Aad. Si. U.S.A. 6, Tang, J.-G., Wang, C.-C. & Tsou, C.-L. (1988) Biohem. J. 255, Varandani, P. T. & Nafz, M. A. (197) Arh. Biohem. Biophys. 141, Wang, C.-C. & Tsou, C.-L. (1986) Biohemistry 25, Wang, Z.-X., Ju, M. & Tsou, C.-L. (1987) J. Theor. Biol. 124, Zahn, H., Gutte, B., Pfeiffer, E. F. & Ammon, J. (1966) Liebigs Ann. Chem. 691, Reeived 2 Otober 1989/8 Deember 1989; aepted 18 Deember 1989