HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 45, 43443-43453, November 8, 2002

This Article Abstract Full Text (PDF) Supplemental Data All Versions of this Article: 277/45/43443    most recent M206107200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Hua, Q.-x. Articles by Weiss, M. A. Search for Related Content PubMed PubMed Citation Articles by Hua, Q.-x. Articles by Weiss, M. A. Social Bookmarking                      What's this?

Mechanism of Insulin Chain Combination

ASYMMETRIC ROLES OF A-CHAIN -HELICES IN DISULFIDE PAIRING^*,

Qing-xin Hua, Ying-Chi Chu^§, Wenhua Jia, Nelson F. B. Phillips, Run-ying Wang^§, Panayotis G. Katsoyannis^§¶, and Michael A. Weiss

From the  Department of Biochemistry, Case Western Reserve School of Medicine, Cleveland, Ohio 44106-4935 and the ^§ Department of Pharmacology and Biological Chemistry, Mount Sinai School of Medicine, New York University, New York, New York 10029

Received for publication, June 19, 2002, and in revised form, August 14, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The A and B chains of insulin combine to form native disulfide bridges without detectable isomers. The fidelity of chain combination thus recapitulates the folding of proinsulin, a precursor protein in which the two chains are tethered by a disordered connecting peptide. We have recently shown that chain combination is blocked by seemingly conservative substitutions in the C-terminal -helix of the A chain. Such analogs, once formed, nevertheless retain high biological activity. By contrast, we demonstrate here that chain combination is robust to non-conservative substitutions in the N-terminal -helix. Introduction of multiple glycine substitutions into the N-terminal segment of the A chain (residues A1-A5) yields analogs that are less stable than native insulin and essentially without biological activity. ¹H NMR studies of a representative analog lacking invariant side chains Ile^A2 and Val^A3 (A chain sequence GGGEQCCTSICSLYQLENYCN; substitutions are italicized and cysteines are underlined) demonstrate local unfolding of the A1-A5 segment in an otherwise native-like structure. That this and related partial folds retain efficient disulfide pairing suggests that the native N-terminal -helix does not participate in the transition state of the reaction. Implications for the hierarchical folding mechanisms of proinsulin and insulin-like growth factors are discussed.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Insulin is a globular protein containing two chains, A (21 residues) and B (30 residues). The monomer in solution (1, 2) resembles the crystallographic T-state (3), an -helix-rich structure stabilized by three disulfide bridges (Fig. 1A). The hormone is generated in vivo by proteolytic processing of a single-chain precursor, proinsulin (4). Folding proceeds via a preferred disulfide pathway (5, 6). A peptide model is provided by combination of isolated A and B chains. This reaction, designated insulin chain combination, yields native disulfide pairing (7). The absence of disulfide isomers demonstrates that specific folding information resides within the isolated chains (8). Two non-native isomers have been prepared by directed chemical synthesis (9). That such isomers are metastable (converting to insulin in the presence of base; Ref. 9) suggests that the native structure represents the ground state in a space of competing monomeric folds (10). Whereas chain combination has enabled the synthesis of many novel insulin analogs (11), including the first commercial recombinant DNA human insulin (12, 13), disulfide pairing can be blocked by specific amino acid substitutions (14, 15). Marked variations in yield of biosynthetic expression of variant single-chain precursor polypeptides have also been observed in engineered strains of Saccharomyces cerevisiae (16). Because the mechanism of folding is not well characterized, it is not known in either case why some analogs are readily prepared while others are not.

We have recently shown that Leu^A16 makes an essential contribution to the efficiency of insulin chain combination (15). Invariant among vertebrate insulins and insulin-like growth factors (Refs. 3 and 17; see arrow in Fig. 1B), Leu^A16 projects between A and B chains to anchor the C-terminal -helix (Fig. 1C). Its structural environment is constrained on opposite sides by the internal disulfide bridges of the protein (A6-A11 and A20-B19) and is otherwise bounded by the conserved side chains of Ile^A2, Tyr^A19, Leu^B11, and Leu^B15 (3). Non-polar substitutions (Ile^A16, Val^A16, and Phe^A16) were shown to impair disulfide pairing and perturb the thermodynamic stability of the hormone. Once formed, however, A16 analogs retain substantial receptor binding activity, suggesting that Leu^A16 does not itself contact the receptor¹ (15).

A complementary approach toward dissecting determinants of disulfide pairing is provided by "protein undesign": segmental destabilization of discrete structural elements (18). In this article we investigate whether the N-terminal -helix of the A chain is necessary for the success of chain combination. Inspection of crystal structures suggests that this helix is an integral component of the insulin fold (3). Introduction of multiple glycine substitutions into the A1-A5 segment (highlighted in red in Fig. 1B) yields analogs that are inactive and less stable than native insulin. Such substitutions shave side-chain contacts and attenuate helical propensities (19). Glycine substitutions are studied in the context of a variant B chain designed to prevent formation of insulin dimers and hexamers. The engineered monomer (designated DKP-insulin) exhibits enhanced stability and activity (20-22). To correlate synthetic yields with structure, circular dichroism (CD)² and two-dimensional ¹H NMR studies of a representative analog lacking invariant side chains Ile^A2 and Val^A3 (A chain sequence GGGEQCCTSICSLYQLENYCN; substitutions underlined and cysteines in bold) are presented.

View larger version (44K): [in this window] [in a new window]   Fig. 1.   Structure and sequence of insulin. A, ribbon model of T-state crystallographic protomer (2-zinc molecule 1, Chinese nomenclature; accession code 4INS) (3). The A chain is shown in red and B chain in blue. Location of side chains IleA2 (black), LeuA16 (purple), TyrA19 (red), and disulfide bridges are as indicated (sulfur atoms as orange spheres). B, sequences of insulin A chains. Residues A1-A5 are highlighted in red. Invariant leucine at position A16 is boxed in purple (arrow). Cysteines are shown in boldface. C, crystal structure of insulin (stereo pair) highlighting local structure in neighborhood of LeuA16. The A chain is shown in black and B chain in blue. Labeled residues are color-coded as follows: LeuA16 (red; methyl groups are indicated by red balls), IleA2 (purple), disulfide bridges (orange; cysteines are depicted as balls), TyrA19 (purple), and other core residues (green; LeuA13, LeuB11, LeuB15, and ValB18).

We demonstrate that chain combination is robust to multiple glycine substitutions in the N-terminal segment of the A chain and thus that this segment does not participate in the transition state of the pairing reaction. The solution structure of [Gly^A2,Gly^A3]DKP-insulin exhibits local unfolding of the A1-A5 segment in an otherwise native-like structure. Although the supersecondary structure of the B chain is not significantly perturbed by the variant A chain, conformational broadening of amide resonances is reduced relative to DKP-insulin. These observations suggest that packing of the native A1-A8 -helix influences the time scale of fluctuations in the B chain. Non-cooperative detachment of the A1-A5 segment is in accord with the modular partial folds of populated disulfide intermediates in the oxidative folding pathway of proinsulin (1, 6, 18) and insulin-like growth factors (23, 24). Although of critical importance in receptor recognition (1, 5, 25-27), the N-terminal -helix of the A chain plays a peripheral role in the specification of disulfide pairing. Hierarchical folding of structural elements (28, 29) may enable discrete surfaces of insulin to reorganize independently on receptor binding (27, 30, 31).

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- 4-Methylbenzhydrylamine resin (0.6 mmol of amine/g; Star Biochemicals, Inc.) was used as solid support for synthesis of A chain analogs; (N-butoxycarbonyl,O-benzyl)-threonine-PAM resin (0.56 mmol/g; Bachem, Inc.) was used as solid support for synthesis of the DKP B chain analog. tert-Butoxycarbonyl-amino acids and derivatives were obtained from Bachem and Peninsula Laboratories; N,N'-dicyclohexylcarbodiimide and N-hydroxybenzotriazole (recrystallized from 95% ethanol) were from Fluka. Amino acid analyses of synthetic chains and insulin analogs were performed after acid hydrolysis; protein determinations were carried out by the Lowry method using native insulin as standard. Chromatography resins were pre-swollen microgranular carboxymethylcellulose (Whatman CM52), DE53 cellulose (Whatman), and Cellex E (Ecteola cellulose; Sigma); solvents were high performance liquid chromatography (HPLC)-grade.

Synthetic Procedures-- Six insulin analogs were prepared by solid-phase synthesis (Ref. 32; Table I). Crude S-sulfonated A chains were purified by chromatography on a Cellex E column (1.5 × 47 cm) as described (14, 33), dialyzed against distilled water, and lyophilized to yield purified A chain S-sulfonate analogs. Crude S-sulfonated DKP B chain was likewise purified on a cellulose DE53 column (1.5 × 47 cm), dialyzed and lyophilized. Chain combination was implemented by the method of Chance and colleagues (12) for 24 h in 0.1 M glycine (pH 10.6) at 4 °C using S-sulfonate-modified A and B chains as described (14). Concentrations of A and B chains were 1.2 and 0.4 mM, respectively; a molar excess of A chain favors A-B pairing (relative to B chain aggregation) and compensates for formation of cyclic A chains. S-Sulfonate modification enhances peptide solubility and renders peptides refractory to disulfide chemistry. Specific A-B pairing is initiated by addition of dithiothreitol in quantity stoichiometric to the concentration of S-sulfonate groups. Relative yields are given in Table II. In a mock chain-combination reaction lacking the A chain, 10% of the B chain mass remained as HPLC-recoverable monomers after 24 h, whereas 60% was sedimented by microcentrifugation; in a mock reaction lacking B chain, the oxidized A chains remained soluble and recoverable by reverse-phase HPLC (RP-HPLC).

Protein Purification-- Insulin analogs were isolated from the combination mixture as described (14, 33) and purified on a 0.9 × 23-cm carboxymethylcellulose chromatography and RP-HPLC on a Vydac 218 TP column (0.46 × 25 cm); the latter used a flow rate of 0.5 ml/min with 20-80% linear gradient of 80% aqueous acetonitrile containing 0.1% trifluoroacetic acid over 80 min. Re-chromatography of this material on RP-HPLC under the same conditions in each case gave a single sharp peak. Amino acid analyses and mass spectrometry in each case gave expected values. The purity of the insulin analogs was in each case greater than 98% as evaluated by analytical reverse-phase HPLC. Electrospray mass spectra revealed no anomalous molecular masses as contaminants.

Disulfide Pairing-- Native disulfide pairing of Gly^A3-insulin was determined by x-ray crystallography in an R₆ zinc hexamer.³ Native disulfide pairing of G3-DKP-insulin was verified by two-dimensional NMR spectroscopy as follows. A native-like nuclear Overhauser effect (NOE) on the protein surface is observed between the  proton of Cys^A7 and the  proton of Cys^B7; a native-like NOE in the core is observed between the  proton of Cys^A6 and the side chain of Leu^B6; the positions of Cys^A11, Cys^A19, and Cys^B20 are well defined in the core by a native-like network of inter-residue NOEs involving Leu^A16, Leu^B11, and Phe^B24. Imposition of non-native pairing schemes in molecular models would violate one or more of these NOEs. Correct pairing of the remaining analogs was assumed.⁴

Receptor Binding Studies-- Radiolabeled [¹²⁵I-Tyr^A14]human insulin was purchased from Amersham Biosciences. Receptor binding assays were performed as described (34) with minor modifications. Human placental cell membranes were prepared (35), stored at 80 °C in small aliquots, and thawed prior to use. Membrane fragments (0.025 mg of protein/tube) were incubated with ¹²⁵I-labeled insulin (~30,000 cpm) in presence of selected concentrations of unlabeled peptide for 18 h at 4 °C in a final volume of 0.25 ml of 0.05 M Tris-HCl and 0.25% (w/v) bovine serum albumin at pH 8. Subsequent to incubation, each mixture was diluted with 1 ml of ice-cold buffer and centrifuged (10,000 × g) for 5 min at 4 °C. The supernatant was then removed by aspiration, and the membrane pellet counted for radioactivity. Data were corrected for nonspecific binding (amount of radioactivity remaining membrane-associated in the presence of 1 µM human insulin). Each determination was performed with three or four replicates (see Table I); values are reported as mean and standard deviation.

Circular Dichroism-- Far-ultraviolet (UV) CD spectra of each analog were obtained using an Aviv spectropolarimeter equipped with thermister temperature control and automated titration unit for guanidine denaturation studies. CD samples for wavelength spectra contained 25-50 µM insulin analog in (a) 50 mM KCl and 10 mM potassium phosphate (pH 7.4) or (b) 100 mM KCl and 10 mM glycine (pH 10.5); samples were diluted to 5 µM for equilibrium denaturation studies. Data were obtained at 4 °C unless otherwise indicated.

Thermodynamic Stabilities-- Guanidine denaturation data were fitted by non-linear least squares (36). In brief, CD data were fitted to Equation 1.

(Eq. 1)

x is the concentration of guanidine hydrochloride, and _A and _B are base-line values in the native and unfolded states. These base lines were approximated by pre- and post-transition lines _A(x) =  + m_Ax and _B(x) =  + m_Bx . Fitting CD data and base lines simultaneously circumvents artifacts associated with linear plots of G as a function of denaturant according to G^o(x) = G + m_Bx (for review, see Ref. 36). The m values obtained in fitting variant unfolding curves are in each case significantly lower than that obtained in fitting the wild-type unfolding curve (see Table II). Although this situation can be associated with an underestimate of stability (19), a lower m value may reflect greater exposed hydrophobic surface in absence of denaturant in accord with NMR studies of 3G analog.

NMR Spectroscopy-- Samples were prepared in 50 mM KCl and 10 mM potassium phosphate (pH 7.0) or 20% v/v deuterated acetic acid (pH 1.9) as described (1). Sequential assignment was obtained at 600 MHz based on homonuclear double-quantum-filtered correlated spectroscopy (DQF-COSY), total correlation spectroscopy (mixing time 55 ms), and NOESY (mixing times 100 and 200 ms) experiments in D₂O and 90% H₂O. NOE restraints were classified as strong (<2.7 Å), medium (<3.4 Å), or weak (<4.3 Å) as described in the supplemental material (available in the on-line version of this article). Distance geometry (DG)/simulated annealing calculations were performed using the program DG-II; restrained molecular dynamics (RMD) calculations were performed using X-PLOR.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

To test the importance of residues A1-A5 in chain combination, successive glycine substitutions were introduced as underlined (analogs 1G-5G): 1G, GIVEQCCTSICSLYQLENYCN (wild type); 2G, GGVEQCCTSICSLYQLENYCN; 3G, GGGEQCCTSICSLYQLENYCN; 4G, GGGGQCCTSICSLYQLE- NYCN; 5G, GGGGGCCTSICSLYQLENYCN.

A variant A chain containing the single substitution Val^A3  Gly (similar to that described by Kaarsholm and co-workers (Ref. 26)) was also prepared. To avoid possible confounding effects of insulin assembly, studies employ the "DKP" B chain (see caption to Table I) to obtain engineered monomers (1, 21). The DKP B chain combines with the wild-type A chain with native efficiency. Surprisingly, despite the non-conservative glycine substitutions, combination of variant A chains 2G-5G in each case retains essentially native yield (column 2 in Table II). As expected (25), the resulting analogs are essentially without biological activity (receptor binding affinities < 0.2% relative to human insulin; Table I).

CD spectra of DKP-insulin analogs 2G-5G exhibit decreased -helix content as illustrated in Fig. 2A. The far UV spectrum of [Gly^A2,Gly^A3]DKP-insulin (3G-DKP-insulin; diamonds in Fig. 2A) differs from that of DKP-insulin (solid line) in the magnitude of helix-associated deflections at 195 and 222 nm and in the ratio of ellipticities at 208 and 222 nm. Respective values of mean residue ellipticity at 222 nm suggest helical unfolding of 6-9 residues. The extent of helical attenuation is similar among each of the 2G-5G analogs⁵; the spectrum of 5G-DKP-insulin is shown in Fig. 3B. Because CD does not resolve specific sites of structural perturbation, the solution structure of 3G-DKP-insulin was investigated by ¹H NMR spectroscopy (1). The one-dimensional ¹H NMR spectrum of the analog is shown in Fig. 4 in relation to that of DKP-insulin. Native and variant spectra exhibit similar non-random dispersion of chemical shifts. Such similarity suggests that the analog retains a folded overall structure but does not exclude segmental destabilization and/or small distributed perturbations. To obtain residue-specific probes, complete sequential assignment was obtained at neutral pH and in 20% v/v deuteroacetic acid by homonuclear two-dimensional NMR methods (see supplemental material, available on-line). Main-chain d_N connectivities are delineated in the fingerprint region of the NOESY spectrum (Fig. 5, A and B, respectively). These and other connectivities are outlined in Wüthrich format in Fig. 6; the A chain (upper panel) is notable for the absence of helix-related connectivities between A1 and A8, whereas the B chain (lower panel) exhibits native secondary structure (1). Although overall patterns of secondary shifts (deviations from random-coil values) are similar in 3G-DKP-insulin and DKP-insulin, widespread small perturbations are observed throughout the protein (Table III). Resonances of residues A1-A4 exhibit motional narrowing, similar in extent to that of terminal B chain residues B1, B2, B29, and B30.

View larger version (39K): [in this window] [in a new window]   Fig. 2.   Structure and stability of 3G-DKP-insulin. A, far-UV CD spectrum of 3G-DKP-insulin () exhibits a lower helix content than does native DKP-insulin (solid line); spectra obtained at 20 °C and pH 7.4. B, guanidine unfolding transition of 3G-DKP-insulin at 4 °C () is shifted to the left relative to DKP-insulin (solid line), indicating decreased thermodynamic stability (see Table I). Ellipticities (222; arbitrary units) are not normalized per residue. C, solution structure of 3G-DKP-insulin (stereo pair) exhibits unfolding of residues A1-A5 in an otherwise native-like partial fold. A chain residues A1-A8 are shown in red and A9-A21 in black; the B chain is shown in blue. An ensemble of 20 DG/RMD structures is shown; restraint information and statistical parameters are given in Table IV.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Effect of pH on CD-detected structure and stability. A, far-UV CD spectra of DKP-insulin at pH 7.4 (line) and 10.5 () at 4 °C. B, spectra of 5G-DKP-insulin at pH 7.4 () and 10.5 () relative to DKP-insulin at pH 7.4 (line) at 4 °C. C, guanidine denaturation studies of DKP-insulin at pH 7.4 () and 10.5 () at 4 °C. D, guanidine denaturation studies of 5G-DKP-insulin at pH 7.4 () and 10.5 () at 4 °C. Inferred thermodynamic parameters are given in Table II.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Comparison of one-dimensional 1H NMR spectra of 3G-DKP-insulin (A) and DKP-insulin (B) in aqueous solution (pH 7) and 25 °C. Selected assignments are shown. Left-hand spectra were obtained in 90% H2O and 10% D2O; right-hand spectra in 100% D2O. The protein concentration was in each case 2 mM.

View larger version (48K): [in this window] [in a new window]   Fig. 5.   Two-dimensional NMR analysis of 3G-DKP-insulin. A and B, fingerprint region of NOESY spectrum of analog. Sequential assignments of A and B chains are outlined in panels A and B, respectively. Additional connectivities are summarized in Fig. 6. Conditions were as in Fig. 4; mixing time was 200 ms. C and D, diagonal plot of inter-residue NOEs in DKP-insulin (C) and 3G-DKP-insulin (D). In axes the A chain is designated 1-21 and B chain 22-51. Boxed areas indicate NOEs within B chain (a), between the C-terminal A chain -helix (A13-A19) and B chain (b), and between the A1-A8 segment and B chain (c). Arrow (in panel C) denotes key NOE between TyrA19 and IleA2 in native monomer. Asterisks indicate helix-related (i, i + 4) NOEs present in native A1-A8 -helix but absent in 3G variant. NOEs from TyrB26 to IleA2 and ValA3 (d in panel C) indicate attachment of C-terminal -strand of B chain to N-terminal -helix of A chain. Main-chain/main-chain contacts are shown in upper left (relative to diagonal); main-chain/side-chain and side-chain/side-chain contacts in lower right.

View larger version (9K): [in this window] [in a new window]   Fig. 6.   Summary of sequential 1H NMR assignment in Wüthrich format. Connectivities in A and B chains are outlined in upper and lower panels, respectively. Absence of (i, i + 3) NOEs in A1-A8 segment indicates unraveling of helix. Asterisks indicate sites of glycine substitution at residues A2 and A3. Arrows denote "DKP" substitutions in B chain that prevent self-association (21, 56).

The three-dimensional structure of 3G-DKP-insulin was defined by analysis of inter-residue NOEs. Such analysis demonstrates that spatial relationships characteristic of native insulin are largely retained in the variant; inter-residue NOEs are shown by diagonal plot in Fig. 5C (DKP-insulin) and Fig. 5D (3G-DKP-insulin). In these plots contacts between main-chain atoms are shown at upper left (relative to the diagonal) whereas contacts involving side chains are shown at lower right. Retained features include maintenance of native-like B chain super-secondary structure and C-terminal A chain -helix (A13-A19). Despite such similarities, the variant lacks local -helix-specific NOEs in the A1-A8 segment (asterisks in Fig. 5, C and D) and long range NOEs between this segment and other parts of the insulin molecule (box c). Of particular note is the absence of NOEs from residues A2 and A3 to either the C-terminal -helix (Leu^A16 or Tyr^A19) or C-terminal B chain -strand (residue Tyr^B26). Retained in the variant are other native-like NOEs between chains, including His^B5-Ile^A10 and Leu^B15-Tyr^A19. The B5-A10 contact constrains the N-terminal segment of the B chain along the surface of the A chain, whereas the B15-A19 contact reflects internal helix-helix packing.

The observed inter-residue NOEs, selected dihedral, and hydrogen-bond-related restraints (Table IV; see "Experimental Procedures") were used to obtain an ensemble of structural models by distance geometry and restrained molecular dynamics (DG/RMD). Coordinates have been deposited with the Protein Data Bank (accession code 1LKQ). In accord with qualitative features of the NMR spectrum, 3G-DKP-insulin exhibits segmental unfolding of the N-terminal -helix of the A chain in an otherwise native-like fold (Fig. 2C). DG/RMD models thus rationalize the attenuated CD spectrum of the analog (Fig. 2A). Similarities among CD spectra of 2G-5G analogs suggest that such segmental destabilization is shared by this series of mutant insulins. The partial hydrophobic core of 3G-DKP-insulin is similar to that of DKP-insulin and native insulin (Fig. 7). Despite detachment of the A1-A5 segment, the majority of core side chains remain inaccessible to solvent (Table V, part B). Thus, although native internal packing of Ile^A2 is absent, desolvation of the remaining core is achieved by specific packing among the side chains of Leu^A16, Tyr^A19, Leu^B6, Leu^B11, Leu^B15, Val^B18, and Phe^B24 in relation to internal cystines A6-A11 and A20-B19 (Fig. 7A). Exposure of side chains in the N- and C-terminal segments of the B chain is enhanced by destabilization of their interface with the A chain. Enhanced exposure of hydrophobic atoms at these interfaces is in accord with the lower m value obtained in two-state fitting of guanidine denaturation studies (Table II, column 6; see above).

View larger version (31K): [in this window] [in a new window]   Fig. 7.   Hydrophobic core of 3G analog resembles native insulin. A, ribbon model (left) and stereo pair (right) showing key side chains (heavy atoms only) in core of 3G-DKP-insulin. The ensemble was aligned according to main-chain atoms of residues A12-A19 and B9-B24. B and C, analogous structural features in DKP-insulin (B) and collection of 29 crystallographic protomers (C) (accession codes 4INS, 2INS, 1APH, 1BPH, 1CPH, 1DPH, 1TRZ, 1TYL, 1TYM, 1ZNI, 1ZNJ, 1LPH, 1G7A, and 1EV6). In each panel the following coloring scheme is employed: the A chain is shown in red and B chain in black (left-hand panels), A16 and B24 side chains in magenta, A19 in red, B12 in blue, B15 in black, and A20-B19 disulfide bridge in auburn. Corresponding percentages of solvent accessibilities in DKP-insulin and 3G analog are given in Table V.

A qualitative difference between 3G-DKP-insulin and DKP-insulin is apparent from inspection of one-dimensional ¹H NMR spectra (Fig. 4). Whereas DKP-insulin (like native insulin) exhibits significant variation in amide line widths, such resonances in the variant are in general sharper and more uniform.⁶ Such narrowing occurs in the folded moiety (as illustrated in Fig. 4 for the downfield amide resonance of Cys^A11) and so is not a consequence of conventional motional narrowing. Rather, whereas amide line widths are anomalously broadened in DKP-insulin, those of 3G-DKP-insulin are more uniformly appropriate given the rotational correlation time of a small globular protein. Hence, DQF-COSY H_N-H cross-peaks in the B9-B19 and A13-A19 helices, essentially unobservable in DKP-insulin because of antiphase cancellation (37), are largely present in the spectrum of 3G-DKP-insulin. Similar trends in amide line widths were observed in comparative studies of insulin and a fragment lacking residues B26-B30 (des-pentapeptide[B26-B30]insulin (DPI); Refs. 38 and 39). We ascribe the "improved" NMR features of 3G-DKP-insulin and DPI to more complete averaging of chemical shifts as a result of fast exchange among conformational substrates (see "Discussion").

Thermodynamic stabilities of 3G-DKP-insulin and related glycine analogs were evaluated by CD-detected guanidine titrations at pH 7.4 and 10.5 (Figs. 2B and 3D). Analysis by the two-state formalism (36) in each case indicates decreased stability (column 3 in Table II). At neutral pH (Table II, part A) 2G-, 3G-, and 4G-DKP-insulin are each ~2.0 kcal/mol less stable than DKP-insulin; 5G-DKP-insulin is ~ 2.4 kcal/mol less than DKP-insulin (Table II, part A). Because the His^B10  Asp substitution (within the DKP B chain; see legend to Table I) enhances the stability of an insulin monomer (22), the 5G A chain was also combined with the wild-type B chain. Like 5G-DKP-insulin, 5G-insulin is ~2.4 kcal/mol less stable than its parent structure (insulin). Guanidine denaturation studies at pH 10.5 correspond to the conditions of insulin chain combination (Table II, part B). Insulin and DKP-insulin are less stable at basic pH than at neutral pH. Under these conditions their CD spectra exhibit a small perturbation (11% decrement in mean residue ellipticity at 222 nm; Fig. 3A) consistent with partial helical instability (40). Thermodynamic decrements caused by the glycine substitutions remain significant at pH 10.5 (G_u 0.4-0.7 kcal/mol; column 4 in Table II, part B) but are less marked under these conditions than at neutral pH. Attenuation of these thermodynamic differences at pH 10.5 reflects the robustness of the glycine analogs to base-induced destabilization; the 3G-5G analogs are in fact more stable at pH 10.5 than at pH 7.4 (Fig. 3D and Table II). CD spectra of these analogs are unperturbed by basic pH (Fig. 3B); similar helical attenuation is observed at pH 7.4 (filled triangles) and pH 10.5 (open triangles) relative to DKP-insulin (solid line). Insensitivity of the 3G-5G analogs to further helical attenuation at basic pH suggests that the corresponding perturbation of insulin and DKP-insulin is the result of loosening of the A1-A8 -helix. Instability of this helix under conditions of chain combination is consistent with the native pairing yields of the 2G-5G analogs. Unlike at neutral pH, the 3G-5G m values obtained at pH 10.5 are similar to those of insulin and DKP-insulin (column 6 of Table II, part B).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Insulin contains three -helices, one in the B chain and two in the A chain (Fig. 1A). Inspection of crystal structures suggests that the N-terminal A chain -helix (residues A1-A8) is integral to the insulin fold (3). This conserved amphipathic -helix (sequence GIVEQCCT; Fig. 1B) contributes to both the protein surface and hydrophobic core. Long range contacts are made by the side chains of Ile^A2 (which packs against Leu^A16, Tyr^A19, Leu^B11 Leu^B15, Tyr^B26, and Pro^B28), Val^A3 (which packs against Tyr^B26), and Cys^A6 (part of internal cystine A6-A11; Fig. 1C). The low activities of analogs containing substitutions or chemical modifications at residues A1-A3 suggest that these residues contact the insulin receptor (1, 16, 25, 41). The present study of multiple glycine substitutions in the N-terminal A chain -helix extends a previous NMR study of a Gly^A3 variant of a monomeric insulin analog by Kaarsholm and colleagues (26). The monomeric B chain template contained four modifications (Glu^B1, Glu^B10, Glu^B16, Glu^B27, and des-B30) and so differs from the DKP template employed here. The Gly^A3 analog was shown to retain a native-like fold with subtle perturbations to the structure and stability of the A1-A8 helix. Unlike the present structure of 3G-DKP-insulin, the Gly^A3 variant retains close contacts between a partially ordered A1-A8 helix and the C-terminal B chain -strand (26). Here, we have employed multiple glycine substitutions to induce a gross perturbation to this helix and effect its detachment. Such "undesign" was motivated by our recent study of Ala^A2-DKP-insulin, wherein loss of critical packing interactions involving Ile^A2 was found to destabilize the N-terminal A chain segment (42). These observations raised the question of whether this segment, seemingly so integral to the structure of insulin, contributes to the mechanism of disulfide pairing.

The solution structure of 3G-DKP-insulin is remarkable for detachment of the A1-A5 segment in an otherwise native-like fold. By conventional criteria of sensitivity and resolution, the quality of the variant spectra is superior to that of DKP-insulin or native insulin (1, 39). This "improvement" (a seeming paradox in light of the partial fold of the analog and its very low activity) reflects a reduction in the extent of anomalous broadening otherwise characteristic of insulin and insulin analogs (43, 44). Because the amide line widths of 3G-DKP-insulin are appropriate for a small globular protein, less extensive anti-phase cancellation is observed in DQF-COSY spectra (37) and more efficient total correlation spectroscopy transfer is achieved. Similar technical improvement in NMR quality was observed on deleting residues B26-B30 in NMR studies of DPI (38, 39). We suggest that the detached (A1-A5) or deleted (B26-B30) elements are coupled to non-local motions in the core of insulin; in native insulin and DKP-insulin, the millisecond time scale of such motions leads to incomplete averaging of NMR chemical shifts ("conformational broadening"; Refs. 43 and 44). We surmise that, in the absence of these "damping" elements, such motions are accelerated into the microsecond time regime, leading to more complete averaging of chemical shifts. Uniform amide line widths are also observed in NMR studies of R₆ insulin hexamers, presumably because their assembly damps such slow motions and restricts the range of accessible conformations (45).

Insulin chain combination is severely impaired by selected substitutions and chemical modifications in the C-terminal segment of the A chain (15, 46). Like the N-terminal A chain -helix of the A chain, the C-terminal helix (residues A13-A19) is amphipathic (LYQLENY; Leu^A16 underlined) and seemingly integral to the structure of insulin (3). Whereas alanine substitutions at positions A12-A15 and A17 are functionally well tolerated (16), the para hydroxyl group of Tyr^A19 is solvent exposed and proposed to contact the receptor (3). The C-terminal helix is anchored to the hydrophobic core by Leu^A16 (underlined above), which packs between A and B chains near internal disulfide bridges A6-A11 and A20-B19 (Fig. 1C). Substitution of Leu^A16 by alanine (16) or valine⁷ causes an essentially complete block to biosynthetic expression of a single-chain insulin precursor in yeast. In chain combination substitution of Leu^A16 by valine or phenylalanine impedes disulfide pairing by at least 25-fold (15). Inefficient pairing is not caused by accelerated off-pathway events such as formation of disulfide isomers (9, 10) or variant A chain polymers. Despite such low yields, the brute force of mass action enabled Val^A16- and Phe^A16-DKP-insulin analogs to be obtained in quantities small but sufficient to demonstrate that these substitutions are consistent with substantial receptor binding activity (40% relative to insulin). Although "unfoldable" A16 analogs exhibit attenuated -helix content and reduced stability (15), these perturbations seem similar in extent to those observed among the present N-terminal glycine series. We speculate that C-terminal A chain packing interactions contribute to the stability of both the native and transition states, whereas N-terminal structure is absent or inessential in the transition state.

Insulin chain combination is typically implemented at pH values between 9.5 and 11 (the present studies employ 0.1 M glycine at pH 10.6). Such basic pH leads to deprotonation of thiolate moieties and enhances yield by limiting aggregation of reduced B chains, a competing off-pathway process that is extremely rapid at pH values less than 8.6, the pK_a of cysteine. Aggregation of B chains also occurs at pH 10.6 but its rate is similar to that of chain combination, enabling isolation of insulin and insulin analogs. We may view relative yields among the 2G-5G series as reflecting kinetic competition between disulfide pairing and B chain aggregation, which defines a common clock in each reaction. The five A chains tested (Table I) are equally effective in this competition. The success of these syntheses may also have been facilitated by the robustness of these analogs to base-induced perturbation in structure or stability. The partial folds of the 3G-5G analogs are resistant to further unfolding at pH 10.5, and their stabilities are somewhat higher than at neutral pH. Relative contributions of kinetic guidance versus end product stability to the yield of chain combination have not been defined.

These observations have potential implications for how proinsulin folds (47). Like chain combination, the yield of proinsulin folding is also enhanced at pH 9.5-11 because of reduced aggregation of unfolded and partially folded species (4, 48). Redox-coupled folding mechanisms of a single-chain proinsulin analog (porcine insulin precursor; PIP) and IGF-I are notable for populated one- and two-disulfide intermediates identified by peptide mapping (5, 49, 50). A key role is played in each case by formation of cystine A20-B19 (or IGF-I homolog 18-61). Whereas in the IGF-I pathway this is the only one-disulfide species to accumulate, cystine A20-B19 is formed in PIP by parallel branches (i.e. either before or after formation of A6-A11; Fig. 8A). Local pairing of A6-A11 is thought to be peripheral to the initial conformational search (6). We propose that cystine A20-B19, uniquely among the disulfide bridges of proinsulin, forms part of a specific folding nucleus⁸ (SFN; Refs. 6, 23, 24, and 51). Insulin analogs containing pairwise cysteine substitutions have previously been characterized as peptide models of such populated intermediates (1, 6, 18). Although these analogs are flexible, their predominant conformational features suggest that formation of the A20-B19 disulfide bridge is directed by a cluster of surrounding non-polar side chains (Fig. 8, B and C). Native-like interactions are observed involving the central B chain -helix (Leu^B11 and Leu^B15), B chain -strand (Phe^B24), and C-terminal A chain -helix (Leu^A16 and Tyr^A19). In the absence of either cystine A6-A11 or A7-B7, the N-terminal segment of the A chain is not well ordered (dashed lines in Fig. 8, B and C), suggesting that its -helical folding is modular and occurs late in the pathway.

View larger version (28K): [in this window] [in a new window]   Fig. 8.   Proposed disulfide pathway of proinsulin. A, disulfide intermediates based on biochemical mapping studies of a proinsulin-like precursor (5). U designates unfolded and reduced polypeptide; N designates native state with three disulfide bridges (A6-A11, A7-B7 and A20-B19). Intermediate species are designated by cystine(s), e.g. [A20-B19] indicates a species containing one disulfide bridge. Cystine A20-B19 is highlighted in red in each panel. Designations fast or slow refer to kinetic study of Feng and co-workers (5). Because technical limitations in peptide mapping precluded unambiguous assignment of disulfide pairing schemes, pathway is in part hypothetical. B and C, cylinder models of two-disulfide insulin analogs. B, DKP-des-[A7-B7]Ser; C, DKP-des-[A6-A11]Ser/Ala based on NMR data (1, 6, 18). Dashed lines indicate disordered regions. Positions of cystines are indicated by filled circles and bar.

Conformational features of two-disulfide analogs suggest that nascent structure in the B chain provides a template for folding of the A chain in a hierarchical process. We interpret the present results from a kinetic perspective. The contrasting effects of N- and C-terminal A chain substitutions on chain combination provide evidence for the existence of a polarized transition state (52) ensemble leading to formation of cystine A20-B19. Because the proposed SFN containing cystine A20-B19 represents the first stable substructure to appear, mutations that destabilize this substructure are also likely to destabilize the transition state. We imagine that packing of Leu^A16 assists in orienting Cys^A20 and Cys^B19 for proper disulfide pairing. In such a mechanism, modular detachment of the A1-A5 segment among the 2G-5G analogs would destabilize the native state but permit efficient disulfide pairing. We imagine that off-nucleus contacts involving this segment are optional in the polarized transition state and so may be bypassed. N-terminal analogs, although less stable than native insulin, are thus readily prepared. The proposed mechanism, although speculative, would rationalize why it is possible to achieve efficient biosynthetic expression and folding of PIP variants containing pairwise substitution of either "off-nucleus" cystine (A6-A11 or A7-B7) but not of cystine A20-B19⁹ (53). Likewise, syntheses of two-disulfide insulin analogs lacking either cystine A6-A11 or A7-B7 (Fig. 8, B and C) were accomplished with high yield (1, 6). Efficient pairing of such analogs seems remarkable in light of their marked instabilities (rows 9 and 10 in Table II, part A).

Testing the hypothesized mechanism of folding will require thermodynamic studies of proinsulin analogs in parallel with kinetic studies of oxidative folding. Of particular interest will be assessment of the consistency between these findings and yields of chain combination, i.e. whether mutations that impair folding of proinsulin would also impair chain combination and vice versa. Parallel findings would suggest that, despite differences in reaction conditions and chain topology, folding of proinsulin and chain combination employ analogous kinetic mechanisms (5, 6, 49). In particular, despite the branched disulfide pathway of proinsulin (Ref. 5; Fig. 8A), we propose that a similar SFN forms in each arm to nucleate formation of the A20-B19 disulfide bridge. The yield of either reaction is limited in vitro by competition between protein folding and aberrant protein aggregation (48). Such aggregation is presumably circumvented in vivo by chaperones (including disulfide isomerase) and sequestration of the folded state in zinc-stabilized hexamers (53). How proinsulin folds and misfolds may contribute to the pathogenesis of type II diabetes mellitus; imbalance between chaperone-mediated protein folding and aggregation in the  cell is proposed to activate a stress pathway leading to apoptosis and diminished  cell reserve (54, 55). The robustness of insulin chain combination to drastic changes in the sequence and structure of the N-terminal A chain segment suggests asymmetric encoding of folding information in the A domain of proinsulin. Analogous studies to define essential and inessential residues in the B chain are in progress. Together, such studies promise to delineate distinct determinants of foldability, stability, and function.

	ACKNOWLEDGEMENTS

We thank S.-Q. Hu for assistance with chain combination studies; S. H. Nakagawa (University of Chicago, Chicago, IL) for assistance with receptor-binding studies and discussion; M. DeFelippis (Eli Lilly and Co.) for gift of human insulin and isolated chains; T. Sosnick for advice regarding CD analysis; D. Poruban for assistance with figures; E. Collins for preparation of the manuscript; Prof. V. Anderson, B. H. Frank, and T. Sosnick for helpful discussion; and E. Blout and M. Karplus for encouragement.

	FOOTNOTES

^* This work was supported in part by a grant from the National Institutes of Health (to M. A. W. and P. G. K.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The on-line version of this article (available at http://www.jbc.org) contains one figure showing histograms of chemical shift changes in 3G-DKP-insulin and four tables providing chemical shift information for 3G-DKP-insulin and DG/RMD restraints.

The atomic coordinates and the structure factors (code 1LKQ) have been deposited in the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ (http://www.rcsb.org/).

^¶ To whom correspondence may be addressed. Tel.: 212-241-9350; Fax: 212-996-7214; E-mail: panayotis.katsoyannis@mssm.edu.

To whom correspondence may be addressed. Tel.: 216-368-5991; Fax: 216-368-3419; E-mail: weiss@biochemistry.cwru.edu.

Published, JBC Papers in Press, August 23, 2002, DOI 10.1074/jbc.M206107200

¹ It is not known whether impaired chain combination of A16 analogs reflects their thermodynamic instability or a kinetic block to disulfide pairing. Efforts to obtain an Ala^A16 analog through biosynthetic expression of a single-chain precursor have been unsuccessful (16). The yield of a Val^A16 mini-proinsulin analog expressed in S. cerevisiae is likewise negligible (see Footnote 7).

³ G. D. Smith, S. Nakagawa, H. S. Tager, and M. A. Weiss, manuscript in preparation.

⁴ Insulin chain combination in each case was observed to generate a unique product and not a mixture of isomers, which typically exhibit very different HPLC mobilities (10). To our knowledge, in no instance have mutations in insulin caused mispairing in synthesis. Gross substitution of the entire insulin B chain by the B domain of insulin-like growth factor I (IGF-I) yields native and non-native isomers in equilibrium (57) in accord with the anomalous refolding properties of IGF-I (49, 50).

⁵ Unlike the 2G-5G series, the CD spectrum of Gly^A3-DKP-insulin exhibits only a small decrement in helix content ([]₂₂₂ value approximately midway between its values in DKP-insulin and 3G-DKP-insulin) in accord with NMR studies of a related analog (see "Discussion" and Ref. 26).

⁶ Anomalous variation among amide line widths in insulin has been ascribed to incomplete averaging of chemical shifts because of millisecond motions in the protein (43, 44). Among analogs containing incomplete cores (such as DPI or 3G-DKP-insulin), such motions may be accelerated, leading to fast (rather than intermediate) exchange on the time scale of NMR chemical shifts (39).

⁷ Y.-M. Feng, personal communication.

⁸ In non-oxidative folding, the SFN defines a minimal stable element of structure whose assembly leads to the native state (58). In the case of disulfide-linked folding of proteins containing multiple cystines, distinct nuclei may exist for each barrier crossing (59).

⁹ Biosynthetic expression studies of single-chain insulin analogs (60) suggest (but do not establish) that formation of cystine A20-B19 is an obligatory kinetic step. We cannot exclude that substitutions in A20 and B19 lead to so profound a decrement in end-product stability as to preclude formation of a native-like fold, i.e. thermodynamic control of folding yield relative to off-pathway aggregation and/or degradation in yeast.

	ABBREVIATIONS

The abbreviations used are: CD, circular dichroism; DG, distance geometry; RMD, restrained molecular dynamics; DKP-insulin, monomeric insulin analog containing three substitutions in the B chain (Asp^B10, Lys^B28, and Pro^B29); DKP-des-[A7-B7]^Ser, two-disulfide analog of DKP-insulin containing substitutions Ser^A7 and Ser^B7; DKP-des-[A6-A11]^Ser, two-disulfide analog of DKP-insulin containing substitutions Ser^A6 and Ser^A11; DQF-COSY, double-quantum-filtered correlated spectroscopy; DPI, des-pentapeptide[B26-B30]insulin; IGF-I, insulin-like growth factor I; NOE, nuclear Overhauser effect; NOESY, nuclear Overhauser effect spectroscopy; PIP, porcine insulin precursor, a single-chain insulin precursor polypeptide; RP, reverse phase; HPLC, high performance liquid chromatography; SFN, specific folding nucleus; 2G-5G, A-chain analogs containing N-terminal glycine substitutions; 3G-DKP-insulin, [Gly^A2,Gly^A3]DKP-insulin, analog of DKP-insulin in which Ile^A2 and Val^A3 are each replaced by glycine. Amino acids are designated by standard one- and three-letter codes.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				M. A. Weiss Proinsulin and the Genetics of Diabetes Mellitus J. Biol. Chem., July 17, 2009; 284(29): 19159 - 19163. [Abstract] [Full Text] [PDF]

				Q.-x. Hua, B. Xu, K. Huang, S.-Q. Hu, S. Nakagawa, W. Jia, S. Wang, J. Whittaker, P. G. Katsoyannis, and M. A. Weiss Enhancing the Activity of a Protein by Stereospecific Unfolding: CONFORMATIONAL LIFE CYCLE OF INSULIN AND ITS EVOLUTIONARY ORIGINS J. Biol. Chem., May 22, 2009; 284(21): 14586 - 14596. [Abstract] [Full Text] [PDF]

				Z.-l. Wan, K. Huang, S.-Q. Hu, J. Whittaker, and M. A. Weiss The Structure of a Mutant Insulin Uncouples Receptor Binding from Protein Allostery: AN ELECTROSTATIC BLOCK TO THE TR TRANSITION J. Biol. Chem., July 25, 2008; 283(30): 21198 - 21210. [Abstract] [Full Text] [PDF]

				Q.-x. Hua, S. H. Nakagawa, W. Jia, K. Huang, N. B. Phillips, S.-q. Hu, and M. A. Weiss Design of an Active Ultrastable Single-chain Insulin Analog: SYNTHESIS, STRUCTURE, AND THERAPEUTIC IMPLICATIONS J. Biol. Chem., May 23, 2008; 283(21): 14703 - 14716. [Abstract] [Full Text] [PDF]

				Q.-x. Hua, J. P. Mayer, W. Jia, J. Zhang, and M. A. Weiss The Folding Nucleus of the Insulin Superfamily: A FLEXIBLE PEPTIDE MODEL FORESHADOWS THE NATIVE STATE J. Biol. Chem., September 22, 2006; 281(38): 28131 - 28142. [Abstract] [Full Text] [PDF]

				Q.-x. Hua, M. Liu, S.-Q. Hu, W. Jia, P. Arvan, and M. A. Weiss A Conserved Histidine in Insulin Is Required for the Foldability of Human Proinsulin: STRUCTURE AND FUNCTION OF AN ALAB5 ANALOG J. Biol. Chem., August 25, 2006; 281(34): 24889 - 24899. [Abstract] [Full Text] [PDF]

				Q.-x. Hua, S. Nakagawa, S.-Q. Hu, W. Jia, S. Wang, and M. A. Weiss Toward the Active Conformation of Insulin: STEREOSPECIFIC MODULATION OF A STRUCTURAL SWITCH IN THE B CHAIN J. Biol. Chem., August 25, 2006; 281(34): 24900 - 24909. [Abstract] [Full Text] [PDF]

				S. H. Nakagawa, Q.-x. Hua, S.-Q. Hu, W. Jia, S. Wang, P. G. Katsoyannis, and M. A. Weiss Chiral Mutagenesis of Insulin: CONTRIBUTION OF THE B20-B23 beta-TURN TO ACTIVITY AND STABILITY J. Biol. Chem., August 4, 2006; 281(31): 22386 - 22396. [Abstract] [Full Text] [PDF]

				K. Huang, J. Dong, N. B. Phillips, P. R. Carey, and M. A. Weiss Proinsulin Is Refractory to Protein Fibrillation: TOPOLOGICAL PROTECTION OF A PRECURSOR PROTEIN FROM CROSS-{beta} ASSEMBLY J. Biol. Chem., December 23, 2005; 280(51): 42345 - 42355. [Abstract] [Full Text] [PDF]

				M. Liu, Y. Li, D. Cavener, and P. Arvan Proinsulin Disulfide Maturation and Misfolding in the Endoplasmic Reticulum J. Biol. Chem., April 8, 2005; 280(14): 13209 - 13212. [Abstract] [Full Text] [PDF]

				Y. Chen, R. Jin, H.-Y. Dong, and Y.-M. Feng In Vitro Refolding/Unfolding Pathways of Amphioxus Insulin-like Peptide: IMPLICATIONS FOR FOLDING BEHAVIOR OF INSULIN FAMILY PROTEINS J. Biol. Chem., December 31, 2004; 279(53): 55224 - 55233. [Abstract] [Full Text] [PDF]

				Q.-x. Hua and M. A. Weiss Mechanism of Insulin Fibrillation: THE STRUCTURE OF INSULIN UNDER AMYLOIDOGENIC CONDITIONS RESEMBLES A PROTEIN-FOLDING INTERMEDIATE J. Biol. Chem., May 14, 2004; 279(20): 21449 - 21460. [Abstract] [Full Text] [PDF]

				Z.-S. Qiao, C.-Y. Min, Q.-X. Hua, M. A. Weiss, and Y.-M. Feng In Vitro Refolding of Human Proinsulin. KINETIC INTERMEDIATES, PUTATIVE DISULFIDE-FORMING PATHWAY, FOLDING INITIATION SITE, AND POTENTIAL ROLE OF C-PEPTIDE IN FOLDING PROCESS J. Biol. Chem., May 9, 2003; 278(20): 17800 - 17809. [Abstract] [Full Text] [PDF]

				M. Liu, J. Ramos-Castaneda, and P. Arvan Role of the Connecting Peptide in Insulin Biosynthesis J. Biol. Chem., April 18, 2003; 278(17): 14798 - 14805. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Supplemental Data All Versions of this Article: 277/45/43443    most recent M206107200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Hua, Q.-x. Articles by Weiss, M. A. Search for Related Content PubMed PubMed Citation Articles by Hua, Q.-x. Articles by Weiss, M. A. Social Bookmarking                      What's this?