Chemical Synthesis of Insulin Analogs through a Novel Precursor

Dec 12, 2013 - ... Ando , Masaki Okumura , Kenji Inaba , Hironobu Hojo , Michio Iwaoka ... Emrah Altindis , Weikang Cai , Masaji Sakaguchi , Fa Zhang ...
0 downloads 0 Views 4MB Size
Articles pubs.acs.org/acschemicalbiology

Chemical Synthesis of Insulin Analogs through a Novel Precursor Alexander N. Zaykov, John P. Mayer, Vasily M. Gelfanov, and Richard D. DiMarchi* Indiana University, Department of Chemistry, Bloomington, Indiana 47405, United States of America S Supporting Information *

ABSTRACT: Insulin remains a challenging synthetic target due in large part to its two-chain, disulfide-constrained structure. Biomimetic single chain precursors inspired by proinsulin that utilize short peptides to join the A and B chains can dramatically enhance folding efficiency. Systematic chemical analysis of insulin precursors using an optimized synthetic protocol identified a 49 amino acid peptide named DesDi, which folds with high efficiency by virtue of an optimized structure and could be proteolytically converted to bioactive two-chain insulin. In subsequent applications, we observed that the folding of the DesDi precursor was highly tolerant to amino acid substitution at various insulin residues. The versatility of DesDi as a synthetic insulin precursor was demonstrated through the preparation of several alanine mutants (A10, A16, A18, B12, B15), as well as ValA16, an analog that was unattainable in prior reports. In vitro bioanalysis highlighted the importance of the native, hydrophobic residues at A16 and B15 as part of the core structure of the hormone and revealed the significance of the A18 residue to receptor selectivity. We propose that the DesDi precursor is a versatile synthetic intermediate for the preparation of diverse insulin analogs. It should enable a more comprehensive analysis of function to insulin structure than might not be otherwise possible through conventional approaches.

D

superior folding properties relative to existing SCI precursors. Structural analysis indicates residue B28 to be in close proximity to the N-terminus of the A-chain suggesting that a direct link between the two might provide a linear peptide capable of folding to the correct insulin conformation (Figure 1). Versatility of the new precursor was illustrated through the synthesis and characterization of insulin analogs with alanine substitutions at positions B12, B15, A10, A16, and A18. These analogs represent a set of mutant insulins that were not reported in the otherwise comprehensive alanine scanning mutagenesis completed by Kristensen and colleagues.16 In the present report five of the omitted alanine mutant precursors were chemically synthesized, folded, and enzymatically processed by Lys-C to yield two-chain insulin analogs that were biochemically characterized.

espite sizable advances in the relationship of insulin structure to function, synthesis remains a rate-limiting step to the discovery of novel insulin analogs.1 Chemical synthesis and rDNA biosynthesis continue to be the primary methods in insulin analog production, and yet, both possess inherent drawbacks.2−4 Classical chemical routes that utilize synthetic insulin A- and B-chains combined under optimal oxidative folding conditions afford yields of less than 10% of the correctly folded hormone.5−7 One way to enhance correct insulin disulfide bond formation is through the use of orthogonal cysteine protecting groups enabling unambiguous formation of individual disulfides, but overall synthetic yield remains low.1,8 The Kent group has enhanced the folding of linear insulin precursors through the use of a nonpeptide tether as a means to improve folding efficiency resulting in a total yield of 12% based on purified peptide fragments.2,7,9,10 Miniproinsulin (2) and porcine insulin precursor (PIP) (3) are single-chain insulins (SCI), which have been extensively utilized in the biosynthesis of insulin. They have been shown to efficiently fold and convert to a full-activity, des-B30 human insulin through the action of lysine-specific protease (LysC).11−15 PIP and related sequences were chemically synthesized by Hoeg-Jensen and co-workers and modified with anionic sequence extensions to enhance aqueous solubility.15 Folding and subsequent treatment with Lys-C simultaneously removed the extension and linker sequences to yield des-B30 human insulin. While our initial efforts sought to optimize Fmoc-based solidphase assembly of the linear precursors, a seminal improvement was realized through the identification of a novel precursor with © 2013 American Chemical Society



RESULTS AND DISCUSSION DesDi Precursor and Comparison to Miniproinsulin and PIP. Improvement in sequence assembly (Supporting Information), while important, was one component of our strategy. A profound enhancement in synthetic yield was realized only through the selection of a novel linear precursor, which could be efficiently folded, and enzymatically converted to the active two-chain form. One such precursor “DesDi” (4), obtained through deletion of ProB28 from miniproinsulin, was assessed for folding efficiency relative to the well-established Received: October 15, 2013 Accepted: December 12, 2013 Published: December 12, 2013 683

dx.doi.org/10.1021/cb400792s | ACS Chem. Biol. 2014, 9, 683−691

ACS Chemical Biology

Articles

70% yield with an improvement to 90% for solubility-tag modified peptides. The effect of the solubility tags was clearly evident from the much sharper chromatographic peaks exhibited by the extended peptides (Figure 2). Side-by-side comparison indicates that the new precursor folds as well or better than miniproinsulin, or PIP. Effect of B-chain Truncations on Folding. Further studies examining sequential C-terminal B-chain truncation of SCI (peptides 8−12, Table 1) revealed that only the des-B30 (9) and des-B29,B30 (10) sequences were capable of efficient folding. Additional truncation exerted a negative effect on folding (Figure 3), supporting our expectation that shortened sequences would interfere with achieving a proper insulin conformation. The failure of the B30-A1 sequence (8) to properly fold reflects the severe steric demands of the folding process and is consistent with the empirical rule proposed by Ammerer et al.17 They concluded that for a general sequence B(1−29)-(XnY)m-(A1−21) it is preferable for the integer n to be at least one in case of Y being Lys or Arg. The N-terminal extension of the B-chain with a polyglutamate linker of general sequence -GEnK- made no apparent difference on folding within this set of peptides although it noticeably improved handling of each molecule (Supporting Information).15 Tolerance of the DesDi Precursor to Substitutions. An additional set of polyglutamate-extended DesDi analogs were prepared to assess the effect of substitutions in the B26−B28 region on folding efficiency (peptides 13−18, Table 1, Figure 4). Substitutions to Lys at B26, B27 (peptides 14, 15), and B28 (3) or to Ala at B28 (16) were accompanied with high folding efficiency. On the other hand, analogs LysB27,ArgB28 (17) or ArgB28 (18) were both associated with lower folding efficiency. This observation is characteristic of an Arg residue at the connection site. In contrast to DesDi, several analogs based on the miniproinsulin sequence exhibited significantly diminished folding yields (peptides 19−25, Table 1). Analogs with substitutions LysB28,ArgB29 (20); LysB27,ArgB29 (21); LysB26,ArgB29 (22); GlyB28 (23); LysB28,ProB29 (24); and ArgB29 (25) all failed to produce properly folded product (Figure 4). These observations suggest the DesDi precursor is preferentially tolerant to substitution and potentially more suited to the synthesis of structurally diverse insulin analogs. This difference between DesDi and miniproinsulin is most pronounced in the substitution of TyrA19 with Ala (peptides 26−27, Table 1), a mutation known to dramatically reduce potency. Tyrosine at A19 is not only crucial for receptor interaction, it also participates in extended contacts with residues IleA2, LeuB15, LeuA16 and PheB25, all of which are involved in the hydrophobic core packing of insulin. We were unable to fold an AlaA19 peptide derived from miniproinsulin, while the DesDi AlaA19 tolerated the mutation quite well (Figure 5). To further validate the versatility of the DesDi precursor three insulin analogs with substitutions at the hydrophobic core A16 residue were synthesized (peptides 28−30, Table 1). Previous studies demonstrated that substitution of the native Leu at A16 with Val thwarted attempts to the analog via the miniproinsulin precursor, as well as by classical chain combination.18 Folding of the A16Val-DesDi (28) analog (Figure 6) proceeded efficiently confirming that even synthetically challenging mutations are well tolerated with this precursor. Similarly, while Ala substitution at A16 precluded folding as a miniproinsulin precursor (29), the folding of the corresponding DesDi Ala-analog (30) in marked contrast was

Figure 1. (A) Proximity of carbonyl carbons of B28 and B29 residues (yellow) to N-terminal amine on A1 (red); the distance reported in Ångstrom and obtained from the PDB structure 3INS. (B) Conceptual representation of the B28-A1 amide bond in DesDi precursor.

miniproinsulin and PIP. For this purpose, two sets of peptides (sequences 2−7, Table 1) were synthesized, folded, and purified. The set included the three single-chain insulin precursor sequences as well a version of each modified with the (Ac-EEEEEK-) extension peptide, a strategy devised by Tofteng and co-workers.15 Use of the DesDi precursor afforded superior yields in comparison to miniproinsulin and PIP precursors, a reflection of improved folding efficiency. In addition, the isolated yields of the Ac-EEEEEK extended peptides proved to be greater than their unmodified counterparts (Supporting Information), a phenomenon attributed to better solubility and chromatographic resolution. For a more direct comparison, purified peptides were reduced to their linear forms with DTT, quickly purified by size-exclusion chromatography to remove the reducing agent, and refolded by air-oxidation. The efficiency of the refolding step was assessed by analytical HPLC using an internal standard (Figure 2). The peptides demonstrated uniformly high conversion, as indicated by a single peak corresponding to a correctly folded product. The yields obtained in this refolding procedure followed a similar trend to that originally observed before reduction and refolding. The three core sequences converted in approximately 684

dx.doi.org/10.1021/cb400792s | ACS Chem. Biol. 2014, 9, 683−691

ACS Chemical Biology

Articles

Table 1. Single-Chain Insulin Peptides Produced in This Studya

a

For biochemical characterization peptides were converted to two-chain form by treatment with Lys-C.

receptor interaction surface (binding site 1) and to play a key role in insulin dimerization.21,22 Alanine substitution at this position significantly reduces potency relative to native insulin. While the AlaB12 mutation was reported to be deleterious to analog expression,16 its synthesis via chain combination was not impaired. This suggests that poor expression may results from inefficient folding under physiological conditions.20 With respect to AlaB15, we are unaware of any successful attempts to prepare this analog, which implies the importance of the native leucine to achieve proper folding. Replacement of leucine at B15 with valine does not result in significant potency loss. Nonetheless, it has been reported to interfere with insulin folding.18 The remaining alanine substitutions at residues A10 and A18 were expected to have little impact on folding as each of their native side-chains project to the insulin surface. The

unaffected by this substitution (Figure 6). These results collectively demonstrate that direct connection of residues B28 and A1 provides a structure that enhances correct disulfide bond formation, even in the presence of otherwise destabilizing mutations. Alanine Scan. In the reported alanine scanning mutagenesis of insulin, several analogs were omitted intentionally or due to technical issues related to expression, or folding.16 Specific Alaanalogs not included were A10, A16, A18, B12, B15, B27, and B28. Some of these, notably B12, B27, and B28, have since been prepared.19,20 We chose to synthesize and characterize four analogs A10, A18, B12, and B15 in addition to A16 (discussed above) using the DesDi precursor (peptides 30−34, Table 1, Figure 7A). Position B12 is particularly significant, as it has been shown to comprise a part of the primary insulin 685

dx.doi.org/10.1021/cb400792s | ACS Chem. Biol. 2014, 9, 683−691

ACS Chemical Biology

Articles

Figure 2. HPLC of crude reactions: SCI peptide reduction (blue trace) and refolding (black trace). Conversions were estimated by HPLC: PIP 72%, Miniproinsulin 63%, DesDi 70%, AcE5K-PIP 79%, AcE5K-Miniproinsulin 93%, AcE5K-DesDi 92%.

Figure 3. Folding efficiency screening of peptides with various degree of B-chain truncation. The general structure of peptides used was Ac-E5R[B1−B25,AB22]-connector-[A1−21]; typical retention time of properly folded peptide is between 6.5 and 7.0 min under the HPLC conditions described in the Supporting Information (acidic-pH analytical HPLC).

In Vitro Evaluation of Alanine Analogs. Analysis of in vitro biological activity by receptor binding and phosphorylation assays (Figure 7, Table 2) indicated a lower potency (60−70%) for the two-chain DesDi analog (derived by Lys-C processing of a single-chain precursor) relative to native insulin. Therefore, comparison of the other analogs is provided relative to DesDi. The B12Ala analog (31) exhibited 2% binding potency relative to DesDi, consistent with prior observations that highlight the importance of this single residue. The B15Ala analog (32) showed ∼2/3 the binding potency of DesDi and A10Ala (33) was ∼1/3 as active. The A18Ala mutant (34) displayed nearly twice the affinity toward the A versus the B isoform of the insulin receptor. Its binding is enhanced ∼50% to IRA relative to the DesDi analog but slightly weaker at IRB.

hydrophobic isoleucine at A10 is not believed to be involved in direct interaction with the insulin receptor, yet a 5-fold reduction in receptor affinity was recently reported when serine, the corresponding residue in insulin-like growth factors, was introduced at this position.23 Based on similar observations with the neighboring A8 and B5 residues, this region is considered important for achieving selectivity between insulin and IGF-1 receptors.24,25 AsnA18 is highly conserved in insulin among diverse species suggesting that it is essential, but replacement with methionine or threonine (the corresponding native residues in IGFs) has only modest impact on potency.23,24,26 To our knowledge, no other single-site substitutions at A18 have been reported. 686

dx.doi.org/10.1021/cb400792s | ACS Chem. Biol. 2014, 9, 683−691

ACS Chemical Biology

Articles

Figure 4. Tolerance of DesDi (a) and miniproinsulin (b) constructs to various substitutions at C-terminal end of B-chain. Peptides are of general sequence AcE5R-[B1−B25,AB22]-connector-[A1−A21]-OH.

Figure 5. Comparison of folding reactions of peptides with alanine substitution at A19. The general structure of peptides used was AcE5R-[B1− B25,AB22]-connector-[A1−21,AA19].

Figure 6. Comparison of folding reactions of SCI’s with mutations at position A16. The general structure of peptides used was [B1−B25]-connector[A1−21,XA16].

less potent at the IGF-1 receptor than the insulin receptor. Additionally, all analogs with the exception of A18Ala appeared to lose more potency at the IGF-1 receptor relative to the insulin receptor when compared to the DesDi parent sequence (Figure 7F and Table 2). The A16Ala analog was observed to be the most preferentially selective in insulin activity relative to IGF1, but the extremely low activity at the latter receptor does not allow precise quantitation. Discussion. Insulin represents a molecule of historical significance to biochemistry and remains of utmost daily importance to millions of insulin-dependent diabetics. Over the last fifty years, there have been steady improvements in the synthesis of this hormone that have led to a deeper understanding in the relationship of structure to function.

Finally, the A16Ala (30) substitution produced an analog with 5-fold lower binding affinity demonstrating that destabilization of the insulin hydrophobic core profoundly influences bioactivity. Biochemical signaling as measured by receptor phosphorylation correlated well with binding data (Figure 7 and Table 2). DesDi insulin proved to be a full agonist of slightly lower potency than native insulin. The B12 mutation was found to be the most disabling change, followed by A16 and B15. The potency of the A10Ala analog relative to that of DesDi in the phosphorylation assay appears 2-fold enhanced as compared to what was observed in the binding assay. The analog A18Ala demonstrated similar selectivity for the IRA isoform in comparable fashion to the binding data. All peptides were 687

dx.doi.org/10.1021/cb400792s | ACS Chem. Biol. 2014, 9, 683−691

ACS Chemical Biology

Articles

Figure 7. (A) Indication of alanine mutation sites (PyMOL, PDB structure 3INS). (B, C) Inhibition of 125I-insulin binding with DesDi insulin analogs at IR-A and IR-B receptors. (D, E, F) Phospho-specific ELISA with HEK293 cells overexpressing IR-A, IR-B, and IGF1-R receptors.

Table 2. Characterization Data for Alanine-Scan Analogs (See Table 1 for Sequences)a mutation IC50 IRA, nM IC50 IRB, nM EC50 IRA, nM EC50 IRB, nM EC50 IGF1-R, nM

DesDi

B12-Ala

B15-Ala

A10-Ala

A16-Ala

A18-Ala

insulin

1.4 ± 0.2 (100%) 1.5 ± 0.3 (100%) 0.40 ± 0.07 (100%) 0.50 ± 0.13 (100%) 48 ± 41 (100%)

65 ± 18 (2%) 91 ± 15 (2%) 27 ± 5 (2%) 18 ± 3 (3%) n.d.

1.8 ± 0.2 (77%) 2.2 ± 0.3 (71%) 0.67 ± 0.16 (59%) 0.75 ± 0.12 (66%) 300 ± 250 (16%)

3.2 ± 0.3 (43%) 4.7 ± 0.7 (32%) 0.36 ± 0.05 (109%) 0.65 ± 0.02 (77%) 200 ± 50 (23%)

7.0 ± 2.7 (20%) 6.7 ± 0.2 (23%) 1.3 ± 0.1 (30%) 1.9 ± 0.5 (26%) 2000 ± 1300 (2%)

1.0 ± 0.2 (140%) 2.0 ± 0.2 (78%) 0.31 ± 0.04 (126%) 0.60 ± 0.18 (82%) 17 ± 6 (290%)

0.85 ± 0.07 (160%) 1.2 ± 0.1 (134%) 0.29 ± 0.08 (139%) 0.33 ± 0.08 (150%) 69 ± 42 (69%)

a

Affinity measured via inhibition of 125I-insulin binding and EC50 values for activity were obtained from phospho-specific ELISA with HEK293 cells (phosphorylation assay). Values and corresponding standard errors are reported from average of three experiments. Percentage in parentheses represents relative potency/activity in comparison to DesDi analog.

more amino acids from the C-terminus of B-chain completely abolished the ability of single-chain peptides to properly fold (peptides 8−12, Table 1Extending the DesDi sequence by one amino acid constitutes miniproinsulin (B29-A1 connection), which folds efficiently but proved less versatile than DesDi. This suggests the C-terminal region of B-chain participates in folding and there is an optimal length. The DesDi precursor successfully accommodated multiple substitutions at residues B26−B28 (peptide 13−18). The strongest support for the superior versatility of DesDi as an insulin precursor derives from direct comparison of the alanine mutation at positions A16 and A19 (peptide 29, 30 and 26, 27), as well as preparation of the A16Val analog (peptide 28) that was previously deemed “unfoldable.18 The near-native biological properties of DesDi obtained by in vitro receptor binding and phosphorylation analyses are consistent with previous observations that the C-terminal pentapeptide of insulin B-chain makes little contribution to potency of the hormone.13 Alanine substitution at position B12 demonstrated reduced potency by 2 orders of magnitude, as reported by others.19 Significant loss of potency due to

Nonetheless, synthetic yields remain modest due to the challenging structure of the hormone and many analogs cannot be made. This limits the search for a more efficacious and safer insulin-based therapy. Through a structure-based approach, we identified a more optimal linear single-chain insulin precursor that proved of high efficiency and versatility in the preparation of previously unattainable analogs. A set of insulin analogs were synthesized using an optimized procedure and isolated as homogeneous single-chain folded peptides in 15−25% yields (see Supporting Information). The comparatively higher synthetic yields of these multiple insulin mutants that we could not otherwise prepare using conventional precursors attests to the versatility of the DesDi sequence. The folding conditions utilized in this work reflect “state of the art” oxidative, alkaline conditions.15,27 High-yield folding was observed for miniproinsulin, PIP, and DesDi peptides with only subtle differences between the three (Figure 2) when employing native insulin sequences (peptides 2−7). As an insulin precursor, DesDi was uniquely preferred relative to similar sequences. The further deletion of one or 688

dx.doi.org/10.1021/cb400792s | ACS Chem. Biol. 2014, 9, 683−691

ACS Chemical Biology

Articles

could not be previously prepared. This was enabled through the identification of DesDi as a superior precursor that folds with high efficiency and exhibits tolerance to destabilizing mutations. The utility of this precursor was validated through the preparation and characterization of several synthetically challenging insulin analogs. This establishes a foundation for more broadly interrogating the relationship of insulin structure to function.

mutation of valine at B12 to alanine is directly related to the importance of this residue in its interaction with the insulin receptor. Although residue B12 is positioned near the hydrophobic core of insulin, it does not significantly contribute to the packing of the molecule and thus does not alter the foldability of the analog. This conclusion is also supported by the computational analysis, which indicates that alanine substitution produces no meaningful void or perturbation in the insulin structure.28 Alanine substitutions of native leucine residues at A16 and B15 could not previously be prepared. Each is involved in packing of the insulin hydrophobic core and typically such mutations interfere with correct folding. These mutants were synthesized and characterized as DesDi insulin analogs. They both demonstrate how destabilization of the hydrophobic core lowers the biological activity of the insulin molecule at the studied receptors. Mutation at B15 seems to produce a less pronounced loss in affinity than the same mutation at A16. Based on analysis of the insulin structure, the B15 residue is involved in the packing of the flexible B-chain C-terminal fragment that detaches from the body of the insulin molecule upon binding.29 We speculate that destabilization caused by this mutation might therefore be offset by facilitation of the fragment dissociation. In other words, the structure of insulin analog more closely resembles its receptor-bound state and requires lesser rearrangement upon binding. Mutation at A16 seems to have a relatively larger impact on at the IGF1 receptor, where selectivity versus insulin receptors is enhanced by nearly 5-fold in comparison to native insulin (Table 2). Given the controversy surrounding insulin mutagenicity this observation is worthy of additional investigation.30−33 Hydrophobic isoleucine at A10 is not believed to be involved in direct interaction with the insulin receptor, but nonetheless, we observed an appreciable loss in binding affinity. Similarly, a 5-fold reduction in receptor affinity was recently reported when serine, the corresponding residue in homologous IGF-1 and IGF-2, was introduced at this position.23 These observations stand in contrast to the nearly full potency observed in the phosphorylation assay which highlights the ability of these assays to discriminate inherent differences in these analogs. All other analogs demonstrated a consistent correlation across both assays. Similar to the previous results with threonine substitution at A18, alanine mutation at this position had little apparent effect on insulin potency.23,26 However, with deeper inspection this mutation displays an interesting preference for insulin-based growth associated receptors. Its potency was enriched approximately 50% at IRA and 3-fold at IGF1 relative to IRB. Asparagine at A18 of insulin is fairly conserved across various species, while in IGF-1 and IGF-2 it is respectively methionine and threonine.25 These observations suggest that A18 has a role in maintaining selectivity among insulin-like receptors. Structurally, this residue is neighboring both insulin recognition surfaces (denoted as site 1 and site 2).34,35 It is uncertain whether A18 is directly involved in receptor interaction or indirectly influences proximal residues. In this regard, TyrA19 is absolutely crucial for insulin binding (site 1) and A17 when mutated to alanine produces a 2-fold reduction in potency (site 2).16 Additional studies directed at A18 could assist in achieving insulin receptor isoform-selective peptides to help elucidate the respective biological function of the IRA and IRB receptors.36 Conclusion. We present here an improved chemical synthesis of insulin analogs through a single-chain form that



METHODS

Synthesis of Single-Chain Insulin (SCI) Analogs. Peptides were synthesized using either an Applied Biosystems 433A or Symphony Peptide Synthesizer. The ABI-433A protocol utilized standard 0.1 mmol Fmoc cycles, diisopropylcarbodiimide (DIC)/ 6-Cl-hydroxybenzotriazole (6-ClHOBt) activation in NMP and 20% piperidine in NMP for Fmoc deprotection. AsnA21 was introduced in the form of the α-tBu-ester of Fmoc-Asp, which was coupled to Rink-amide ChemMatrix resin as the first cycle of the automated protocol. N-terminal acetylation of poly glutamate extended peptides was performed manually with 40 eq. of acetic anhydride and N,N-diisopropylethylamine in 10 mL of DCM for 1−2 h. The Symphony protocol used DIC/6-Cl-HOBt activation in DCM/ DMF (1:2), a 10-fold excess of Fmoc amino acid to resin and a 1.5 h coupling cycle. Deprotection was accomplished with two 10 min treatments of 20% piperidine in DMF. N-terminal acetylation was carried out using acetic acid in the last coupling cycle. All peptides were cleaved with a TFA cocktail (20 mL for 0.1 mmol of resin) containing 2.5% v/v of each: β-mercaptoethanol, triisopropylsilane, anisole, and water. The cleavage mixture was precipitated with a 10-fold volume of ether and the solid was isolated by centrifugation. The precipitate was washed twice more with ether and dried in vacuo for 1 h. Due to poor solubility of linear chain SCI’s, no characterization was performed at this stage and peptides were advanced directly to the folding step. Folding and Purification of SCI Analogs. Crude peptides (110−130 mg, 0.1 mM) and cysteine hydrochloride (70 mg, 2 mM) were combined in 250 mL Erlenmeyer flask and dissolved in 200 mL of glycine buffer (20 mM, pH 10.5). The pH of the solution was adjusted to 10.5 and reaction was stirred in open air at 4 °C overnight. For HPLC analysis, a 200 μL aliquot was withdrawn and 10 μL of a 1 mM tryptophan solution was added as a reference, the pH was adjusted to 8.0−8.5 with 0.1 M HCl, and precipitate, which occasionally formed, was removed by centrifugation. The pH of the solution was then adjusted to 9.0 with TFA and the folded peptide was isolated by preparative HPLC. Individual fractions were assessed for purity by MALDI-MS and analytical HPLC, pooled, and lyophilized. The homogeneity of the lyophilized materials was confirmed by MALDI-MS and HPLC. Typical yields of isolated peptides were consistently in the 10−25% range. Reduction and Refolding of Purified SCI Analogs. Purified peptides (6 mg) were dissolved in 250 μL of reducing buffer (6 M guanidinium-HCl, 0.1 M Tris-HCl, 0.1 M DTT, pH 8.7) and slowly stirred for 1−2 h. Conversion was confirmed by HPLC and the reduced peptides were purified by size exclusion chromatography on a Superdex-75 column using ammonium bicarbonate buffer (25 mM, pH 9.0) as eluent. Concentration of peptide in the collected fractions was determined by UV−vis absorbance at 280 nm. Folding reactions were conducted by diluting the peptides to 40−50 μM concentration with 50 μL of glycine buffer (400 mM, pH 11.0), 5 μL cysteine solution (200 mM, 1 mM final concentration), approximately 30 μL of 1 M NaOH (volume needed to adjust pH to 10.5) and an appropriate volume of 20 mM glycine buffer (pH 10.5) to a final volume of 1 mL. Peptides were folded overnight at 4 °C and analyzed by HPLC. 10 μL of 1 mM tryptophan solution was added to each 200 μL aliquot as an internal reference standard. The amount of folded peptide was determined through the ratio of the peak areas corresponding to peptide and tryptophan. 689

dx.doi.org/10.1021/cb400792s | ACS Chem. Biol. 2014, 9, 683−691

ACS Chemical Biology

Articles

Conversion of SCI Analogs to the Two-Chain Form (TCI) with Lys-C Endoproteinase. Peptides were dissolved in ammonium bicarbonate buffer (25 mM, pH 8.0) at 0.9-mg mL−1 concentrations. 10 μL Lys-C (the same buffer, 3UN/100 μL) was added to 200 μL of peptide solution and reactions were incubated initially at 37 °C for 4 h, and overnight at RT. Reactions that appeared sluggish at the 4-h time point were treated with additional aliquots of enzyme. Completion of the reactions was confirmed by MALDI-MS and HPLC. Crude peptides were diluted 5-fold, and their concentrations determined by absorbance at 280 nm. Alanine substituted analogs were characterized directly without purification. Affinity Measurements Using Scintillation Proximity Binding Assay. A series of 5-fold dilutions of the alanine analogs, DesDi and insulin were prepared in the assay buffer (50 mM Tris pH 7.5, 150 mM NaCl, 1 mM CaCl2, 5 mM MgCl2, 0.1% protease-free BSA, sterilized and stored at 4 °C) in 96-well format. An approximately 0.2 nM (0.5 μCi/mL) solution of radiolabeled I-125 insulin was prepared in the assay buffer. Cell membranes with insulin receptors A and B were prepared from transfected HEK293 cells. The stock solutions of membranes were diluted in the assay buffer at the concentration determined from calibration experiments and supplemented with protease inhibitor cocktail (Sigma, P-2714). 50 μL of each ligand dilution, 50 μL of the I-125 insulin solution, and 50 μL of membrane solutions were transferred into the 96-well assay plates (Costar #3632) in the described order. PVT PEI-treated Wheat Germ Agglutinin Type A SPA Beads (Perkin-Elmer) were suspended in the assay buffer at 5 mg mL−1 and 50 μL of suspension was added to each well. The plates were sealed, mixed for 5 min on a plate shaker, incubated 12 h at RT, and read on a MicroBeta scintillation counter (Perkin-Elmer). Activity Measurements with Phospho-specific ELISA. To measure receptor phosphorylation insulin, A, -B, or IGF-1 receptor transfected HEK293 cells were plated in 96 well tissue culture plates (Costar #3596, Cambridge, MA) and cultured in Dulbecco’s modified Eagle medium (DMEM) supplemented with 100 IU/ml penicillin, 100 μg/mL streptomycin, 10 mM HEPES and 0.25% bovine growth serum (HyClone SH30541, Logan, UT) for 16−20 h at 37 °C, 5% CO2, and 90% humidity. Serial dilutions of insulin or IGF-1 analogs were prepared in DMEM supplemented with 0.5% bovine serum albumin (Roche Applied Science #100350, Indianapolis, IN) and added to the wells with adhered cells. After 15 min incubation at 37 °C in humidified atmosphere with 5% CO2 the cells were fixed with 5% paraformaldehyde for 20 min at RT, washed twice with phosphate buffered saline pH 7.4, and blocked with 2% bovine serum albumin in PBS for 1 h. The plate was then washed three times and filled with horseradish peroxidase-conjugated antibody against phosphotyrosine (Upstate biotechnology #16-105, Temecula, CA) reconstituted in PBS with 2% bovine serum albumin per manufacturer’s recommendation. After overnight incubation at 4 °C, the plate was washed 4 times and 0.1 mL of TMB single solution substrate (Invitrogen, #00-2023, Carlbad, CA) was added to each well. Color development was stopped 5 min later by adding 0.05 mL 1N HCl. Absorbance at 450 nm was measured on Titertek Multiscan MCC340 (ThermoFisher, Pittsburgh, PA). Data Analysis and Nonlinear Fits. Data from affinity measurements and phosphorylation assays were processed in Excel using Solver add-in for nonlinear least-squares fitting. The equations used in the fitting procedures are given below. For competitive inhibition of radio-labeled ligand binding, Y = Ymin +

Variables: X, concentration of a ligand. Fitted parameters: Ymin and Ymax, lower and upper asymptotes; EC50, half-maximal concentration of the response from activation by first ligand; IC50, half-maximal concentration of inhibition caused by binding of the second ligand (to account for insulin negative cooperativity). To simplify the analysis, the equation assumes that the binding of the second ligand results in inactive receptor.



* Supporting Information Details of analytical and preparative HPLC, size-exclusion chromatography (FPLC), mass spectrometry, and UV−vis spectroscopy. This material is available free of charge via the Internet at http://pubs.acs.org.



*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We thank J. Karty for support with MS analysis and J. Levy for support with peptide synthesis. Funding was partially provided by Merck Research Laboratories and Indiana University.



IC50 n X

( )

Ymax − Ymin 1+

X EC50

+

REFERENCES

(1) Mayer, J. P., Zhang, F., and DiMarchi, R. D. (2007) Insulin Structure and Function. Pept. Sci. 88, 687−713. (2) Sohma, Y., Hua, Q.-X., Whittaker, J., Weiss, M. A., and Kent, S. B. H. (2010) Design and Folding of [GluA4(OβThrB30)]Insulin (“Ester Insulin”): A Minimal Proinsulin Surrogate that Can Be Chemically Converted into Human Insulin. Angew. Chem. 122, 5621−5625. (3) Belgi, A., Hossain, M. A., Tregear, G. W., and Wade, J. D. (2011) The Chemical Synthesis of Insulin: From the Past to the Present. Immunol., Endocr. Metab. Agents Med. Chem. 11, 40−47. (4) Ladisch, M. R., and Kohlmann, K. L. (1992) Recombinant Human Insulin. Biotechnol. Prog. 8, 469−478. (5) Hua, Q.-X., Chu, Y.-C., Jia, W., Phillips, N. F. B., Wang, R.-Y., Katsoyannis, P. G., and Weiss, M. A. (2002) Mechanism of Insulin Chain Combination: Asymmetric Roles of A-chain α-helices in Disulfide Pairing. J. Biol. Chem. 277, 43443−43453. (6) Katsoyannis, P. G. (1964) The Synthesis of the Insulin Chains and Their Combination to Biologically Active Material. Diabetes 13, 339−348. (7) Avital-Shmilovici, M., Mandal, K., Gates, Z. P., Phillips, N. B., Weiss, M. A., and Kent, S. B. H. (2013) Fully Convergent Chemical Synthesis of Ester Insulin: Determination of the High Resolution X-ray Structure by Racemic Protein Crystallography. J. Am. Chem. Soc. 135, 3173−3185. (8) Liu, F., Luo, E. Y., Flora, D. B., and Mayer, J. P. (2013) Concise Synthetic Routes to Human Insulin. Org. Lett. 15, 960−963. (9) Sohma, Y., and Kent, S. B. H. (2009) Biomimetic Synthesis of Lispro Insulin via a Chemically Synthesized “Mini-Proinsulin” Prepared by Oxime-Forming Ligation. J. Am. Chem. Soc. 131, 16313−16318. (10) Fang, G.-M., Li, Y.-M., Shen, F., Huang, Y.-C., Li, J.-B., Lin, Y., Cui, H.-K., and Liu, L. (2011) Protein Chemical Synthesis by Ligation of Peptide Hydrazides. Angew. Chem., Int. Ed. 50, 7645−7649. (11) Markussen, J. A. N. (1985) Comparative Reduction/Oxidation Studies with Single Chain Des-(B30) Insulin and Porcine Proinsulin. Int. J. Pept. Protein Res. 25, 431−434. (12) Markussen, J., Fiil, N., Ammerer, G., Hansen, M. T., Thim, L., Norris, K., and Voigt, H. O. (1996) Insulin Precursors, Novo, Nordisk A/S.

Variables: X, concentration of a ligand. Fitted parameters: Ymin and Ymax, lower and upper asymptotes; IC50, half-maximal inhibitory concentration; n, Hill coefficient. For ELISA phosphorylation activity measurements, Y = Ymin +

AUTHOR INFORMATION

Corresponding Author

Ymax − Ymin 1+

ASSOCIATED CONTENT

S

IC50 X

690

dx.doi.org/10.1021/cb400792s | ACS Chem. Biol. 2014, 9, 683−691

ACS Chemical Biology

Articles

(13) Leyer, S., Gattner, H.-G., LeithaUser, M., Brandenburg, D., Wollmer, A., and Hocker, H. (1995) The Role of the C-terminus of the Insulin B-chain in Modulating Structural and Functional Properties of the Hormone. Int. J. Pept. Protein Res. 46, 397−407. (14) Home, P., Bartley, P., Russell-Jones, D., Hanaire-Broutin, H., Heeg, J.-E., Abrams, P., Landin-Olsson, M., Hylleberg, B., Lang, H., and Draeger, E. (2004) Insulin Detemir Offers Improved Glycemic Control Compared With NPH Insulin in People With Type 1 Diabetes: A randomized clinical trial. Diabetes Care 27, 1081−1087. (15) Tofteng, A. P., Jensen, K. J., Schäffer, L., and Hoeg-Jensen, T. (2008) Total Synthesis of desB30 Insulin Analogues by Biomimetic Folding of Single-Chain Precursors. ChemBioChem 9, 2989−2996. (16) Kristensen, C., Kjeldsen, T., Wiberg, F. C., Schaffer, L., Hach, M., Havelund, S., Bass, J., Steiner, D. F., and Andersen, A. S. (1997) Alanine Scanning Mutagenesis of Insulin. J. Biol. Chem. 272, 12978− 12983. (17) Ammerer, G., Fiil, N., Hansen, M. T., Markussen, J., Norris, K., Thim, L., and Voigt, H. O. (1990) DNA-Sequence Encoding Biosynthetic Insulin Precursors and Process for Preparing the Insulin Precursors and Human Insulin, Novo Industri A/S, Novo Industri A/ S, A Danish Corporation. (18) Liu, M., Wan, Z.-l., Chu, Y.-C., Aladdin, H., Klaproth, B., Choquette, M., Hua, Q.-x., Mackin, R. B., Rao, J. S., De Meyts, P., Katsoyannis, P. G., Arvan, P., and Weiss, M. A. (2009) Crystal Structure of a ″Nonfoldable″ Insulin: Impaired Folding Efficiency Despite Native Activity. J. Biol. Chem. 284, 35259−35272. (19) Nakagawa, S. H., Tager, H. S., and Steiner, D. F. (2000) Mutational Analysis of Invariant Valine B12 in Insulin: Implications for Receptor Binding. Biochemistry 39, 15826−15835. (20) Chen, H., Shi, M., Guo, Z.-Y., Tang, Y.-H., Qiao, Z.-S., Liang, Z.H., and Feng, Y.-M. (2000) Four new monomeric insulins obtained by alanine scanning the dimer-forming surface of the insulin molecule. Protein Eng. 13, 779−782. (21) Vinther, T. N., Norrman, M., Strauss, H. M., Huus, K., Schlein, M., Pedersen, T. Å., Kjeldsen, T., Jensen, K. J., and Hubálek, F. (2012) Novel Covalently Linked Insulin Dimer Engineered to Investigate the Function of Insulin Dimerization. PLoS One 7, e30882. (22) Baker, E. N., Blundell, T. L., Cutfield, J. F., Cutfield, S. M., Dodson, E. J., Dodson, G. G., Hodgkin, D. M. C., Hubbard, R. E., Isaacs, N. W., Reynolds, C. D., Sakabe, K., Sakabe, N., and Vijayan, N. M. (1988) The Structure of 2Zn Pig Insulin Crystals at 1.5 Å Resolution. Philos. Trans. R. Soc., B 319, 369−456. (23) Gauguin, L., Klaproth, B., Sajid, W., Andersen, A. S., McNeil, K. A., Forbes, B. E., and De Meyts, P. (2008) Structural Basis for the Lower Affinity of the Insulin-like Growth Factors for the Insulin Receptor. J. Biol. Chem. 283, 2604−2613. (24) Conlon, M. J. (2000) Molecular Evolution of Insulin in Nonmammalian Vertebrates. Am. Zool. 40, 200−212. (25) Wan, Z., Xu, B., Huang, K., Chu, Y.-C., Li, B., Nakagawa, S. H., Qu, Y., Hu, S.-Q., Katsoyannis, P. G., and Weiss, M. A. (2004) Enhancing the Activity of Insulin at the Receptor Interface: Crystal Structure and Photo-cross-linking of A8 Analogues. Biochemistry 43, 16119−16133. (26) Palsgaard, J. (2003) Receptor binding properties of 25 different hystricomorph-like human insulin analogs. Master Degree Thesis, University of Copenhagen, Copenhagen, Denmark. (27) Qiao, Z. S., Guo, Z. Y., and Feng, Y. M. (2001) Putative Disulfide-Forming Pathway of Porcine Insulin Precursor during Its Refolding In Vitro. Biochemistry 40, 2662−2668. (28) Zoete, V., and Meuwly, M. (2006) Importance of Individual Side Chains for the Stability of a Protein Fold: Computational Alanine Scanning of the Insulin Monomer. J. Comput. Chem. 27, 1843−1857. (29) Menting, J. G., Whittaker, J., Margetts, M. B., Whittaker, L. J., Kong, G. K. W., Smith, B. J., Watson, C. J., Zakova, L., Kletvikova, E., Jiracek, J., Chan, S. J., Steiner, D. F., Dodson, G. G., Brzozowski, A. M., Weiss, M. A., Ward, C. W., and Lawrence, M. C. (2013) How Insulin Engages Its Primary Binding Site on the Insulin Receptor. Nature 493, 241−245.

(30) Ruiter, R., Visser, L. E., Herk-Sukel, M. P. P., Coebergh, J. W. W., Haak, H. R., Geelhoed-Duijvestijn, P. H., Straus, S. M. J. M., Herings, R. M. C., and Stricker, B. H. C. (2012) Risk of Cancer in Patients on Insulin Glargine and Other Insulin Analogues in Comparison with Those on Human Insulin: Results from a Large Population-Based Follow-up Study. Diabetologia 55, 51−62. (31) Sciacca, L., Le Moli, R., and Vigneri, R. (2012) Insulin Analogs and Cancer. Front. Endocrinol. 3, 21. (32) Zib, I., and Raskin, P. (2006) Novel Insulin Analogues and Its Mitogenic Potential. Diabetes, Obes. Metab. 8, 611−620. (33) Cordera, R., Salani, B., and Briatore, L. (2009) Insulin Analogues: Fears, Facts, and Fantasies. Diabetes/Metab. Res. Rev. 25, 50−51. (34) De Meyts, P., and Whittaker, J. (2002) Structural Biology of Insulin and IGF1 Receptors: Implications for Drug Design. Nat. Rev. Drug Discovery 1, 769−783. (35) De Meyts, P. (2004) Insulin and Its Receptor: Structure, Function, and Evolution. BioEssays 26, 1351−1362. (36) Glendorf, T., Stidsen, C. E., Norrman, M., Nishimura, E., Srensen, A. R., and Kjeldsen, T. (2011) Engineering of Insulin Receptor Isoform-Selective Insulin Analogues. PLoS One 6, e20288.

691

dx.doi.org/10.1021/cb400792s | ACS Chem. Biol. 2014, 9, 683−691