Mutations at the dimer, hexamer, and receptor-binding surfaces of insulin independently affect insulin-insulin and insulin-receptor interactions.

Mutagenesis of the dimer- and hexamer-forming surfaces of insulin yields analogues with reduced tendencies to aggregate and dramatically altered pharmacokinetic properties. We recently showed that one such analogue, HisB10----Asp, ProB28----Lys, LysB29----Pro human insulin (DKP-insulin), has enhanced affinity for the insulin receptor and is useful for studying the structure of the insulin monomer under physiologic solvent conditions [Weiss, M. A., Hua, Q. X., Lynch, C. S., Frank, B. H., & Shoelson, S. E. (1991) Biochemistry 30, 7373-7389]. DKP-insulin retains native secondary and tertiary structure in solution and may therefore provide an appropriate baseline for further studies of related analogues containing additional substitutions within the receptor-binding surface of insulin. To test this, we prepared a family of DKP analogues having potency-altering substitutions at the B24 and B25 positions using a streamlined approach to enzymatic semisynthesis which negates the need for amino-group protection. For comparison, similar analogues of native human insulin were prepared by standard semisynthetic methods. The DKP analogues show a reduced tendency to self-associate, as indicated by 1H-NMR resonance line widths. In addition, CD spectra indicate that (with one exception) the native insulin fold is retained in each analogue; the exception, PheB24----Gly, induces similar perturbations in both native insulin and DKP-insulin backgrounds. Notably, analogous substitutions exhibit parallel trends in receptor-binding potency over a wide range of affinities: D-PheB24 greater than unsubstituted greater than GlyB24 greater than SerB24 greater than AlaB25 greater than LeuB25 greater than SerB25, whether the substitution was in a native human or DKP-insulin background. Such "template independence" reflects an absence of functional interactions between the B24 and B25 sites and additional substitutions in DKP-insulin and demonstrates that mutations in discrete surfaces of insulin have independent effects on protein structure and function. In particular, the respective receptor-recognition (PheB24, PheB25), hexamer-forming (HisB10), and dimer-forming (ProB28, LysB29) surfaces of insulin may be regarded as independent targets for protein design. DKP-insulin provides an appropriate biophysical model for defining structure-function relationships in a monomeric template.     

Atomic force microscopy of crystalline insulins: the influence of sequence variation on crystallization and interfacial structure.

The self-association of proteins is influenced by amino acid sequence, molecular conformation, and the presence of molecular additives. In the presence of phenolic additives, LysB28ProB29 insulin, in which the C-terminal prolyl and lysyl residues of wild-type human insulin have been inverted, can be crystallized into forms resembling those of wild-type insulins in which the protein exists as zinc-complexed hexamers organized into well-defined layers. We describe herein tapping-mode atomic force microscopy (TMAFM) studies of single crystals of rhombohedral (R3) LysB28ProB29 that reveal the influence of sequence variation on hexamer-hexamer association at the surface of actively growing crystals. Molecular scale lattice images of these crystals were acquired in situ under growth conditions, enabling simultaneous identification of the rhombohedral LysB28ProB29 crystal form, its orientation, and its dynamic growth characteristics. The ability to obtain crystallographic parameters on multiple crystal faces with TMAFM confirmed that bovine and porcine insulins grown under these conditions crystallized into the same space group as LysB28ProB29 (R3), enabling direct comparison of crystal growth behavior and the influence of sequence variation. Real-time TMAFM revealed hexamer vacancies on the (001) terraces of LysB28ProB29, and more rounded dislocation noses and larger terrace widths for actively growing screw dislocations compared to wild-type bovine and porcine insulin crystals under identical conditions. This behavior is consistent with weaker interhexamer attachment energies for LysB28ProB29 at active growth sites. Comparison of the single crystal x-ray structures of wild-type insulins and LysB28ProB29 suggests that differences in protein conformation at the hexamer-hexamer interface and accompanying changes in interhexamer bonding are responsible for this behavior. These studies demonstrate that subtle changes in molecular conformation due to a single sequence inversion in a region critical for insulin self-association can have a significant effect on the crystallization of proteins.     

Effects of the phenylalanine-22----leucine, glutamic acid-49----methionine, glycine-234----aspartic acid, and glycine-234----lysine mutations on the folding and stability of the alpha subunit of tryptophan synthase from Escherichia coli

The effects of four single amino acid replacements on the stability and folding of the alpha subunit of tryptophan synthase from Escherichia coli have been investigated by ultraviolet differences spectroscopy. In previous studies [Miles, E. W., Yutani, K., & Ogasahara, K. (1982) Biochemistry 21, 2586], it had been shown that the urea-induced unfolding at pH 7.8, 25 degrees C, proceeds by the initial unfolding of the less stable carboxyl domain (residues 189-268) followed by the unfolding of the more stable amino domain (residues 1-188). The effects of the Phe-22----Leu, Glu-49----Met, Gly-234----Asp, and Gly-234----Lys mutants on the equilibrium unfolding process can all be understood in terms of the domain unfolding model. With the exception of the Glu-49----Met replacement, the effects on stability are small. In contrast, the effects of three of the four mutations on the kinetics of interconversion of the native form and one of the stable partially folded intermediates are dramatic. The results for the Phe-22----Leu and Gly-234----Asp mutations indicate that these residues play a key role in the rate-limiting step. The Glu-49----Met mutation increases the stability of the native form with respect to that of the intermediate but does not affect the rate-limiting step. The Gly-234----Lys mutation does not affect either the stability or the kinetics of folding for the transition between native and intermediate forms. The changes in stability calculated from the unfolding and refolding rate constants agree quantitatively with those obtained from the equilibrium data. When considered with the results from a previous study on the Gly-211----Glu replacement [Matthews, C. R., Crisanti, M. M., Manz, J. T., & Gepner G. L. (1983) Biochemistry 22, 1445], it can be concluded that the rate-limiting step in the conversion of the intermediate to the native conformation involves either domain association or some other type of molecule-wide phenomenon.     

The use of Fmoc-Lys(Pac)-OH and penicillin G acylase in the preparation of novel semisynthetic insulin analogs.

In this paper, we present the detailed synthetic protocol and characterization of Fmoc-Lys(Pac)-OH, its use for the preparation of octapeptides H-Gly-Phe-Tyr-N-MePhe-Thr-Lys(Pac)-Pro-Thr-OH and H-Gly-Phe-Phe-His-Thr-Pro-Lys(Pac)-Thr-OH by solid-phase synthesis, trypsin-catalyzed condensation of these octapeptides with desoctapeptide(B23-B30)-insulin, and penicillin G acylase catalyzed cleavage of phenylacetyl (Pac) group from Nepsilon-amino group of lysine to give novel insulin analogs [TyrB25, N-MePheB26,LysB28,ProB29]-insulin and [HisB26]-insulin. These new analogs display 4 and 78% binding affinity respectively to insulin receptor in rat adipose membranes.     

Physicochemical basis for the rapid time-action of LysB28ProB29-insulin: dissociation of a protein-ligand complex.

The rate-limiting step for the absorption of insulin solutions after subcutaneous injection is considered to be the dissociation of self-associated hexamers to monomers. To accelerate this absorption process, insulin analogues have been designed that possess full biological activity and yet have greatly diminished tendencies to self-associate. Sedimentation velocity and static light scattering results show that the presence of zinc and phenolic ligands (m-cresol and/or phenol) cause one such insulin analogue, LysB28ProB29-human insulin (LysPro), to associate into a hexameric complex. Most importantly, this ligand-bound hexamer retains its rapid-acting pharmacokinetics and pharmacodynamics. The dissociation of the stabilized hexameric analogue has been studied in vitro using static light scattering as well as in vivo using a female pig pharmacodynamic model. Retention of rapid time-action is hypothesized to be due to altered subunit packing within the hexamer. Evidence for modified monomer-monomer interactions has been observed in the X-ray crystal structure of a zinc LysPro hexamer (Ciszak E et al., 1995, Structure 3:615-622). The solution state behavior of LysPro, reported here, has been interpreted with respect to the crystal structure results. In addition, the phenolic ligand binding differences between LysPro and insulin have been compared using isothermal titrating calorimetry and visible absorption spectroscopy of cobalt-containing hexamers. These studies establish that rapid-acting insulin analogues of this type can be stabilized in solution via the formation of hexamer complexes with altered dissociation properties.     

Effect of Ssc protein mutations on the outer membrane permeability barrier function in Salmonella typhimurium: a study using ssc mutant alleles made by site-directed mutagenesis

We have previously discovered and characterized a novel essential enterobacterial protein, the Ssc protein of Salmonella typhimurium and found that the mutation Val291----Met in this protein inhibits bacterial growth at 42 degrees C and the function of its outer membrane permeability barrier at 37 degrees C [7]. In the present paper we prepared, by site-directed mutagenesis, a series of novel plasmid-encoded Ssc mutant proteins and tested their ability to compensate the loss of wild-type Ssc. The mutant proteins Met288----Lys and Gly289----Asp completely lacked this ability, and accordingly, were very defective. Ssc mutants Met288----Leu, Met290----Lys, and Met292----Lys were partially defective. Mutants Met290----Leu and Met292----Leu were non-defective as were also four randomly made mutant proteins with mutations outside the 288-292 region. The S. typhimurium derivative which contained both the chromosomally encoded Ssc Val291----Met and the plasmid-encoded Ssc Gly289----Asp had an outer membrane defect more severe than that caused by SscMet291 only. The mutant Ssc proteins had very little, if any, effect on the outer membrane function in the presence of wild-type Ssc. Even though the function of Ssc is not yet known, our results indicate that region 288-292 is important and that SscAsp289 is thus far the most defective mutant Ssc.     

The influence of amino substitutions within the mature part of an Escherichia coli outer membrane protein (OmpA) on assembly of the polypeptide into its membrane

The membrane part of the 325-residue outer membrane protein OmpA of Escherichia coli encompasses residues 1-177. This part is thought to cross the membrane eight times in antiparallel beta-strands, forming four loops of an amphipathic beta-barrel. With the aim of gaining some insight into the mechanism of sorting, i.e. the way the protein recognizes and assembles into its membrane, a set of point mutants in the ompA gene has been generated. Selection for toxicity of ompA expression following mutagenesis with sodium bisulfite yielded genes with multiple base pair substitutions, the majority of which resulted in amino acid substitutions in the membrane moiety of the protein. None of the altered proteins was blocked in membrane incorporation. A proline residue exists at or near each of the presumed turns at the inner side of the outer membrane. Using oligonucleotide-directed mutagenesis, each of them was replaced by a leucine residue which is thought to be a turn blocking residue. None of these proteins had lost the ability to be incorporated into the membrane. Apparently, leucine residues are tolerated at turns in this protein. To interfere with the formation of antiparallel beta-strands, four double mutants were prepared: ompA-ON3 (Ala11----Pro, Leu13----Pro), -ON4 (Ala11----Asp, Leu13----Pro), -ON5 (Gly160----Val, Leu162----Arg), and -ON6 (Leu164----Pro, Val166----Asp). The former three proteins and even quadruple mutants consisting of a combination of ompA-ON2 or -ON4 with -ON5 were not defective in membrane assembly. In contrast, the OmpA-ON6 protein was translocated across the plasma membrane but could not be incorporated into the outer membrane. It is concluded that at least one rather small area of the polypeptide is of crucial importance for the assembly of OmpA into the outer membrane.     

Structural and thermodynamic consequences of burying a charged residue within the hydrophobic core of T4 lysozyme.

To determine the energetic and structural consequences of placing a charged group within the core of a protein, two "buried charge" mutants, Met 102----Lys (M102K) and Leu 133----Asp (L133D) were constructed in phage T4 lysozyme. Both proteins fold at neutral pH, although they are substantially less stable than wild type. The activity of M102K is about 35% that of wild type, while that of L133D is about 4%. M102K could be crystallized, and its structure was determined at high resolution. The crystal structure (at pH 6.8) of the mutant is very similar to that of wild type except for the alpha-helix that includes residues 108-113. In wild-type lysozyme, one side of this helix is exposed to solvent and the other contacts Met 102. In the M102K structure this alpha-helix becomes much more mobile, possibly allowing partial access of Lys 102 to solvent. The stability of M102K, determined by monitoring the unfolding of the protein with CD, is pH-dependent, consistent with the charged form of the substituted amino acid being more destabilizing than the uncharged form. The pKa of Lys 102 was estimated to be 6.5 both by differential titration and also by NMR analysis of isotopically labeled protein with 13C incorporated at the C epsilon position of all lysines. As the pH is lowered below pH 6.5, the overall three-dimensional structure of M102K at room temperature appears to be maintained to pH 3 or so, although there is evidence for some structural adjustment possibly allowing solvent accessibility to the protonated form of Lys 102.     

Comparative 2D NMR studies of human insulin and des-pentapeptide insulin: sequential resonance assignment and implications for protein dynamics and receptor recognition

The solution structure and dynamics of human insulin are investigated by 2D 1H NMR spectroscopy in reference to a previously analyzed analogue, des-pentapeptide(B26-B30) insulin (DPI; Hua, Q.X., & Weiss, M.A. (1990) Biochemistry 29, 10545-10555). This spectroscopic comparison is of interest since (i) the structure of the C-terminal region of the B-chain has not been determined in the monomeric state and (ii) the role of this region in binding to the insulin receptor has been the subject of long-standing speculation. The present NMR studies are conducted in the presence of an organic cosolvent (20% acetic acid), under which conditions both proteins are monomeric and stably folded. Complete sequential assignment of human insulin is obtained and leads to the following conclusions. (1) The secondary structure of the insulin monomer (three alpha-helices and B-chain beta-turn) is similar to that observed in the 2-Zn crystal state. (2) The folding of DPI is essentially the same as the corresponding portion of intact insulin, in accord with the similarities between their respective crystal structures. However, differences between insulin and DPI are observed in the extent of conformational broadening of amide resonances, indicating that the presence or absence of residues B26-B30 influences the overall dynamics of the protein on the millisecond time scale. (3) Residues B24-B28 adopt an extended configuration in the monomer and pack against the hydrophobic core as in crystallographic dimers; residues B29 and B30 are largely disordered. This configuration differs from that described in a more organic milieu (35% acetonitrile; Kline, A.D., & Justice, R.M., Jr. (1990) Biochemistry 29, 2906-2913), suggesting that the conformation of insulin in the latter study may have been influenced by solvent composition. (4) The insulin fold is shown to provide a model for collective motions in a protein with implications for the mechanism of protein-protein recognition. To our knowledge, this paper describes the first detailed analysis of a protein NMR spectrum under conditions of extensive conformational broadening. Such an analysis is made possible in the present case by comparative study of an analogue (DPI) with more tractable spectroscopic properties.     

Two-dimensional NMR and photo-CIDNP studies of the insulin monomer: assignment of aromatic resonances with application to protein folding, structure, and dynamics

The aromatic 1H NMR resonances of the insulin monomer are assigned at 500 MHz by comparative studies of chemically modified and genetically altered variants, including a mutant insulin (PheB25----Leu) associated with diabetes mellitus. The two histidines, three phenylalanines, and four tyrosines are observed to be in distinct local environments; their assignment provides sensitive markers for studies of tertiary structure, protein dynamics, and protein folding. The environments of the tyrosine residues have also been investigated by photochemically induced dynamic nuclear polarization (photo-CIDNP) and analyzed in relation to packing constraints in the crystal structures of insulin. Dimerization involving specific B-chain interactions is observed with increasing protein concentration and is shown to depend on temperature, pH, and solvent composition. In the monomer large variations are observed in the line widths of amide resonances, suggesting intermediate exchange among conformational substates; such substates may relate to conformational changes observed in different crystal states and proposed to occur in the hormone-receptor complex. Additional evidence for multiple conformations in solution is provided by comparative studies of an insulin analogue containing a peptide bond between residues B29 and A1 (mini-proinsulin). This analogue forms dimers and higher-order oligomers under conditions in which native insulin is monomeric, suggesting that the B29-A1 peptide bond stabilizes a conformational substate favorable for dimerization. Such stabilization is not observed in corresponding studies of native proinsulin, in which a 35-residue connecting peptide joins residues B30 and A1; this extended tether is presumably too flexible to constrain the conformation of the B-chain. The differences between proinsulin and mini-proinsulin suggest a structural mechanism for the observation that the fully reduced B29-A1 analogue folds more efficiently than proinsulin to form the correct pattern of disulfide bonds. These results are discussed in relation to molecular mechanics calculations of insulin based on the available crystal structures.     

Hierarchical protein "un-design": insulin's intrachain disulfide bridge tethers a recognition alpha-helix

A hierarchical pathway of protein folding can enable segmental unfolding by design. A monomeric insulin analogue containing pairwise substitution of internal A6-A11 cystine with serine [[Ser(A6),Ser(A11),Asp(B10),Lys(B28),Pro(B29)]insulin (DKP[A6-A11](Ser))] was previously investigated as a model of an oxidative protein-folding intermediate [Hua, Q. X., et al. (1996) J. Mol. Biol. 264, 390-403]. Its structure exhibits local unfolding of an adjoining amphipathic alpha-helix (residues A1-A8), leading to a 2000-fold reduction in activity. Such severe loss of function, unusual among mutant insulins, is proposed to reflect the cost of induced fit: receptor-directed restoration of the alpha-helix and its engagement in the hormone's hydrophobic core. To test this hypothesis, we have synthesized and characterized the corresponding alanine analogue [[Ala(A6),Ala(A11),Asp(B10),Lys(B28), Pro(B29)]insulin (DKP[A6-A11](Ala))]. Untethering the A6-A11 disulfide bridge by either amino acid causes similar perturbations in structure and dynamics as probed by circular dichroism and (1)H NMR spectroscopy. The analogues also exhibit similar decrements in thermodynamic stability relative to that of the parent monomer as probed by equilibrium denaturation studies (Delta Delta G(u) = 3.0 +/- 0.5 kcal/mol). Despite such similarities, the alanine analogue is 50 times more active than the serine analogue. Enhanced receptor binding (Delta Delta G = 2.2 kcal/mol) is in accord with alanine's greater helical propensity and more favorable hydrophobic-transfer free energy. The success of an induced-fit model highlights the applicability of general folding principles to a complex binding process. Comparison of DKP[A6-A11](Ser) and DKP[A6-A11](Ala) supports the hypothesis that the native A1-A8 alpha-helix functions as a preformed recognition element tethered by insulin's intrachain disulfide bridge. Segmental unfolding by design provides a novel approach to dissecting structure-activity relationships.     

1H NMR spectrum of the native human insulin monomer. Evidence for conformational differences between the monomer and aggregated forms

The effects of high dilution on the 1H Fourier transform NMR spectrum of native human insulin at pH* 8.0 and 9.3 have been examined at 500 MHz resolution. The dependence of the spectrum on concentration and comparison with the spectrum of a biologically highly potent monomeric insulin mutant (SerB9----Asp) establish that at 36 microM (pH* 9.3) or 18 microM (pH* 8) and no added buffer or salts, human insulin is monomeric. Under these conditions of dilution, ionic strength, and pH*, human insulin and the SerB9----Asp mutant exhibit nearly identical 1H NMR spectra. At higher concentrations (i.e. greater than 36 microM to 0.91 mM), native human insulin dimerizes, and this aggregation causes a change in insulin conformation. Although there are many changes in the spectrum, the TyrB26 ring H3,5 proton signals located at 6.63 ppm and the methyl signal located at 0.105 ppm (characteristics of monomeric insulin) are particularly distinct signatures of the conformation change that accompanies dimerization. Magnetization transfer experiments show that the 0.105 ppm methyl signal shifts downfield to a new position at 0.45 ppm. We conclude that the 0.105 ppm methyl signal is due to a conformation in which a Leu methyl group is centered over and in van der Waals contact with the ring of an aromatic side chain. Dimerization causes a conformation change that alters this interaction, thereby causing the downfield shift. Nuclear Overhauser studies indicate that the methyl group involved is located within a cluster of aromatic side chains and that the closest ring-methyl group interaction is with the ring of PheB24.     

Mutagenesis of the NH2-terminal domain of elongation factor Tu

Mutagenesis was carried out in the N-terminal domain of elongation factor Tu (EF-Tu) to characterize the structure-function relationships of this model GTP binding protein with respect to stability, the interaction with GTP and GDP, and the catalytic activity. The substitutions were introduced in elements around the guanine nucleotide binding site or in the loops defining this site, in the intact molecule or in the isolated N-terminal domain (G domain). The double substitution Val88----Asp and Leu121----Lys, two residues situated on two vicinal alpha-helices, influences the basic activities of the truncated factor to a limited extent, probably via long-range interactions, and induces a destabilisation of the G domain structure. The functional alterations brought about by substitutions on the consensus sequences 18-24 and 80-83 highlight the importance of these residues for the interaction with GTP/GDP and the GTPase activity. Mutations concerning residues interacting with the guanine base lead to proteins in large part insoluble and inactive. In one case, the mutated protein (EF-TuAsn135----Asp) inhibited the growth of the host cell. This demonstrates the crucial role of the base specificity for the active conformation of EF-Tu. The obtained results are discussed in the light of the three-dimensional structure of EF-Tu.     

Inhibition of agonist-induced platelet aggregation, Ca2+ mobilization and granule secretion by guanosine 5''-[beta-thio]diphosphate and GDP in intact platelets. Evidence for an inhibitory mechanism unrelated to the inhibition of G-protein-GTP interaction.

(1) We [Muir, Offord & Davies (1986) Biochem. J. 237, 631-637 and Davies, Muir & Offord (1986) Biochem. J. 240, 609-612] have previously identified a major product in the degradation of insulin by insulin proteinase (the N-terminal fragment produced by cleavage between residues LeuA13 and TyrA14, SerB9 and HisB10) together with evidence for a minor cleavage site between HisB10 and LeuB11 or between LeuB11 and ValB12. (2) We now present evidence for minor sites of cleavage between TyrA14 and GlnA15, GluB13 and AlaB14 as well as HisB10 and LeuB11.     

An insulin analogue with gamma-amino butyric acid substitution for A13Leu-A14Tyr.

An insulin A chain analogue, [A13-14 GABA, A21 Ala]A chain, for which the dipeptide Leu-Try at A13-A14 was substituted by a non-coded amino acid, gamma-amino butyric acid (GABA) and A21 Asn by Ala, was prepared by stepwise Fmoc solid-phase manual synthesis and then combined with the natural B chain of porcine insulin to yield an insulin analogue, [A13-14 GABA, A21Ala] porcine insulin (GABA substituted insulin). This insulin analogue still retains 50% in vivo biological activity and 59% in receptor binding capacity. It can also be crystallized. These results indicate that its overall conformation is similar to the native form and that the side chains of A13Leu and A14Tyr are not essential for insulin activity. In addition, the replacement of a normal C-N peptide bond by an unnatural C-C bond may have general meaning in structure and function studies of other proteins.     

Toward the solution structure of human insulin: sequential 2D 1H NMR assignment of a des-pentapeptide analogue and comparison with crystal structure

2D 1H NMR studies are presented of des-pentapeptide-insulin, an analogue of human insulin lacking the C-terminal five residues of the B chain. Removal of these residues, which are not required for function, is shown to reduce conformational broadening previously described in the spectrum of intact insulin [Weiss et al. (1989) Biochemistry 28, 9855-9873]. This difference presumably reflects more rapid internal motions in the fragment, which lead to more complete averaging of chemical shifts on the NMR time scale. Sequential 1H NMR assignment and preliminary structural analysis demonstrate retention in solution of the three alpha-helices observed in the crystal state and the relative orientation of the receptor-binding surfaces. These studies provide a foundation for determining the solution structure of insulin.     

Synthesis and evaluation of photolabile insulin prodrugs

We have developed two photolabile insulin prodrugs, insulin-2P and insulin-3P. These prodrugs were synthesized by protecting GlyA1 (N(alphaA1)), and one or both of the PheB1 (N(alphaB1)) and LysB29 (N(epsilonB29)) amino groups in insulin using 5'-(alpha-methyl-nitro-piperonyl)oxy-carbonyl as the protecting group. These insulin prodrugs were efficiently activated by exposure to longwave UV light to produce insulin quantitatively. Using 2-deoxyglucose uptake assays, both di- and tri-protected compounds were less active than native insulin in the protected state, and showed comparable activity to native insulin upon photoactivation.     

Crystal structure of allo-Ile(A2)-insulin, an inactive chiral analogue: implications for the mechanism of receptor binding

The crystal structure of an inactive chiral analogue of insulin containing nonstandard substitution allo-Ile(A2) is described at 2.0 A resolution. In native insulin, the invariant Ile(A2) side chain anchors the N-terminal alpha-helix of the A-chain to the hydrophobic core. The structure of the variant protein was determined by molecular replacement as a T(3)R(3) zinc hexamer. Whereas respective T- and R-state main-chain structures are similar to those of native insulin (main-chain root-mean-square deviations (RMSD) of 0.45 and 0.54 A, respectively), differences in core packing are observed near the variant side chain. The R-state core resembles that of the native R-state with a local inversion of A2 orientation (core side chain RMSD 0.75 A excluding A2); in the T-state, allo-Ile(A2) exhibits an altered conformation in association with the reorganization of the surrounding side chains (RMSD 0.98 A). Surprisingly, the core of the R-state is similar to that observed in solution nuclear magnetic resonance (NMR) studies of an engineered T-like monomer containing the same chiral substitution (allo-Ile(A2)-DKP-insulin; Xu, B., Hua, Q. X., Nakagawa, S. H., Jia, W., Chu, Y. C., Katsoyannis, P. G., and Weiss, M. A. (2002) J. Mol. Biol. 316, 435-441). Simulation of NOESY spectra based on crystallographic protomers enables the analysis of similarities and differences in solution. The different responses of the T- and R-state cores to chiral perturbation illustrates both their intrinsic plasticity and constraints imposed by hexamer assembly. Although variant T- and R-protomers retain nativelike protein surfaces, the receptor-binding activity of allo-Ile(A2)-insulin is low (2% relative to native insulin). This seeming paradox suggests that insulin undergoes a change in conformation to expose Ile(A2) at the hormone-receptor interface.     

How insulin binds: the B-chain alpha-helix contacts the L1 beta-helix of the insulin receptor

Binding of insulin to the insulin receptor plays a central role in the hormonal control of metabolism. Here, we investigate possible contact sites between the receptor and the conserved non-polar surface of the B-chain. Evidence is presented that two contiguous sites in an alpha-helix, Val(B12) and Tyr(B16), contact the receptor. Chemical synthesis is exploited to obtain non-standard substitutions in an engineered monomer (DKP-insulin). Substitution of Tyr(B16) by an isosteric photo-activatable derivative (para-azido-phenylalanine) enables efficient cross-linking to the receptor. Such cross-linking is specific and maps to the L1 beta-helix of the alpha-subunit. Because substitution of Val(B12) by larger side-chains markedly impairs receptor binding, cross-linking studies at B12 were not undertaken. Structure-function relationships are instead probed by side-chains of similar or smaller volume: respective substitution of Val(B12) by alanine, threonine, and alpha-aminobutyric acid leads to activities of 1(+/-0.1)%, 13(+/-6)%, and 14(+/-5)% (relative to DKP-insulin) without disproportionate changes in negative cooperativity. NMR structures are essentially identical with native insulin. The absence of transmitted structural changes suggests that the low activities of B12 analogues reflect local perturbation of a "high-affinity" hormone-receptor contact. By contrast, because position B16 tolerates alanine substitution (relative activity 34(+/-10)%), the contribution of this neighboring interaction is smaller. Together, our results support a model in which the B-chain alpha-helix, functioning as an essential recognition element, docks against the L1 beta-helix of the insulin receptor.     

Synthesis and characterization of poly(ethylene glycol)-insulin conjugates

Human insulin was modified by covalent attachment of short-chain (750 and 2000 Da) methoxypoly (ethylene glycol) (mPEG) to the amino groups of either residue PheB1 or LysB29, resulting in four distinct conjugates: mPEG(750)-PheB1-insulin, mPEG(2000)-PheB1-insulin, mPEG(750)-LysB29-insulin, and mPEG(2000)-LysB29-insulin. Characterization of the conjugates by MALDI-TOF mass spectrometry and N-terminal protein sequence analyses verified that only a single polymer chain (750 or 2000 Da) was attached to the selected residue of interest (PheB1 or LysB29). Equilibrium sedimentation experiments were performed using analytical ultracentrifugation to quantitatively determine the association state(s) of insulin derivatives. In the concentration range studied, all four of the conjugates and Zn-free insulin exist as stable dimers while Zn(2+)-insulin was exclusively hexameric and Lispro was monomeric. In addition, insulin (conjugate) self-association was evaluated by circular dichroism in the near-ultraviolet wavelength range (320-250 nm). This independent method qualitatively suggests that mPEG-insulin conjugates behave similarly to Zn-free insulin in the concentration range studied and complements results from ultracentrifugation studies. The physical stability/resistance to fibrillation of mPEG-insulin conjugates in aqueous solution were assessed. The data proves that mPEG(750 and 2000)-PheB1-insulin conjugates are substantially more stable than controls but the mPEG(750 and 2000)-LysB29-insulin conjugates were only slightly more stable than commercially available preparations. Circular dichroism studies done in the far ultraviolet region confirm insulin's tertiary structure in aqueous solution is essentially conserved after mPEG conjugation. In vivo pharmacodynamic assays reveal that there is no loss in biological activity after conjugation of mPEG(750) to either position on the insulin B-chain. However, attachment of mPEG(2000) decreased the bioactivity of the conjugates to about 85% of Lilly's HumulinR formulation. The characterization presented in this paper provides strong testimony to the fact that attachment of mPEG to specific amino acid residues of insulin's B-chain improves the conjugates' physical stability without appreciable perturbations to its tertiary structure, self-association behavior, or in vivo biological activity.