Simulation of a complex protein structural change: the T {leftrightarrow} R transition in the insulin hexamer

The T R transition in the insulin hexamer is an outstandingmodel for protein structural changes in terms of its extentand complexity: the limiting structures T₆, T₃R₃ and R₆ havebeen defined by X-ray crystallography. The transition occurscooperatively within trimers. It involves displacements of >30Å and a secondary structural rearrangement of 15% of thepeptide chain between extended and helical conformations. Experimentaldata for the transition are plentiful. Theoretical methods tosimulate pathways without constraints would never succeed withsuch substantial transitions. We have developed two approaches,targeted energy minimization (TEM) and targeted molecular dynamics(TMD). Previously successful in simulating the T R transitionof the insulin monomer, these procedures are also shown hereto be effective in the hexamer. With TMD, more conformationalspace is explored and pathways are found at 500 kJ/mol lowerenergy than with TEM. Because the atoms have to meet distanceconstraints in sum rather than individually, a high degree ofconformational freedom and independence. is implied. T₆ T₃R₃and T₃R₃ T₆ pathways do not coincide because the transformationis directed. One subunit enters a dead end pathway in one directionof the TMD simulation, which shows that constraint and freedomare critically balanced. The ensemble of productive pathwaysrepresents a plausible corridor for the transition. A videodisplay of the transformations is available     

Simulations of the T {leftrightarrow} R conformational transition in aspartate transcarbamylase

Aspartate transcarbamylase (ATCase) from Escherichia coli isone of the best known allosteric enzymes. In spite of numerousexperiments performed by biochemists, no consensus model forthe cooperative transition between the tensed (T) and the relaxed(R) forms exists. It is hypothesized, however, that changesin the quaternary structure play a key role in the allostericproperties of oligomeric proteins such as ATCase. Previous normalmode calculations of the two states of ATCase illustrated thetype of motions that could be important in initiating the transition.In this work four pathways for the transition were calculatedusing the targeted molecular dynamics (TMD) method without constrainton the symmetry of the system. The most important quaternarystructure changes are the relative rotation and translationof the catalytic trimers and the rotations of the regulatorydimers. The simulations show that these quaternary changes startimmediately and finish when about 70% of the transition is completedwhereas there are tertiary changes throughout the transition.In agreement with the work of Lipscomb et al., it was foundthat the relative translation between the catalytic trimersappears to play a central role in allowing the transition tooccur. In all the simulations differences are observed in theopening and closing behaviours of the domains in the catalyticand regulatory chains that could provide a structural interpretationfor the results of certain site-directed mutagenesis experiments.Overall the motions of the subunits are concerted even thoughthe constraint imposed on the TMD method does not explicitlyrequire that this be so.     

Energetic approach to the folding of four {alpha}-helices connected sequentially

The packing of four -helices, which each consist of 12 Ala residuesand are sequentially connected to each other by a segment of10 Ala residues, has been investigated by means of energy minimizations.For the lowest energy structure thus obtained, the followingfeatures have been found: (i) the four -helices are intimatelypacked to form an assembly with an approximately square section;(ii) the distances of closest approach between two adjacentinterhelix axes are 7.7±0.2 Å and those betweentwo diagonal interhelix axes are 11.2±0.2 Å; (iii)the adjacent interhelix angles are -163±2°; and (iv)the diagonal interhelix angles are 24±4°. These resultsindicate that the polypeptide chain, driven by energetics (nonbondedand electrostatic interactions), is folded into a typical left-handedtwisted four-helix bundle with an {small tilde}4-fold symmetricarray, as observed in most four -helix proteins. Furthermore,it has been found that the interaction between the loops formedby the connecting segments and the other part of molecule playsa significant role in stabilizing such a bundle structure. Thetechnology developed here and the relevant knowledge obtainedthrough this study are very useful for the study of modelingfour-helix bundle proteins.     

Restoration of pore-forming activity in staphylococcal {alpha}-hemolysin by targeted covalent modification

In previous studies, the replacement of His35 in the pore-formingprotein -hemolysin (HL) with Leu, lie, Pro, Arg, Ser, Thr orCys yielded inactive polypeptides. Here, we show that modificationof the inactive single-cysteine mutant HL-H35C with iodoacetamide,to form H35CamC, generates significant pore-forming activity.The closely related polypeptides H35N and H35Q have, respectively,essentially no activity and greatly reduced activity. The modifiedresidue in H35CamC, S-carboxamidomethylcysteine, mimicks histidinein volume, polarity and hydrogen bonding potential, but is unableto ionize. Unmodified H35C is defective in the final step ofpore formation: the conversion of an inactive heptameric membrane-boundassembly intermediate to a structure containing open channels.It is this step in assembly, that is ameliorated in H35CamC.     

The semisynthesis of octadeutero-PheB1-octadeutero-ValB2]-procine insulin and its characterization by mass spectrometry

Insulin analogues labelled with stable isotopes (e.g. deuterium,¹⁸O, ¹⁵N, etc.) are authentic (the native structure is rigorouslymaintained), non-radioactive (preferred for injection into man)and can easily be distinguished from endogenous insulin by massspectrometry by virtue of their molecular masses. Appropriatecombinations of amino-protecting groups (methylsulphonylethyloxycarbonyland t-butoxy carbonyl), Edman degradation and chemical couplingwere used to produce [octadeutero-Phe^B1]-porcine insulin and[octadeutero-Phe^B1-octadeutero-Val^B2]-porcine insulin. The analogueswere characterized by electrospray ionization mass spectrometry.Standard mixtures of labelled and unlabelled insulins were successfullystudied by mass spectrometry. Isotope dilution mass spectrometrycould therefore provide a useful direct measure of insulin undertrue physiological conditions, without many of the drawbacksof existing methods. In this regard, the analogue with 16 deuteriumswas more suitable than the octadeuterated analogue, since thegreater mass difference between the labelled and unlabelledforms enabled a lower mass spectrometric resolution to be used,resulting in higher sensitivity     

Reaction mechanism of trypsin-catalysed semisynthesis of human insulin studied by fast atom bombardment mass spectrometry

The production of semisynthetic human insulin for therapeuticpurposes is of considerable importance. During trypsin-catalysedtransformation of pig insulin into an ester of insulin of humansequence, the alanyl residue at position B³⁰ is removed andreplaced with an esterified residue of threonine. We have carriedout this transformation in a medium enriched in ¹⁸OH₂ and studiedthe product by MS. In contrast to a previous report, we findthat incorporation of label into the B²⁹–B³⁰ peptide bondoccurs during the transformation with threonine methyl esterin aqueous N, N-dimethylacetamide. Quantitative data are presentedand the implications of these findings are discussed.     

The elusive role of the N-terminal extension of {beta}A3- and {beta}Al-crystallin

ß-Crystallins are structural lens proteins with aconserved two-domain structure and variable N- and C-terminalextensions. These extensions are assumed to be involved in quaternaryinteractions within the ß-crystallin oligomers orwith other lens proteins. Therefore, the production of ßA3-and ßAl-crystallin from the single ßA3/A1mRNA by dual translation initiation is of interest. These crystallinsare identical, except that ßAl has a much shorterN-terminal extension than ßA3. This rare mechanismhas been conserved for over 250 million years during the evolutionof the ßA3/A1 gene, suggesting that the generationof different N-terminal extensions confers a selective advantage.We therefore compared the stability and association behaviourof recombinant ßA3- and ßAl-crystallin.Both proteins are equally stable in urea- and pH-induced denaturationexperiments. Gel filtration and analytical ultracentrifugationestablished that ßA3 and ßA1 both form homodimers.In the water-soluble proteins of bovine lens, ßA3and ßA1 are present in the same molecular weight fractions,indicating that they oligomerize equally with other ß-crystallins.¹H-NMR spectroscopy showed that residues Met1 to Asn22 of theN-terminal extension of ßA3 have great flexibilityand are solvent exposed, excluding them from protein interactionsin the homodimer. These results indicate that the differentN-terminal extensions of ßA3 and ßA1 donot affect their homo- or heteromeric interactions.     

Light-induced denaturation of bacteriorhodopsin solubilized by octyl-{beta}-glucoside

The structural stability of bacteriorhodopsin (bR) solubilizedby octyl-ß-glucoside was studied by measuring the denaturationkinetics under visible light irradiation and in the dark. Thedenaturation of bR solubilized by 50 mM octyl-ß-glucosidewas very slow at room temperature when it was left in the dark.However, its spontaneous denaturation was accelerated when thesolubilized bR was irradiated by visible light. The denaturationkinetics under visible light irradiation and in the dark couldbe well described by a single decay constant. The activationenergy for the denaturation of bR was estimated from the temperaturedependence of decay time constants. The activation energy undervisible light irradiation was 12.5 kcal/mol, which was muchsmaller than the corresponding value in the dark, 26.2 kcal/mol.These results strongly suggest that some of the photointermediatestates are less stable than the ground state of bR. The criticaltemperature and the activation energy for denaturation of bRin the solubilized state were much lower than those in the 2Dcrystalline state. Comparing the denaturation behavior in the2D crystalline state and that in the octyl-ß-glucoside-solubilizedstate, our findings suggest that protein–protein interactioncontributes to the stability of this protein.     

Display of {beta}-lactamase on the Escherichia coli surface: outer membrane phenotypes conferred by Lpp'-OmpA'-{beta}-lactamase fusions

Bacterial cell-surface exposure of foreign peptides and solubleproteins has been achieved recently by employing a fusion proteinmethodology. An Lpp'–OmpA(46–159)–Bla fusionprotein has been shown previously to display the normally periplasmicenzyme ß-lactamase (Bla) on the cell surface of theGram-negative bacterium Escherichia coli. Here, we have investigatedthe role of the OmpA domain of the tripartite fusion proteinin the surface display of the passenger domain (Bla) and havecharacterized the effects of the fusion proteins on the integrityand permeability of the outer membrane. We show that in additionto OmpA(46–159), a second OmpA segment, consisting ofamino acids 46–66, can also mediate the display of Blaon the cell surface. Other OmpA domains of various lengths (aminoacids 46–84, 46–109, 46–128, 46–141and 46–145) either anchored the Bla domain on the periplasmicface of the outer membrane or caused a major disruption of theouter membrane, allowing the penetration of antibodies intothe cell. Detergent and antibiotic sensitivity and periplasmicleakage assays showed that changes in the permeability of theouter membrane are an unavoidable consequence of displayinga large periplasmic protein on the surface of E.coli. This isthe first systematic report on the effects that cell surfaceengineering may have on the integrity and permeability propertiesof bacterial outer membranes.     

Improvement of turn structure prediction by molecular dynamics: a case study of {alpha}1-purothionin

Because of the problems in predicting a correct conformationfor loop regions in homology-based prediction, disagreementsare often found between the predicted models and the refinedX-ray structures of the same protein in loop regions. Such asituation has been encountered for ₁-purothionin (₁-PT). Hence,attempts have been made to improve the predicted model of ₁PTby limited molecular dynamics using both AMBER and XPLOR. Withmolecular dynamics, the previously predicted incorrect turnregion reverts to the correct conformation as seen in the X-rayrefined structure. In contrast to the model which is not subjectedto molecular dynamics, the improved model refines with the X-raydata of ₁PT in fewer cycles, without any manual rebuilding andwith comparable or better refinement statistics. Also, the improvedmodel serves as a better starting model in the determinationof the structure with the molecular replacement methods.     

The preparation of porcine (Tyr-A14-2H) insulin and its characterization by mass spectrometry

The title compound was prepared by catalytic deiodination of(Tyr-A¹⁴-3-I) insulin with deuterium gas. Under special conditions(large excess of PdO in CH₃OD/D₂O at pH 9.2) developed to minimizeproblems due to poisoning of the catalyst, deuterated insulinwas produced in yields of about 35% after purification by reversedphase HPLC. For analysis, the deuterated insulin was digestedwith V8 protease and was shown by mass spectrometry to haveincorporated deuterium to an extent of 96.6 atom %, exclusivelyin a pentapeptide containing Tyr A¹⁴. The title compound shouldprove useful to those workers studying insulin by mass spectrometry,and use of the method with tritium gas in place of deuteriumshould permit the preparation of a specifically labelled radioactiveinsulin analogue which behaves identically to the natural hormone.     

The structure of a designed peptidomimetic inhibitor complex of {alpha}-thrombin

Thrombin displays remarkable specificity, effecting the removalof fibrinopeptides A and B of fibrinogen through the selectivecleavage of two Arg–Gly bonds between the 181 Arg/Lys–Xaabonds in fibrinogen. Significant advances have been made inrecent years towards understanding the origin of the specificityof cleavage of the Argl6–Gly17 bond of the A-chain ofhuman fibrinogen. We have previously proposed a model for thebound structure of fibrinopeptide A_7–16 (FPA), based uponNMR data, computer-assisted molecular modeling and the synthesisand study of peptidomimetic substrates and inhibitors of thrombin.We now report the structure of the ternary complex of an FPAmimetic (FPAM), hirugen and thrombin at 2.5 Å resolution(R-factor = 0.138) and specificity data for the inhibition ofthrombin and related trypsin-like proteinases by FPAM. The crystallographicstructures of FPA and its chloromethyl ketone derivative boundto thrombin were determined. Although there are differencesbetween these structures in the above modeled FPA structureand that of the crystal structure of FPAM bound to thrombin,the , angles in the critical region of P1–P2–P3in all of the structures are similar to those of bovine pancreatictrypsin inhibitor (BPTI) in the BPTI–trypsin complex andD–Phe–Pro–Arg (PPACK) in the PPACK–thrombinstructure. A comparison between these and an NMR-derived structureis carried out and discussed.     

Inhibition of {beta}S-chain dependent polymerization by synergistic complementation of contact site perturbations of {alpha}-chain: application of semisynthetic chimeric {alpha}-chains

Mouse _1–30-horse _31–141 chimeric -chain, a semisyntheticsuper-inhibitory -chain, inhibits ß^S-chain dependent polymerizationbetter than both parent -chains. Although contact site sequencedifferences are absent in the _1–30 region of the chimericchain, the four sequence differences of the region _17-22 couldinduce perturbations of the side chains at ₁₆, ₂₀ and ₂₃, thethree contact sites of the region. A synergistic complementationof such contact site perturbation with that of horse _31–141probably results in the super-inhibitory activity of the chimeric-chain. The inhibitory contact site sequence differences, bythemselves, could also exhibit similar synergistic complementation.Accordingly, the polymerization inhibitory activity of Hb Le-Lamentin(LM) mutation [His20()Gln], a contact site sequence difference,engineered into human–horse chimeric -chain has been investigatedto map such a synergistic complementation. Gln20() has littleeffect on the O₂ affinity of HbS, but in human–horse chimeric-chain it reduces the O₂ affinity slightly. In the chimeric-chain, Gln20() increased sensitivity of the ßßcleft for the DPG influence, reflecting a cross-talk betweenthe ₁ß₁ interface and ßß cleft in this semisyntheticchimeric HbS. In the human -chain frame, the polymerizationinhibitory activity of Gln20() is higher compared with horse_1–30, but lower than mouse _1–30. Gln20() synergisticallycomplements the inhibitory propensity of horse _31–141.However, the inhibitory activity of LM–horse chimeric-chain is still lower than that of mouse–horse chimeric-chain. Therefore, perturbation of multiple contact sites inthe _1–30 region of the mouse–horse chimeric -chainand its linkage with the inhibitory propensity of horse _31–141has been now invoked to explain the super-inhibitory activityof the chimeric -chain. The `linkage-map' of contact sites canserve as a blueprint for designing synergistic complementationof multiple contact sites into -chains as a strategy for generatingsuper-inhibitory antisickling hemoglobins for gene therapy ofsickle cell disease.     

Strong inhibition of fibrinogen binding to platelet receptor {alpha}IIb{beta}3 by RGD sequences installed into a presentation scaffold

In order to probe the structural constraints on binding of RGDsequences to the platelet receptor _IIbß₃ we have usedrecombinant DNA techniques to install the RGD sequence into‘presentation scaffolds’, small proteins of known3-D structure chosen to present guest sequences in constrainedorientations. Using Escherichia coli expression systems we madesequence variants in which loop residues of the immunoglobulinV_L domain REI and of human interleukin-1ß were replaced(without changing polypeptide length) by the RGD sequence atpositions predicted, based on small molecule studies, to orientthe RGD moiety into an active conformation. These variants donot compete for fibrinogen binding to _IIbß₃ up toalmost 1 mM concentration. Unfolded or proteolytically fragmentedforms of these same proteins do compete, however, showing thatthe RGD sequences in the mutants must be prohibited from bindingby constraints imposed by scaffold structure. To suppress theeffects of such structural constraints we constructed two sequencevariants in which RGD-containing sequences 42–57 or 44–55from the snake venom platelet antagonist kistrin were inserted(this increasing the length of the loop) into the third complementaritydetermining loop of REI. Both of these variants compete stronglyfor fibrinogen binding with IC₅₀s in the nM range. These results,plus data on kistrin-related peptides also presented here, suggestthat the molecular scaffold REI is capable of providing to aninstalled sequence a structural context and conformation beneficialto binding. The results also suggest that in order to bind wellto _IIbß₃, RGD sequences in protein ligands must eitherproject significantly from the surface of the scaffold and/orretain a degree of conformational flexibility within the scaffold.Molecular scaffolds like REI should prove useful in the elucidationof structure-function relationships and the discovery of newactive sequences, and may also serve as the basis for noveltherapeutic agents.     

Molecular modeling of coiled-coil {alpha}-tropomyosin: analysis of staggered and in register helix--helix interactions

In register and staggered models of tropomyosin coiled-coilwere built from X-ray C coordinates and refined via moleculardynamics. The two models show similar structural features withthe X-ray structure of GCN4 leucine zipper. Empirical energeticmethods used to compare the in register and staggered modelsindicate that both are equally probable. The two models havesimilar profiles of solvation free energy of folding for residuesat positions a and d of the repeating heptad, indicating thatresidues at these positions are as well buried in an in registerstructure as in a staggered one. Neither the in register northe 14 residues staggered structure can be ruled out based onhydrophobic or e–g' (g–e') electrostatic interactionswhich are not able to distinguish between the two models andare therefore not selective. However, the eg–b'c' electrostaticinteractions, although smaller in magnitude, are in favor ofthe in register model. Furthermore, analysis of hydrophobicand electrostatic interactions along the tropomyosin sequenceshows that bulky residues in positions a and d prevent the formationof inter-chain salt bridges.     

The role of the sequence extensions in {beta}-crystallin assembly

The modular construction of the eye lens ß-crystallinsmakes them good candidates for protein engineering to ascertainthe rules of assembly of oligomers. X-ray studies have shownthat although the polypeptide chains of ßB2-crystallinand -crystallins fold to form similar N- and C-terminal domains,the conformation of the connecting peptides are such that the-crystallins are monomers and the ß-crystallin isa dimer. Unlike -crystallins, the numerous -crystallins haveextensions of variable sequence from the globular domains. Wehave tested the effect of removing the N- and C-terminal extensionsfrom rat ßB2-crystallin using a bacterial expressionsystem. Abundant proteins were produced in Escherichia coliusing the pET or pQE vectors. Full-length and truncated proteinswere purified and checked for refolding using circular dichroism.Sizing of the truncated proteins using gel filtration chroma-tographyshowed that the absence of either the N- or C-terminal extensiondoes not affect dimerization of ßB2-crystallin.     

Bipartite organization of the Bacillus subtilis endo-{beta}-1,4-glucanase revealed by C-terminal mutations

The C-terminal boundary of primary sequence of the Bacillussubtilis PAP115 endo-ß-1,4-glucanase (EG) requiredfor stable catalytic activity has been mapped by site-directedmutagenesis using Escherichia coli as host. The 52 kDa cel geneproduct, EG₄₇₀ and a 33 kDa mutant (EG₃₀₀), lacking 170 residuesthrough a nonsense mutation at the leucine-330 codon of thegene, exhibited similar patterns of enzymatic activity and pHoptima using cellooligopentaose as substrate.CD spectra indicatedthat the bulk of the -helical secondary structure in EG₄₇₀ wascontained within EG₃₀₀. However, relative to EG₄₇₀, the specificactivity of EG₃₀₀ was 3- to 4-fold lower with amorphous celluloseas substrate and {small tilde}4-to5-fold higher with carboxymethylcellulose(soluble cellulose).These results along with data which showthat EG₄₇₀ binding capacity to mirocrystalline cellulose is{small tilde} 11 times more than that of EG₃₀₀, demonstratethe importance of residues 330–499 for non-catalytic bindingof cellulose. A construct of the cel gene carrying a deletionof codons 330–499 and an insertion of a nonsense codonat leucine-330, was further used to make mutants EG₂₉₆ and EG₂₉₁with nonsense codon substitutions at arginine and serine-321,respectively.Western analysis using EG-specific antiserum revealedthat relative losses in enzymatic activity of EG₂₉₆ (50%) andEG₂₉₁ (95%) could be accounted for by the extent of their proteolysis,signifying a marked destabilization of these enzymes by removalof only a few amino acids.     

Calcium binding mode of {gamma}-carboxyglutamic acids in conantokins

Conantokin-T (con-T) and conantokin-G (con-G) are two highlyhomologous peptide toxins found in Conus venom. The former isa 21-residue peptide with four -carboxyglutamic acid (Gla) residues(at positions 3, 4, 10 and 14), while the latter is a 17-residuepeptide with five -carboxyglutamic acid residues (at positions3, 4, 7, 10 and 14). Despite the apparent similarity in numberand relative positions of the -carboxyglutamic acid residues,¹¹³Cd-NMR studies indicated a distinct metal binding behaviorfor con-G and con-T. There appears to be four binding sitesin con-G in contrast to one metal binding site in con-T. Toelucidate the mode of calcium binding by the -carboxyglutamicacid residues in these conantokins, we designed various analogouspeptides with their -carboxyglutamic acid replaced by otheramino acid residues. ¹¹³Cd-NMR experiments on conantokin analoguesreveal that the major difference in the number of metal bindingsites between con-G and con-T is due to the residue at position7. We also performed molecular simulations to calculate therelative binding free energies of several potential bindingsites. Based on our theoretical and experimental results, wepropose a `four-site' binding model for conantokin-G and a `single-site'binding model for conantokin-T.     

6-Phospho-{beta}-galactosidases of Gram-positive and 6-phospho-{beta}-glucosidase B of Gram-negative bacteria: comparison of structure and function by kinetic and immunological methods and mutagenesis of the lacG gene of staphylococcus aureus

The 6-phospho-ß-galactosidase of Staphylococcus aureus,Lactococcus lactis and Lactobacillus casei and 6-phospho-ßglucosidaseB of Escherichia coli build a subfamily inside a greater enzymefamily, named the glycosal hydrolase family 1, which, hi addition,contains nine ß-glycosidases of different origins.Kinetic and immunological evidence is provided in this reportwhich strengthens the relationship of the four 6-phospho-ß-glycosidases.It is shown that the 6-phospho-ß-galactosidases and6-phospho-ß-glucosidase B are able to split aromaticß-galactoside phosphates and ß-glucosidephosphates. The turnover numbers of hydrolysis of substrateswith different epimerization at C-4 of the glycon vary up to15-fold only. Two polydonal antisera, one derived against thenative 6-phospho-ß-galactosidase from S.aureus andthe other derived against the 6-phospho-ß-glucosidaseB, cross-reacted with both enzymes. Peptides of the proteinswere separated by reverse phase HPLC. The cross-reacting peptideswere sequenced and shown to be localized at almost the sameposition in the aligned primary structures of both enzymes.An insertion of nine amino adds near these antigenic domainsis unique for the 6-phospho-ß-glycosidases and missingwithin the sequences of the ß-glycoside-specific membersof the family. The lacG gene of a 6-phospho-ß-galactosidasenegative S.aureus mutant was doned into E.coli and sequenced.In the totally inactive mutant protein only the glycine at position332 was changed to an arginine. This amino acid is part of thesequence insertion near the antigenic domain reacting with bothantisera. These data support the assumption that the regionis of great importance for the function of the enzymes and thatit is possible it determines the specificity of the phosphorylatedform of the substrates. In addition, the 6-phospho-ß-galactosidaseof S.aureus was modified by sitedirected mutagenesis of thecorresponding lacG gene hi order to replace residues Glul60and Glu375, which were suspected of being involved hi the generalacid catalysis of substrate hydrolysis, with glutamine residues.The mutant protein 160EQ retained some catalytic activity whilethe protein 375EQ was totally inactive. Glu375 is the activesite nudeophile of the 6-phospho-ß-galactosidase ofS.aureus. It is located in the sequence motif ENG where Glu358was identified as the catalytkally active nudeophile hi theß-glucosidase of Agrobacterium.     

The crystal structure of {alpha}-bungarotoxin at 2.5 ? resolution: relation to solution structure and binding to acetylcholine receptor

We report collection of 2.5 ? resolution X-ray diffraction datafrom newly grown crystals of the rare ‘small unit cell’form of the long snake neurotoxin, -bungarotoxin. The previousmodel of the molecule has been rebuilt, and refined using least-squaremethods to a crystallographic residual of 0.24 at 2.5 ? resolution.-Bungarotoxin's crystal structure is compared with the crystalstructures of two other snake neurotoxins (cobratoxin and erabutoxin),and with its solution structure inferred from spectroscopicstudies. Significant differences include less ß-sheetin bungarotoxm's crystal structure than in solution, or in thecrystal structures of other neurotoxins, and an unusual orientationin the crystal of the invariant tryptophan. The functional,binding surface of bungarotoxin is described; it consists primarilyof hydrophobic and hydrogen-bonding groups and only a few chargedside chains. The structure is compared with experimental bindingparameters for neurotoxins.