All-Atom Protein Sequence Design Based on Geometric Deep Learning.

Designing sequences for specific protein backbones is a key step in creating new functional proteins. Here, we introduce GeoSeqBuilder, a deep learning framework that integrates protein sequence generation with side chain conformation prediction to produce the complete all-atom structures for designed sequences. GeoSeqBuilder uses spatial geometric features from protein backbones and explicitly includes three-body interactions of neighboring residues. GeoSeqBuilder achieves native residue type recovery rate of 51.6%, comparable to ProteinMPNN and  other leading methods, while accurately predicting side chain conformations. We first used GeoSeqBuilder to design sequences for thioredoxin and a hallucinated three-helical bundle protein. All the 15 tested sequences expressed as soluble monomeric proteins with high thermal stability, and the 2 high-resolution crystal structures solved closely match the designed models. The generated protein sequences exhibit low similarity (minimum 23%) to the original sequences, with significantly altered hydrophobic cores. We further redesigned the hydrophobic core of glutathione peroxidase 4, and 3 of the 5 designs showed improved enzyme activity. Although further testing is needed, the high experimental success rate in our testing demonstrates that GeoSeqBuilder is a powerful tool for designing novel sequences for predefined protein structures with atomic details. GeoSeqBuilder is available at https://github.com/PKUliujl/GeoSeqBuilder.

Influence of Side Chain Conformation on the Activity of Glycosidase Inhibitors.

Substrate side chain conformation impacts reactivity during glycosylation and glycoside hydrolysis and is restricted by many glycosidases and glycosyltransferases during catalysis. We show that the side chains of gluco and manno iminosugars can be restricted to predominant conformations by strategic installation of a methyl group. Glycosidase inhibition studies reveal that iminosugars with the gauche,gauche side chain conformations are 6- to 10-fold more potent than isosteric compounds with the gauche,trans conformation; a manno-configured iminosugar with the gauche,gauche conformation is a 27-fold better inhibitor than 1-deoxymannojirimycin. The results are discussed in terms of the energetic benefits of preorganization, particularly when in synergy with favorable hydrophobic interactions. The demonstration that inhibitor side chain preorganization can favorably impact glycosidase inhibition paves the way for improved inhibitor design through conformational preorganization.

Identifying the elusive link between amino acid sequence and charge selectivity in pentameric ligand-gated ion channels.

Among neurotransmitter-gated ion channels, the superfamily of pentameric ligand-gated ion channels (pLGICs) is unique in that its members display opposite permeant-ion charge selectivities despite sharing the same structural fold. Although much effort has been devoted to the identification of the mechanism underlying the cation-versus-anion selectivity of these channels, a careful analysis of past work reveals that discrepancies exist, that different explanations for the same phenomenon have often been put forth, and that no consensus view has yet been reached. To elucidate the molecular basis of charge selectivity for the superfamily as a whole, we performed extensive mutagenesis and electrophysiological recordings on six different cation-selective and anion-selective homologs from vertebrate, invertebrate, and bacterial origin. We present compelling evidence for the critical involvement of ionized side chains-whether pore-facing or buried-rather than backbone atoms and propose a mechanism whereby not only their charge sign but also their conformation determines charge selectivity. Insertions, deletions, and residue-to-residue mutations involving nonionizable residues in the intracellular end of the pore seem to affect charge selectivity by changing the rotamer preferences of the ionized side chains in the first turn of the M2 α-helices. We also found that, upon neutralization of the charged residues in the first turn of M2, the control of charge selectivity is handed over to the many other ionized side chains that decorate the pore. This explains the long-standing puzzle as to why the neutralization of the intracellular-mouth glutamates affects charge selectivity to markedly different extents in different cation-selective pLGICs.

Prediction of protein mutation effects based on dehydration and hydrogen bonding - A large-scale study.

Reliable computational prediction of protein side chain conformations and the energetic impact of amino acid mutations are the key aspects for the optimization of biotechnologically relevant enzymatic reactions using structure-based design. By improving the protein stability, higher yields can be achieved. In addition, tuning the substrate selectivity of an enzymatic reaction by directed mutagenesis can lead to higher turnover rates. This work presents a novel approach to predict the conformation of a side chain mutation along with the energetic effect on the protein structure. The HYDE scoring concept applied here describes the molecular interactions primarily by evaluating the effect of dehydration and hydrogen bonding on molecular structures in aqueous solution. Here, we evaluate its capability of side-chain conformation prediction in classic remutation experiments. Furthermore, we present a new data set for evaluating "cross-mutations," a new experiment that resembles real-world application scenarios more closely. This data set consists of protein pairs with up to five point mutations. Thus, structural changes are attributed to point mutations only. In the cross-mutation experiment, the original protein structure is mutated with the aim to predict the structure of the side chain as in the paired mutated structure. The comparison of side chain conformation prediction ("remutation") showed that the performance of HYDEprotein is qualitatively comparable to state-of-the art methods. The ability of HYDEprotein to predict the energetic effect of a mutation is evaluated in the third experiment. Herein, the effect on protein stability is predicted correctly in 70% of the evaluated cases. Proteins 2017; 85:1550-1566. © 2017 Wiley Periodicals, Inc.

Assessment of protein side-chain conformation prediction methods in different residue environments.

Computational prediction of side-chain conformation is an important component of protein structure prediction. Accurate side-chain prediction is crucial for practical applications of protein structure models that need atomic-detailed resolution such as protein and ligand design. We evaluated the accuracy of eight side-chain prediction methods in reproducing the side-chain conformations of experimentally solved structures deposited to the Protein Data Bank. Prediction accuracy was evaluated for a total of four different structural environments (buried, surface, interface, and membrane-spanning) in three different protein types (monomeric, multimeric, and membrane). Overall, the highest accuracy was observed for buried residues in monomeric and multimeric proteins. Notably, side-chains at protein interfaces and membrane-spanning regions were better predicted than surface residues even though the methods did not all use multimeric and membrane proteins for training. Thus, we conclude that the current methods are as practically useful for modeling protein docking interfaces and membrane-spanning regions as for modeling monomers.

Preferred side-chain conformation of arginine residues in a triple-helical structure.

The crystal structure of the triple-helical peptide (Pro-Hyp-Gly)3 -Pro-Arg-Gly-(Pro-Hyp-Gly)4 (POG3-PRG-POG4) was determined at 1.45 Å resolution. POG3-PRG-POG4 was designed to permit investigation of the side-chain conformation of the Arg residues in a triple-helical structure. Because of the alternative structure of one of three Arg residues, four side-chain conformations were observed in an asymmetric unit. Among them, three adopt a ttg(-) t conformation and the other adopts a tg(-) g(-) t conformation. A statistical analysis of 80 Arg residues in various triple-helical peptides showed that, unlike those in globular proteins, they preferentially adopt a tt conformation for χ1 and χ2 , as observed in POG3-PRG-POG4. This conformation permits van der Waals contacts between the side-chain atoms of Arg and the main-chain atoms of the adjacent strand in the same molecule. Unlike many other host-guest peptides, in which there is a significant difference between the helical twists in the guest and the host peptides, POG3-PRG-POG4 shows a marked difference between the helical twists in the N-terminal peptide and those in the C-terminal peptide, separated near the Arg residue. This suggested that the unique side-chain conformation of the Arg residue affects not only the conformation of the guest peptide, but also the conformation of the peptide away from the Arg residue.

Structure of racemic duloxetine hydro-chloride.

Duloxetine hydro-chloride (trade name Cymbalta) is marketed as a single enanti-omer (S)-N-methyl-3-(naphthalen-1-yl-oxy)-3-(thio-phen-2-yl)propyl-am-in-ium chloride, C18H20NOS+·Cl-, which is twice as effective as the (R)-enanti-omer in serotonin uptake. Here, we report the crystal structure of duloxetine hydro-chloride in its racemic form (space group Pna21), where it shows significant differences in the mol-ecular conformation and packing in its extended structure compared to the previously reported (S)-enanti-omer crystal structure. Mol-ecules of this type, comprising aromatic groups with a single side chain terminated in a protonated secondary amine, are commonly found in active anti-depressants. A Cambridge Structural Database survey of mol-ecules with these features reveals a strong correlation between side-chain conformation and the crystal packing: an extended side chain leads to mol-ecules packed into separated layers of hydro-phobic and ionic hydro-philic phases. By comparison, mol-ecules with bent side chains, such as racemic duloxetine hydro-chloride, lead to crystal-packing motifs where an ionic hydro-philic phase is encapsulated within a hydro-phobic shell.

Side-chain conformers to allow conversion from normal to isoaspartate in age-related proteins and peptides.

Dß (or D-iso)- and Lß- (or iso)- aspartyl (Asp) residues are accumulated in aged lens crystallins and amyloid beta (Aß) proteins, respectively, as a result of spontaneous, nonenzymatic isomerization of normal Lα-Asp. To explore why such uncommon Asp isomers are accumulated, the stability of Lα-, Lß-, and Dß-Asp was compared in view of the staggered side-chain conformers. By using cylindrin (KVKVLGD7VIEV) from αB-crystallin and Aß17-25 (L17VFF20AED23)VG25) containing Asp isomers, the vicinal spin-spin coupling constants of Asp Hα-Hß1 and Hα-Hß2 were quantified by high-resolution solution 1H NMR. It was found that the trans conformer was extremely preferred in Dß-Asp7 side-chain of cylindrin. In Aß17-25, the side chain of Lß-Asp23 was likely to adopt trans conformer, while gauche conformers were rather rich in Lα-Asp23. In gauche conformers, the close distance between Asp carboxylate carbon (CCOO-) and backbone nitrogen (N) next to Asp is advantageous to the intramolecular cyclization to form succinimide intermediate, followed by the conversion from α- to ß-Asp. The cyclization is limited in the trans conformer because of the long distance between CCOO- and N, to keep Dß- or Lß-Asp stable. This would be the reason for the site specificity of Asp isomerization in proteins. The higher population of trans conformer in Asp side chain, the less isomerization of Asp as shown as Asp76 in αA-crystallin. The stability and less reactivity of normal Asp and its isomers are the potential factors to determine whether or not the abnormal accumulation is permitted in aged crystallins and Aß.

Stereospecific synthesis of methyl 2-amino-2-deoxy-(6S)-deuterio-α,ß-d-glucopyranoside and methyl 2,6-diamino-2,6-dideoxy-(6R)-deuterio-α,ß-d-glucopyranoside: Side chain conformations of the 2-amino-2-deoxy and 2,6-diamino-2,6-dideoxyglucopyranosides.

The stereospecifically labeled 6-monodeuterio methyl 2,6-diamino-2,6-dideoxy-α- and ß- d-glucopyranosides were synthesized with a view to determining their side chain conformations. NMR studies in D2O at pH 5 and pH 11 reveal both anomers to adopt very predominantly the gt conformation consistent with the gauche conformation of 2-aminoethanol and its acetate salt. In contrast, as also revealed with the help of stereospecifically-labelled monodeuterio isotopomers, the methyl 2-amino-2-deoxy-α- and ß- d-glucopyranosides are an approximately 1:1 mixture of gg and gt conformers as is found in glucopyranose itself.

Staggered side-chain conformers of aspartyl residues prerequisite to transformation from L-α- to D-ß-aspartate 58 in human-lens αA-crystallin fragment.

D-ß-aspartyl (Asp) residues are found in aged human-lens αA-crystallin. To explore why the uncommon D-ß-Asp is accumulated, the stability of L-α-, D-α-, and D-ß-Asp residues is compared in view of the staggered side-chain conformers. By using αA-crystallin fragment, T(55)VLD(58)SGISEVR(65), composed of Asp58 isomers, the vicinal spin-spin coupling constants of Asp58 Hα-Hß1 and Hα-Hß2 are quantified by high-resolution solution NMR. The trans conformer is most preferred in the D-ß-Asp side-chain, whereas gauche+ and gauche- are abundant in L-α- and D-α-Asp. The close distance between Asp58 carboxylate carbon (CCOO-) and Ser59 nitrogen (N) in gauche+ and gauche- is advantageous to the intramolecular cyclization to form succinimide intermediate, followed by the transformation from α- to ß-Asp. The cyclization is not allowed in the trans conformer because of the long distance between CCOO- and N, to keep D-ß-Asp stable. The change from gauche to trans conformer in D-ß-Asp is exothermic and enthalpy-driven.