David A. Langs

Molecular structure and mechanisms of action of cyclic and linear ion transport antibiotics

Ionophores are antibiotics that induce ion transport across natural and artificial membranes. The specific function of a given ionophore depends upon its selectivity and the kinetics of ion capture, transport, and release. Systematic studies of complexed and uncomplexed forms of linear and cyclic ionophores provide insight into molecular mechanisms of ion capture and release and the basis for ion selectivity. The cyclic dodecadepsipeptide valinomycin, cyclo[(-L-Val-D-Hyi-D-Val-L-Lac)3-], transports potassium ions across cellular membrane bilayers selectively. The x-ray crystallographic and nmr spectroscopic data concerning the structures of Na+, K+, and Ba+2 complexes are consistent and provide a rationale for the K+ selectivity of valinomycin. Three significantly different conformations of valinomycin are observed in anhydrous crystals, in hydrated crystals grown from dimethylsulfoxide, and in crystals grown from dioxane. Each of these conformations suggests a different mechanism of ion capture. One of the observed conformations has an elliptical structure stabilized by four 4<--1 intramolecular hydrogen bonds and two 5<--1 hydrogen bonds. Ion capture could be readily achieved by disruption of the 5<--1 hydrogen bonds to permit coordination to a potassium ion entering the cavity. The conformation found in crystals obtained from dimethyl sulfoxide is an open flower shape having three petals and three 4<--1 hydrogen bonds. Complexation could proceed by a closing up of the three petals of the flower around the desolvating ion. In the third form, water molecules reside in the central cavity of a bracelet structure having six 4<--1 hydrogen bonds. Two of these bracelets stack over one another with their valine-rich faces surrounding a dioxane molecule. The stacked molecules form a channel approximately 20 A in length, suggesting that under certain circumstances valinomycin might function as a channel. A series of analogues of valinomycin differing in ring composition and size have been synthesized and their transport properties tested. Peptide substitution and chiral variation in the dodecadepsipeptide can result in stabilization or modification of the different conformers. While contraction of the ring size results in loss of ion transport properties, expansion of the ring size permits complexation of larger ions and small positively charged molecules. Gramicidin A is a pentadecapeptide that functions as a transmembrane channel for transporting monovalent cations. Crystal structures of the cesium chloride complex and two uncomplexed forms of gramicidin A have been reported. In all three structures the gramicidin A molecule is a left-handed, antiparallel, double-stranded helical dimer. In the cesium complex the beta 7.2-helix has 6.4 residues per turn with an internal cavity large enough to accommodate cesium ions. In the uncomplexed structures the channel is 31 A long and has 5.6 amino acids per turn. Because the helix is too tightly wound to permit ion transport, ion transport would require breaking and reforming of hydrogen bonds.

Heterodimer Formation and Crystal Nucleation of Gramicidin D

The linear pentadecapeptide antibiotic gramicidin D is a heterogeneous mixture of six components. Precise refinements of three-dimensional structures of naturally occurring gramicidin D in crystals obtained from methanol, ethanol, and n-propanol demonstrate the unexpected presence of stable left-handed antiparallel double-helical heterodimers that vary with the crystallization solvent. The side chains of Trp residues in the three structures exhibit sequence-specific patterns of conformational preference. Tyr substitution for Trp at position 11 appears to favor beta ribbon formation and stabilization of the antiparallel double helix that acts as a template for gramicidin folding and nucleation of different crystal forms. The fact that a minor component in a heterogeneous mixture influences aggregation and crystal nucleation has potential applications to other systems in which anomalous behavior is exhibited by aggregation of apparently homogeneous materials, such as the enigmatic behavior of prion proteins.

Ab initio structure determination and refinement of a scorpion protein toxin.

The structure of toxin II from the scorpion Androctonus australis Hector has been determined ab initio by direct methods using SnB at 0.96 A resolution. For the purpose of this structure redetermination, undertaken as a test of the minimal function and the SnB program, the identity and sequence of the protein was withheld from part of the research team. A single solution obtained from 1 619 random atom trials was clearly revealed by the bimodal distribution of the final value of the minimal function associated with each individual trial. Five peptide fragments were identified from a conservative analysis of the initial E-map, and following several refinement cycles with X-PLOR, a model was built of the complete structure. At the end of the X-PLOR refinement, the sequence was compared with the published sequence and 57 of the 64 residues had been correctly identified. Two errors in sequence resulted from side chains with similar size while the rest of the errors were a result of severe disorder or high thermal motion in the side chains. Given the amino-acid sequence, it is estimated that the initial E-map could have produced a model containing 99% of all main-chain and 81% of side-chain atoms. The structure refinement was completed with PROFFT, including the contributions of protein H atoms, and converged at a residual of 0.158 for 30 609 data with F >or= 2sigma(F) in the resolution range 8.0-0.964 A. The final model consisted of 518 non-H protein atoms (36 disordered), 407 H atoms, and 129 water molecules (43 with occupancies less than unity). This total of 647 non-H atoms represents the largest light-atom structure solved to date.

Statistical expectation value of the Debye-Waller factor and E(hkl) values for macromolecular crystals.

If the unit-cell distribution of atomic mean-square displacement parameters B = 8pi(2) is assumed to be normal, with mean micro = and variance sigma(2) = <(B-)(2)>, the statistical expectation value of the Debye-Waller factor W(2) = exp(-2Bs(2)), where s = (sin theta)/lambda, is = exp[-2( micro - sigma(2)s(2))s(2)]. This result has been incorporated into procedures for scaling and normalizing measured Bragg intensities to their Wilson expectation values. The procedures can determine both isotropic micro (B) and sigma(B) and anisotropic micro (U(ij)) and sigma(U(ij) distribution parameters. Tests with experimental data and refined structural models for several protein crystals show that the procedures yield reliable normalized structure-factor amplitudes for direct-methods applications, with values of R = summation operator (h)||E(o)| - |E(c)||/ summation operator (h)|E(o)| averaging approximately 5%.

The phase problem in neutron crystallography.

The straightforward solution of the crystal structure of cyclosporin (C(62)H(111)N(11)O(12).H(2)O) by a modified Shake-and-Bake procedure, using experimental neutron diffraction data alone, shows that the positivity of the density function is not a necessary prerequisite for solving the phase problem. The initial applications suggest the intriguing possibility that positivity may actually be a hindrance.

Progress on the direct-methods solution of macromolecular structures using single-wavelength anomalous-dispersion (SAS) data

In the past few years, a number of strategies have been outlined to resolve the SAS phase ambiguity given that unique estimates omega (h, k) of the triple invariants are available. A new least-squares method is described that can in principle resolve the phase ambiguity to determine macromolecular phases provided that omega (h, k) estimates are unbiased. Limitations of the method in practical applications are discussed. An example is given where the correct solution can be identified by use of the SAS tangent formula in the instance that traditional SAS phasing methods have lead to an incorrect heavy-atom substructure.

Rational genomics I: antisense open reading frames and codon bias in short-chain oxido reductase enzymes and the evolution of the genetic code.

The short‐chain oxidoreductase (SCOR) family of enzymes includes over 6000 members, extending from bacteria and archaea to humans. Nucleic acid sequence analysis reveals that significant numbers of these genes are remarkably free of stopcodons in reading frames other than the coding frame, including those on the antisense strand. The genes from this subset also use almost entirely the GC‐rich half of the 64 codons. Analysis of a million hypothetical genes having random nucleotide composition shows that the percentage of SCOR genes having multiple open reading frames exceeds random by a factor of as much as 1 × 106. Nevertheless, screening the content of the SWISS‐PROT TrEMBL database reveals that 15% of all genes contain multiple open reading frames. The SCOR genes having multiple open reading frames and a GC‐rich coding bias exhibit a similar GC bias in the nucleotide triple composition of their DNA. This bias is not correlated with the GC content of the species in which the SCOR genes are found. One possible explanation for the conservation of multiple open reading frames and extreme bias in nucleic acid composition in the family of Rossman folds is that the primordial member of this family was encoded early using only very stable GC‐rich DNA and that evolution proceeded with extremely limited introduction of any codons having two or more adenine or thymine nucleotides. These and other data suggest that the SCOR family of enzymes may even have diverged from a common ancestor before most of the AT‐rich half of the genetic code was fully defined. Proteins 2005.

Reinvestigation of the Use of Patterson Maps to Extrapolate Data to Higher Resolution

Many years ago, Karle & Hauptman proposed that the Patterson function could be used for data extrapolation beyond the observed range of the actual measured data. Few people have subsequently attempted to exploit this interesting idea, which might suggest possible limitations of this method, even in structural applications of modest complexity. This appears not to be the case, however, but the original ideas for implementing the extrapolation can be significantly improved. New calculation protocols indicate that Patterson maps may be used to extend observed data sets from 1.0 to approximately 0.5 A resolution with reasonably good precision. Correlation coefficients between the extrapolated F(hkl)s and their structure-computed expected values typically range between 0.40 and 0.70 across the unobserved range, even for structures containing as many as 600 non-H light atoms in the asymmetric unit. The method is equally good at extrapolating F values for small zones of data that may not have been recorded within the observed resolution range of the diffraction experiment. Furthermore, triplet phase invariants that incorporate one or two extrapolated terms are nearly as reliable as those formed using only the observed data.

TDSIR phasing: direct use of phase-invariant distributions in macromolecular crystallography.

A new strategy for employing three phase triples invariant estimates from Hauptmans single isomorphous replacement (SIR) and anomalous dispersion (SAS) joint probability distribution formulae is outlined which produces a single unique phase-invariant solution in the case where the positions of the heavy-atom scatterers is known. A similar but non-identical result is obtained for the phase invariants of a structure for which a molecular-replacement solution has been obtained. It is important to note that the values of the individual native/derivative phases can be determined directly from the probability distribution formulae without having to utilize the phase-invariant estimates in an active way. Elimination of the multisolution aspect of utilizing phase-invariant estimates should have important repercussions with regard to phasing macromolecular sets of derivatized data. Trial calculations based on experimentally measured 2.5 A data for three derivatives of cytochrome c550 are encouraging. The average of the three SIR maps resolves a number of structural ambiguities seen in the published multiple isomorphous replacement (MIR) map obtained from eight derivatives.

Globbic approximation in low-resolution direct-methods phasing

Probabilistic direct-methods phasing theory, originally based on a uniform atomic distribution hypothesis, is shown to be adaptable to a non-uniform bulk-solvent-compensated globbic approximation for protein crystals at low resolution. The effective number n(g) of non-H protein atoms per polyatomic glob increases with decreasing resolution; low-resolution phases depend on the positions of only N(g) = N(a)/n(g) globs rather than N(a) atoms. Test calculations were performed with measured structure-factor data and the refined structural parameters from a protein crystal with approximately 10 000 non-H protein atoms per molecule and approximately 60% solvent volume. Low-resolution data sets with d(min) ranging from 15 to 5 A gave n(g) = ad(min) + b, with a = 1.0 A(-1) and b = -1.9 for the test case. Results of tangent-formula phase-estimation trials emphasize that completeness of the low-resolution data is critically important for probabilistic phasing.