Search

For several years, Science policy has required that published X-ray crystallographic coordinates be deposited with a publically accessible database (e.g., Brookhaven Protein Data Bank). That policy recognized the high costs and risks of solving important structures to accord those who have made such investments successfully the time to reap their intellectual benefits. Wlodawer et al. (below) assert that recent technical enhancements have reduced “the time needed to solve a structure (to less than) than the allowed hold period” and have called for immediate release of the data. In cooperative discussions with our colleagues at Nature and other journals, Science wishes to evaluate how best to fulfill our responsibilities to the community. We urge readers to respond to a survey now underway (see ) and to us, so that future policy may be modified on the basis of scientific judgement. # Immediate Release of Crystallographic Data: A Proposal {#article-title-2} There has been an incredibly rapid increase in the rate of determination of three-dimensional (3D) structures of biomacromolecules, as reflected by the deposition of a new structure in the Protein Data Bank (PDB) at the Brookhaven National Laboratory ([1][1]), on average, every 5 hours. Unfortunately, in parallel, an increasing proportion of depositors take advantage of the PDB's policy of allowing structures to be kept “on hold” for up to a year after coordinate deposition. Despite a recent drop in the number of structures put “on hold,” nearly half the entries deposited are not released immediately. The policy of the PDB is based on rules drawn up by the International Union of Crystallography (IUCr) in the late 1980s for papers published in IUCr journals. These rules ([2][2]) also provide the basis for the policies of most other scientific journals and of a number of government funding agencies, such as the National Institutes of Health, for work undertaken with grant support. It is time to consider whether this policy is still appropriate. When it was debated and accepted by the community 10 years ago, the time needed to solve a macromolecular structure was often measured in years and was rarely less than 1 year. The time needed for detailed analysis of such structures was also fairly long. The 1-year hold on coordinates was therefore instituted to allow the authors to reap the fruit of their tremendous investment of time and effort. Because of recent advances in protein expression and purification, crystallization procedures, x-ray instrumentation, and computer software, the time needed to solve a structure is often shorter than the allowed hold period. In light of such developments, it is difficult to justify withholding coordinates for any period once the paper has been published. Biomolecular structure analysis has succeeded in bringing 3D structures to the forefront of molecular biological research. This success has expanded both the interest in and utility of the information being deposited in the PDB. The molecular modeling community has grown and evolved considerably, due to the expansion of this source of experimental data. The value of the data rests in their availability to the broader community. Methods are continuously being developed to analyze new structures and their relationships to the collection of existing structures. New uses for these data, such as statistical potentials for folding and threading calculations, and interface recognition tools, are evolving rapidly. No single research group can fully exhaust this wealth of information. The value of the resource grows proportionally to the timeliness of the data and to the number of scientists who have access to them. Three-dimensional structural information is also a crucial link elucidating the role of a translated region of a DNA sequence of unknown function. The time has come to change the rules of deposition so as to ensure that the coordinates are released concomitantly with publication of the paper (or papers) describing the structure. We are convinced that without access to the coordinates, the structures cannot be used for comparison with other proteins, for theoretical analysis or, more and more important, for drug design. We propose that coordinates deposited at the PDB should be marked as either “for immediate release” or “to be released upon publication.” We also recommend that the maximum hold for primary data (that is, x-ray structure factors, and nuclear magnetic resonance proton-proton distance and dihedral angle restraints) be reduced from 4 years to 1 year. The PDB is already working on a “layered approach” to deposition so that it will be possible to release entries as submitted, after the authors have checked for outliers and errors through the PDB's WEB-based AutoDep procedure, on the same day that they are deposited (with the permission of authors), or when the article related to the structure is published, if the authors request “release on publication date” rather than “hold for a year.” It is clear that such a change in policy will require cooperation of both the granting agencies and the scientific journals, as well as the overwhelming support of the scientists doing the research. It should be stressed that even the current policy is not uniformly enforced. These changes would bring macromolecular crystallography into line with the requirements of other fields, such as gene sequencing, which have never allowed extended hold periods. We hope that this proposed change in the deposition policy will be publicly debated and ultimately accepted. 1. [↵][3] . 2. [↵][4]International Union of Crystallography, Acta Crystallogr. A45, 658 (1989). [OpenUrl][5] [1]: #ref-1 [2]: #ref-2 [3]: #xref-ref-1-1 "View reference 1 in text" [4]: #xref-ref-2-1 "View reference 2 in text" [5]: {openurl}?query=rft.jtitle%253DActa%2BCrystallogr.%26rft.volume%253DA45%26rft.spage%253D658%26rft.atitle%253DACTA%2BCRYSTALLOGR%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx

The present study attempts to quantitate the contribution of electrostatic interactions to stabilization of the catalytic transition states of cholinesterases (ChEs) and to the traffic of reversible ligands and reactive species within the active site gorge. For this purpose, we calculate the dependence on ionic strength of the electric potential in the region of the active site gorge of several ChEs, and assess the contribution of negatively charged residues. We also calculate the difference in free energy of binding to AChE of the charged transition state analog m-(N, N, N-trimethylammonio)-2,2,2-trifluoroacetophenone(TFK+) and of its neutral isoster m-(2, 2, 2-trimethyl)trifluoroacetophenone(TFK°) both with WT AChE and with the W84A mutant at various ionic strengths. Based on these results we employ absolute rate reaction theory to calculate the electrostatic potential at the site of binding.

The concept of an electrostatic motif on the surface of biological macromolecules as a definite topographical pattern of electrostatic potentials in 3D space, provides a powerful tool for identification of functionally important regions on the surface of structurally related macromolecules. We employ this tool to analyze the electrostatic properties of cholinesterases (ChEs) from various species and show that all these structures reveal a negative external surface potential in an “annular” area around the entrance to the active-site gorge that becomes more negative as one approaches the rim of the gorge. These potentials are highly correlated among the structures examined, down to a sequence identity as low as 35%, indicating that they are a conserved property of the cholinesterase family. We further suggest that the coincidence between a shared topological arrangement of the macromolecular chain and a conserved electrostatic motif can be used to identify proteins sharing a common recognition mechanism. Using this approach, we identify a functional region common to ChEs and to gliotactin (GLI), neurotactin (NRT), and neuroligin-1 (NL-1), a set of neural cell-adhesion proteins which have been suggested to be structurally related to ChEs due to their high sequence similarity, but lacking the key catalytically active serine.

The Protein Data Bank (PDB) is an archive of experimentally determined three-dimensional structures of proteins, nucleic acids, and other biological macromolecules with a 25 year history of service to a global community. PDB is being replaced by 3DB, the Three-Dimensional Database of Biomolecular Structures that will continue to operate from Brookhaven National Laboratory. 3DB will be a highly sophisticated knowledge-based system for archiving and accessing structural information that combines the advantages of object oriented and relational database systems. 3DB will operate as a direct-deposition archive that will also accept third-party supplied annotations. Conversion of PDB to 3DB will be evolutionary, providing a high degree of compatibility with existing software.

A formal discussion of the archival journal requirements for data deposition was held at the International Seminar-cum-School on Macromolecular Crystallographic Data at Calcutta, India in November 1995.The current policy of the International Union of Crystallography (IUCr) is that on publication of a crystal structure determination of a macromolecule the atomic parameters used or represented in the publication must be deposited in the Brookhaven Protein Data Bank. The deposition of structure amplitudes is recommended but not obligatory. The policy provides crystallographers with the option to delay the release of atomic parameters for one year and of structure amplitudes for up to four years from the date of publication. Participants strongly supported this policy and felt it should be strictly applied by the journals.Recent developments in X-ray crystallographic experimental and refinement techniques and the huge expansion in computing power and networking, however, necessitate the review of deposition arrangements.It was noted that the new validation procedures are much more effective but require the experimental structure amplitudes as well as the atomic parameters. In addition, the technical arrangements for deposition, analysis and validation of macromolecular crystal structures are now much easier.We consider it vital for macromolecular crystallographers to respond to these developments in their deposition practices. We recommend, therefore, that publication of macromolecular crystal structures should be accompanied by deposition of atomic parameters and also structure amplitudes. Amongst the many reasons identified for this practice, the following two are critical.(1) Rigorous validation of the structure determination results can only be carried out using both atomic parameters and experimental structure amplitudes. It is important that journals ensure referees have sufficient information to prevent incorrect structures being published.(2) Archiving of these data will ensure that they are not lost. There were numerous reports at this meeting of data being lost. This probably reflects a widespread problem in the crystallographic community.

Three proteins from Halobacterium marismortui, malate dehydrogenase (hMDH), glutamate dehydrogenase (hGDH) and ferredoxin (hFD) were purified and characterized with respect to their molecular masses, amino acid composition and, for hFD only, primary structure. Striking features of halophilic proteins are: the high excess of acidic over basic residues; acidic clusters in the sequence. Low-salt concentration causes inactivation and changes in structural parameters of hMDH and hGDH. Reactivation of hMDH involves long-lived stable intermediates. The salt concentration optimum of enzymic activity is independent of salt nature. The high capacity of halophilic proteins to retain water and salt is due to unique molecular properties, studied by physico-chemical techniques.

The atomic coordinates are given for yeast phenylalanine transfer RNA in the orthorhombic crystal form. The structure has been refined by fitting to successively improved electron density maps at 2.7 A resolution. The model fitting has been accomplished by using an interactive computer graphics system to minimize the errors inherent in manual model building and coordinate measurements, using an optical comparator. The atomic coordinates have then been “idealized” to make bond distances, bond angles, steric conformation and non-bonded contacts close to standard values, while constraining the model to fit the electron density maps.

We present the results of the Jimal stage of the X-ray crystallographic studies of yeast phenylalanine transfer RNA in an orthorhombic crystal form. The crystal structure of the transfer KNA has been refined by a least-squares procedure to minimize the difference between the observed (F,) and calculated (F’,) structure factors from X-ray diffraction patterns. The final crystallographic discrepancy index, R = IIF,, - Fcl/IF,, is 0.198, based upori 8426 structure factors with magnitudes over twice the estimated standard deviation, corresponding to 96.4% of the complete set of data with resolutions up t,o 2.7 A. During the refinement, bond lengths and angles within each phosphate group and each rmcleoside (base plus sugar) were constrained exactly to their appropriate standard values, while those for the linkages between the nucleosides and phosphates were elastically restrained close to their standard values. The details of the application of the constraint-restraint least-squares (CORELS) refinement method to the crystal structure of yeast phenylalanine tRNA are described in this paper. A complete list of atomic co-ordinates and the rigid group thermal factors are presented. The st,ereochemical details of this structrire and their frictional implications are described in the following paper.

X-ray crystallagraphic studies studies indicate that there are at least four site-specifically bound hydrated Mg2+ ions, [Mg(H2O)n]2+, in yeast tRNAPhe. The size and the octahedral coordination geometry, rather than the charge, of [Mg(H2O)N]2+ appear to be the primary reasons for the specificity of magnesium ions in site-binding and in the stabilization of the tertiary structure of tRNA.

The use of a macromolecular model building program, CORELS, which fits a series of constrained groups to a set of approximate guide coordinates, while simultaneously restraining the stereochemistry between groups, is described for protein molecules. The results of an application of this program to hen egg white lysozyme are discussed.

A method of structure-factor least-squares refinement of constrained groups linked by distance restraints has been developed for the refinement of macromolecular structures. Each constrained group can have any number of variable dihedral rotation parameters within the group in addition to the rigid-body translational, rotational and thermal parameters. The matrix of normal equations may be either full or sparse and provision is made for solution by matrix inversion or the conjugate-gradient iterative method. This procedure has been successfully used for 3 A data and should be applicable even for lower-resolution data and especially for cases with a poorer data-per-atom ratio. The structure of yeast phenylalanine tRNA has been refined with this procedure from a starting crystallographic R value of 42% to a final R value of 25% with isotropic 'group" thermal parameters and 22% with isotropic atomic thermal parameters for 8207 independent reflections at 2.7 A resolution. The proper stereochemistry of bond distances, angles and van der Waals contacts for the restrained atoms was maintained within reasonable limits throughout the refinement. Although originally developed for nucleic acids, this procedure is directly applicable to the refinement of protein structures. In addition, a combination of applying distance restraints between groups and least-squares fitting of these groups to target coordinates has been used purely as an idealization process for imposing proper stereochemistry on an approximate model.

Further analysis of the three-dimensional electron density map of yeast phenylalanine tRNA is presented. Attention is focused on the several types of unique hydrogen bonding that are found in the molecule and a number of sections of the electron density map are presented. These sections are compared with an electron density map of a dinucleoside phosphate. The bases in the helical stem regions are all involved in Watson-Crick hydrogen bonding interactions with the exception of the guanine-uracil base pair. Several additional tertiary hydrogen bonding interactions are described.

We present a method which permits comparison of the preferential use of degenerate codons within any gene. The method makes use of the triplet frequencies in the noncoding frames to assess whether a preference is specific to the reading frame. Preference is given a statistical meaning by use of the analysis of variance coupled to Duncan's multiple range test. Preferential use of degenerate codons is gene-specific and independent of gene size. The data suggest that any correlation between codon frequency distribution and tRNA levels is unreliable. In those animal genes examined, codons ending in C or G are preferred; in animal viruses tested, codons ending in U or A are preferred. Similarly, the bacterial genes and the genes of single-stranded DNA phages that we analyzed differed from each other as well as from eukaryotic genes in the third base of the codon.

Several long nucleotide sequences were analysed to find out if any of them are unusual in terms of the possible formation of “hairpins” (pairs of complementary runs of nucleotides forming double-stranded structures) as compared to random sequences. The RNA of MS2 bacteriophage has more potential hairpins with short loops (up to 10 bases in a loop) than found in random sequences with the same length and base composition. Other analyzed nucleotide sequences (SV40, φX174, Fd and 16S rRNA) behaved very much as their corresponding random ones. Most of the extra hairpins of the MS2 RNA are estimated to be thermodynamically stable. These potential hairpins might play some role in the function of the MS2 RNA.

A model for the complex between double-stranded DNA and a beta-ribbon (a two-stranded antiparallel beta-sheet) of proteins is proposed as one of the possible modes of structural recognition between DNA and proteins. In this model, the contact occurs on the narrow groove of DNA, and the symmetry elements as well as the repeat distances of DNA and the beta-ribbon coincide, thus providing favorable complementary contacts.

Structural modelling techniques using energy minimization and molecular dynamics have been employed to generate kinked models for the solution structure of two DNA tridecamer sequences containing inserted adenosines: d(CGCAGAATTCGCG)2 and d(CGCAGAGCTCGCG)2. These models are consistent with NMR studies of these sequences in solution. The overall shapes of the two models are similar, consisting of three B-DNA sections: two outer segments on the same side of the central portion, with the additional adenosines acting as wedges to kink the structure. An alternative scheme for the hydrogen bond pairing at the kink site is suggested as a way for the additional adenosines to be stabilized in the duplex.

Background: Fasciculin (FAS), a 61-residue polypeptide purified from mamba venom, is a three-fingered toxin which is a powerful reversible inhibitor of acetylcholinesterase (AChE). Solution of the three-dimensional structure of the AChE/FAS complex would provide the first structure of a three-fingered toxin complexed with its target.Results The structure of a complex between Torpedo californica AChE and fasciculin-II (FAS-II), from the venom of the green mamba (Dendroaspis angusticeps) was solved by molecular replacement techniques, and refined at 3.0 å resolution to an R-factor of 0.231. The structure reveals a stoichiometric complex with one FAS molecule bound to each AChE subunit. The AChE and FAS conformations in the complex are very similar to those in their isolated structures. FAS is bound at the ‘peripheral’ anionic site of AChE, sealing the narrow gorge leading to the active site, with the dipole moments of the two molecules roughly aligned. The high affinity of FAS for AChE is due to a remarkable surface complementarity, involving a large contact area (∼2000 å2) and many residues either unique to FAS or rare in other three-fingered toxins. The first loop, or finger, of FAS reaches down the outer surface of the thin aspect of the gorge. The second loop inserts into the gorge, with an unusual stacking interaction between Met33 in FAS and Trp279 in AChE. The third loop points away from the gorge, but the C-terminal residue makes contact with the enzyme.Conclusion Two conserved aromatic residues in the AChE peripheral anionic site make important contacts with FAS. The absence of these residues from chicken and insect AChEs and from butyrylcholinesterase explains the very large reduction in the affinity of these enzymes for FAS. Several basic residues in FAS make important contacts with AChE. The complementarity between FAS and AChE is unusual, inasmuch as it involves a number of charged residues, but lacks any intermolecular salt linkages.

The primary sequence of esterases from Acinetobacter lwoffii RAG-1 and A. calcoaceticus BD413 were compared with linearized structural sequences of two hundred proteins selected from Brookhaven Protein DataBank using a modified version of the Bowie et al. algorithm [3]. Significant structural homology was found to α/β proteins and specifically to those with the α/β-hydrolase fold for which the crystal structure was reported. No such homology was detected using common primary sequence alignment programs such as FASTA or BLAST.

Summary A preferred strategy for preventing nerve agents intoxication is catalytic scavenging by enzymes that hydrolyze them before they reach their targets. Using directed evolution, we simultaneously enhanced the activity of a previously described serum paraoxonase 1 (PON1) variant for hydrolysis of the toxic S P isomers of the most threatening G-type nerve agents. The evolved variants show ≤340-fold increased rates and catalytic efficiencies of 0.2-5 × 10 7 M −1 min −1 . Our selection for prevention of acetylcholinesterase inhibition also resulted in the complete reversion of PON1's stereospecificity, from an enantiomeric ratio (E) −4 in favor of the R P isomer of a cyclosarin analog in wild-type PON1, to E > 2,500 for the S P isomer in an evolved variant. Given their ability to hydrolyze G-agents, these evolved variants may serve as broad-range G-agent prophylactics.

Zalpha (Zα) domains bind to left-handed Z-DNA and Z-RNA. The Zα domain protein family includes cellular (ADAR1, ZBP1 and PKZ) and viral (vaccinia virus E3 and cyprinid herpesvirus 3 (CyHV-3) ORF112) proteins. We studied CyHV-3 ORF112, which contains an intrinsically disordered region and a Zα domain. Genome editing of CyHV-3 indicated that the expression of only the Zα domain of ORF112 was sufficient for normal viral replication in cell culture and virulence in carp. In contrast, its deletion was lethal for the virus. These observations revealed the potential of the CyHV-3 model as a unique platform to compare the exchangeability of Zα domains expressed alone in living cells. Attempts to rescue the ORF112 deletion by a broad spectrum of cellular, viral, and artificial Zα domains showed that only those expressing Z-binding activity, the capacity to induce liquid-liquid phase separation (LLPS), and A-to-Z conversion, could rescue viral replication. For the first time, this study reports the ability of some Zα domains to induce LLPS and supports the biological relevance of dsRNA A-to-Z conversion mediated by Zα domains. This study expands the functional diversity of Zα domains and stimulates new hypotheses concerning the mechanisms of action of proteins containing Zα domains.

A survey of all available double-stranded RNA crystal structures shows that there is a considerable range of variation in local conformation of a given base-pair doublet, but that there is no significant correlation between base-pair sequence and RNA local conformation.

The 2 Fe-ferredoxin from the Halobacterium of the Dead Sea has been crystallized. The space group is P6322 with one protein molecule per asymmetric unit. The cell parameters are a = b = 60.6 A , c = 127.8 A . The crystals are stable under radiation and diffract to high resolution.

THEORIES of the molecular structure of nucleic acids have so far been based on evidence from the crystal structures of monomeric units such as nucleosides and mononucleotides, the interpretation of diffraction patterns of oriented nucleic acid fibres and molecular model building1–6. Such approaches can help to suggest structures of periodic molecules such as helices, but they are insufficient for predicting and understanding nonrepetitive structures such as the loops in transfer RNA (tRNA), presumably associated with many of the functions of tRNA. To understand the geometry of nucleic acids and possible constraints on their conformation, it is therefore essential to know the detailed conformation of the sugar residues and the conformational relationship between the sugar residue, the base and the phosphate group7–9. The simplest molecule which contains this information is a 3´5´-dinucleoside phosphate. We now report the structure of uridine-3´,5´-adenosine phosphate (UpA). This is the first naturally occurring dinucleoside phosphate whose crystal structure has been determined by X-ray diffraction. The only other dinucleoside phosphate with known crystal structure is adenosine-2´,5´-uridine phosphate10, but it does not have the naturally occurring 3´5´ sugar phosphate linkage.

The crystal structure of yeast tRNAPhe (in the orthorhombic form) is being refined at 3 A resolution in order to obtain a set of atomic coordinates that: (1) fit the X-ray diffraction data and (2) are consistent with the geometry of standard nucleotides and do not violate non-bonded interactions.

A correlation analysis of chromatin DNA nucleotide sequences reveals the clear tendency of some of the dinucleotides to be repeated along the sequences with periods of 3 and about 10.5 bases. This latter period, which is equal within experimental error to recent estimates of the pitch of the DNA double helix [Wang, J. (1979) Proc. Natl. Acad. Sci. USA 76, 200-203; Trifonov, E. & Bettecken, T. (1979) Biochemistry 18, 454-456] is interpreted as a reflection of the deformational anisotropy of the DNA molecule that facilitates its smooth folding in chromatin.

Crystals of the DNA tridecamer d(C-G-C-A-G-A-A-T-T-C-G-C-G) have been grown by the vapor-diffusion technique with 2-methyl-2,4-pentanediol as precipitant. They are monoclinic space group (C2, with a = 79.6 A , b = 43.1 A , c = 24.9 A and β = 98.7 ° . Previous nuclear magnetic resonance studies predicted that this tridecamer forms a duplex similar to the B DNA dodecamer, d(C-G-C-G-A-A-T-T-C-G-C-G), except for an extra adenosine residue that is stacked within the helix but remains unpaired: Preliminary X-ray diffraction studies confirmed that the tridecamer is in the B DNA conformation, consistent with the nuclear magnetic resonance results.

We present a fast algorithm to produce a graphic matrix representation of sequence homology. The algorithm is based on lexicographical ordering of fragments. It preserves most of the options of a simple naive algorithm with a significant increase in speed. This algorithm was the bais for a program, called DNAMAT, that has been extensively tested during the last three years at the Weizmann Institute of Science and has proven to be very useful. In addition we suggest a way to extend our approach to analyse a series of related DNA or RNA sequences, in order to determine certain common structural features. The analysis is done by 'summing' a set of dot-matrices to produce an overall matrix that displays structural elements common to most of the sequences. We give an example of this procedure by analysing tRNA sequences.

Publisher Summary The CORELS (COnstrained-REstrained Least-Squares) program is developed for proteins and nucleic acids to take advantage of the intrinsic rigid groups found in the molecules and to overcome the relatively low resolution of the X-ray data from their crystals. CORELS combines Scheringer's rigid groups constraints, extended to allow for variable torsion angles, with distance restraints to maintain stereochemistry between groups within a specified error limit. Even though allowing variable internal dihedral angles introduces torsional degrees of freedom within the otherwise constrained group, it can reduce the total number of structural parameters in the structure by decreasing the number of groups. The advantages of this approach include a large increase in the data-to-parameter ratio over the restrained refinement methods; automatic maintenance of group stereochemistry: within the group, all bond lengths and bond angles are constrained. The use of a constrained-restrained least-squares procedure has proven to be extremely useful in refining macromolecular structures, especially when the initial model has severe errors. This method inherently has many fewer degrees of freedom than restrained refinement procedures and therefore is applicable at extremely low resolution with a very large radius of convergence.

Atomic coordinates of E. Coli tRNA 1 Val have baen generated from the X-ray crystal structure of Yeast tRNA Phe by base substitution followed by idealization. The NMR spectrum of E. Coli tRNA 1 Val was then calculated using these coordinates and ring current calculations. The similarity between the calculated and observed NMR spectra of this tRNA gives credence to the procedure used previously to compare the solution and X-ray crystal structure of Yeast tRNA Phe and further suggests a strong similarity between the structures of Yeast tRNA Phe and E. Coli IRNA 1 Val both in solution and in the crystal. The results presented here indicate that a combination of NMR and atomic coordinates generated from a tRNA model may be a powerful procedure for determining three dimensional structure of a tRNA in solution.

The idea that DNA in chromatin can be packed smoothly, without breaking of base-stacking interactions, has been explored by both energetical estimations and stereochemical model building. A model of deformed DNA is built that fits reasonably to the known dimensions of a nucleosome. The model has slightly changed torsion angles in the sugarphosphate moieties relative to B DNA, varying gradually along the chains. The angle between planes of adjacent base pairs varies between 1 and 7 degrees. This model of deformed DNA does not have any unusually close nonbonded contacts and is evidently not the only possible model of smooth packing of DNA in chromatin. An energetical estimation of the critical radius of curvature of a smoothly bent DNA molecule is made using approximate potential functions for different van der Waals contracts in the B DNA structure. The critical radius of curvature of the deformed DNA-axis is close to the radius of a nucleosome (similar to 50 A). The smooth packing is a good alternative to models of kinked folding of DNA in nucleosomes.

Structural modelling techniques are employed to explore the energetic requirements for the transformation of classical B DNA into unwound yet double-stranded DNA structures. Structural idealization using CORELS computer program of Sussman et al. followed by energy minimization using the EREF program of Levitt, leads to two regular non-helical models. In both models, the bases are conventionally paired and stacked, yet there is no net rotation between successive base pairs. One model, N1, has a 1-bp repeating unit; the second, N2, has a 2-bp repeating unit. The dihedral angles of the backbone all have values found either in the B or the Z form of DNA, except for the P-O5'-C5'-C4' angle, which is in the unprecedented g+ or g- domains. The energy difference found between the two N form models and B form DNA are 6.6 and 3.4 kcal/mol/nucleotide for N1 and N2 respectively. These relatively low energy differences encourage the idea that non-helical forms of DNA may contribute to the alternate DNA structures found in S1 nuclease sensitive and other regulatory regions of active genes.

The 3-angstrom electron density map of crystalline yeast phenylalanine transfer RNA has provided us with a complete three-dimensional model which defines the positions of all of the nucleotide residues in the moleclule. The overall features of the molecule are virtually the same as those seen at a resolution of 4 angstroms except that many additional details of tertiary structure are now visualized. Ten types of hydrogen bonding are identified which define the specificity of tertiary interactions. The molecule is also stabilized by considerable stacking of the planar purines and pyrimidines. This tertiary structure explains, in a simple and direct fashion, chemical modification studies of transfer RNA. Since most of the tertiary interactions involve nucleotides which are common to all transfer RNA 's, it is likely that this three-dimensional structure provides a basic pattern of folding which may help to clarify the three-dimensional structure of all transfer RNA's.