Showing posts with label structural biology. Show all posts
Showing posts with label structural biology. Show all posts

Sunday, October 30, 2011

3DNA base-mutation functionality for the study of pretein-DNA interactions

In my last blog post titled "mutate_bases, a new 3DNA tool for in silico base mutations of 3D nucleic acid structures", I outlined the availability of the mutate_bases program and some areas of its possible applications, including to "perform base-pair mutations in DNA-protein complexes".

In the September 6, 2011 issue of PNAS (vol. 108, no. 36), AlQuraishi and McAdamsa from Stanford University published a high-profile article "Direct inference of protein–DNA interactions using compressed sensing methods" (pp. 14819–14824). See also the thoughtful Commentary by Vijay Pande, "(Compressed) sensing and sensibility". Essentially, by combining concepts from compressed sensing and statistical mechanics, AlQuraishi and McAdamsa have developed a novel approach to determine the energy potential of protein–DNA complexes, leading to "an impressive advance in predictive capability" -- ~90% accuracy compared with ~60% for the best-performing alternative computational methods.

I am so glad to find that 3DNA (ref nos. 12 and 13) was used in the work:
In silico mutagenesis was carried out using the 3DNA software package (12, 13), which maintains the backbone atoms of the DNA molecule, but replaces the base pair atoms in a way that is consistent with the backbone orientation in the crystal. (p.14821, middle of the right column)

Hopefully, the new and novel base mutation functionality in 3DNA will find more applications and be in wider use. Surely, I'll respond promptly to users' feedback and refine related 3DNA tools as necessary.

Tuesday, October 25, 2011

mutate_bases, a new 3DNA tool for in silico base mutations of 3D nucleic acid structures

In response to repeated requests from 3DNA users over the years, I recently (well, a few months ago) released a new utility program, mutate_bases, to the 3DNA suite of software package. As its name implies, mutate_bases can be used for mutating bases in a nucleic acid structure, with two key and unique features: (1) the sugar-phosphate backbone conformation is untouched; (2) the base reference frame (position and orientation) is reserved, i.e., the mutated structure shares the same base-pair/step parameters as the original one.

mutate_bases is a standalone ANSI C program, on a par with other major 3DNA programs (e.g., find_pair, analyze, rebuild, fiber etc). As seen from the help message below, it can be used for any nucleic-acid-containing structures (DNA, RNA, or their complexes) in PDB format:
------------------------------------------------------------------------
Usage: mutate_bases mutinfo pdbfile outfile
    'mutinfo' can contain upto 5 fields for each mutation:
        [name=residue_name] [icode=insertion_code]
        chain=chain_id seqnum=residue_number
        mutation=residue_name
    Alternatively, 'mutinfo' can be specified with the '-l' option
        followed by a mutation file: '-l mutations.dat'
    o The five fields per mutation can be in any order or CaSe
    o Each field can be abbreviated to its first character
    o Multiple mutations are separated by ';'
    o Fields in [] (i.e., name and icode) are optional
    o Mutation info should be QUOTED to be taken as one entry

Examples:
    mutate_bases 'c=a s=2 m=DA' 355d.pdb 355d_G2A.pdb
    # mutate G2 in chain A of B-DNA 355d to A
    mutate_bases 'c=a s=2 m=DA; c=B s=23 m=DT' 355d.pdb 355d_mutfile.pdb
    # mutate base-pair G-C (C23 in chain B) to A-T
        # also create file 'mutations.dat', see below
    mutate_bases -l mutations.dat 355d.pdb 355d_mutfile.pdb
        # 355d_mutfile.pdb and 355d_mutfile.pdb would be identical
    mutate_bases 'c=A s=74 m=U' 1evv.pdb 1evv_C74U.pdb
    # mutate C74 in chain A of tRNA 1evv to U
------------------------------------------------------------------------

mutate_bases is designed to solve the base mutation problem in a practical sense: robust and efficient, getting its job done and then out of the way. The program can have many possible applications: in addition to perform base-pair mutations in DNA-protein complexes, it should also prove handy RNA modeling and in providing initial structures for QM/MM/MD energy calculations, and in DNA/RNA modeling studies.

Thursday, July 14, 2011

GpU dinucleotide platform, the smallest unit with key RNA structural features

Compared to DNA, RNA has three salient structural features: it contains ribose sugar, uracil, and is normally single-standed. The O2'(G)...O2P(U) H-bond stabilized GpU dinucleotide platform may turn out to be the smallest unit with all those RNA hallmarks (see Figure below).


Firstly, it must have the guanosine ribose to form the O2'(G)...O2P(U) H-bond.

Secondly, the methyl group in position 5 of thymine would cause steric clash with guanosine, thus disrupting the N2(G)...O4(U) base-base H-bond to form the GpU dinucleotide platform.


Thirdly, a dinucleotide, by definition, is single-standed. The two H-bonds, plus the covalent linkage, makes the GpU platform extremely rigid (see Figure 1 of our 2010 NAR paper).

Moreover, the GpU platform is directional: swapping the two bases while keeping the sugar-phosphate backbone fixed does not allow for a base-base H-bond, thus no UpG dinucleotide platform.

Sunday, June 26, 2011

PDB v3.3 and partial atom occupancy

From the PDB mailing list, I know of the recent announcement "PDB Archive Version 4.0 to be Released July 13, 2011". This "ambitious" review of the PDB archive has resulted in a new set of corrected files in ten categories, including biological assemblies, residual B factors, peptide inhibitors/antibiotics, polymers containing nonstandard polymer linkages, and partial occupancy etc. "These data reflect the wwPDB's continuing commitment to providing accurate and detailed data to users worldwide."

I am interested in changes in the PDB format, and read the PDF document "Description of Changes and Corrections for PDB July 2011 Remediation Release":
PDB format files updated in this remediation release comply with PDB Format Version 3.30. PDBx and PDBML data files comply with the PDB Exchange Dictionary v.4.0, and PDBML XML Schema V4.0, respectively.
Specifically, I checked carefully the section on "Partial occupancy", which is quoted in full below:
Problem

In the 2009 remediation, occupancies were corrected in 490 X-ray
and neutron entries. A mistake was made in 104 of these entries:
for atoms with alternate conformer labels and with summed total
occupancy less than 1.0, the occupancies were re-scaled as 1.0/n,
where n is the number of conformers.

Approach

The originally deposited occupancies of the affected atoms were
restored and the remediation was then carried out properly, via:
  • Atoms with multiple conformations but identical coordinates and B-values were merged and their occupancies were summed.
  • Atoms which now have (total) occupancies <= 1.0 were left as deposited.
  • Atoms with (total) occupancies > 1 were rescaled proportionally to a sum of 1.0
Results

The occupancies have been corrected in these entries.

This partial occupancy issue reminds me of an extensive and very informative discussion a few months back in CCP4BB, under the thread "what to do with disordered side chains" and its derivatives, about setting "zero" occupancy and/or high B values for PDB ATOM/HETATM records in disordered regions. Over the course of the discussion, Frances Bernstein made the following comment:
I am absolutely positive that there is software that does its voodoo on ATOM/HETATM records and pays absolutely no attention to anything beyond the x, y, z coordinates (i.e. beyond column 54).
3DNA does pay attention beyond column 54 (up to 80, actually) for ATOM/HETATM records, but internally it does not make use of the occupancy info. In future releases of 3DNA, I am planning to take occupancy/B-factor into consideration, probably through configurable parameters.

Reading through the "Description of Changes and Corrections for PDB July 2011 Remediation Release", I noticed cases of re-corrections of previous corrections in wwPDB remediation efforts. A concrete example is about partial occupancy in the 2009 remediation: among the 490 corrected X-ray and neutron entries, a mistake was made in 104 of them. As a side note, there was a post by Morten Kjeldgaard, titled "PDB changes data in entries?" early this year in CCP4BB, where partial occupancy was used as an example.

Sunday, June 5, 2011

Lower case chain identifiers in PDB format

First formulated in early 1970s, the PDB format is rigid with fixed columns for designated contents in its ATOM/HETATM records. Specificlly, a single column, #22, is assigned for the chain identifier (id). Traditionally, the 26 upper case letters of English alphabet (A-Z), space (i.e., ' '), and the single digits (0-9) have been used as chain ids. Up until the ribosomal structures came up, I guess, those 26 + 1 + 10 = 37 characters had been sufficient for the chain ids.

To the best of my knowledge, for a long time, most PDB parsers assume upper case chain ids. Indeed, 3DNA v1.5 automatically converts each ATOM/HETATM records to upper cases. The first time I became aware of lower case chain ids was when I saw a post in the 3DNA forum, titled "Small bug in find_pair", where a user reported the 'w' vs 'W' chain ids in PDB entry 1VSP. Then I refined 3DNA so that the case of chain ids can be preserved, through an undocumented command line option (as a feature for internal testing purpose).

My view to make 3DNA chain ids case-sensitive has been reinforced when I read the article "Crystal structures of CGG RNA repeats with implications for fragile X-associated tremor ataxia syndrome" recently published in Nucleic Acids Research. The asymmetric unit of the unmodified CGG-repeats-containing duplex (GCGGCGGC)2, NDB entry NA1017 / PDB entry 3R1C, contains a total of 36 chains: designated as A-Z, plus a-j. Without distinguishing cases of the chain ids, the 3DNA output would become quite confusing.

Thus, in future releases of 3DNA, the default would be switched to preserve the case of chain ids. This chain id 'case' serves as an excellent example that scientific software products, unlike publications per se, are not fixed but need continuous care and maintenance to meet the challenges of an evolving world.

Sunday, March 13, 2011

Review article on NMR analysis of protein–DNA interactions by Milon et al

Through Google scholar, I became aware of a recent review article by Milon et al., titled "Nuclear magnetic resonance analysis of protein–DNA interactions" in the journal J. R. Soc. Interface:
This review focuses on the experimental strategies currently employed to solve structures of protein–DNA complexes and to analyse their dynamics. It highlights how these approaches can help in understanding detailed molecular mechanisms of target recognition.
I browsed through the text to get myself more familiar with NMR the methodology and its applications in protein-DNA recognition. I was surprised that 3DNA was cited in the article, especially with respect to its unique analyze/rebuild complementarity:
In addition, several software programs have been developed to model DNA bending such as the 3DNA program, which allows analysis of DNA structural parameters and enables it to be rebuilt with customized DNA models [76]. Several Web servers have been created recently and provide interesting tools to analyse and rebuild DNA models [77,78].
I am only wishing that 3DNA's neat features could be more widely recognized; hopefully I'd have the opportunity to further refine 3DNA and move it to the next level.

Saturday, October 23, 2010

Chi (χ) torsion angle characterizes base/sugar relative orientation

Except for pseudouridine, a nucleoside in DNA/RNA contains an N-glycosidic bond that connects the base to the sugar. The chi (χ) torsion angle, which characterizes the relative base/sugar orientation, is defined by O4'-C1'-N1-C2 for pyrimidines (C, T and U), and O4'-C1'-N9-C4 for purines (A and G).

Normally (as in A- and B-form DNA/RNA duplex), χ falls into the ranges of +90° to +180°;  –90° to –180° (or 180° to 270°), corresponding to the anti conformation (Figure below, top). Occasionally, χ has values in the range of –90° to +90°, referring to the syn conformation (Figure below, bottom). Note that in left-handed Z-DNA with CG repeating sequence, the purine G is in syn conformation whilst the pyrimidine C is anti.




Presumably, the χ-related anti/syn conformation is a very basic and simple concept. In essence, though, the N-glycosidic bond and the corresponding χ torsion angle illustrate that the base and sugar are two separate entities, i.e. there is an internal degree of freedom between them. In this respect, it is worth noting that the Leontis-Westhod sugar edge for base-pair classification corresponds to the anti form only. When a base is flipped over into the syn conformation,  the "sugar edge", defined in connection with the minor (shallow) groove side of a nitrogenous bases, simply does not exist.

Base-flipping (anti/syn conformation switch) is one of the factors associated with the two possible relative orientations in a base pair, characterized explicitly in 3DNA as of type M+N or M–N since the 2003 NAR paper (Figure 2, linked below). I reemphasized this distinction in our 2010 GpU dinucleotide platform paper (in particular, see supplementary Figure S2). Unfortunately, this subtle (but crucial, in my opinion) point has never been taken seriously (or at all) by the RNA community, even with 3DNA's wide adoption. However, as people know 3DNA deeper/better and take RNA base-pair classification more rigorously, I have no doubt they will begin to appreciate the simplicity of this explicit distinction and the resultant full quantification of each and every possible base pair using standard geometric parameters.




On a related issue, current versions of 3DNA (v1.5 and v2.0) output only the χ torsion angle without providing the anti/syn classification. This defect, and many others, will hopefully be rectified in future releases of 3DNA.

Sunday, September 19, 2010

New NDB entries have IDs start with the prefix NA?

Recently, while reading the article titled "Designing Triple Helical Fragments: The Crystal Structure of the Undecamer d(TGGCCTTAAGG) Mimicking T.AT Base Triplets" by Van Hecke (Crystal Growth & Design), I noticed the following:
The atomic coordinates and structure factors have been deposited in the Protein Data Bank41 and Nucleic Acid Database42 (PDB and NDB entry codes 3L1Q and NA0392, respectively).
The NDB id NA0392 reminded me of an email communication I had with Dr. Olson early this year when she told me of the id change of new NDB entries. Occupied with other issues, I did not pay much attention to this point until recently.

Over the years, NDB has established itself prominently as "a repository of three-dimensional structural information about nucleic acids". Traditionally, an NDB id has some associated "meanings", with noticeably exceptions, as discussed one year ago in my blog post "PDB id vs NDB id". Thus, it is not surprising that NDB decided to make changes in naming new ids. However, I cannot find any announcement on the id change in the NDB website; a Google search on "NDB id" did not uncover anything new either. Luckily, NDB provides search by release date ("Released Since"). After a few tries, I traced that the new id policy began to be implemented from around March 2010.

The new NDB id convention appears to be NA (presumably standing for for Nucleic Acid) followed by 4 digits. As more entries available, it is not that hard to imagine that the number of digits must be expanded to 5, 6, or even more. Naturally, PDB id, with a fixed 4-character length (up to now), is (far) more consistent than the NDB id is.

Sunday, September 5, 2010

Identification of C-H...N/O H-honds using 3DNA

Recently, I came across an interesting article by Kiliszek et al., titled "Atomic resolution structure of CAG RNA repeats: structural insights and implications for the trinucleotide repeat expansion diseases". In addition to its biological implications, this paper uses 3DNA to deal with non-canonical-base-pair-containing helical structures in a sensible way:
The helical parameters were calculated using 3DNA (29). Sequence-independent measures were used, based on vectors connecting the C1' atoms of the paired residues, to avoid computational artefacts arising from non-canonical base pairing.
Another significant point, which is the focus of this post, is the observation that "All the adenosines are in the anti-conformation and the only interaction within each A-A pair is a single C2-H2...N1 hydrogen bond." (Figure below) Given the 0.95 Å ultrahigh resolution of structure 3nj6/na0608, it is likely that this type of A-A is real.


The find_pair program in the currently distributed versions of 3DNA (v2.0 and before), however, does not identify this A-A pair (thus the corresponding NDB list of "Base Pair Step Parameters" is incomplete) for the following two reasons:
  1. No H-bond exists between N/O base atoms – currently a requirement for a base-pair.
  2. By default, only N/O atoms are used in defining H-bonds (see tag hb_atoms in file "misc_3dna.par"). Nevertheless, by adding C as a possible atom in forming H-bond, and manually editing find_pair generated input file to analyze, 3DNA structural parameters can be calculated as usual.
I have updated find_pair in 3DNA to identify such C-H...N/O H-bond automatically (see below for entry 3nj6). Upon further refinements and validations, future releases of 3DNA will have this functionality available.

    1   95  #    1 | ...1>A:...1_:[..G]G-----C[..C]:..10_:A<...2
    2   94  #    2 | ...1>A:...2_:[..G]G-----C[..C]:...9_:A<...2
    3   93  #    3 | ...1>A:...3_:[..C]C-----G[..G]:...8_:A<...2
    4   92  #    4 | ...1>A:...4_:[..A]A-**--A[..A]:...7_:A<...2
    5   91  #    5 | ...1>A:...5_:[..G]G-----C[..C]:...6_:A<...2
    6   90  #    6 | ...1>A:...6_:[..C]C-----G[..G]:...5_:A<...2
    7   89  #    7 | ...1>A:...7_:[..A]A-**--A[..A]:...4_:A<...2
    8   88  #    8 | ...1>A:...8_:[..G]G-----C[..C]:...3_:A<...2
    9   87  #    9 | ...1>A:...9_:[..C]C-----G[..G]:...2_:A<...2
   10   86  #   10 | ...1>A:..10_:[..C]C-----G[..G]:...1_:A<...2
##### Criteria: 4.00  0.00  15.00  2.50  65.00  4.50  7.50   [ O N C]
##### 2 non-Watson-Crick base-pairs, and 1 helix (0 isolated bps)
##### Helix #1 (10): 1 - 10

The Jmol paper: a paradigm shift in crystallographic visualization

From the Jmol mailing list, I noticed Prof. Robert Hanson's announcement titled "Jmol and crystallography: the paper". Finally, an "official" paper has been published on Jmol, an increasing popular molecular visualization tool.

While the paper is categorized under "teaching and education" and summarized as below:
Recent advances in molecular and crystallographic visualization methods are allowing instructors unprecedented opportunities to enhance student learning using virtual models within a familiar web-browser context. In step with these advances, the latest versions of the Jmol molecular visualization applet offer capabilities that hold potential for revolutionizing the way students learn about symmetry, uncertainty and the overall enterprise of molecular structure determination.
Jmol is certainly as useful for research as other commonly used molecular visualization tools, such as PyMOL and RasMol. Specifically, Jmol applets have become dominated in web-based applications/resources (including RCSB PDB). In fact, the HTML version of the paper itself makes extensive use of Jmol applet for interactive manipulation of over 20 figures.

Over the past couple of years, I have witnessed the many new features added to Jmol, driven by an active and friendly user community. I have been particularly impressed by Prof. Hanson's quick and to-the-point responses to users' requests for new features and bug reports, including to my request for Jmol's "Support of 'alchemy' format for rectangular schematic base-pair geometry". Overall, the paper presents a nice overview of some of Jmol's new functionality, and could become highly cited.

Reading through the paper, I was just a bit surprised that PyMOL is not mentioned. Nowadays, Jmol and PyMOL are apparently the top two molecular graphics software tools, with common and complementary functionality.

Saturday, August 21, 2010

Calculation of the chi (χ) torsion angle of pseudouridines in RNA structures

In RNA structures, chi (χ) is defined as the torsion angle about the N-glycosidic bond between the ribose sugar and the canonical A/C/G/U bases or their modified variants. Specifically, for pyrimidines (C and U), χ is defined by O4'-C1'-N1-C2; and for purines (A and G) by O4'-C1'-N9-C4. See the following figure from the IMB Jena Image Library:



Pseudouridine (5-ribosyluracil, PSU) was the first identified modified nucleoside in RNA and is the most abundant. PSU is unique in that it has a C-glycosidic bond instead of the N-glycosidic bond common to all other nucleosides, canonical or modified. It thus poses a problem as to how to calculate the χ torsion angle: should it be O4'-C1'-C5-C4 reflecting the actual glycosidic bond connection, or should the conventional definition O4'-C1'-N1-C2 still be applied literally? As a concrete example, the figure below shows the (slightly) different numerical values, as given by the two definitions, for PSU 6 on chain A of PDB/NDB entry 3cgp/ar0093.


Needless to say, definition of the χ torsion angle of PSU in RNA structures is a very subtle/minor point, and I am not aware of any discussion on this issue in literature (I'd appreciate your sharing of related information in the comment). In 3DNA, PSU is identified explicitly, and χ is defined by O4'-C1'-C5-C4. In NDB and a couple of other tools (I've played with), χ for PSU is defined by O4'-C1'-N1-C2. Again using 3cgp/ar0093 (figure above) as an example, 3DNA gives -162.7°, whilst NDB gives -163.9°. Hopefully, this post will help clarify a confusion for those who care about such little details.

Saturday, August 14, 2010

PDB format, how many variants are there?

In the field of structural biology of macromolecules (proteins and nucleic acids), the PDB format is still the primary one, even though it has shown its age/limitations, and mmCIF and PDBML have been introduced as alternatives.

In 3DNA, PDB is currently the only accepted format for the analysis routines (analyze/cehs), the primary format for rebuilt structures (rebuild/fiber), and various other output structure files (e.g., bestpairs.pdb etc.). By and large, 3DNA follows the PDB format documented at the RCSB website. In my maintaining and supporting of 3DNA and the forum over the years , I have come across other variants of the PDB format. Software packages customize PDB in subtle, sometimes substantial, ways that cause problems with 3DNA. Some examples follow:
  • Recently, a user tried to analyze an RNA duplex using w3DNA, but found that "the server seems not to recognize the file". It turned out that the so-called PDB format does not have residue name/number and chain id aligned in columns properly. For example, among other issues, the chain id in the problematic PDB is not at column #22 as specified by the standard.
  • A couple of months back, another user asked how to handle AMBER-introduced ribonucleotide abreviations (RA/RC/RG/RU). Compared to the above "PDB" variant, the AMBER PDB is well-behavored, and can be easily recognized by 3DNA after introducing corresponding one letter matches in file baselist.dat. Interestingly, a couple of years ago, DA/DC/DG/DT was introduced into the PDB format to distinguish deoxyribonucleotide (DNA) from ribonucleotide A/C/G/U (RNA). The AMBER convention, by using RA/RC/RG/RU for RNA, appears to be more symmetric.
  • From my (limited) experience in helping a 3DNA user several years ago, I know that CHARMM uses three-letter residue names for DNA (e.g., ADE/CYT/GUA/THY). There is actually a short Perl utility program (x3dna2charmm_pdb) in 3DNA for converting 3DNA generated PDB file to that recognized by CHARMM.
  • As noted in another post, "PDB ATOM coordinates record", Babel/Open Babel converted PDB files (mostly of small molecules) do not follow atom naming conventions.
As widely used and standard as PDB format is, there are still many variants in common use. Thus, when you meet a problem related to a PDB file, it helps to check if it conforms to the standard. With 3DNA, I've always tried to incorporate as many PDB variants as possible, so users can run 3DNA on them transparently.

Saturday, August 7, 2010

Calculation of DNA bending angle

Once in a while, I receive the question (via email or on the 3DNA forum) on how to calculate DNA bending angle, mostly associated with DNA-protein complexes. This question certainly qualifies as an FAQ; Indeed, I did consider writing an entry on it in the 3DNA website. However, I've hesitated to do so because the topic is more subtle and complicated than it appears.

On its face, an angle is defined by two vectors; let's call them a and b, and if each is normalized, then the angle (in degrees) between them is: acos(dot(a, b)) * 180/pi. Geometrically, after moving the tails of the vectors into the same position (e.g., origin), the heads would normally define a plane, unless a and b are strictly parallel (0°) or anti-parallel (180°).

DNA structures are three-dimensional, normally far more complicated than a single number can quantify. The concept of DNA bending angle, as I understand it, is only applicable to DNA structures with two relatively straight fragments (as in CAP-DNA complexes). Under such situations, one can fit a least-squares (ls) linear helical axis to each of the two fragments, and calculate the angle between them. Towards this end, 3DNA outputs the following section when it judges that the input structure is not strongly curved. Using 355d/bdl084, which is distributed with 3DNA, as an example:
Global linear helical axis defined by equivalent C1' and RN9/YN1 atom pairs                                    
Deviation from regular linear helix: 3.30(0.52)
Helix:    -0.127  -0.275  -0.953
HETATM 9998  XS    X X 999      17.536  25.713  25.665
HETATM 9999  XE    X X 999      12.911  15.677  -9.080
Average and standard deviation of helix radius:
P: 9.42(0.82), O4': 6.37(0.85),  C1': 5.85(0.86)
Where the Helix: line gives the normalized vector along the "best-fit" helical axis. The two HETATM records provides the two end points on the helix, and they are directly related to the Helix: line by a simple equation. Following the above example, we have (Octave/Matlab code):

XE = [12.911 15.677 -9.080];
XS = [17.536 25.713 25.665];

dd = XE - XS
% -4.6250 -10.0360 -34.7450

Helix = dd / norm(dd)
% -0.12685 -0.27526 -0.95296 ==> [-0.127 -0.275 -0.953]

With the two HETATM records, one can easily add them into the original PDB file to display the helical axis using a molecular graphics programs (e.g., RasMol, Jmol, PyMol). Moreover, the two helix vectors can be used to reorient the original PDB structure into a view so that one helical fragment lies along the x-axis, and the other in the xy-plane. As documented in detail in recipes #4 on "Automatic identification of double-helical regions in a DNA–RNA junction" (2008 3DNA Nature Protocols paper), "The chosen view allows for easy visualization and protractor measurement of the overall bending angle between the two relatively straight helices."

The following points are well worth noting:
  • The ls fitting procedure used in 3DNA follows SCHNAaP, which was based on the algorithm in the well-known NewHelix program, maintained by Dr. Richard Dickerson upto the 1990s. While fitting a global linear helical axis to strongly curved DNA structures clearly makes no much sense with derived parameters (NewHelix itself has been replaced by FreeHelix, also from Dickerson), I do believe it is meaningful to fit a linear helix with relatively straight DNA fragment. That's why I have kept this functionality in SCHNArP and 3DNA; 3DNA bending angle calculation serves as an example illustrating the point – it provides an "intuitive" way for biologists to understand how the bending angle is calculated; it can actually be measured directly.
  • Instead of directly ls-fitting a linear helical axis with 3DNA, one can alternatively superimpose a regular fiber model into the DNA fragment, and then derive the straight helical axis from the fitted coordinates. The two approaches normally gives slightly different numerical values, as would be expected.
  • Overall, bending angle is (at most) an approximate measure of DNA curvature. In my opinion, the concept is only applicable for comparing a set of structures, each with two relatively straight fragments. Even in such cases, the relative spatial relationship between two helical fragments is more complicated than a simple (bending) angle could quantify. Specially, it is not meaningful to make a selling point in slight differences of DNA bending angles.

Sunday, August 1, 2010

Principle axes of inertia as a mean to reorient a molecule

In structural biology, it is often helpful to reorient a molecule in a specific view for easy visualization and comparison. With no a priori knowledge of a feature, the (objective) principle axes of inertia can be used as a mean to automatically reset a structure into its most extended orientation (and two other orthogonal views). In 3DNA, rotate_mol has such a functionality which is explored by blocview to generate the characteristic, simplified molecular images adopted by the RCSB PDB and NDB.

Recently, the thread (initiated by Dr. Pascal Auffinger) at the 3DNA forum titled "rotate_mol ambiguous orientations" has led me to think the topic deeper. As always, the issue is best illustrated with an example. The file 355d_A6_T19.pdb contains the coordinates of the base pair A6-T19 of PDB entry 355d. Let xyz be a variable in Octave (Matlab) containing the coordinates of all atoms in file 355d_A6_T19.pdb, the following m-file provides the detailed procedure:
xyz = [
13.954 12.080 15.071
13.092 11.246 15.953
14.769 11.453 14.020
13.056 13.202 14.406
12.184 13.965 15.222
11.308 14.846 14.376
12.033 15.991 13.896
10.654 14.188 13.168
9.262 14.480 13.256
11.373 14.820 11.974
11.810 16.185 12.511
13.051 16.742 11.953
14.253 16.097 11.802
15.197 16.853 11.305
14.585 18.084 11.106
15.066 19.322 10.605
16.329 19.520 10.209
14.198 20.357 10.535
12.938 20.162 10.946
12.370 19.048 11.444
13.259 18.034 11.497
14.865 29.619 5.710
14.172 30.844 5.281
16.107 29.695 6.515
13.846 28.701 6.510
12.684 28.174 5.864
11.909 27.311 6.826
12.626 26.090 7.129
11.584 27.960 8.173
10.189 27.833 8.395
12.409 27.166 9.184
12.544 25.808 8.521
13.716 24.979 8.867
13.496 23.688 9.310
12.381 23.202 9.467
14.631 22.971 9.551
15.943 23.388 9.393
16.852 22.613 9.632
16.107 24.759 8.938
17.488 25.309 8.745
14.995 25.475 8.711 ]

cov_mtx = cov(xyz)

% cov_mtx =
% 3.4094 2.3860 -1.4222
% 2.3860 33.7089 -15.7175
% -1.4222 -15.7175 7.8786

[v, lambda] = eig(cov_mtx)

% v =
% 0.094830 0.992835 -0.072709
% 0.419520 -0.106092 -0.901525
% 0.902779 -0.054989 0.426575

% lambda =
% 0.42535 0 0
% 0 3.23319 0
% 0 0 41.33836

x = v(:, 2);
y = v(:, 3);
z = cross(x, y);

% z =
% -0.094830
% -0.419520
% -0.902779

num = size(xyz, 1);
xyz2 = (xyz - ones(num, 1) * mean(xyz)) * [x y z]

% xyz2 =
% 1.148136 10.032849 -0.514731
% 0.332293 11.223635 -0.879359
% 2.081609 10.090517 0.619843
% .................................
% 2.277813 -4.170312 -0.501254
% 3.601180 -4.848890 -0.688714
% 1.110301 -4.831784 -0.491249
Among other subtitles, note that:
  1. lambda 3 has the largest value (41.33836), i.e., with the biggest variance, and its corresponding eigenvector is set as the vertical y-axis.
  2. lambda 2 shows intermediate variation (3.23319), and its corresponding eigenvector is set as the horizontal x-axis.
  3. The z-axis is derived as the cross-product of x- and y-axes to ensure a right-handed frame. Note that it is the opposite of eigenvector 1.
  4. The original xyz coordinates are moved to its geometric center [xyz - ones(num, 1) * mean(xyz)], and then projected onto the preferred frame to get the transformed coordinates (xyz2) with the molecule in the most extended view (vertical then horizontal).

Sunday, June 27, 2010

Get all torsion angles in a nucleic acid structure

In the field of nucleic acid structure analysis, a commonly calculated set of parameters is the torsion angles. Specifically, they include the main chain and chi (χ) torsion angles defined as follows:
  • α: O3'(i-1)-P-O5'-C5'
  • β: P-O5'-C5'-C4'
  • γ: O5'-C5'-C4'-C3'
  • δ: C5'-C4'-C3'-O3'
  • ε: C4'-C3'-O3'-P(i+1)
  • ζ: C3'-O3'-P(i+1)-O5'(i+1)
  • χ: for pyrimidines (Y, i.e., T, U, C): O4'-C1'-N1-C2; for purines (R, i.e., A, G): O4'-C1'-N9-C4
A related set of parameters characterizes the sugar conformation:
  • ν0: C4'-O4'-C1'-C2'
  • ν1: O4'-C1'-C2'-C3'
  • ν2: C1'-C2'-C3'-C4'
  • ν3: C2'-C3'-C4'-O4'
  • ν4: C3'-C4'-O4'-C1'
  • tm: amplitude of pseudorotation of the sugar ring
  • P: phase angle of pseudorotation of the sugar ring
These torsion angles are clearly defined and are readily available from various informatics programs. Not surprisingly, 3DNA also provides a complete set of DNA/RNA backbone torsions, calculated robustly and efficiently. The key is the "-s" option of find_pair" program, which is nevertheless little used, mostly because it is not the default.

Using the Haloarcula marismortui 50S large ribosomal subunit as an example (1jj2), the output file 1jj2.outs from the following command contains all the above mentioned main chain and sugar conformational parameters:
find_pair -s 1jj2.pdb stdout | analyze
Please see the Jena "Nucleic acid backbone parameters" website for a diagram (based on Saenger's book) defining the various backbone torsion angles.

Sunday, May 23, 2010

MODEL/ENDMDL records in PDB X-ray crystal structures

Over the years, I have always been using the PDB format for nucleic acid structures. Naturally I've thought that I know the format well, especially the "Coordinate Section." According to the PDB documentation,
MODEL/ENDMDL records are used only when more than one structure is presented in the entry, as is often the case with NMR entries.
In my experience, I have always connected the MODEL/ENDMDL pair only with the different models in an NMR entry. However, I was recently bitten by a subtlety in PDB format that is related to the MODEL/ENDMDL records in X-ray crystal structures which contain more than one asymmetric unit in their biological assembly.

As always, the point is best illustrated with an example – here I am using the four-stranded DNA Holliday junction X-ray crystal structure (1zf4/ud0061) solved by Ho and colleagues [PNAS 2005 May 17;102(20):7157-62]. As shown in the NDB website, the asymmetric unit of 1zf4/ud0061 contains only two chains; it is the biological assembly that has the four-stranded DNA junction.

I downloaded the biological assembly coordinates (in PDB format) from the NDB (named it '1zf4.pdb'), and ran blocview on it: blocview -i=1zf4.png 1zf4.pdb. However, the generated image [see Figure (A) below] has only half of what expected – as if the downloaded file contains only coordinates of an asymmetric unit. The mystery was gone when I checked the PDB file and realized that the MODEL/ENDMDL records now also apply to X-ray crystal structures to delineate symmetric-related units. Since 3DNA is designed to handle only one structure – it stops processing whenever an END or ENDMDL record is encountered. Simply change each 'ENDMDL' record to ' ENDMDL' (i.e., adding a space, or any character, for that matter) and run blocview again will get the expected image [Figure (B)].

Note that by default, RasMol also only displays the first model in a PDB file. To see multiple structures, the option '-nmrpdb' must be specified in the command line.

Friday, April 2, 2010

What's special about the GpU dinucleotide platform?

Recently, I (together with Drs. Wilma Olson and Harmen Bussemaker – a team with a unique combination of complementary expertise) published a new article in Nucleic Acids Research (NAR): "The RNA backbone plays a crucial role in mediating the intrinsic stability of the GpU dinucleotide platform and the GpUpA/GpA miniduplex". The key findings of this work are summarized in the abstract:
The side-by-side interactions of nucleobases contribute to the organization of RNA, forming the planar building blocks of helices and mediating chain folding. Dinucleotide platforms, formed by side-by-side pairing of adjacent bases, frequently anchor helices against loops. Surprisingly, GpU steps account for over half of the dinucleotide platforms observed in RNA-containing structures. Why GpU should stand out from other dinucleotides in this respect is not clear from the single well-characterized H-bond found between the guanine N2 and the uracil O4 groups. Here, we describe how an RNA-specific H-bond between O2'(G) and O2P(U) adds to the stability of the GpU platform. Moreover, we show how this pair of oxygen atoms forms an out-of-plane backbone ‘edge’ that is specifically recognized by a non-adjacent guanine in over 90% of the cases, leading to the formation of an asymmetric miniduplex consisting of ‘complementary’ GpUpA and GpA subunits. Together, these five nucleotides constitute the conserved core of the well-known loop-E motif. The backbone-mediated intrinsic stabilities of the GpU dinucleotide platform and the GpUpA/GpA miniduplex plausibly underlie observed evolutionary constraints on base identity. We propose that they may also provide a reason for the extreme conservation of GpU observed at most 5'-splice sites.

As a nice surprise, this publication was selected by NAR as a featured article! According to the NAR website:
Featured Articles highlight the best papers published in NAR. These articles are chosen by the Executive Editors on the recommendation of Editorial Board Members and Referees. They represent the top 5% of papers in terms of originality, significance and scientific excellence.

I feel very gratified with the "extra" recognition. From my own perspective, I can easily rank this paper as the top one in my publication list: from the very beginning, I has been struck by the simplicity and elegance of the GpU story. Hopefully, time will verify the validity of this scientific contribution.

Behind the hood, though, there is a long, complex (sometimes perplexing), yet interesting story associated with this work. Here is how it got started. While writing the 3DNA 2008 Nature Protocols (NP) paper, I selected the (previously undocumented) "-p" option of "find_pair" to showcase its capability to identify higher-order base associations, using the large ribosomal subunit (1JJ2) as an example. I noticed the unexpected O2'(G)⋅⋅⋅O2P(U) H-bond within the GpU dinucleotide platform in the pentaplet shown left in Figure A below. I was well aware of Leontis-Westholf's pioneering work on "Geometric nomenclature and classification of RNA base pairs" which involves three distinct edges – the Watson-Crick edge, the Hoogsteen edge, and the Sugar edge, yet without taking into consideration of possible sugar-phosphate backbone interactions (Figure B below). So I decided to double-check, just to be sure that the H-bond was not spurious due to defects in the H-bond detecting scheme of "find_pair", and the results were very surprising.

The following section was re-added into the 3DNA NP paper in the very last revision:
It is also worth noting that the G1971–U1972 platform is stabilized not only by the well-characterized G(N2)⋅⋅⋅U(O4) H-bond interaction, but also by a little-noticed G(O2’)⋅⋅⋅U(O2P) sugar-phosphate backbone interaction (Fig. 6a). Examination of the 50S large ribosomal unit (1JJ2) alone reveals ten such double H-bonded G–U platforms, far more occurrences than those registered by any other dinucleotide platform (including A–A) in this structure. Apparently, the G–U platform is more stable than other platforms with only a single base–base H-bond interaction. We are currently investigating this overrepresented G–U dinucleotide platform in other RNA structures. (p.1226)

Saturday, March 13, 2010

Hoogsteen base-pair

The A·U (or A·T) Hoogsteen pair is a well-known base pair (bp), named after the scientist who discovered it. As shown in the Figure below (left), in the Hoogsteen bp scheme, adenine uses its N7 and N6 atoms (at the major groove edge) to form two H-bonds with the N3 and O4 atoms from uracil, respectively. Interestingly, if the uracil base ring is flipped around the N7(A)…N3(U) H-bond by 180 degrees, N6(A) can also form an H-bond with O2(U), i.e., N6(A)…O2(U): this pairing scheme is called the reverse Hoogsteen bp (right).


I first came to know about the Hoogsteen bp from Saenger's book ("Principles of Nucleic Acid Structure"). Over the years, I have read many articles mentioning the Hoogsteen bp and touched this topic myself in the 2003 3DNA NAR publication. However, I have never read Hoogsteen's two original publications on this topic until recently:
  1. The two-page long preliminary report, titled "The structure of crystals containing a hydrogen-bonded complex of 1-methylthymine and 9-methyladenine", was published in Acta Cryst. (1959). 12, 822-3. The paper contained only a single reference to the Watson-Crick DNA structure paper, published in Nature in 1953. I found it very revealing to understand why Hoogsteen used the methyl-ed derivatives of thymine and adenine, and how the failed initial interpretation of the experimental "vector-density map" using the Watson-Crick A-T bp led to the discovery of the new base-pairing scheme:
    The fact that the first trial structure could not be refined led to a more critical scrutiny of the generalized projection and a greater emphasis on the significance of certain spurious peaks and on relatively large variations in the heights of peaks that were assumed to represent atoms. The correct structure was finally discovered by changing the positions of a few atoms in the 9-methyladenine portion of the asymmetric unit.
  2. The more extensive account of the Hoogsteen bp story, titled "The Crystal and Molecular Structure of a Hydrogen-Bonded Complex Between 1-Methylthymine and 9-Methyladenine", published in Acta Cryst. (1963) 16, 907-16.
I like these two papers, and more generally those focused-articles, where authors get directly to a point and addressed it thoroughly and clearly. Most publications nowadays are very ambitious, trying to solve "big problems": the papers are generally far more complicated and often have "reproducibility" problems.

As a side note, the term Hoogsteen "edge" appears quite frequently in today's publications of RNA structures: in the Leontis-Westhof bp classification scheme, the term simply means the major groove edge in what would be a Watson-Crick bp geometry.

Saturday, October 31, 2009

How to calculate torsion angle?

Given the x-, y-, and z-coordinates of four points in 3-dimensional (3D) space, how to calculate torsion angle? Overall, this is a well-solved problem in structural biology, and one can find detailed description of the algorithm in text books and on-line documents. The algorithm for calculating torsion angle is implementated in virtually every software package in structural chemistry or biology.

Basic as it is, however, in my experience, it is very important to have a detailed appreciation of how the method works in order to really get into the 3D world. Here is a worked example using Octave/Matlab of my simplified, geometry-based implementation of how to calculate torsion angle, including how to determine its sign. No theory or (complicated) mathematical formula, just a step-by-step illustration of how I solve this problem.
  1. Coordinates of four points in variable abcd:
    abcd = [ 21.350  31.325  22.681
    22.409 31.286 21.483
    22.840 29.751 21.498
    23.543 29.175 22.594 ];
  2. Two auxiliary functions: norm_vec() to normalize a vector; get_orth_norm_vec() to get the orthogonal component (normalized) of a vector with reference to another vector, which should have been normalized.
    function ovec = norm_vec(vec)
    ovec = vec / norm(vec);
    endfunction

    function ovec = get_orth_norm_vec(vec, vref)
    temp = vec - vref * dot(vec, vref);
    ovec = norm_vec(temp);
    endfunction
  3. Get three vectors: b_c is the normalized vector b→c; b_a_orth is the orthogonal component (normalized) of vector b→a with reference to b→c; c_d_orth is similarly defined, as the orthogonal component (normalized) of vector c→d with reference to b→c.
    b_c = norm_vec(abcd(3, :) - abcd(2, :))
    % [0.2703158 -0.9627257 0.0094077]
    b_a_orth = get_orth_norm_vec(abcd(1, :) - abcd(2, :), b_c)
    % [-0.62126 -0.16696 0.76561]
    c_d_orth = get_orth_norm_vec(abcd(4, :) - abcd(3, :), b_c)
    % [0.41330 0.12486 0.90199]
  4. Now the torsion angle is the angle between the two vectors, b_a_orth and c_d_orth, and can be easily calculated by their dot product. The sign of the torsion angle is determined by the relative orientation of the cross product of the same two vectors with reference to the middle vector b→c. Here they are in opposite direction, thus the torsion angle is negative.
    angle_deg = acos(dot(b_a_orth, c_d_orth)) * 180 / pi
    % 65.609
    sign = dot(cross(b_a_orth, c_d_orth), b_c)
    % -0.91075
    if (sign < 0)
    ang_deg = -angle_deg % -65.609
    endif
A related concept is the so-called dihedral angle, or more generally the angle between two planes. As long as the normal vectors to the two corresponding planes are defined, the angle between them is easy to work out.

Moreover, the method to calculate twist angles of helical nucleic acid structures in SCHNAaP and 3DNA is essentially the same.

Sunday, October 18, 2009

How to quantify the relative geometry between two base-pairs? — Part 1

In the field of double helical DNA (and RNA) structures, the following two sets of parameters to normally used to quantify the relative geometry of the two base pairs in a dinucleotide step:
  • shift, slide, rise, tilt, roll, and twist — which I call them stacking or simple step parameters
  • x-displacement, y-displacement, helical rise, inclination, tip, and helical twist — which I call them helical parameters
3DNA calculates all of these parameters. Over the years, I have been approached quite a few times on the local helical parameters interpretation question. In literature, I've noticed numerous times of confusions the community still has over the two set of parameters, especially with regard to the issue of "twist vs helical twist" and "rise vs. helical rise". This series of blog posts is aimed to clarify the problem by providing some background information and step-by-step worked examples illustrating how the parameters are actually calculated in 3DNA (which follows SCHNAaP/SCHNArP). In my experience, knowing how the parameters are derived is the key to understand what they mean, and thus to avoid misinterpretations.

Part 1 — background information

  1. The Calladine and El-Hassan Scheme (CEHS) calculates only the six step parameters, i.e., shift, slide, rise, tilt, roll, and twist. The original CEHS implementation was a few FORTRAN subroutines taking advantage the code base of Dickerson's well-known NewHelix program. In developing SCHNAaP/SCHNArP, a collaborative project between the then Sheffield and Cambridge groups, I also went through the source code of NewHelix to get a thorough knowledge of its internals in order to integrate its nice features into SCHNAaP.

    One thing I noticed was x-displacement, a parameter characterizing the "hole" of A-DNA in top view. The global helical axis from NewHelix was (still is) appealing to me, especially for a short, relatively straight fragment. Even for a clearly non-straight duplex, e.g., in super helical nucleosome core particle DNA, or a severely kinked DNA, the deviation from regular linear helix serves as a good parameter to quantify its overall non-linearity. As a side note, a single so-called bending angle is often misleading — for example, the bending angle as commonly cited in literature (mostly calculated from Curves), is strongly influenced by the two terminal base-pairs.

    So I devised a parallel set of global helical parameters following the CEHS scheme of angle combination: here instead of roll-tilt, I used tip-inclination. The definition of global helical axis and the point the helix passes through are based on my simplified implementation in ANSI C of the algorithms used by NewHelix.

    Thus SCHNAaP calculates both a set of local CEHS step parameters, and a set of global helical parameters — a unique combination of CEHS and NewHelix, made possible only after a thorough understanding of both methods. SCHNArP was developed to complete the circle, i.e., to rebuild a structure given a set of parameters, either the local step parameters: shift, slide, rise, tilt, roll, and twist, or the global helical parameters: x-displacement, y-displacement, helical rise, inclination, tip, and helical twist.

  2. The RNA (Running Nucleic Acids) program by Babcock et al. calculates a set of local helical parameters. While working at Rutgers in Dr. Olson's group, I was interested in knowing how the local helical axis and the point its passes through were defined in the RNA program, since it would naturally substitute for the global one in SCHNAaP/SCHNArP to make two sets of purely local parameters. As it turned out, however, the algorithm was finally implemented in the 3DNA software package, as summarized in the email I sent out before the 13th Conversation at Albany:
    The way to calculate the local helical axis and helical parameters in 3DNA is *essentially the same* as in RNA, but expressed in a much simpler way. First, the local helical axis is calculated as dx-times-dy, following Bansal, which gives the *same* result as the "single rotation axis" detailed in the RNA paper. I still could not figure out WHY, but verified this numerically. The location where the helix passes through is based directly on the RNA paper (i.e., following Chasles's theorem). The procedure to calculate helical parameters, i.e., tip/inclination/x-disp/y-disp/etc, following the SCHNAaP paper (i.e, with tip-inclination combination, which is consistent with the roll-tilt, propeller-buckle combinations in 3DNA). What amazed us is that this much simpler tip-inclination implementation in 3DNA gives *exactly* the same numerical values as the original RNA algorithm.
  3. In summary, the two sets of local parameters as implemented in 3DNA are based on a unique combination of nice features from several programs well-known in the community, including SCHNAaP/SCHNArP (CEHS), NewHelix, RNA, and NUPARM. However, the 3DNA calculated parameters are numerically different from any of them, partially because 3DNA adopts the standard base reference frame. To the best of my knowledge, 3DNA is the only software that allows for rigorous conversion between the two sets of local parameters — a utility program in 3DNA called step_hel illustrates this simple fact. Again, as I wrote before the 13th Conversation at Albany:
    Take each base-pair as a rigid block, only 6 parameters are required to relate one bp to the other. Two sets are commonly used: one set is "shift, slide, rise, tilt, roll and twist", and the other set is "x-displacement, y-displacement, helical rise, inclination, tip, and helical twist". Obviously, these two sets of parameters should be directly convertible, as demonstrated by Calladine and Drew in their 1984 JMB B-to-A transition paper, and illustrated in 3DNA manuscript.
  4. Links to other programs referred to in this section: