I think that you are overlooking at the fact that you do not know what the hydrogen bond lengths are between the base pairs.

You should make sure that the covalent geometry is correct (tighten or loosen the matrix as you do!) and then when you

have a well refined structure, analyze your DNA for deviations from 'ideality' eg ideal A or B DNA.

Longer-than-text-book hydrogen bonds, as well as the often ignored pucker amplitudes of your sugars, carry valuable

information about DNA conformation, likely showing:

a. the strain due of your bound protein (I am assuming here you have protein bound to it).

b. the properties of the particular sequence (assuming no protein bound).

The B-DNA-like sugar puckers, must be a very common mistake in the PDB, especially for bend DNA, legacy of an 'energy term'

causing the sugar puckers to be more tightly restrained to B-DNA than covalent bonds to Engh-Huber values, in (older versions?) of CNS.

That, together with retraining the hydrogen bonds in DNA as far as I recall, made DNA look good, but did not necessarily made it more correct.

Good old PROTIN/PROLSQ had an even worse bug ignoring chiral centers, but then not many people refined DNA with PROTIN/PROLSQ.

Refmac, Phenix and Buster seem to handle DNA geometry right (for my taste, eg they leave puckers and h bonds unrestrained)

(I am not sure what CNS does these days).

Sorry for the more general review. The short answer is that I would leave the H-bonds alone, make sure the rest is Ok,

and see what is the underlying reason they are longer than usual: this is likely to be informative at the end.

On Aug 5, 2009, at 11:52, Kristof Van Hecke wrote:

Dear,

When (trying) to refine a DNA-structure (resolution 2.5) using Refmac_5.5.0072 (CCP4 6.1.1), some of the H-bonds between Watson-Crick bases are becoming too large.

Reducing the Matrix weighting term to tighten the geometry, doesn't effect these H-bond distances much.
Reducing the "VDW SIGMA HBOND" also doesn't solve the problem.

Adding "external distance restraints" does the trick, but some B-factors (not the ones involved in H-bonding!) blow up completely.

Hence, what's actually the best way of tighten H-bond restraints in Refmac, or am I overlooking some other issues here..?

Thank you very much.

Regards

Kristof Van Hecke

--------------------------------------
Kristof Van Hecke, PhD
Biomoleculaire Architectuur
Celestijnenlaan 200 F
B-3001 Heverlee (Leuven)
Tel: +32(0)16327477
--------------------------------------

P please don't print this e-mail unless you really need to

Anastassis (Tassos) Perrakis, Principal Investigator / Staff Member

Department of Biochemistry (B8)

Netherlands Cancer Institute,

Dept. B8, 1066 CX Amsterdam, The Netherlands

Tel: +31 20 512 1951 Fax: +31 20 512 1954 Mobile / SMS: +31 6 28 597791