Yes, that's true: coiled-coils are a nightmare, especially for molecular
replacement! Apart from potential twinning problems, the internal
symmetry and often very tight packing makes it extremely awful for MR
replacement trials. I would recommend to create at least seven models
for molecular replacement, each shifted by one residue along the
coiled-coil helix axis, because of the seven-residue-long
heptad-repeats. Even if there is no overall-bend of the coiled-coil
helix, it will still be difficult.
Anyway, good luck!
Dirk.
Am 26.01.10 12:41, schrieb Phil Evans:
I would guess that coiled-coil structures might be difficult to solve by MR
because of multiple false solutions in a repetitive structure.
There's a lot to be said for experimental phases
Phil
On 25 Jan 2010, at 17:50, Michele Lunelli wrote:
Dear all,
I am trying to solve a structure at 2.05 A resolution by molecular replacement.
The space group
seems to be P21, with unit cell dimension 52.63, 29.43, 104.970 and beta =
95.60.
Only one copy of the protein should be present in the asymmetric unit, with 58%
of solvent content.
The search model used for MR is a truncated construct of the same protein,
comprising more that 60%
of the residues. However, no convincing MR solution is found (I used phaser,
molrep, epmr and also
mr.bump). No solutions refine to R and Rfree lower than 51-52%.
The CCP4 documentation about twinning states that "Monoclinic with na + nc ~ a
or na + nc ~ c can be
twinned". This is not clear to me, but I have c = 2a, and therefore n = 2/3.
Nevertheless all the tests run by ctruncate (and sfcheck) exclude twinning. The
observed cumulative
distribution for |L| almost overlap the expected untwinned, and the observed
cumulative intensity
distribution is not sigmoidal at all (actually it is growing faster that the
theoretical). Also the
acentric and centric moments exclude twinning, for example the acentric:
<E> = 0.858 (Expected value = 0.886, Perfect Twin = 0.94)
<E**3> = 1.442 (Expected value = 1.329, Perfect Twin = 1.175)
<E**4> = 2.438 (Expected value = 2, Perfect Twin = 1.5)
Both ctruncate and sfcheck found a pseudo-translation vector:
ctruncate (0.050, 0.000, 0.957), ratio 0.23
sfcheck (0.954, 0.000, 0.040), ratio 0.218
However a second copy cannot be present in the asymmetric unit (there would be
16% of solvent
content). Since the protein is expected to form a coiled-coil, I think that the
detected
pseudo-translation arises from the helices.
Alternatively, it is possible that the space group is wrong? And if so, how can
I figure out the
correct one?
Thank you in advance,
Michele
--
*******************************************************
Dirk Kostrewa
Gene Center, A 5.07
Ludwig-Maximilians-University
Feodor-Lynen-Str. 25
81377 Munich
Germany
Phone: +49-89-2180-76845
Fax: +49-89-2180-76999
E-mail: [email protected]
WWW: www.genzentrum.lmu.de
*******************************************************
