With a pseudo translation vector like that the SG could be any of the 8
orthorhombic SGs; P222 P21 22 P21212 P212121 P2 21 2 P2 21 21 P 2 2 21
Test them all, and see if any give a dect solution..
Eleanor
Alison Li wrote:
We recently collected a complete 2.5A MAD dataset. However, finding a
solution has not been as straightfoward for reasons unclear to us. We would
be grateful for any helpful advice or suggestions.
The thin plate shaped crystal was grown from a relatively small protein (90
residues).
The crystal diffracted well with visible and defined reflections up to ~2.8A
range.
With both HKL2000 and mosflm, initial indexing indicated orthorhombic unit cell
with dimension of 52 x 82 x 100.
Systematic absences along a and b axis were observed thus the dataset was
scaled in P21212 space group.
The unit cell dimension and space group suggested 4 protein chains per ASU.
There is a pseudo-translation with 55 % peak at a fractional coordinate of
0.5, 0.5, 0.48.
There are 7 methionines per chain. Thus we expect 28 Se per ASU. Mass
spectroscopy and fluoresence scan both confirmed successful incorporation of
Se-methionine in the crystal.
According to xtriage, anomalous signal extended to 3.0A at least in the peak
dataset (The table of measurability as a function of resolution is shown
below).
unused: - 43.0868 [ 0/5 ]
bin 1: 43.0868 - 5.3953 [2751/2902] 0.5838
bin 2: 5.3953 - 4.2836 [2893/2905] 0.4575
bin 3: 4.2836 - 3.7425 [2891/2899] 0.2820
bin 4: 3.7425 - 3.4005 [2888/2896] 0.1898
bin 5: 3.4005 - 3.1568 [2864/2898] 0.1222
bin 6: 3.1568 - 2.9708 [2835/2898] 0.0653
bin 7: 2.9708 - 2.8220 [2777/2906] 0.0458
bin 8: 2.8220 - 2.6992 [2714/2902] 0.0226
bin 9: 2.6992 - 2.5953 [2550/2865] 0.0168
bin 10: 2.5953 - 2.5057 [2307/2914] 0.0071
unused: 2.5057 - [ 0/0 ]
However, HA search using hkl2map, Autosol, and Autosharp resulted in only
3~4 HA sites. When hkl2map was used, most HA sites had poor CC and
Patterson FOM and did not clearly stand out as they normally should.
Structural homologs suggest that the protein has a compact single core
domain comprised of 4 a-helices. The positions of most HAs are unlikely to be
located in the flexible region.
If any abnormalies are seen with the dataset, it's during scaling step in
HKL2000.
Chi2 is unusually high at lower resolution (Chi2 is >3 from 3.5A as shown
below) and there is a relatively high percentage of rejections (>1.5 %).
Shell Lower Upper Average Average Norm. Linear Square
limit Angstrom I error stat. Chi**2 R-fac R-fac
50.00 5.38 4299.0 77.3 49.0 7.886 0.076 0.085
5.38 4.27 2938.7 52.7 35.0 5.843 0.083 0.090
4.27 3.73 2314.2 45.9 33.5 3.935 0.082 0.084
3.73 3.39 1245.3 34.4 28.9 2.838 0.101 0.094
3.39 3.15 658.6 28.1 25.9 1.957 0.132 0.127
3.15 2.96 451.9 27.1 25.7 1.322 0.157 0.138
2.96 2.82 307.2 27.1 26.3 1.001 0.201 0.169
2.82 2.69 253.1 28.8 28.1 0.866 0.225 0.193
2.69 2.59 199.3 31.5 31.0 0.801 0.262 0.233
2.59 2.50 159.5 34.7 34.4 0.688 0.292 0.261
All reflections 1312.9 39.0 31.9 2.748 0.100 0.089
Xtriage also complains that there are abnormal intensities at some resolution
ranges.
Finally, the crystallization requires CdSO4, so we suspect that cadmium ions
are incorporated in the crystal. If so, we suspect there may be weak
anomalous signal contribution from Cd as well.
In summary, we appear to have a complete dataset that shows strong
anomalous signal. However it appears that we have overlooked something or
there is an unusual crystallographic issue that we are not aware of. Any
suggestions will be very much appreciated.
Alison Li
Graudate student, Dr. Mark Paetzel's group
Simon Fraser University, BC, Canada
