Biomedical Engineering Reference
In-Depth Information
4.3.4.1 Advantages
PCS experiments are amenable to perdeuterated protein structures and thusto
high-molecular-weight systems with slow rotational tumbling.
4.3.4.2
Disadvantages
PCS are methodologically challenging since a spin-label has to be attached,
and new chemical shifts assigned. Similar to RDC data their use in de novo
structure determination is challenging since they are not instructive unless
reasonable starting models can be produced to fit tensor parameters.
4.3.4.3
The Bottom Line
PCS data are a promising tool for structure determination of high-molecular-
weight systems.
76,127
Although their use for structure determination was
proposed nearly two decades ago,
75,77
only a handful of structures have been
solved or refined with PCS data which reflects the technological difficulties
involved on the experimental as well as on the structure-calculation side.
4.3.5
Chemical Shift
The chemical shift is the effect of the local magnetic and electronic
environment on the resonance frequency of a spin and is thus a sensitive
probe of local structure. Structural fluctuations on the microsecond timescale
are averaged out which leads to sharp resonance frequencies in the absence of
motion on the millisecond timescale (exchange broadening). Assigned chemical
shifts are a necessary prerequisite for all other NMR experiments that yield
structural restraints and are therefore the easiest NMR parameter to obtain.
The relationship between local structure and chemical shifts is complex
and thus challenging to compute. Quantum mechanical (QM) methods
(SHIFTS
78,79
and CheShift
80,81
) have been explored, but are computationally
rather expensive. A major hurdle for QM-based methods is the motional
averaging of chemical shifts on the millisecond timescale. To properly capture
this effect, QM calculations have to be carried out on structural ensembles that
reflect motion on the microsecond timescale.
79,82
Accordingly, the currently
most successful methods are empirically trained on large datasets of high-
resolution X-ray structures whose chemical shifts are available experimentally
(SHIFTX,
83
SHIFTX2,
84
PROSHIFT,
85
CamShift,
86
SPARTA+,
87,88
and
shAIC
89
). Motional averaging is thus modelled implicitly. For methyl chemical
shifts the existing data yields less good statistics than for backbone chemical
shifts, yet promising results were obtained.
90
The main contributions to the backbone chemical shifts N, H
N
,C9,C
a
,C
b
,
H
a
are torsion angles at residue i and neighbouring residues i 21, i + 1,
hydrogen bonds for H
N
,H
a
,N,C9 and ring currents for H
a
and H.
N87
This is
