Biomedical Engineering Reference
In-Depth Information
1.3.1.2 'Domain Parsing' of Multi-Domain Proteins
Many proteins are composed of individually folding domains or regions
which can be either structurally (and/or functionally) independent, or which
can interact to form a modular pseudo-quaternary structure. In the first
case, the individual functional or structural units can be produced and
studied independently.
46
In the latter example, with a suitable system it can
be possible to breakdown the full-length protein into tractable pieces that
can be subjected to traditional NMR assignment procedures. The resonance
assignments obtained of the smaller fragments are then transferred to the
full-length protein. This approach was applied to the 204 kDa homodimeric
protein, SecA (2 6 901 amino acids; ref. 29). Three constructs of SecA with
increasing length were prepared with combinations of [d
1
-
13
CH
3
]isoleucine-,
[
13
CH
3
,
12
CD
3
]-leucine/valine- and [e-
13
CH
3
]methionine-specific labelling.
Two-dimensional (
1
H,
13
C) HMQC spectra of each construct were recorded
and compared with the full-length protein. Methyl group assignments
obtained from smaller constructs were transferred upwards to allow almost
complete methyl-group assignment of the full-length homodimer. The three
truncated constructs were monomeric as they lacked a C-terminal region
that mediated dimerisation and were thus more suitable for standard
backbone- and side-chain-based assignment procedures. It is interestingto
note that while good quality (
1
H,
15
N) correlation spectra could only be
recorded of the shorter monomeric constructs, equivalent spectra of the full-
length dimer revealed fewer than 10% of the expected peaks. In contrast,
high quality, interpretable (
1
H,
13
C) correlation spectra of methyl-labelled
samples could be recorded, even in a 204 kDa homodimer protein with 228
accessible methyl probes.
1.3.2 Resonance Assignment by Mutagenesis
The isotope-labelling protocols described in Section 1.2.2 allow methyl groups
to be protonated in a perdeuterated background in a residue-specific way with
essentially no isotopic scrambling. By this measure, mutating an isotope-
labelled residue (e.g., an isoleucine) to an isopolar one that is not labelled (e.g.,
a valine) would cause, in the simplest case, the loss of single methyl resonance
(Figure 1.6). Comparison of 2D (
1
H,
13
C) correlation spectra of wild-type and
mutant proteins would reveal a single missing peak that could then be assigned
to the mutated residue.
The practice of assignment-by-mutagenesis has been used since the dawn of
protein NMR spectroscopy. More recently this technique has been imple-
mented to aid or conduct methyl group assignment in large proteins. The
assignment examples described in Sections 1.3.1.1 and 1.3.1.2 benefited from
proteins that could withstand changes in oligomeric state or protein chain
length. This level of tolerance to truncations or mutations, which is necessary
to start the assignment process, is not always possible. The NMR data used in
an
(
1
H,
13
C)
assignment-by-mutagenesis
strategy
is
a
collection
of
2D
