not. Protein folding is then observed with the dispersion of the
resonance peaks, reflecting the structure of the protein (large dispersion ¼ stable folding), on 1D proton spectra or on
two-dimensional (2D)
1
H15
N correlation spectra if the protein is
15 N-labeled.
In the 1D
1
H spectrum of a well-folded protein, the peaks are
narrow and sharp and distributed over a large range of chemical
shifts meaning a good signal dispersion; signals can be especially
found at
1 H resonance values lower than 0.5 ppm or higher than
8.5 ppm corresponding to high field-shifted methyl group protons
and downfield-shifted amide protons respectively. In contrast, the
peaks are broader and not as widely dispersed in the spectrum of an
unfolded or partially folded protein [44].
For example, in a fast 1D NMR spectrum of MAST2-PDZ
(Fig. 6a), we observed sharp and narrow peaks that cover a large
range of chemical shifts, that is, peaks in the negative ppm range
between 0 and -1 ppm corresponding to upfield methyl protons
located in the hydrophobic core of the domain, and peaks out near
10 ppm corresponding to downfield backbone amide protons. This
is a good indication that MAST2-PDZ is folded. This is a rather fast
and qualitative technique to assess protein folding. Also, all the
protons of the sample are seen in the 1D spectrum. In the past
decades, many methods have been developed to reduce the water
signal around 4.7 ppm; however, signals from protons from the
buffer can drastically complicate the analysis of spectra. For example, the 1D NMR spectrum of MAST2-PDZ (Fig. 6a) displays
intense Tris peaks around 3 ppm, but the peaks of the protein are
still visible.
The 2D
1 H15
N correlation HSQC spectrum [45] provides the
“fingerprint” of a protein as the dispersion of cross-peaks is unique
to each protein folding and sequence. On the contrary, unfolded
proteins have all similar easily recognizable restrained patterns with
a very limited dispersion of cross peaks. In addition, the peak width
is related to the molecular mass of the molecule. We except for
MAST2-PDZ quite narrow peaks consistent with the small size of
the domain (Fig. 6b). Unexpected large peaks can be observed for
autoassociated domains or/and for proteins in conformational
exchange. For example, in the 2D spectrum of the free MAST2PDZ, we do not observe the expected number of peaks and the
peaks are broad due to an autoassociation and an exchange in the
low-to-medium NMR timescales between monomers and dimers
while upon binding of the ligand many additional peaks (in red)
appear and peaks are sharper when MAST2-PDZ recovers its
monomeric form [8, 36]. In conclusion, NMR spectrum is highly
sensitive to the folded state, oligomerization and aggregation state
of a protein, and is thus efficient for the quality control of protein
production [46].
118
Ce ´ lia Caillet-Saguy et al.
resonance peaks, reflecting the structure of the protein (large dispersion ¼ stable folding), on 1D proton spectra or on
two-dimensional (2D)
1
H15
N correlation spectra if the protein is
15 N-labeled.
In the 1D
1
H spectrum of a well-folded protein, the peaks are
narrow and sharp and distributed over a large range of chemical
shifts meaning a good signal dispersion; signals can be especially
found at
1 H resonance values lower than 0.5 ppm or higher than
8.5 ppm corresponding to high field-shifted methyl group protons
and downfield-shifted amide protons respectively. In contrast, the
peaks are broader and not as widely dispersed in the spectrum of an
unfolded or partially folded protein [44].
For example, in a fast 1D NMR spectrum of MAST2-PDZ
(Fig. 6a), we observed sharp and narrow peaks that cover a large
range of chemical shifts, that is, peaks in the negative ppm range
between 0 and -1 ppm corresponding to upfield methyl protons
located in the hydrophobic core of the domain, and peaks out near
10 ppm corresponding to downfield backbone amide protons. This
is a good indication that MAST2-PDZ is folded. This is a rather fast
and qualitative technique to assess protein folding. Also, all the
protons of the sample are seen in the 1D spectrum. In the past
decades, many methods have been developed to reduce the water
signal around 4.7 ppm; however, signals from protons from the
buffer can drastically complicate the analysis of spectra. For example, the 1D NMR spectrum of MAST2-PDZ (Fig. 6a) displays
intense Tris peaks around 3 ppm, but the peaks of the protein are
still visible.
The 2D
1 H15
N correlation HSQC spectrum [45] provides the
“fingerprint” of a protein as the dispersion of cross-peaks is unique
to each protein folding and sequence. On the contrary, unfolded
proteins have all similar easily recognizable restrained patterns with
a very limited dispersion of cross peaks. In addition, the peak width
is related to the molecular mass of the molecule. We except for
MAST2-PDZ quite narrow peaks consistent with the small size of
the domain (Fig. 6b). Unexpected large peaks can be observed for
autoassociated domains or/and for proteins in conformational
exchange. For example, in the 2D spectrum of the free MAST2PDZ, we do not observe the expected number of peaks and the
peaks are broad due to an autoassociation and an exchange in the
low-to-medium NMR timescales between monomers and dimers
while upon binding of the ligand many additional peaks (in red)
appear and peaks are sharper when MAST2-PDZ recovers its
monomeric form [8, 36]. In conclusion, NMR spectrum is highly
sensitive to the folded state, oligomerization and aggregation state
of a protein, and is thus efficient for the quality control of protein
production [46].
118
Ce ´ lia Caillet-Saguy et al.