472 Fluorine in Medicinal Chemistry and Chemical Biology
fi t can only occur for one structure and will thus give an indication that this is the actual
structure inside the membrane.
In principle, the peptide orientation is fully defi ned by the tilt and azimuthal angles
described in Figure 18.3 . However, peptides interacting with a membrane are mobile, and
to account for this motion a simplifi ed order parameter, S
mol
, is introduced, which has the
effect of scaling all the calculated splittings by a factor between 0 and 1 [4, 23, 44 – 47,
49 – 51] . S
mol
= 0 would correspond to complete isotropic averaging, where all the orienta-
tional information would be lost, and S
mol
= 1 corresponds to a completely immobile
peptide. The peptides in our studies usually have S
mol
values between 0.6 and 0.8. This
gives information about the mobility of the peptides and can also be used to estimate the
size of aggregates. More elaborate motional models have recently been investigated (E.
Strandberg et al. , submitted), but have shown that the calculated values of tilt and azi-
muthal angles for the systems investigated were virtually identical to those obtained using
the more simple approach described here. For our peptides, we thus determine the tilt
angle, the azimuthal angle, and S
mol
. In principle, three constraints are needed to determine
these three parameters, each constraint being obtained from a
19
F NMR measurement on
a singly
19
F - labeled peptide. In practice, at least four
19
F - labeled peptides are used in order
to get a more reliable result, to make sure all the data are consistent, and to rule out any
possible structural perturbations due to introduction of
19
F - labeled amino acids.
After measuring several local dipolar splittings, these data are used to calculate the global
orientation of the full peptide. From the known structure of the peptide, theoretical curves show
which dipolar couplings are expected for different labeled positions, depending on tilt angle τ ,
azimuthal angle ρ , and S
mol
. The calculated values of splittings are then compared with the
experimentally obtained values, and the root mean square deviation (RMSD) is calculated.
An example of how the procedure works is shown in Figure 18.5 . The peptide in this
example is MSI - 103, which has been labeled with
19
F at the four positions indicated in
Figure 18.5 a. Four different peptides were synthesized where a single amino acid (Ala - 7,
Ile - 9, Ala - 10, or Ile - 13) was replaced with CF
3
- Phg. For each label a
19
F NMR spectrum
is recorded (see Figure 18.5 a); in this example, spectra are from MSI - 103 in DMPC (1,2 -
dimyristoyl - sn - glycero - 3 - phosphocholine) at a peptide - to - lipid molar ratio (P/L) = 1 : 400,
Figure 18.5 (a) The peptide MSI - 103 was labeled with CF
3
- Phg at four positions, marked in
lighter gray. For each label a
19
F NMR spectrum was recorded in DMPC at P/L = 1 : 400. From
these spectra the dipolar couplings were measured, giving the values shown next to each
spectrum. (b) The measured dipolar couplings are compared with theoretical curves for
different orientations of the peptide. The best - fi t curve is here shown together with experimental
data for the different labeled positions (fi lled squares). (c) The RMSD plot shows the root
mean square deviation between experimental and calculated splittings, for all possible
combinations of tilt and azimuthal angles. In this case, the best - fi t tilt angle ( τ ) is 101 ° and
the azimuthal angle ( ρ ) is 130 ° . (d) Side view of the peptide in the membrane, which is
represented by a gray box. The tilt angle defi nes the angle between the peptide long axis and
the membrane normal. (e) View of the peptide along the helical axis. The azimuthal angle
defi nes how much the peptide is rotated around its axis, with the starting point defi ned as
the vector from the helical axis to C
α
of residue Lys - 12 being parallel with the bilayer
surface.