116 INSTRUMENTAL METHODS
relaxation can also produce an exchange of magnetization through the NOE
effect. In either case, nuclei that are close together in space may produce cross
peaks in the NOESY experiment that, along with additional structural infor-
mation (see COSY and TOCSY in Section 3.4.9 ), leads to three - dimensional
structure determination of large molecules in solution.
3.4.11 Multidimensional NMR,
Over the last 10 – 15 years, multidimensional (3D and 4D) NMR spectroscopic
and computational techniques have been developed to solve structures of
larger proteins in solution. This technology has the advantage of resolving the
severe overlap in 2D NMR spectra for proteins having > 100 residues. Usually,
multidimensional NMR experiments require isotopic labeling of some or all
nuclei of interest. For example, signal overlap in a two - dimensional^1 H^1 H cor-
relation spectrum can be resolved in the dimension of the heteronucleus — that
is,^15 N or^13 C attached to the protons of relevance. If fast relaxation is a problem,
all nonexchangeable protons can be deuterated as well, leading to sharper
signals. In a year 2000 publication by the Kurt W ü thrich group, one fi nds this
example.^22 Transverse relaxation - optimized spectroscopy (TROSY) - type
triple resonance NMR experiments were used to determine the secondary
structure of an octameric 110 - kDa protein. The fi nal structure determined
from the uniformly^2 H,^13 C,^15 N - labeled 7,8 - dihydroneopterinaldolase (DHNA),
a lyase, from Staphylococcus aureus , showed 20 - fold to 50 - fold sensitivity
gains when compared to the corresponding conventional triple resonance
NMR experiments. Complete sequence - specifi c assignments of the 121 - residue
(13.7 - kDa) polypeptide chain in the 110 - kDa octamer were obtained in aqueous
solution at 20 ° C. Additionally, the secondary structures determined in the solu-
tion by NMR were found to coincide nearly identically with those in the crystal
structure of the DHNA octamer. It is important to note that secondary and
tertiary structures of monomeric or multicomponent enzymes in the 110 - kDa,
∼ 900 - amino acid size range have not been accomplished using multidimen-
sional solution-NMR to date in 2007. Portions of these larger molecules have
been studied by solution NMR, and examples will be found in Chapters 5 , 6,
and 7.
Currently the upper limit of applicability of multidimensional NMR
methods may be for monomeric proteins up to 50 kDa, 250 – 300 amino acid
residues, but this may increase with future technological advances. Using these
methods, one is able to resolve protein structure at the same level as X - ray
crystallographic data that has been resolved to approximately 2.5 Å. While it
may be true that a carefully refi ned X - ray structure of a given protein (a solid -
state picture) may not be identical to the “ physiologically true ” solution struc-
ture determined by NMR, in most cases where structures from both methods
are available, strong agreement in secondary and tertiary structure has been
found. Evidence reinforcing the comparison comes from calculations of three
bond coupling (^3 J Hα HN ) constants from well - refi ned crystallographic data that