ChemInform Abstract: Principles and Progress in Ultrafast Multidimensional Nuclear Magnetic Resonance

Article (PDF Available)inAnnual Review of Physical Chemistry 60(1):429-48 · May 2008with14 Reads
DOI: 10.1146/annurev.physchem.040808.090420 · Source: PubMed
Abstract
Multidimensional acquisitions play a central role in the progress and applications of nuclear magnetic resonance (NMR) spectroscopy. Such experiments have been collected traditionally as an array of one-dimensional scans, with suitably incremented delay parameters that encode along independent temporal domains the nD spectral distribution being sought. During the past few years, an ultrafast approach to nD NMR has been introduced that is capable of delivering any type of multidimensional spectrum in a single transient. This method operates by departing from the canonical nD NMR scheme and by replacing its temporal encoding with a series of spatial manipulations derived from magnetic resonance imaging. The present survey introduces the main principles of this subsecond approach to spectroscopy, focusing on the applications that have hitherto been demonstrated for single-scan two-dimensional NMR in different areas of chemistry.
ANRV373-PC60-21 ARI 25 February 2009 18:3
Principles and Progress in
Ultrafast Multidimensional
Nuclear Magnetic Resonance
Mor Mishkovsky and Lucio Frydman
Department of Chemical Physics, Weizmann Institute of Science, 76100 Rehovot, Israel;
email: Lucio.Frydman@weizmann.ac.il
Annu. Rev. Phys. Chem. 2009. 60:429–48
First published online as a Review in Advance on
December 1, 2008
The Annual Review of Physical Chemistry is online at
physchem.annualreviews.org
This article’s doi:
10.1146/annurev.physchem.040808.090420
Copyright
c
2009 by Annual Reviews.
All rights reserved
0066-426X/09/0505-0429$20.00
Key Words
NMR spectroscopy, multidimensional acquisitions, ultrafast methods,
spatial encoding, hyperpolarized samples
Abstract
Multidimensional acquisitions play a central role in the progress and applica-
tions of nuclear magnetic resonance (NMR) spectroscopy. Such experiments
have been collected traditionally as an array of one-dimensional scans, with
suitably incremented delay parameters that encode along independent tem-
poral domains the nD spectral distribution being sought. During the past few
years, an ultrafast approach to nD NMR has been introduced that is capable
of delivering any type of multidimensional spectrum in a single transient.
This method operates by departing from the canonical nD NMR scheme
and by replacing its temporal encoding with a series of spatial manipulations
derived from magnetic resonance imaging. The present survey introduces
the main principles of this subsecond approach to spectroscopy, focusing
on the applications that have hitherto been demonstrated for single-scan
two-dimensional NMR in different areas of chemistry.
429
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
MRI: magnetic
resonance imaging
Fourier transform
(FT): the
mathematical
procedure that enables
one to discern which
frequency
contributions compose
a time-dependent
signal response
function
Free induction
decay: traditional
term used to denote
time-domain signals of
the kind described in
Equation 1 and arising
upon exciting a spin
ensemble in an NMR
experiment
T
2
: spin-spin
relaxation time,
describing the
decoherence lifetimes
of the spins’ signal
following their
excitation into a
transverse evolution
plane
1. INTRODUCTION
Nuclear magnetic resonance (NMR) is a spectroscopic technique based on monitoring the pre-
cession of spin-endowed atomic nuclei when placed in a strong external magnetic field. With its
origin in curiosity-driven investigations about the nature of quantum mechanical phenomena,
NMR has over the years transformed into an indispensable applied tool that impacts a remarkably
wide range of scientific disciplines (1). High-resolution NMR of dissolved molecules, for example,
serves as the eyes of chemists involved in organic, pharmaceutical, and natural-products research
(2). When employed in combination with high-resolution solid-state techniques, NMR provides a
unique window to study the chemical structure of challenging heterogeneous materials, including
polymorphic mixtures, polymers, glasses, and catalysts (3). It is one of the few methods avail-
able for determining the structure and the dynamics of proteins and nucleic acids in their native
solution state at individual, site-resolved levels (4). Moreover, even if concealed under different
names [e.g., magnetic resonance spectroscopy or magnetic resonance imaging (MRI)], NMR has
evolved into a widely used in vivo tool capable of diagnosing and imaging malignancies, evaluating
metabolic status, angiographing noninvasively, and revealing the activation of human brains even
to the smallest of stimuli (5, 6).
Despite this outstandingly wide scope of applications, broadly speaking, one common mea-
surement protocol underlies all these different uses of the quantum-mechanic spin-precession
phenomenon: the pulsed Fourier transform (FT) NMR method, whose purpose is to measure
inductively the different Bohr evolution frequencies allowed to the spin ensemble under obser-
vation within its quantized energy manifold (1–7). These in turn are governed by isotropic shift
or J-coupling interactions in liquid-state experiments, by these couplings plus a variety of spin
anisotropies when dealing with solids, and by a combination of the above plus external ad hoc
fields in the case of NMR imaging. Regardless of which parameters or interactions are actually
sought, NMR extracts the frequency distributions originated by these couplings by monitoring
the responses that they impart on the spins’ time evolution. The result arising in these measure-
ments upon subjecting spins to an excitation impulse is the so-called free induction decay, a signal
(voltage) given by a weighted sum of oscillating functions,
S(t) =
acting
I() exp(it) exp(t/T
2
), (1)
defined by the allowed single-quantum spectral distribution of allowed transitions I(), as well as
by a T
2
relaxation decay. S(t) can then clearly provide the I() spectrum sought by FT versus the
single variable t, defining the time domain that supports it.
Although pulsed NMR experiments initially involved only this kind of data collection, as a
function of a single time axis (8), it was soon realized that significant benefits could result by
correlating the spins’ evolution along multiple time domains (9, 10). These nD NMR experiments
would then not just measure but also correlate and separate different contributions to the overall
spin-precession frequencies; this could improve the resolution of the experiment, as well as extract
information that would be simply unavailable in the single-quantum one-dimensional (1D) trace.
The canonical scheme for such experiments follows Jeener and Ernst’s seminal 2D NMR proposal,
which laid the foundations for multidimensional spectroscopy based on the four blocks (9, 10)
Preparation Evolution (t
1
) Mixing Acquisition (t
2
). (2)
By suitable incrementation of the t
1
and t
2
time variables, an NMR experiment of this kind can
then yield a 2D time-domain signal,
s (t
1
, t
2
) =
all
2
d
2
all
1
d
1
I(
1
,
2
)e
i
1
t
1
e
t
1
/T
2
e
i
2
t
2
e
t
2
/T
2
, (3)
430 Mishkovsky
·
Frydman
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
from which correlations between so-called indirect- and direct-domain NMR frequencies
1
and
2
can be extracted by 2D FT analysis:
I(ν
1
2
)
all t
2
dt
2
all t
1
dt
1
S(t
1
, t
2
)e
iν
1
t
1
e
iν
2
t
2
. (4)
At first glance, this approach to the retrieval of I(
1
,
2
) appears to be a simple extension of the
1D time-domain NMR experiment to two dimensions. The multiple times involved in these ac-
quisitions, however, actually possess different origins. t
2
is a physical acquisition time along which
the signal is directly digitized, akin to the t time involved in the 1D NMR free induction decay
of Equation 1; the spins’ behavior along this direct-domain axis therefore can be characterized
by FT of data digitized throughout a single-scan experiment. The remaining time axes of the
experiment, however, cannot be sampled in the same fashion. This is resolved in the Jeener-Ernst
paradigm by monitoring these domains indirectly, i.e., by carrying out N
i
discrete incrementa-
tions of certain time delays t
i
within the sequence throughout a series of independent experiments.
Herein lies the brilliance, yet also a potential weakness, of this kind of acquisition. Indeed, re-
gardless of sensitivity considerations, nested encoding schemes such as the one represented by
Equation 2 require monitoring tens or hundreds of independent increments along each of the
n – 1 indirect-domain time axes to properly characterize its internal evolution frequencies. More-
over, because each point along these n 1 indirect time domains is associated with an independent
signal acquisition, this results in an exponential increase of the minimum experimental acquisition
time with dimensionality n. The unambiguous gains resulting from expanding NMR from a 1D
to an nD experiment may therefore come at a price.
Driven by this reality, and stimulated by an increasing reliance of all the above-mentioned
contemporary NMR applications on high-dimensional experiments, a growing number of alter-
natives that depart from the traditional sampling principles embodied by Equation 2 have emerged
during the past few years (11, 21). These include routes that process the acquired data by non-
Fourier methods (13–16), acquisitions that incorporate frequency-based manipulations (17), and
accordioned derivations (18) whereby multiple indirect-domain time delays are incremented si-
multaneously (19–21). Among these new proposals is also an ultrafast approach, departing both
from the traditional temporal encoding and from the alternatives just mentioned, that enables the
acquisition of arbitrary multidimensional data sets within a single scan without requiring any a pri-
ori information (22–24). At the heart of this proposal is a departure from the classical Jeener-Ernst
serial incrementation scheme, which is replaced by a parallel encoding of the indirect-domain time
information along a spatial dimension. This review discusses the basic features and potential ap-
plications of this ultrafast NMR method, particularly as it is used within the context of rapid 2D
NMR acquisitions. We summarize first the physical principles underlying ultrafast 2D NMR,
continue with an overview of some potential applications of this method, discuss its combination
with a variety of schemes enabling its extension to higher dimensions, and conclude by review-
ing additional applications that have been recently demonstrated based on the spatial-encoding
concepts on which ultrafast NMR spectroscopy relies.
2. SPATIAL ENCODING AND THE SINGLE-SCAN ACQUISITION
OF 2D NUCLEAR MAGNETIC RESONANCE SPECTRA
Ultrafast 2D NMR methods depart from traditional schemes in that, instead of triggering the
indirect-domain evolutions homogenously, for all sites and positions within the analyzed sample
at once, they encode the evolution frequencies being sought in a spatially heterogeneous fashion.
Several alternatives have been proposed for achieving such heterogeneous evolution (22, 25–30),
www.annualreviews.org
Ultrafast Multidimensional NMR 431
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
a majority of which impart different evolution durations prior to the mixing period on spins
positioned at different coordinates within the sample. All these schemes share the application of
a train of either continuous or discrete frequency-swept radiofrequency (RF) pulses, acting in
combination with suitably echoed magnetic field gradients. These procedures are tuned so as to
impart indirect-domain evolution times proportional to the spins’ coordinate along a particular
direction z. In other words, they rely on a spatial encoding of the spin interactions, whereby
the indirect-domain evolution time of a 2D NMR experiment is made t
1
Cz, where C is a
spatio-temporal constant under the experimentalist’s control. (This C parameter is usually given
by
t
max
1
L
, the maximum t
1
evolution time imparted, divided by the overall sample length L.) The
evolution coupling parameters
1
being sought thus impart over these periods an effective spatially
dependent precession phase φ(z)
1
t
1
= C
1
z. It follows that within the framework of a Bloch
space in which the spins’ magnetizations precess within a common rotating frame (Figure 1)
every chemical site within the sample subtends a shift-induced helical pattern along the z spatial
coordinate:
M
x
(z) + iM
y
(z)
I(
1
)
L
exp(iC
1
z) exp(Cz/T
2
). (5)
The helical winding of magnetizations represented by this equation is characterized by a pitch
depending on the
1
parameter one is attempting to measure but in general does not lead to any
net observable signal as these helices are characterized by a destructive interference among their
constituent spin packets when considered over a macroscopic sample. Nevertheless, their encoded
information can be preserved throughout the various coherent mixing processes involved in a 2D
NMR pulse sequence and, at a final acquisition stage, can be read out by applying a z-dependent
gradient acting in combination with the data sampling. Indeed, field gradients have the ability to
wind and unwind spin-magnetization patterns of their own according to
M
x
γ Gz·t
−− M
x
cos(γ Gzt) + M
y
sin(γ Gzt)
= M
x
cos
(
kz
)
+ M
y
sin
(
kz
)
. (6)
Under suitable conditions, therefore, these succeed to successively unwind the various helices
that were subtended by the individual chemical sites, leading to observable echoes arising from
constructive interference phenomena among spins positioned throughout the sample L (Figure 1).
Moreover, as the timing of such echoes depends on the strengths of the
1
internal interactions that
created each site’s winding, this allows one to map the indirect-domain spectral information being
sought by monitoring the positions of the resulting echo peaks. Mathematically, these spectral
peak positions are characterized by the values taken by the wave numbers k = γ
a
t
0
G
a
(t
)dt
,
representing the action of the unwinding acquisition gradient G
a
. Assuming that this MRI-like
acquisition process begins in a spatially encoded magnetization pattern of the kind described by
Equation 5, the integral of the observable signal over the sample length is then
S[k(t)]
1
I(
1
)
L
exp[iC
1
z] exp[Cz/T
2
] exp[ik(t)z]dz
1
I(
1
)δ[C
1
+ k]. (7)
This latter spectral sum includes well-behaved δ-functions such as sinc-, Lorentzian- or Gaussian
lineshapes, leading to peaks whenever –k/C matches an existing precession frequency. This ratio
therefore becomes the equivalent, in this kind of experiment, to the indirect-domain frequency
scale ν
1
in a conventional 2D acquisition.
A procedure such as the one summarized in Figure 1 allows one to read out an NMR spectrum
with the aid of a gradient-driven action—without the use of a numerical FT. A special feature of
432 Mishkovsky
·
Frydman
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
Ω
1
Ω
2
G
a
k/ν
1
z
Spin 1
Shift Ω
1
Spin 2
Shift Ω
2
k
1,2
= –C
1,2
)
RF
ab c
G
z
z
Spatial
encoding
z
Homogeneous
sequence
+G
e
G
e
Gradient-driven
readout
Initial states in (M
x
, M
y
, M
z
) Bloch space
Spatially encoded states in (M
x
, M
y
, M
z
)
Mixing
Mixing
Mixing
Figure 1
Imparting and reading out NMR spectral data by gradient-driven processes. (Top panel ) A summary of the
pulses and gradients applied to the spins. (Bottom panel ) The idealized behavior imparted by these
manipulations on two chemically inequivalent sites, each represented by an array of magnetization vectors as
a function of their spatial z coordinates throughout the sample. (a) Spatial-encoding stage incorporating
frequency-swept pulses and suitably refocused magnetic field gradients, G
e
. The radiofrequency (RF)
achieves a sequential excitation of spins along the direction of the gradient; because these are applied in a
sign-alternating fashion, no phase related to the spatial position of spins is retained. The result is a shift-
driven winding of the magnetizations along the gradient’s direction (bottom panel ). (b) This can be followed by
conventional, homogenous sequences, for instance, those involved in arbitrary mixing processes. (c) Finally,
data are collected while in the presence of an acquisition gradient, G
a
, capable of unwinding the shift-induced
magnetization spirals encoded during the excitation. The sharp echoes that are then generated unveil an
array of peaks, delivering the NMR spectrum that acted on the spins during the stage shown in panel a.
such gradient-driven readout is that it can be implemented over a very short time, on the order of
T
a
t
max
1
2π SW
1
γ
a
G
a
L
, where SW
1
denotes the window of spectral frequencies one is seeking to charac-
terize. These gradient effects can then be reversed immediately, simply by reversing the currents
flowing through the gradient’s amplifier, and can therefore be repeated multiple times over the
course of a t
2
direct-domain acquisition time. This constitutes the second principal ingredient
enabling the completion of the 2D experiment within a single scan: By rapidly alternating the
sign of the decoding acquisition gradients, one can observe the I(
1
) spectrum repetitively and
www.annualreviews.org
Ultrafast Multidimensional NMR 433
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
a Oscillating-gradient acquisition
Time t
+G
a
Δt
2
= T
a
+G
a
G
a
G
a
N
2
t
2
k/ν
1
k/ν
1
FID(t) = S(k/ν
1
,t
2
)
FT along t
2
ν
2
b Rearrangement and processing
FID(t)
1
4
3
2
1
2
3 4
Figure 2
The extension of the single gradient–driven refocusing process illustrated in Figure 1 to a multi-echo
process capable of yielding full 2D NMR spectra within a single scan. (a) By performing multiple, rapid
± G
a
oscillations, one can read out the spatially encoded indirect-domain spectrum I(
1
) many (N
2
) times
separated by relatively short intervals t
2
= T
a
. The phase modulation then affecting the different echoes
(continuous lines) is given by their respective direct-domain evolution frequencies
2
.(b) A full 2D NMR
spectrum can therefore be obtained by rearranging the single-scan interferogram I(k/ν
1
, t
2
) represented by
this free induction decay (FID) into its proper position within a 2D place, followed by its 1D Fourier
transform (FT) as a function of t = t
2
.
thereby monitor the phase modulation of the indirect-domain frequency peaks arising as a func-
tion of a detection time t
2
. The time-domain signal S(t) arising during the course of such an
oscillating-gradient procedure thereby constitutes a 2D interferogram in the space subtended by
the ν
1
= k/C indirect-domain frequency axis and the t
2
= t direct-domain acquisition time.
Signals collected throughout this process and rearranged into their proper positions within such
mixed frequency-/time-domain space can therefore lead to the desired 2D NMR spectrum if these
echo signals are subjected to a final 1D FT process along the direct domain (Figure 2), all of this
within a single scan.
3. POTENTIAL APPLICATIONS OF ULTRAFAST 2D NUCLEAR
MAGNETIC RESONANCE
Further details underlying ultrafast NMR’s spatial encoding and ways to implement such processes
in contemporary NMR commercial hardware have been summarized recently in a comprehensive
article (31). Therefore, we turn to a discussion of some potential applications and extensions that
have been demonstrated based on this new approach to the single-transient collection of 2D NMR
spectra.
3.1. Real-Time Ultrafast 2D Nuclear Magnetic Resonance
of Samples Subject to Constant Flow
The capability of gaining structural information using 2D NMR transforms the single-scan ap-
proach described above into a potential candidate for identifying compounds and tracking chemical
separations in real time. Indeed the past decade has witnessed a growing number of applications
that use NMR spectroscopy to monitor compounds passing through its main observation coil,
both as a metabonomics diagnostic tool (32) and in combination with liquid-chromatography
procedures (33). When considering this kind of NMR observation, two main options arise: One
434 Mishkovsky
·
Frydman
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
789
7
8
9
789
7
8
9
789
7
8
9
789
7
8
9
1.85 min 3.7 min 9.25 min
1
H shift (ppm)
Real time 2D TOCSY NMR
acquisitions on a mixture
subject to continuous
chromatographic elution
1
H shift (ppm)
1
H shift (ppm)
1
H shift (ppm)
1
H shift (ppm)
O
O
0.62 min
Background Compound #1 Compound #2 Compound #3
Br
Br Br
NO
2
Figure 3
Real-time identification of mixed components subject to a continuous flow separation via 2D ultrafast
1
H
total correlation–spectroscopy NMR (35). 2D NMR spectra were constantly acquired 32 s apart over a
15-min elution period; the figure concentrates on spectra corresponding to the drawn compounds appearing
at the indicated elution times.
Total correlation
spectroscopy: a
common type of
homonuclear 2D
experiment revealing
connectivities between
neighboring sites
involves performing the measurements in a stopped-flow mode, whereby fractions are collected
separately and inserted one by one into the NMR spectrometer for individual examination; the
other involves monitoring the NMR spectra as analytes flow continuously through the observation
coil. The latter experimental setup is more amenable to analyzing the large number of samples
that one needs to monitor in either metabolic profiling or analytical separation applications, yet
this flow mode limits its applicability to the collection of relatively short 1D experiments. Ultrafast
2D analyses, conversely, could open new possibilities toward the application of multidimensional
NMR during the few seconds that samples typically spend within the NMR observation coil un-
der typical elution conditions. In this context, given sufficient spectral sensitivity, the ultrafast 2D
NMR protocol can measure homonuclear
1
H total correlation spectroscopy spectra (7, 34) of sam-
ples subjected to continuous flow through the NMR sample coil (Figure 3); further applications
of these principles are envisioned.
3.2. Following Chemical and Biophysical Transformations
by Ultrafast 2D Nuclear Magnetic Resonance
Another opportunity that may arise from rapid 2D NMR approaches concerns the possibility of
monitoring chemical and/or biophysical transformations in real time. In fact, NMR spectroscopy
offers a number of different avenues toward monitoring dynamic chemical and biophysical process
(36–39). We can gain insight into kinetics occurring on fast (10
7
–10
3
s) timescales through a
variety of relaxation time measurements, and environmental changes in the 10
2
–10
0
s range
can be revealed by lineshape variations. However, longer timescales could be probed directly by
monitoring changes in the appearance of 1D or 2D NMR spectra as a function of time, a standard
approach for examining the nonequilibrium kinetics of chemical, biophysical, and in vivo processes
(5, 38, 39). The timescales of conventional 2D NMR approaches, however, are not ideally suited
to this kind of determination owing to their intrinsic multiscan acquisition modes; by shortening
the minimum times required to complete the collection of the 2D NMR data, ultrafast methods
could help alleviate this limitation. The data in Figure 4 illustrate this ability with a series of 2D
www.annualreviews.org
Ultrafast Multidimensional NMR 435
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
678
6
7
8
678
6
7
8
41.4 s
678
6
7
8
11.5 s
4.6 s
NO
2
O
2
N
CN
+ MeO
+
i
ii
i
ii
ii
NO
2
H OMe
O
2
N
H
H
CN
NO
2
H
OMe
O
2
N
H
CN
H
1
H shift (ppm)
1
H shift (ppm)
1
H shift (ppm)
1
H shift (ppm)
(i) (ii)
Figure 4
Real-time monitoring of a chemical reaction using ultrafast 2D total correlation spectroscopy NMR (40),
illustrating the appearance of compounds (bottom) at the indicated times since triggering the reaction inside
the NMR magnet. Both reactants were prepolarized prior to the reaction: One reactant was loaded into the
5-mm NMR tube, whereas the other was held within a capillary prior to its sudden injection.
Heteronuclear
multiple-quantum
correlation: a
common 2D NMR
approach for
determining the
connectivity between
heteronuclear pairs
(e.g.,
1
H and
15
N)
NMR snapshots collected while a chemical reaction was triggered in situ within the NMR tube.
Data acquisition in this kind of experiment began slightly before triggering the chemical process,
and a series of 2D total correlation spectroscopy NMR spectra recorded every 2.3 s could reflect
even transient stages of the ensuing Messenhimer complexation reaction (40).
If and when sensitivity suffices, biophysical transformations such as H/D exchange processes
or folding events could also become amenable to this kind of measurement. In these biomolecular
cases (in which sensitivity is always a challenge), it is often convenient to couple the ultrafast
2D protocol with other acquisition methods capable of minimizing the experiment’s recycle de-
lay and thereby maximize the number of 2D NMR frames available per unit acquisition time.
One such recent proposal is the band-selective, optimized flip-angle, short-transient, heteronu-
clear multiple-quantum correlation NMR experiment (41, 42), which enables the reduction of
the interscan repetition delay down to 100 ms while preserving high sensitivity by relying on
an accelerated spin-lattice relaxation imparted from a selective excitation in combination with
optimized flip angles to enhance the steady-state signal arising from the excited spins (43–45).
This in return allows the recording of conventionally sampled 2D
1
H-
15
Nor
1
H-
13
C correlation
spectra within minimal experimental times of just several seconds. Moreover, if combined with
ultrafast 2D NMR, this methodology can yield spectra at very high frame rates, compatible with
the continuous following of biophysical processes. Figure 5 illustrates an example of these spec-
tra (46), with an H/D exchange process followed under experimental conditions that combine
the very short interscan delay afforded by the optimized flip-angle protocol, with the single-scan
capabilities of ultrafast NMR. This method is capable of collecting 2D NMR correlation spectra
at 1-mM protein concentrations with frame rates of approximately 1 Hz; this could become an
important aid in studying subsecond processes in which peak positions or intensities change owing
to protein or nucleic acid folding, binding, or dynamics.
436 Mishkovsky
·
Frydman
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
2.4 s
4.8 s
9.6 s
14.4 s
77.588.59
110
115
120
125
77.588.59
110
115
120
125
77.588.59
110
115
120
125
77.588.59
110
115
120
125
19.2 s
24 s
31.2 s
40.8 s
77.588.59
110
115
120
125
77.588.59
110
115
120
125
77.588.59
110
115
120
125
77.588.59
110
115
120
125
S65
H68
I61
I13
1
H shift (ppm)
15
N shift (ppm)
15
N shift (ppm)
15
N shift (ppm)
15
N shift (ppm)
1
H shift (ppm)
Figure 5
Representative series of real-time 2D ultrafast heteronuclear multiple-quantum correlation NMR spectra
recorded on an ubiquitin solution, using the band-selective, optimized-flip-angle, short-transient protocol,
following the dissolution of an initially fully protonated lyophilized powder onto a D
2
O-based buffer as a
function of the time delay elapsed since the dissolution (46). The repetition time between full recording was
2.4 s, and the data were monitored over a 20-min interval. Red circles indicate selected NH resonances
that rapidly disappear owing to the H D exchange.
3.3. Ultrafast 2D Nuclear Magnetic Resonance on Prepolarized Samples
In recent years, we have witnessed the emergence of a variety of NMR prepolarization methods
capable of building up nuclear polarizations that exceed their thermal counterparts by several
orders of magnitude (47–50). In many instances, these methodologies are constrained to extract
their superspectra within a single or at most a few transients, making them poor starting points
for conventional 2D NMR experiments requiring the collection of a large set of scans. Such
a scenario presents another instance in which new possibilities could arise through the use of
www.annualreviews.org
Ultrafast Multidimensional NMR 437
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
a Single-scan: dark b Single-scan: light
c-Ala
d
-Ala-Tyr-Tyr-Glu-Glu-Ala
d
-Ala] = 0.5 mM
H3/H5
1
1
2
3
4
5
6
7
234567
1
H shift (ppm)
1234567
1
H shift (ppm)
1
H shift (ppm)
Tyrosine residues
Tyrosine residues
Tyrosine residues
Figure 6
Potential benefits resulting from combining chemically induced dynamic nuclear polarization (CIDNP)
prepolarization and ultrafast 2D methods. Both panels illustrate single-scan 2D total correlation
spectroscopy
1
H NMR spectra recorded on a cyclic octapeptide dissolved in D
2
O at a 0.5-mM
concentration. (a) Spectrum under standard conditions. (b) Spectrum resulting from pre-irradiating the
sample for 0.5 s using a 480-nm light source (2W), generating a CIDNP enhancement of tyrosine’s aromatic
protons (52).
Dynamic nuclear
polarization (DNP):
a generic method to
enhance nuclear spin
polarizations by
saturating the
resonances of nearby
electrons
ultrafast 2D NMR methodologies. An early example of such an integration was the use of chem-
ically induced dynamic nuclear polarization (CIDNP), a prepolarization process involving the
light irradiation of a suitable photoexcitable molecule (51). When this irradiation is carried out
in the presence of a suitable peptide or protein sample, the generated radicals may have the ca-
pability of affecting the steady-state polarization of the nuclear spins of certain aromatic residues
(e.g., tryptophan, tyrosine, and histidine) via nuclear-electron hyperfine interactions. As usual,
when implementing NMR studies on peptides or proteins, it would be desirable to carry out
this sensitivity-enhancement procedure while spreading the affected resonances throughout a 2D
frequency spectrum. CIDNP, however, has a limited compatibility with 2D NMR owing to sig-
nificant photobleaching effects that set in after the first few light irradiation cycles. Ultrafast 2D
NMR can help complete photo-CIDNP acquisitions on such prepolarized samples within a single
scan and thus avoid such complications altogether; Figure 6 shows a preliminary step in that
direction with a single-scan 2D total correlation spectroscopy spectrum acquired at submillimolar
concentrations following a brief period of CIDNP polarization enhancement (52).
An even more promising integration combines dynamic nuclear polarization (DNP) as the mag-
netization build-up process, with heteronuclear ultrafast 2D correlation spectroscopies. DNP is
one of the earliest (53–55) and arguably one of the most general hyperpolarization mechanisms, as
it only requires the irradiation of a small amount of free radical mixed with the target molecule to
achieve its goals. Although the DNP process itself is usually carried on a frozen glass, one can now
rapidly melt and transfer such a hyperpolarized mixture into a conventional NMR spectrometer,
thereby benefiting from a very large signal enhancement also in the liquid state (56, 57). The avail-
ability of a commercial instrument capable of performing this single-shot DNP hyperpolarization
of liquid samples has caused substantial excitement in the fields of NMR and MRI (58–62). Once
again, the ultrafast approach seems well suited for making such a single-shot approach compatible
with the acquisition of 2D NMR data. Figure 7 illustrates a variety of heteronuclear ultrafast 2D
experiments collected on hyperpolarized samples, depicting some of the possibilities that could
arise from this integration of single-scan 2D NMR with DNP (63, 64).
438 Mishkovsky
·
Frydman
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
a
15
N-
1
H ultrafast HSQC
6-mM natural abundance
methyl salicylic acid
0.51.552
125
130
60-μM
15
N choline chloride
0.22-mM natural
abundance
o,m,p-xylene/toluene
200-nM
15
N urea
5.5 4.5
3.5
60
55
40
50
45
40
35
7.5 7 66.5
5.5
60
55
40
50
b
13
C-
1
H ultrafast HSQC
c
15
N-
1
H ultrafast HMBC d
13
C-
1
H ultrafast HMBC
1
H shift (ppm)
1
H shift (ppm)
1
H shift (ppm)
13
C shift (ppm)
15
N shift (ppm)
15
N shift (ppm)
1
H shift (ppm)
13
C shift (ppm)
Figure 7
Heteronuclear 2D correlation NMR spectra measured on different dynamic nuclear polarization–enhanced
small molecules, showing the compatibility of ultrafast methods with the ex situ liquid-phase
hyperpolarization approach. All spectra were measured within 1 s after the sudden dissolution and injection
of the targeted molecules, leading to the indicated final effective concentrations. HMBC, heteronuclear
multiple-bond correlation; HSQC, heteronuclear single-quantum coherence experiments.
4. EXTENDING ULTRAFAST 2D NUCLEAR MAGNETIC RESONANCE
TO HIGHER DIMENSIONALITIES
Higher-dimensional spectroscopy extends the principles embodied in Equation 2, through the
addition of an extra mixing process and an additional indirect domain (7, 65). As an example,
3D NMR must deal with the independent incrementation of two indirect-domain time variables,
t
1
and t
2
, each followed by fixed sequences Mixing
1
and Mixing
2
, respectively, concluding with
data acquisition as a function of a direct-domain time t
3
. The need to increment two nested
indirect-domain delays raises even further challenges in terms of the minimum times required to
www.annualreviews.org
Ultrafast Multidimensional NMR 439
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
1
H
15
N
13
C
G
z
F
1
spatial
encoding
F
2
temporal
encoding
F
3
signal
digitization
F
2
/
13
C
(±1000 Hz)
(±1000 Hz)
F
2
/
13
C (±1000 Hz)
a Hybrid 3D pulse sequence b 3D HNCO 2-mM U-
13
C/
15
N LAF
2
ττ ττ
ΤΤ Τ
Τ
π/2π/2
(ii)
(ii)
(i)
(i)
(ii)
(i)
C
α
+G
e
G
e
+G
a
G
a
N
1
t
2
2
t
2
t
1
max
t
1
max
t
1
max
2T
a
Acquisition time = 85 s
Acquisition time = 85 s
Acquisition time = 85 s
F
3
/
1
H
(±1000 Hz)
(±1000 Hz)
F
1
/
1
5
N
(±1187 Hz)
(±1187 Hz)
F
3
/
1
H (±1000 Hz)
F
1
/
15
N (±1187 Hz)
Figure 8
(a) Pulse sequence assayed for the accelerated acquisition of 3D HNCO spectra, with the
15
N dimension spatially encoded along the z
axis and
13
C interactions monitored using a conventional temporal encoding. Data in the resulting ultrafast 2D planes are then suitably
rearranged and Fourier transformed along the t
2
, t
3
temporal dimensions. Narrow and wide bars denote nonselective π/2 and π pulses,
respectively; also indicated are the frequency-swept spatial-encoding pulses (π/2 labels). All
13
C pulses affect the CO resonance region
except for a C
α
-specific π inversion in the middle of the t
2
evolution (shaped pulse) for homonuclear decoupling purposes.
(b) Experimental 3D spectrum collected with the sequence shown in panel a on a uniformly enriched leucine-alanine-phenylalanine
peptide solution. For the
13
C dimension, t
2
= 500 μs and 16 time increments were used; overall, 64 scans with a recycle delay of 1.3 s
were accumulated.
HNCO: 3D NMR
experiment correlating
1
H and
15
N peaks of a
given amide group
with the carbonyl
13
C
peak belonging to the
previous residue
complete the acquisition of these higher-dimensional data sets; this section briefly discusses new
opportunities in this area that may arise from the advent of spatial-encoding methods.
The simplest way to incorporate the concepts introduced in Section 2 into a 3D NMR se-
quence is to spatially encode one of the indirect domains, while retaining a conventional temporal
incrementation to monitor the other. The experimental times required for acquiring such 3D
NMR data set then correspond to those normally associated with a 2D acquisition, a substantial
reduction compared with conventional counterparts. Figure 8 shows such a hybrid option as ap-
plied to the collection of a 3D HNCO NMR data set. Out of the various alternatives, the scheme
shown in Figure 8a incorporates a spatially encoded chemical-shift evolution of
1
H-enhanced
15
N sites along the sample’s z axis, while
13
C interactions are monitored in a multiscan fashion
using a conventional t
2
incrementation. The resulting HNCO spectrum confirms that 3D NMR
data indeed can be obtained in this fashion at approximately 2-mM concentration levels on a 500-
MHz spectrometer equipped with a conventional room-temperature probe, within approximately
90-s acquisition times (Figure 8b). Similar timescales and concentrations are also amenable when
dealing with small proteins (66).
Further accelerations of high-dimensional experiments could arise by combining the single-
scan capabilities of ultrafast 2D NMR with methods capable of performing a more efficient scan-
ning of nD time domains based on taking a small number of 2D slices within it (19–21). In the
context of 3D NMR acquisition, these projection-reconstruction (PR) methods yield an efficient
route to the sampling of the indirect-domain evolution domains. PR substitutes the independent
incrementation of the (t
1
,t
2
) times with a limited number of 2D NMR acquisitions involving fixed
ratios α = tan
1
(
t
2
t
1
) of these variables. These in turn represent different α-projections of the spec-
tral data within the 3D frequency space, which can yield the full F
1
F
2
F
3
3D spectrum being sought
440 Mishkovsky
·
Frydman
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
TOCSY-based 3D UF PR
n-butyl chloride
a HSQC-based 3D UF PR
15
N FMOC-Ala/
15
N FMOC-Val
b
1
0
–1
4
3
2
7
6
9
987654
F
2
(H1, ppm)
F
3
(H-1, ppm)
F
1
(H-1, ppm)
F
2
(N-15, ppm)
F
1
(H-1, ppm)
32101
8
3.75
2.25
0.75
5
4.0
5.5
5.0
4.5
7.0
6.5
116 112 108
F
3
(H-1,
ppm
)
104 100 98 92
7.5
5.2
5.0
4
.8
6.0
Figure 9
Examples of ultrafast (UF)/projection-reconstruction (PR) 3D NMR acquisitions based on (a) homonuclear
and (b) heteronuclear correlations. Each data set was reconstructed from three single-scan ultrafast 2D
projections, leading to a total experimental time of just 5 s (67). HSQC, heteronuclear single-quantum
coherence; TOCSY, total correlation spectroscopy.
if inverse-radon-transformed for each discrete direct-domain F
3
frequency. As the individual pro-
jected planes needed by this protocol involve the collection of 2D time-domain signals, these in
turn can be greatly accelerated by translating the PR principles into the spatio-temporal terms
involved in ultrafast NMR. Assuming that the full sample length L is employed when applying
these joint indirect-domain (t
1
,t
2
) encodings, the original PR conditions can be cast into these
terms by demanding that the spatio-temporal constants {C
i
}
i = 1,2
associated with each indirect
domain be incremented according to α = tan
1
(
C
2
C
1
) (67). Then, 3D NMR spectra can be recorded
by combining a small number of 2D projections, each of which in turn is collected within a single
scan. Figure 9 illustrates this integration with 3D correlation data reconstructed using only three
single-scan 2D NMR projections.
Section 2 describes how frequency-swept RF pulses applied in combination with field gradients
could be used to spatially encode and subsequently read out the indirect domain of a 2D NMR
experiment. Provided that linearly independent gradient geometries are used, these arguments
can be extended further to include an arbitrary number of indirect dimensions (24) and thereby to
acquire arbitrarily high nD NMR data within a single transient. For example, Figure 10a incorpo-
rates two separate gradients arranged along linearly independent z and y geometries to implement
the consecutive encodings of the spin evolution that would be needed for a 3D NMR acquisition.
The first of these processes induces an
1
t
1
-dependent winding of the spin packets along the z
direction, whereas for each z coordinate, a second gradient generates an
2
t
2
-dependent encoding
along the y axis. Because of the ensuing double-winding of spin packets, the overall bulk magne-
tization is reduced again to zero, and an acquisition process implemented on the resulting sample
is associated with a null initial signal. Moreover, only the simultaneous application of suitable G
z
a
,
G
y
a
acquisition gradients can succeed in aligning the spin packets throughout the sample’s volume,
leading to a definition of the peak’s indirect-domain frequencies as a function of two independent
variables: k
z
/C
z
G
z
a
(t)dt(L
z
/t
max
1
) and k
y
/C
y
G
y
a
(t)dt(L
y
/t
max
2
). Oscillation of these two
wave numbers as a function of a direct-domain acquisition time t
3
yields a rasterization of the
3D F
1
F
2
t
3
mixed domain, and 1D FT as a function of t
3
can thereby provide the full set of peak
www.annualreviews.org
Ultrafast Multidimensional NMR 441
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
Τ
Τ
Τ
Τ
ττττ
ππ
ππ
1
H
15
N
13
C
G
z
+G
e
z
–G
e
z
+G
e
y
–G
e
y
–G
a
y
–G
a
y
+G
a
y
+G
a
y
N
1
N
1
N
2
+G
a
z
–G
a
z
+G
a
z
–G
a
z
G
y
G
x
a 3D HNCO with full spatial encoding
t
1
max
t
2
max
2T
a
2T
a
(±290 Hz)
15
(±650 Hz)
13
(±770 Hz)
F
3
/
1
H (±290 Hz)
b 3D HNCO 2-mM U-
13
C/
15
N LAF
(ii)
(i)
(ii)
(i)
Acquisition time = 85 s
Acquisition time = 2 s
F
1
/
1
5
N (±650 Hz)
F
2
/
13
C (±770 Hz)
Figure 10
(a) 3D ultrafast HNCO based on a pulse sequence involving spatial encoding of
15
N(F
1
) along the z axis and
13
C(F
2
) along the y axis.
(b) Data set collected on U-
15
N/
13
C LAF in d
6
-DMSO/H
2
O in only two scans (for the sake of phase cycling away artifacts) on an
800-MHz spectrometer using a room-temperature probe.
coordinates along all the 3D spectral coordinates for all intervening sites—within a single scan.
The resulting approach suffers from an even further decrease in sensitivity over its 2D counterpart
because of its need to concomitantly sample multiple spectral domains per dwell along the direct
time domain. Nevertheless, if sensitivity is sufficient, a dramatic reduction in the overall acquisition
time required for completing a 3D NMR experiments can be achieved. Figure 10b illustrates this
for a 3D HNCO example on the same peptide that was analyzed in Figure 8; this time, only a cou-
ple of phase-cycled scans lead to the desired trace in an approximately 2-s overall acquisition (66).
5. ADDITIONAL SPECTROSCOPIC OPPORTUNITIES PROVIDED
BY SPATIALLY SELECTIVE MANIPULATIONS
Similar to its NMR imaging counterpart, ultrafast NMR operates under the presence of extensive
magnetic field gradients. In contrast to MRI-oriented methods seeking to obtain spatially resolved
information, however, the techniques described above look for a purely spectroscopic, coherent
NMR evolution. Still, we believe that these applications within an nD NMR context constitute
just part of the potential of these novel concepts. Indeed, in addition to a variety of applications
demonstrated in MRI-oriented experiments (68–71), spatial encoding has also been used to speed
up other types of purely spectroscopic NMR measurements that—although not necessarily mon-
itoring a coherent indirect-domain evolution—still demand the acquisition of a series of spectra
for their execution. This section briefly surveys some of these instances.
A well-known example of a measurement requiring multiscan acquisitions is the inversion-
recovery experiment, which aims to measure the spin-lattice relaxation time T
1
(2, 7). This method
monitors signals arising upon subjecting spins to a delay-180
-τ -90
sequence, as a function of
various τ values. The most time-consuming step in this kind of experiment is the initial recycle de-
lay, during which the magnetization is expected to recover to its full equilibrium value after waiting
for an a priori unknown delay, 5T
1
. By replacing the final 90
hard pulse with a spatially selective
excitation coupled to a single-scan data-acquisition period, this experiment can be compressed
from an array into a single scan (72). In the resulting single-scan inversion-recovery sequence, a
single 180
pulse is thus required to invert the full magnetization; as magnetization relaxes back
442 Mishkovsky
·
Frydman
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
2 3
2
3
10 ms
2 3
2
3
450 ms
2 3
2
3
1.25 s
H
3
CH
3
C
C
O
CH
3
N
CH
3
C
O
CH
3
N
CH
3
2 3
2
3
850 ms
1
H shift (ppm)
1
H shift (ppm)
1
H shift (ppm)
1
H shift (ppm)
1
H shift (ppm)
1
1
2
2
31
3
2
3
Figure 11
Array of single-scan 2D exchange NMR spectra collected on a dimethylacetamide/D
2
O solution as a
function of the indicated mixing periods. All these 2D spectra rely on a common initial spatial-encoding
process along the sample’s z axis; each spectrum was then recalled and measured following different exchange
delays, using 90
final excitation pulses and selecting different planes along the sample’s x direction. The
time required to collect each of the 2D NMR experiments was 66 ms plus the indicated duration of the
mixing times; the acquisition of the full series of 2D NMR spectra required only 1.499 s, including all the
nested mixing delays (73).
to equilibrium, different slices of the sample are then successively excited and probed as a function
of different recovery durations τ —thus completing the measurement of the full array in a single
transient. A somewhat similar idea, incorporating an array of a 2D set of exchange experiments as
a function of varying mixing periods (73), has also been demonstrated in this fashion. Figure 11
illustrates how, in very short acquisition times, the effects of chemical exchange can be clearly
followed by the growth of the ensuing off-diagonal exchange-derived cross peaks.
Another spatial manipulation principle that—although it does not include the encoding of spin
coherences—also compresses an arrayed acquisition into a single scan has been discussed recently
in connection with the 2D diffusion ordered spectroscopy experiment (74, 75). This shift-resolved
characterization of molecular diffusion is used widely, among other applications, for separating
NMR peaks according to the individual chemical components partaking of a complex mixture and
for extracting the approximate hydrodynamic radii of molecules in solution (76–78). A suitable
application of swept pulses and refocused gradients also manages, in this case, to impart along a
sample’s axis an array of diffusion-dependent q-encoding values of the kind needed to extract a
full 2D diffusion ordered spectroscopy spectral set. Figure 12 summarizes the type of results that
can become available by this single-scan methodology.
An additional example of how spatial-encoding strategies can be exploited to compress what
usually entails a series of different scans into a single transient has been demonstrated recently
within the context of the phase cycling of RF pulses for coherence selection and/or artifact suppres-
sion purposes. Phase cycling is particularly onerous scan-wise within the context of conventional
2D acquisitions, given the large number of pulses involved, the numerous coherence transfer
pathways thus potentially created, and the need to couple their suitable filtration to the sampling
of numerous incremented indirect-domain evolution delays (2, 7). It is consequently particularly
attractive to employ spatially selective strategies to compress complex cycling schemes effectively
into a single scan. The feasibility of this approach was demonstrated by the implementation of the
various RF manipulations that would normally be executed throughout the independent step of the
cycling scheme, at once and throughout different positions along the sample. The joint detection
of the signals from all slices and their suitable combination could then yield the desired coher-
ence pathway contributions to the final observable signal (79). In another development related
to artifact suppression in high-resolution 2D NMR, spatially selective manipulations have been
applied to eliminate zero-quantum coherence (ZQC) contributions often arising upon executing
www.annualreviews.org
Ultrafast Multidimensional NMR 443
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
78
1
H NMR shift dimension (ppm)
Single-scan 2D DOSY experiment
D domain (*10
6
mm
2
ms
–1
)
910
0
1
2
3
Figure 12
Single-scan 2D diffusion ordered spectroscopy (DOSY) behavior observed for a mixture of tetraphenyl-
porphine, benzaldehyde, and diphenylether (50 mM each) in CDCl
3
at 25
C. The horizontal axis
corresponds to the high-resolution trace afforded by this sequence along the
1
H dimension; the vertical axis
measures the diffusion coefficients extracted for each peak in the spectrum as a function of chemical shift.
Notice the clear separation between the slower diffusivity shown by the porphyrin molecule peaks
(D 1.2 × 10
6
mm
2
ms
1
) and the remaining, smaller aromatic molecules (75).
2D nuclear Overhauser enhancement spectroscopy–type NMR sequences. These experiments are
challenged by difficulties in differentiating longitudinal M
z
-type magnetizations and/or spin-order
states from ZQCs—given that both behave identically when subject to conventional phase-cycling
or gradient-purging procedures. Conversely, the action of a frequency-swept 180
pulse in the
presence of a magnetic field gradient enables one to eliminate ZQC-derived artifacts while re-
taining the remaining longitudinal components because of the destructive interference between
ZQC contributions that then derive from different parts in the sample (80, 81).
6. CONCLUDING REMARKS
This review presents the basic principles of and some opportunities that could arise from recently
proposed NMR spatial-encoding procedures, particularly for compressing what would normally
require the collection of multiple 1D NMR spectra, into a single scan. One main feature of this
MRI-derived approach to spectroscopy is that, inasmuch as nD NMR acquisitions are concerned,
it is general: Therefore, provided that sufficient sensitivity is available, it can be applied to acquire
a variety of homo- and heteronuclear 2D NMR spectra within a single scan. The availability of
a protocol capable of delivering such rich information within short timescales could thus benefit
certain chemical and biological studies that have hitherto been impractical with the aid of 2D
spectroscopy. Potential applications discussed within this context include the use of 2D NMR as
a real-time tool for following chemical and biophysical processes, for monitoring the fingerprints
of samples undergoing continuous flow, and for exploiting the sensitivity that can be provided
by single-shot nuclear hyperpolarization methods. A number of spectroscopic extensions of these
ideas are also discussed, including their use in accelerating higher-dimensional (3D) NMR ex-
periments; in speeding up diffusion, exchange, and relaxation measurements; and in accelerating
and perfecting the elimination of spectral artifacts.
Despite the exiting new opportunities that arise when considering the various applications of
these concepts, we stress that these benefits can only materialize in practice if there is sufficient
spectral sensitivity. Indeed, whereas spatial encoding enables the compression of a wide variety
444 Mishkovsky
·
Frydman
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
of arrayed NMR experiments into a single scan, useful results can only be extracted from such
acquisitions if there is sufficient sensitivity to observe the desired information. In this respect, on
a per-scan basis, ultrafast nD NMR methods prove less sensitive than conventional counterparts
because of their need to rapidly sample multiple dimensions at once. This in turn brings about a
need to further expand the NMR’s receiver bandwidth, thereby causing an increase in the incoming
spectral noise. All this highlights the importance of enhancing sensitivity as a major goal to focus
on, if intending to bring out the full potential of these accelerated acquisition methods.
SUMMARY POINTS
1. Imaging-derived concepts can be used to encode the time evolution normally undergone
by spins over the course of an NMR experiment, along a spatial domain.
2. This spatial-encoding approach enables the compression of arrayed NMR acquisitions,
particularly those involved in multidimensional NMR experiments, into a single scan.
3. This compression enables the reduction of the overall acquisition times of arbitrary
homo- or heteronuclear 2D NMR experiments by several orders of magnitude, resulting
in an ultrafast approach to multidimensional spectroscopy that may enable new applica-
tions in chemical, biophysical, and in vivo NMR.
4. Among the possible applications for this new acquisition mode are real-time 2D NMR
measurements of analytes subject to continuous flow, the following of transient chemi-
cal and biophysical rearrangements, 2D measurements on hyperpolarized samples, and
extensions to higher-dimensional experiments.
5. The success of these new potential applications hinges on the ability ofachieving sufficient
sensitivity to observe the desired multidimensional NMR signals within a single, or at
most a few, transients.
DISCLOSURE STATEMENT
The authors are not aware of any biases that might be perceived as affecting the objectivity of this
review.
ACKNOWLEDGMENTS
We are grateful to Boaz Shapira, Yoav Shrot, and Maayan Gal for the insight they provided
throughout the research work hereby described. This research was supported by the U.S.-Israel
Binational Science Foundation (BSF 2004298), the Israel Science Foundation (ISF 1206/05), the
European Commission (EU-NMR contract no. 026145), and the generosity of the Perlman Family
Foundation.
LITERATURE CITED
1. Grant DM, Harris RK, eds. 1996. Encyclopedia of NMR. Chichester, NY: Wiley & Sons
2. Derome AE. 1987. Modern NMR Techniques for Chemistry Research. Oxford: Pergamon
3. Frydman L. 2001. Perspectives in solid state NMR: spin-1/2 and beyond. Annu. Rev. Phys. Chem. 52:463–98
4. Cavanagh J, Fairbrother WJ, Palmer AG, Skelton NJ. 1996. Protein NMR Spectroscopy: Principles and
Practice. San Diego: Academic
www.annualreviews.org
Ultrafast Multidimensional NMR 445
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
5. deGraaf R. 2007. In Vivo NMR Spectroscopy: Principles and Techniques. Chichester, NY: Wiley & Sons
6. Buxton RB. 2001. An Introduction to Functional MRI: Principles and Techniques. Cambridge: Cambridge
Univ. Press
7. Ernst RR, Boudenhausen G, Wokaun A. 1987. Principles of Nuclear Magnetic Resonance in One and Two
Dimensions. Oxford: Clarendon
8. Ernst RR, Anderson WA. 1966. Application of Fourier transform spectroscopy to magnetic resonance.
Rev. Sci. Instrum. 37:93–102
9. Jeener J. 1994 (1971). Ampere Summer School lecture notes. In NMR and More in Honor of Anatole
Abragam, ed. M Goldman, M Porneuf, pp. 1–379. Les Ulis, France: Ed. Phys.
10. Aue WP, Bartholdi E, Ernst RR. 1976. Two dimensional spectroscopy: application to nuclear magnetic
resonance. J. Chem. Phys. 64:2229–46
11. Kup
ˇ
ce E, Nishida T, Freeman R. 2003. Hadamard NMR spectroscopy. Prog. Nucl. Magn. Reson. Spectrosc.
42:95–122
12. Atreya HS, Szyperski T. 2005. Rapid NMR data collection. Methods Enzymol. 394:78–108
13. Mandelshtam VA. 2000. The multidimensional filter diagonalization method I. Theory and numerical
implementation. J. Magn. Reson. 144:343–56
14. Hoch JC, Stern AS. 2001. Maximum entropy reconstruction, spectrum analysis and deconvolution in
multidimensional nuclear magnetic resonance. Methods Enzymol. 338:159–78
15. Orekhov Y, Ibraghimov I, Billeter M. 2003. Optimizing resolution in multidimensional NMR by three-
way decomposition. J. Biomol. NMR 27:165–73
16. Br
¨
uschweiler R, Zhang F. 2004. Covariance nuclear magnetic resonance spectroscopy. J. Chem. Phys.
120:5253–60
17. Kup
ˇ
ce E, Freeman R. 2003. Frequency-domain Hadamard spectroscopy. J. Magn. Reson. 162:158–65
18. Bodenhausen G, Ernst RR. 1981. The accordion experiment, a simple approach to three-dimensional
NMR spectroscopy. J. Magn. Reson. 45:367–73
19. Szyperski T, Wider G, Bushweller JH, Wuthrich K. 1993. Reduced dimensionality in triple-resonance
NMR experiments. J. Am. Chem. Soc. 115:9307–8
20. Kim S, Szyperski T. 2003. GFT NMR, a new approach to rapidly obtain precise high-dimensional NMR
spectral information. J. Am. Chem. Soc. 125:1385–93
21. Kup
ˇ
ce E, Freeman R. 2004. Projection-reconstruction technique for speeding up multidimensional NMR
spectroscopy. J. Am. Chem. Soc. 126:6429–40
22. Frydman L, Scherf T, Lupulescu A. 2002. The acquisition of multidimensional NMR spectra within a
single scan. Proc. Natl. Acad. Sci. USA 99:15858–62
23. Frydman L, Scherf T, Lupulescu A. 2003. Principles and features of single-scan two-dimensional NMR
spectroscopy. J. Am. Chem. Soc. 125:9204–17
24. Shrot Y, Frydman L. 2003. Single-scan NMR spectroscopy at arbitrary dimensions. J. Am. Chem. Soc.
125:11385–96
25. Pelupessy P. 2003. Adiabatic single scan two-dimensional NMR spectroscopy. J. Am. Chem. Soc.
125:12345–50
26. Shrot Y, Shapira B, Frydman L. 2004. Ultrafast 2D NMR spectroscopy using continuous spatial encoding
of the spin interactions. J. Magn. Reson. 171:163–70
27. Tal A, Shapira B, Frydman L. 2005. A continuous phase-modulated approach to spatial encoding in
ultrafast 2D NMR spectroscopy. J. Magn. Reson. 176:107–14
28. Andersen NS, K
¨
ockenberger W. 2005. A simple approach for phase-modulated single-scan 2D NMR
spectroscopy. Magn. Reson. Chem. 43:795–97
29. Shapira B, Shrot Y, Frydman L. 2006. Symmetric spatial encoding in ultrafast 2D NMR spectroscopy.
J. Magn. Reson. 178:33–41
30. Giraudeau P, Akoka S. 2007. A new detection scheme for ultrafast 2D J-resolved spectroscopy. J. Magn.
Reson. 186:352–57
31. Shrot Y, Frydman L. 2008. Spatial encoding strategies for ultrafast multidimentional nuclear magnetic
resonance. J. Chem. Phys. 128:052209
32. Lindon JC, Holmes E, Nicholson JK. 2004. Toxicological applications of magnetic resonance. Progr. Nucl.
Magn. Reson. Spectrosc. 45:109–43
446 Mishkovsky
·
Frydman
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
33. Albert K, ed. 2002. On-Line LC-NMR and Related Techniques. Chichester, NY: Wiley & Sons
34. Braunschweiler L, Ernst RR. 1983. Coherence transfer by isotropic mixing: applications to proton cor-
relation spectroscopy. J. Magn. Reson. 53:521–28
35. Shapira B, Karton A, Aronzon D, Frydman L. 2004. Real-time 2D NMR identification of analytes un-
dergoing continuous chromatographic separation. J. Am. Chem. Soc. 126:1262–65
36. Dayie KT, Wagner G, Lef
`
evre J-F. 1996. Theory and practice of nuclear spin relaxation in proteins.
Annu. Rev. Phys. Chem. 47:242–83
37. Bain AD. 2003. Chemical exchange in NMR. Prog. Nucl. Magn. Reson. Spectrosc. 43:63–103
38. Dobson CM, Hore PJ. 1998. Kinetic studies of protein folding using NMR spectroscopy. Nat. Struct.
Biol. 5:504–7
39. Ishima R, Torchia DA. 2000. Protein dynamics from NMR. Nat. Struct. Biol. 7:740–43
40. Gal M, Mishkovsky M, Frydman L. 2006. Real-time monitoring of chemical transformation by ultrafast
2D NMR spectroscopy. J. Am. Chem. Soc. 128:951–56
41. M
¨
uller L. 1979. Sensitivity enhanced detection of weak nuclei using heteronuclear multiple quantum
coherence. J. Am. Chem. Soc. 101:4481–84
42. Schanda P, Brutscher B. 2005. Very fast two-dimentional NMR spectroscopy for real-time investigation
of dynamic events in proteins on the time scale of seconds. J. Am. Chem. Soc. 127:8014–15
43. Pervushin K, Voegeli B, Eletsky A. 2002. Longitudinal
1
H relaxation optimization in TROSY NMR
spectroscopy. J. Am. Chem. Soc. 124:12898–902
44. Freeman R, Hill HDW. 1971. Phase and intensity abnormalities in Fourier transform NMR. J. Magn.
Reson. 4:366–83
45. Ross A, Salzmann M, Senn H. 1997. Fast-HMQC using Ernst angle pulses: an efficient tool for screening
of ligand binding to target protein. J. Biomol. NMR 10:289–96
46. Gal M, Schanda P, Brutsher B, Frydman L. 2007. UltraSOFAST HMQC NMR and the repetitive
acquisition of 2D protein spectra at Hz rates. J. Am. Chem. Soc. 129:1372–77
47. Bowers CR, Weitekamp DP. 1986. Transformation of symetrization order in nuclear magnetic resonance.
Phys. Rev. Lett. 57:2645–48
48. Eisenschmid TC, Kirss RU, Deutsch PP, Hommeltoft SI, Eisenberg R, et al. 1987. Parahydrogen induced
polarization in hydrogenation reaction. J. Am. Chem. Soc. 109:8089–91
49. Albert MS, Cates GD, Driehuys B, Happer W, Saam B, et al. 1994. Biological magnetic resonance imaging
using laser polarized
129
Xe. Nature 370:199–201
50. Navon G, Song YQ, Room T, Appelt S, Taylor RE, Pines A. 1996. Enhancement of solution NMR and
MRI with laser polarized xenon. Science 271:1848–51
51. Muus LT, Atkins PW, McLauchlen KA, Pedersen JB. 1977. Chemically Induced Magnetic Polzarization.
Dordrecht: D. Reidel
52. Shapira B, Morris E, Muszkat KA, Frydman L. 2004. Sub-second 2D NMR spectroscopy at submilimolar
consentrations. J. Am. Chem. Soc. 126:11756–57
53. Carver TR, Slichter CP. 1953. Polarization of nuclear spin in metals. Phys. Rev. 92:212–13
54. Hausser KH, Stehilk D. 1968. Dynamic nuclear polarization in liquids. Adv. Magn. Reson. 3:79–139
55. Abragam A, Goldman M. 1982. Nuclear Magnetism: Order and Disorder. Oxford: Oxford Univ. Press
56. Ardenkjaer-Larsen JH, Fridlund B, Gram A, Hansson G, Hansson L, et al. 2003. Increase in signal-to-
noise ratio of >10,000 times in liquid-state NMR. Proc. Natl. Acad. Sci. USA 100:10158–63
57. Joo CG, Hu KN, Bryant JA, Griffin RG. 2006. In situ temperature jump high-frequency dynamic nu-
clear polarization experiments: enhanced sensitivity in liquid-state NMR spectroscopy. J. Am. Chem. Soc.
128:9428–32
58. Golman K, Ardenkjær-Larsen JH, Petersson JS, M
˚
ansson S, Leunbach I. 2003. Molecular imaging with
endogenous substances. Proc. Natl. Acad. Sci. USA 100:10435–39
59. Golman K, Zandt R, Thaning M. 2006. Real time metabolic imaging. Proc. Natl. Acad. Sci. USA 103:11270–
75
60. Chen AP, Albers MJ, Cunningham CH, KohlerSJ, Yen YF, et al. 2007. Hyperpolarized C-13spectroscopic
imaging of the TRAMP mouse at 3T: initial experience. Magn. Reson. Med. 58:1099–106
61. Day SE, Kettunen MI, Gallagher FA, Hu DE, Lerche M, et al. 2007. Detecting tumor response to
treatment using hyperpolarized
13
C magnetic resonance imaging. Nat. Med. 13:1382–87
www.annualreviews.org
Ultrafast Multidimensional NMR 447
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
ANRV373-PC60-21 ARI 25 February 2009 18:3
62. Gallagher FA, Kettunen MI, Day SE, Hu DE, Ardenkjaer-Larsen JH, et al. 2008. Magnetic resonance
imaging of pH in vivo using hyperpolarized
13
C-labelled bicarbonate. Nature 457:940–43
63. Frydman L, Blazina D. 2007. Ultrafast two-dimensional nuclear magnetic resonance spectroscopy of
hyperpolarized solutions. Nat. Phys. 3:415–19
64. Mishkovsky M, Frydman L. 2008. Progress in hyperpolarized ultrafast 2D nuclear magnetic resonance.
Chem. Phys. Chem. 16:2340–48
65. Sattler M, Schleucher J, Griesinger C. 1999. Heteronuclear multidimensional NMR experiments for the
structure determination of proteins in solution employing pulsed field gradients. Progr. Nucl. Magn. Reson.
Spectrosc. 34:93–158
66. Mishkovsky M, Gal M, Frydman L. 2007. Spatially-encoded strategies in the execution of biomolecular-
oriented 3D NMR experiments. J. Biomol. NMR 39:291–301
67. Mishkovsky M, Kupce E, Frydman L. 2007. Ultrafast based projection-reconstruction three-dimensional
nuclear magnetic resonance spectroscopy. J. Chem. Phys. 127:034507
68. Shrot Y, Frydman L. 2005. Spatially-encoded NMR and the acquisition of 2D magnetic resonance images
within a single scan. J. Magn. Reson. 172:179–90
69. Tal A, Frydman L. 2006. Spatial encoding and the acquisition of high definition MR images in inhomo-
geneous magnetic fields. J. Magn. Reson. 181:179–94
70. Chamberlain R, Park JY, Corum C, Yacoub E, Ugurbil K, et al. 2007. RASER: a new ultrafast magnetic
resonance imaging method. Magn. Reson. Med. 58:794–99
71. Tal A, Frydman L. 2007. Spectroscopic imaging from spatially-encoded single-scan multidimensional
MRI data. J. Magn. Reson. 189:46–58
72. Bhattacharyya R, Kumar A. 2003. A fast method for the measurement of long-spin lattice relaxation times
by single scan inversion recovery experiment. Chem. Phys. Lett. 383:99–103
73. Shapira B, Frydman L. 2003. Arrayed acquisition of 2D exchange NMR spectra within a single experiment.
J. Magn. Reson. 165:320–24
74. Thrippleton MJ, Loening NM, Keeler J. 2003. A fast method forthe measurement of diffusion coefficients:
one-dimensional DOSY. Magn. Reson. Chem. 41:441–47
75. Shrot Y, Frydman L. 2008. Single scan 2D DOSY NMR. J. Magn. Reson. 195:226–31
76. Morris KF, Johnson CS. 1993. Resolution of discrete and continuous molecular-size distributions by
means of diffusion-ordered 2D NMR spectroscopy. J. Am. Chem. Soc. 115:4291–99
77. Barjat H, Morris GA, Smart S, Swanson AG, Williams SCR. 1995. High-resolution diffusion-ordered 2D
spectroscopy (HR-DOSY): a new tool for the analysis of complex mixtures. J. Magn. Reson. B 108:170–72
78. Johnson CS. 1999. Diffusion-ordered NMR spectroscopy: principles and applications. Prog. NMR Spec-
trosc. 34:203–55
79. Parish DN, Szyperski T. 2008. Simultaneously cycled NMR spectroscopy. J. Am. Chem. Soc. 130:4925–33
80. Thrippleton JM, Keeler J. 2003. Elimination of zero quantum interference in two dimensional NMR
spectra. Angew. Chem. Int. Ed. Engl. 42:3938–41
81. Cano E, Thrippleton JM, Keeler J, Shaka AJ. 2004. Cascaded z-filters for efficient single-scan suppression
of zero-quantum coherence. J. Magn. Reson. 167:291–97
448 Mishkovsky
·
Frydman
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
AR373-FM ARI 25 February 2009 17:55
Annual Review of
Physical Chemistry
Volume 60, 2009
Contents
Frontispiece ppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppp
xiv
Sixty Years of Nuclear Moments
John S. Waugh ppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppp1
Dynamics of Liquids, Molecules, and Proteins Measured with Ultrafast
2D IR Vibrational Echo Chemical Exchange Spectroscopy
M.D. Fayer pppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppp21
Photofragment Spectroscopy and Predissociation Dynamics of Weakly
Bound Molecules
Hanna Reisler pppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppp39
Second Harmonic Generation, Sum Frequency Generation, and χ
(3)
:
Dissecting Environmental Interfaces with a Nonlinear Optical Swiss
Army Knife
Franz M. Geiger pppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppp61
Dewetting and Hydrophobic Interaction in Physical and Biological
Systems
Bruce J. Berne, John D. Weeks, and Ruhong Zhou ppppppppppppppppppppppppppppppppppppppppp85
Photoelectron Spectroscopy of Multiply Charged Anions
Xue-Bin Wang and Lai-Sheng Wang pppppppppppppppppppppppppppppppppppppppppppppppppppppp105
Intrinsic Particle Properties from Vibrational Spectra of Aerosols
´
Omar F. Sigurbjörnsson, George Firanescu, and Ruth Signorell ppppppppppppppppppppppppp127
Nanofabrication of Plasmonic Structures
Joel Henzie, Jeunghoon Lee, Min Hyung Lee, Warefta Hasan, and Teri W. Odom pppp147
Chemical Synthesis of Novel Plasmonic Nanoparticles
Xianmao Lu, Matthew Rycenga, Sara E. Skrabalak, Benjamin Wiley,
and Younan Xia pppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppp167
Atomic-Scale Templates Patterned by Ultrahigh Vacuum Scanning
Tunneling Microscopy on Silicon
Michael A. Walsh and Mark C. Hersam pppppppppppppppppppppppppppppppppppppppppppppppppp193
DNA Excited-State Dynamics: From Single Bases to the Double Helix
Chris T. Middleton, Kimberly de La Harpe, Charlene Su, Yu Kay Law,
Carlos E. Crespo-Hernández, and Bern Kohler pppppppppppppppppppppppppppppppppppppppppppp217
viii
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
AR373-FM ARI 25 February 2009 17:55
Dynamics of Light Harvesting in Photosynthesis
Yuan-Chung Cheng and Graham R. Fleming ppppppppppppppppppppppppppppppppppppppppppppp241
High-Resolution Infrared Spectroscopy of the Formic Acid Dimer
¨
Ozgür Birer and Martina Havenith ppppppppppppppppppppppppppppppppppppppppppppppppppppppp263
Quantum Coherent Control for Nonlinear Spectroscopy
and Microscopy
Yaron Silberberg ppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppp
277
Coherent Control of Quantum Dynamics with Sequences of Unitary
Phase-Kick Pulses
Luis G.C. Rego, Lea F. Santos, and Victor S. Batista ppppppppppppppppppppppppppppppppppppp293
Equation-Free Multiscale Computation: Algorithms and Applications
Ioannis G. Kevrekidis and Giovanni Samaey pppppppppppppppppppppppppppppppppppppppppppppp321
Chirality in Nonlinear Optics
Levi M. Haupert and Garth J. Simpson ppppppppppppppppppppppppppppppppppppppppppppppppppp345
Physical Chemistry of DNA Viruses
Charles M. Knobler and William M. Gelbart ppppppppppppppppppppppppppppppppppppppppppppp367
Ultrafast Dynamics in Reverse Micelles
Nancy E. Levinger and Laura A. Swafford pppppppppppppppppppppppppppppppppppppppppppppppp385
Light Switching of Molecules on Surfaces
Wesley R. Browne and Ben L. Feringa ppppppppppppppppppppppppppppppppppppppppppppppppppppp407
Principles and Progress in Ultrafast Multidimensional Nuclear
Magnetic Resonance
Mor Mishkovsky and Lucio Frydman pppppppppppppppppppppppppppppppppppppppppppppppppppppp429
Controlling Chemistry by Geometry in Nanoscale Systems
L. Lizana, Z. Konkoli, B. Bauer, A. Jesorka, and O. Orwar ppppppppppppppppppppppppppppp449
Active Biological Materials
Daniel A. Fletcher and Phillip L. Geissler ppppppppppppppppppppppppppppppppppppppppppppppppp469
Wave-Packet and Coherent Control Dynamics
Kenji Ohmori pppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppp487
Indexes
Cumulative Index of Contributing Authors, Volumes 56–60 ppppppppppppppppppppppppppp513
Cumulative Index of Chapter Titles, Volumes 56–60 pppppppppppppppppppppppppppppppppppp516
Errata
An online log of corrections to Annual Review of Physical Chemistry articles may be
found at http://physchem.annualreviews.org/errata.shtml
Contents ix
Annu. Rev. Phys. Chem. 2009.60:429-448. Downloaded from arjournals.annualreviews.org
by Weizmann Institute of Science on 06/22/09. For personal use only.
    • "A variety of recent developments have succeeded in this direction by speeding up normal timedomain sampling and the pulsing repetition rate (see reviews [61, 62]). Optimized spectral aliasing [63] , spectral reconstruction algorithms that can handle sparse, non-uniformly sampled data sets646566, Hadamard-encoding techniques [67], and spatially encoded, single-scan methods [68,69] use sparse data sampling for fast recording of real-time NMR data. The advent of non-uniform sampling paved the way into a new era of multidimensional NMR spectroscopy in short time (more details can be found in reviews707172). "
    [Show abstract] [Hide abstract] ABSTRACT: During protein folding reactions toward the native structure, short-lived intermediate states can be populated. Such intermediates expose hydrophobic patches and can self-associate leading to non-productive protein misfolding. A major focus of current research is the characterization of short-lived intermediates and how molecular chaperones enable productive folding. Real-time NMR spectroscopy, together with the development of advanced methods, is reviewed here and the potential these methods have to characterize intermediate states as well as interactions with molecular chaperone proteins at single-residue resolution is highlighted. Various chaperone interactions can guide the protein folding reaction and thus are important for protein structure formation, stability, and activity of their substrates. Chaperone-assisted protein folding, characterization of intermediates, and their molecular interactions using real-time NMR spectroscopy will be discussed. Additionally, recent advances in NMR methods employed for characterization of high-energy intermediates will be discussed. Real-time NMR combines high-resolution with kinetic information of protein reactions, which can be employed not only for protein folding studies and the characterization of folding intermediates but also to investigate the molecular mechanisms of assisted protein folding. Real-time NMR spectroscopy remains an effective tool to reveal structural details about the interaction between chaperones and transient intermediates. Methodologically, it provides in-depth understanding of how kinetic intermediates and their thermodynamics contribute to the protein folding reaction. This review summarizes the most recent advances in this field. This article is part of a Special Issue entitled Proline-directed Foldases: Cell Signaling Catalysts and Drug Targets. Copyright © 2014. Published by Elsevier B.V.
    Full-text · Article · Dec 2014
    • "Although the principles of ultrafast 2D MRS spectroscopy have been described in detail (12,18,19 ), it is convenient to revisit their use for the specific purpose of this study, addressing joint SPSP selectivity. SPEN departs from traditional schemes in that instead of triggering the chemical shift evolution homogenously, it allows various chemical sites to accrue their spin evolution in a spatially heterogeneous fashion. "
    [Show abstract] [Hide abstract] ABSTRACT: PURPOSE: To introduce a method that provides simultaneous spatial and spectral selectivity, whose implementation is less demanding than-and quality comparable to-conventional 2D spectral-spatial counterparts. THEORY: Spatiotemporal encoding concepts lead to a spatially selective, chemical-shift-dependent echo, with simultaneous dephasing of all other off-resonant species. The approach only requires applying a pair of suitable radiofrequency-swept pulses, and allows arbitrary shaping of the spatial profiles. METHODS: Based on arguments derived for chirp pulses operating in the sequential-sweep approximation, quadratic-phase SLR excitation and refocusing waveforms were designed and used to collect 2D slice- and shift-selective images on a 7 T microimaging system (phantoms). The same strategy was used to obtain multi-slice echo-planar and spin-echo images of breast on human volunteers in a 3 T scanner. RESULTS: The method managed to deliver excellent shift-selective multi-slice images in phantoms and human volunteers. Simultaneous water and fat images were also collected in a single, interleaved acquisition mode on both platforms, using straightforward sequence and reconstruction modifications of the basic scheme. CONCLUSION: A new way to achieve chemical shift selectivity with high quality spatial profiling is achieved, without the usual requirements for playing out fast oscillating gradients in conjunction with carefully timed radiofrequency pulses. Magn Reson Med, 2013. © 2013 Wiley Periodicals, Inc.
    Full-text · Article · Feb 2014
    • "Of course we have barely covered the tip of the 'iceberg'. There is a huge ongoing effort to further increase the efficiency of the NMR experiments with important contributions from the groups of Wüthrich et al. [66], Wagner et al. [67], Mishkovsky and Frydman [68], Markley et al. [69], Brutcher et al. [70], Zhang and Bruschweiler [71], Orekhov et al. [72], Hoch et al. [73], Celik and Shaka [74], Kozminski et al. [75] and many others. Cross-fertilization of ideas with similar developments in magnetic resonance imaging [75,76] will undoubtedly produce new techniques and new applications in this rapidly developing field. "
    [Show abstract] [Hide abstract] ABSTRACT: The recent introduction of NMR spectrometers with multiple receivers permits spectra from several different nuclear species to be recorded in parallel, and several standard pulse sequences to be combined into a single entity. It is shown how these improvements in the flow and quality of spectral information can be significantly augmented by compressive sensing techniques--controlled aliasing, Hadamard spectroscopy, single-point evaluation of evolution space (SPEED), random sampling, projection-reconstruction, and hyperdimensional NMR. Future developments of these techniques are confidently expected to mitigate one of the most serious limitations in multidimensional NMR--the excessive duration of the measurements.
    Full-text · Article · Aug 2011
Show more

Recommended publications

Discover more