Ann. Magn. Reson. Vol. 1, Nos 1 and 2, 1-31, 2003 AUREMN ©

REVIEW


- 1 -
Applications and Advances in Cryogenic NMR Probes and
Computer-Assisted Structure Elucidation



Gary E. Martin* and David J. Russell
Pfizer Corporation, Global Research and Development; Worldwide Pharmaceutical Sciences;
Kalamazoo, MI 49001-0199


Kirill A. Blinov and Mikhail E. Elyashberg
Advanced Chemistry Development; Bakuleva 6, str. 1; Moscow 117513, Russian Federation

Antony J. Williams
Advanced Chemistry Development, Inc.; 90 Adelaide St., West, Suite 600
Toronto, Ontario M5H 3V9, Canada


* To whom inquiries should be directed
Gary E. Martin
gary.e.martin@pharmacia.com
+(269)833-6283 (Fax) +(269)833-2030
Pfizer Corporation
Global Research & Development
Worldwide Pharmaceutical Sciences
Rapid Structure Characterization Group
MS# 4821-259-277
7000 Portage Road
Kalamazoo, MI 49001-0199


Final revision July 9, 2003









Abstract: There have been significant advances over approximately the last decade in terms of NMR
probes designed for dealing with progressively smaller samples. Probe diameters and correspondingly
sample volumes have progressively decreased. Recently, cryogenic NMR probes have started to be
installed in some NMR facilities providing dramatic further increases in sensitivity, which have pushed
sample requirements still lower. The impact of advances with cryogenic NMR probes will be discussed
in the context of sample requirements for the structural characterization of complex molecules. In
parallel, the development of computer-assisted structure elucidation (CASE) methods has likewise
continued producing innovative new solutions to overcome hurdles that existed in the earliest CASE
systems. In this work advanced computer algorithms to supplement human capabilities and intuition,
coupled with vastly improved sensitivity through the advent of cryogenic NMR probes will be discussed
and the characterization of several indoloquinoline alkaloids from Cryptolepis sanguinolenta will be
used as case studies.

Martin et al Ann. Magn. Reson.
- 2 -
N
N
H
N
N
CH3

x1x x2x

Introduction

Small volume NMR probes are by no
means a new idea. Various researchers have
explored various aspects of this idea for a number
of years.1,2 Renewal of active interest in this area of
NMR began in late 1991 and was detailed in reports
published in early 1992 by one of the authors and
his co-workers.3,4 In the intervening years, 2.5 or 3
mm “micro” NMR probes have become a mainstay
of individuals who routinely engage in small sample
NMR studies. More recently, magic angle and
conventional 1.7 and 1.0 mm “submicro” NMR
probes have also been described. Together, these
developments formed the basis for a recent review.5
The development of new high sensitivity,
small volume NMR probe technology, coupled with
the development of new 2D NMR experimental
methods and broader access to high field NMR
instruments have greatly facilitated the structural
characterization of new and often scarce natural
products, degradants and impurities. Indeed, micro
NMR probe technology was extensively utilized in
the structural characterization of many members of
the family of indoloquinoline alkaloids isolated from
the shrub Cryptolepis sanguinolenta indigenous to
Ghana and used in folkloric medicine.

Cryptolepis Alkaloid Structure Characterization
Quindoline represents a relatively rare
example of a molecule that was known synthetically
before it was identified as a natural product.6 Two
members of the indolo- [3,2-b]quinoline family of
alkaloids, cryptolepine (1) and quindoline (2), have
been known for many years7,8 and extensively
studied using modern NMR methods.9-11
Ongoing studies of C. sanguinolenta during
the period from 1991-1994 led mainly to the
isolation and structural characterization of a number
of 11-substituted analogues of the indolo[3,2-
b]quinoline skeleton.12 One notable exception
was the 1992 report of the isolation and structural
characterization of the complex, spiro nonacyclic
alkaloid cryptospirolepine (3).13
N
N
CH3
O
N
NH
CH3


x3x
NN
CH3
N
N
CH3
x4x x5x

Vol. 1, Nos 1 and 2, 2003 Applications and Advances in Cryogenic NMR Probes and
Computer-Assisted Structure Elucidation

- 3 -
Around the same time, another C31
alkaloid, given the designation TC-6, was also
isolated but efforts to characterize its structure
were frustratingly fruitless. During late 1994 and
1995 several additional alkaloids were
isolated from C. sanguinolenta that demonstrated
the biogenetic versatility of this plant species
in that two new indoloquinoline ring systems
and an indolobenzazepine were identif ied.
These included cryptosanguinolentine (4, an indolo
[3,2-c]quinoline), cryptotackieine (5, an indolo[2,3-b]
quinoline), and homocryptolepinone (6, an indolo




















[3,2-b] benzazepinone), the latter constituting a
portion of the complex structure of cryptospirolepine
(3) and possibly a component of the biosynthetic
pathway leading to the formation of that molecule.14-17
Beginning in 1995, dimeric and mixed dimeric
alkaloids began to be isolated and identified from C.
sanguinolenta. The first compound of this group was a
mixed “dimer” identified as cryptolepicarboline (7),
comprised of cryptolepine and β-carboline that was
reported in 1995.18 The first symmetrical dimer,
cryptomisrine (8), was reported the following year.19
Subsequently, biscryptolepine (9) and cryptoquindoline
(10) were reported, the latter suggested to be a
dimerization artifact of the isolation scheme.15,16 Very
recently, cryptolepinone and the mixed dimer
cryptoquindolinone (11) were identified as two of the 26
chromatographically separable degradation products of
cryptospirolepine (3) that formed on long term storage of
a NMR sample in DMSO-d6.20 Cryptolepinone and its
oxidation product, the 5-N-oxide, were isolated and
identified previously in a series of reports.12,21-24




N
N
H
O
CH3

x6x
N
N
N
N
CH3
N
NH
O
NHN

x7x x8x

N
N
CH3
N
N
N
N
CH3
N
N
CH3
x9x x10x

Martin et al Ann. Magn. Reson.
- 4 -
The Evolving Role of Computer-Assisted Structure
Elucidation (CASE) Methods in Natural Product
Structure Characterization


The isolation and structural characterization
of cryptoquindolinone (11) is particularly interesting in
that the structure elucidation was carried out in
parallel by an investigator intimately familiar with this
alkaloid family and by computer-assisted structure
elucidation (CASE) methods.20 Hallmarks of any
successful structure elucidation effort are:
1) complete 1H and 13C NMR spectral assignments
based on 2D-NMR data; 2) mass spectral fragment
ions that can be explained by rational fragmentation
pathways that are consistent with the proposed
structure; and 3) vibrational spectra that can be
reasonably explained on the basis of the functional
groups present in the molecule being studied.
Cryptoquindolinone (11), as noted above,
was the 2nd most abundant of the 26
chromatographically resolvable degradation products
formed on prolonged storage of a 2.5 mg NMR
sample of cryptospirolepine (3) in DMSO-d6. It is
interesting to contrast the characterization of this
molecule by human investigators working in parallel
with the use of CASE methods. Familiarity with
Cryptolepis indoloquinoline alkaloid family allowed
human intuition to make what might be called a
quantum “leap” forward in the elucidation process
by recognizing from a color shift to deep purple
when the sample was made basic with ammonia
gas and from 13C NMR shifts contained in the

N
N
CH3
N
N
O
H3C

x11x

N
N
CH3

x12x
HMBC spectrum. Based on these observations, it
was deduced that the degradant molecule must
contain an 11-cryptolepinyl (12) fragment in its structure.
Practically, this knowledge-based deduction
reduced the complexity of the structure elucidation
substantially by removing about half of the resonances
from consideration, allowing the investigator to focus
attention on the remaining half of the molecule. The
problem was solved within 72 hrs of the completion
of NMR data acquisition by the human investigator,
the balance of the structure identified as a
quindolinone (13) moiety substituted via the indole
nitrogen at the 10-position.
Contrasting the work done by the human
investigator in elucidating the structure of
cryptoquindolinone (11) with results from a CASE
approach provides an interesting comparison. The
Structure Elucidator v6.08 program package was
used for the computer-assisted structure elucidation
portion of the study; data consisting of an accurate
mass and empirical formula, a 1H reference
spectrum, and gradient 2D COSY, HSQC, phase-
sensitive HMBC, and a 500 msec ROESY spectrum
were also input. No constraints were imposed on
the program, which gave ~3,300 structures in 8 hr.
After removal of duplicate structures, 355 structures
remained. Sorting these structures on the basis of
accurate calculated 13C chemical shifts, the correct
structure, cryptoquindolinone (11) was in the second
position in the accurate 13C chemical shift sorted
(dA13C) list of structures. It is important to note that
the initial computational run did not have the
advantage of the intuitive deduction that the
molecule contained an 11-cryptolepinyl moiety that
the human investigator had available. When the 11-
cryptolepinyl fragment was included as input for
Structure Elucidator, the program was able to output
1268 structures in less than 10 minutes, which

Vol. 1, Nos 1 and 2, 2003 Applications and Advances in Cryogenic NMR Probes and
Computer-Assisted Structure Elucidation

- 5 -
reduced to 111 structures after filtering to remove
duplicates. Again, when the data were examined,
the correct structure was in the second position of
the dA 13C-sorted list of structures. As seen from
this comparison, 72 hrs for human interpretation of
the data vs. as little as 10 minutes when Structure
Elucidator was provided the same knowledge-based
constraints as the human investigator, underscores
the powerful synergy of the interaction between
competent investigators and CASE programs.
In studies of the versatility of the Structure
Elucidator program, results obtained by applying the
program to the elucidation of ~150 structures
reported in the Journal of Natural Products has
been reported.25 The size of the structures selected
ranged from 15 to 90 skeletal atoms; molecular
weights for the molecules studied ranged from 200 to
1280 Da. Unambiguous structures were determined for
91 of the 100 molecules investigated; 75% of the
structures were solved in < 1 min and 95% were
solved in < 30 min of computer time. Another
interesting observation was that only very rarely was
the correct structure not in the first five structures of
the dA13C-sorted list of structures and that in 85% of
cases, the correct structure had a dA13C < 4 ppm.25,26
Examining the case studies further, the
number of problems encountered involving
connectivities of non-standard length (e.g. nJCH, nJHH n
> 3, when unanticipated) was 50. For those 50
structural problems, the number of cases when the
program detected the non-standard connectivities
automatically (algorithmically) was 44/50 (90%).
Perhaps more importantly, in those 44 cases when
the program was able to detect non-standard
connectivities, the program was able to resolve the
contradictions automatically in 22/44 cases (44%
overall) allowing the automated solution of the
structures in question. Some problems involving
non-standard connectivities, 6/50 (~10%), were not
detected by the program. Half of the problems involving
algorithmically undetected, non-standard connectivities
were indirectly identified by large dA 13C values (>5.5
ppm) associated with the solution. For the remaining
three cases, two involved hidden contradictions and one
was a case where a –OH and –CH3 group were
permuted, so it is not possible to unequivocally say
which structure would be correct. Overall, 49/50 (98%)
problems involving the presence of non-standard
connectivities were detected by a combination of direct
(algorithmic) and indirect means.27, 28
The results highlighted above are especially
important since many structures can now be solved
using 2D NMR data without verification of the data for
the presence of non-standard connectivities. For some
molecules, 10 or more such connectivities can
sometimes be observed using some of the new
accordion-optimized long-range heteronuclear shift
correlation experiments, making this an important
capability.29 In 50% of the cases, the presence of non-
standard long-range correlations was detected
automatically and the contradictions resolved, allowing
the successful characterization of the structure. In the
long term, this capability may obviate the need for
repeating statically-optimized long-range experiments
(HMBC) with different optimizations. Conversely, the
algorithmic capabilities now extant and still being refined
in Structure Elucidator may encourage more extensive
usage of accordion-optimized long-range correlation
experiments by investigators who might otherwise be
hesitant to use them because of the significant increase
in the information content of these experiments in some
cases. In the case of some heternouclide pairs,
specifically 1H-15N long-range heteronuclear shift
correlation experiments, several studies have shown
that consistently more reliable data are obtained when
accordion-optimized experiments are used because of
the variability of nJHN (n ≥ 2) heteronuclear couplings.30-32



N
N
O
H3C

x13x

Martin et al Ann. Magn. Reson.
- 6 -
Elucidation of the Structure of Quindolinocryptotackieine
Initial considerations.

One of the alkaloidal component fractions
isolated in 1990 in Ghana at the same time as the
isolation of the fraction containing cryptospirolepine (3)13
contained a C31 alkaloid that was given the notebook
designation TC-6. The original 400 MHz NMR data
showed the molecule to have severely congested 1H
and 13C NMR spectra, with significant potential for
ambiguities due to resonance overlap when trying to
establish connectivity networks and correct
resonance assignments during structure elucidation.
The molecule was fully aromatic with the exception of
an N-methyl group. The molecule gave a molecular
weight of 448 Da and HRMS data consistent with an
empirical formula of C31H20N4. LC/MS data were not
available at the time, and inexplicably for an aromatic
compound, successive losses of 14 Da were
observed in the direct insertion probe MS data that
were recorded, rendering the utility of these data
somewhat suspect.
The structure of cryptospirolepine (3) was
elucidated immediately prior to beginning work on the
TC-6 fraction.13 Undoubtedly, our thinking was biased
in the direction of a “complex” molecular structure
given that we were dealing with a 31 carbon alkaloid.
At the time, only indolo[3,2-b]quinolines such as
cryptolepine (1) and quindoline (2) and simple 11-
substituted analogues were known from C.
sanguinolenta.12 Other ring fusions such as the
indolo[3,2-c]- and –[2,3-b]quinolines, represented by
cryptotackieine (4) and cryptosanguinolentine (5),
respectively, would remain undiscovered for several
more years.14-17 Likewise, reports of the structural
characterization of dimeric and mixed dimeric alkaloids
such as cryptolepicarboline (7), cryptomisrine (8),
biscryptolepine (9), and cryptoquindoline (10) also would
not appear for several more years.15,16,18,19 These
biases and gaps in the knowledge of the diverse
range of alkaloids biosynthesized by C. sanguinolenta
were further exacerbated by the fact that the initial
data were acquired before the advent of gradient 2D
NMR experiments, meaning that there was a low
likelihood of observing longer range correlations in
the HMBC experiment (nJCH, n > 3). Additionally,
long-range 1H-15N heteronuclear shift correlation
experiments had not been developed, precluding the
use of these data in the structure elucidation process.


Elucidating the structure from the original 1991
NMR data.

Data available from the initial investigation
of this alkaloid fraction in 1991 included 1D 1H and
13C reference NMR spectra, and an ensemble of 2D
spectra that included non-gradient, COSY, ROESY,
HMQC, and HMBC spectra. As will be quickly noted
from even a cursory inspection of the 1D reference
spectra shown in Figure 2, there was considerable
overlap in both the proton and carbon spectra, that
could lead to significant numbers of ambiguities in
terms of reliable resonance assignments from the
available 2D NMR data. Nevertheless, these data
were used as a starting point for the CASE process.
Quaternary carbon resonances, which are
frequently used to link various substructural
fragments of a molecule via long-range correlations
observed in HMBC spectra, provide a useful
example of the type of ambiguities that are referred
to above. In a correctly assigned aromatic four-spin
system, 14 for example, the protons labeled HB and
HD both would be expected to exhibit long-range
correlations to quaternary carbon CA. In contrast, as
shown by 15, if the assignment of HB1 is permuted
with HB3 as shown, then HB3 and HD1 will not
correlate correctly to the CA1 quaternary carbon
resonance. Rather, they will correlate to two
different quaternary carbons, one of which, CA3, is
not associated correctly with this spin system,
creating ambiguity.
To illustrate how such ambiguities arise,
consider the paired COSY and HMQC spectra of
quindolinocryptotackieine (TC-6) shown in Figure 3.
As shown in the concerted interpretation of these
data, the proton/carbon resonant pair resonating
furthest upfield is correlated in the COSY spectrum
to a proton resonance at ~7.62 ppm. Unfortunately,
there are three protons substantially overlapped at

Vol. 1, Nos 1 and 2, 2003 Applications and Advances in Cryogenic NMR Probes and
Computer-Assisted Structure Elucidation

- 7 -
the point in the 400 MHz 1H spectrum, as shown by
the boxed region of the HMQC spectrum in the
figure. If there is not some way in which to resolve
which of the three carbons in question is associated
with the proton vicinal to the proton resonating at
~7.12 ppm that served as the starting point for this
example, then an ambiguity in the assignment is
created. While on the subject of these data, it is
also worth noting that the HMQC experiment, the
most commonly used heteronuclear shift correlation
experiment among the natural products chemistry
community, offers significant lower resolution than
can be obtained using the HSQC experiment. The
HSQC experiment is now standard in most pulse
sequence libraries and should be the standard used
nowadays. A similar statement can be made for the
comparison of the HMQC- and HSQC-TOCSY
experiments.37
Despite the possibility of assignment
ambiguities, to provide preliminary proton/carbon
resonant pair assignments for input into Structure
Elucidator, the COSY and HMQC spectra shown in
Figure 3 were used to establish the pairwise
assignments shown by 16 – 19.
Structure Elucidator uses a convention
analogous to the substructural fragments shown
above, called a Molecular Connectivity Diagram
(MCD). Connectivity across one or more bonds can
be shown in this construct, which also allows
the incorporation of quaternary carbon and
heteroatoms. The starting MCD for the elucidation
of the structure of quindolinocryptotackieine is
shown in Figure 4. Carbon-carbon single bonds are
denoted in by thick solid (bold) lines, correlations
across three bonds (3JCH) established from the
HMBC spectrum are shown by medium solid lines,
and correlations across more than three bonds
(nJCH, n >3) are denoted in grey. Working through
the data examining it for the possibility of
ambiguities leads to the MCD shown in Figure 5.
Ambiguous correlations are denoted in this diagram
by dashed lines and obviously a very complex and
confusing situation arises due to the numerous
potentially ambiguous assignments in the proton
and carbon spectral data.

CA3
HA1
HD1
HB3 CA1CA
HA
HB
HD

x14x x15x


8.88/119.64
8.23/133.85
7.76/129.19
7.86/125.65
7.57/114.86
7.85/135.95
7.59/123.39
8.86/127.21
x16x x17x

Martin et al Ann. Magn. Reson.
- 8 -
To resolve ambiguities in preparing the
data for structure elucidation, it is useful to start with
closely similar carbon resonance assignments, for
example the pair of carbons resonating at 129.19
and 129.06 ppm. As will be noticed from Figure 6,
these resonances are located in rings B and D,
respectively. After the pairwise resonance assignment
swap, the data are checked for ambiguities and,
in this case, those associated with this pair of
resonances has been resolved, completing the
assignment of the contiguous protonated carbons in
rings A, B, and C. Successive, pairwise resonance
assignment exchanges are made checking for
ambiguities after each swap until all ambiguities are
removed from the data set. In the case of this
problem, the process consumed approximately 48 hr
of investigator interaction with the program package,
finally affording the MCD shown in Figure 7.
Beginning from the MCD shown in Figure
7, the Structure Elucidator v6.08 program generated
355 structures within 10 sec, which reduced to 266
non-isomorphic structures after the program output
was filtered. Obviously, this task was accomplished
working from a well-defined and constrained MCD
that benefited substantially from investigator
familiarity with the indoloquinoline alkaloid family.
After structure generation, the 266 retained
structures were sorted on the basis of accurately
calculated 13C shifts vs. the observed 13C shifts (dA
13C). The first twelve sorted structures are shown in
Figure 8. Only the first four structures had dA 13C <
5 ppm. As discussed above, >85% of the structure
elucidation problems solved using Structure
Elucidator gave dA 13C < 4 ppm.25,26 Clearly at this
point, criteria that will allow the successful
identification of the correct structure from the sorted
output list must be defined.








Defining selection criteria to evaluate sorted
output

It is risky at best to assume that the
structure with the lowest dA13C value will
necessarily be the correct structure in a given
Structure Elucidator computational run
although frequently, the correct structure is
indeed first in the output list. On this basis,
the program was limited in terms of the data
provided to it to solely the 1D reference
spectra and the COSY, HMQC, and HMBC
data. The ROESY data were held back for use
as a potential selection criterion. Another
possible criterion for sorting through
structures is one based on mass spectral
fragmentation. Quindolinocryptotackieine gives
a parent ion, M+, and prominent ions associated
with the loss of a methyl group, which is not
particularly diagnostic, and major fragment ions
at 232, 231, and 217 Da that are diagnostically
useful. These alternative but complimentary
approaches to selecting the most likely
candidate from Structure Elucidator program
output are discussed in next page.

N
N
CH3 HH
H H


x20x
7.11/111.85
7.58/132.06
7.52/123.69
8.68/123.58
7.80/128.93
7.53/127.34
7.79/129.24
8.31/129.06


x18x x19x

Vol. 1, Nos 1 and 2, 2003 Applications and Advances in Cryogenic NMR Probes and
Computer-Assisted Structure Elucidation

- 9 -
Selection on the basis of NMR criteria.

As already noted, the ROESY data were
held in abeyance for use in evaluating molecules
generated by Structure Elucidator. Consider the
well-characterized molecule cryptolepine (20) in
terms of key ROE correlations.
The N-methyl group gives ROE correlation
responses to the H-4 and H-6 protonated aromatic
resonances flanking the N-methyl group in the
structure. The H-11 singlet aromatic proton exhibits
an ROE correlation only to H-1, the correlation in the
other direction is precluded by the indole annular
nitrogen. In comparison, quindolinocryptotackieine
exhibits single ROE correlations from both the
N-methyl and isolated aromatic proton
resonances. Based on this observation, any
molecule contained in the output of Structure
Elucidator, after sorting on the basis of dA13C,
that incorporates a cryptolepinyl moiety can be
excluded from consideration on the basis of
expected ROE or NOE correlation responses.
Referring to the sorted structures
shown in Figure 8, the first structure with the
lowest dA13C value of 3.540 ppm, is a
quindolinocryptolepine (21). This molecule can
be eliminated from consideration in that the
ROESY spectrum of this molecule, as just noted,
would be expected to exhibit two ROE
correlations from the N-methyl and a single
correlation from the isolated aromatic singlet,
which is contrary to the single ROE’s from
both structural subcomponents observed for
quindolinocryptotackieine.
The second molecule in the sorted list of
structures is quindolinocryptotackieine, withdA13C =
4.449. This molecule fits the ROE correlation
pattern required of the molecule being investigated,
with only one ROE correlation for both the N-methyl
group and the aromatic singlet. Consequently,
quindolinocryptotackieine (22) must be considered a
viable possibility for the identified structure of this molecule.
Continuing, the third structure in the sorted
list is a cryptolepinyl-desmethyl-cryptotackieine,
which gave a dA13C result of 4.570 ppm. There are
several features of this structure that facilitate its
exclusion from consideration. First, although by no
means implausible, N-desmethylcryptotackieine is
not a known biosynthetic product of C.
sanguinolenta. More importantly, both the N-methyl
and the isolated aromatic singlet in this structure
would be expected to exhibit two NOE or ROE
correlation responses, which would clearly
invalidate this as a structural candidate. The fourth


N
N
N
N
CH3
N
N
N
N
CH3

x21x x22x

Martin et al Ann. Magn. Reson.
- 10 -
and final structure in the output list with dA13C < 5
ppm is cryptoquindoline (10), a known compound.
Here again, the structure can be ruled out on the
basis of the two NOE or ROE correlations that
would be observed from the N-methyl group.
Thus, based on the preceding discussion, it
is possible to logically sift through the dA13C sorted
output list of Structure Elucidator to identify the
correct structure, which in this case was
quindolinocryptotackieine in the 2nd position of the
output with dA13C = 4.449 ppm. Although fused ring
structures were not positioned high enough in the
dA13C-sorted output of Structure Elucidator to be
considered as viable candidates for the structure of
TC-6, it is also worth considering the use of mass
spectral fragmentation as an orthogonal means of
sifting through the output of Structure Elucidator.
Efforts in this direction are presented in the following
section.

Selection on the basis of mass spectral
fragmentation.
An important criterion of any successful
structure elucidation effort is the establishment of
a molecular structure that can fragment in the
mass spectrometer in a manner consistent with
likely fragmentation pathways. In the case of
quindolinocryptotackieine (22), an MS/MS experiment
gave prominent fragment ions at m/z 232, 231, and
217, as well as an ion for M+-CH3. These fragment
ions, like NMR chemical shifts, also can provide the
basis for sorting the output of Structure Elucidator.
Using the program MS Fragmenter, the 266 non-
isomorphic structures generated by Structure
Elucidator were subjected to fragmentation and then
sorted on the basis of accurate 13C chemical shift
calculation (dA13C). The first 20 structures after this
operation are shown in Figure 9. The list of
structures with the respective fragment ions was
then sorted on the basis of whether or not the
molecule could generate fragment ions at m/z 232
and 217, consistent with the two major halves of the
molecule, cryptotackieine (5) and quindoline (2),
respectively. After this initial sort, a second sort
on the basis of dA13C was again performed.
This process afforded a list of structures, the first 12
of which are shown in Figure 10. As will be noted
from even a cursory comparison of the output lists
shown in Figures 8 and 10, while the former dA13C
sorted list had a number of fused ring structures, the
latter has none since none of the fused ring
structures generated by Structure Elucidator would
be capable of undergoing facile fragmentation to
give either the m/z 232 or 217 ions as major
fragment ions.
Using ROE correlations as a selection
criterion, quindolinocryptotackieine was again in the
2nd position of the sorted Structure Elucidator
output. It is important to have available to the
investigator alternative means of sorting and
evaluating the output of Structure Elucidator when
dealing with complex unknown structures. In the
present case, the data set sorted on the basis of
plausible mass spectral fragments was not
necessary, but it could be vital in other cases when
NMR criteria lead to equivocal results.

Advances in Cryogenic NMR Probe Technology
and Applications
As noted in the introduction, a variety of
strategies have been employed to circumvent the
inherent insensitivity of NMR to facilitate the study of
ever smaller samples, and/or to make it feasible
to conduct potentially very informative NMR
experiments that suffer from their own inherent lack
of sensitivity.5 In this regard, cryogenic NMR probe
technology has taken center stage in recent years
and is now a primary area of development effort on
the part of NMR instrument vendors, and the subject
of a recent review.38 Although there have been a
scant few applications of cryogenic NMR probe
technology in the elucidation of natural product
structures, that will undoubtedly change as access
to this new and highly sensitive probe technology
grows in the coming years.20, 39-42
Cryogenic NMR probes take advantage of
the substantial reduction in noise that can be
achieved by reducing the temperature of the probe
rf coils to near the temperature of liquid helium.
Additional benefits are obtained with a cold preamp

Vol. 1, Nos 1 and 2, 2003 Applications and Advances in Cryogenic NMR Probes and
Computer-Assisted Structure Elucidation

- 11 -
but the preamp is usually not cooled to the extremes
of the rf coil. The rf coil of the cryogenic NMR probe
currently in use in the author’s laboratory is
maintained at 25 K and typically the preamp is
around 50 K. Using this approach, gains in
sensitivity relative to an identical conventional
“warm” NMR will generally be a factor of 3X or
more. At a 3-fold increase in sensitivity, measuring
time to constant signal-to-noise ratio is decreased
by a factor of 9 relative to a conventional probe.
Experiments that are demonstrably facilitated
by the use of cryogenic NMR probes are
experiments such as 1H-15N long-range
heteronuclear shift correlation experiments,33,34
hyphenated 2D NMR experiments such as HSQC-
TOCSY, -NOESY, and –ROESY, which are
generally considered to have only a fraction of the
sensitivity of HMBC experiments.43 The latter
applications, HSQC-NOESY and –ROESY, have
not yet been reported with a cryogenic NMR probe
in the published literature; examples of the former is
presented below. For investigators with access to
cold carbon-detect cryogenic NMR probe technology,
an obvious benefit to improved sensitivity is found
when INADEQUATE spectra are measured.41 The
sensitivity of INADEQUATE is notoriously low and
sample requirements correspondingly high. By
using cryogenic NMR probe technology to deal with
such problems, sample requirements are somewhat
mitigated and measurement times are correspondingly
decreased.


Potential impacts of cryogenic NMR probe
technology on CASE strategies.
A popular experimental strategy for
structure elucidation frequently combines a
homonuclear correlation experiment such as COSY
or TOCSY with a direct heteronuclear shift
correlation experiment. Traditionally, the natural
products community has relied on the HMQC
experiment for this purpose although a number of
studies clearly demonstrate the superior resolution
that can be obtained by using HSQC.35-37,44,45
Irrespective of the heteronuclear experiment used, a
situation like that presented in Figure 3, in which
closely spaced carbon resonances can be
associated with a given proton, are not uncommon.
The pair of experiments shown in Figure 3 can be
replaced by an inverted direct response HSQC-
TOCSY spectrum as shown in Figure 11.37,45,46 As
discussed above, ambiguities of the type highlighted
by the COSY/HMQC data in Figure 3 complicate the
elucidation of chemical structures whether by a
human investigator or when using CASE methods.
The three possibilities for the 13C shift associated
with the vicinal neighbor proton shown in Figure 3
must be resolved before the elucidation of the
structure can progress. The data shown in Figure 3
were both acquired overnight using an ~2.5 mg
sample of quindolinocryptotackieine (22) in a
conventional 5 mm probe at 400 MHz. In
comparison, the inverted direct response HSQC-
TOCSY data shown in panel B) in Figure 11 were
acquired using ~1 mg of 22 in a 3 mm NMR tube
run coaxially in a 5 mm 500 MHz cryoprobe in ~4
hrs. These data serve to highlight the sensitivity
advantage inherent to cryoprobe technology and
obviate some of the ambiguities associated with the
more traditional 2D NMR experimental strategies.
Reynolds and Enriquez have recently reviewed the
topic experiment selection, processing strategies,
and related concepts germane to natural product
structure, to which the interested reader is
referred.48 While there is no way to circumvent the
ambiguities inherent to the COSY/HMQC combination
in Figure 3, the vicinal neighbor proton/carbon pair
is unequivocally identified in the HSQC-TOCSY
experiment shown in panel B) in Figure 11.


Sensitivity comparisons – 5 mm 500 MHz NMR
cryoprobe vs. conventional 3 mm 600 MHz
NMR data
Performance comparisons of new probe
technology vs. older better understood hardware are
always interesting. Several examples will be
discussed. Since the advantage inherent to using
HSQC-TOCSY data was already discussed, it is
perhaps appropriate to begin by comparing results

Martin et al Ann. Magn. Reson.
- 12 -
obtained with this experiment in conventional vs.
cryogenic NMR probes. The comparisons that
follow will all utilize 3 mm NMR samples that are
probably the best size for dealing with any
scarce sample. Again using a sample of ~1 mg
quindolinocryptotackieine (22) dissolved in 170 µL
of methanol-d4 in a 3 mm sample tube, 4 hr
inverted direct response HSQC-TOCSY data were
acquired using the same sample in both a 3 mm
gradient inverse triple resonance probe on a 600
MHz NMR spectrometer and the 5 mm cryoprobe in
a 500 MHz instrument.37,46,47 These data are shown
in Figure 12. As will be observed from the data in
panel A, the 600 MHz data gave a signal-to-noise
ratio of 8:1 in 4 hrs, as measured from the trace for
the carbon resonating at ~119 ppm plotted with the
contour plot. In contrast, in the same period of time,
despite the filling factor losses associated with
running a 3 mm sample coaxially in a 500 MHz 5
mm cryogenic NMR probe, data with a signal-to-
noise ratio of 16.7:1 were obtained. For even
modest samples, the combination of cryogenic NMR
probe technology and HSQC-TOCSY provides a
valuable means of dealing with highly congested
spectra in a less equivocal fashion. In past work,
we have shown that even when several resonances
in the same spin system are overlapped that it is
possible to disentangle the connectivity network by
using HSQC-TOCSY.49 Hopefully, as more
widespread access to cryogenic NMR probe
technology becomes available, more natural
products investigators will begin to utilize this less
sensitive but very powerful 2D NMR experiment.
Pulsed field gradients have seemingly
become ubiquitous in acquiring 2D NMR data. The
generally accepted belief is that it is always better to
use gradients if the spectrometer has that capability.
Unfortunately, that “truth” is quite incorrect. When
gradients are used in lieu of phase cycling to select
coherence pathways, the experiment is discarding
approximately half of the signal, and when that is a
very small signal from a scarce sample to begin
with, that puts the investigator at a severe
disadvantage. Generally speaking, if the sample
size is such that it will be necessary to run more
than the typical phase cycle for a given 2D NMR
experiment to generate adequate signal-to-noise
ratios for the investigation being conducted, an
investigator is better off using non-gradient phase-
cycled experiments. We have observed this
experimentally and this observation has also
recently been treated in the work of Reynolds and
Enriquez.50
Comparison non-gradient HMBC spectra
are shown in Figures 13 and 14 for a 19 µg sample
of the model alkaloid retrorsine (23) dissolved in 150
µL of methanol-d4 in a 3 mm sample tube. Identical
15 hr data acquisitions are shown in Figure 13. The
data acquired overnight for this sample using a 5
mm 500 MHz cryoprobe are shown in panel A. In
comparison, there are only very few responses that
would be reliably considered to be real in the
corresponding data acquired using a conventional 3
mm 500 MHz gradient inverse triple resonance
probe in panel B.
Figure 14 compares the overnight (15 hr)
cryogenic NMR probe data to a 120 hr acquisition in
a conventional 3 mm gradient inverse triple
resonance NMR probes. Despite what would
normally be considered to be an unacceptably long
acquisition, the conventional data (panel B) are still
at a disadvantage relative to the cryoprobe
data (panel A). Not only are not all of the responses
observed in the cryoprobe visible in the
conventional 120 hr data, but some of the legitimate
N
O
O
CH2
O
O
OH
OH
CH3H
H


x23x

Vol. 1, Nos 1 and 2, 2003 Applications and Advances in Cryogenic NMR Probes and
Computer-Assisted Structure Elucidation

- 13 -
responses in that spectrum are sufficient weak
(relative to the cryoprobe data) that a cautious
investigator might question the authenticity of the
response if they were dealing with an unknown
structure. Hence, while a sample of this size is still
quite manageable using cryogenic NMR probe
technology in reasonable periods of time (overnight
to a weekend), this sample could reasonably be
considered intractable using conventional 3 mm
NMR probe technology. Resorting to 1.7 mm
submicro NMR probe technology, if available, would
ameliorate the problem, but that probe format is by
no means routinely available. From a cost
consideration standpoint, 1.7 mm submicro NMR
probe technology might be a viable option for
laboratories faced with having to routinely deal with
small samples when faced by budgetary constraints;
1.7 mm probes cost about one fifth what a cryogenic
NMR probe would cost, and have no continued
maintenance costs such as those associated with
the closed loop cooling systems used with cryogenic
NMR probes.
Another NMR experiment that benefits
from the availability of cryogenic NMR probe
technology is in the acquisition of long-range 1H-15N
heteronuclear shift correlation data at natural
abundance. This area of NMR research has been
the subject of several recent reviews.33,34 Given
that >80% of all pharmaceuticals contain nitrogen in
addition to many natural products, it is obvious that
there will be continued interest in exploiting 15N, an
insensitive, low natural abundance nuclide that is
not generally amenable to direct observation except
when very large samples are available to an
investigator. A comparison of the acquisition of
long-range 1H-15N 2D NMR data using a
conventional 3 mm gradient inverse triple resonance
NMR probe and a 5 mm 500 MHz Chiliprobe using
a 2 mg sample of the oxazolidinone antibiotic
eperezolid (24) dissolved in 150 µL of acetonitrile-d3
in a 3 mm NMR tube was recently reported.31 These
results are shown in Figure 15. The contour plot
displayed in panel A show the results of a 4 hr
acquisition in the conventional 3 mm 500 MHz
gradient inverse triple resonance probe. The signal-
to-noise ratio of the projection through F1 (15N
frequency domain) was 49:1. The corresponding F1
projection is shown in panel C. In comparison, the
results of a 26 min experiment are shown in panel
B. The signal-to-noise ratio of the projection of these
data, shown in panel D, was 101:1.


Application of cryoprobe technology for the
acquisition of long-range 1H-15N data for
quindolinocryptotackieine (22)

One of the problems inherent to long-range
1H-15N experiments when performed on molecules
such as the indoloquinoline alkaloids is the very
small size of the long-range 1H-15N couplings to the
nitrogen resonances. In the case of cryptolepine
itself, these coupling constants have been
measured and mandate a long-range optimization in
the range of 2.5-3 Hz if there is to be any chance of
successfully observing the long-range correlations


109.2
53.8
100.0112.6
6"
4"
3" 2"
1"
5' 6'
3' 2'
1'
97
5
3 1
F
NN O
O
NH
CH3
ON
OH
O

x24x

Martin et al Ann. Magn. Reson.
- 14 -
sought in the experiment.51 For a sample of
cryptospirolepine (3, 750 µg) acquiring these data
using a 1.7 mm submicro probe at 600 MHz
consumed a significant amount of spectrometer
time.52,53 In part this was due to the small size
of the sample time. There are, however, also
considerable signal losses associated with the
length of the magnetization transfer delays when the
experiment is optimized for very small couplings.
Consequently, the advantages of acquiring these
data using a cryogenic NMR probe are obvious.
Figure 16 shows the results of an overnight
acquisition of a 3 Hz gradient HMBC spectrum for a
sample of ~1 mg of quindolinocryptotackieine (22) in
a 3 mm NMR tube. Correlations to three of the four
annular nitrogens in the structure of this alkaloid are
observed in the experiment. Acquiring comparable
data using a conventional 3 mm probe at 500 or 600
MHz probably would have consumed about a week
of spectrometer time, which would preclude the
acquisition of these data in most laboratories.


15N Chemical Shift Range Predictions

With the much greater sensitivity of
cryogenic NMR probes and the consequent
improvement in our ability to acquire long-range
1H-15N chemical shift correlation data, more
individuals will have access to this valuable
structural information as cryogenic NMR probe
technology becomes more widespread. Typical
chemical shift ranges for various types of 15N
resonances are not as intuitively obvious to the
average investigator at present as a 13C chemical
shift would be since most investigators have not had
the opportunity to develop any degree of familiarity
with 15N chemical shift data. In this regard, CASE
programs also have a valuable application in an
investigator’s ability to use CASE program databases
to predict the 15N chemical shifts of a new class of
molecules. These chemical shift predictions can, in
turn, be used to establish the F1 spectral windows for
the acquisition of the long-range 15N chemical shift
correlation spectra.
Using the approach just described,
15N chemical shifts can be predicted for
quindolinocryptotackieine (22) using theACD/NNMR
prediction program,which is an additional module
that can be integrated into the Structure Elucidator
program package. Predicted shifts ranged from 317
to 100 ppm; the actual chemical shift range for 22
was 107-278 ppm. The F1 chemical shift range of
80-320 ppm used to acquire the long-range 1H-15N
data shown in Figure 16 was selected based on the
author’s familiarity with the 15N chemical shift
behavior in this series of alkaloids. When faced with
a molecule from a class of compounds with which
an investigator has no prior experience, using the
predicted 15N chemical shift range from NNMR and
the expanding that range by about 20 ppm on each
end to allow for prediction errors should afford a
reasonable F1 window for the long-range 1H-15N
experiment in which all of the resonances will likely
be observed without having to worry about F1
folding. The NNMR can also be trained with
experimental data obtained by a laboratory thereby
providing feed-forward effects for future predictions
and reducing the opportunities for errors in setting
the spectral width.

Conclusions

Significant advances have been made in
the development of both Computer-Assisted
Structure Elucidation (CASE) program packages
and cryogenic NMR probe technology.
We have shown the application of the former
in the elucidation of the structure of a complex,
indoloquinoline mixed dimeric alkaloid,
quindolinocryptotackieine (22). Because of significant
resonance overlaps in both the proton and carbon
spectra, there were numerous chemical shift
assignment ambiguities that precluded the
assignment of the structure of this molecule by
human investigators for more than a decade.
However, using the Structure Elucidator v6.08
software package, it was possible after approximately
48 hrs of investigator effort to successively resolve
and remove pairwise assignment exchanges, to

Vol. 1, Nos 1 and 2, 2003 Applications and Advances in Cryogenic NMR Probes and
Computer-Assisted Structure Elucidation

- 15 -
establish a molecular connectivity diagram (MCD)
that allowed the successful solution of the structure.
Strategies for selecting the correct structure from
the Structure Elucidator program output have been
described, as has a new approach to sorting data
based on the MS Fragmenter program.
In parallel with the advances in CASE
programs, cryogenic NMR probes are beginning
to provide unprecedented experimental access to
data that might otherwise be time prohibitive
for some samples. Using a reisolate of
quindolinocryptotackieine (22) an example was
shown of the use of inverted direct response HSQC-
TOCSY to circumvent one type of assignment
ambiguity problem. These data compliment the
utilization of CASE programs by reducing investigator
time for ambiguity resolution. We have also shown
examples of the increased efficiency of using
cryoprobes for the acquisition of data on extremely small
samples, and the acquisition of long-range 1H-15N
heteronuclear shift correlation data. In a complimentary
fashion, the 15N prediction capability of the NNMR
program component of the Structure Elucidator v6.08
program package can be used to predict 15N chemical
shift ranges for new molecules, facilitating the selection
of appropriate F1 spectral windows.
In conclusion, CASE programs supplemented
with the significantly enhanced sensitivity of cryogenic
NMR probe technology, are quite complimentary and
will likely play a significant role in the elucidation of new,
complex molecular structures.


Acknowledgements:

The authors would like to acknowledge: R.
C. Crouch, Varian NMR Instruments, Palo Alto, CA;
E. R. Martirosian, S. G. Molodtsov, V. Lashin, I.
Troitskiy, and M. Bayliss of Advanced Chemistry
Development and the Novosibirsk Institute of Organic
Chemistry; Professor P. L. Schiff, Jr. of the University
of Pittsburgh School of Pharmacy, Pittsburgh, PA;
and A. N. Tackie, Centre for Research into Plant
Medicine, Mampong-Akwapim, Ghana.

References

1. Odelblad, E.; ‘Nordisk Forening for Obsterik
och Gynekologi,’ Karolinska Institute,
Stockholm, 1966.
2. Shoolery, J. N.; Small Coils for NMR
Microsamples; in Topics in Carbon-13 NMR
Spectroscopy, Levy, G. C., Ed.; Wiley, New
York, 1979.
3. Crouch, R. C.; Martin, G. E.; J. Nat. Prod.
1992, 55, 1343-1347.
4. Crouch, R. C.; Martin, G. E.; Magn. Reson.
Chem., 1992, 30, S66-S70.
5. Martin, G. E.; Microprobes and Methodologies
for Spectral Assignments: Applications; in
Encyclopedia of Nuclear Magnetic Resonance,
vol 9. Advances in NMR, Grant, D. M. and
Harris, R. K., Eds.; Wiley & Sons, Ltd.,
Chichester, 2002.
6. Fichter, F.; Boehringer, R.; Chem. Ber. 1906,
39, 3932-3940.
7. Gellert, E.; Raymond-Hammet; Schlittler, E.;
Helv. Chim. Acta, 1951, 34, 642-651.
8. Dwuma-Badu, D.; Ayim, J. S. K.; Fiagbe, N. I.
Y.; Knapp, J. E.; Schiff, P. L., Jr.; J. Pharm.
Sci., 1978, 67, 433-434.
9. Ablordeppey, S. Y.; Hufford, C. D.; Bourne, R.
F.; Dwuma-badu, D.; Planta Med., 1990, 56,
416-417.
10. Tackie, A. N.; Sharaf, M. H. M.; Schiff, P. L.,
Jr.; Crouch, R. C.; Martin, G. E.; J.
Heterocyclic Chem., 1991, 28 1429-1435.
11. Spitzer, T. D.; Crouch, R. C.; Martin, G. E.;
Sharaf, M. H. M.; Schiff, P. L., Jr.; Tackie, A.
N.; Boye, G. L.; J. Heterocyclic Chem., 1991,
28, 2065-2070.
12. Sharaf, M. H. M.; Alkaloids of Cryptolepis
sanguinolenta (Lindl.) Schlechter (Asclepiadaceae);
Ph.D. Dissertation, University of Pittsburgh,
School of Pharmacy, 1993, 220 pp.
13. Tackie, A. N.; Boye, G. L.; Sharaf, M. H. M.;
Schiff, P. L., Jr.; Crouch, R. C.; Spitzer, T. D.;
Johnson, R. L.; Dunn, J.; Minick, D.; Martin, G.
E.; J. Nat. Prod., 1993, 56, 653-670.

Martin et al Ann. Magn. Reson.
- 16 -
14. Sharaf, M. H. M.; Schiff, P. L., Jr.; Tackie, A.
N.; Phoebe, C. H., Jr.; Martin, G. E.; J.
Heterocyclic Chem., 1996, 33, 239-243.
15. Cimanga, K.; De Bruyne, T.; Pieters, L.;
Claeys, M.; Vlietinck, A.; Tetrahedron Lett.,
1996, 37, 1703-1706.
16. Paulo, A.; Gomes, E. T.; Houghton, P. J.; J.
Nat. Prod., 1995, 58, 1485-1491.
17. Sharaf, M. H. M.; Schiff, P. L., Jr.; Tackie, A.
N.; Davis, A. O.; Andrews, C. W.; Crouch, R.
C.; Martin, G. E.; J. Heterocyclic Chem., 1996,
32, 1631-1636.
18. Sharaf, M. H. M.; Schiff, P. L., Jr.; Tackie, A.
N.; Phoebe, C. H., Jr.; Howard, L.; Meyers, C.;
Hadden, C. E.; Wrenn, S. K.; Davis, A. O.;
Andrews, C. W.; Minick, D.; Johnson, R. L.;
Shockcor, J. P.; Crouch, R. C.; Martin, G. E.;
Magn. Reson. Chem., 1995, 33, 767-778.
19. Sharaf, M. H. M.; Schiff, P. L., Jr.; Tackie, A.
N.; Phoebe, C. H., Jr.; Johnson, R. L.; Minick,
D.; Andrews, C. W.; Crouch, R. C.; Martin, G.
E.; J. Heterocyclic Chem., 1996, 33, 789-797.
20. Martin, G. E.; Hadden, C. E.; Russell, D. J.;
Kaluzny, B. D.; Guido, J. E.; Duholke, W. K.;
Stiemsma, B. A.; Thamann, T. J.; Crouch, R. C.;
Blinov, K.; Elyashberg, M.; Martirosian, E. R.;
Molodtsov, S. G.; Williams, A. J.; Schiff, P. L., Jr.;
J. Heterocyclic Chem., 2002, 39, 1241-1250.
21. Cooper, M. M.; Lovell, J. A.; Joule, J. A., Tet.
Lett.; 1996, 37, 4283-
22. Sharaf, M. H. M.; Schiff, P. L., Jr.; Tacki, A. N.;
Martin, G. E.; J. Heterocyclic Chem., 1998, 35,
1365-1369.
23. Hadden, C. E.; Duholke, W. K.; Guido, J. E.;
Robins, R. H.; Martin, G. E.; Sharaf, M. H. M.;
Schiff, P. L., Jr.; J. Heterocyclic Chem., 1999,
36, 525-531.
24. Martin, G. E.; Hadden, C. E.; Blinn, J. R.;
Sharaf, M. H. M.; Tackie, A. N.; Schiff, P. L.,
Jr.; Magn. Reson. Chem., 1999, 37, 1-6.
25. Elyashberg, M. E.; Blinov, K. A.; Williams, A.
J.; Martirosian, E. R.; Molodtsov, S. G.; J. Nat.
Prod., 2002, 65, 693-703.

26. Blinov, K. A.; Carlson, D.; Elyashberg, M. E.;
Martin, G. E.; Martirosian, E. R.; Molodtsov, S.
G.; Williams, A. J.; Magn. Reson. Chem.,
2003, 41, 359-372.
27. Elyashberg, M. E.; Blinov, K. A.; Williams, A.
J.; Molodtsov, S. G.; Martin, G. E.; Martirosian,
E. R.; J. Comp. Inf. Comp. Sci., 2003,
submitted for publication.
28. Molodtsov, S. G.; Elyashberg, M. E.; Blinov, K.
A.; Williams, A. J.; Martirosian, E. R.; J. Comp.
Inf. Comp. Sci., 2003, manuscript in
preparation.
29. Martin, G. E.; Exploitation of Long-range
Heteronuclear Coupling Constants – New
Techniques for Long-Range Heteronuclear
Shift Correlation Spectroscopy ; in Ann. Rep.
NMR Spectrosc.; Webb, G. A., Ed., Academic
Press, New York, 2002, 46, 37-100.
30. Martin, G. E.; Hadden, C. E.; Magn. Reson.
Chem., 2000, 38, 251-256.
31. Crouch, R. C.; Llanos, W.; Mehr, K. G.;
Hadden, C. E.; Russell, D. J.; Martin, G. E.;
Magn. Reson. Chem., 2001, 39, 555-558.
32. Kline, M.; Cheatham, S.; Magn. Reson.
Chem., 2003, 41, 307-314.
33. Martin, G. E.; Hadden, C. E.; J. Nat. Prod.,
2000, 63, 543-585.
34. Marek, R.; Ly…ka, A.; Curr. Org. Chem., 2002,
6, 35-66.
35. Martin, G. E.; Crouch, R. C.; Inverse-detected
2D-NMR applications in alkaloid chemistry; in
Modern Methods of Plant Analysis; Linskens,
H. F.; Jackson, J. F., Eds.; Springer-Verlag,
Berlin, 1994, 15, 25-89.
36. Reynolds, W. F.; McLean, S.; Tay, L.-L.;
Yu, M.; Enriquez, R. G.; Estwick, D. M.;
Pascoe, K. O.; Magn. Reson. Chem.,
1997, 35, 455-462.
37. Hadden, C. E.; Martin, G. E.; Luo, J.-K.;
Castle, R. N.; J. Heterocylic Chem., 1999,
36, 533-539.

Vol. 1, Nos 1 and 2, 2003 Applications and Advances in Cryogenic NMR Probes and
Computer-Assisted Structure Elucidation

- 17 -
38. Martin, G. E.; Cryogenic NMR Probes:
Applications; in Encyclopedia of Nuclear
Magnetic Resonance, vol. 9. Advances in
NMR, Grant, D. M. and Harris, R. K., Eds.;
Wiley & Sons, Ltd., Chichester, 2002.
39. Logan, T. M.; Murali, N.; Wang, G.; Jolivet, C.;
Magn. Reson. Chem., 1999, 37, 762-765.
40. Russell, D. J.; Hadden, C. E.; Martin, G. E.;
Gibson, A. A.; Zens, A. P.; Carolan, J. L.; J.
Nat. Prod., 2000, 63, 1047-1049.
41. Bringmann, G.; Wohlfarth, M.; Rischer, H.;
Grüne, M.; Schlauer, J.; Angew. Chem., Intl.
Edn., Engl., 2000, 35, 1464-1466.
42. Rashid, M. A.; Gustafson, K. R.; Crouch, R. C.;
Groweiss, A.; Pannell, L. K.; Van, Q. N.; Boyd,
M. R.; Organic Lett., 2002, 4, 3293-3296.
43. Martin, G. E.; Hadden, C. E.; Russell, D. J.;
Solution NMR Spectroscopy, in Handbook of
Spectroscopy, in Gauglitz: Handbook of
Spectroscopy, vol. 1, Wiley-VCH, New York,
2003, pp 209-268.
44. Eastwick, D. M.; Pascoe, K. O.; Magn. Reson.
Chem., 1997, 35, 455-462.
45. Claridge, T. D. W. High Resolution NMR
Techniques in Organic Chemistry, Pergammon,
Amsterdam, 1999, pp. 229-230.
46. Domke, T.; J. Magn. Reson., 1991, 95, 174-177.
47. Crouch, R. C.; Davis, A. O.; Martin, G. E.;
Magn. Reson. Chem., 1995, 33, 889-892.
48. Reynolds, W. F.; Enriquez, R. G.; J. Nat.
Prod., 2002, 65, 221-244.
49. Hadden, C. E.; Duholke, W. K.; Guido, J. E.;
Robins, R. H.; Martin, G.E.; Sharaf, M. H. M.;
Schiff, P. L., Jr.; J. Heterocyclic Chem., 1999,
36, 525-531.
50. Reynolds, W. F.; Enriquez, R. G.; Magn.
Reson. Chem., 2001, 39, 531-538.
51. Martin, G. E.; Crouch, R. C.; Sharaf, M. H. M.;
Schiff, P. L., Jr.; J. Nat. Prod., 1996, 59, 2-4.
52. Martin, G. E.; Hadden, C. E.; Tackie, A. N.;
Sharaf, M. H. M.; Schiff, P. L., Jr.; Magn.
Chem. Reson., 1999, 37, 529-537.
53. Hadden, C. E.; Martin, G. E.; Tackie, A. N.;
Schiff, P. L., Jr.; J. Heterocyclic Chem., 1999,
36, 1115-1117.

Martin et al Ann. Magn. Reson.
- 18 -





































Figure 1. Structures generated by Structure Elucidator v6.08 during the elucidation of the structure of
cryptoquindolinone (11). No constraints were imposed on the program, which generated ~3,300
structures in 8 hr of computer time, which reduced to 355 structures after the removal of duplicates.
Sorting that list on the basis of accurately calculated 13C shifts (dA) gave the results shown here. The
correct structure is in the 2nd position of the output list with dA 13C = 3.947. It should be noted,
however, that on the basis of calculated 1H shifts, this structure would be ranked first.

Vol. 1, Nos 1 and 2, 2003 Applications and Advances in Cryogenic NMR Probes and
Computer-Assisted Structure Elucidation

- 19 -

A





B


Figure 2. 1D 1H and 13C reference spectra of TC-6 recorded at a proton observation
frequency of 400 MHz in d4-methanol.

Martin et al Ann. Magn. Reson.
- 20 -


Figure 3. COSY/HMQC composite presentation. Interpreting the data beginning with the proton/carbon
resonant pair furthest upfield shows that the proton resonating at ~7.12 ppm is vicinally correlated to
a proton resonating at ~7.62 ppm, which is overlapped with two other protons. Following the
correlation back into the HMQC spectrum, if the wrong 13C resonance in the boxed region of the
spectrum is assigned to the vicinal neighbor proton at ~7.62 ppm, ambiguity is introduced into the
spectral assignment, which can in turn lead to difficulties in correctly assembling the structure of an
unknown molecule.

Vol. 1, Nos 1 and 2, 2003 Applications and Advances in Cryogenic NMR Probes and
Computer-Assisted Structure Elucidation

- 21 -
CH3(43.10)(5.28)
CH
(114.86)(7.57)
C
116.30
CH
(119.64)(8.88)
CH
(123.39)(7.59)
C
125.52CH
(125.65)(7.86)
CH
(127.21)(8.86)
C
130.02
CH
(133.85)(8.23)
CH
(135.95)(7.85)
C
138.49
C
143.16
C
148.95
CH
(111.85)(7.11)
CH
(115.79)(7.90)
CH
(123.58)(8.68)
CH
(123.69)(7.52)
CH
(127.34)(7.53)
C
128.48
CH
(129.06)(7.80)(8.31)
CH
(132.06)(7.58)
C
135.33
C
145.87
C
146.07
C
147.43
C
123.76
CH
(128.93)(7.80)(8.31)
CH
(129.19)(7.76)(7.79)
CH
(129.24)(7.76)(7.79)
C
134.19
N
N
N
N
A B
C D


Figure 4. Starting Molecular Connectivity Diagram (MCD) for the elucidation of the structure of
quindolinocryptotackieine using Structure Elucidator v6.08.
CH3(43.10)(5.28)
CH
(111.85)(7.11)
CH
(114.86)(7.57)
CH
(115.79)(7.90)
C
116.30
CH
(119.64)(8.88)
CH
(123.39)(7.59)
CH
(123.58)(8.68)
CH
(123.69)(7.52)
C
123.76
C
125.52CH
(125.65)(7.86)
CH
(127.21)(8.86)
CH
(127.34)(7.53)
C
128.48
CH
(128.93)(7.80)(8.31)
CH
(129.06)(7.80)(8.31)
CH
(129.19)(7.76)(7.79)
CH
(129.24)(7.76)(7.79)
C
130.02
CH
(132.06)(7.58)
CH
(133.85)(8.23)
C
135.33
CH
(135.95)(7.85)
C
138.49
C
143.16
C
145.87
C
146.07
C
147.43
C
148.95
C
134.19
N
N
N
N
A B
C D


Figure 5. MCD showing possibly ambiguous assignments pairwise by dashed lines. Obviously, from the
number of potentially ambiguous assignments shown, this structure elucidation problem is a
very complex and challenging one. In part, the number of potential permutations in the initial
assignments helps to explain part of the difficulty in initial attempts to assign the structure of this
molecule when the study of it was first undertaken in 1991.

Martin et al Ann. Magn. Reson.
- 22 -

CH3(43.10)(5.28)
CH
(114.86)(7.57)
C
116.30
CH
(119.64)(8.88)
CH
(123.39)(7.59)
C
125.52CH
(125.65)(7.86)
CH
(127.21)(8.86)
CH
(129.19)(7.79) C
130.02
CH
(133.85)(8.23)
CH
(135.95)(7.85)
C
138.49
C
143.16
C
148.95
CH
(111.85)(7.11)
CH
(115.79)(7.90)
CH
(123.58)(8.68)
CH
(123.69)(7.52)
CH
(127.34)(7.53)
C
128.48
CH
(128.93)(7.80)(8.31)
CH
(129.06)(7.80)(8.31)
CH
(129.24)(7.76)
CH
(132.06)(7.58)
C
135.33
C
145.87
C
146.07
C
147.43
C
123.76
C
134.19
N
N
N
N
A B
C D

Figure 6. Molecular connectivity diagram (MCD) showing the starting point for the pairwise
ambiguity resolution. The process was begun with the pair of carbons resonating at
129.19 and 129.06 ppm located in rings B and D, respectively. In this case, the protons
associated with the carbons must be exchanged to correct the assignment.
CH3(43.10)(5.28)
CH
(114.86)(7.57)
C
116.30
CH
(119.64)(8.88)
CH
(123.39)(7.59)
C
125.52CH
(125.65)(7.86)
CH
(127.21)(8.86)
CH
(129.19)(7.79) C
130.02
CH
(133.85)(8.23)
CH
(135.95)(7.85)
C
138.49
C
143.16
C
148.95
N
CH
(111.85)(7.11)
CH
(115.79)(7.90)
CH
(123.58)(8.68)
CH
(123.69)(7.52)
C
123.76
CH
(127.34)(7.53)
C
128.48
CH
(128.93)(7.80)
CH
(129.06)(8.31)
CH
(129.24)(7.76)
CH
(132.06)(7.58)
C
135.33
C
145.87
C
146.07
C
147.43
C
134.19
N
N
N
ROESY 4JCH
A B
C D


Figure 7. Final MCD obtained by the removal of all chemical shift assignment ambiguities in a
pairwise fashion. The significant number of assignment perturbations due to the
highly overlapped proton and carbon spectra was a significant factor in preventing the
determination of the structure when the data were first recorded.

Vol. 1, Nos 1 and 2, 2003 Applications and Advances in Cryogenic NMR Probes and
Computer-Assisted Structure Elucidation

- 23 -



CH3
N
N
N
N
dA(
13C): 3.540 (4.507)
dA(
1H): 0.351 (0.507)
1 (ID:86)
CH3
N
N
N
N
dA(
13C): 4.449 (6.855)
dA(
1H): 0.381 (0.553)
2 (ID:85)
CH3
N
N
N N
dA(
13C): 4.570 (6.296)
dA(
1H): 0.375 (0.565)
3 (ID:36)
CH3
N
N
N
N
dA(
13C): 4.793 (6.089)
dA(
1H): 0.406 (0.544)
4 (ID:334)
CH3
N
N
N
N
dA(
13C): 5.342 (8.180)
dA(
1H): 0.415 (0.617)
5 (ID:35)
CH3
N
N
N
N
dA(
13C): 5.385 (7.437)
dA(
1H): 0.509 (0.684)
6 (ID:25)
CH3
N
N
N
N
dA(
13C): 5.424 (7.277)
dA(
1H): 0.566 (0.814)
7 (ID:232)
CH3
N
N
N
N
dA(
13C): 5.442 (7.451)
dA(
1H): 0.492 (0.659)
8 (ID:92)
CH3
N
N
N
N
dA(
13C): 5.485 (6.919)
dA(
1H): 0.703 (0.996)
9 (ID:41)
CH3
N
N
N
N
dA(
13C): 5.487 (7.038)
dA(
1H): 0.377 (0.533)
10 (ID:179)
CH3
N
N
N
N
dA(
13C): 5.676 (7.254)
dA(
1H): 0.675 (0.980)
11 (ID:84)
CH3
N
NN
N
dA(
13C): 5.679 (7.714)
dA(
1H): 0.573 (0.797)
12 (ID:231)


Figure 8. The first twelve structures from the Structure Elucidator program output sorted on the basis of
increasing dA13C. Experience has demonstrated that in > 80% of cases, the correct structure
exhibits a dA13C < 4 ppm and in the present case, only four of the twelve structures have dA13C
values < 5 ppm.

Martin et al Ann. Magn. Reson.
- 24 -

Figure 9. Structure elucidator output (266 non-isomorphic structures) was subjected to mass spectral
fragmentation using the MS Fragmenter program. After fragmentation, the output list was next sorted
on the basis of accurate 13C chemical shift calculation, dA13C. The first 20 structures in the output list
after sorting on the basis of dA13C are shown. Cursory inspection shows that in addition to mixed
“dimeric” indoloquinoline structures, fused ring structures are contained in the output list, some with
quite low dA13C values, e.g. structure number 2. The third tabulated value for each molecule is d(MS).
Molecules in the sorted list with d(MS) = 0.905 are indoloquinoline-derived that met the fragmentation
criteria of 232, 231, and 217 Da. Molecules with d(MS) = 0.751 are generally fused ring analogues that
would be able to partially met the observed fragment pattern of quindolinocryptotackieine (22).
CH3
N
N
N
N
dA(
13C): 3.540 (4.507)
dF(
13C): 5.703 (6.914)
d(MS): 0.905
1 (ID:86)
CH3
N
N
N
N
dA(
13C): 4.111 (6.831)
dF(
13C): 4.036 (5.783)
d(MS): 0.751
2 (ID:191)
CH3
N
N
N
N
dA(
13C): 4.449 (6.855)
dF(
13C): 5.189 (7.377)
d(MS): 0.905
3 (ID:85)
CH3
N
N
N N
dA(
13C): 4.570 (6.296)
dF(
13C): 6.181 (7.982)
d(MS): 0.905
4 (ID:36)
CH3
N
N
N
N
dA(
13C): 4.735 (7.327)
dF(
13C): 4.166 (6.130)
d(MS): 0.751
5 (ID:187)
CH3
N
N
N
N
dA(
13C): 4.793 (6.089)
dF(
13C): 6.073 (7.424)
d(MS): 0.905
6 (ID:334)
CH3
N
N
N
N
dA(
13C): 5.252 (6.930)
dF(
13C): 5.860 (7.345)
d(MS): 0.751
7 (ID:43)
CH3
N
N
N
N
dA(
13C): 5.263 (7.437)
dF(
13C): 5.478 (7.517)
d(MS): 0.751
8 (ID:88)
CH3
N
N
N
N
dA(
13C): 5.342 (8.180)
dF(
13C): 5.645 (8.486)
d(MS): 0.905
9 (ID:35)
CH3
N
N N
N
dA(
13C): 5.348 (8.056)
dF(
13C): 5.055 (6.781)
d(MS): 0.751
10 (ID:230)
CH3
N
N
N
N
dA(
13C): 5.385 (7.437)
dF(
13C): 5.704 (7.960)
d(MS): 0.751
11 (ID:25)
CH3
N
N
N
N
dA(
13C): 5.396 (6.954)
dF(
13C): 4.980 (6.275)
d(MS): 0.751
12 (ID:203)
CH3
N
N
N
N
dA(
13C): 5.424 (7.277)
dF(
13C): 5.496 (7.263)
d(MS): 0.751
13 (ID:232)
CH3
N
N
N
N
dA(
13C): 5.442 (7.451)
dF(
13C): 6.371 (9.112)
d(MS): 0.905
14 (ID:92)
CH3
N
N
N
N
dA(
13C): 5.485 (6.919)
dF(
13C): 6.405 (7.624)
d(MS): 0.751
15 (ID:41)
CH3
N
N
N
N
dA(
13C): 5.487 (7.038)
dF(
13C): 5.658 (7.071)
d(MS): 0.905
16 (ID:179)
CH3
N N
N
N
dA(
13C): 5.589 (7.484)
dF(
13C): 5.191 (7.139)
d(MS): 0.751
17 (ID:198)
CH3
N
N
N
N
dA(
13C): 5.676 (7.254)
dF(
13C): 6.083 (7.563)
d(MS): 0.751
18 (ID:84)
CH3
N
N
N
N
dA(
13C): 5.678 (8.559)
dF(
13C): 5.421 (8.333)
d(MS): 0.751
19 (ID:350)
CH3
N
NN
N
dA(
13C): 5.679 (7.714)
dF(
13C): 5.266 (7.129)
d(MS): 0.751
20 (ID:231)

Vol. 1, Nos 1 and 2, 2003 Applications and Advances in Cryogenic NMR Probes and
Computer-Assisted Structure Elucidation

- 25 -


CH31
2
3
4 5
6
7
8
9
10
1112
1314
15
16
17
18
19 20
21
22
23
24
25
26
27
28
2930
31N
32
N
33
N
34 N
35
dA(
13C): 3.540 (4.507)
dA(
1H):
dF(
13C): 5.703 (6.914)
1 (ID:76)
CH31
2
3
4
5
6
7
8
9
10
1112
13
14
15
16
17
18
19 20
21
22
23
24
25
26
27
28
2930
31N
32
N
33
N
34
N
35
dA(
13C): 4.449 (6.855)
dA(
1H):
dF(
13C): 5.189 (7.377)
2 (ID:75)
CH31
2
3
4
5
6
7
8
9
10
11
12
13
14
1516
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
N
32
N
33
N
34
N
35
dA(
13C): 4.570 (6.296)
dA(
1H):
dF(
13C): 6.181 (7.982)
3 (ID:36)
CH3 1
23
4
5
6
7
8
9
10
11 12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
N
32
N
33
N
34
N
35
dA(
13C): 4.793 (6.089)
dA(
1H):
dF(
13C): 6.073 (7.424)
4 (ID:235)
CH31
2
3
4
5
6
7
8
9
10
11
12
13
14
1516
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
N
32
N
33
N
34
N
35
dA(
13C): 5.342 (8.180)
dA(
1H):
dF(
13C): 5.645 (8.486)
5 (ID:35)
CH31
2
3
4
5
6
7
8
9
10
1112
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
2930
31
N
32
N
33
N
34
N
35
dA(
13C): 5.442 (7.451)
dA(
1H):
dF(
13C): 6.371 (9.112)
6 (ID:78)
CH31
2
3
4
5
6
7
8
9
10
1112
13
14
1516
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
N
32
N
33
N
34 N
35
dA(
13C): 5.487 (7.038)
dA(
1H):
dF(
13C): 5.658 (7.071)
7 (ID:161)
CH31
2
3
4
5
6
7
8
9
10
11
12
13
14
1516
17
18
19
20
21
22
23
24
2526
27
28
29
30
31
N
32
N
33
N
34
N
35
dA(
13C): 6.156 (7.767)
dA(
1H):
dF(
13C): 5.757 (8.669)
8 (ID:163)
CH31
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
2223
24
25
26
27
28
2930
31
N
32 N
33
N
34
N
35
dA(
13C): 6.225 (9.703)
dA(
1H):
dF(
13C): 6.084 (8.439)
9 (ID:80)
CH31
2
3
4
5
6
7
8
9
10
11
12
13
14
1516
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
N
32
N
33
N
34
N
35
dA(
13C): 6.266 (8.347)
dA(
1H):
dF(
13C): 6.555 (9.627)
10 (ID:38)
CH31
2
3
4
5
6
7
8
9
10
1112
1314
1516
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
N
32
N
33
N
34
N
35
dA(
13C): 6.415 (8.677)
dA(
1H):
dF(
13C): 5.285 (7.421)
11 (ID:160)
CH3 1
2
3
4
5
6
7
8
9
10
11
12
13
14
1516
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
N
32
N
33
N
34
N
35
dA(
13C): 6.554 (8.559)
dA(
1H):
dF(
13C): 6.309 (8.375)
12 (ID:37)
Figure 10. Structure output of Structure Elucidator subject to fragmentation using MS Fragmenter and then
sorted on the presence of fragment ions at m/z 232 and 217. After sorting on the basis of the selected
fragment ions, the list was reordered on the basis of increasing dA13C.

Martin et al Ann. Magn. Reson.
- 26 -

A
B
Figure 9. Structure output of Structure Elucidator subject to fragmentation using MS Fragmenter and then sorted on the
presence of fragment ions at m/z 232 and 217. After sorting on the basis of the selected fragment ions, the list was
reordered on the basis of increasing dA13C. Figure 11. A.) Conventional approach to natural product structure
elucidation favored in many laboratories using COSY or TOCSY to establish the proton-proton connectivity network and
HMQC to correlate protons to their directly bound carbon resonances. These data were both acquired in a conventional
400 MHz NMR probe using an ~2.5 mg sample of quindolinocryptotackieine (22) overnight. B.) Inverted direct response
HSQC-TOCSY spectrum of 22 acquired in ~4 hr using an ~ 1 mg sample in a 3 mm NMR tube in a 5 mm cryogenic 500
MHz NMR probe. The ambiguity associated with the three carbon resonances in the boxed region of panel A) is
circumvented by the unequivocal correlation to the resonance furthest downfield of the boxed trio of resonance in the
HSQC-TOCSY spectrum.

Vol. 1, Nos 1 and 2, 2003 Applications and Advances in Cryogenic NMR Probes and
Computer-Assisted Structure Elucidation

- 27 -


A








B

Fi
gu
re
1
2.

C
om
pa
ra
tiv
e
4
hr
in
ve
rte
d
di
re
ct
r
es
po
ns
e
H
S
Q
C
-T
O
C
S
Y
s
pe
ct
ra
r
ec
or
de
d
us
in
g
an
~
1
m
g
sa
m
pl
e
of
q
ui
nd
ol
in
oc
ry
pt
ot
ac
ki
ei
ne
(
22
)
di
ss
ol
ve
d
in
1
70
µL
o
f
m
et
ha
no
l-d
4
in
a
3
m
m
N
M
R
s
am
pl
e
tu
be
.
A
.)
S
pe
ct
ru
m
r
ec
or
de
d
on
a
6
00
M
H
z
N
M
R
in
st
ru
m
en
t
eq
ui
pp
ed
w
ith
a
3
m
m
g
ra
di
en
t
in
ve
rs
e
tri
pl
e
re
so
na
nc
e
N
M
R
p
ro
be
in
4
h
r.
T
he
s
ig
na
l-t
o-
no
is
e
ra
tio
f
or
t
he
t
ra
ce
s
ho
w
n
co
rr
es
po
nd
in
g
to
t
he
p
ro
to
na
te
d
ca
rb
on
r
es
on
at
in
g
at
~
11
9
pp
m
w
as
8
:1
.
B
.)
S
pe
ct
ru
m

re
co
rd
ed
fr
om
th
e
sa
m
e
3
m
m
s
am
pl
e
of
2
2
ru
n
co
ax
ia
lly
in
a
5
m
m
5
00
M
H
z
V
ar
ia
n
C
hi
lip
ro
be

in
4
h
r.
T
he
s
ig
na
l-t
o-
no
is
e
ra
tio
fo
r
th
e
sa
m
e
pr
ot
on
at
ed

ca
rb
on
tr
ac
e
as
p
lo
tte
d
in
p
an
el
A
w
as
1
6.
7:
1.

Martin et al Ann. Magn. Reson.
- 28 -

A









B

Fi
gu
re
1
3.

C
om
pa
ris
on
n
on
-g
ra
di
en
t H
M
B
C
s
pe
ct
ra
o
f a
1
9
µg
s
am
pl
e
of
th
e
al
ka
lo
id
re
tro
rs
in
e
(2
3)
d
is
so
lv
ed
in
1
50
µL
o
f m
et
ha
no
l-d
4 i
n
a
3
m
m
N
M
R
tu
be
.
A.
) O
ve
rn
ig
ht
(1
5
hr
) 6

H
z
op
tim
iz
ed
H
M
B
C
s
pe
ct
ru
m
o
f r
et
ro
rs
in
e
ac
qu
ire
d
us
in
g
a
V
ar
ia
n
5
m
m
5
00
M
H
z
C
hi
lip
ro
be
.
T
he
s
pe
ct
ru
m
is
p
lo
tte
d
at
th
e
no
is
e-
flo
or
. T
he
m
aj
or
ity
o
f t
he
p
er
tin
en
t
re
sp
on
se
s
ar
e
ob
se
rv
ed
in
th
is
1
5
hr
d
at
a
ac
qu
is
iti
on
d
es
pi
te
th
e
si
ze
o
f t
he
s
am
pl
e.
H
ig
h
qu
al
ity
d
at
a
w
ou
ld
o
bv
io
us
ly
b
e
re
co
rd
ed
o
ve
r a
w
ee
ke
nd
. B
.)
C
or
re
sp
on
di
ng

ov
er
ni
gh
t (
15
h
r)
H
M
B
C
s
pe
ct
ru
m
o
f r
et
ro
rs
in
e
ac
qu
ire
d
us
in
g
a
co
nv
en
tio
na
l 3
m
m
N
al
or
ac
Z
•S
P
E
C

50
0
M
H
z
gr
ad
ie
nt
in
ve
rs
e
tri
pl
e
re
so
na
nc
e
N
M
R
p
ro
be
. O
nl
y
a
fe
w
o
f t
he
m
or
e
in
te
ns
e
co
rre
la
tio
ns
a
re
re
lia
bl
y
ob
se
rv
ed
in
th
e
sp
ec
tru
m
.
A
t a
m
in
im
um
, a
cq
ui
rin
g
us
ab
le
d
at
a
on
a
s
am
pl
e
th
is
s
iz
e
in
a
c
on
ve
nt
io
na
l 3
m
m
g
ra
di
en
t
in
ve
rs
e
pr
ob
e
at
5
00
M
H
z
w
ou
ld
c
on
su
m
e
a
lo
ng
w
ee
ke
nd
.

Vol. 1, Nos 1 and 2, 2003 Applications and Advances in Cryogenic NMR Probes and
Computer-Assisted Structure Elucidation

- 29 -
Fi
gu
re
1
4.

C
om
pa
ris
on
n
on
-g
ra
di
en
t
H
M
B
C
s
pe
ct
ra
o
f
a
19
µg
s
am
pl
e
of
t
he
a
lk
al
oi
d
re
tro
rs
in
e
(2
3)
d
is
so
lv
ed
in
1
50
µL
o
f
m
et
ha
no
l-d
4
in
a
3
m
m
N
M
R
t
ub
e.
A
.)
O
ve
rn
ig
ht
(
15
h
r)
6
H
z
op
tim
iz
ed
H
M
B
C
s
pe
ct
ru
m
o
f r
et
ro
rs
in
e
ac
qu
ire
d
us
in
g
a
V
ar
ia
n
5
m
m
5
00
M
H
z
C
hi
lip
ro
be
.
T
he
s
pe
ct
ru
m
is
p
lo
tte
d
at
th
e
no
is
e-
flo
or
.
Th
e
m
aj
or
ity
o
f
th
e
pe
rti
ne
nt
r
es
po
ns
es
a
re
o
bs
er
ve
d
in
t
hi
s
15
h
r
da
ta
a
cq
ui
si
tio
n
de
sp
ite
t
he
s
iz
e
of
t
he
s
am
pl
e.
B
.)
H
M
B
C
s
pe
ct
ru
m

ob
ta
in
ed
u
si
ng
w
ha
t w
ou
ld
n
or
m
al
ly
b
e
an
u
na
cc
ep
ta
bl
y
lo
ng
1
20
h
r
ac
qu
is
iti
on
in
a
c
on
ve
nt
io
na
l 3
m
m
N
al
or
ac
Z
•S
P
E
C

50
0
M
H
z
gr
ad
ie
nt
in
ve
rs
e
tri
pl
e
re
so
na
nc
e
N
M
R
p
ro
be
.
M
or
e
of
th
e
lo
ng
-r
an
ge
c
or
re
la
tio
ns
a
re
r
el
ia
bl
y
ob
se
rv
ed
in
th
e
sp
ec
tru
m
b
ut
th
er
e
ar
e
st
ill
s
ig
ni
fic
an
tly
fe
w
er
r
es
po
ns
es

th
an
w
er
e
ob
se
rv
ed
in
th
e
ov
er
ni
gh
t (
15
h
r)
H
M
B
C
s
pe
ct
ru
m
re
co
rd
ed
in
th
e
C
hi
lip
ro
be
.


A








B

Martin et al Ann. Magn. Reson.
- 30 -



A









C
B









D

Fi
gu
re
1
5.

C
om
pa
ra
tiv
e
lo
ng
-ra
ng
e
1 H
-1
5 N
C
IG
A
R
-H
M
B
C
2
D
N
M
R
s
pe
ct
ra
o
f t
he
o
xa
zo
lid
in
on
e
an
tib
io
tic
e
pe
re
zo
lid
(2
4)
a
cq
ui
re
d
us
in
g
a
2
m
g
sa
m
pl
e
di
ss
ol
ve
d
in
1
50
µL

ac
et
on
itr
ile
-d
3
in
a
3
m
m
N
M
R
tu
be
. T
he
s
pe
ct
ru
m
s
ho
w
n
in
p
an
el
A
w
as
a
cq
ui
re
d
in
4
h
r u
si
ng
a
c
on
ve
nt
io
na
l 3
m
m
5
00
M
H
z
gr
ad
ie
nt
in
ve
rs
e
tri
pl
e
re
so
na
nc
e
N
M
R
p
ro
be
. T
he
p
ro
je
ct
io
n
th
ro
ug
h
F 1
(1
5 N
) o
f t
he
se
d
at
a
is
s
ho
w
n
in
p
an
el
C
. T
he
s
ig
na
l-t
o-
no
is
e
ra
tio
fo
r t
he
e
xp
er
im
en
t w
as
4
9:
1.
D
at
a
ac
qu
ire
d
fo
r t
he
s
am
e
sa
m
pl
e
in
2
6
m
in
u
si
ng
a
5
m
m
5
00
M
H
z
C
hi
lip
ro
be

ar
e
sh
ow
n
in
p
an
el
B
. T
he
c
or
re
sp
on
di
ng
F
1
pr
oj
ec
tio
n,
s
ho
w
n
in
p
an
el
D
, g
av
e
a
si
gn
al
-to
-n
oi
se
ra
tio
o
f
10
1:
1.
T
he
re
sp
on
se
a
t ~
68
p
pm
in
b
ot
h
pr
oj
ec
tio
ns
is
th
e
fo
ld
ed
a
ce
to
ni
tri
le
n
itr
og
en
re
so
na
nc
e
co
rre
la
te
d
by
th
e
m
et
hy
l.

Vol. 1, Nos 1 and 2, 2003 Applications and Advances in Cryogenic NMR Probes and
Computer-Assisted Structure Elucidation

- 31 -


278.1
107.4
168.5
N N
CH3
N
N
10'
5'
5
6


Figure 16. Overnight 3 Hz optimized 1H-15N HMBC spectrum of an ~ 1 mg sample of quindolinocryptotackieine
(22) dissolved in 150 µL methanol-d4. Correlations to three of the four nitrogens are observed as
shown.

Martin et al Ann. Magn. Reson.
- 32 -