Regiochemical tagging: A new tool for structural characterization of isomeric components in combinatorial mixtures

Article abstract of DOI:10.1021/ja993844v

In this contribution we present a new combined synthetic and analytical strategy (regiochemical tagging) that allows facile determination of complete structure, including substituent position and regiochemistry, of mass- redundant components in complex co

Full text of DOI:10.1021/ja993844v

3358
J. Am. Chem. Soc. 2000, 122, 3358-3366
Regiochemical Tagging: A New Tool for Structural Characterization
of Isomeric Components in Combinatorial Mixtures
Noureddin Nazarpack-Kandlousy,† Igor V. Chernushevich,‡ LingJie Meng,§
Ying Yang,§ and Alexey V. Eliseev*,†
Contribution from the Department of Medicinal Chemistry, State UniVersity of New York at Buffalo,
Buffalo, New York 14260, MDS Sciex, 71 Four Valley Dr., Concord, Ontario, Canada L4K 4V8, and Mass
Spectrometry Laboratory of MMRC, Faculty of Medicine, UniVersity of Toronto, Toronto ON M5S 1A8
ReceiVed October 28, 1999
Abstract: In this contribution we present a new combined synthetic and analytical strategy (regiochemical
tagging) that allows facile determination of complete structure, including substituent position and regiochemistry,
of mass-redundant components in complex combinatorial mixtures. The libraries of components (oxime ethers)
are formed by the reaction of a mixture of substituents (aldehydes) with a scaffold containing several chemically
similar attachment points (aminooxy groups). The structure of the resulting library components can then be
determined from a combination of single MS and the tandem (MS/MS) spectra. Determination of the unique
isomeric motif for each component is made possible via the following features of library design: (1) part of
the scafforld moiety, “transferable group” (the nitrogen atom from the oxime group) is transferred to the
substituent during fragmentation in the tandem experiment, (2) transferable groups on the scaffold differ from
each other by either isotopic labels or fragmentation energies, and (3) mass-redundant substituents are isotopically
labeled to create at least a 2 mass unit difference between them. The components of the resulting library thus
become labeled with different mass- and energy tags, which allows for precise regiochemical assignment of
the functional group positions on the scaffold and substituents by mass spectrometry. The approach has been
used to create and analyze a mixture of 27 isomeric compounds, each containing three boronic acid groups.
The combination of the MS and MS/MS spectra of the tagged mixture has yielded a unique and structurally
definitive signature of each component. Applications of the regiochemical tagging techniques to rapid synthesis
and screening of combinatorial mixtures are discussed.
Introduction
The two most common ways to address the identification
problem are tagging techniques3 and deconvolution strate-
gies.2a,b,e,3a,4 However, complications arise when a library
contains multiple components of the same molecular weight and
similar structural features.
The explosive development of combinatorial chemistry over
the past decade has successfully addressed the needs of applied
and fundamental research for rapid generation and screening
of large numbers of compounds.1
One of the most convenient and rapid ways to explore
functional diversity space of the mixture-based libraries is via
reactions of templates (scaffolds) containing two or more
chemically similar attachment points with a variety of substit-
uents (see Scheme 1).4,5 The resulting “branched” library
components cover multiple tiny structural variations, such as
isomerism of the functional substituents on the scaffold, that
play an important role in the recognition of biomolecular targets.
In designing a combinatorial library, an investigator always
faces the choice of working either with arrays of individual
compounds prepared via parallel syntheses or with mixtures
generated by one or several one-pot protocols.1e,2 While mixtures
are readily accessible and allow one to build up a vast diversity
of compounds for a short time, they have apparent drawbacks,
such as difficulty with the identification of active components.
† State University of New York.
‡ MDS Sciex.
(3) (a) Janda, K. D. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 10779-10785.
(b) Burbaum, J. J.; Ohlmeyer, M. H. J.; Reader, J. C.; Henderson, I.; Dillard,
L. W.; Li, G.; Randle, T. L.; Sigal, N. H.; Chelsky, D.; Baldwin, J. J. Proc.
Natl. Acad. Sci. U.S.A. 1995, 92, 6027-6031. (c) Still, W. C. Acc. Chem.
Res. 1996, 29, 155-163. (d) Czarnik, A. W. Proc. Natl. Acad. Sci. U.S.A.
1997, 94, 12738-12739. (e) Wentworth, P.; Janda, K. D. Curr. Opin.
Biotechnol. 1998, 9, 109-115.
(4) (a) Boger, D. L.; Goldberg, J.; Jiang, W. Q.; Chai, W. Y.; Ducray,
P.; Lee, J. K.; Ozer, R. S.; Andersson, C. M. Bioorg. Med. Chem. 1998, 6,
1347-1378. (b) Boger, D. L.; Chai, W. Y.; Jin, Q. J. Am. Chem. Soc. 1998,
120, 7220-7225. (c) Boger, D. L.; Chai, W. Y. Tetrahedron 1998, 54,
3955-3970.
(5) (a) Carell, T.; Wintner, E. A.; Bashirhashemi, A.; Rebek, J., Jr. Angew.
Chem., Int. Ed. Engl. 1994, 33, 2059-2061. (b) Carell, T.; Wintner, E. A.;
Sutherland, A. J.; Rebek, J., Jr.; Dunayevskiy, Y. M.; Vouros, P. Chem.
Biol. 1995, 2, 171-183. (c) Shipps, G. W.; Pryor, K. E.; Xian, J.; Skyler,
D. A.; Davidson, E. H.; Rebek, J., Jr. Proc. Natl. Acad. Sci. U.S.A. 1997,
94, 11833-11838.
§ University of Toronto.
(1) (a) Lowe, G. Chem. Soc. ReV. 1995, 24, 309. (b) Thompson, L. A.;
Ellman, J. A. Chem. ReV. 1996, 96, 555-600. (c) Williard, X.; Pop, I.;
Bourel, L.; Horvath, D.; Baudelle, R.; Melnyk, P.; Deprez, B.; Tartar, A.
Eur. J. Med. Chem. 1996, 31, 87-98. (d) Balkenhohl, F.; Von dem
Busschehunnefeld, C.; Lansky, A.; Zechel, C. Angew. Chem., Int. Ed. Engl.
1996, 35, 2289-2337. (e) Maehr, H. Bioorg. Med. Chem. 1997, 5, 473-
491.
(2) (a) Konings, D. A. M.; Wyatt, J. R.; Ecker, D. J.; Freier, S. M. J.
Med. Chem. 1996, 39, 2710-2719. (b) Wilsonlingardo, L.; Davis, P. W.;
Ecker, D. J.; Hebert, N.; Acevedo, O.; Sprankle, K.; Brennan, T.; Schwarcz,
L.; Freier, S. M.; Wyatt, J. R. J. Med. Chem. 1996, 39, 2720-2726. (c)
Herman, L. W.; Tarr, G.; Kates, S. A. Mol. DiVersity 1997, 2, 147-155.
(d) Neustadt, B. R.; Smith, E. M.; Lindo, N.; Nechuta, T.; Bronnenkant,
A.; Wu, A.; Armstrong, L.; Kumar, C. Bioorg. Med. Chem. Lett. 1998, 8,
2395-2398. (e) Boutin, J. A.; Lambert, P. H.; Bertin, S.; Volland, J. P.;
Fauchere, J. L. J. Chromatogr. B 1999, 725, 17-37.
10.1021/ja993844v CCC: $19.00 © 2000 American Chemical Society
Published on Web 03/24/2000

Regiochemical Tagging
J. Am. Chem. Soc., Vol. 122, No. 14, 2000 3359
Scheme 1
Scheme 2
particularly convenient to form the fragments directly in the
mass spectrometer by breaking the component molecular ions
by the collision-induced dissociation in the tandem MS (MS/
MS) experiment.9 The mass tags, such as isotopes, are intro-
duced into the transferable groups at the stage of the scaffold
synthesis. Once the unique tag of each linker is transferred to
its substituent upon fragmentation, the substituent fragment will
“remember” its attachment point via its specific mass or isotopic
distribution. Hence, the substitution motif in the parent com-
pound can be reconstructed from the MS/MS fragment peaks.
This approach has been tested in this work on model mixtures
of compounds containing oxime ether linkages. This class of
compounds is interesting from several standpoints. Oxime ethers
can be easily formed from commercially available or readily
synthesized alkoxyamines, in which the ONH2 groups serve as
attachment points and aldehydes serve as substituents. As shown
previously, the latter reaction is relatively insensitive to the
electronic and steric properties of the building blocks and results
in good representation of most of the library components.10
Finally, oxime ethers may be used in dynamic combinatorial
libraries11 wherein the effective components can be amplified
in a combined synthesis-screening process.12
As a tradeoff, these minor differences severely complicate the
identification of a specific component regiochemistry.
Combinatorial mixtures are commonly analyzed by mass
spectrometry as one of the most sensitive and informative
techniques.2e,6 However, information from the mass spectrum
does not generally yield the complete structure of a “branched”
library component, such as relative position of substituents on
the scaffold, even when coupled to identification of the
component fragments via a tandem (MS/MS) experiment.7
We introduce here a new combined synthetic/analytical
strategy of library design, referred to as regiochemical tagging,
that allows complete identification of structure, including
regiochemistry, of each component by mass spectrometry.
Unlike existing tagging techniques, including isotopic mass
tagging,8 our approach is based on labeling building blocks and
their position in the resulting components, rather than whole
molecules.
Results and Discussion
As described later, the regiochemical tagging method can be
used with scaffolds containing more than two attachment points.
However, to test the approach we first synthesized compound
1 containing just two ONH2 groups. The mixture components
were generated by the reaction of scaffold 1 with substituted
benzaldehydes. As one of the systems of interest, we used the
aldehydes containing boronic acid units in ortho, meta, and para
positions (Scheme 2). Boronic acids have recently been a subject
of extensive studies in the molecular-recognition community13
due to their ability to reversibly form cyclic esters with
carbohydrate units in aqueous solutions, thereby providing one
of the few effective ways to selectively recognize sugars in
water.14 Because the binding strength in such cyclic esters is
The basic idea of the regiochemical tagging strategy is
outlined in Scheme 1. The library is synthesized by a divergent
approach from the central scaffold and substituents attached to
it via similar linkages. The main requirement of the linking
group is its ability to form fragments that contain one or several
atoms transferred from the scaffold to substituent moieties
(transferable groups). Such fragmentation may generally result
from any chemical reaction, but for analytical purposes, it is
(6) (a) Dunayevskiy, Y.; Vouros, P.; Carell, T.; Wintner, E. A.; Rebek,
J., Jr. Anal. Chem. 1995, 67, 2906-2915. (b) Dunayevskiy, Y. M.; Vouros,
P.; Wintner, E. A.; Shipps, G. W.; Carell, T.; Rebek, J., Jr. Proc. Natl.
Acad. Sci. U.S.A. 1996, 93, 6152-6157. (c) Winger, B. E.; Campana, J. E.
Rapid Commun. Mass Spectrosc. 1996, 10, 1811-1813. (d) Chu, Y. H.;
Dunayevskiy, Y. M.; Kirby, D. P.; Vouros, P.; Karger, B. L. J. Am. Chem.
Soc. 1996, 118, 7827-7835. (e) Gao, J. M.; Cheng, X. H.; Chen, R. D.;
Sigal, G. B.; Bruce, J. E.; Schwartz, B. L.; Hofstadler, S. A.; Anderson, G.
A.; Smith, R. D.; Whitesides, G. M. J. Med. Chem. 1996, 39, 1949-1955.
(f) Berlin, K.; Jain, R. K.; Tetzlaff, C.; Steinbeck, C.; Richert, C. Chem.
Biol. 1997, 4, 63-77. (g) Pomerantz, S. C.; McCloskey, J. A.; Tarasow, T.
M.; Eaton, B. E. J. Am. Chem. Soc. 1997, 119, 3861-3867. (h) VanBree-
men, R. B.; Huang, C. R.; Nikolic, D.; Woodbury, C. P.; Zhao, Y. Z.;
Venton, D. L. Anal. Chem. 1997, 69, 2159-2164. (i) Zhao, Y. Z.;
Vanbreemen, R. B.; Nikolic, D.; Huang, C. R.; Woodbury, C. P.; Schilling,
A.; Venton, D. L. J. Med. Chem. 1997, 40, 4006-4012. (j) Loo, J. A. Eur.
Mass Spectrosc. 1997, 3, 93-104.
(9) Busch, K. L.; Glish, G. L.; McLuckey, S. A. Mass Spectrometry/
Mass Spectrometry; VCH Publishers: Weinheim, Germany, 1988.
(10) Nazarpack-Kandlousy, N.; Zweigenbaum, J.; Henion, J.; Eliseev,
A. V. J. Comb. Chem. 1999, 1, 199-206.
(11) Polyakov, V. A.; Nelen, M. I.; Nazarpack-Kandlousy, N.; Ryabov,
A. D.; Eliseev, A. V. J. Phys. Org. Chem. 1999, 12, 357-363.
(12) (a) Ganesan, A. Angew. Chem., Int. Ed. 1998, 37, 2828-2831. (b)
Eliseev, A. V.; Lehn, J.-M. Comb. Chem. Biol. 1999, 243, 159-172. (c)
Lehn, J.-M. Chem. Eur. J. 1999, 5, 2455-2463.
(13) (a) James, T. D.; Sandanayake, K.; Shinkai, S. Angew. Chem., Int.
Ed. Engl. 1996, 35, 1911-1922. (b) James, T. D.; Shinmori, H.; Takeuchi,
M.; Shinkai, S. Chem. Commun. 1996, 705-706. (c) James, T. D.; Linnane,
P.; Shinkai, S. Chem. Commun. 1996, 281-288. (d) Shinkai, S.; Takeuchi,
M. Trends Anal. Chem. 1996, 15, 188-194.
(14) Davis, A. P.; Wareham, R. S. Angew. Chem., Int. Ed. 1999, 38,
2978-2996.
(7) Blom, K. F. Anal. Chem. 1997, 69, 4354-4362.
(8) Geysen, H. M.; Wagner, C. D.; Bodnar, W. M.; Markworth, C. J.;
Parke, G. J.; Schoenen, F. J.; Wagner, D. S.; Kinder, D. S. Chem. Biol.
1996, 3, 679-688.

3360 J. Am. Chem. Soc., Vol. 122, No. 14, 2000
Nazarpack-Kandlousy et al.
Scheme 3
tagged library would yield different masses for the substituents
cleaving from the ortho and para positions.
The overall labeling approach is thus designed to provide
double resolution of the library mass spectra: resolution in the
single MS spectrum, via the substituent tagging, and MS/MS
resolution, via the scaffold regiochemical tagging (Scheme 3).
In this way, the combination of MS and MS/MS analysis of
the labeled library should show a unique trace of each
component in the mixture, from which complete isomeric motif
of each component could be determined.
To verify the fragmentation pattern of the library components
and provide references for the MS/MS fragments, we first tested
a number of individual compounds, such as o-15N-8 (Figure 1)
and the standards of library components C1, C4, and C9
(Scheme 4). As expected, the MS/MS spectrum of the molecular
peak of compound o-15N-8 showed major fragments at 353.2
and 354.2 atomic mass units (amu) (Figure 1) that one could
assign to the cleavage of the ArCHN fragments from the labeled
and unlabeled positions on the scaffold, respectively.
strongly dependent upon relative positions of the boronate
moieties, it is likely that in these libraries one can find and
isolate by affinity selection new specific binders for certain
carbohydrate motifs. This approach may be used for fine-tuning
recognition properties of the boronate-based artificial receptors.
However, it came as a surprise that in the MS/MS spectrum
of p-15N-8, a similar compound based on the oppositely labeled
scaffold, the corresponding fragments were shifted by 1 amu,
as if they both had been derived from the cleavage of a lighter
substituent. Furthermore, we examined the MS/MS of the
residual peak of an unlabeled compound that was present in
small amounts in the single MS of o-15N-8. It showed a pattern
similar to that of o-15N-8, with major fragment peaks at 353.2
and 354.2. The latter two observations pointed at the possibility
of an alternative fragmentation mechanism, perhaps, involving
the breakage of the N-O bond only in the ortho position to
the alkylamino substituent. This mechanism was further eluci-
dated from the spectra of compounds C4 and C9 (Figure 2).
The reaction between scaffold 1 and three isomeric boronic
acids results in the formation of nine isomeric compounds 2
that offer a double challenge from the analysis standpoint. Not
only are all components isobaric (i.e., have the same mass),
and therefore show up as a single peak in the MS spectrum,
but also, due to the isomerism of the starting aldehydes, the
molecular fragments that form upon fragmentation of 2 in a
collision-induced dissociation are isobaric as well. Thus, the
combination of MS and MS/MS analysis is incapable of
detecting each individual component in the mixture and
identifying structure of an individual component.
The molecular peaks of both C4 and C9 (505.2 and 513.2,
respectively) fragmented at 10 eV to yield two major peaks in
the MS/MS spectrum. As with compounds p-15N-8 and o-15N-
8, the lower-mass fragment corresponded in each case to the
cleavage of the O-N bond in the ortho position. However, the
mass unit difference between the lower- and higher-mass
fragments was 2 amu, as opposed to 1 amu in the non-deuterated
compounds. The QqTOF instrument used for the mass measure-
ment yielded the precise mass difference of 2.014 between the
two fragments. This value matches with the mass of one
deuterium atom and most probably corresponds to partial
transfer of the imine hydrogen to the scaffold during the
fragmentation at the ortho-N-O bond. The precise mass
measurement also allowed us to eliminate other fragment
combinations that would yield similar mass differences (such
as the (D + N15) - (H + N14) ) 2.003 that might emerge
from some cleavage pattern from both ortho and para positions).
To address the problem of the substituent mass-redundancy,
we labeled the corresponding aldehydes with the isotopes of
hydrogen and boron to introduce the mass-difference of at least
2 units between the isomers (Scheme 3a). The labeled aldehydes
were synthesized from commercially available enriched precur-
sors (see Supporting Information). The mass spectrum of the
library formed by the reaction of 1 with the labeled aldehydes
would thus contain six molecular peaks (three of them for
individual components and three for pairs of isomers, see Figure
3). The above labeling scheme resulted in a mass unit difference
of at least 2 between the molecular peaks, which allowed us to
avoid their interference with the intense M + 1 isotopic peaks
to simplify the subsequent MS/MS analysis. The interference
with the M + 2 peaks, which are significantly smaller, results
only in minor extra signals present in many MS/MS spectra.
The labeling of the substituents yields no information about
their position on the scaffold, that is, about isomerism of
compounds 2. To address this issue, we introduced an isotopic
label in one of the aminooxy groups (Scheme 3b).
All of the above observations indicate that the O-N bond in
the ortho position to the aminoalkyl group undergoes fragmenta-
tion at significantly lower energy than the one in para position
and that its fragmentation is accompanied by an unusual
hydrogen transfer process from the imine group of the substitu-
ent to the scaffold. Because the probability of any intermolecular
assistance to the fragmentation reaction in the gas phase is low,
the only difference in the environment of both groups is the
presence of the alkylammonium group next to the N-O bond
that breaks more readily. It is therefore reasonable to assume
that the cleavage is assisted by the ammonium group, for
example via the six-membered transition state shown in Scheme
5. This mechanism also accounts for the partial transfer of the
imine proton/deuterium. The degree of this transfer decreases
at higher fragmentation energies (see Figure 1). The possibility
The sequence of reactions used for the scaffold synthesis
(Scheme 4) allowed us to introduce the two aminooxy groups
in separate steps. The group in the para position could thus be
labeled with the 15N isotope. Upon propagation of the labeled
scaffold 7 to the library components C1-C9, the substituents
in each component become attached to the nitrogen atoms with
different masses. Preliminary experiments showed that the
collision-induced dissociation of molecular peaks of a variety
of oxime ethers in the MS/MS experiment consistently yielded
fragments resulting from the breakage of the O-N bond, as
one of the weakest bonds in the molecule. Corresponding
fragments obtained from each component of the regiochemically

Regiochemical Tagging
J. Am. Chem. Soc., Vol. 122, No. 14, 2000 3361
Scheme 4
(a) K2CO3, DMF, rt, 86% total; (b) K2CO3, DMF, rt, 55%; (c) Ti(O-i-Pr)4, NaBH3CN, CHCl3/MeOH, 46%; (d) N2H4‚H2O, MeOH/CHCl3, 64%;
(e) 1% TFA in DMSO.
also exists that the preferred fragmentation in the ortho position
is attributable to steric effects of the neighboring group. We
are currently performing control experiments with some struc-
tural analogues to verify the transfer mechanism.
The relative stability of the para-substituent indicates that
the unassisted collision-induced dissociation of the N-O bond
requires intrinsically higher energies. This is indirectly supported
by the MS/MS of the protected scaffold 6 in which noticeable
fragments have been detected only at the energies of 30 eV
and above. It is also probable that the breakage of the p-O-N
bond is cooperative with the ortho, because the peak at 208.1,
corresponding to cleavage of both substituents, was always
observed in the MS/MS spectrum.
Overall, the experiments with the individual compounds
revealed an unexpected fragmentation mechanism and consider-
able difference in the stability of the ortho- and para-
substituent-scaffold linkages. As a result, introduction of the
labeled nitrogen into scaffold 7 did not yield the required
regiochemical information. Importantly, however, the originally
stated goal to detect and identify every single component in
the mixture by the MS analysis could be achieved, due to the
different stability of the N-O linkers. The molecular peak
cluster of the library of nine compounds (Figure 3) showed all
six expected peaks, three of which represented pairs of two
different components. The MS/MS of these molecular ions
yielded distinct traces of each individual component, identifiable
as the fragments resulting from the ortho cleavage.
15N isotope. Thus, one of the attachment points on the scaffold
bore the “energy” tag, while the other one was labeled with a
mass tag that made it different from the third, untagged point.
The MS/MS spectrum of an individual standard compound
CT8 (see Table 1) based on the new scaffold indeed showed
the expected fragmentation pattern (Figure 4a). Its comparison
with the unlabeled analogue (Figure 4b) proved that one of the
fragments resulted from the cleavage of the 15N-labeled position.
The reaction of scaffold 12 with the three labeled formyl
boronic acids resulted in the formation of a mixture library of
27 components listed in Table 1. Because of the molecular mass
redundancy, the library includes nine different masses, all of
which were detected in the molecular cluster in the single MS
spectrum (Figure 5). Again, the component molecular peaks here
differ by at least 2 mass units, which helps avoid their
interference with the M + 1 molecular peaks of the closest
neighbors.
Each of the molecular ions was characterized by its MS/MS
spectrum. We chose to perform fragmentation at 20 eV to
minimize the peaks resulting from the proton transfer to the
scaffold, as discussed above. The results shown in Figure 5
represent an error-free proof of the tagging strategy design in
that all 36 actually observed MS/MS peaks were predicted on
the basis of library structure. Combination of the MS and MS/
MS spectra obtained from a single injection has thus provided
unique signature of each of the 27 components that, unlabeled,
are all isomers and yield nearly identical spectra!
The unexpected observation of different O-N bond frag-
mentation energies prompted us to design a new regiochemical
tagging scheme that was implemented in a trisubstituted scaffold
12 (Scheme 6). Compound 12 contains two aminooxy groups
in the ortho position to the central nitrogen atom, which were
expected to fragment at similar energies, and one aminooxy
group in the para position, which should not show significant
fragmentation. To distinguish between the fragments resulting
from different ortho groups, one of them was labeled with the
The most complex peak in the molecular cluster of the CT
library corresponds to six isomeric components. In the MS/MS
spectrum of that and some other peaks, the signals originating
from different components are sometimes overlapped. However,
if each component were injected individually, its unique
signature would unambiguously determine its isomeric motif
and distinguish it from other 26 isomers.
This study concentrated primarily on the boronic acid-
containing substituents, as those that allowed us to push the

3362 J. Am. Chem. Soc., Vol. 122, No. 14, 2000
Nazarpack-Kandlousy et al.
Figure 2. Single MS spectrum of the molecular ion clusters and MS/
MS spectra (10 eV) of individually injected C4 (a, b) and C9 (c, d).
Scheme 5
Figure 1. Segments of the MS/MS spectra corresponding to the N-O
bond cleavage in the molecular ions of o-15N-8 (solid lines) and p-15N-8
(dashed lines) at various energies.
limits of our approach applicability for structural characterization
of very similar compounds in their mixtures. Apparently, the
regiochemical tagging schemes can be used for a variety of
substituents. We briefly tested several libraries formed from
scaffold 7 and p-hydroxy and p-nitrobenzaldehydes, as well as
the formyl boronic acids, and observed all component peaks
and their easily identifiable fragments in the MS/MS spectra.
This indicates that the method can be used in libraries formed
from aldehydes with varying electronic properties. It is essential
for unambiguous structure determination of a particular com-
ponent based on scaffold 12 that the mass difference between
closest substituents be at least 2 amu, as it was implemented in
the labeled formyl boronic acids. In this case, the combination
of the MS and MS/MS spectra is unique for each component.
Concluding Remarks
motif of the strong binders isolated, for example, by affinity
selection15 can be determined by single injection in the mass
spectrometer. The method could thus be used for rapid geometric
and functional mapping of the target binding sites.
The fact that only isotopes are used as labels simplifies the
synthetic part of the technique, which essentially only requires
The method of regiochemical tagging offers a unique ap-
proach to complete structural characterization of specially
designed combinatorial mixture components, which would be
difficult or impossible to achieve through other techniques. The
combination of the tagging strategy with the facile mixture
generation from the scaffold and substituent building blocks can
be very useful for rapid screening of potentially very diverse
libraries for binding to a particular target. A complete isomeric
(15) (a) Eliseev, A. V. Curr. Opin. Drug DiscoVery DeV. 1998, 1, 106-
115. (b) Woodbury, C. P.; Venton, D. L. J. Chromatogr. B 1999, 725, 113-
137.

Regiochemical Tagging
J. Am. Chem. Soc., Vol. 122, No. 14, 2000 3363
Scheme 6
(a) K2CO3, DMF, rt, 86% total; (b) K2CO3, DMF, rt, 55%; (c) Ti(O-
i-Pr)4, NaBH3CN, CHCl3/MeOH, 46%; (d) N2H4‚H2O, MeOH/CHCl3,
64%; (e) 1% TFA in DMSO.
Experimental Section
Mass Spectrometry. Most of the experiments were performed on a
prototype of a tandem quadrupole time-of-flight mass spectrometer
(QqTOF) built at SCIEX.16 The high resolution and mass accuracy of
the QqTOF instrument makes it possible to remove many ambiguities
in isotopic labeling. In brief, the instrument consists of four main
parts: an electrospray ion source, a quadrupole mass filter Q1, a
collision cell Q2 (with argon as collision gas), and a time-of-flight
(TOF) mass analyzer with orthogonal injection of ions. In single MS
mode of operation, Q1 is operated in radio frequency (rf)-only (non-
analyzing) mode, and precursor ions are recorded by TOF. In MS/MS
mode, precursors of interest are mass-selected with unit resolution in
Q1 and fragmented in Q2, and fragment ions are recorded by TOF.
Mass resolution of TOF is ∼10 000 (full width at half-maximum
definition). Typical mass accuracy is ∼20 ppm with external calibration
(when calibrated in a separate experiment), and 3 ppm with internal
calibration (when calibrated with the known peaks in the same
spectrum).
Solutions were prepared in water/acetonitrile (75:25) with sample
concentration ranging from 1 to 5 µM, and injected into the mass
spectrometer at a flow rate of 1 µL/min. Typically, a single MS
spectrum was recorded in 30 s, and the MS/MS spectrum was recorded
in 1-3 min.
Some spectra of precursors and individual compounds were acquired
on an API III Plus triple quadrupole mass spectrometer (PE Sciex,
Thornhill, Canada) fitted with an articulated pneumatically assisted
nebulization probe. The spectrometer was operated at unit resolution
Figure 3. Single MS spectrum of the molecular ion cluster (top) and
MS/MS spectra (10 eV) of the molecular ions (marked in the right top
corners) of the mixture of components C1-C9.
repetition of known synthesis with the labeled units. It is
important that the tagged components fully retain all chemical
properties of their untagged analogues.
One can easily visualize further various modifications and
extensions of the principal regiochemical tagging scheme. For
example, it may first appear that the number of “mass-taggable”
positions in the scaffold is limited by the number of available
stable isotopes. However, mixtures of isotopes can be entered
in the tagged positions (e.g., as 50:50 14N:15N, or in different
ratios) to label multiple attachment points with unique fractions
of the mass tags. Likewise, different scaffold-substituent linkers
can be explored that follow the transfer-fragmentation rule
depicted in Scheme 1.
(16) Shevchenko, A.; Chernushevich, I.; Ens, W.; Standing, K. G.; Wilm,
M.; Mann, M. Rapid Commun. Mass Spectrom. 1997, 11, 1015-1024.

3364 J. Am. Chem. Soc., Vol. 122, No. 14, 2000
Nazarpack-Kandlousy et al.
Table 1. Molecular Weights of Library Components Based on
Scaffold 12 and Their Key Fragments
Synthesis. All commercial reagents were purchased from Aldrich,
except for the 15N-labeled hydroxylamine that was obtained from
Cambridge Isotope Laboratories. NMR spectra were recorded on Varian
Unity 300, 400, and 500 MHz instruments and processed using the
original manufacturer software or the SwanMR shareware for Macin-
tosh.17 FAB MS spectra were recorded on a VG Analytical 70-SE
instrument.
a
a
a
component no. R1
R2
R3
Mb
M-R1N M-R315N
CT1
CT2
CT3
CT4
CT5
CT6
CT7
CT8
o
o
m
o
m
o
m
m
o
p
o
o
o
o
o
o
m
o
m
m
m
o
o
p
m
p
o
p
o
m
m
p
m
p
o
p
p
m
p
o
m
o
699.2
701.2
701.2
701.2
703.2
703.2
703.2
705.2
705.2
705.2
705.2
707.2
707.2
707.2
707.2
707.2
707.2
709.2
709.2
709.2
711.2
711.2
711.2
713.2
713.2
713.2
717.2
552.2
554.2
552.2
554.2
554.2
556.2
554.2
556.2
558.2
552.2
558.2
560.2
560.2
558.2
558.2
554.2
554.2
560.2
560.2
556.2
564.2
558.2
558.2
564.2
560.2
560.2
564.2
551.2
551.2
553.2
553.2
553.2
553.2
555.2
555.2
551.2
557.2
557.2
553.2
557.2
553.2
559.2
557.2
559.2
555.2
559.2
559.2
557.2
557.2
563.2
559.2
559.2
563.2
563.2
o
endo-15N-Hydroxy-5-norbornene-2,3-dicarboximide (X-15N-OH,
Scheme 4). The compound was synthesized in a modification of
previously described procedure.18 cis-5-Norbornene-endo-2,3-dicar-
boxylic anhydride (1.307 g, 7.72 mmol) was added to a solution of
0.495 g (7.02 mmol) of hydroxylamine-15N‚HCl (98% 15N) in 5 mL of
water, and the mixture was heated at 60 °C for 3 h. After the mixture
cooled to room temperature, the pH was adjusted to 4.0 with glacial
acetic acid. After removal of the solvent, purification on a silica gel
column (10% MeOH/CH2Cl2) yielded 1.153 g (91%) of the product.
1H NMR (δ ppm, 500 MHz, DMSO-d6) 10.76 (br s, 1H), 6.05 (t, J )
1.5 Hz, 2H), 3.25(m, 4H), 1.58 (d, J ) 10 Hz, 1H), 1.5 (d, J ) 10 Hz,
1H). 13C NMR (δ ppm, 125 MHz, DMSO-d6) 172.9 (d, J ) 11.9 Hz),
134.4, 51.0 (d, J ) 1.4 Hz), 43.8, 42.1 (d, J ) 8.3 Hz). FAB MS:
181.2 ([M + H]+).
m
m
o
m
p
o
o
p
m
p
o
m
o
CT9
CT10
CT11
CT12
CT13
CT14
CT15
CT16
CT17
CT18
CT19
CT20
CT21
CT22
CT23
CT24
CT25
CT26
CT27
m
m
p
p
m
m
p
o
p
p
m
p
p
p
p
m
m
p
p
o
p
p
m
p
Aldehydes 3 and 4. A solution of endo-N-hydroxy-5-norbornene-
2,3-dicarboximide (1.69 g, 9.14 mmol) and anhydrous potassium
carbonate (1.27 g, 9.14 mmol) in anhydrous DMF (25 mL) was stirred
for 30 min at room temperature. Then 2,4-difluorobenzaldehyde (1.00
mL, 9.14 mmol) was added. The reaction progress was monitored by
TLC (silica gel, 25% acetone in petroleum ether) performed after
express workup of the reaction mixture aliquots (dilution with water
and extraction with CH2Cl2). After 48 h, TLC showed completion of
the reaction and formation of two spots having very close Rf values
(0.27 and 0.20, 20% acetone/petroleum ether). Crude NMR showed
the presence of 3 and 4 in the ratio of 78/22. The residue was partitioned
between saturated aqueous NaCl and CH2Cl2. The water layer was
additionally extracted with CH2Cl2, and the combined organic layers
were washed with brine and dried over Na2SO4. After the solvent was
removed in vacuo, the two isomers were separated on silica gel (20%
acetone in petroleum ether) to give 86% overall yield. Minor isomer
(4): 1H NMR (δ ppm, 300 MHz, CDCl3) 10.44 (s, 1H), 7.91 (dd, J1 )
6.6 Hz, J2 ) 8.7 Hz, 1H), 6.91 (td, J1 ) 2.4 Hz, J2 ) 8.2 Hz, 1H),
6.66 (dd, J1 ) 2.4 Hz, J2 ) 9.6 Hz, 1H), 6.32 (t, J ) 1.8 Hz, 2H), 3.52
(br s, 2H), 3.40(m, 2H), 1.85 (dt, J1 ) 1.5 Hz, J2 ) 9.0 Hz, 1H), 1.58
(d, J ) 9.0 Hz, 1H). 13C NMR (δ ppm, 75 MHz, CDCl3) 186.20, 170.54,
168.60, 165.17, 161.0, 135.37, 131.08, 130.94, 121.2, 112.53, 112.24,
p
a See notation in Scheme 3. b Protonated ion.
1
102.49, 102.12, 51.66, 44.93, 43.16). Major isomer (3): H NMR (δ
ppm, 500 MHz, CDCl3) 10.22 (s, 1H), 7.85 (dd, J1 ) 7.5 Hz, J2 ) 8.5
Hz, 1H), 6.89 (dd, J1 ) 2.0 Hz, J2 ) 8.5 Hz, 1H), 6.78 (dd, J1 ) 2.5
Hz, J2 ) 11 Hz, 1H), 6.33(t, J ) 2.0 Hz), 3.52 (br, s, 2H), 3.42 (m,
2H), 1.86(d, J ) 9.0 Hz, 1H), 1.60 (d, J ) 9.0 Hz). 13C NMR(δ ppm,
125 MHz, CDCl3) 185.46, 185.41, 170.33, 166.44, 164.37, 163.12,
163.03, 135.26, 130.48, 130.45, 120.63, 120.56, 110.24, 110.21, 102.25,
102.04, 51.58, 44.88, 43.15).
Protected 2,4-15N-Diaminooxy Benzaldehyde 5. A solution of
endo-15N-hydroxy-5-norbornene-2,3-dicarboximide (180 mg, 1.0 mmol)
and anhydrous potassium carbonate (138 mg, 1.0 mmol) in anhydrous
DMF (20 mL) was stirred for 30 min at room temperature, and then 4
(603 mg, 2.0 mmol) was added. After the solution stirred for 4 days,
the residue was partitioned between saturated aqueous NaCl and CH2-
Cl2. The water layer was extracted two more times with CH2Cl2. The
combined organic layers were washed with brine and dried over Na2-
SO4. After removal of the solvent in vacuo, purification by silica gel
chromatography (step gradient from 1% EtOAc in CH2Cl2 to 5% EtOAc
in CH2Cl2) yielded 253.8 mg (55%) of TLC-pure 5. 1H NMR was the
same as for the nonlabeled analogue.10 13C NMR (δ ppm, 125 MHz,
DMSO-d6): 186.27, 171.24, 171.08 (d, J ) 8.8 Hz), 162.94, 160.12,
135.12, 134.98, 131.92, 119.65, 108.12, 99.70, 51.17, 44.26, 44.23,
42.89. 15N NMR (δ ppm, 51 MHz, DMSO-d6, formamide in DMSO-
d6 ) 90.0 ppm) 201.29. FAB MS: 462.1 ([M + H]+).
Figure 4. Segments of the MS/MS spectra corresponding to the N-O
bond cleavage in the molecular ions of individually injected CT8 (a,
parent ion 705.2) and its unlabeled analogue (b, parent ion 704.2) at
20 eV.
(50% valley definition) over the mass range m/z 50-2400. Samples
were introduced into the electrospray ionization source at a flow rate
of 5 µL/min with a Harvard syringe pump. The electrospray needle
was operated at 4.8 kV, the orifice voltage was set at 55 V, and nitrogen
was used as the nebulization gas. Full scan mass spectra were acquired
over the mass range m/z 200-1500 by scanning the thirst mass
spectrometer, Q1, using a m/z 0.2 step size and a 1 ms dwell time.
Product ion mass spectra (MS/MS) were acquired by colliding the Q1
selected precursor ion with argon gas (99.999%) at a collision target
gas thickness of 9 × 1014 molecules/cm2 in Q2 operated in rf-only
mode and scanning the second mass spectrometer, Q3, from m/z 15 to
just above the mass of the precursor ion.
Piperidine Derivative 6. The compound was synthesized as the
1
nonlabeled analog10 and showed identical H NMR. Yield 48%. 13C
(17) Balacco, G. J. Chem. Inf. Comput. Sci. 1994, 34, 1235-1241.
(18) Bauer, L.; Miarka, S. V. J. Org. Chem. 1959, 24, 1293-1296.

Regiochemical Tagging
J. Am. Chem. Soc., Vol. 122, No. 14, 2000 3365
Figure 5. Single MS spectrum of the molecular ion cluster (top) and MS/MS spectra (20 eV) of the molecular ions (marked in the right top
corners) of the mixture of components CT1 - CT27.
NMR (δ ppm, 125 MHz, CDCl3) 171.13 (d, J ) 5 Hz), 170.98, 157.41,
156.41, 135.16, 134.88, 131.30, 123.68, 111.08, 102.75, 55.39, 54.29,
51.44, 51.35, 44.77, 43.00, 42.88, 42.81, 29.17, 25.87, 24.14. ESI MS
(CH3CN/water) 531.2 ([M + H]+). ESI MS/MS of the M + H ion:
465.2, 369.2.
from 1% MeOH in CH2Cl2 to 20% MeOH in CH2Cl2) resulted in 425.4
mg (45%) of 9. H NMR (δ ppm, 500 MHz, CDCl3) 7.37 (d, J ) 8.5
1
Hz, 1H), 6.80 (dd, J1 ) 2.5 Hz, J2 ) 8.0, 1H), 6.74 (d, J ) 2.5 Hz,
1H), 6.33 (t, J ) 2.0 Hz, 2H), 6.25 (t, J ) 1.8 Hz, 2H), 3.98 (s, 2H),
3.50 (br. m, 4H), 3.35 (dm, J ) 14.5 Hz, 4H), 2.49 (s, 3H), 1.84 (tm,
J ) 12.5 Hz, 2H), 1.58 (t, J ) 9.0 Hz, 2H). 13C NMR (δ ppm, 75
MHz, CDCl3) 171.54, 170.93, 159.13, 156.87, 135.05, 134.91, 133.19,
117.54,110.90, 103.26, 51.32, 50.17, 46.42, 44.65, 42.85, 32.65.
Aminooxy Scaffold 7. The compound was synthesized from 6 using
the previously described deprotection procedure, and the 1H NMR was
identical to that of the nonlabeled analogue.10 Yield 64%. ESI MS (CH3-
CN/water) 239.2 ([M + H]+).
Synthesis and analytical data for the oppositely labeled standard
compound o-15N-8 as well as for the labeled formyl boronic acids are
given in the Supporting Information.
Protected 215N-Aminooxy Benzaldehyde 10. A solution of endo-
15N-hydroxy-5-norbornene-2,3-dicarboximide (360 mg, 2.0 mmol) and
anhydrous potassium carbonate (276 mg, 2.0 mmol) in anhydrous DMF
(10 mL) was stirred for 30 min at room temperature. Then 2-fluo-
robenzaldehyde (652 µL, 6.0 mmol) was added, and the reaction mixture
was stirred at 65 °C for 48 h. The reaction progress was monitored by
TLC (silica gel, 5% EtOAc in CH2Cl2) performed after express workup
of the reaction mixture aliquots (dilution with water and extraction with
CH2Cl2). The pure product was obtained in 62% yield using the same
workup procedure as for 4. 1H NMR (δ ppm, 500 MHz, CDCl3) 10.57
(s, 1H), 7.86 (dd, J1 ) 2.0 Hz, J2 ) 8.0 Hz, 1H), 7.52 (dt, J1 ) 1.5 Hz
J2 ) 8.0 Hz, 1H), 7.21 (dt, J1 ) 1.0 Hz, J2 ) 8.0 Hz, 1H), 6.96 (d, J
) 8.0 Hz, 1H) 6.27 (t, J ) 2.0 Hz, 2H), 3.49 (m, 2H), 3.38(m, 2H),
Amine 9. To a solution of the unlabeled aldehyde 5 (1.01 g, 2.2
mmol) in CHCl3 (150 mL) was added titanium(IV) isopropoxide (1.24
mL, 4.0 mmol). After the mixture stirred for 20 min, methylamine
hydrochloride (135 mg, 2.0 mmol was added. The mixture was refluxed
for 30 min and then kept at room temperature for another 30 min,
followed by addition of methanol (50 mL) and NaBH3CN (800 mg,
6.0 mmol, in two portions within 1 h). The solvent was then removed,
and the residue was partitioned between 200 mL of CH2Cl2 and 50
mL of brine and stirred for 1 h. The mixture was then filtered, the
aqueous layer was extracted twice more with CH2Cl2, and the combined
organic layers were washed with brine and dried over Na2SO4. Removal
of the solvent followed by silica gel chromatography (step gradient
1.81 (dd, J1 ) 1.5 Hz, J2 ) 9.0 Hz, 1H), 1.56 (d, J ) 9.5 Hz, 1H). 13
NMR (δ ppm, 125 MHz, CDCl3) 187.68, 170.80 (d, J ) 9.1 Hz),
C

3366 J. Am. Chem. Soc., Vol. 122, No. 14, 2000
Nazarpack-Kandlousy et al.
159.59, 135.41, 135.02, 128.30, 124.82,124.60, 114.41, 51.49, 44.80,
43.02, 42.95. FAB MS: 285.3 ([M + H]+).
2.4 Hz, J2 ) 8.4 Hz, 1H), 3.64 (s, 2H), 3.59(s, 2H), 2.20 (s, 3H). ESI
MS (CH3CN/water 1/3 v/v) 306.2 ([M + H]+).
Protected Triaminooxy Amine 11. The compound was synthesized
from 9 and 10 using the same reductive amination procedure as for 9.
Yield 45%. 1H NMR (δ ppm, 500 MHz, CDCl3) 7.51 (d, J ) 8.5 Hz,
2H), 7.15 (t, J ) 7.8 Hz, 1H), 7.08 (t, J ) 7.5 Hz, 1H), 6.78 (td, J1 )
General Procedure for Library Formation. Stock solutions of the
aldehydes in DMSO (0.0125 M each) were mixed in equal volumes,
and then 1% TFA (v/v) was added. The mixture was then added to a
0.0100 M solution of 7 or 12 to reach the stoichiometry of 1.25 equiv
of aldehyde per aminooxy group. The resulting solution was diluted
and injected into the mass spectrometer no earlier than 24 h.
1.0 Hz, J ) 8.0 Hz, 2H) 6.68 (t, J ) 2.0 Hz, 1H), 6.33 (d, J ) 1.5
2
Hz, 2H), 6.27 (d, J ) 1.5 Hz, 2H),), 6.22 (d, J ) 1.5 Hz, 2H), 3.79 (s,
2H), 3.74 (s, 2H), 3.47 (br, s, 6H), 3.31 (m, 6H), 2.25 (s, 3H), 1.79 (t,
J ) 8.5 Hz, 3H), 1.53 (t, J ) 9.5 Hz, 3H). 13C NMR (δ ppm, 125
MHz, CDCl3) 171.19 (d, J ) 9.5 Hz), 171.09, 170.92 157.41, 156.30,
155.95, 135.21, 135.15, 135.02, 134.89, 131.18, 130.49, 127.68, 127.23,
124.30, 113.12, 111.16, 102.76, 54.86, 54.22, 53.42, 51.52, 51.44, 51.38,
44.78, 44.75, 43.04, 43.00, 42.97, 42.87, 42.27. ESI MS (CH3CN/water
1/3 v/v) 744.3 ([M + H]+).
Acknowledgment. This authors are grateful to Department
of Medicinal Chemistry, SUNY at Buffalo and NIH (Grant
GM6035201) for financial support of this work.
Supporting Information Available: Synthetic procedures
and analytical data for labeled formyl boronic acids and
compound o-15N-8 (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
Scaffold 12. The compound was synthesized from 11 by using the
1
previously described deprotection procedure.10 Yield 55%. H NMR
(δ ppm, 400 MHz, 1%D2O/CD3CN v/v) 7.39 (dd, J1 ) 1.2 Hz, J2 )
6.8 Hz, 1H), 7.25-7.30 (m, 2H), 7.21(d, J ) 2.8 Hz, 1H), 7.14 (d, J
) 8 Hz, 1H), 6.913 (t,d, J1 ) 1.6 Hz, J2 ) 7.6 Hz, 1H), 6.61 (dd J1 )
JA993844V