Novel ring chemistry of vitamin B6 with singlet oxygen and an activated ene: Isolated products and identified intermediates suggesting an operable [3 + 2] cycloaddition

Article abstract of DOI:10.1039/c2ob26067k

Pyridoxine reaction with ¹O₂ in aqueous solution at neutral pH resulted in oxidation at the 2- and 6-positions of the pyridine ring and unprecedented ring contraction. Kinetic and low temperature studies provided observable intermediates by NMR spectroscopy. In addition, novel cycloaddition between pyridoxine and N-methylmaleimide, without N-alkylation and in water, suggest a common [3 + 2] cycloaddition with the 3-hydroxypyridine ring.

Full text of DOI:10.1039/c2ob26067k

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Organic &
Biomolecular
Chemistry
Cite this: Org. Biomol. Chem., 2012, 10, 7278
www.rsc.org/obc
COMMUNICATION
Novel ring chemistry of vitamin B6 with singlet oxygen and an activated ene:
isolated products and identified intermediates suggesting an operable [3 + 2]
cycloaddition†
David Samuel, Kirsten Norrell and David G. Hilmey*
Received 3rd June 2012, Accepted 7th August 2012
DOI: 10.1039/c2ob26067k
1
substrate and utilize the 3-hydroxypyridine ring as an electron
sink to catalyze select group of chemical reactions
Pyridoxine reaction with O2 in aqueous solution at neutral
pH resulted in oxidation at the 2- and 6-positions of the pyri-
dine ring and unprecedented ring contraction. Kinetic and
low temperature studies provided observable intermediates
by NMR spectroscopy. In addition, novel cycloaddition
between pyridoxine and N-methylmaleimide, without
N-alkylation and in water, suggest a common [3 + 2] cyclo-
addition with the 3-hydroxypyridine ring.
a
including decarboxylation, transamination, racemization, and
deamination.4
In recent studies, vitamin B6 has been implicated as an anti-
oxidant, countering the known effects of various reactive oxygen
species (ROS).5,6 For example, pyridoxamine has been shown to
inhibit the formation of 4-hydroxynonenal (4-HNE) and malon-
aldehyde (MDA), markers for lipid peroxidation and protein
glycosylation in red blood cells.7,8 A. thaliana and the fungus
C. keidi have been shown to upregulate vitamin B6 expression
under singlet oxygen (1O2) induced stress.9,10
Introduction
Specifically looking at singlet oxygen-related stress, pyridox-
1
ine has been shown to be an efficient quencher of O2 with a
Vitamin B6 (Fig. 1), in its active form pyridoxal-5′-phosphate
(PLP), is a well known cofactor necessary for primary metab-
olism in both eukaryotic and prokaryotic organisms.1 It has been
estimated that 1.5% of the genes in a typical E. coli bacterium
encode PLP-dependent enzymes.2 These enzymes function in a
variety of settings including amino acid metabolism, hormone
biosynthesis, neurotransmitter biosynthesis, glycogen degra-
dation, and lipid biosynthesis.3 Much of the chemistry of PLP
relates to its ability to form a Schiff base with a potential
second order rate constant of 107 to 108 M−1 s−1 under high
intensity light stress.6,11 Of particular interest, endoperoxide 2
(Scheme 1) was observed when a di-TBS analog of pyridoxine
was oxidized at reduced temperatures in CD2Cl2 and shown to
be in equilibrium with the hydroperoxide 1.11
Scheme 1 Products by NMR of singlet oxygen induced oxidation of a
di-TBS protected pyridoxine analog.11
Results and discussion
We report novel products of reaction between pyridoxine and
singlet oxygen in protic solvents at room and reduced tempera-
tures. The isolated final product is an unprecedented ring con-
tracted α-hydroxyketone. Probing the reaction mechanism by
intermediate identification uncovered a six-membered ring
keto-lactam through time-dependent NMR studies. Temperature-
dependent experiments also revealed a transient endoperoxide
bicycle in methanol. Pyridoxine was also found to undergo a
previously undescribed reaction with N-methylmaleimide under
similar conditions providing an analogous bicyclic product. We
Fig. 1 All forms of vitamin B6.
Department of Chemistry, St. Bonaventure University, St. Bonaventure,
NY 14778, USA. E-mail: dhilmey@sbu.edu; Fax: +1 716 375 7618;
Tel: +1 716 375 2603
†Electronic supplementary information (ESI) available: Experimental
detail, 1D and 2D NMR data, and X-ray structure files. CCDC
894615–894616. For ESI and crystallographic data in CIF or other elec-
tronic format see DOI: 10.1039/c2ob26067k
7278 | Org. Biomol. Chem., 2012, 10, 7278–7281
This journal is © The Royal Society of Chemistry 2012

View Article Online
propose that both reactions proceed through a formal or con-
certed [3 + 2] dipolar cycloaddition.
investigation of the underlying pathway associated with this par-
ticular transformation.
In order to determine if product 4 was an artifact of purifi-
cation, we observed its formation directly by NMR. The exper-
iment involved a time course run in one hour intervals in D2O
under the same conditions as in Scheme 2. Fig. 3 illustrates the
reaction progression over a total of three hours. Pyridoxine is
converted to 4, with an additional intermediate formed. 2D
NMR analysis of this compound supports lactam 5 (Fig. 4). The
data obtained through HMBC analysis shows long range coup-
ling between the 2-position carbonyl (162.7 ppm) and the
3′-methylene as well as coupling between the 5-position
carbonyl (196.2 ppm) and the 6′-methyl at 1.59 ppm. Integration
of corresponding peaks indicate a lag in the production of 4 rela-
tive to 5 suggesting the initial formation of 5 with subsequent
conversion to 4.
When the reaction solvent was changed to CH3OH, product 6
was isolated; identical to that described by Foote (Scheme 3A).11
The similarity between compounds 5 in water and 6 from metha-
nol is noteworthy, as is the inability of 6 to contract to compound
4 in CH3OH. With analogous products of singlet oxygen
addition formed in CH3OH and aqueous buffer, reduced temp-
erature NMR studies were carried out in methanol-d4 to observe
the initial adduct between 1O2 and pyridoxine (Scheme 3B).
When the reaction was run at −78 °C in CD3OD, a single
product was seen by NMR. The endoperoxide 7 matches the
obtained NMR data. Fig. 5 illustrates the 13C and pertinent 2D
data. The pyridoxine derived bicycle 7 is analogous to com-
pound 2 observed by Foote in CD2Cl2 and closely matches the
reported NMR spectra.11 Of particular note is the peak at
5.93 ppm representing the C1 proton. This signal is coupled to
both alkene carbons as well as the opposite bridgehead carbon.
No analog to the reported hydroperoxide 1 was found in this
experiment indicating no observable equilibrium between an
open hydroperoxide adduct and the observed endoperoxide
bicycle 7.
Singlet oxygen addition to pyridoxine
1
Pyridoxine was subjected to O2 through the photosensitization
of rose bengal (RB) in aqueous solution.12 Reactions performed
in phosphate buffer, NH4OAc buffer, and in unbuffered water,
all at near neutral pH, produced the same product shown in
Scheme 2. These results suggest that the salt content of the
aqueous solution is not important to the reaction outcome. The
crystalline compound 4, produced as the major product in 43%
yield, was characterized by 1D and 2D NMR and then further
confirmed by X-ray crystallography (Fig. 2). When the reaction
was run without RB or light, no reaction was observed over a
24 h time period when monitored by TLC or H NMR.13 The
ring contracted product shows additional oxygenation at the orig-
inal 2- and 6-positions. This unusual ring contraction led to an
1
Scheme 2 Isolated ring contraction product of 1O2 treated pyridoxine
in buffered water in the presence of 510 nm filtered light for 4 h.
An analogous NMR experiment replacing CD3OD with
CH3OH did not support the incorporation of an OCH3 from
solvent into the intermediate, further supporting 7 (ESI†). A
temperature-dependent experiment was conducted to observe the
Fig. 2 ORTEP plot of lactam product 4.
1
stability of 7 (Fig. 6). H NMR reveals that as the reaction is
warmed, 7 is stable to −40 °C. It then converts to a set of com-
pounds from −30 to −10 °C. Eventually the transient intermedi-
ate converts to one major product, 8, at room temperature,
which, again matches the final product previously described.11
Fig. 3 1H NMR kinetic study of pyridoxine in the presence of singlet
oxygen shown in 1 h intervals. Signals are labelled according to the
reaction scheme in (A). After one hour, the main product is compound
5. Over three hours, 3 is converting to 4 (green) and 5 (blue).
Fig. 4 Summary of data for compound 5. 13C NMR of mixture of 4
and 5 and summary of HMBC data for 5 presented in 2D spectra.
Summary of 1H NMR and HMQC data for 5 listed in columns.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 7278–7281 | 7279

View Article Online
Scheme 4 3,2-Dipolarcycloaddition of 3-hydroxypyridinium betaines
with electron poor ene.
Scheme 3 (A) Reaction of pyridoxine in methanol. (B) In CD3OD the
bicyclic adduct 7 represents the product of the clean and specific conver-
sion at reduced temperatures after 4 h.
Addition of an electron deficient ene to unactivated pyridoxine
Scheme 5 Successful [3
aqueous buffer.
+ 2] cycloaddition with pyridoxine in
1
Imagining a mechanism for endoperoxide formation with O2,
pyridoxine’s calculated and experimentally determined zwitterio-
nic nature in pH 7 aqueous solvent (Fig. 7) presents an electronic
arrangement which could both donate and accept electronic
density in the aromatic ring.14,15 This electronic arrangement is
analogous to N-alkylated 3-hydroxypyridines under basic con-
ditions, which are utilized in the [3 + 2] dipolar cycloaddition of
electron deficient enes (Scheme 4).16 This zwitterionic structure,
along with the identification of structure 7 provides mechanistic
clues suggesting a possible formal or concerted [3 + 2] cyclo-
addition reaction between pyridoxine and singlet oxygen at
physiological pH. The endoperoxide 7 would be the product of
such a reaction.
Cycloadditions with singlet oxygen are well known17 and a
mechanistic hypothesis has been suggested in a previous compu-
tational study on singlet oxygen addition to pyridoxine, but not
investigated.18 Similar bicyclic endoperoxides, through this
1
hypothesized mechanism, have been reported though in the O2
oxidation of pyrroles and furans.19
To further validate a proposed [3 + 2] cycloaddition mechan-
ism, pyridoxine was treated with N-methylmaleimide at room
temperature in pH 7.5 aqueous phosphate buffer (Scheme 5).
Surprisingly, a single diastereomer was isolated in 63% yield
and determined by X-ray crystallography to be the exo product
9. The hypothesized zwitterionic nature of pyridoxine in neutral
water electronically allows the observed cycloaddition, support-
ing the ability of vitamin B6 to react specifically at the 2- and
6-position. When the reaction was extended to 3-hydroxypyri-
dine, no product was observed.20 The apparent superiority of the
pyridoxine structure will warrant additional investigation of the
pyridine substitution necessary for robust reaction. It is note-
worthy that this is the first described cycloaddition reaction
between a 3-hydroxypyridine and ene in water and the first
without prior nitrogen alkylation.
Fig. 5 Summary of NMR data for endoperoxide 7 at −78 °C in
CD3OD. 1H and HMQC data for 7 listed in right columns. 13C and
HMBC data for 7 presented in 2D spectral cutouts.
Given the intermediates and products formed, a mechanistic
hypothesis is proposed for 1O2 quenching by pyridoxine
(Scheme 6) consistent with the Kornblum–DeLaMare rearrange-
ment.21 Initial cycloaddition with the pyridoxine zwitterion pro-
vides the endoperoxide bicycle and is followed by proton
removal from the bridgehead with concomitant peroxide clea-
vage and carbonyl formation. The resulting lactam (capable of
exchanging in methanol to provide product 8) then opens and
recloses on the 1,2-diketone to produce the ring contracted
product. The nature of ring contraction in aqueous solutions is
uncertain and may proceed through a stepwise or concerted
mechanism. Additional work has begun looking at similar reac-
tions with the other forms of vitamin B6 and in determining the
source of oxygen atoms in the final product.
Fig. 6 Intermediate 7 (peaks associated shown in red) appear stable to
temperatures of −40 °C. An unidentified intermediate (blue) appears at
approximately −30 °C but disappears at room temperature. The final
product, 8 (green), appears at −30 °C and grows until it is the major
product at room temperature.
Fig. 7 Experimentally and computationally determined charge state of
pyridoxine at differing pH values.
7280 | Org. Biomol. Chem., 2012, 10, 7278–7281
This journal is © The Royal Society of Chemistry 2012

View Article Online
Notes and references
1 C. K. Mathews, K. E. Van Holde and K. G. Ahern, Biochemistry,
Addison Wesley Longman, Inc., San Francisco, CA, 3rd edn, 2000.
2 R. Percudani and A. Peracchi, EMBO Rep., 2003, 4, 850–854.
3 J. McMurry, Organic Chemistry with Biological Applications 2e. Brooks/
Cole, Belmont, CA, 2011.
4 J. McMurry and T. Begley, The Organic Chemistry of Biological Path-
ways, Roberts and Company Publishers, Englewood, CO, 2005.
5 M. Ehrenshaft, P. Bilski, M. Y. Li, C. F. Chignell and M. E. Daub, Proc.
Natl. Acad. Sci. U. S. A., 1999, 96, 9374–9378.
6 P. Bilski, M. Y. Li, M. Ehrenshaft, M. E. Daub and C. F. Chignell, Photo-
chem. Photobiol., 2000, 71, 129–134; J. Natera, W. Massad and
N. A. Garcia, Photochem. Photobiol. Sci., 2012, 11, 938–945.
7 J. M. Onorato, A. J. Jenkins, S. R. Thorpe and J. W. Baynes, J. Biol.
Chem., 2000, 275, 21177–21184; S. K. Jain and G. Lim, Free Radical
Biol. Med., 2001, 30, 232–237.
Scheme 6 Hypothesized [3 + 2] cycloaddition of singlet oxygen to
pyridoxine in protic solvents.
8 R. G. Khalifah, P. Todd, A. A. Booth, S. X. Yang, J. D. Mott and
B. G. Hudson, Biochemistry, 1996, 35, 4645–4654.
9 M. Havaux, B. Ksas, A. Szewczyk, D. Rumeau, F. Franck, S. Caffarri
and C. Triantaphylidès, BMC Plant Biol., 2009, 9, 130–152.
10 M. Ehrenshaft, A. E. Jenns, K. R. Chung and M. E. Daub, Mol. Cell,
1998, 1, 603–609.
Conclusions
Pyridoxine has been shown to undergo a hypothesized [3 + 2]
cycloaddition reaction to give bridged bicycles with singlet
oxygen and N-methylmaleimide. The reactivity, in aqueous
buffer, is believed to proceed through the pyridoxine zwitterion
leading to unprecedented reaction with the 3-hydroxypyridine
ring. The described reactions are the first reported products of
1O2 reaction in aqueous systems with production of a novel ring-
contracted lactam. We also describe a novel product of reaction
between pyridoxine and maleimide as the first dipolar cyclo-
addition with a 3-hydroxypyridine in water and without prior
nitrogen alkylation.
The authors would like to acknowledge support from The
Arnold T. Borer Fellowship and Clare College, St. Bonaventure
University. Special thanks to Jeremiah W. Hanes for helpful dis-
cussions and Bill Brennessel at the University of Rochester for
X-ray crystallography support.
11 B. K. Ohta and C. S. Foote, J. Am. Chem. Soc., 2002, 124, 12064–12065.
12 D. C. Neckers, J. Photochem. Photobiol., A, 1989, 47, 1–29.
1
13 Data not shown. Starting material observed by TLC and H NMR analy-
sis of reaction mixture.
14 M. Ristila, J. M. Matxain, A. Strid and L. A. Eriksson, J. Phys. Chem. B,
2006, 110, 16774–16780.
15 D. E. Metzler and E. E. Snell, J. Am. Chem. Soc., 1955, 71, 2431.
16 A. R. Katritzky and N. Dennis, Chem. Rev., 1989, 89, 827–861.
17 E. L. Clennan, Tetrahedron, 2000, 56, 9151–9179.
18 J. M. Matxain, M. Ristil, I. Strid and L. A. Eriksson, Chem.–Eur. J.,
2007, 13, 4636–4642.
19 K. Gollnick and A. Griesbeck, Tetrahedron, 1985, 41, 2057–2068;
M. N. Alberti, G. C. Vougioukalakis and M. Orfanopoulos, J. Org.
Chem., 2009, 74, 7274–7282.
20 NMR monitoring with 3-hydroxypyridine and N-methylmaleimide
resulted in no reaction and recovery of starting material during 24 h
period.
21 N. Kornblum and H. E. DeLaMare, J. Am. Chem. Soc., 1951, 73,
880–881.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 7278–7281 | 7281