NMR studies of the protonation states of pyridoxal-5′-phosphate in water

Article abstract of DOI:10.1016/j.molstruc.2010.03.032

We have measured the ¹³C NMR spectra of the cofactor pyridoxal-5′-phosphate (vitamin B₆, PLP) at 278 K in aqueous solution as a function of pH. By ¹³C enrichment of PLP in the C-4′ and C-5′ positions we were able to measure spectra down to pH 1. From the dependence of the ¹³C chemical shifts on pH, the pK_a values of PLP could be determined. In particular, the heretofore uncharacterized protonation state of PLP, in which the phosphate group as well as the pyridine ring and the phenolic groups are fully protonated, has been analyzed. The corresponding pK_a value of 2.4 indicates that the phosphate group is solely involved in the first deprotonation step. The ¹⁵N chemical shifts of the pyridine ring of PLP published previously are in good agreement with the new results. These shifts contain information about the tautomerism of the different protonation states of PLP. The implications of these findings for the biological function of PLP are discussed.

Full text of DOI:10.1016/j.molstruc.2010.03.032

Journal of Molecular Structure 976 (2010) 282–289
Contents lists available at ScienceDirect
Journal of Molecular Structure
journal homepage: www.elsevier.com/locate/molstruc
NMR studies of the protonation states of pyridoxal-50-phosphate in water
Monique Chan-Huot a, Christiane Niether a, Shasad Sharif a, Peter M. Tolstoy a, Michael D. Toney b,
Hans-Heinrich Limbach a,
*
a Institut für Chemie und Biochemie, Freie Universität Berlin, Takustraße 3, D-14195 Berlin, Germany
b Department of Chemistry, University of California-Davis, 95616 Davis, USA
a r t i c l e i n f o
a b s t r a c t
We have measured the 13C NMR spectra of the cofactor pyridoxal-50-phosphate (vitamin B6, PLP) at 278 K
in aqueous solution as a function of pH. By 13C enrichment of PLP in the C-40 and C-50 positions we were
able to measure spectra down to pH 1. From the dependence of the 13C chemical shifts on pH, the pKa
values of PLP could be determined. In particular, the heretofore uncharacterized protonation state of
PLP, in which the phosphate group as well as the pyridine ring and the phenolic groups are fully proton-
ated, has been analyzed. The corresponding pKa value of 2.4 indicates that the phosphate group is solely
involved in the first deprotonation step. The 15N chemical shifts of the pyridine ring of PLP published pre-
viously are in good agreement with the new results. These shifts contain information about the tautom-
erism of the different protonation states of PLP. The implications of these findings for the biological
function of PLP are discussed.
Article history:
Received 28 January 2010
Received in revised form 8 March 2010
Accepted 8 March 2010
Available online 16 March 2010
Dedicated to Prof. Austin Barnes on the
occasion of his 65th birthday
Keywords:
Vitamin B6
Pyridoxal-50-phosphate
13C Labeling
Protonation states
13C NMR
Ó 2010 Elsevier B.V. All rights reserved.
15N NMR
1. Introduction
The interconversion takes place on the second timescale and hence
contributes two different sets of signals in the NMR spectra [7].
Pyridoxal-50-phosphate (PLP, Fig. 1) is a cofactor of enzymes
that are responsible for various amino acids transformations such
as racemization and transamination [1–3]. A large number of dif-
ferent protonation states and tautomers relevant to its biological
function are found. The aldehyde group can be free, hydrated or
Recently, some of us have explored the acid–base properties of
PLP labeled in the pyridine ring with 15N using 15N NMR spectros-
copy [8]. This method is sensitive to the protonation state of the
pyridine ring which, in turn, depends on the protonation state
and the equilibrium constants KII and KIII of tautomerism in proton-
ation states II and III. These could be measured, as well as several
pKa values. However, no information has been obtained to date
concerning the first protonation state 0, nor has the ratio between
the aldehyde form 1a and the hydrate 1b been quantified in this
state. Therefore, as part of a series of studies concerning the func-
tion of PLP in various environments [8,9] we were interested in
detecting the presence of the first protonation state 0 and deter-
mining its pKa0 value. For that purpose, 13C NMR spectroscopy of
13C labeled PLP is an appropriate stratagem that might also be use-
ful in subsequent studies of model Schiff bases. Therefore, we syn-
thesized PLP labeled at the C-40 and C-50 positions with 13C,
form Schiff bases with primary amines (e.g., enzymic
e-amino
groups of lysine residues or -amino groups of amino acid sub-
a
strates). Furthermore, it contains three functional groups which
may adopt different protonation states: a phosphate group (AH2),
a pyridine ring (B), and an OH group (XH) near the aldehyde
function.
According to Fig. 1 one can conceive five protonation states 0 to
IV, which dominate at different pH values. Besides the fully proton-
ated state 0 = AH2BHXH which contains four protons and the fully
deprotonated state ABX, each protonation state can adopt different
tautomers. The pKa values of protonation states I–IV have been
determined by UV–Vis [4], 1H NMR [5], and 13C NMR [6] spectros-
copy. The aldehyde form is present in the whole pH range whereas
the hydrate is formed only below pH 6, and dominates below pH 4.
13
abbreviated as C2-PLP (1) (Fig. 2). Position C-40 seemed to us to
be the best diagnostic site for the hydration and protonation states
of PLP. For the synthesis we used the procedure of O’Leary and Pay-
ne [10] by which 13C is also introduced into position C-50. This later
proved to be valuable as both positions were needed to determine
the pKa values of 1a and of 1b as a function of pH.
* Corresponding author. Tel.: +49 3083855375; fax: +49 3083855310.
E-mail address: limbach@chemie.fu-berlin.de (H.-H. Limbach).
0022-2860/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.molstruc.2010.03.032

M. Chan-Huot et al. / Journal of Molecular Structure 976 (2010) 282–289
283
OH
O
H
HO
O
P
O
P
C
CH
H
O
a
b
CH2
CH2
OH
CH3
O
O
O
O
HO
HO
N
H
CH3
N
H
PLP aldehyde form 1a
PLP hydrate form 1b
R
R
R
OH
O
O
2-
HO3PO
O PO
HO3PO
3
+
+
+
N
H
CH3
N
CH3
N
CH3
H
H
R
R
AHBHXH
KI
AHBHX
ABHX
O
OH
2-
O PO
3
H2O3PO
0
I
II
III
IV
KII
KIII
p
Ka3
CH3
p
Ka0
p
Ka1
p
Ka2
N
ABX
N
H
CH3
AH2BXH
AHBXH
ABXH
AH2BHXH
R
R
R
OH
OH
OH
2-
H2O3PO
HO3PO
O PO
3
N
CH3
N
CH3
N
CH3
Fig. 1. (a) Aldehyde form 1a and hydrate form 1b of pyridoxal-50-phosphate (PLP). (b) Protonation states of 1a and 1b.
After evaporation of the solvent, 20 ml of water was added to the
resulting oil. The aqueous phase was then washed with dichloro-
methane (3 ꢀ 30 ml). After evaporation of the solvent, the residue
was distilled under vacuum (Bp = 58 °C at 0.3 mbar) to obtain b as a
clear transparent oil. Yield: 2.9 g (17 mmol), 84%, clear transparent
4'
4'
H
HO
O
OH
13C
13CH
O
P
O
H
O
13CH2
13CH2
OH
P
HO
HO
HO
O
O
5'
HO
5'
oil. Rf = 0.88 (DCM:MeOH, 10:0.1). IR (KBr):
m
= 2985, 1685, 1639,
N
CH3
N
CH3
1269, 1200, 1150, 1030, 965, 865, 830, 800 cmꢁ1
.
13C {1H} NMR
(125 MHz, chloroform-d): d = 164.9 (s, COOEt, no 1J(13C,13C) is ob-
served), 129.5 (ABX system, splitting 73 Hz), 60.9 (t, CH2–CH3),
13.7 (q, CH2–CH3). 1H NMR (270 MHz, chloroform-d): d = 6.2
(AA0XX0, 2H, splitting 10 Hz and 6 Hz), 4.2 (q, 4H, CH2–CH3), 1.3
(t, 3J(1H, 1H) = 7.3 Hz, 6H, CH2–CH3). MS (EI): m/z (%) = 174.1
([M]+, 0.1), 145.1 ([MꢁCH2–CH3]+, 4.6), 129.0 ([MꢁO–CH2–CH3]+,
24), 100.0 ([Mꢁ13CO2–CH2–CH3]+).
13C2-PLP (1a)
13C2 -PLP (1b)
Fig. 2. 13C labeled PLP studied in this work. Isotope enrichment either 25% or 100%
in positions C-40 and C-50.
2. Experimental
2.1. Synthesis of 13C enriched PLP
2.1.2. Ethyl-N-formyl-D,L-alaninate (d)
A mixture containing L-alanine (5 g, 40 mmol, 1 eq) and fresh
13C2-PLP was 25% isotopically enriched at the C-40 and C-50 posi-
tions according to Fig. 3, which is a modified version of the synthe-
sis by O’Leary and Payne [10]. The main modification concerned
the formation of the diethyl maleate ester b which was obtained
by direct esterification of the commercially available 1,4-di-13C-
anhydride maleic acid (Eurisotop). In a second modification, p-
toluidine was used as the amine to react with the oxidized pyri-
doxine hydrochloride g instead of p-phenitidine, and finally the
crude product of the phosphorylation reaction was hydrolyzed at
formic acid (8 ml, 220 mmol, 5.5 eq) in dry ethanol (60 ml) was
heated inside a high pressure autoclave HR-100 (made of steel
from Berghof GmbH) at 200 °C for 24 h. At the end of the reaction
a pressure of ꢂ5 bar had built up. After cooling slowly to room
temperature and reaching atmospheric pressure, the solvent was
evaporated under vacuum. The residue was distilled under vacuum
(Bp = 87–90 °C at 0.3 mbar) to obtain d as a clear transparent oil
and was directly used for the next step without storing. Yield:
5.5 g (38 mmol), 95%, clear transparent oil.
room temperature overnight instead of heating under nitrogen.
alanine was commercially available from Aldrich Sigma.
L-
13C {1H} NMR (125 MHz, chloroform-d): d = 172.4 (s, COOEt),
160.7 (d, CHO), 61.4 (t, CH2–CH3), 46.6 (d, NH–CH–CH3), 18.0 (q,
NH–CH–CH3), 13.8 (q, CH2–CH3). 1H NMR (500 MHz, chloroform-
d): d = 5.1 (s, 1H, CHO), 6.9 (b, 1H, NH), 4.5 (qt, 3J(1H,
1H) = 7.3 Hz, 1H, NH–CH–CH3), 4.1 (q, 3J(1H, 1H) = 7.1 Hz, 2H,
CH2–CH3), 1.3 (d, 3J(1H, 1H) = 7.3 Hz, 3H, CH–CH3), 1.2 (t, 3J(1H,
1H) = 7.1 Hz, 3H, CH2–CH3). MS (FAB (+), matrix chloroform-d/m-
NO2–benzyl–OH): m/z (%) = 145.8 ([M + H]+, 48), 72.0 ([MꢁCO2–
CH2–CH3]+, 31), 44.1 ([MꢁCH3–CH–CO2–CH2–CH3]+, 65).
2.1.1. Diethyl-di-13C-maleate ester (b)
A 1:3 mixture of 1,4-di-13C-anhydride maleic acid a (2 g,
20 mmol, isotope enrichment 99%) and non 13C enriched maleic
acid anhydride was dissolved in a 2:1 mixture of dry ethanol
(52 ml) and toluene (26 ml). The reaction was catalyzed with
0.5 ml concentrated H2SO4. The mixture was refluxed overnight.

284
M. Chan-Huot et al. / Journal of Molecular Structure 976 (2010) 282–289
13C O
O
OH
13C
H2N
a
c
O
x HCl
O
H2SO4 cat.
EtOH
HCOOH
EtOH
O
P2O5
N
O
O
O
+
EtO213C 13CO2Et
H
N
H
CH2Cl2
EtO
b
e
d
Δ, P
13CH2OH
13CH2OH
13CO2Et
OEt
LiAlH4 HO
Et2O
N
HCl HO
13CO2Et
O
13CO2Et
13CO2Et
g
N
H
f
N
H
Cl-
Cl-
O
H
13C
N
13CH
13CH2OP
H3PO4/P2O5
HCl
NH2
MnO2
HO
HO
13CH2OH
N
N
H
h
1a
Cl-
Fig. 3. Synthesis of 13C enriched PLP according to O’Leary and Payne [10].
2.1.3. 5-ethoxy-4-methyloxazole (e)
CH2-CH3), 63.2 (t, CH2–CH3), 15.6 (q, C-20), 13.8 (q, CH2–CH3),
13.5 (q, CH2–CH3). 1H NMR (500 MHz, chloroform-d): d = 8.3 (s,
1H, H-6), 4.4 (dq, 3J(1H, 1H) = 7.3 Hz, 3J(1H, 13C) = 2.6 Hz, 4H, CH2–
CH3), 2.9 (s, 3H, H-20), 1.4 (dt, 3J(1H, 1H) = 7.3 Hz, 3J(1H,
13C) = 2.7 Hz, 6H, CH2–CH3). MS (FAB (+), matrix acetone/m-NO2–
benzyl–OH): m/z (%) = 256.3 ([MꢁCl]+, 100), 209.8 (31), 182.0
(36). MS (FAB (–), matrix acetone/m-NO2–benzyl–OH): m/z
(%) = 288.2 ([MꢁH]ꢁ, 2), 254.4 ([MꢁH–HCl]ꢁ, 100), 226.0 (9.65).
Small portions of P2O5 (23.5 g, 160 mmol, 4 eq) were added to a
solution of ethyl-N-formyl-D,L-alaninate (d) (5.5 g, 35 mmol, 1 eq)
in dichloromethane (250 ml). The reaction mixture was refluxed
for 48 h. After cooling to room temperature, a solution of 35%
NaOH in water was added slowly through the refrigerant. The
aqueous phase was then washed with dichloromethane
(3 ꢀ 150 ml). After evaporation of the solvent, the residue was dis-
tilled under vacuum to obtain e as a clear transparent oil. Yield:
2.3 g (18 mmol), 50%, clear transparent oil. IR (KBr):
m
= 3450,
2.1.5. Pyridoxine HCl (g)
2985, 1670, 1515, 1335, 1220, 1130, 1020, 655 cmꢁ1
.
13C {1H}
[40,50-di-13C]-4,5-Bis-ethoxycarbonyl-3-hydroxy-2-methyl-
pyridinium chloride (f) (500 mg, 1.7 mmol, 1 eq) was added via a
Soxhlet extractor overnight to a solution of LiAlH4 (520 mg,
14 mmol, 8 eq) in 100 ml of dry ether. The reaction mixture was
treated with 40 ml of water, added drop-wise initially. The mixture
was filtered through a frit and washed with boiling water (2 ꢀ 60
ml). After evaporation of half of the volume, CO2 was bubbled
through the aqueous clear yellow solution for 1 h. After drying un-
der vacuum, the white powder was extracted with boiling metha-
nol (2 ꢀ 50 ml), and a clear yellow solution is obtained. The solvent
was evaporated and methanolic HCl (10 ml, 3 N) was added to the
yellowish oil. After evaporation of the solvent and drying, pyridox-
ine hydrochloride (g) was obtained as a powder. Yield: 325 mg
(1.6 mmol), 92%, white powder. Rf = 0.5 (unprotonated form), 0.1
NMR (125 MHz, chloroform-d): d = 154.2 (s, C-5), 142.1 (d, C-2),
112.1 (s, C-4), 70.1 (t, CH2–CH3), 14.9 (q, CH2–CH3), 9.8 (q, CH3).
1H NMR (500 MHz, chloroform-d): d = 7.4 (s, 1H, CH), 4.1 (q,
3J(1H, 1H) = 7.1 Hz, 2H, CH2–CH3), 2.0 (s, CH3), 1.3 (t, 3J(1H,
1H) = 7.1 Hz, 3H, CH2–CH3). MS (FAB (+), matrix chloroform-d/m-
NO2–benzyl–OH): m/z (%) = 145.8 ([M + H]+, 48), 72.0 ([MꢁCO2–
CH2–CH3]+, 31), 44.1 ([MꢁCH3–CH–CO2–CH2–CH3]+, 65).
2.1.4. [40,50-di-13C]-4,5-Bis-ethoxycarbonyl-3-hydroxy-2-methyl-
pyridinium chloride (f)
A mixture of 5-ethoxy-4-methyloxazole (e) (500 mg, 3.9 mmol,
and 1 eq) and diethyldi-13C-maleate ester (b) (680 mg, 3.9 mmol,
1 eq) was stirred in a closed vial and heated at 70 °C for 3 days.
The reaction mixture turned orange. After cooling down to room
temperature, the mixture was dissolved in methanolic HCl
(10 ml, 3N) and stirred for 4 h. Pressure crystallization with cold
ether was performed to obtain product (f) as needle crystals after
refrigerating at ꢁ23 °C for 24 h. Yield: 670 mg (2.3 mmol), 59%,
light yellow needles. Rf = 0.8 (DCM:MeOH, 10:0.4). IR (KBr):
(protonated form) (DCM: MeOH, 10: 0.2). IR (KBr):
m = 3325,
2825, 1625, 1545, 1385, 1280, 1215 cmꢁ1 13C {1H} NMR
.
(125 MHz, D2O): d = 155.2, 145.1, 142.8, 139.0 (s, C-2 or C-3 or C-
4 or C-5), 131.9 (d, C-6), 60.4 (t, C-40 or C-50, 3J(13C, 13C) = 4 Hz),
59.1 (t, C-50 or C-40, 3J(13C, 13C) = 4 Hz). 1H NMR (500 MHz, D2O):
d = 8.0 (s, 1H, H-6), 4.9 (d, 1J(1H, 13C) = 150 Hz, 2H, CH2–OH), 4.7
(d, 1J(1H, 13C) = 150 Hz, 2H, CH2–OH), 2.5 (s, 3H, CH3). MS (FAB
(+), matrix H2O/glycerol): m/z (%) = 206.9 ([M]+, 1), 172.3 ([MꢁCl]+,
29), 153.8. MS (FAB (–), matrix H2O/glycerol): m/z (%) = 205.9
([MꢁH]ꢁ, 29), 170.3 ([MꢁH–HCl]ꢁ, 51), 151.8, 129.
m
= 2545, 2045, 1695, 1530, 1265, 1040 cmꢁ1 13C {1H} NMR
.
(125 MHz, chloroform-d): d = 164.5 (s, 3J(13C, 13C) = 1 Hz, COO–
CH2–CH3), 161.8 (s, COO–CH2–CH3), 154.1 (s, C-3), 149.2 (s, C-2),
130.9 (d, C-6), 128.2 (s, C-4 or C-5), 127.8 (s, C-5 or C-4), 64.1 (t,

M. Chan-Huot et al. / Journal of Molecular Structure 976 (2010) 282–289
285
2.1.6. N-(pyridoxylidene)-tolylamine hydrochloride (h)
was used as external reference. Some routine NMR spectra of b
to g were measured using a 270 MHz NMR spectrometer.
Pyridoxine HCl (g) (1.6 g, 7.9 mmol, 1 eq) was dissolved in ca.
30 ml of water then oxidized with manganese dioxide (MnO2)
which was freshly prepared from potassium permanganate
(KMnO4, 1.2 g, 7.6 mmol, 1 eq), sodium bisulfite (NaHSO3, 1.6 g,
15 mmol, 10 eq) and 50% sulfuric acid H2SO4 (4 ml) diluted in
60 ml of water. After stirring for 4 h, the solution was diluted by
adding 500 ml of water and p-toluidine (1.2 g, 11 mmol, 1.4 eq)
was added. The pH was adjusted to 7.5 with a 1 N NaHCO3 solution
and left to stir overnight. The Schiff base was filtered and washed
with water then with ether. The product was dried under high vac-
uum. Yield: 1 g (3.5 mmol), 44%, dark yellow powder. Rf = 0.2
3. Results
A
13C NMR titration of 13C2-PLP in H2O (5 mM) was performed
between pH 1 and 12. Typical {1H}13C spectra obtained are de-
picted in Fig. 4. Only the peaks arising from the isotopically en-
riched carbon sites C-40 and C-50 were analyzed, not the small
peaks arising from the carbon sites containing 13C at natural abun-
dance. 1a gives a signal typical for the aldehyde position C-40,
around 194.5 ppm, and a second signal of equal height around
61.0 ppm. The signal of C-40 of 1b is shifted to about 87 ppm be-
cause of the change of the hybridization at this carbon. By contrast,
the chemical shift of C-50 is altered little. At low pH, the hydrate
form dominates and hence the two signals of 1b are larger than
those of 1a, in agreement with previous 15N NMR studies [8]. By
contrast, the signal of 1a dominates above pH 4. The equilibrium
constants for hydration were calculated from the mole fraction ra-
tios, which were derived by signal integration. The results are
assembled in Table 1.
(DCM:MeOH, 10:0.2). IR (KBr):
m = 3115, 2825, 1405, 1005, 815,
495 cmꢁ1 13C {1H} NMR (125 MHz, DMSO-d6): d = 160.7 (d, C-40,
.
3J(13C, 13C) = 2 Hz), 153.2, 148.2, 144.6, 138.2, 137.7, 133.6, 119.9,
114.0, 110.5 (aromatic carbons), 58.9 (t, C-50, 3J(13C, 13C) = 2 Hz),
18.7 (q, CH3).1H NMR (500 MHz, DMSO-d6): d = 9.17 (s, 1H, H-40),
7.98 (s, 1H, H-6), 7.42 and 7.32 (d, H-aromatic of toluidine part),
4.77 (d, 1J(1H, 13C) = 150 Hz, 2H, CH2-OH), 3.33 (s, 3H, CH3). MS
(FAB (+), matrix H2O/glycerol): m/z (%) = 259.4 ([M + H]+, 6),
153.7, 151.8 ([MꢁN–Toluene–H]+, 100), 116.9. MS (FAB (–), matrix
H2O/glycerol): m/z (%) = 257.4 ([MꢁH]ꢁ, 2), 221.8 ([MꢁH–HCl]ꢁ,
2), 152.8 ([MꢁN–Toluene]ꢁ, 100).
½1bꢃ
Kh
¼
ð1Þ
½1aꢃ
13
2.1.7. C-pyridoxal-50-phosphate (1a)
Table 2 collects all 13C chemical shifts measured at the different
A mixture of phosphorus pentoxide (P2O5, 12 g, 84 mmol, 13 eq)
and phosphoric acid (85%, 16 g, 140 mmol, 21 eq) was prepared
and cooled down to room temperature since the mixing evolves
heat. Subsequently, N-(pyridoxylidene)–tolylamine hydrochloride
(h) (1.6 g, 6.6 mmol, 1 eq) was added. The honey-like mixture
was incubated at 45 °C for 7 h. The reaction was quenched with
an aqueous solution of HCl (3.5 ml, 0.1 M) and left overnight. The
crude product was loaded on an ion exchange column (Amberlite
IR 120) and eluted using water as solvent. After lyophilizing, PLP
was obtained as a yellow solid. Yield: 470 mg, 24%, yellow solid.
pH values. Generally, average chemical shifts of species subject to
different protonation states can be expressed as a function of pH
using the Henderson–Hasselbalch equation [11] adapted for NMR
spectroscopic methods in the fast proton exchange regime [12].
For PLP, which has five protonation states 0 to IV according to
Fig. 1, this equation can be written in the following form [8].
10pH—pK
ai
X
dobs ¼ d0 þ
ðdiþ1 ꢁ diÞ
;
i ¼ 0 to IV
ð2Þ
1 þ 10pH—pK
ai
i
IR (KBr):
m .
= 3365, 2925, 1630 cmꢁ1 13C {1H} NMR (125 MHz,
As usual, pKai represents the pH values where protonation
states i and i + 1 are at the same concentration. di represents the
limiting chemical shift of protonation state i. Eq. (2) is valid for
all nuclei of PLP.
D2O): d = 195.6 (C-40, 3J(13C, 13C) not observable, aldehyde form),
163.6 (d, C-6), 151.3 (s, C-5), 135.4 (s, C-4), 125.5 (s, C-2), 123.6
(s, C-3), 89.4 (d, C-40, 3J(13C, 13C) not observable, acetal form),
61.7 (t, C-30), 15.6 (q, C-20). 1H NMR (500 MHz, D2O): d = 10.4 (d,
The 13C chemical shifts of the aldehyde form 1a and of the hy-
drated form 1b of PLP in Table 2 are plotted in Fig. 5 as a function of
pH. The solid lines were calculated using Eq. (2), optimizing the pKa
values and the limiting 13C chemical shifts. The results are given in
Tables 3 and 4.
13
1J(1H, C) = 185 Hz, 1H, H-40 – aldehyde form), 8.1 (s, 1H, H-6 –
aldehyde form), 6.4 (d, 1J(1H, C) = 170 Hz, 1H, H-40 – hydrate
13
form), 5.2 (dd, 1J(1H, 13C) = 150 Hz, 2J(1H, H) = 5 Hz, 2H, H-30 –
1
aldehyde form), 5.1 (dd, 1J(1H, 13C) = 150 Hz, 2J(1H, 1H) = 5 Hz, 2H,
H-30 – hydrate form), 2.6 (s, 3H, H-20 – aldehyde form), 2.5 (s,
3H, H-20 – hydrate form).
4. Discussion
2.2. Sample preparation
In the first part of this section, we discuss the pH-dependent
equilibrium between the aldehyde form 1a and the hydrate form
1b. In the second part, we discuss the different protonation states
observed here, and compare these results with those obtained pre-
viously [8] using 15N NMR of PLP labeled with 15N in the pyridine
ring.
Aqueous solutions of PLP (5 mM) were prepared using water or
heavy water degassed and stored under argon in order to remove
oxygen and carbon dioxide. The pH values of the solutions were
adjusted before each spectroscopic measurement by addition of
degassed 3 M, 1 M or 0.1 M sodium hydroxide or hydrochloric acid
solutions. For that purpose we used a HANNA HI 9025 pH meter
equipped with a HAMILTON Spintrode P electrode. The pH values
were measured after the experiments and showed an average error
of 0.15.
4.1. Equilibrium between aldehyde and hydrate form
In Fig. 6, the mole fractions of the aldehyde form 1a and of the
hydrated form 1b are plotted as a function of pH. Qualitatively, it
has been known for a long time that 1b dominates at low pH
and 1a at high pH. However, the 13C experiments here allowed
us to elucidate the pH of 4.2 where the equilibrium constant of
hydration Kh is unity.
2.3. Spectroscopic methods
NMR spectra of pyridoxal-50-phosphate were measured using a
Bruker AMX 500 spectrometer (500.13 MHz for 1H, 125 MHz for
13C) at 278 K. The 13C spectra were recorded in the inverse gated
1H-decoupled mode using H2O as solvent, with field locking using
a D2O containing capillary. The recycle delay was set to 10 s. TMS
A consequence of the hydration equilibrium is that it is difficult
13
to determine the C chemical shifts of C-501a at low and of 1b at
high pH.

286
M. Chan-Huot et al. / Journal of Molecular Structure 976 (2010) 282–289
pH
12
C-4’
C-5’
9
7
6
5
4
3
2
1
1a
1b
40
120
δ13C(ppm)
100
80
60
200
180
160
140
Fig. 4. {1H}13C NMR spectra of 13C2-PLP (1) in H2O at 278 K as a function of pH.
Table 1
Table 2
Mole fractions of the aldehyde 1a and hydrate 1b obtained by 13C NMR.
13
13C chemical shifts in ppm of C2-PLP at the position C-40 and C-50.
C-40
C-50
pH
Mole fraction
Kh
pH
C-40 1a
C-40 1b
1a
1b
1a
1b
1.2
2.0
3.1
3.6
3.9
5.0
5.9
7.0
8.0
0.13
0.14
0.19
0.32
0.41
0.77
0.83
0.87
0.92
0.97
0.96
0.96
1.0
0.87
0.86
0.81
0.68
0.59
0.23
0.17
0.13
0.08
0.03
0.04
0.04
0
6.70
6.03
4.2
1.0
1.5
2.0
3.2
3.6
3.9
5.0
5.9
7.0
8.0
9.0
10.1
11.1
12.0
193.5
193.5
193.6
193.9
194.5
194.7
195.0
195.2
195.7
195.9
196.0
195.9
196.0
196.0
86.5
86.5
86.6
86.6
87.0
87.1
87.6
88.1
88.4
88.6
89.1
89.8
89.7
89.7
60.1
60.1
60.2
60.2
60.4
60.5
60.5
60.3
59.8
59.8
60.1
61.0
62.1
62.3
2.15
1.42
0.30
0.21
0.15
0.09
0.03
0.04
0.04
0
60.8
61.0
61.2
61.7
61.6
61.1
61.2
61.7
61.9
61.9
61.9
9.0
10.1
11.1
12.0
4.2. Determination of the pKa values of PLP by 13C NMR
and of 1b at high pH. However, all four pKa values could be deter-
mined by the combination of the spin probes C-40 and C-50.
The results of our 13C NMR titration study of PLP in aqueous
solution are illustrated in Fig. 5. Two 13C spin probes were present,
at C-40 and C-50 (Fig. 2). For the aldehyde form 1a, the solid lines
reproduce well the experimental data. Between two pKa values,
illustrated by vertical dashed lines, the chemical shifts exhibit little
dependence on pH, while near the pKa values a strong dependence
is observed. Position C-40 is most diagnostic both in the case of 1a
and of 1b since its signals can be observed even when the corre-
sponding species represents a minor form. An exception is 1a at
high pH: the chemical shifts of III and IV are very similar, but the
corresponding pKaIII value could be well determined from the sig-
nal of position C-50. Problems arise with the determination of the
chemical shifts of C-50 of the minor forms, i.e. of 1a at low pH
4.3. Determination of the equilibrium constants of tautomerism of PLP
by 15N NMR
In a previous study [8], Eq. (2) was used to analyze the pH-
dependent 15N chemical shifts of PLP labeled with 15N in the pyri-
dine ring. It was shown that the chemical shift di of a given proton-
ation state i obtained by the Henderson–Hasselbalch fit of the
experimental data to Eq. (2) depends on the equilibrium constants
Ki of the OH/NH tautomerism (Fig. 1b) given by
1
Ki
1 þ Ki
di ¼
dNH
þ
dN
ð3Þ
1 þ Ki

M. Chan-Huot et al. / Journal of Molecular Structure 976 (2010) 282–289
287
δ13C-5’ 1a
δ 13C-4’ 1a
a
b
62.0
61.5
61.0
60.5
60.0
59.5
196.0
195.5
195.0
194.5
194.0
193.5
193.0
192.5
0
I
II
III
IV
0
I
II
III
IV
0
2
4
6
8
10
12
14
0
2
4
6
8
10
12
14
pH
pH
δ13C-4’ 1b
δ13C-5’ 1b
c
d
62.5
62.0
61.5
61.0
60.5
60.0
59.5
59.0
90.5
90.0
89.5
89.0
88.5
88.0
87.5
87.0
86.5
0
I
II
III
IV
0
I
II
III
IV
0
2
4
6
8
10
12
14
0
2
4
6
8
10
12
14
pH
pH
13
Fig. 5. C chemical shifts of the C-40 and C-50positions of the aldehyde form 1a and the hydrate form 1b in water as a function of pH. The dashed vertical lines indicate pKa
values. The solid lines were calculated using the Henderson–Hasselbalch equation (2) and the parameters reported in Tables 2 and 3.
Table 3
pKa values of PLP in water.
1.0
1a
PLP
pKa0
pKaI
pKaII
pKaIII
0.8
0.6
0.4
0.2
0.0
1a
15N NMR correcteda
13C-40 NMRb
13C-50 NMRb
Literature
2.4
2.4
n.o.
3.6
3.6
6.4
6.4
6.4
6.1d
8.3
n.o.
3.6
8.3
3.1–3.7c
8.3–8.9e
1b
15N NMR
13C-40 NMR
13C-50 NMR
Literature
n.o.
2.1
2.1
4.4
4.4
4.4
4.1c
n.o
6.2
6.2
6.1b
8.7
8.7
8.7
8.7c
1b
n.o.: Not observed.
a
2
4
6
8
10
12
Ref. [8].
This study.
Refs. [4,14].
Refs. [2,15].
b
pH
c
d
Fig. 6. Mole fractions evaluated by integration of the C-40 signals of the aldehyde
form 1a and of the hydrate form 1b of PLP in water at 278 K as a function of pH.
e
Ref. [16].
dNH and dN represent the chemical shifts of the protonated and
deprotonated ring nitrogen, respectively. The two values are very
different, but exhibit little dependence on the protonation state.
Thus, only if Ki differs for two adjacent protonation states, the cor-
responding pKa value can be obtained by 15N NMR.
Table 4
Limiting 13C and 15N NMR chemical shifts (in ppm) of PLP obtained from Fig. 5.
PLP
d0
dI
dII
dIII
dIV
By inspection of Fig. 1b, it is clear that KI is very small, i.e. the
assumption is plausible [8] that the tautomeric form AH2BXH of
protonation state I is not present in view of the large acidity of
phosphoric acid. For protonation state IV the equilibrium does
not exist. Protonation state 0 was not considered previously [8],
but it is clear that only a single tautomer can be formed. On the
other hand, KII and KIII could be obtained previously [8] by fitting
the 15N chemical shifts to Eqs. (2) and (3).
1a
15N
C-40
C-50
170.4
193.4
60.0
184.4
194.0
60.0
198.4
195.1
61.8
199.4
196.0
60.8
264
196.0
61.9
1b
15N
C-40
C-50
162.6
86.5
60.1
162.6
86.5
60.4
168.1
87.9
60.5
168.1
88.3
59.6
264
89.6
62.2

288
M. Chan-Huot et al. / Journal of Molecular Structure 976 (2010) 282–289
By contrast, it is difficult to derive the equilibrium constants of
O
O
H
H
O
P
O
P
tautomerism Ki of a given protonation state i from its average 13C
chemical shift di listed in Table 4, as the 13C chemical shift changes
caused by the tautomerism depend on the protonation state, in
contrast to the 15N chemical shifts. However, by the combination
of 13C and 15N NMR this problem can be solved. We took the pKa
values obtained by 13C NMR (dashed vertical lines) and reanalyzed
the 15N chemical shifts reported previously [8]. The results are de-
picted in Fig. 7. The solid curves were calculated using the chemical
shifts dNH and dN (dashed horizontal lines) as well as the equilib-
rium constants of tautomerism of protonation states II and III of
1a and 1b, KII and KIII previously [8] (see also caption of Fig. 7).
For reference, we added the dashed curves which were calculated
assuming that the equilibrium constants of tautomerism KII =
KIII = 0 (lower curve) or infinite (upper curve).
C
C
N
H
O
K=0.4
O
O
O
O
O
O
O
RO
RO
N
H
CH
CH
3
3
II: R = H
III: R = –
OH
OH
OH
HO
HO
O
P
O
P
K=0.06
O
O
O
RO
RO
N
H
CH
N
CH
3
3
Fig. 8. Tautomerism of 1a and of 1b in protonation states II and III. For further
discussion see text.
4.4. Acid–base properties of PLP and biological function
What do the pKa values and equilibrium constants of tautomer-
ism tell us about the acid–base properties of the two forms of PLP?
At pH values below 2.4 both 1a and 1b are fully protonated, i.e. the
protonation state 0 (Fig. 1b) is formed. This state has not been ob-
served before. The new pKa0 values of 2.4 for 1a and of 2.1 for 1b
are very close to the corresponding value of 2.15 for H3PO4 [13].
Thus, when pH is increased, at pH 3 the phosphoric group is mono-
anionic in 1a and 1b, and protonation state I is formed, but the pyr-
idine ring still protonated. When pH is further increased, first 1a
looses a second proton below and 1b above pH 4, i.e. protonation
state II is formed. This is also the region where the hydrate 1b
looses its dominance. Thus, 1b dominates only in protonation state
I, whereas 1b dominates in the higher protonation states. In a nar-
row pH region around 4, the conversion of 1a to 1b is then associ-
ated with a protonation, i.e. a conversion from II to I. By 15N NMR it
is shown that the tautomeric equilibrium of 1b is shifted almost
entirely into the direction of the protonated pyridine form,
whereas 1a forms a substantial amount of the deprotonated pyri-
dine form. This observation can be rationalized in terms of a stron-
ger intramolecular OHO-hydrogen bond in 1a as compared to 1b as
illustrated in Fig. 8. We speculate that this hydrogen bond plays a
decisive role for the dominance of 1a at higher pH, but not at low
pH. At physiological pH 7.35 1b is negligible, i.e. protonation state
III of 1a seems to be of most physiological importance. The remain-
ing proton is almost equally distributed between the pyridine ring
and the phenolic function. Complete deprotonation and formation
of state IV does not seem to play a physiological role.
a
280
δN
1a
260
240
220
200
180
160
140
0
I
II
III
IV
δNH
δ
0
2
4
6
8
10
12
14
pH
280
260
240
220
200
180
160
b
δN
1b
However, in the active site of PLP dependent enzymes the acid–
base behavior can be modified compared to the aqueous environ-
ment. Thus, recently [9d] it was shown that the Schiff base formed
by PLP with the e-amino group of a lysine residue in aspartate ami-
0
I
II
III
IV
notransferase behaves as if it were embedded in a highly polar but
aprotic organic solvent. In this environment, the acid–base proper-
ties of PLP are unique. Thus, further studies of the interaction of
PLP with amines in water, model environments and in enzymes
will be fruitful in the future.
δ
δNH
5. Conclusions
Using 13C NMR of the cofactor pyridoxal-50-phosphate (PLP) la-
13
beled with C in positions C-40 and C-50 it has been possible to
0
2
4
6
8
10
12
14
pH
determine the pKa values of the different protonation states 0 to
IV depicted in Fig. 1b for the aldehyde form 1a as well as for the
hydrate form 1b. The results are in good agreement with those
published previously as illustrated by Table 3. Here, for the first
time, measurements could be performed at lower pH where PLP
is not very soluble. Because of the 13C spin probes which exhibit
a high signal intensity and a high sensitivity to pH changes we
were able to detect protonation state 0 which is of the AH2BHXH
type. Furthermore, we speculate that the dominance of 1a at high-
Fig. 7. 15N chemical shifts measured previously [8] of the 15N-labeled pyridine ring of
1a and 1b as a function of pH. The dashed vertical lines indicate pKa values, the dashed
horizontal lines the limiting 15N chemical shifts of the non-protonated and the
protonated pyridine ring. The solid curves were obtained by using the pKa values of
PLP determined here by 13C NMR (Table 3) and the following parameters determined
previously [8]: dN(1a) = dN(1b) = 264 ppm, dNH(1a) = 170.4 ppm, dNH(1b) = 162.6
ppm (dashed horizontal lines), KII(1a) = KIII(1a) = 0.4, KII(1b) = KIII(1b) = 0.06. The
dashed curves were calculated assuming that the equilibrium constants of tautom-
erism KII = KIII = 0 (lower curve) or infinite (upper curve).

M. Chan-Huot et al. / Journal of Molecular Structure 976 (2010) 282–289
289
[6] (a) H. Witherup, E.H. Abbott, J. Org. Chem. 40 (1975) 2229–2233;
(b) C. Harruff, W.T. Jenkins, Org. Magn. Res. 8 (1976) 548–557.
[7] M.L. Ahrens, G. Maass, P. Schuster, H. Winkler, J. Am. Chem. Soc. 92 (1970)
6134–6139.
[8] S. Sharif, M. Chan-Huot, P.M. Tolstoy, M.D. Toney, K.H.M. Jonsson, H.H.
Limbach, J. Phys. Chem. B 111 (2007) 3869–3876.
[9] (a) S. Sharif, G.S. Denisov, M.D. Toney, H.H. Limbach, J. Am. Chem. Soc. 129
(2007) 6313–6327;
er pH is associated with the stabilizing effect of the intramolecular
OHO-hydrogen bond, and that this effect is less important in pro-
tonation states 0 and I. The new results lead to a consistent picture
of the protonation states of PLP in aqueous environments. The
present work will guide future work aimed at understanding better
the interaction of PLP with amines in model systems and in en-
zyme environments.
(b) S. Sharif, D. Schagen, M.D. Toney, H.H. Limbach, J. Am. Chem. Soc. 129
(2007) 4440–4455;
(c) S. Sharif, G.S. Denisov, M.D. Toney, H.H. Limbach, J. Am. Chem. Soc. 128
(2006) 3375–3387;
Acknowledgements
(d) S. Sharif, E. Fogle, M.D. Toney, G.S. Denisov, I.G. Shenderovich, P.M. Tolstoy,
M. Chan-Huot, G. Buntkowsky, H.H. Limbach, J. Am. Chem. Soc. 129 (2007)
9558–9559.
This work has been supported by the Deutsche Forschungs-
gemeinschaft and the Fonds der Chemischen Industrie, Frankfurt.
[10] M.H. O’Leary, J.R. Payne, J. Biol. Chem. 251 (1976) 2248–2254.
[11] (a) R. De Levie, J. Chem. Educ. 80 (2003) 146;
(b) H.N. Po, N.M. Senozan, J. Chem. Educ. 78 (2001) 1499–1503.
[12] F. Blomberg, W. Maurer, H. Rüterjans, J. Am. Chem. Soc. 99 (1977) 1849–1859.
[13] J.S. Elliot, R.F. Sharp, L. Lewis, J. Phys. Chem. 62 (1958) 686–689.
[14] (a) O. Manousek, P. Zuman, Collect. Czech. Chem. Commun. 29 (1964) 1432–
1457;
(b) P. Zuman, O. Manousek, Collect. Czech. Chem. Commun. 29 (1961) 2134–
2143;
(c) O. Manousek, P. Zuman, Biochim. Biophys. Acta 44 (1960) 393–394.
[15] C.M. Harris, J. Johnson, D.E. Metzler, Biochim. Biophys. Acta 421 (1976) 181–
194.
References
[1] D.E. Metzler, Biochemistry. The Chemical Reactions of Living Cells, vol. 1,
Academic Press, New York, 1977. pp. 444–461.
[2] P. Christen, D.E. Metzler, Transaminases, first ed., John Wiley & Sons, New York,
1985.
[3] E.E. Snell, S.J. Di Mari, The enzymes, in: P.D. Boyer (Ed.), Kinetics and
Mechanism, vol. 2, third ed., Academic Press, New York, 1970, pp. 335–362.
[4] D.E. Metzler, E.E. Snell, J. Am. Chem. Soc. 77 (1955) 2431–2437.
[5] (a) W. Korytnyk, R.P. Singh, J. Am. Chem. Soc. 85 (1963) 2813–2817;
(b) W. Korytnyk, H. Ahrens, Methods Enzymol. 18 (1970) 475–483.
[16] J.W. Bridges, D.S. Davies, R.T. Williams, Biochem. J. 98 (1966) 451–468.