Concerning the photodiastereomerization and protic equilibria of urocanic acid and its complex with human serum albumin

Article abstract of DOI:10.1007/s00706-004-0219-1

Photodiastereomerization of urocanic acid and its human serum albumin complex (its binding constant was estimated to amount 4.1·10 ^-4dm³·mol^-1) was investigated. It was found that although the photodiastereomerization rates were similar, the photoequilibrium positions differed significantly ((E):(Z) = 33:77 for free urocanic acid, and 50:50 for the complex). This is thought to be due to a different stabilization of the corresponding orthogonal excited states. The thermal barrier of diastereomerization was estimated to amount to more than 250 kJ·mol^-1 making it a very unlikely process under physiological and photodiastereomerization conditions. The various prototropic species of the two diastereomers at various pH values were analyzed by means of a mathematical model and from these results a novel photoinduced pH-jump methodology allowing for fast, persistent, diffusion controlled, and bidirectional jumps is proposed. Springer-Verlag 2004.

Full text of DOI:10.1007/s00706-004-0219-1

Monatshefte f €u r Chemie 135, 1297–1304 (2004)
DOI 10.1007/s00706-004-0219-1
Concerning the Photodiastereomerization
and Protic Equilibria of Urocanic Acid
and its Complex with Human Serum
Albumin
Bettina Schwarzinger and Heinz Falkꢀ
Institute of Organic Chemistry, Johannes Kepler University Linz, A-4040 Linz, Austria
Received April 30, 2004; accepted May 4, 2004
Published online July 23, 2004 # Springer-Verlag 2004
Summary. Photodiastereomerization of urocanic acid and its human serum albumin complex (its
ꢂ4
3
ꢂ1
binding constant was estimated to amount 4.1ꢁ10 dm ꢁmol ) was investigated. It was found that
although the photodiastereomerization rates were similar, the photoequilibrium positions differed
significantly ((E):(Z) ¼ 33:77 for free urocanic acid, and 50:50 for the complex). This is thought to
be due to a different stabilization of the corresponding orthogonal excited states. The thermal barrier of
diastereomerization was estimated to amount to more than 250 kJꢁmolꢂ1 making it a very unlikely
process under physiological and photodiastereomerization conditions. The various prototropic species
of the two diastereomers at various pH values were analyzed by means of a mathematical model and
from these results a novel photoinduced pH-jump methodology allowing for fast, persistent, diffusion
controlled, and bidirectional jumps is proposed.
Keywords. Diastereomerization; Photochemistry; Protic equilibria; Thermal stability; pH-Jump.
Introduction
(E)-Urocanic acid (3-(1H-imidazol-4-yl)propenoic acid, 1) has been first isolated by
Jaffe from the urine of a peculiar dog in 1874 [1]. It is known as a deamination
metabolite of histidine [2] that accumulates in the stratum corneum where it is a
major UV-chromophore representing about 0.7% of the dry weight of the human
skin [3]. Quite early it has been recognized that 1 could constitute a natural sun-
screen since its UV-absorption spectrum (ꢀmax ¼ 275 nm) overlaps with the major
erythmogenic region (290–310 nm) [4]. Upon irradiation with UV-light the natu-
rally occuring (E)-1 isomerizes partly to its diastereomer (Z)-1, which absorbs in this
region as well [5]. In the 1980’s De Fabo et al. demonstrated that (Z)-1 is a mediator
ꢀ Corresponding author. E-mail: heinz.falk@jku.at

1
298
B. Schwarzinger and H. Falk
of UV-induced immunosuppression [6]. This immunosuppressive effect has also
been linked to UV-induced skin cancer [7]. There are many investigations available
in literature concerning the photodiastereomerization reaction of 1 [8]. Thus, e.g. the
isomerization quantum yields at different wavelengths have been studied by
Morrison et al. [9, 10], and more recently by Hanson et al. [5] and Mohammad
[11]. In addition, the proportions of the diastereomers in the photostationary states
have been demonstrated by Laihia et al. [12] to be dependent on pH values.
Such studies were exclusively focused upon free 1. However, under physiolo-
gical conditions a plethora of compounds including proteins will be present in
addition to 1. Pondering the constitution of 1, one can imagine strong interactions
or even complexation of 1 with constituents of such a matrix, and accordingly, the
photodiastereomerization characteristics of free and associated 1 could be quite
different. Given the physiological importance of the photodiastereomerization
reaction we explored this topic in some detail and report the findings below.
Results and Discussions
Protic Equilibria of 1
Before setting out to investigate the photodiastereomerization of free and complexed
1 we decided to study the protonation–deprotonation equilibria of the diastereomers
of 1 in some detail. This seemed to be important in peculiar, as it has been shown that
Fig. 1. Protonation states of 1; the species possible in principle are the neutral (1), zwitterionic
ꢂþ þ ꢂ 2ꢂ
(
1
), protonated (1 ), monodeprotonated (1 ), and dideprotonated (1 ) forms

Photodiastereomerization and Protic Equilibria
1299
the photodiastereomerization reaction is dependent on pH [12] and wavelength [6].
Thus we constructed a quantitative picture of the various species present at different
pH values by calculating the equilibrium concentrations of the different species of
(
E)- and (Z)-1 as given in Fig. 1. Starting with the total concentration c of the
Ua
þ
þꢂ
ꢂ
2ꢂ
pertinent diastereomer of 1 as given by c ¼ [1 ] þ [1] þ [1 ] þ [1 ] þ [1 ]
Ua
þ
þ
þ
ꢂ
and deducing from Fig. 1 the relations K ¼ [H ][1]=[1 ], K ¼ [H ][1 ]=[1],
1
2
þ
2ꢂ
ꢂ
þꢂ
K ¼ [H ][1 ]=[1 ], and K ¼ [1]=[1 ], the matrix equation system (1) could
3
T
be derived. Using the pK values for (E)-1 (pK ¼ 4.85 [13], pK ¼ 4.67 [13],
1
2
pK ¼ 1.16 [13], pK ¼ 14.52 was approximated by assigning to it the pK value of
T
3
a
imidazole [14]) and (Z)-1 (pK ¼ 3.00 [13], pK ¼ 6.70 [13], pK ¼ 1.25 [13], and
1
2
T
because there should be no significant difference between the two diastereomers in
their imidazole NH deprotonation behaviour, pK ¼ 14.52) the different equilibrium
3
concentrations of the species of (E)- and (Z)-1 could then be calculated from (1) by
means of Mathematica .
+
0
1 0
1
0
1
þ
1
1
1
0
1
0
1
0
0
½1 ꢃ
cUa
0
þ
B
B
B
B
@
C B
C B
C
C
C
C
A
B
B
B
B
@
C
C
C
C
A
ꢂK ½H ꢃ
½1ꢃ
1
þ
ꢂþ
C B
ꢁ
0
0
0
ꢂK
0
½H ꢃ
½1
ꢃ
¼
0
ð1Þ
2
C B
þ
ꢂ
A @
0
0
ꢂK ½H ꢃ
½1 ꢃ
0
3
2
ꢂ
1
ꢂKT
0
0
½1 ꢃ
0
The results of these calculations are visualized in Fig. 2, and they demonstrate
that under physiological conditions (pH ꢄ 5–7) up to four species will coexist.
Their relative concentrations vary strongly with the pH values. In addition, these
concentrations will strongly differ for the two diastereomers for a given pH value.
ꢂþ
Thus, e.g. at pH 6 we find for (E)-1 equal amounts of zwitterion 1 and mono-
ꢂ
deprotonated species 1 (ꢄ48% both) together with a rather small amount (ꢄ2%)
þ
of the protonated species 1 , whereas for (Z)-1 the main species is the zwitterion
1
ꢂþ
ꢂ
þ
(ꢄ82%) and the mono-deprotonated form 1 and the protonated species 1
are present in about 13 and 5% only. Accordingly, a highly complex photodiaster-
eomerization behavior with respect to pH and irradiation wavelength could be
envisaged because a series of different species with different photochemical prop-
erties will be present in strongly varying relative concentrations depending on the
pH of the solution. Of course, additional complications could arise from a variety
of rotamers and tautomers possible for the individual prototropic species.
1
Although there have been several studies of H NMR spectra of the diastereomers of 1 (e.g. Ref.
[
15]) no consistently assigned set of signals was available for aqueous solutions. Thus, to acquire a
reliable in situ analytical tool to investigate quantitative aspects of photodiastereomerization such a set
Fig. 3) was derived by 2D methods. By tracing the various shifts of these signals in dependence of pH
(
a direct tool to assess the photodiastereomerization was established. Because it turned out that direct
observation of the photodiastereomerization of complexed systems was not possible due to the low
1
concentrations of 1 used, and moreover, to severe signal overlaps, we used the H NMR data to
calibrate the UV spectroscopic method.
Photodiastereomerization of 1
The first problem to solve was to determine whether there is indeed complexation
between the diastereomers of 1 and human serum albumin (HSA), which we chose

1
300
B. Schwarzinger and H. Falk
Fig. 2. Calculated pH-dependence of the concentrations of the protic species of (E)- and (Z)-
ꢂþ
þ
ꢂ
urocanic acid (1): the neutral 1, the zwitterion 1 , the protonated 1 , the monodeprotonated 1 ,
ꢂ
and the dideprotonated 12
as a simple protein model for the complex matrix of the skin. To estimate the
complex binding properties the protein fluorescence decrease (ꢀEX ¼ 280 nm,
ꢀEM ¼ 340 nm) induced by ligand binding was used [16]. Accordingly, the

Photodiastereomerization and Protic Equilibria
1301
1
Fig. 3. Assignment of the H NMR shifts of (Z)- and (E)-1 in aqueous (D O) solution, pH ¼ 5.0
2
fluorescence of HSA was measured in the presence and absence of ligand at various
ꢂ4
ꢂ3
concentrations of (E)-1 (0–4.8ꢁ10 molꢁdm ) in aqueous solutions (pH ¼ 7.0) at
ꢂ6
ꢂ3
a protein concentration of 3.9ꢁ10 molꢁdm . The binding constant K of (E)-1 to
b
HSA was then calculated from a Scatchard linearization [17]: [B]=[F] ¼ K ꢁ[R ]–
b
T
K ꢁ[B], where [B] and [F] are the bound and the free ligand concentrations and [R ]
b
T
is the total receptor (HSA) concentration. The plot displayed in Fig. 4 led to a
4
ꢂ3
binding constant K ¼ 4.1ꢁ10 molꢁdm . Although this value was derived for the
b
(E)-diastereomer we infer that both diastereomers of 1 will be bound similarly
tightly to HSA.
Irradiation of free aqueous (E)-1 and a solution of its complex with HSA
ꢂ5
ꢂ3
(
c ¼ 3.7ꢁ10 molꢁdm and a ratio (E)-1 to HSA of 50:1 in water, pH ¼ 7.0)
1
resulted in a decrease of the UV absorbances because of the lower extinction
Fig. 4. Scatchard plot for the complexation of (E)-1 with HSA

1
302
B. Schwarzinger and H. Falk
coefficient of the resulting (Z)-diastereomer. NMR detection for estimating the
ratio of the diastereomers was not possible due to the low concentration used in
these experiments. However, it was used to assess absolute concentrations of the
diastereomers for concentrated solutions and to derive the corresponding relative
absorptions upon dilution. Because the " value for (E)-1 was found to be non-linear
for the concentration region used, the logarithmic adjustment " ¼ 17750 þ
ꢂx=1:1188
was applied, where x is the molar concentration. The measured
1
022.3ꢁe
absorbance is the sum of the absorbance of the two diastereomers: E ¼ E þ
E
ꢂ5
E ¼ " ꢁc ꢁd þ " ꢁc ꢁd, where d is 1 cm and the sum of c and c is 3.7ꢁ10
Z
E
E
ꢂ3
Z
Z
E
Z
molꢁdm (the initial concentration of (E)-1). By means of these relationships
the concentrations of (E)- and (Z)-1 during and after irradiation when reaching
the photostationary equilibrium could be derived. Although the photodiastereomer-
ization rates did not differ significantly, the observed ratio of the two diastereomers
in the free state was found to differ significantly amounting to 33:67 for the aque-
ous solution, but to 50:50 for the HSA-complexed state. This difference does not
derive from the one observed for solutions of different pH values [12] because the
pH values of the free and complexed system were kept at 7.0. Obviously, the
differential stabilization of the orthogonal state(s) [12] leading to the observed
photodiastereomer ratio is strongly dependent on the complexation of the diaste-
reomeric states. This result documents clearly that the influence of the matrix in the
physiological system is not negligible as has been inferred recently [12], and has to
be taken into account in investigations of the photochemical behavior of 1. Inter-
estingly enough, the photoequilibrium ratio of (Z)-1:(E)-1 in sun-exposed skin has
been found to amount to 50:50 [7]. Thus the complexation of 1 by skin constituents
is modeled perfectly by its complexation by HSA.
Thermal Diastereomerization Reaction
Because of the importance of the photodiastereomerization reaction the question of
a possible thermal isomerization accompanying the photoreaction is highly inter-
esting and has not yet been addressed explicitly to our knowledge. To evaluate this
possible influence, we thermally equilibrated samples of various mixtures of (E)-
and (Z)-1 dissolved in D O. The concentrations of the diastereomers were followed
2
1
by H NMR spectroscopy of the temperated samples. The first mixture investigated
consisted of 67% (Z)-1 and 33% (E)-1 (the photostationary state concentrations),
the other one contained 15% (Z)-1 and 85% (E)-1 (which is far away from the
ꢅ
photostationary equilibrium). After thermal equilibration for 32 h at 80 C samples
ꢅ
were withdrawn and the temperature was set to 90 C for additional 14 h. No sig-
nificant concentration changes could be detected in all samples. From this observa-
tion we extrapolated a lower limit for the Arrhenius activation energy of
ꢂ1
>
250 kJꢁmol . Thus the kinetic stabilities of the diastereomers of 1 were found
to be quite extraordinarily high. Accordingly, it can be concluded that the (Z)-
diastereomer of urocanic acid ((Z)-1) found in the human skin [8] cannot originate
from a thermal equilibration, but is exclusively due to photodiastereomerization. In
addition, a disturbance of photodiastereomerization experiments by a thermal back-
reaction of the (E)-diasteromer can thus be ruled out. The latter reaction could have
been envisaged from the thermodynamic instability of the (E)-diastereomer,

Photodiastereomerization and Protic Equilibria
1303
ꢂ1
which we estimated by AM1 calculations to amount to about 9 kJꢁmol (DH ¼
f
ꢂ1
ꢂ1
ꢂ115 kJꢁmol for the (E)-1 and ꢂ124 kJꢁmol for the (Z)-1 diastereomer – cf.
Ref. [18]).
Epilogue: Light-Induced pH-Jump
As discussed in the first chapter (see Fig. 2), the diastereomers of 1 display
different and sensitive prototropic behaviors with respect to the pH values within
the physiological pH region. This triggered the idea that the photodiastereomer-
ization reaction could be used to initiate a photoinduced persistent pH jump
derived from the photoinduced shift of the proportion of the two diastereomers
ꢂ3
at a given pH value. Indeed, when aqueous solutions of (E)-1 (c ¼ 1ꢁ10
ꢂ3
molꢁdm ) with different starting pH values were irradiated with UV-light under
stirring and concomitant pH-measurement the pH values changed. Thus, e.g.
when starting at pH 3.48, the solution equilibrated upon irradiation at pH 3.36,
and when starting at pH 8.32, pH 8.46 was attained. Of course, by exchanging the
(E)-diastereomer by the (Z) configured one a pH jump in the opposite direction
will be obtained. Since photodiastereomerizations are very fast reactions (ps to
ns), this behavior opens the possibility to achieve laser induced ultrafast pH
jumps yielding acidification or basification of solutions of pertinent biomolecules,
like proteins or polynucleotides. Such jumps using, as proposed, one of the dia-
stereomers of urocanic acid (1) will be ultrafast (diffusion controlled), persistent,
and bidirectional depending on which diastereomer is initially used at a certain
pH value. Compared with this new methodology the conventional techniques are
rather limited. Thus, commonly pH jumps are induced by means of stopped-flow
experiments [19] in the slow reaction domain and by means of the photodissocia-
tion of o-nitrobenzaldehyde in the fast reaction area [20]. Although the latter
method yields persistent jumps it is unidirectional allowing for acidification of
a system only.
Experimental
(
E)-Urocanic acid ((E)-1) and HSA (M ¼ 66439) were of commercial origin (Sigma). Photodiastereo-
merizations of the stirred samples were executed by means of a Hanau TQ 150 Z1 UV-lamp in SiO
2
cells under Ar atmosphere in aqueous solution (dist. H O, D O). (Z)-Urocanic acid ((Z)-1) was isolated
2
2
from such photodiastereomerization batches by means of preparative HPLC (Dionex: P 608 pump,
ASI-100 automated sample injector, PDA detector, LiChrosorb RP 18 (7ꢁm) column, d ¼ 10 mm,
1
l ¼ 250 mm, distilled H O). H NMR spectra were recorded by means of a Bruker Avance 200 MHz,
2
2
D correlation experiments (COSY) were set up using the commercially available tools. The UV and
fluorescence spectra were obtained on the Varian Cary 100 Bio and Hitachi F-4010 spectrometer using
SiO cells, d ¼ 1 cm. Thermal isomerizations were conducted using a mgw Lauda Thermo-Boy and pH
2
measurements were done by means of a Orion Model 290 A with a Hamilton Slimtrode or Hamilton
+
Slintrode (directly inserted into the NMR tube). Calculations were executed using Mathematica .
Acknowledgment
We are grateful to M. Kitzm €u ller for assistance in the mathematical calculations.

1
304
B. Schwarzinger and H. Falk: Photodiastereomerization and Protic Equilibria
References
[1] Jaffe M (1874) Ber dtsch chem Ges 7: 1669
[2] Baden HP, Pathak MA, Butler D (1966) Nature 210: 732; Baden HP, Pathak MA (1967) J Invest
Derm 48: 11
[
[
[
[
[
[
3] Tabachnick J (1957) Arch Biochem Biophys 70: 295
4] Zenisek A, Kral J, Hais I (1955) Biochim Biophys Acta 18: 589
5] Hanson KM, Li B, Simon JD (1997) J Am Chem Soc 119: 2715
6] De Fabo EC, Noonan FP (1983) J Exp Med 158: 84
7] Norval M (2001) J Photochem Photobiol B63: 28
8] for reviews see: Mohammad T, Morrison H, HogenEsch H (1999) Photochem Photobiol 69: 115;
Simon JD (2000) Acc Chem Res 33: 307
[9] Morrison H, Bernasconi C, Pandey G (1984) Photochem Photobiol 40: 549
[
[
[
[
[
10] Morrsion H, Avnir D, Bernasconi C, Fagan G (1980) Photochem Photobiol 32: 711
11] Mohammad T (2002) Tetrahedron Lett 43: 8897
12] Laihia JK, Lemmetyinen H, Pasanen P, Jansen CT (1996) J Photochem Photobiol B 33: 211
13] Halle JC, Pichon C, Terrier F (1984) J Biol Chem 259: 4142
14] Barton D, Ollis D (eds) (1979) Comprehensive Organic Chemistry. In: Sammes PG (ed)
Heterocyclic Compounds, vol 4. Pergamon Press
[
[
[
[
[
[
15] Neuvonen H, Neuvonen K (1998) J Chem Soc Perkin Trans 2, 1665
16] Sulkowska A, Bojko B, Rownicka J, Pentak D, Sulkowski W (2003) J Mol Struct 651–653: 237
17] Hiratsuka T (1990) J Biol Chem 265: 18786
€
18] Lahti A, Hotokka M, Neuvonen K, Ayr €a s P (1995) J Mol Struct Theochem 331: 169
€
19] e.g.: Nagel JHA, Gultyaev AP, Oist €a m €o KJ, Gerdes K, Pleij CWA (2002) Nucl Acids Res 30: 63
20] Viappiani C, Bonetti G, Carcelli M, Ferrari F, Sternieri A (1998) Rev Sci Instrum 69: 270;
Abbruzzetti S, Crema E, Masino L, Vecli A, Viappiani C, Small JR, Libertini LJ, Small E (2000)
Biophys J 78: 405; Choi J, Hirota N, Terazima M (2001) J Phys Chem 105: 12