Rapid nitrogen inversion pathway in the cis/trans isomerization of selenoxo peptide bonds

Article abstract of DOI:10.1002/chem.201203721

Nitrogen inversion is discussed as a potentially stabilizing step in the acylation pathway of serine proteases, and the highly acidic carbonyl-amine form of the Asn side chain amide is discussed to play an essential role for amide group activation in the oligosaccharyltransferasecatalyzed Asn glycosylation. Upon irradiation with 288 nm UV light, the cis isomer content increases, as characterized by a red-shift of the UV spectrum, and then recovers to the ground state cis/trans ratio after dark equilibration. The deprotonation of the C atom attached to the C-Se bond could also be excluded, since this would lead to an epimerization. he apparent rate constant for the cis to trans decay of the peptide at a given pH value contains a rotational and an inversion term for the selenoxo and the selenoimidate anion. The selenoxo peptide bond is of special interest, since the formation of the selenoimidate anion results in an enhancement of peptide bond cis/trans isomerization dynamics via a reaction pathway previously undetected in aqueous solution.

Full text of DOI:10.1002/chem.201203721

COMMUNICATION
DOI: 10.1002/chem.201203721
Rapid Nitrogen Inversion Pathway in the cis/trans Isomerization of Selenoxo
Peptide Bonds
Yun Huang, Gꢀnther Jahreis, Christian Lꢀcke, and Gunter Fischer*[a]
The peptide bonds of a protein backbone confer confor-
mational constraints, of which the slow cis/trans isomeriza-
tion defines an enzymatically targeted conformational inter-
conversion of high biological significance.[1] The majority of
experimental data for the spontaneous isomerization sup-
port the assumption that the barrier of cis/trans interconver-
sion is rotational and not one controlled by nitrogen inver-
sion in aqueous solution.[2] Moreover, the transition state of
the enzymatically catalyzed isomerization is associated with
motions resulting from peptide bond rotation.[3] Despite the
findings described above, non-enzymatic pathways to en-
hance the conformational flexibility of the protein backbone
by manipulating the electronic nature of the peptide bond
are under debate.
proach to the investigation of these properties, because they
couple higher NH acidity with photocontrol of the cis/trans
equilibrium. The thioxo substitution, due to its minimal per-
turbation of the peptide backbone,[11] is considered an ideal
backbone modulation to serve as a hydrogen bonding
probe,[12] reversible photoswitch[13] and potent fluorescence
quencher.[14] The selenium-substituted analogues, selenoxo
peptides, have also been recently synthesized and they fea-
ꢀ
ture a higher C N rotational barrier, enhanced NH acidity
and lower photoswitching energy than the thioxo pepti-
des.[15] A further study of the oxo, thioxo and selenoxo pep-
tide bonds has revealed that chalcogen atomic polarizability
is the predominant factor determining several of their physi-
cochemical properties.[16]
ꢀ
ꢀ
ꢀ
Notably, the extreme pKa of the peptide bond [CO NH ]
might not completely prevent the transient formation of its
imidate or iminol forms by electrostatic interactions in
buried regions of a protein. This kind of behavior is related
to the specific base-catalyzed dissociation of amide protons,
which is the reason for the chemical exchange under basic
conditions and is useful for probing protein folding and sta-
bility.[4] Peptide bond tautomerism is expected to participate
in the unidirectional proton-transfer pathway in cytochro-
me c oxidase.[5] Nitrogen inversion is discussed as a poten-
tially stabilizing step in the acylation pathway of serine pro-
teases,[6] and the highly acidic carbonyl-amine form of the
Asn side chain amide is discussed to play an essential role
for amide group activation in the oligosaccharyltransferase-
catalyzed Asn glycosylation.[7] By applying density function
(DFT)-based simulations, the cis/trans isomerization is
found to exhibit distinct pathways in the oxoamide and the
iminol forms of polyglycine.[8] Although the characterization
of peptide-bond-derived imidates is crucial to the under-
standing of proteinaceous amide reactivity and conforma-
tional interconversions, their short lifetime[9] and low abun-
dance at moderate basic pH values[10] prevent a detailed
characterization in aqueous solution. Chalcogen-substituted
isosteres of the peptide bond may provide an alternative ap-
Spectrophotometric titration of Bz-VGAy
G
NH2 (Bz=benzoyl, y indicates a pseudo-peptide bond)
yielded a pKa of 9.1 for the selenoxo amide proton dissocia-
tion (Figure 1). The ionic form featured a loss of UV absorb-
ꢀ
Figure 1. The pH-dependent UV spectra of Bz-VGAy
G
in aqueous solution at 20.08C. The absorbance at 296 nm was fitted to
the Henderson–Hasselbalch equation, resulting in a pKa value of 9.1 for
the selenoxo amide group.
[a] Y. Huang, Dr. G. Jahreis, Dr. C. Lꢀcke, Prof. G. Fischer
Max Planck Research Unit for Enzymology of Protein Folding
Weinbergweg 22, 06120 Halle/Saale (Germany)
Fax : (+49)345-5511972
ance at about 296 nm but an increase of absorbance at
around 260 nm. The loss of the 296 nm band, which is attrib-
uted to p–p* transition of C=Se, clearly indicated that the
double bond between carbon and selenium does not exist in
the dissociated form. The new 260 nm band may result from
E-mail: fischer@enzyme-halle.mpg.de
fischer@enzyme-halle.mpg.de
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201203721.
ꢀ
ꢀ
the n–s* transition of [C
U
Chem. Eur. J. 2013, 19, 1179 – 1183
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1179

longer wavelengths compared to the ionized selenol group
in selenocysteine.[17] The negative charge should be localized
at the Se due to its high atomic polarizability,[16] thus leading
to the formation of a selenoimidate anion. Upon irradiation
with 288 nm UV light, the cis isomer content increased, as
characterized by a red-shift of the UV spectrum, and then
recovered to the ground state cis/trans ratio after dark equi-
libration (Figure 2). Notably, only the selenoxo moiety but
Figure 2. UV spectra characterizing the reversible photoisomerization of
Figure 3. Time course of the cis decay after UV irradiation (288 nm,
ꢀ
Bz-VGAy
pH 9.0, at 20.08C.
G
ꢀ
4 min) observed for Bz-VGAy
G
(1 mm phosphate buffer), and b) pH 9.9 (1 mm Gly-NaOH buffer) at
20.08C. The dark reaction was monitored at 320 nm. From the time
course at pH 7.1, a first-order rate constant of 2.58ꢂ10ꢀ3 sꢀ1 was calculat-
ed. The time course at pH 9.9 was fitted by Equation (1), resulting in the
not any other ionized species contributed to the time-de-
pendent UV absorbance at 320 nm. Typically, kinetic traces
were recorded in the dark after the photostationary state
was reached, which is characterized by a cis content of
following parameters: A00 =0.001 AU, b=42.3, DA=0.0011 AU, kapp
=
0.053 sꢀ1. The solid lines represent least-square fits. The plots of the resid-
uals are given above each graph.
ꢀ
about 20% for the [CSe NH] moiety. However, a lower
value was observed at high pH under the same irradiation
conditions.
the rate versus pH profile, the highest interconversion rate
could be observed in a pH region corresponding to the com-
plete formation of the selenoimidate anion (Figure 4).
Given the structural features of this species, which must
make rotational movements more energy demanding, nitro-
gen inversion represents an alternative pathway for the
rapid cis isomer decay. Notably, the E/Z interconversion of
the N-arylformimidate anion has been discussed to take
In contrast to the relatively inert response of the cis/trans
isomerization of oxoamides on pH variation,[18] the cis to
trans isomerization rate of selenoxo peptides was dramati-
cally accelerated at high pH (Figure 3). Intriguingly, time
courses of the cis isomer decay varied depending on pH, fol-
lowing strict first-order kinetics under neutral conditions
(Figure 3a), whereas sigmoidal kinetics was observed at al-
kaline pH (Figure 3b). Comparison of the shapes of the
ꢀ
time courses obtained for Bz-VGAyACTHNUTRGNE[NUG CSe NH]A-OMe and
ꢀ
Bz-VGAy
G
Supporting Information) did not reveal any dependence on
peptide structure.
Next, it was established that the cis to trans isomerization
ꢀ
ꢀ
of the selenoxo imidic bond [CSe N<] in Bz-VGAyACTHUNTGRNEUNG[CSe
N]P-OMe was not accelerated at high pH values (Figure S3
in the Supporting Information). Therefore, we concluded
that OHꢀ ion attack on the selenocarbonyl carbon in the se-
lenoxoamide state of the peptides was unsuitable for ex-
plaining the high isomerization rates found in Figure 3b.
The deprotonation of the Ca atom attached to the C=Se
bond could also be excluded, since this would lead to an epi-
merization that was not observed, as evidenced by the con-
servation of the HPLC profile and by NMR spectra (Fig-
ures S4 and S5 in the Supporting Information). Considering
Figure 4. The pH-dependent cis to trans isomerization of Bz-VGAy-
ꢀ
N
phosphate buffer (1 mm), and data for pHꢁ8 were measured in Gly-
NaOH buffer (1 mm). The pH–kapp plot was fitted by Equation (2) with
the experimental restrictions kcis-to-trans =0.0022 sꢀ1 and pKtrans =9.1, result-
ing in the parameters koEb-sto-Z =0.077 sꢀ1 and pKcis =9.3.
1180
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 1179 – 1183

Isomerization of Selenoxo Peptide Bonds
COMMUNICATION
dark is negligible,[21] and the E content should also be very
low at equilibrium due to a strong lone pair repulsion in the
E configuration. Hence, we assumed in the analyses that kcis-
to-trans @ktrans-to-cis and kE-to-Z @kZ-to-E, thus allowing us to ignore
the reverse reactions in Scheme 1.
Solving the rate equations for the cis isomer decay (see
the Supporting Information) based on the mechanism de-
picted in Scheme 1 results in Equations (1) and (2):
kapp
ꢀ
ꢁ
A320
¼
¼
þ A00
appt
ð1Þ
ð2Þ
kapp
DA
b þ
ꢀ b ek
Scheme 1. The cis/trans isomerization, nitrogen inversion and protona-
tion/deprotonation of selenoxo peptides after leaving the photostationary
state for the dark reaction. Putative transition states are shown in brack-
ets.[22]
ðpHꢀpKcis
Þ
10
1þ10
obs
k
EꢀtoꢀZ þ kcisꢀtoꢀtrans
ðpKtrans ꢀpHÞ
kapp
ðpHꢀpKcis
Þ
1 þ 10
where DA is the amplitude of the absorbance increase after
irradiation, A00 is the absorbance at infinite time, b is a con-
stant at a given pH (see the Supporting Information), kapp is
the apparent rate constant for cis to trans relaxation, kcis-to-
is the rate constant of cis to trans isomerization via the
place via a nitrogen inversion pathway in an organic sol-
vent.[19] Since the proton transfer steps are fairly fast and
not rate-limiting for the slow conformational interconver-
sions in aqueous solution, the trans and cis conformers could
rapidly equilibrate with the Z and E selenoimidate anions,
respectively, in the photostationary state (Scheme 1). In ad-
dition, the sigmoidal decay curves at high pH values (Fig-
ure 3b, and Figures S1 and S2) and the linear concentration
dependence of isomerization rates (Figure 5) indicate the
traꢀns
C N rotational pathway, koEb-sto-Z is the observed rate constant
of E to Z conversion in the selenoimidate anion, and t is the
reaction time.
The sigmoidal character of the time course depicted in
Figure 3b is perfectly described by Equation (1). Based on
Equation (2), the apparent rate constant for the cis to trans
decay of the peptide at a given pH value contains a rotation-
al and an inversion term for the selenoxo and the selenoimi-
date anion, respectively. Consequently, the pH–kapp plot
(Figure 4) could be fitted by Equation (2), resulting in an
observed rate constant of 0.077 sꢀ1 for E to Z conversion of
the selenoimidate anion, which represents a 35-fold rate in-
ꢀ
crease as compared to the C N rotational pathway that ap-
plies for the neutral selenoxo peptide. The solvent deuteri-
obs
E-to-Z
um kinetic isotope effect (SKIE), defined as k
(H2O)/
obs
k
(D2O)=1.13ꢂ0.04, was measured with Bz-VGAy-
E-to-Z
ꢀ
[CSe NH]FA-NH2 (4.5 mm) by using the same ratio between
the ionized and unionized species (pH 9.2/pD 9.85, Figure S6
in the Supporting Information). This SKIE value indicates
that solvent reorganization occurs in the transition state of
nitrogen inversion to a much higher extent than in the case
of the rotational pathway, for which the value of kcis-to-trans
(H2O)/kcis-to-trans (D2O)=1.004ꢂ0.008 is consistent with the
SKIE obtained for cis/trans isomerization via the rotational
pathway in oxoamide bonds.[18a,23]
obs
E-to-Z
Figure 5. Concentration dependence of k
obtained for Bz-VGAy-
ꢀ
[CSe NH]A-OMe in Gly-NaOH buffer (1 mm, pH 9.7) at 20.08C. The
obs
k
values were derived from Equation (2), by applying kcis-to-trans =
E-to-Z
0.003 sꢀ1, pKtrans =9.8 and pKcis =9.7 (Figure S1b). The linear fit resulted
in an intercept corresponding to kuEn-ctaot-Z =0.10 sꢀ1
.
possibility of an autocatalytic term in the cis isomer decay,
which most likely is due to the transient presence of the
Z selenoimidate anion for nucleophilic attack on the imine
carbon atom of the E conformer, thus adding to the rate of
apparent cis isomer decay (Scheme 1). The resulting tetrahe-
dral adduct has a disrupted amide resonance and will thus
exhibit a very low energy barrier to isomerization. This type
of catalysis is consistent with data indicating a very high nu-
cleophilicity of anionic selenium compounds.[20]
From the concentration dependence (Figure 5), the E to
uncat
E-to-Z
Z inversion rate without the autocatalytical term (k
)
could be extrapolated. Therefore, measuring the tempera-
uncat
ture dependence of k
allowed the determination of the
E-to-Z
activation parameters DGꢀ, DHꢀ and DSꢀ of the nitrogen
ꢀ
inversion pathway in comparison to those found for C N ro-
tation. Based on the linear Eyring plot (Figure S7 in the
Supporting Information), the activation energy DGꢀ for E
ꢀ
In the experiment described in Figure 4, the intermolecu-
lar term of cis isomer decay was held small. Furthermore,
for secondary amide peptide bonds the cis content in the
to Z inversion in Bz-VGAy
G
termined to be 18.4 kcalmolꢀ1 (Table S1 in the Supporting
Information), which is 2.3 kcalmolꢀ1 lower than the C N ro-
Chem. Eur. J. 2013, 19, 1179 – 1183
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1181

G. Fischer et al.
tation barrier of 20.7 kcalmolꢀ1 measured at pH 6.5. The
large negative entropy of the E to Z inversion indicates
marked solvent participation and reorganization.
For the sake of comparison, the pH-dependence of cis to
ꢀ
obvious from the intramolecular catalysis of peptidyl prolyl
isomerization, peptide bond properties might change consid-
erably owing to amino acid sequence specific interac-
tions.[23,26]
trans decay in the thioxo peptide Bz-VGAyACHTNUTRGNE[UNG CS NH]FA-
NH2 was also evaluated. The pKa of the thioxo group was
determined spectrophotometrically at 266 nm to be equal to
11.0 (Figure S8 in the Supporting Information). Surprisingly,
the presence of thioimidate anion showed that the cis
isomer decay is even slower than that of the unionized thio-
xoamide and the time courses did not reveal autocatalysis
(Figure 6 and Figure S8b). It is therefore reasonable to
Acknowledgements
The authors are grateful for financial support from the BMBF (RP
ProNet3). Monika Seidel is acknowledged for excellent technical assis-
tance and Dr. C. Jia for helpful discussions.
Keywords: chalcogens · isomerization · nitrogen inversion ·
selenoimidate anion · selenoxo peptides
[1] a) G. Fischer, Angew. Chem. 1994, 106, 1479–1501; Angew. Chem.
Int. Ed. Engl. 1994, 33, 1415–1436; b) W. J. Wedemeyer, E. Welker,
H. A. Scheraga, Biochemistry 2002, 41, 14637–14644; c) C. Schiene-
Fischer, T. Aumꢀller, G. Fischer, Top. Curr. Chem. 2011, 1–33;
d) K. P. Lu, G. Finn, T. H. Lee, L. K. Nicholson, Nat. Chem. Biol.
2007, 3, 619–629; e) Y.-C. Liou, X. Z. Zhou, K. P. Lu, Trends Bio-
chem. Sci. 2011, 36, 501–514; f) C. Schiene-Fischer, J. Habazettl,
F. X. Schmid, G. Fischer, Nat. Struct. Mol. Biol. 2002, 9, 419–424.
[2] a) K. B. Wiberg, Acc. Chem. Res. 1999, 32, 922–929; b) D. M.
Schnur, Y. H. Yuh, D. R. Dalton, J. Org. Chem. 1989, 54, 3779–
3785.
[3] U. Doshi, L. C. McGowan, S. T. Ladani, D. Hamelberg, Proc. Natl.
Acad. Sci. USA 2012, 109, 5699–5704.
[4] a) Y. Bai, T. Sosnick, L. Mayne, S. Englander, Science 1995, 269,
192–197; b) T. M. Raschke, S. Marqusee, Curr. Opin. Biotechnol.
1998, 9, 80–86; c) S. W. Englander, Annu. Rev. Biophys. Biomol.
Struct. 2000, 29, 213–238.
Figure 6. The pH-dependence of the cis to trans decay measured for Bz-
ꢀ
VGAy
N
radiation with 260 nm UV light for 4 min. The cis decay was subsequently
monitored at 290 nm. Data for pH<8 and pH>10.8 were measured in
phosphate buffer (1 mm); data for 8ꢃpH<10.8 were measured in Gly-
NaOH buffer (1 mm).
[5] K. Shimokata, Y. Katayama, H. Murayama, M. Suematsu, T. Tsuki-
hara, K. Muramoto, H. Aoyama, S. Yoshikawa, H. Shimada, Proc.
Natl. Acad. Sci. USA 2007, 104, 4200–4205.
[6] a) B. Liu, C. J. Schofield, R. C. Wilmouth, J. Biol. Chem. 2006, 281,
24024–24035; b) P.-O. Syrꢃn, K. Hult, ChemCatChem 2011, 3, 853–
860.
[7] C. Lizak, S. Gerber, S. Numao, M. Aebi, K. P. Locher, Nature 2011,
474, 350–355.
[8] K. Kamiya, M. Boero, K. Shiraishi, A. Oshiyama, J. Phys. Chem. B
2006, 110, 4443–4450.
assume a nitrogen inversion rate that is similar or slightly
slower than the rate of the rotational pathway. The E to Z
inversion barrier of the N-benzylformimidate anion in
[D6]DMSO has been determined as 23.3 kcalmolꢀ1,[24] and
in case of the N-benzylthioformimidate anion in MeOD as
19.6 kcalmolꢀ1.[25] Both values are above the inversion barri-
er of the selenoimidate anion in aqueous solution of
18.6 kcalmolꢀ1.
[9] J. S. Anderson, G. Hernandez, D. M. LeMaster, Curr. Org. Chem.
2010, 14, 162–180.
[10] G. P. Connelly, Y. W. Bai, M. F. Jeng, S. W. Englander, Prot. Struct.
Funct. Genet. 1993, 17, 87–92.
[11] a) J. H. Miwa, A. K. Patel, N. Vivatrat, S. M. Popek, A. M. Meyer,
Org. Lett. 2001, 3, 3373–3375; b) D. R. Artis, M. A. Lipton, J. Am.
Chem. Soc. 1998, 120, 12200–12206.
[12] A. Bachmann, D. Wildemann, F. Praetorius, G. Fischer, T. Kiefhab-
er, Proc. Natl. Acad. Sci. USA 2011, 108, 3952–3957.
[13] a) R. Frank, M. Jakob, F. Thunecke, G. Fischer, M. Schutkowski,
Angew. Chem. 2000, 112, 1163–1165; Angew. Chem. Int. Ed. 2000,
39, 1120–1122; b) J. Zhao, D. Wildemann, M. Jakob, C. Vargas, C.
Schiene-Fischer, Chem. Commun. 2003, 2810–2811; c) D. Wilde-
mann, C. Schiene-Fischer, T. Aumꢀller, A. Bachmann, T. Kiefhaber,
C. Lꢀcke, G. Fischer, J. Am. Chem. Soc. 2007, 129, 4910–4918; d) Y.
Huang, Z. Cong, L. Yang, S. Dong, J. Pept. Sci. 2008, 14, 1062–1068;
e) V. Cervetto, R. Pfister, J. Helbing, J. Phys. Chem. B 2008, 112,
3540–3544.
Even though the polarity of the solvents used is quite dif-
ferent, a linear decreasing trend is clearly observed in the
DGꢀE-to-Z versus chalcogen polarizability profile (Figure S9 in
the Supporting Information), which is in agreement with
previous findings that the chalcogen atomic polarizability
ꢀ
ꢀ
dominates many electronic properties of [CX NR ] (X=O,
S, Se; R =H, alkyl).[16] Certainly, it is clear that the introduc-
tion of photoresponsive chalcogens in a polypeptide back-
bone opens up a convenient way to manipulate and to char-
acterize backbone dynamics of a polypeptide chain in a site-
specific manner. The selenoxo peptide bond, due to its low
pKa and rapid E/Z inversion, is of special interest, since the
formation of the selenoimidate anion results in an enhance-
ment of peptide bond cis/trans isomerization dynamics via a
reaction pathway previously undetected in aqueous solution.
At this point, our data do not support a similar isomeriza-
tion rate enhancement for the oxoimidate species potentially
formed in the interior of proteins. However, as it became
[14] J. M. Goldberg, S. Batjargal, E. J. Petersson, J. Am. Chem. Soc.
2010, 132, 14718–14720.
[15] Y. Huang, G. Jahreis, C. Lꢀcke, D. Wildemann, G. Fischer, J. Am.
Chem. Soc. 2010, 132, 7578–7579.
1182
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 1179 – 1183

Isomerization of Selenoxo Peptide Bonds
COMMUNICATION
[16] Y. Huang, G. Jahreis, G. Fischer, C. Lꢀcke, Chem. Eur. J. 2012, 18,
9841–9848.
sociation the configuration of the iminolate anions is defined by the
Z/E notation.
[17] R. E. Huber, R. S. Criddle, Arch. Biochem. Biophys. 1967, 122, 164–
173.
[23] U. Reimer, N. E. Mokdad, M. Schutkowski, G. Fischer, Biochemistry
1997, 36, 13802–13808.
[18] a) R. K. Harrison, R. L. Stein, Biochemistry 1990, 29, 1684–1689;
b) R. K. Harrison, R. L. Stein, J. Am. Chem. Soc. 1992, 114, 3464–
3471.
[19] C. L. Perrin, J. D. Thoburn, J. Org. Chem. 1989, 54, 764–767.
[20] E. S. J. Arnꢃr, Exp. Cell Res. 2010, 316, 1296–1303.
[21] G. Scherer, M. L. Kramer, M. Schutkowski, U. Reimer, G. Fischer, J.
Am. Chem. Soc. 1998, 120, 5568–5574.
[24] C. L. Perrin, C. P. Lollo, C. S. Hahn, J. Org. Chem. 1985, 50, 1405–
1409.
[25] W. Walter, E. Schaumann, Chem. Ber. 1971, 104, 4–10.
[26] S. Wawra, G. Fischer, in cis-trans Isomerization in Biochemistry
(Ed.: C. Dugave), Wiley-VCH, Weinheim, 2006, pp. 167–193.
[22] Herein, the conventional cis/trans notation is used to define the con-
figuration of the peptide bonds, whereas after the amide proton dis-
Received: October 17, 2012
Published online: December 11, 2012
Chem. Eur. J. 2013, 19, 1179 – 1183
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1183