Effects of Glycosylation of (2
S
,4
R
)-4-Hydroxyproline on the Conformation,
Kinetics, and Thermodynamics of Prolyl Amide Isomerization
Neil W. Owens, Craig Braun, Joe D. O’Neil, Kirk Marat, and Frank Schweizer*
Department of Chemistry, UniVersity of Manitoba, Winnipeg, Manitoba, R3T 2N2 Canada
Received May 16, 2007; E-mail: schweize@cc.umanitoba.ca
Glycosylation is a common post-translational modification of
proteins implicated in cellular recognition processes and controlling
protein conformation.
1
Typically, carbohydrates are O-linked to
serine (Ser) and threonine (Thr) or N-linked to asparagine (Asn).
Glycosylation of (2S,4R)-4-hydroxyproline (Hyp) is widespread in
the plant kingdom and occurs in Hyp-rich glycoproteins (HRGPs)
that are associated with the cell walls of algae and flowering plants.
2
HRGPs are characterized by extensively glycosylated Hyp se-
quences that contain O-glycosidic linkages to the pyranose
D-
galactose or the furanose
L-arabinose.
2
Although HRGPs are broadly
implicated in many aspects of plant growth, development
3
and cell
wall stability,
2
no information is available about the structural and
conformational implications of Hyp-glycosylation on peptide
backbone conformation.
Hyp and proline (Pro) are unique among the proteinogenic amino
acids since they are characterized by limited rotation of the ø
dihedral angle (Figure 1) as their side chain is fused to the peptide
backbone. As a consequence, there is a reduction in the energy
difference between the prolyl amide cis and trans isomers, making
them nearly isoenergetic; this leads to higher cis N-terminal amide
content relative to the other amino acids. Moreover, the isomer-
ization of the prolyl amide bond has been shown to be the rate-
determining step in the folding pathways of many peptides and
proteins.
4
Herein we describe the effects of galactosylation of Hyp on the
conformation as well as the thermodynamics and kinetics of prolyl
N-terminal amide isomerization. Compounds 4a AcHyp(R-D-Gal)-
NHMe and 4b AcHyp(β-D-Gal)NHMe were selected as glycopep-
tide mimics, while AcProNHMe 1, AcHypNHMe 2, and AcHyp(O-
tert-butyl)NHMe 3 served as non-glycosylated reference compounds
(Figure 1). The trans rotamers in compounds 1-4b were assigned
on the basis of higher C
δ
atom NMR chemical shifts relative to
the cis rotamer
5
and nOe transfer between H-δ of proline and the
N-acyl protons in selective 1D GOESY experiments.
6
The relative
amounts of cis and trans isomers were determined by integrating
and averaging as many distinct proton signals as possible for both
the major and minor isomers in the
1
H NMR spectra.
7
In D
2
Oat
37 °C, the trans/cis isomer ratio equilibrium constant (K
t/c
) for 4a
(3.41 ( 0.30) and 4b (3.37 ( 0.28) are nearly identical to those of
2 (3.52 ( 0.05) and 3 (3.34 ( 0.15), and the observed differences
are within the experimental errors (Table 3).
The kinetics of cis/trans isomerization for compounds 1-4b were
determined by
1
H NMR spectroscopy inversion transfer experi-
ments
8
in D
2
O at elevated temperature. At 67 °C, the cis-to-trans
rate constant of isomerization (k
ct
)oftheR-glycosylated Hyp model
peptide 4a (k
ct
) 0.83 s
-1
) is very similar compared to the
hydroxyproline model peptide 2 (k
ct
of 0.73 s
-1
) and 3 (k
ct
) 0.77
s
-1
), while the β-anomer 4b gave slightly lower rates (k
ct
) 0.61
s
-1
). A similar trend was observed for the trans-to-cis rate constants
of compounds 1-4b (Table 1). At physiological temperature the
kinetic rates are too slow to be differentiated by this assay.
The effects of temperature on k
ct
and k
tc
were analyzed by Eyring
plots
9
(Supporting Information) and values for ∆H
‡
and ∆S
‡
were
calculated from linear least-squares fits of the data in these plots
and are presented in Table 2. The activation parameters demonstrate
that the free-energy barriers to isomerization of compounds 1-4b
Figure 1.
Cis-trans isomerization of reference diamides 1-3 and the
galactosylated Hyp model amides 4a and 4b.
Table 1.
Rates of Prolyl Amide Isomerization for 1-4b
amide
k
ct
a
(s
-
1
)
k
tc
b
(s
-
1
)
1 0.81 ( 0.02 0.31 ( 0.02
2 0.73 ( 0.01 0.25 ( 0.01
3 0.77 ( 0.02 0.27 ( 0.01
4a 0.83 ( 0.05 (0.82 ( 0.06)
c
0.27 ( 0.02 (0.27 ( 0.03)
c
4b 0.61 ( 0.04 (0.57 ( 0.05)
c
0.21 ( 0.02 (0.19 ( 0.02)
c
a
Carried out in D
2
Oat67°C.
b
Calculated from k
ct
and equilibrium
(K
t/c
).
c
Carried out in phosphate buffer pH ) 7.4 at 67 °C.
Table 2.
Activation Enthalpies (∆
H
‡
) and Entropies (∆
S
‡
)as
Derived from Eyring Plots in D
2
O for 1-4b
cis
to
trans
a
trans
to
cis
a
amide
∆
H
‡
kcal/mol
∆
S
‡
cal/mol
‚
K
∆
G
‡
300K
kcal/mol
∆
H
‡
kcal/mol
∆
S
‡
cal/mol
‚
K
∆
G
‡
300K
kcal/mol
1 20.6 1.2 20.2 21.3 1.4 20.9
2 20.2 0.1 20.2 21.1 0.5 21.0
3 21.8 4.7 20.4 22.7 5.4 21.1
4a 20.4 0.6 20.2 20.6 1.0 20.3
4b 22.4 6.2 20.6 23.3 6.6 21.3
a
Error limits obtained from the residuals of the linear least-squares fitting
of the data to equation ln(k/T) ) (-∆H
‡
/R)(1/T) + ∆S
‡
/R + ln(k
B
/h) were
1-2% for compounds 1 and 2, and 3-6% for compounds 3, 4a, and 4b.
Table 3.
Thermodynamic Parameters for Isomerization of 1-4b
Amide
∆
H
°
a
(kcal/mol)
∆
S
°
a
(cal/mol
‚
K) K
t
/
c
b
(37
°
C)
1 -0.95 ( 0.01 -0.87 ( 0.02 3.03 ( 0.08
2 -1.33 ( 0.03 -1.76 ( 0.09 3.52 ( 0.05
3 -1.29 ( 0.02 -1.72 ( 0.06 3.34 ( 0.15
4a -1.27 ( 0.02 -1.64 ( 0.07 3.41 ( 0.30
4b -1.30 ( 0.03 -1.77 ( 0.10 3.37 ( 0.28
a
Error limits obtained by linear least-squares fitting the data of the van’t
Hoff plots to equation ln K
t/c
) (-∆H°/R)(1/T) + ∆S°/R;
b
Carried out in
D
2
O; ( SE determined by integration of two or more sets of trans/cis
isomers.
Published on Web 09/01/2007
11670
9
J. AM. CHEM. SOC. 2007,
129
, 11670-11671 10.1021/ja073488d CCC: $37.00 © 2007 American Chemical Society
are enthalpic in origin. The effects of temperature on the values of
K
t/c
) (k
ct
/k
tc
) were measured directly by NMR spectroscopy, and
the resulting data were analyzed by van’t Hoff plots (Supporting
Information). Values for ∆H° and ∆S° were calculated from linear
least-squares fits of these plots (Table 3). In all cases studied, the
trans isomer of 1-4b is more stable than the cis isomer. Moreover,
the values of K
t/c
for 1-4b are dependent on temperature such that
the trans isomer becomes increasingly favored as the temperature
decreases.
The pucker of Hyp (2S,4R) in model peptide 2 in solution has
been previously assigned to the Cγ-exo conformation on the basis
of observed J-coupling constants.
10
The prolyl ring pucker for
compounds 3, 4a, and 4b were similarly established as the Cγ-exo
conformation on the basis of
1
H NMR coupling constants by
comparison to literature values. For example, in compound 4a we
observed both
3
J
Rβ1
and
3
J
Rβ2
couplings constants to be 8.2 Hz.
The expected coupling constants for the Cγ-exo conformer are
7-10 Hz and 7-11 Hz, respectively, whereas those for Cγ-endo
are 6-10 Hz and 2-3 Hz, respectively. Similarly, other couplings
show characteristic patterns for the Cγ-exo pucker (Supporting
Information).
Previous reports have shown that inductive effects in the
γ-position of proline have significant structural consequences on
the thermodynamics and kinetics of prolyl amide bond isomeriza-
tion.
11
To assess the inductive effect caused by glycosylation of
hydroxyproline we determined the
13
C NMR chemical shifts of the
C
γ
atom, which can indicate electron withdrawal by pendant
function groups,
12
and have previously been used to correlate the
electron-withdrawing effects in various C
γ
-substituted proline
analogues.
11c
The observed
13
C NMR chemical shifts (δ
Cγ(trans)
)
indicate that electron withdrawal increases in the order hydroxyl
(δ
Cγ
) 69.9) (2) < tert-butoxyl (δ
Cγ
) 70.1) (3) <R-galactosyl
(δ
Cγ
) 76.5) (4a) < β-galactosyl (δ
Cγ
) 77.9) (4b) (see Table 1 in
the Supporting Information).
To determine the extent of interaction between the galactose and
prolyl rings, we performed nOe transfer experiments with com-
pounds 4a and 4b in D
2
O. Selective inversion of one of the H-β
protons in 4a by a selective GOESY
6
experiment resulted in a 1.5%
resonance transfer to a peak at δ ) 3.83 ppm corresponding to the
overlapped signals of both H-4 and H-5 of galactose. By compari-
son, selective inversion of H-2 in 4b produced nOe transfer (3.0%)
to H-R of Hyp. These results suggest that galactosylation of Hyp
induces close contacts between distant positions in the carbohydrate
and pyrrolidine rings.
In summary, we have found that glycosylation of Hyp in
compounds 4a and 4b has no apparent effect on the isomer
equilibrium (K
t/c
) nor on the rate of isomerization (k
tc
, k
tc
) in water
between the cis and trans isomers when compared to unglycosylated
reference compounds 1-3. However, glycosylation of Hyp provides
an inductive electron withdrawing effect on the prolyl ring. The
magnitude of the change in
13
C NMR chemical shift (6.5-8 ppm)
has been correlated to the strength of the inductive effect and is
similar to a downfield shift of 8 ppm observed when a trifluoro-
acetate group was attached to Hyp in AcHypOMe.
11c
The effect is
slightly larger for AcHyp(β-D-Gal)NHMe 4b (by 1.5 ppm)
compared to the R-anomer 4a. It is known that (4R)-electronegative
substituents stabilize the Cγ-exo pucker of proline in peptide
mimics
11
and contribute to enhanced stability of the triple helix in
collagen.
13
Moreover, it has been established both from computa-
tional
14
and experimental studies
15
that stabilization of the Cγ-exo
ring pucker favors the prolyl trans amide isomer. Our results
therefore suggest that glycosylation of Hyp may lead to additional
stabilization of the Cγ-exo ring pucker of Hyp. However, this
stabilization does not translate in a measurable increase on K
t/c
for
peptide mimics 4a and 4b when compared to unglycosylated 2.
Most likely the stabilization of the trans isomer in compounds 4a
and 4b is too small to be differentiated and remains within the
experimental error. Any stabilization of the Cγ-exo ring pucker
may be more apparent in larger glycopeptides that contain two or
more glycosylated Hyp units, where the effect may be additive.
Perhaps more importantly, nOe experiments show that the glyco-
sylation of 2 results in distant contacts between the proline and
galactose rings, suggesting that galactosylation of Hyp induces
conformational constraint into glycopeptides. Very likely this is
the result of increased steric strain induced upon glycosylation.
Additionally, glycosylation of Hyp may also affect other properties
including solvation, solubility, and thermostability. In conclusion,
while there is no significant influence on prolyl N-terminal amide
isomerization, the presence of both an enhanced inductive effect
and Gal-Pro contacts between distant positions in 4a and 4b
suggests that glycosylation of Hyp will have important implications
on peptide backbone conformation in HRGPs and glycosylated Hyp-
containing peptides.
Acknowledgment. The authors thank the National Science and
Engineering Council of Canada (NSERC) and the University of
Manitoba for financial support.
Supporting Information Available: Synthetic procedures,
1
H
NMR,
13
C NMR, 1D-GOESY spectra, Eyring plots, van’t Hoff plots,
plots of intensity versus mixing time for the magnetization transfer
experiments of 1-4b. This material is available free of charge via the
Internet at http://pubs.acs.org.
References
(1) (a) Varki, A. Glycobiology 1993, 3,97-130. (b) Dwek, R. A. Chem.
ReV. 1996, 96, 683-720.
(2) Lamport, D. T. A. Recent AdV. Phytochem. 1977, 11,79-115.
(3) (a) Knox, R. B.; Clarke, A.; Harrison, S.; Smith, P.; Marchalonis, J. J.
Proc. Natl. Acad. Sci. U.S.A. 1976, 73, 2788-2792. (b) Wang, H.; Wu,
H. M.; Cheung, A. Y. Plant Cell 1993, 5, 1639-1650. (c) Keller, B.;
Lamb, C. J. Genes DeV. 1989, 3, 1639-1646. (d) Sadava, D.; Chrispeels,
M. J. Biochem. Biophys. Acta 1973, 227, 278-287. (e) Esquerre-Tugaye,
M. T.; Lamport, D. T. A. Plant Physiol. 1979, 64, 314-319.
(4) Reviewed in Fischer, G.; Schmid, F. X. Biochemistry 1990, 29, 2205-
2212.
(5) Beausoleil, E.; Sharma, R.; Michnick, S. W.; Lubell, W. D. J. Org. Chem.
1998, 63, 6572-6578.
(6) GOESY is a selective 1D NOESY experiment, see: Stonehouse, J.; Adell,
P.; Keeler, J.; Shaka, A. J. J. Am. Chem. Soc. 1994, 116, 6037-6038.
(7) Taylor, C. M.; Hardre, R.; Edwards, P. J. B.; Park, J. H. Org. Lett. 2003,
5, 4413-4416.
(8) Larive, C. K.; Rabenstein, D. L. J. Am. Chem. Soc. 1993, 115, 2833-
2836.
(9) Eyring, H. J. Chem. Phys. 1935, 3, 107-115.
(10) Cai, M.; Huang, Y.; Liu, J.; Krishnamoorthi, R. J. Biomol. NMR 1995, 6,
123-128.
(11) (a) Eberhardt, E. S.; Panasik, N., Jr.; Raines, R. T. J. Am. Chem. Soc.
1996, 118, 12261-12266. (b) Renner, C.; Alefelder, S.; Bae, J. H.; Budisa,
N.; Huber, R.; Moroder, L. Angew. Chem., Int. Ed. 2001, 40, 923-925.
(c) Jenkins, C. L.; McCloskey, A. I.; Guzei, I. A.; Eberhardt, E. S.; Raines,
R. T. Biopolymers 2005, 80,1-8.
(12) Friebolin, H. Basic One- and Two-Dimensional NMR Spectroscopy, 3rd
ed.; Wiley-VCH: New York, 1998.
(13) Sakakibara, S.; Inouye, K.; Shudo, K.; Kishida, Y.; Kobayashi, Y.;
Prockop, D. J. J. Biochem. Biophys. Acta 1973, 303, 198-202.
(14) Improta, R.; Benzi, C.; Barone, V. J. Am. Chem. Soc. 2001, 123, 12568-
12577.
(15) DeRider, M. L.; Wilkens, S. J.; Waddell, M. J.; Bretscher, L. E.; Weinhold,
F.; Raines, R. T.; Markley, J. L. J. Am. Chem. Soc. 2002, 124, 2497-2505.
JA073488D
COMMUNICATIONS
J. AM. CHEM. SOC.
9
VOL. 129, NO. 38, 2007 11671