Amide Cis-trans isomerization in aqueous solutions of methyl N -Formyl- d -glucosaminides and Methyl N -Acetyl- d -glucosaminides: Chemical equilibria and exchange kinetics

Article abstract of DOI:10.1021/ja9086787

Amide cis-trans isomerization (CTI) in methyl 2-deoxy-2-acylamido-d- glucopyranosides was investigated by ¹H and ¹³C NMR spectroscopy. Singly ¹³C-labeled methyl 2-deoxy-2-formamido-d- glucopyranoside (MeGlcNFm) anomers provided standard ¹H and ¹³C chemical shifts and ¹H-¹H and ¹³C-¹³C spin-coupling constants for cis and trans amides that are detected readily in aqueous solution. Equipped with this information, doubly ¹³C-labeled methyl 2-deoxy-2-acetamido-d-glucopyranoside (MeGlcNAc) anomers were investigated, leading to the detection and quantification of cis and trans amides in this biologically important aminosugar. In comparison to MeGlcNFm anomers, the percentage of cis amide in aqueous solutions of MeGlcNAc anomers is small (～23% for MeGlcNFm versus ～1.8% for MeGlcNAc at 42 °C) but nevertheless observable with assistance from ¹³C-labeling. Temperature studies gave thermodynamic parameters δG°, δH°, and δS° for cis-trans interconversion in MeGlcNFm and MeGlcNAc anomers. Cis/trans equilibria depended on anomeric configuration, with solutions of α-anomers containing less cis amide than those of β-anomers. Confirmation of the presence of cis amide in MeGlcNAc solutions derived from quantitative ¹³C saturation transfer measurements of CTI rate constants as a function of solution temperature, yielding activation parameters E_act, δG° ^?, δH°^-, and δS° ^? for saccharide CTI. Rate constants for the conversion of trans to cis amide in MeGlcNFm and MeGlcNAc anomers ranged from 0.02 to 3.59 s^-1 over 31-85 °C, compared to 0.24-80 s^-1 for the conversion of cis to trans amide over the same temperature range. Energies of activation ranged from 16-19 and 19-20 kcal/mol for the cis → trans and trans → cis processes, respectively. Complementary DFT calculations on MeGlcNFm and MeGlcNAc model structures were conducted to evaluate the effects of an acyl side chain and anomeric structure, as well as C2-N2 bond rotation, on CTI energetics. These studies show that aqueous solutions of GlcNAc-containing structures contain measurable amounts of both cis and trans amides, which may influence their biological properties.

Full text of DOI:10.1021/ja9086787

Published on Web 03/12/2010
Amide Cis-Trans Isomerization in Aqueous Solutions of
Methyl N-Formyl-D-glucosaminides and Methyl
N-Acetyl-D-glucosaminides: Chemical Equilibria and
Exchange Kinetics
Xiaosong Hu,† Wenhui Zhang,† Ian Carmichael,‡ and Anthony S. Serianni*,†
Department of Chemistry and Biochemistry and Radiation Laboratory, UniVersity of Notre
Dame, Notre Dame, Indiana 46556
Received October 12, 2009; E-mail: aseriann@nd.edu
Abstract: Amide cis-trans isomerization (CTI) in methyl 2-deoxy-2-acylamido-D-glucopyranosides was
investigated by 1H and 13C NMR spectroscopy. Singly 13C-labeled methyl 2-deoxy-2-formamido-D-
glucopyranoside (MeGlcNFm) anomers provided standard 1H and 13C chemical shifts and 1H-1H and
13C-13C spin-coupling constants for cis and trans amides that are detected readily in aqueous solution.
Equipped with this information, doubly 13C-labeled methyl 2-deoxy-2-acetamido-D-glucopyranoside (MeGlcNAc)
anomers were investigated, leading to the detection and quantification of cis and trans amides in this
biologically important aminosugar. In comparison to MeGlcNFm anomers, the percentage of cis amide in
aqueous solutions of MeGlcNAc anomers is small (∼23% for MeGlcNFm versus ∼1.8% for MeGlcNAc at
42 °C) but nevertheless observable with assistance from 13C-labeling. Temperature studies gave
thermodynamic parameters ∆G°, ∆H°, and ∆S° for cis-trans interconversion in MeGlcNFm and MeGlcNAc
anomers. Cis/trans equilibria depended on anomeric configuration, with solutions of R-anomers containing
less cis amide than those of ꢀ-anomers. Confirmation of the presence of cis amide in MeGlcNAc solutions
derived from quantitative 13C saturation transfer measurements of CTI rate constants as a function of solution
q
q
q
temperature, yielding activation parameters Eact, ∆G° , ∆H° , and ∆S° for saccharide CTI. Rate constants
for the conversion of trans to cis amide in MeGlcNFm and MeGlcNAc anomers ranged from 0.02 to 3.59
s-1 over 31-85 °C, compared to 0.24-80 s-1 for the conversion of cis to trans amide over the same
temperature range. Energies of activation ranged from 16-19 and 19-20 kcal/mol for the cis f trans and
trans f cis processes, respectively. Complementary DFT calculations on MeGlcNFm and MeGlcNAc model
structures were conducted to evaluate the effects of an acyl side chain and anomeric structure, as well as
C2-N2 bond rotation, on CTI energetics. These studies show that aqueous solutions of GlcNAc-containing
structures contain measurable amounts of both cis and trans amides, which may influence their biological
properties.
Introduction
comparatively small energy difference between the cis and trans
conformations caused by the absence of bulky R-groups.4,5
Amide cis-trans isomerization (CTI) is an important ex-
change process in biological systems.1 In proteins, especially
those that are proline-rich (e.g., collagen), the conformation of
Xaa-Pro peptide bonds influences the pathway of protein
folding and determines the final protein tertiary structure.2 Prolyl
cis-trans isomerases are well-known folding accessory proteins
responsible for enzyme-catalyzed CTI of NXaa-ProC peptide
bonds in ViVo.3 Gly-Gly peptide bonds may also undergo
spontaneous amide bond isomerization in ViVo due to the
Recently, secondary amide peptide bond cis-trans isomerases
(APIases) have been reported to catalyze amide CTI in nonprolyl
peptide bonds; an example of this class of enzyme is the Hsp60
chaperone, DnaK.6 Amide CTI may also play a role in the
chemo-mechanical cycle of motor proteins.7
Amide CTI in glycobiology is less prominent than in
proteobiology, largely because glycosidic linkages are involved
in the assembly of oligo/polysaccharides rather than peptide
(amide) bonds. However, some biologically important saccha-
rides contain amide bonds as part of a side chain substituent,
† Department of Chemistry and Biochemistry.
‡ Radiation Laboratory.
(1) cis-trans Isomerization in Biochemistry; Dugave, C., Ed.; J Wiley &
Sons: New York, 2006.
(3) Voet, D.; Voet, J. G. Biochemistry, 3rd ed.; J. Wiley & Sons: New
York, 2004; p 290.
(2) (a) Stein, R. L. AdV. Protein Chem. 1993, 44, 1–24. (b) Schmid, F. X.;
Mayr, L. M.; Mu¨cke, M.; Scho¨nbrunner, E. R. AdV. Protein Chem.
1993, 44, 25–66. (c) Hur, S.; Bruice, T. C. J. Am. Chem. Soc. 2002,
124, 7303–7313. (d) Agarwal, P. K.; Geist, A.; Gorin, A. Biochemistry
2004, 43, 10605–10618. (e) Fangha¨nel, J.; Fischer, G. Front. Biosci.
2004, 9, 3453–3478. (f) Lu, K. P.; Finn, G.; Lee, T. H.; Nicholson,
L. K. Nat. Chem. Biol. 2007, 3, 619–629.
(4) Scherer, G.; Kramer, M. L.; Schutkowski, M.; Reimer, U.; Fischer,
G. J. Am. Chem., Soc. 1998, 120, 5568–5574.
(5) Schiene-Fischer, C.; Fischer, G. J. Am. Chem. Soc. 2001, 123, 6227–
6231.
(6) Schiene-Fischer, C.; Habazettl, J.; Schmid, F. X.; Fischer, G. Nat.
Struct. Biol. 2002, 9, 419–424.
(7) Tchaicheeyan, O. FASEB J. 2004, 18, 783–789.
9
10.1021/ja9086787  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 4641–4652 4641

A R T I C L E S
Hu et al.
Scheme 1
column was eluted with distilled water at 2.7 mL/min, 20-mL
fractions were collected, and the fractions were assayed for
aminosugar using phenol-sulfuric acid.10,11 Methyl 2-acetamido-
2-deoxy-R-D-glucopyranoside eluted first (fractions 17-19) (870
mg, 3.70 mmol, 27.2%), followed by methyl 2-acetamido-2-deoxy-
ꢀ-D-glucopyranoside (fractions 21-25) (1.15 g, 4.89 mmol, 36.0%).
Scheme 2
Methyl 2-acetamido-2-deoxy-R-D-glucopyranoside (830 mg, 3.53
mmol) was dissolved in 30 mL of H2O, and barium oxide12 was
added batchwise to adjust the solution pH to ∼13. The reaction
mixture was refluxed for 12 h. After cooling to room temperature,
the precipitate was removed by filtration, and the filtrate was treated
with dry ice to precipitate barium salts. After filtration to remove
the salts, the filtrate was concentrated to a syrup at 30 °C in Vacuo
to give methyl 2-amino-2-deoxy-R-D-glucopyranoside (677 mg, 3.53
mmol, 100%).
Methyl 2-amino-2-deoxy-R-D-glucopyranoside (670 mg, 3.47
mmol) was dissolved in a minimal volume of H2O, and EtOH (25
mL) was added. Sodium [13C]formate (99 atom % 13C; Cambridge
Isotope Laboratories, Inc.) (335 mg, 4.86 mmol, dissolved in 6 mL
of H2O) and EEDQ10 (2-ethoxy-1-ethoxycarbonyl-1,2-dihydro-
quinone) (857 mg, 3.47 mmol dissolved in 20 mL of EtOH) were
added to the aminosugar solution with stirring. The reaction vessel
was sealed with a rubber septum and covered with aluminum foil,
and the solution was stirred at 30 °C for 24 h. The reaction mixture
was then evaporated in Vacuo at 30 °C to a syrup to remove ethanol
and water. Distilled water (20 mL) was added to dissolve the
aminosugar syrup, and the solution was extracted with 6 × 15 mL
of CH2Cl2. The resulting aqueous solution was deionized with
batchwise additions of excess Dowex 50 × 8 (20-50 mesh) (H+)
and Dowex 1 × 8 (20-50 mesh) (OAc-) ion-exchange resins. After
filtration to remove the resins, the deionized filtrate was concentrated
at 30 °C in Vacuo to remove acetic acid. The crude methyl 2-deoxy-
2-[13C]formamido-R-D-glucopyranoside (1) was purified by chro-
matography on a column (1.5 cm × 68 cm) containing Dowex 50
notably, the N-acetyl side chains of N-acetyl-D-glucosamine
(GlcNAc; NAG) and N-acetyl-neuraminic acid (Neu5Ac) (Scheme
1). While it is commonly held that amide conformation in
GlcNAc monomers is exclusively trans, it is known that, in
some glycoprotein X-ray structures, NAG residues are observed
with the N-acetyl side chain in the cis conformation.8 This
observation led us to ponder whether the amide cis-trans
equilibrium can be detected and quantified in aqueous solutions
of methyl glycosides of GlcNAc and GlcNAc-related com-
pounds (Scheme 2). In this report, we describe NMR approaches
to detect and quantify cis and trans amides in aqueous solutions
of these glycosides. Using saturation-transfer 13C NMR methods,
first-order rate constants and kinetic activation parameters have
been measured for CTI in saccharides. We show that the
cis-trans equilibrium and exchange kinetics depend on sac-
charide anomeric configuration and side chain structure (N-acetyl
Vs N-formyl) (Scheme 2), demonstrating the importance of local
structure on this side chain rearrangement.
Experimental Section
Methyl 2-Deoxy-2-[13C]formamido-r-D-glucopyranoside (1)
(r-MeNFG) and Methyl 2-Deoxy-2-[13C]formamido-ꢀ-D-glu-
copyranoside (2) (ꢀ-MeNFG). 2-Acetamido-2-deoxy-D-glucose
(3.00 g, 13.6 mmol) was dissolved in 30 mL of absolute MeOH,
and Dowex HCR-W2 (H+) resin was added with gentle stirring.
The reaction mixture was refluxed for 1 h, cooled, and filtered to
remove the resin, and the filtrate was concentrated to a syrup at 30
°C in Vacuo to give a crude mixture of glycoside anomers (2.87 g,
12.2 mmol). The syrup was dissolved in a minimal volume of H2O,
and the solution applied to a column (3.0 cm × 50 cm) containing
Dowex 1 × 8 (200-400 mesh) (OH-) ion-exchange resin.9,10 The
(9) Austin, P. W.; Hardy, F. E.; Buchanan, J. C.; Baddiley, J. J. Chem.
Soc. 1963, 5350–5353.
(10) Zhu, Y.; Pan, Q.; Thibaudeau, C.; Zhao, S.; Carmichael, I.; Serianni,
A. S. J. Org. Chem. 2006, 71, 466–479.
(11) Hodge, J. E.; Hofreiter, B. T. Methods Carbohydr. Chem. 1962, 1,
380–394.
(12) Neumann, J.; Weingarten, S.; Thiem, J. Eur. J. Chem. 2007, 1130-
1144. (These authors used Ba(OH)2 octahydrate in this prior work;
BaO was used as a substitute in the current work.)
(8) Bode, W.; Wei, A. Z.; Huber, R.; Meyer, E.; Travis, J.; Neumann, S.
EMBO J. 1986, 5, 2453–2458.
9
4642 J. AM. CHEM. SOC. VOL. 132, NO. 13, 2010

Amide Cis-Trans Isomerization
A R T I C L E S
× 8 (200-400 mesh) ion-exchange resin in the Ca2+ form13 to
afford 560 mg (2.52 mmol, 72.6%) of 1.
Hz spectral window, 20 s (1/2) and 30 s (3/4) relaxation delays,
1.8 s acquisition time. FIDs were treated with a 2-Hz exponential
line-broadening function and were zero-filled once to give final
spectral digital resolutions of ∼0.14 Hz/pt. Peak areas of the cis
and trans carbonyl or C2 resonances were determined by computer
integration. Signals were carefully phased and starting and ending
points selected to minimize integration errors. For each signal, the
integration was performed three times, and the average value was
used in the determination of solution percentages, from which trans/
cis equilibrium constants, Ktrans/cis, were calculated.
Methyl 2-deoxy-2-[13C]formamido-ꢀ-D-glucopyranoside (2) (ꢀ-
MeNFG) was prepared by a procedure similar to that used for 1.
A detailed description is available in Supporting Information.
Methyl 2-[1-13C]Acetamido-2-deoxy-r-D-[2-13C]glucopyrano-
side (3) (r-MeNAG) and Methyl 2-[1-13C]Acetamido-2-deoxy-
ꢀ-D-[2-13C]glucopyranoside (4) (ꢀ-MeNAG). 2-Amino-2-deoxy-
D-[2-13C]glucose hydrochloride (400 mg, 1.85 mmol; Omicron
Biochemicals, Inc.) was dissolved in a minimal volume of distilled
H2O, and the solution pH was adjusted to 6.6 with the addition of
Dowex 1 × 8 (200-400 mesh) (OH-) ion-exchange resin. The
resin was removed by filtration and washed with water, giving a
total filtrate volume of ∼6 mL to which was added 25 mL of EtOH.
Sodium [1-13C]acetate (99 atom % 13C; Cambridge Isotope Labo-
ratories, Inc.) (230 mg, 2.77 mmol, dissolved in a small amount of
distilled H2O) and EEDQ10 (2-ethoxy-1-ethoxycarbonyl-1,2-dihy-
droquinone) (684 mg, 2.77 mmol, dissolved in 10 mL of EtOH)
were added to the aminosugar solution with stirring. The reaction
vessel was sealed with a rubber septum and covered with aluminum
foil, and the reaction solution was stirred at 30 °C for 24 h. The
reaction solution was then concentrated in Vacuo at 30 °C to a stiff
syrup to remove ethanol and water. Distilled water (20 mL) was
added to dissolve the syrup, and the solution was extracted with 6
× 15 mL of CH2Cl2. The resulting aqueous solution was deionized
with batchwise additions of excess Dowex 50 × 8 (20-50 mesh)
(H+) and Dowex 1 × 8 (20-50 mesh) (OAc-) ion-exchange resins,
and the deionized solution was concentrated at 30 °C in Vacuo to
remove acetic acid. The crude yield of 2-[1-13C]acetamido-2-deoxy-
D-[2-13C]glucopyranoses was 390 mg (1.75 mmol, 94.5%).
The 2-[1-13C]acetamido-2-deoxy-D-[2-13C]glucopyranoses (360
mg, 1.61 mmol) were dissolved in 30 mL of absolute MeOH, and
Dowex HCR-W2 (H+) resin (∼0.5 g) was added with gentle stirring.
The reaction mixture was refluxed for 1 h, cooled, and filtered to
remove the resin, and the filtrate was concentrated to a syrup of
crude glycosides (300 mg, 1.27 mmol). The syrup was dissolved
in a minimal volume of distilled H2O, and the solution was applied
to a column (3.0 cm × 50 cm) containing Dowex 1 × 8 (200-400
mesh) (OH-) ion-exchange resin.9,10 The column was eluted with
distilled water at 2.7 mL/min, and 20-mL fractions were collected.
Column fractions were assayed for aminosugar using a 1% (w/v)
CeSO4/2.5% (w/v) (NH4)6Mo7O24/10% aq H2SO4 reagent14 (fraction
samples spotted on a silica gel TLC plate, followed by spraying
with the reagent and charring on a hot plate). Methyl 2-[1-
13C]acetamido-2-deoxy-R-D-[2-13C]glucopyranoside (3) eluted first
(fractions 17-19) (98 mg, 0.414 mmol, 25.7%), followed by methyl
2-[1-13C]acetamido-2-deoxy-ꢀ-D-[2-13C]glucopyranoside (4) (frac-
tions 21-25) (138 mg, 0.582 mmol, 36.1%).
1D 1H NMR spectra were obtained on a 600 MHz NMR
spectrometer using a 90° pulse width, a spectral window of 9615.4
Hz, a relaxation delay of 0.6 s, and a 3.4 s acquisition time. FIDs
were zero-filled twice to give final spectral digital resolutions of
∼0.04 Hz/pt. 2D 1H-1H gCOSY15 and 13C-1H gHSQC16 spectra
1
were performed using standard NMR software to confirm H and
13C signal assignments.
Measurement of CTI Rate Constants by 13C Saturation-
Transfer NMR Spectroscopy. First-order rate constants for CTI
in 1-4 were measured as a function of temperature by 13C
saturation-transfer NMR (ST-NMR)17-20 using compounds 1-4
selectively labeled with 13C at the carbonyl carbon of the N-acyl
side chain and/or the C2 ring carbon.
kcisftrans
cis amide y
z trans amide
ktransfcis
The Bloch equation, modified to account for chemical exchange
and describing the change in intensity (Mz) of the carbonyl carbon
(or C2) resonance as a function of time, t, after the application of
a nonselective 90° pulse, is as follows:
dMtzrans
dt
-[Mtzrans(t) - Mztrans(0)]
)
- ktransfcisMztrans(t) +
T1,trans
cis
z
k
M (t)
(1)
cisftrans
Single-resonance saturation experiments were conducted at 150
MHz (13C). The selected cis resonance was saturated for times
ranging from 0.005 to 50 s before application of a 90° observe
pulse, and a relaxation delay of 20-30 s was used to allow for
complete relaxation. In the saturation-transfer experiment, the cis
resonance, denoted Mzcis, is saturated, which causes Mzcis f 0 and
cis
thus k
M
(t) f 0. The resulting equation containing terms
cisftrans
z
for Mztrans, T1,trans, and ktransfcis can be integrated to the following
form:
Methyl 2-[1-13C]Acetamido-2-deoxy-r-D-[1,2-13C2]glucopyrano-
side (3′) (r-MeNAG) and Methyl 2-[1-13C]Acetamido-2-deoxy-
ꢀ-D-[1,2-13C2]glucopyranoside (4′) (ꢀ-MeNAG). Compounds 3′
and 4′ were prepared by a procedure similar to that used for 3 and
4. A detailed description is available in Supporting Information.
NMR Spectroscopy. Detection and Quantification of Amide
cis and trans Isomers of 1-4 in Aqueous Solution. 13C NMR
samples were prepared by dissolving ∼20 mg of 1-4 in 0.7 mL of
2H2O to give ∼0.13 M solutions. The p2H values of the NMR solutions
were adjusted to 8.1 ( 0.1 with Dowex HCR-W2 (H+) ion-exchange
resin or dilute aqueous NaO2H to eliminate potential effects of different
[2H+] on the measured equilibrium and rate constants.
τ1,trans
τ1,trans
-τ
τ1,trans
Mtzrans(τ) ) Mztrans(0)
exp
+
(2)
(
)
(
)
τtrans
T1,trans
where Mztrans(τ) is the intensity of the trans resonance at time τ
after onset of saturation of the cis resonance, and Mztrans(0) is the
intensity of the trans resonance in the absence of saturation.
Furthermore,
(15) (a) von Kienlin, M.; Moonen, C. T. W.; van der Toorn, A.; van Zijl,
P. C. M. J. Magn. Reson. 1991, 93, 423–429. (b) Canet, D. Prog.
NMR Spectrosc. 1997, 30, 101–135.
(16) (a) Vuister, G. W.; Boelens, R.; Kaptein, R.; Hurd, R. E.; John, B.;
van Zijl, P. C. M. J. Am. Chem. Soc. 1991, 113, 9688–9690. (b) Kay,
L. E.; Keifer, P.; Saarinen, T. J. Am. Chem. Soc. 1992, 114, 10663–
10665.
The percentages of cis and trans forms in aqueous (2H2O)
solutions of 1-4 at different temperatures were determined by
13C{1H} NMR spectroscopy (150 MHz 13C). Data acquisition and
processing conditions were as follows: 90° pulse width; 36764.7
(17) Forse´n, S.; Hoffman, R. A. J. Chem. Phys. 1963, 39, 2892–2901.
(18) Campbell, I. D.; Dobson, C. M.; Ratcliffe, R. G.; Williams, R. J. P.
J. Magn. Reson. 1978, 29, 397–417.
(13) Angyal, S. J.; Bethell, G. S.; Beveridge, R. J. Carbohydr. Res. 1979,
73, 9–18.
(14) Tropper, F. D.; Andersson, F. O.; Grand-Maitre, C.; Roy, R.
Carbohydr. Res. 1992, 229, 149–154.
(19) Serianni, A. S.; Pierce, J.; Huang, S.-G.; Barker, R. J. Am. Chem.
Soc. 1982, 104, 4037–4044.
(20) Snyder, J. R.; Johnston, E. R.; Serianni, A. S. J. Am. Chem. Soc. 1989,
111, 2681–2687.
9
J. AM. CHEM. SOC. VOL. 132, NO. 13, 2010 4643

A R T I C L E S
Hu et al.
in the calculation by using RADII)UAHF in the input file to ensure
convergence during energy minimization.
1
1
τtrans
1
)
+
(3)
τ1,trans
T1,trans
where τtrans ) 1/ktransfcis and T1,trans is the spin-lattice relaxation
time of the trans resonance.
For 3 and 4 (MeNAG anomers), the small percentage of cis form
in solution at equilibrium (<3%) allowed for rapid (i.e., within 0.005
s) and selective saturation of the cis C1′ (carbonyl) carbon
resonance. Consequently, ktransfcis was measured from a plot of
ln(Mztrans(τ) - Mztrans(∞)) Versus saturation time τ, where
Mztrans(∞) is the intensity of the trans C1′ carbon resonance an
-1
“infinite” time (defined as 5(ktransfcis + 1/T1,trans
)
after applying
the saturating irradiation, giving a line with slope ) -1/τ1,trans. The
relaxation time was determined from the relationship
Mtzrans(∞)
Mtzrans(0)
τ1,trans
)
(4)
T1,trans
Model structures 5-8 were used to mimic compounds 1-4,
respectively. Calculations were conducted on 5-8 in two limiting
conformations about R, defined as the H2-C2-N2-H torsion
angle; these conformations are denoted R-syn and R-anti. In both
R conformations, the ꢀ torsion angle, defined as H-N2-C1′-O
in 5/6 and C2-N2-C1′-C2′ in 7/8, was rotated in 15° increments
through 360°, giving a total of 24 optimized structures each for
5-8 (total of 96 optimized structures). Exocyclic torsion angles in
both R-syn and R-anti forms were constrained as follows: the
C2-C1-O1-C7 and the C2-C3-O3-H torsion angles were fixed
at 180°. In addition, the C3-C2-N2-C1′ torsion angle was fixed
at 60° for R-syn and -120° for R-anti, the C1-C2-N2-C1′ torsion
angle was fixed at -60° for R-syn and 120° for R-anti, and the
H2-C2-N2-H torsion angle was fixed at 0° for R-syn and 180°
for R-anti. All other geometric parameters were allowed to optimize.
Energy data obtained for 5/6 and 7/8 were normalized within each
pair; for example, the lowest energy structure within the GlcNAc
series was 7 (solvated) in the R-anti conformation with a
C2-N2-C1′-C2′ torsion angle of 180° (trans amide). The total
energy of this structure was arbitrarily set at 0 kcal/mol, and the
energies of the remaining 95 structures within the series are reported
relative to this reference structure. Data for 5/6 were treated
similarly.
The values of Mztrans(0) and Mztrans(∞) were determined using τ
values of 0.005 and 50 s, respectively. Therefore, both τ1,trans and
T1,trans could be computed with precision, and eq 3 applied to
determine τtrans. From τtrans, the rate constant ktransfcis was calculated.
From the Ktrans/cis equilibrium constant and ktransfcis at a given
temperature, kcisftrans was determined.
For 1 and 2 (MeNFG anomers), the relatively large percentage
of cis form in solution (∼20%) prevented rapid and selective
saturation of the cis signal within a short time period (i.e., 0.005
s). Therefore a different procedure was followed to determine
ktransfcis values. The Mztrans(∞)/Mztrans(0) ratio was determined by
conducting two saturation transfer experiments with identical
saturation times (20 s), with saturation at the cis carbonyl carbon
resonance frequency (on-resonance) in the first experiment and with
saturation at a frequency far from all other resonances in the
spectrum (130 ppm; off-resonance) in the second. A relaxation delay
of 20 s was used in both experiments. These two experiments
yielded the ratio, Mztrans(∞)/Mztrans(0).. T1,trans was then determined
directly via an inversion-recovery experiment.21,22 With Mztrans(∞)/
Mztrans(0) and T1,trans available, τ1,trans was determined from eq 4,
and τtrans determined from eq 3. From τtrans, ktransfcis was calculated.
Using the equilibrium constant measured at the same temperature,
kcisftrans was calculated.
Results and Discussion
q
q
q
Values of Eact, ∆G° , ∆H° , and ∆S° for cisftrans and
transfcis reactions of 1-4 were calculated as described in
Supporting Information.
General Experimental Strategy. The primary aim of this
investigation was to determine whether both cis and trans amide
conformations of GlcNAc glycosides 3 and 4 are detectable in
aqueous solution by NMR spectroscopy. It was expected that
very minor amounts of the cis form would be present, making
its identification by NMR difficult (i.e., weak signals might arise
from the presence of contaminants rather than from the cis
isomer). To address these concerns, methyl glycosides of
N-formyl derivatives 1 and 2 were prepared, since prior work
had shown that cis and trans amides coexist in aqueous solution
in comparable amounts in related structures.29-32 Our underlying
assumption was that the effects of amide conformation on the
13C chemical shifts of 3 and 4 would be similar to those observed
in 1 and 2. In addition, 3 and 4 were prepared with 13C-labeling
Calculations
Geometry optimizations were conducted with the use of Gauss-
ian03,23 and density functional theory (DFT) was employed with
the B3LYP functional24 and the 6-31G* basis set25 (B3LYP/6-
31G*). The effect of solvent water was evaluated using the self-
consistent reaction field (SCRF),26 the integral equation formalism
(polarizable continuum) model (IEFPCM),27 and the above-noted
B3LYP/6-31G* computational method as implemented in Gauss-
ian03. The united atom topological model (UAHF)28 was applied
(21) Vold, R. L.; Waugh, J. S.; Klein, M. P.; Phelps, D. E. J. Chem. Phys.
1968, 48, 3831–3832.
(22) Weiss, G. H.; Ferretti, J. A. Prog. NMR Spectrosc. 1988, 4, 317–335.
(23) Frisch, M. J.; et al. Gaussian03, ReVision A.1; Gaussian, Inc.:
Pittsburgh, PA, 2003.
(29) Katzenellenbogen, E.; Romanowska, E.; Kocharova, N. A.; Shashkov,
A. S.; Knirel, Y. A.; Kochetkov, N. K. Carbohydr. Res. 1995, 273,
187–195.
(24) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
(25) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972, 56,
2257–2261.
(30) Muldoon, J.; Shashkov, A. S.; Senchenkova, S. N.; Tomshich, S. V.;
Komandrova, N. A.; Romanenko, L. A.; Knirel, Y. A.; Savage, A. V.
Carbohydr. Res. 2001, 330, 231–239.
(26) Cance´s, M. T.; Mennucci, B.; Tomasi, J. J. Chem. Phys. 1997, 107,
3032–3041.
(31) Kocharova, N. A.; Maszewska, A.; Zatonsky, G. V.; Bystrova, O. V.;
Ziolkowski, A.; Torzewska, A.; Shashkov, A. S.; Knirel, Y. A.;
Rozalski, A. Carbohydr. Res. 2003, 338, 1425–1430.
(32) Breazeale, S. D.; Ribeiro, A. A.; McClerren, A. L.; Raetz, C. R. H.
J. Biol. Chem. 2005, 280, 14154–14167.
(27) Cammi, R.; Mennucci, B.; Tomasi, J. J. Phys. Chem. A 2000, 104,
5631–5637.
(28) Barone, V.; Cossi, M.; Tomasi, J. J. Chem. Phys. 1997, 107, 3210–
3221.
9
4644 J. AM. CHEM. SOC. VOL. 132, NO. 13, 2010

Amide Cis-Trans Isomerization
A R T I C L E S
Figure 1. DFT-calculated total energies for 5 (A), 6 (B), 7 (C), and 8 (D) as a function of rotation of the N2-C1′ amide bond (0° ) pure cis; 180° ) pure
trans). Solid symbols ) in Vacuo energies; open symbols ) solvated energies. Red symbols ) R-syn; green symbols ) R-anti.
at both C2 and C1′ (the side chain carbonyl carbon) to allow
detection of the geminal 2JC2,C1′ spin-coupling constant between
the labeled carbons. Observation of internally consistent 2JC2,C1′
splitting patterns would thus provide further evidence for the
cis isomer, because weak signals from contaminants would
likely show either no splitting or a splitting pattern different
from that expected. Upon detection of signals attributable to a
putative cis isomer based on chemical shift and JCC arguments,
13C-labeling at either the side chain carbonyl carbon or the C2
ring carbon would then be exploited in quantitative 13C
saturation-transfer experiments to confirm the presence of
cis a trans chemical exchange and allow measurements of first-
order rate constants as a function of temperature, from which
kinetics parameters could be determined. The magnitude of the
latter values would provide indirect evidence for cis a trans
equilibria in solution, since these parameters have been measured
previously in related systems.4,5,33
to substantial amide hydrolysis (Scheme S1, Route A; see
Supporting Information). Consequently, an alternate route was
followed (Scheme S1, Route B; see Supporting Information)
in which GlcNAc was first converted to its methyl glycopyra-
nosides and the latter were hydrolyzed with aqueous BaO to
give intermediate glycosides of the free amine. The latter
hydrolysis occurred relatively smoothly and in reasonable yield.
The free amine glycosides were then converted to the N-formyl
derivatives using sodium [13C]formate and EEDQ, and the
glycopyranoside anomers were purified by chromatography.
Energies of the Cis and Trans Isomers of 5-8. DFT-Derived
energies of 5-8 as a function of conformation about the R and
ꢀ bonds are shown in Figure 1 and are summarized in Table
S1 (see Supporting Information). In Vacuo and solvated energies
were calculated and compared (Figure 1). Inclusion of solvent
lowered absolute energies significantly as expected, but relative
energies were largely unchanged. Activation barriers for the
interconversion of cis and trans isomers range from 18 to 29
kcal/mol, and in most but not all cases, inclusion of solvent
enhanced the barriers. Activation barriers for cis f trans differ
slightly depending on the pathway of rotation, since the system
is asymmetric. In general, barriers are slightly higher for the
-90° f 180° pathway than for the 90° f 180° pathway in
5/6; for 7/8, no general pattern emerges from the data (Table
S1). Experimental energies of activation (Eact) for cis f trans
of 65-82 kJ/mol (15.5-19.6 kcal/mol) have been reported
previously for the peptide bonds of oligopeptides.4
Synthesis of 13C-labeled GlcNAc methyl glycosides was ac-
complished using methods described previously.10 2-Amino-2-
deoxy-D-glucose was converted to its 2-deoxy-2-acetamido de-
rivative using sodium acetate and EEDQ. The resulting N-acetyl-
D-glucosamine was subsequently converted to an anomeric mixture
of methyl glycopyranosides, and the glycosides were separated by
chromatography. D-[1-13C]Glucosamine, D-[1,2-13C2]glucosamine,
and/or sodium [1-13C]acetate were used in the protocol to give two
sets of 13C-isotopomers 3/4 and 3′/4′.
Synthesis of the methyl glycosides of NFG was less straight-
forward. Attempts to glycosidate N-formyl-D-glucosamine led
Within the solvated data set for MeNFG anomers, the most stable
structure was 5 (R-anti) in the trans amide conformation, followed
closely by 5 (R-anti) [+0.06 kcal/mol] (cis amide), 6 (R-anti)
(33) Wawra, S.; Fischer, G. In cis-trans Isomerization in Biochemistry;
Dugave, C., Ed.; J. Wiley & Sons: New York, 2006; pp 167-193.
9
J. AM. CHEM. SOC. VOL. 132, NO. 13, 2010 4645

A R T I C L E S
Hu et al.
effects are caused by the shielding anisotropy of the exocyclic
carbonyl group, with its influence on nearby 1H shifts determined
by differences in the orientation of the CdO bond relative to
specific ring protons. Ring protons remote from the side chain are
less subject to this effect, and the small shift changes observed
may be due to subtle effects on ring molecular orbitals mediated
by amide conformation and/or small changes in pyranosyl ring
conformation.
Virtually all 13C NMR signals in 1 and 2 are affected by amide
conformation (Figure 3; Table 1; for 13C-13C spin-couplings,
see Table 3). The largest differences upon conversion of trans
f cis are observed at C2 and the exocyclic formyl carbon (∼4.7
and ∼3.2 ppm downfield shifts, respectively).
In summary, NMR data obtained on the NFG methyl
glycosides reveal chemical shift dependencies on side chain
amide conformation that will be exploited in the following
analyses of NMR data obtained on the NAG methyl glycosides
1
2
Figure 2. 600 MHz H NMR spectrum of 2 in H2O at 22 °C showing
signals arising from the anomeric H1 and side chain formyl hydrogen in
the cis and trans forms of the molecule. The formyl hydrogen signals appear
as doublets as a result of the presence of one-bond 13C-1H spin-coupling
involving the 13C-labeled formyl carbon (1JCH ) 195.6 Hz (cis isomer) and
197.7 Hz (trans isomer).
1
3 and 4. The underlying assumption is that the direction of H
and/or 13C chemical shift changes observed in 1/2 and their
magnitudes will be similar to those in 3/4.
The large difference in 13C chemical shifts of the formyl (C1′)
carbon in the cis and trans forms and the presence of
13C-labeling at C1′ facilitated measurements of Ktrans/cis in 1 and
2 at different solution temperatures (Table 2). Ktrans/cis is affected
by anomeric configuration; at all temperatures studied, the
percentage of cis form in solution is greater in ꢀ-anomer 2 than
in R-anomer 1. For example, at 42°, Ktrans/cis is 4.22 ( 0.03 for
1 and 2.84 ( 0.02 for 2, translating into ∼19% cis for 1 and
∼26% cis for 2. van’t Hoff plots of the data in Table 2 (see
Figure S1 in Supporting Information) gave the following
standard enthalpy changes, ∆Hocisftrans, for the conversion of
cis to trans forms: -1.1 ( 0.1 kcal/mol for 1; -0.8 ( 0.1 kcal/
[+1.37 kcal/mol] (cis amide), and 6 (R-anti) [+1.70 kcal/mol]
(trans amide). For MeNAG anomers, the most stable structure was
7 (R-anti) in the trans amide conformation, followed closely by 8
(R-anti) (+1.98 kcal/mol (trans amide)). The latter findings are
consistent qualitatively with NMR data showing a greater propor-
tion of N-acetyl-R-D-glucosamine than N-acetyl-ꢀ-D-glucosamine
(68% R-pyranose; 32% ꢀ-pyranose) in aqueous solution and with
the commonly held assumption that R-anti conformations are more
stable than R-syn conformations in GlcNAc anomers based on
magnitudes of JH2,NH values.34,35 In MeNFG models 5 and 6,
3
mol for 2. ∆Hocisftrans and standard entropy changes, ∆So
,
R-anti structures appear favored in both anomers, but a more
balanced distribution of cis and trans amide conformations is
predicted in these structures. It should be noted, however, that these
energy calculations will be influenced by subtle intramolecular
H-bonding involving side chain atoms and adjacent OH groups at
C1 and C3, which can skew energies in favor of forms that are
not significantly populated in solution, although the inclusion of a
polarizable dielectric continuum in the calculation probably miti-
gates this effect to some extent. In addition, the present calculations
explore a limited set of exocyclic C-O bond torsions in the vicinity
of the N-formyl/N-acetyl side chain, and thus only portions of the
full energy hypersurfaces were sampled. Therefore, although the
calculations yield reasonable cis-trans interconversion barriers,
relative stabilities of pyranose anomers, and conformational
behavior about R, they are not expected to yield quantitative
energies of cis and trans isomers and thus accurate cis/trans
equilibria.
cisftrans
were determined from van’t Hoff plots and used to calculate
standard free energy changes, ∆Gocisftrans: ∆Go
and
cisftrans
∆So
for 1, -0.9 ( 0.1 kcal/mol and -0.8 ( 0.3 cal/K/
cisftrans
mol; for 2, -0.7 ( 0.1 kcal/mol and -0.6 ( 0.2 cal/K/mol. In
1 and 2, the conversion of cis f trans amide is enthalpically
favored but entropically disfavored, explaining the observed
temperature dependence of Ktrans/cis (higher temperatures reduce
Ktrans/cis and increase the percentage of cis form in solution).
13C NMR spectra of 1 and 2 yielded JCC values involving
the side chain 13C-labeled formyl carbon (C1′) and the natural
abundance carbons of the pyranosyl ring; these spin-couplings
2
3
3
include JC2,C1′, JC1,C1′, and JC3,C1′ (Figure 3, Table 3). While
the latter two couplings are expected to depend strongly on
C2-N2 bond conformation (Karplus dependency) and thus can
vary appreciably in magnitude in different structures, the
2
geminal JC2,C1′ is likely to be less structure-sensitive and thus
NMR Spectra of 1 and 2. 1D 1H NMR spectra of 1 and 2 show
the presence of cis and trans forms of each molecule in aqueous
solution (Figure 2; Table 1; for 1H-1H and 13C-1H spin-couplings,
see Table S2 in Supporting Information). 1H nuclei most affected
by amide conformation are H1, H2, H3, OCH3, and the formyl
hydrogen, with the largest effect observed at H2 (∼0.5 ppm),
followed by the formyl hydrogen (0.1-0.2 ppm) (Table 1). The
direction of chemical shift changes observed upon the conversion
of trans f cis varies with proton site and anomeric configuration.
The H2, H3, and formyl hydrogen signals shift upfield, whereas
the H1 and OCH3 signals shift downfield in 1; similar behavior is
found for 2, except in this case H1 shifts upfield. In contrast, the
chemical shifts of H4-H6a/H6b in 1/2 are minimally affected by
amide conformation (changes < 0.03 ppm). Presumably these
serves as a more reliable probe in the assignment of weak signals
arising from cis forms in 13C NMR spectra of 3 and 4. For this
reason, 3 and 4 were prepared each with two sites of 13C
enrichment to detect the CdO (C1′) and C2 signals selectively;
(34) (a) Horton, D.; Jewell, J. S.; Philips, K. D. J. Org. Chem. 1966, 31,
4022–4025. (b) Neuberger, A.; Fletcher, A. P. Carbohydr. Res. 1971,
17, 79–88. (c) Two arguments have been offered to explain the
predominant R-anomeric configuration of GlcNAc in aqueous solution.
The first invokes a dipole-dipole interaction between the C1-O1 and
the amide CdO bonds that is minimized in the R-pyranose.34b The
second invokes hydrogen-bonding between the amide NAc side chain
hydrogen and O1 as a means of preferentially stabilizing the
R-pyranose (cis-fused five-membered ring).34a Both arguments pre-
sume an approximate R-anti/ꢀ-trans conformation of the NAc side
chain in GlcNAc.
9
4646 J. AM. CHEM. SOC. VOL. 132, NO. 13, 2010

Amide Cis-Trans Isomerization
A R T I C L E S
Table 1. 1H and 13C Chemical Shift Assignmentsa for 1 and 2 in Their Cis and Trans Amide Conformations
cmpd
δH1
δH2
δH3
δH4
δH5
δH6a
δH6b
δOCH
δ(CO)H
3
1/ꢀ-cis
4.797
4.747
4.423
4.454
∼3.47
3.964
3.172
3.714
3.61-3.66
3.669
3.504
3.421
3.450
3.37-3.45
3.37-3.45
3.61-3.66
3.61-3.66
3.37-3.45
3.37-3.45
∼3.84
3.840
3.895
3.895
3.741
3.744
3.708
3.708
3.369
3.357
3.514
3.472
7.970
8.103
7.931
8.153
1/ꢀ-trans
2/ꢀ-cis
2/ꢀ-trans
3.535
cmpd
δC1
δC2
δC3
δC4
δC5
δC6
δOCH
δCdO
3
1/ꢀ-cis
101.48
100.61
104.00
104.28
59.52
54.77
61.56
56.82
73.72
73.77
75.94
76.23
72.17
72.39
72.27
72.44
74.27
74.30
78.41
78.47
63.07
63.10
63.21
63.27
57.74
57.69
59.98
59.69
170.07
166.96
170.68
167.33
1/ꢀ-trans
2/ꢀ-cis
2/ꢀ-trans
a
1H chemical shifts in ppm relative to external dimethyl-2-silapentane-5-sulfonic acid (DSS); 13C chemical shifts in ppm relative to external DSS.
2
Measured at 22 °C in H2O.
Figure 3. Partial 13C{1H} NMR spectra (150 MHz) of 2 in 2H2O at 22 °C. (A) Anomeric C1 signals in the cis and trans amides, showing splittings due to
2
3JC1,C1′. (B, C) C2 signals in the cis and trans amides, showing splittings due to JC2,C1′. (D) C3 signals in the cis and trans amides showing splittings due
3
to JC3,C1′. (E) Intense signals from the labeled formyl (C1′) carbon in the cis and trans amides. See Table 3 for measured J-couplings.
both chemical shifts are sensitive to amide conformation as
shown in 1/2 (Table 1), and the easily measured 2JC2,C1′ between
both 13C-labeled carbons could be used to validate C1′ and C2
signal assignments in cis isomers.
labeled C2 and exocyclic CdO (C1′) carbons of the trans form
(Figure 4). In each spectrum, identical splittings were observed
2
in both signals due to the presence of JC2,C1′, thus confirming
13C2-N2-13C1′ bond connectivity. On the basis of the 13C
chemical shift behaviors established for 1 and 2, a close
examination was made of regions 4-6 ppm downfield of each
NMR Spectra of 3 and 4. 1D 1H NMR spectra of 3 and 4 in
aqueous (2H2O) solution contained signals from the dominant
trans amide form (Table 4). Efforts to detect and assign signals
arising from the cis amides were unsuccessful because of their
weak intensities in spectra containing significant signal overlap.
In contrast, 1D 13C{1H} NMR spectra of the doubly 13C-
labeled 3 and 4 contained strong doublets attributed to the
(35) Mobli, M.; Almond, A. Org. Biomol. Chem. 2007, 5, 2243–2251.
(36) (a) Church, T.; Carmichael, I.; Serianni, A. S. Carbohydr. Res. 1996,
280, 177–186. (b) Serianni, A. S.; Bondo, P. B.; Zajicek, J. J. Magn.
Reson. 1996, 112B, 69–74. (c) Zhao, S.; Bondo, G.; Zajicek, J.;
Serianni, A. S. Carbohydr. Res. 1998, 309, 145–152.
9
J. AM. CHEM. SOC. VOL. 132, NO. 13, 2010 4647

A R T I C L E S
Hu et al.
Table 2. Ktrans/cis for 1 and 2 at Different Solution Temperatures
van’t Hoff plots of the data in Table 5 gave the following
∆Ho
values: for 3, -3.2 ( 0.4 kcal/mol; for 4, -2.6 (
cisftrans
T (°C)
Ktrans/cis (1; R)
Ktrans/cis (2; ꢀ)
0.2 kcal/mol. These values are slightly more negative than those
42.0
52.7
58.0
64.4
69.7
75.1
84.6
4.22 ( 0.03a
3.95 ( 0.02
3.84 ( 0.02
3.76 ( 0.02
3.64 ( 0.02
3.54 ( 0.02
3.37 ( 0.02
2.84 ( 0.02
2.69 ( 0.02
found for 1 and 2. ∆Socisftrans Values were also determined from
van’t Hoff plots, and ∆Go
values were calculated from
cisftrans
∆Hocisftrans and ∆Socisftrans; for 3, ∆Gocisftrans ) -2.7 ( 0.1 kcal/
2.59 ( 0.02
mol, ∆So
) -1.6 ( 1.2 cal/K/mol; for 4, ∆Go
)
cisftrans
cisftrans
2.49 ( 0.02
2.41 ( 0.02
-2.4 ( 0.1 kcal/mol; ∆Socisftrans ) -0.7 ( 0.5 cal/K/mol. Thus,
like 1 and 2, the conversion of cis f trans amide is enthalpically
favored and entropically disfavored in 3 and 4.
a Errors based on variations observed from multiple integration of 13C
NMR signals in a given spectral data set.
Rate Constants for Amide Cis-Trans Isomerization in
1-4. 13C Saturation-transfer (13C ST) experiments were conducted
on 1-4 to test for the presence of chemical exchange and thus
confirm the assignment of cis and trans 13C NMR signals (Figure
4), and to measure amide cis-trans isomerization (CTI) rate
constants and associated kinetic activation parameters.
Rate constants ktransfcis in 1 and 2 at different temperatures,
obtained from steady-state 13C ST experiments (see NMR
Spectroscopy section), are given in Table 6. At 64.4 °C, ktransfcis
for 1 was 0.15 ( 0.04 s-1; this value increased ∼5-fold to 0.77
( 0.05 s-1 at 84.6 °C. In contrast, ktransfcis for 2 was 0.62 (
0.05 s-1 at 64.4 °C, increasing to 3.59 ( 0.16 s-1 at 84.6 °C.
The rate of conversion of trans f cis amide depends on
anomeric configuration, with conversion more favored in 2.
Energies of activation (Eact) for trans f cis were 19.6 ( 3.7
kcal/mol for 1 and 19.6 ( 4.5 kcal/mol for 2.
Rate constants kcisftrans in 1 and 2 (Table 6) were calculated
from ktransfcis and Ktrans/cis values at each temperature. Energies
of activation (cis f trans) were 18.3 ( 3.9 kcal/mol for 1 and
18.8 ( 4.5 kcal/mol for 2. Amide CTI appears more kinetically
favored in 2 than in 1, on the basis of the larger ktransfcis and
kcisftrans values observed in the former at any temperature.
CTI kinetics measurements were conducted on 3 and 4 using
time-dependent 13C ST in which the C1′ (carbonyl) carbon of
the cis form was saturated for varying times and the effect of
this saturation monitored on the C1′ signal intensity of the trans
of these intense signals in an effort to detect signals arising
from cis forms. Weak signals were detected (Table 4; Figure
4). Splittings of the CdO (C1′) and C2 signals attributed to
putative cis forms were identical, suggesting that the signals
arose from carbons in the same molecule.
Further support for the cis and trans signal assignments in 3
and 4 was obtained from 13C NMR spectra of the triply 13C-
labeled isotopomers 3′ and 4′. In both anomers, the labeled
carbons were mutually coupled, giving rise to three-spin systems
containing internally consistent J-couplings; these results are
shown for 4′ in Figure S2 in Supporting Information.
Although the cis signals were weak, the use of 13C-labeled
compounds allowed determinations of the percentages of both
cis and trans amides of 3 and 4 at different solution temperatures
by integration of either the labeled C1′ (CdO) or C2 carbon
signals (Table 5; only data for C2 are shown). In 3, the cis amide
varied from 1.4% to 2.5% as temperature increased from 42.0
to 84.6 °C (Table 5). In contrast, the cis amide varied from
1.9% to 3.1% in 4 over a temperature range of 31.3-75.1 °C.
At 52.7 °C, the percentage of cis form was greater in 4 (2.6%)
than in 3 (1.6%), showing a dependence of Ktrans/cis on anomeric
configuration similar to that found for 1 and 2 (Table 2). These
percentages of cis amide, however, are roughly 10-fold lower
than found for 1 and 2 at similar solution temperatures.
Table 3. 13C-13C Spin-Coupling Constantsa in 1-4
spin-coupling constanta
b
c
cmpd
2JC2,C1′
3JC1,C1′
3JC3,C1′
1JC2,C3
2JC2,C4
1JC1′,CH (acetyl)
3JC2,CH (acetyl)
1JC1,C2
3JC2,OCH (aglycone)
3
3
3
1/ꢀ-cis
1/ꢀ-trans
2/ꢀ-cis
2/ꢀ-trans
3/ꢀ-cis
3/ꢀ-trans
4/ꢀ-cis
4/ꢀ-trans
2.8
1.1
2.5
1.4
1.1
0.8
0.8
1.0
2.1
1.1
1.7
1.1
1.6d
1.1
1.4d
1.3
1.0
1.3
1.4
1.1
44.7d
45.0
1.5
1.1
36.9
37.2
+3.0
+2.3
50.3
50.4
1.6
1.8
2.8
3.1
44.8d
45.0
a In Hz ( 0.1 Hz at 22 °C in 2H2O. b Sign unknown. c Predicted signs based on analogous JC2,C4 in ꢀ-D-glucopyranosyl rings (ref 36). d Couplings
measured from 13C{1H} NMR spectra of 3′ and 4′.
2
Table 4. 1H and 13C Chemical Shift Assignmentsa for 3 and 4 in Their Cis and Trans Amide Conformations
cmpd
δH1
δH2
δH3
δH4
δH5
δH6a
δH6b
δOCH3
δ(CO)CH3
3/ꢀ-trans
4/ꢀ-trans
4.711
4.395
3.866
3.641
3.665
3.481
3.429
3.33-3.43
3.626
3.33-3.43
3.831
3.888
3.734
3.700
3.336
3.457
1.988
1.987
cmpd
δC1
δC2
δC3
δC4
δC5
δC6
δOCH3
δCdO
δ(CO)CH3
3/ꢀ-cis
101.07b
100.66
104.87b
104.49
60.27b
56.21
62.58b
58.00
179.37b
177.04
180.32b
177.25
3/ꢀ-trans
4/ꢀ-cis
73.71
76.51
72.56
72.48
74.25
78.44
63.15
63.29
57.72
59.62
24.43
24.69
4/ꢀ-trans
a
1H chemical shifts in ppm relative to external dimethyl-2-silapentane-5-sulfonic acid (DSS); 13C chemical shifts in ppm relative to external DSS.
Measured at 22 °C in H2O. b Measured from the 13C{1H} NMR spectra of 3′ and 4′.
2
9
4648 J. AM. CHEM. SOC. VOL. 132, NO. 13, 2010

Amide Cis-Trans Isomerization
A R T I C L E S
2
Figure 4. Partial 13C{1H} NMR spectra (150 MHz) of 4 in H2O at 22 °C. (A) Signals from the 13C-labeled carbonyl (C1′) carbon from cis (180.32 ppm)
and trans (177.25 ppm) forms, with insets showing splittings due to 2JC2,C1′. (B, C) Signals from the 13C-labeled C2 carbon in the cis (62.58 ppm) and trans
2
(58.00 ppm) forms, showing JC2,C1′ values internally consistent with values measured from the C1′ signals in A.
Table 5. Ktrans/cis for 3 and 4 at Different Solution Temperatures
Rate constants, kcisftrans, in 3 and 4, calculated from ktransfcis
and Ktrans/cis values at each temperature, are also shown in Table
7. Energies of activation (cis f trans) were 16.0 ( 3.2 kcal/
mol for 3 and 16.7 ( 3.4 kcal/mol for 4, values statistically
similar to those observed for 1 and 2. Like 1/2, amide CTI
appears more kinetically favored in ꢀ-anomer 4 than in
R-anomer 3.
T (°C)
Ktrans/cis (3; R)a
Ktrans/cis (4; ꢀ)
31.3
42.0
52.7
64.4
75.1
84.6
53.0 ( 1.5
45.3 ( 1.1
38.1 ( 0.8
33.8 ( 0.6
31.1 ( 0.5
72.5 ( 2.8
60.3 ( 1.9
51.1 ( 1.4
45.3 ( 1.1
38.8 ( 0.8
Additional activation parameters for the CTI of 1-4 are
q
a Values were determined from the integration of the C2 signals.
Errors are based on variations observed from multiple integration of C2
signals in a given spectral data set.
summarized in Table 8. ∆G° values range from 17 to 21 kcal/
q
mol and ∆H° from 15 to 19 kcal/mol, and the signs of both
q
parameters are positive. In contrast, ∆S° values are uniformly
q
negative, ranging from -3 to -9 cal/mol/K. ∆S° values appear
Table 6. Rate Constants (s-1), ktransfcis and kcisftrans, for 1 and 2 at
Different Temperatures
more negative for R-anomers 1 and 3 than for ꢀ-anomers 2 and
4, but the large errors make this observation inconclusive.
1 (R)
2 (ꢀ)
Conclusions
T (°C)
ktransfcis
kcisftrans
ktransfcis
kcisftrans
42.0
52.7
64.4
69.7
75.1
84.6
0.09 ( 0.05
0.22 ( 0.04
0.62 ( 0.05
0.24 ( 0.14
0.60 ( 0.12
1.61 ( 0.14
N-Acetylation in biological systems is a common chemical
modification that has important functional ramifications. For
example, N-acetylation of L-lysine side chains of histones serves
as a mechanism to reduce positive charge and thereby reduce
electrostatic attraction between histones and their negatively
charged nucleic acid binding partners.37 This reduced affinity plays
an active role in modulating gene expression.38 In a similar vein,
N-acetylation of 2-deoxy-2-aminosugars such as D-glucosamine 6P
eliminates their charge properties and thus affects their biological
behaviors and functions. For example, D-galactosamine is toxic in
animals and thus is normally found in its N-acetylated form in ViVo
(found as UDP-GalNAc produced via C4-epimerization of UDP-
GlcNAc).39 In contrast, D-glucosamine 6P is synthesized in ViVo
from D-fructose 6P by the bifunctional enzyme, glutamine:fructose
6P amidotransferase (GFA1), and can accumulate without toxic
effects. In ViVo conversion of D-glucosamine 6P to N-acetyl-D-
glucosamine 6P is catalyzed by the CoA-dependent enzyme,
D-glucosamine 6P N-acetyltransferase (GNA1).40
0.15 ( 0.04
0.22 ( 0.04
0.36 ( 0.04
0.77 ( 0.05
0.57 ( 0.14
0.79 ( 0.14
1.27 ( 0.14
2.59 ( 0.19
1.57 ( 0.08
3.59 ( 0.16
3.92 ( 0.23
8.66 ( 0.43
form (Figure 5A). Data linearization at different solution
temperatures (Figure 5B) gave ktransfcis (Table 7). For 3, ktransfcis
ranged from 0.02 to 0.9 s-1 over a temperature range of
42.0-84.6 °C, whereas for 4, ktransfcis ranged from 0.05 to 2.6
s-1 over a temperature range of 31.3-75.0 °C. As found for
1/2, ktransfcis in 3/4 depends on anomeric (C1) configuration,
with trans f cis conversion more favored in ꢀ-anomer 4 at all
temperatures investigated. Activation energies (Eact) (trans f
cis) (Figure S3 in Supporting Information) were 19.2 ( 2.8 kcal/
mol for 3 and 19.4 ( 3.0 kcal/mol for 4, values similar to those
found for 1 and 2.
9
J. AM. CHEM. SOC. VOL. 132, NO. 13, 2010 4649

A R T I C L E S
Hu et al.
Figure 5. (A) Plots of C1′trans signal intensity in 3 as a function of saturation time of the C1′cis signal. Data were collected at different temperatures: black
) 42.0 °C; blue ) 52.7 °C; green ) 64.4 °C; red ) 75.1 °C; orange ) 84.6 °C. (B) Linearization of the data shown in A following eq 1, from which ktransfcis
values were determined.
Table 7. Rate Constants (s-1), ktransfcis and kcisftrans, for 3 and 4 at
Different Temperatures
To date, little attention has been paid to amide cis-trans
equilibria and kinetics in N-acylated saccharides, specifically
3 (R)
4 (ꢀ)
in GlcNAc, GalNAc, and N-acetyl-neuraminic acid (Neu5Ac).
It is normally assumed that, in solution, the trans amide
predominates in these and related structures, and the potential
presence of cis isomers and their role in biology have been
overlooked. The latter is especially noteworthy in light of the
distribution of cis and trans GlcNAc residues in crystal
structures of glycoproteins,42 where a broad and nearly continu-
ous span of rotamers about the N-CO bond is observed despite
the relatively high energies required to rotate the partial C-N
double-bond into nonplanar conformations (i.e., geometries other
than the idealized cis and trans forms) (Figure 6A). This energy
likely derives from crystal packing forces, highlighting the high
degree of structural deformation accessible in crystal structures
of conformationally flexible molecules or fragments thereof
under certain packing conditions. Additional analysis of the same
data set with respect to the more easily rotatable C2-N2 bond
(R) shows a preference for R-anti (i.e., H2 anti to NH) (Figure
6B), a result consistent with prior interpretations of 3JH2,NH spin-
couplings in GlcNAc and related structures measured in
solution.35 Correlation between C2-N2 (R) and amide (ꢀ) bond
conformation (Figure 6C) shows a clustering of data points near
180° for both torsions, as expected, but also a discernible
horizontal clustering of points attributable to the cis amide over
a wide range of R conformations. The distribution of torsions
is more uniform for R than for ꢀ, presumably as a result of the
partial double bond character of the latter that destabilizes
torsions between 10° to 60° and -10° to -60°.
T (°C)
ktransfcis
kcisftrans
ktransfcis
kcisftrans
31.3
42.0
52.7
64.4
75.1
84.6
0.05 ( 0.02
0.13 ( 0.02
0.36 ( 0.01
1.09 ( 0.01
2.56 ( 0.01
2.58 ( 1.34
5.70 ( 0.91
13.7 ( 0.7
37.0 ( 0.9
79.7 ( 1.4
0.02 ( 0.01
0.06 ( 0.01
0.16 ( 0.01
0.40 ( 0.01
0.87 ( 0.02
1.65 ( 0.74
3.38 ( 0.61
8.13 ( 0.68
17.9 ( 1.0
33.8 ( 1.5
Table 8. Activation Parameters for Amide Cis-Trans
Interconversion in 1-4
activation parameter (trans f cis)
q
q
q
cmpd
In A
∆G° (kcal/mol)
∆H° (kcal/mol)
∆S° (cal/K/mol)
R-NFG 1 27.3 ( 5.3 20.9 ( 0.5
ꢀ-NFG 2 28.8 ( 6.4 20.0 ( 0.7
R-NAG 3 26.9 ( 4.0 20.7 ( 0.4
ꢀ-NAG 4 28.9 ( 4.4 19.7 ( 0.4
18.9 ( 3.7
18.9 ( 4.5
18.5 ( 2.9
18.7 ( 3.7
-6.5 ( 10.4
-3.5 ( 12.7
-7.4 ( 7.9
-3.2 ( 10.8
activation parameter (cis f trans)
q
q
q
cmpd
In A
∆G°
∆H°
∆S°
R-NFG 1
ꢀ-NFG 2
R-NAG 3
ꢀ-NAG 4
26.7 ( 5.5
28.5 ( 6.5
26.0 ( 4.6
28.6 ( 5.0
19.9 ( 0.6
19.3 ( 0.7
18.0 ( 0.5
17.3 ( 0.4
17.6 ( 3.9
18.1 ( 4.5
15.4 ( 3.1
16.1 ( 3.4
-7.7 ( 10.9
-4.1 ( 12.9
-9.0 ( 9.0
-3.9 ( 9.9
Although N-acetylation is a ubiquitous saccharide modi-
fication in biological systems, the chemical and biochemical
properties conferred by this structural change are not fully
appreciated. It is well established that generic amide bonds
assume both cis and trans conformations in solution, and
considerable study has been directed toward understanding
the thermodynamics and kinetics of amide CTI in simple
systems. For example, free energy differences between cis
and trans formamides and acetamides have been reported to
range from 0.6 to 9.3 kcal/mol, and energy barriers to amide
bond rotation in simple amides range from 18 to 30 kcal/
mol depending on the nature of the substituents appended to
the carbonyl carbon and amide nitrogen.41
Prior studies have shown that aqueous solutions of N-
formylated aminosugars contain comparable amounts of cis and
trans amides.29–32 The small N-formyl hydrogen apparently
stabilizes the cis form (relative to methyl in the N-acetyl
derivative) by minimizing steric interactions with the proximal
(40) Hurtado-Guerrero, R.; Raimi, O. G.; Min, J.; Zeng, H.; Vallius, L.;
Shepherd, S.; Ibrahim, A. F. M.; Wu, H.; Plotnikov, A. N.; van Aalten,
D. M. F. Biochem. J. 2008, 415, 217–223.
(41) Isaacs, N. Physical Organic Chemistry, 2nd ed.; Longman Scientific
and Technical: Harlow, England, 1995; p 350.
(42) This survey was conducted on 1311 pdb files (from the RCSB PDB
database) containing a total of 7994 GlcNAc residues. Within the latter
group, 917 residues contained a H2-C2-N2-H torsion angle between
60° and-60°, producing an ∼11% R-syn population; the remaining
∼89% is R-trans. More pertinent to the present work, 463 residues
contained a C2-N2-C1′-C2′ torsion angle between 60° and-60°,
giving an ∼6% ꢀ-cis population.
(37) Legube, G.; Trouche, D. EMBO Rep. 2003, 4, 944–847.
(38) Verdone, L.; Agricola, E.; Caserta, M.; Di Mauro, E. Briefings Funct.
Genomics Proteomics 2006, 5, 209–221.
(39) McMillan, J. M.; McMillan, D. C. Toxicology 2006, 222, 175–184.
9
4650 J. AM. CHEM. SOC. VOL. 132, NO. 13, 2010

Amide Cis-Trans Isomerization
A R T I C L E S
additional support for cis and trans signal assignments in 3/4
was obtained from studies of internally consistent 13C-13C spin-
couplings in 13C-labeled GlcNAc isotopomers and by the
demonstration and quantification of cis-trans chemical ex-
change by 13C saturation transfer.
The equilibrium between cis and trans amides in the N-formyl
and N-acetyl derivatives of methyl D-glucosaminide depends
on anomeric configuration. The cis amide is less stable in
R-anomers of NFG and NAG methyl glycosides than in the
corresponding ꢀ-anomers. Given the proximity of the N-acyl
side chain to the anomeric center in these structures, this
dependency might be expected. Inspection of molecular models,
however, does not provide a structural rationale for the observed
dependency if preferred conformation about R in the cis amides
3
is anti (as observed in the trans amides based on JH2,NH
values35). On the other hand, if R is syn in cis amides, or partly
so, in solution, then the acyl side chain and the axial C1-O1
bond would be in close proximity in R-anomers, thereby
producing a destabilizing steric interaction.45 This problem and
the effects of acyl chain structure and length on the anomeric
dependency of Ktrans/cis will require further investigation.
The relatively small rate constants for saccharide cis-trans
isomerization call attention to the potential implications of this
process in biological receptor binding and enzyme catalysis. In
hypothetical cases in which the cis amide of GlcNAc is the
productive form for binding, the overall rate of binding could
be determined by the rate of conversion of the trans to cis form.
Whether such a scenario occurs in ViVo and whether biological
systems have evolved enzymes to facilitate CTI in saccharides
in order to circumvent this problem demand further scrutiny.
Figure 6. (A) Distribution of C2-N2-C1′-C2′ dihedral angles observed
in a survey of crystal structures of GlcNAc residues in glycoproteins taken
from the PDB.42 The trans amide predominates (180°), but a continuous
distribution of angles is observed over the full 360° rotation, including that
for the cis amide (0°). (B) Distribution of R dihedral angles in the same
ensemble of X-ray structures as in A. (C) Correlation between conformations
of the R and ꢀ bonds in the same group of GlcNAc structures sampled in
A and B.
Given the importance of anomeric configuration on amide
cis-trans equilibria in 2-deoxy-2-formamido and 2-deoxy-2-
acetamido sugars, it is natural to ponder the effects of other
structural perturbations on the cis-trans equilibria and CTI
kinetics in saccharides. For example, in structures like methyl
ꢀ-chitobioside 9 (and the chitin polymer derived from it) and
hyaluronic acid 10, the multiple N-acetyl groups may display
different cis-trans equilibria and exchange kinetics due to their
different chemical environments and/or states of solvation.
Having established the fundamental NMR characteristics of cis
and trans amides in simple GlcNAc glycosides, and the utility
of selective 13C labeling to enable such studies, it is now feasible
to investigate N-acetyl side chain behavior in more complex
C2-H2 bond.43 In the present work, NMR characterization of
the N-formyl glycosides was a prerequisite to the more chal-
lenging detection and quantification of cis amide in solutions
of GlcNAc glycosides. The success of the latter work hinged
on the assumption that the effects of amide conformation on
1H and 13C chemical shifts for nuclei in the vicinity of the amide
bond were similar in the N-formyl and N-acetyl derivatives, thus
allowing signal assignments in the latter despite the low
abundance of cis amide. While this assumption proved valid,
(45) Preliminary data on the behavior of R in the cis and trans forms of 2
was obtained by measuring 3JH2,NH values in DMSO-d6 solvent (Figure
S4 in Supporting Information). For 2, nearly identical coupling
magnitudes were observed in both amide conformations (9.6 Hz in
ꢀ-cis; 9.5 Hz in ꢀ-trans), suggesting similar bond conformations about
bond R. On the other hand, the effect of amide conformation on R is
probably smaller for N-formyl than for N-acetyl substituents. At
present, similar J-coupling measurements have not been attempted in
MeGlcNAc anomers because of the considerably less intense cis
signals.
(43) Eliel and Wilen have suggested that the preference for trans amide
conformations in solution is due to a combination of steric (attractive
in trans, repulsive in cis) and charge interaction factors. See ref 44
for further discussion.
(44) Eliel, E. L.; Wilen, S. H. Stereochemistry of Organic Compounds;
Wiley: New York: 1994; pp 620-621.
9
J. AM. CHEM. SOC. VOL. 132, NO. 13, 2010 4651

A R T I C L E S
Hu et al.
of 2-amino-2-deoxy-D-[2-13C]glucose hydrochloride and 2-amino-
2-deoxy-D-[1,2-13C2]glucose hydrochloride.
structures. Likewise, studies of N-acetyl side chain behavior in
Neu5Ac, where the side chain resides in a different chemical
environment than exists in GlcNAc, may reveal new structural
properties that contribute to its biological properties. It is
noteworthy that Neu5Ac is commonly found as a terminal (and
thus exposed) residue in the oligosaccharide chains of N-linked
glycoproteins, thus increasing the likelihood that, in some
instances, recognition between Neu5Ac residues and biorecep-
tors may depend on side chain amide conformation.
Supporting Information Available: Routes for the synthesis
of NFG glycosides; DFT-calculated total energies of 5-8 (in
1
Vacuo and solvated); H-1H spin-coupling constants in 1 and
2; van’t Hoff plots for 1 and 2; partial 13C{1H} NMR spectra
of 4′; Arrhenius plot of ktransfcis in 4; partial 1H NMR spectrum
of 2 in DMSO-d6; description of the synthesis of 2; description
of the synthesis of 3′ and 4′; description of calculations of kinetic
Acknowledgment. This work was supported by a grant (to A.S.)
from the National Institutes of Health (GM059239). The Notre
Dame Radiation Laboratory is supported by the Office of Basic
Energy Sciences of the United States Department of Energy. This
is Document No. NDRL-4815 from the Notre Dame Radiation
Laboratory. The authors thank Omicron Biochemicals, Inc. for gifts
q
q
q
activation parameters Eact, ∆G° , ∆H° , and ∆S° ; full ref 23.
This material is available free of charge via the Internet at http://
pubs.acs.org.
JA9086787
9
4652 J. AM. CHEM. SOC. VOL. 132, NO. 13, 2010