NMR-based conformational analysis of sphingomyelin in bicelles

Article abstract of DOI:10.1016/j.bmc.2011.11.001

Sphingomyelin (SM) is a common sphingolipid in mammalian membranes and is known to be substantially involved in cellular events such as the formation of lipid rafts. Despite its biological significance, conformation of SM in a membrane environment remains

Full text of DOI:10.1016/j.bmc.2011.11.001

Bioorganic & Medicinal Chemistry 20 (2012) 270–278
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
NMR-based conformational analysis of sphingomyelin in bicelles
Toshiyuki Yamaguchi a,b, Takashi Suzuki a, Tomokazu Yasuda a, Tohru Oishi a,c, Nobuaki Matsumori a,
,
⇑
Michio Murata a,b,
⇑
a Department of Chemistry, Graduate School of Science, Osaka University, Toyonaka, Osaka 560-0043, Japan
b JST, ERATO, Lipid Active Structure Project, Toyonaka, Osaka 560-0043, Japan
c Department of Chemistry, Graduate School of Sciences, Kyushu University, Higashi-ku, Fukuoka 812-8581, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Sphingomyelin (SM) is a common sphingolipid in mammalian membranes and is known to be substan-
tially involved in cellular events such as the formation of lipid rafts. Despite its biological significance,
conformation of SM in a membrane environment remains unclear because the noncrystalline property
and anisotropic environment of lipid bilayers hampers the application of X-ray crystallography and
NMR measurements. In this study, to elucidate the conformation of SM in membranes, we utilized
bicelles as a substitute for a lipid bilayer membrane. First, we demonstrated through 31P NMR, 2H
NMR, and dynamic light scattering experiments that SM forms both oriented and isotropic bicelles by
changing the ratio of SM/dihexanoyl phosphatidylcholine. Then, we determined the conformation of
SM in isotropic bicelles on the basis of coupling constants and NOE correlations in 1H NMR and found that
the C2–C6 and amide groups of SM take a relatively rigid conformation in bicelles.
Received 14 September 2011
Revised 31 October 2011
Accepted 1 November 2011
Available online 7 November 2011
Keywords:
Sphingomyelin
Bicelle
NMR spectroscopy
Lipid rafts
Ó 2011 Elsevier Ltd. All rights reserved.
Conformation
1. Introduction
constants by NMR experiments.7,8 To mimic the membrane envi-
ronment, micellar media such as sodium dodecyl sulfate are often
Until the late 1990s, plasma membranes were considered to be a
two-dimensional fluid where lipids and proteins diffuse freely.
During the last decade, their heterogeneity and complexity have
been widely recognized; for example, it has been found that sphin-
gomyelin (SM), the most abundant sphingolipid in biomembranes,
forms microdomains called lipid rafts on a plasma membrane in
the presence of cholesterol.1 Lipid rafts are believed to play a key role
in numerous cellular processes such as signal transduction, protein
sorting, and cholesterol shuttling.1,2 In addition, they are recognized
as potential sites for toxin interactions, entryways for pathogens,3
and fusion sites for HIV.4–6 Despite such biological/pathological
significance, the manner in which raft-forming lipids self-assemble
to form a nanometer-scale organization is still unknown, largely
because of the rapid association and dissociation equilibrium of lipid
molecules in rafts. To gain a fundamental understanding of molecu-
lar interactions in lipid rafts, exploration of the structures and
dynamics of SM in membrane environments is crucial.
used; however, as a membrane alternative, micelles are controver-
sial because of their small radius, high curvature, and lack of lamel-
lar structures. Recently, bicelles has been found to better
reproduce a natural membrane environment since they possess
true lipid bilayer portions.9,10 Bicelles are generally comprised of
a long-chain phospholipid such as dimyristoyl phosphatidylcholine
OH
γ
O
O
O
O
β
4
2
Me3N
P
α
1
3
5 C13H27
15H31
HN
C
3'
1'
2'
O
D-erythro-N-stearoyl-sphingomyelin
(SSM)
O
So far, however, the conformation of SM in membrane environ-
ments has not been elucidated, although its conformation in organ-
ic solvents has been extensively analyzed on the basis of coupling
O
O
O
O
P
Me3N
O
C5H11
O
C5H11
O
⇑
Corresponding authors. Address: Department of Chemistry, Graduate School of
1,2-dihexanoyl-3-sn-phosphatidylcholine
Science, Osaka University, Toyonaka, Osaka 560-0043, Japan. Tel.: +81 6 6850 5790;
fax: +81 6 6850 5789 (N.M); tel./fax: +81 6 6850 5774 (M.M).
E-mail addresses: matsmori@chem.sci.osaka-u.ac.jp (N. Matsumori), murata@
chem.sci.osaka-u.ac.jp (M. Murata).
(DHPC)
Figure 1. Structures of stearoylsphingomyelin (SSM) and dihexanoylphosphatidyl-
choline (DHPC).
0968-0896/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2011.11.001

T. Yamaguchi et al. / Bioorg. Med. Chem. 20 (2012) 270–278
271
Figure 2. 31P NMR spectra of bSM/DHPC (4:1) bicelles at 35, 37, and 40 °C. The total lipid concentration was 500 mM (A) and 100 mM (B).
(DMPC), and a short-chain one such as dihexanoyl phosphatidyl-
choline (DHPC, Fig. 1). The size and shape of bicelles can be con-
trolled by the ratio of long-chain phosphatidylcholine (PC) to
short-chain PC, that is, q. At higher q values, bicelles adopt a mag-
netically aligned lamellar bilayer morphology. In contrast, when
the q values are decreased, bicelles become disk-shaped. In this
configuration, the planar region consisting of long-chain lipids is
surrounded by a rim comprised of short-chain lipids. Bicelles with
the q value less than 1 undergo fast tumbling and can be regarded
as isotropic in aqueous solutions, thus allowing high-resolution 1H
NMR measurements. In fact, small bicelles have often been used for
structural determinations of small integral peptides and
membrane-bound agents.11–13
changing the ratio of SM/DHPC. Then, the conformation of SM in
isotropic bicelles was determined on the basis of vicinal 1H–1H
coupling constants and NOEs. Although some NOESY cross peaks
were completely overlapped, we successfully prepared
a
site-selectively deuterated SM and made the NOE assignment for
the central portion of SM (C1–C5).
2. Results and discussion
2.1. Formation of oriented bicelles containing SM
Before dealing with the conformation of SM in bicelles, we
examined whether SM forms bicelles with DHPC, as is the case
with DMPC. 31P NMR is generally used for characterizing the
orientation and morphology of bicelles.14 In oriented bicelles with
In this study, we applied the bicelle technique to explain the
structure of SM in a membrane environment. Since there have
been no reports so far on bicelle formation with SM, we first dem-
onstrated that SM forms both oriented and isotropic bicelles by
the normal perpendicular to the direction of the magnetic field, 31
P
NMR spectra show two signals. The high-field resonance is

272
T. Yamaguchi et al. / Bioorg. Med. Chem. 20 (2012) 270–278
Fig. 3) and measured 2H NMR. Details of the preparation of
100-d2-SSM are given in the Supplementary. Fig. 3 depicts the 2H
NMR spectrum of 100-d2-SSM/DHPC bicelles (q = 4) at the total lipid
concentration of 500 mM and shows a well-resolved doublet with-
out any center peak. It is known that lipid molecules residing in the
bilayer portion of oriented bicelles result in sharp doublet signals
in the 2H NMR spectrum, while those in the rim portion of oriented
bicelles show peaks near the spectral center because of the isotro-
pic nature of the rim portion.16 Fig. 3 clearly demonstrates not only
that SM/DHPC bicelles are magnetically aligned but also that SM
molecules reside exclusively in the bilayer portions of bicelles.
Although 2H splitting of D2O residing in hydration shell of oriented
bicelles was sometimes used for evaluating the formation of
oriented bicelles,17 2H splitting of 100-d2-SSM unequivocally
demonstrates the oriented bicelle formation because it resides in
the bicelles.
2
Figure 3. H NMR spectrum of SSM/100-d2-SSM/DHPC (3:1) bicelle (500 mM) at
37 °C.
2
We also measured H NMR spectra of SSM/100-d2-SSM/DHPC
(3:1) bicelles at various temperatures (Fig. S9, Supplementary),
which unexpectedly suggest that aligned bicelles occur from 30
to 50 °C. This is different from Fig. 2A that indicates that the ori-
ented bicelles forms only at around 37 °C. This discrepancy is prob-
ably caused by the difference of SM used; in Fig. 2 we used brain
SM that is a mixture of SMs with various acyl-chain lengths as de-
scribed earlier, but we used pure SSM in the 2H NMR experiments.
This suggests that oriented bicelles formed by pure SSM is more
temperature-tolerant than those by brain SM. Based on this find-
ing, we will try to prepare more stable oriented bicelles using pure
SM, and the results will be reported in due course.
Next, to estimate the size of bSM/DHPC (4:1) bicelles, we car-
ried out dynamic light scattering experiments by varying temper-
ature and concentration (Fig. 4). All preparations basically showed
a single main peak, thus indicating that SM and DHPC form aggre-
gates of relatively uniform size; that is, coexistence of different-
sized aggregates such as micelles and vesicles does not occur. For
bicelles with q = 4 at 500 mM of the total lipid concentration, the
bicelle size increased with increasing temperature from 35 to
40 °C (Fig. 4A). Taking account of the results of 31P NMR experi-
ments in Fig. 2A, which illustrate that 500 mM bicelles with q = 4
are magnetically aligned only at 37 °C, bicelles around 100 nm in
diameter tend to be oriented, which is comparable to the size re-
ported for aligned DMPC/DHPC bicelles.18 The larger size at 40 °C
(Fig. 4A) suggests the formation of vesicles, which is consistent
with the spectral features of 31P NMR in Fig. 2A. At the low lipid
concentration (100 mM), bicelles with q = 4 become smaller
attributed to long-chain lipids in the planar section, whereas the
downfield resonance is attributed to DHPC on the rim of the
disks.14 Since large quantities of SM are necessary for 31P NMR
experiments, we used bovine brain SM (bSM), which is a standard
SM in biological and biophysical studies and contains about 50%
stearoylsphingomyelin (SSM, Fig. 1).15 We first examined the
orientation of bicelles composed of bSM/DHPC (4:1, q = 4) at
500 mM of the total lipid concentration (bSM + DHPC). As shown
in Figure 2A, 31P NMR spectra showed two signals at 37 °C; high
filed resonance is attributable to the planar part of oriented bi-
celles, and low-field one is due to the rim portion. This implies that
bSM/DHPC bicelles are aligned along the external magnetic field;
On the other hand, a broad isotropic signal observed at 35 °C indi-
cates that magnetically aligned bicelles do not occur, and the spec-
tral feature at 40 °C suggests possible vesicle formation (Fig. 2A).
Bicelles with the same q value (q = 4) at 100 mM of the total lipid
concentration showed no orientation at the tested temperatures
(Fig. 2B). In separate experiments where the total lipid concentra-
tion was varied from 100 mM to 500 mM (Figs. S1–S5, Supplemen-
tary), we observed that bicelles are oriented at concentrations of
more than 200 mM. These data show that alignment of bSM/DHPC
bicelles is significantly influenced by not only temperature but also
total lipid concentration; in particular, bSM/DHPC bicelles with
q = 4 can be oriented in a narrow temperature range at around
37 °C.
To further confirm that SM/DHPC bicelles with q = 4 are mag-
netically oriented, we prepared a deuterated SSM (100-d2-SSM,
Figure 4. Dynamic light scattering results of bSM/DHPC bicelles with q = 4 at 35 °C (dashed line), 37 °C (solid line), and 40 °C (dotted line). Total lipid concentration was
500 mM (A) and 100 mM (B).

T. Yamaguchi et al. / Bioorg. Med. Chem. 20 (2012) 270–278
273
Figure 5. 31P NMR spectra of bSM/DHPC = 1:2 bicelles at 35 °C, 37 °C, and 40 °C. The total lipid concentration was 500 mM (A) and 100 mM (B).
(Fig. 4B), which is also consistent with the isotropic orientation
shown in the 31P NMR experiments. These data indicate that the
size of bicelles with q = 4 considerably depends not only on tem-
perature but also on the concentration of lipids.
present in phosphatidylcholine.19,20 As depicted in Figure 5, 31P
NMR signals of q = 0.5 bicelles are virtually unaffected by temper-
ature (35–40 °C) or lipid concentration (100–500 mM).
Dynamic light scattering experiments also showed that the size
of bSM/DHPC bicelles with q = 0.5 is virtually unaffected by tem-
perature or concentration (Fig. 6), which is consistent with the re-
sults of 31P NMR experiments (Fig. 5). The size of bicelles with
q = 0.5 is around 20 nm in diameter, which is comparable to that
of isotropic DMPC/DHPC bicelles (Fig. S6, Supplementary). These
results suggest that an SM/DHPC system with q = 0.5 forms stable
isotropic bicelles at broad temperature and concentration ranges.
Thus, in the following conformation analysis of SM in isotropic
bicelles, we used SM/DHPC (1:2) bicelles with total lipid concen-
tration of 100 mM and measured NMR spectra at 37 °C.
2.2. Formation of isotropic bicelles containing SM
After finding that SM forms oriented bicelles, we examined the
formation of isotropic bicelles comprised of bSM/DHPC (1:2,
q = 0.5). In the 31P NMR spectra, two sharp signals in close vicinity
(dSM = 2.7 ppm, dDHPC = 2.1 ppm) are clearly observed (Fig. 5), thus
indicating the formation of small and magnetically isotropic bi-
celles. The low-field resonance of the 31P signal of SM is probably
attributable to intramolecular hydrogen bonding, which is not

274
T. Yamaguchi et al. / Bioorg. Med. Chem. 20 (2012) 270–278
Figure 6. Dynamic light scattering of bSM/DHPC bicelles with q = 0.5 at 35 °C (dashed line), 37 °C (solid line), and 40 °C (dotted line). Total lipid concentration was 500 mM
(A) and 100 mM (B).
Table 1
1H chemical shifts (ppm) of SSM in SSM/DHPC (1:2) bicelles at 37 °C
Position
da
Position
da
NH
H5
H4
7.83
5.72
5.41
4.28
4.02
4.02
3.90
Hb
3.66
3.20
2.21
1.98
1.60
1.22
0.80
H
c
H2’
H6
H3’
CH2
H18,18’
Ha
H1
H3
H2
a
Chemical shifts were referenced to DSS.
Table 2
3JH/H values of SSM in bicelles and in methanol
3JH2/H3 (Hz)
3JH3/H4 (Hz)
3JH2/NH (Hz)
Figure 7. Five hundred megahertz 1H NMR spectrum of SSM/DHPC (1:2) bicelles at
37 °C in H2O/D2O (9:1). Total lipid concentration was 100 mM.
In SSM/DHPC (1/2) bicelles
In CD3OD
10.7
8.4
7.7
7.6
9.1
—
a
a
Coupling constant was not observed because of exchanging N–H with N–D.
2.3. Measurements of 1H NMR of SSM-DHPC isotropic bicelles
For the following 1H NMR measurements, we used SSM (Fig. 1)
that was purified by HPLC from bovine bSM. The 1H NMR spectrum
shown in Fig. 7 was measured with SSM/DHPC bicelles (1:2,
100 mM of the total lipid concentration) in H2O/D2O (9:1, v/v) at
37 °C. Although DHPC gave much stronger signals, resonances from
SSM are relatively well resolved and successfully assigned (Table 1).
One-dimensional and E.COSY spectra were recorded to determine
relevant 3JH/H values (Table 2). A large 3JH-2/H-3 value indicates that
the C2–C3 bond largely takes an anti-conformation.21 As compared
the cross peak because of the complete overlapping of H-1 and
H-3 signals. Because the unambiguous assignment of the NOE is
necessary to determine the conformation around the amide por-
tion, we synthesized 3-d-SSM to deplete the overlapping H-3 sig-
nal. The synthesis of 3-d-SSM is shown in Scheme 1, which is a
slight modification of a previous report.23 In short, starting from
L-serine, deuterium was introduced by the reduction of ketone
with LiAlD4, and then olefin metathesis by the Grubbs catalyst, fol-
lowed by the introduction of a choline group and an acyl chain, fur-
nishing 3-d-SSM with 100% deuterium labeling rate. The NOESY
measurement in isotropic bicelles containing 3-d-SSM showed that
the cross peak intensity of concern was significantly weakened
(Fig. 8C); the evaluation of the cross-peak intensity showed that
the cross-peak volume in Fig. 8C is reduced to 49% of that in
Fig. 8B. This means that the correlations for amide-H/H-1 and
amide-H/H-3 contribute equally to the cross-peak in Fig. 8B.
3
3
to the JH,H data in CD3OD (Table 2), the JH-2/H-3 value in bicelles
was shown to be larger than that in CD3OD, indicating that anti-
conformation is more predominant for the C2–C3 bond in bicelles
than in CD3OD. Similarly, judging from a reported Karplus relation
3
3
for JHN/C of peptides,22 the observed JH2/NH (9.1 Hz) implies a
a
H
major anti-rotamer for the C2–N bond. Rotamer around the
C3–C4 bond will be discussed later.
To obtain H–H distance information of SSM, NOESY experi-
ments were carried out for SSM/DHPC (1:2) bicelles (Fig. 8A). We
successfully observed conformationally relevant NOEs such as
H-3/H-5 and H-2/H-4 (Fig. 8A). The amide proton (7.83 ppm) also
gave an NOE cross peak at around 4 ppm corresponding to H-1
and H-3 resonances (Fig. 8B); however, it was difficult to assign
2.4. Conformation of SSM in bicelles
3
As described above, the large JH-2/H-3 (10.7 Hz, Table 2) in
bicelles indicates the anti-conformation for the C2–C3 bond.21
Likewise, judging from the empirical Karplus relation between

T. Yamaguchi et al. / Bioorg. Med. Chem. 20 (2012) 270–278
275
Figure 8. NOESY spectrum of SSM/DHPC (1:2) bicelles at 37 °C in H2O/D2O (9:1) (A) and its expanded spectrum (B). Expanded NOESY spectrum of 3-d-SSM/DHPC (1:2)
bicelles (C). The total lipid concentration was 100 mM in all spectra. The cross peak in the dashed circle in spectrum (B) is markedly weakened in spectrum (C), indicating that
the cross peak is mostly attributable to the NOE arising from the amide and H3.
amide protons and C
a
protons in peptides,22 the observed JNH/H-2
B3LYP/6-31G(d,p) basis set.27
A truncated structure of SM
(9.1 Hz, Table 2) also implies that H-2/NH takes an anti-
orientation.
(Fig. S8, Supplementary) was used for the calculation, and the
C3–C4 bond was rotated through 360° in 30° increments. At each
point, geometry optimization was performed with the rotation of
C3–C4 bond fixed and then the coupling constant was calculated.
The Karplus relation (Fig. S8, Supplementary) was derived by fit-
ting the resultant coupling constants to a general Karplus equation.
On the other hand, to our knowledge, Karplus relation for
C(sp2)–C(sp3) bonds has not been reported so far, probably because
of the scarce occurrence of such bonds in peptides and proteins.
Although the parameters required to define a Karplus relation are
traditionally determined using coupling constants of molecules
with known geometries,24 it is now feasible to calculate coupling
constants using quantum-chemistry methods for any arbitrary
geometry. Recently, it has been shown that the density function
theory can calculate scalar couplings with reasonable accu-
racy25,26; hence, to derive a Karplus relation of the C3–C4 bond
of SSM, we calculated scalar coupling constants of the bond by
3
The observed JH-3/H-4 value for the C3–C4 bond (7.7 Hz, dashed
line in Fig. S8, Supplementary) revealed that anti-orientation for
H-3/H-4 was predominant, which was further supported by the
NOE observed for H-3/H-5.
We then constructed a possible conformation of SM partial
structure (C1–C6 and amide portions) by considering the dihe-
dral angles as well as the NOEs obtained for NH/H-3, H-2/H-4,

276
T. Yamaguchi et al. / Bioorg. Med. Chem. 20 (2012) 270–278
O
successfully elucidated the partial conformation of SM in a mem-
brane environment on the basis of 1H–1H spin coupling constants
and NOE data. To achieve this, it was necessary to synthesize
3-d-SSM to discriminate the overlapping signals in the NOESY
spectrum as well as to establish the calculated Karplus equation
for the C(sp2)–C(sp3) bond to obtain dihedral information for the
C3–C4 bond of SM.
As described earlier, lipid rafts have attracted much attention
because of their crucial roles in biological processes. However,
the mechanism of raft formation, in particular the origin of molec-
ular interactions in lipid rafts, remains elusive. The conformation of
SM in a membrane environment is, therefore, indispensable for
understanding the mode of molecular interactions that occur in li-
pid rafts and will provide a valuable insight into the mechanism of
raft formation. For instance, one may notice that this conforma-
tional feature of SM facilitates intermolecular hydrogen bonds be-
tween SM amide groups, which may lead to stronger molecular
interaction in SM membranes than PC membranes. NMR studies
in this direction are currently underway in our group. In addition,
with the bicelles comprised of SM available, it would become fea-
sible to evaluate the influences of cholesterol on an SM-based
membrane.
O
O
d
a-c
D
OMe
TBSO
HO
OH
TBSO
BocHN
N
NHBoc
NH2
Me
D
OH
OH
g-j
f
e
TBSO
OTBS
TBSO
C13H27
NHBoc
NHBoc
D
D
OR
O
O
O
O
k,l
P
Me3N
OH
HO
C13H27
C13H27
NHBoc
NHBoc
R=H and TBS
D
O
O
O
O
m-o
Me3N
P
C13H27
C17H35
HN
3-d-SSM
O
Scheme 1. Synthesis of 3-d-SSM. Reagents and Conditions: (a) (Boc)2O, NaOH, H2O,
dioxane, rt,6 h; (b) NH(Me)OMeꢁHCl, NMM, EDC, CH2Cl2, ꢂ15 °C, 2 h; (c) TBSCl,
imidazole, DMF, rt, 1 h (74% for three steps); (d) vinyl MgBr, THF, rt, 1.5 h (91%); (e)
LiAlD4, THF, rt, 40 min (15%); (f) 1-pentadecene, Grubbs II catalyst, p-quinone,
CH2Cl2, 55 °C, 1.5 h (60%); (g)TBAF, THF, rt, 25 min (96%); (h) PivCl, pyridine, ꢂ30 °C,
2.5 h (92%); (i) TBSOTf, 2,6-lutidine, CH2Cl2, 0 °C, 30 min; (j) DBU, MeOH, rt, 14 h
(57% for two steps); (k) 2-chloro-1,3,2-dioxaphosphoran-2-oxide, Et3N, DMAP,
benzene, 0 °C, 1 h; (l) Me3N, CH3CN, 85 °C, 72 h (61% for two steps); (m) TFA, CH2Cl2,
0 °C, 30 min (n) p-nitrophenyl octadecanoate, Et3N, DMAP, THF, rt, 13 h (o) TBAF,
THF, rt, 26 h (54% for three steps).
4. Experimental
4.1. Materials
Sphingomyelin (bovine, brain) and 1,2-dihexanoyl-3-sn-phos-
phatidylcholine (DHPC) were purchased from Avanti Polar Lipids
(Alabaster, AL). Deuterium oxide was purchased from Euriso-top,
and deuterium-depleted water was from ISOTEC. All other chemi-
cals were obtained from standard venders. Thin-layer chromatog-
raphy (TLC) was performed on a glass plate precoated with silica
gel (Merck Kieselgel 60 F254).
4.2. Purification of SSM
Because commercially available semi-synthetic D-erythro-N-ste-
aroyl-sphingomyelin (SSM) is a mixture of epimers at C3 of sphin-
gosine, we purified SSM from bovine brain SM (Avanti polar lipid)
by HPLC with a slightly modified procedure of the previous report.28
Bovine brain SM was a mixture of SMs with 16:0 (2%), 18:0 (SSM,
49%), 20:0 (5%), 22:0 (8%), 24:0 (6%) and 24:1 (20%),15 which were
separated by reverse-phase HPLC (Shimadzu LC-10ADvp with an
SPDM10Av photodiode array detector) with the following condi-
tions; column: cosmosil-5C18-AR-II (20 ꢀ 250 mm), eluent: MeOH,
flow rate: 4 mL/min, detection: 205 nm UV absorption. Retention
times are 55 min for C18:0-SM (SSM), 72 min for C20:0 and
90 min for C24:1.
Figure 9. Conformation of the central part of SM in bicelles deduced from 1H NMR
data.
3
and H-3/H-5. The observed JH,H values typical for anti-orienta-
tion suggest that the conformational alteration hardly occurs
with respect to these bonds. This is supported by the fact that
the NOE data are completely consistent with the dihedral angles.
Finally, energy minimization of the initial structure eventually
resulted in the conformation of SM in bicelles, as shown in
Fig. 9. Thus, we successfully determined the conformation of
the partial structure of SM in bicelles and found that the C2–
C6 and amide portions of SM take a relatively rigid conformation
under membrane environments.
4.3. Bicelle preparation
Bicelles were prepared according to a general procedure.13,29
For 31P NMR measurements and dynamic light scattering experi-
ments, commercially available brain SM (bSM) was used without
purification. bSM (ca. 220 mg for q = 4 bicelles or 90 mg for
q = 0.5 bicelles) was dissolved in chloroform/methanol, dried in va-
cuo overnight, rehydrated with H2O (600
245 L for q = 0.5 bicelles), and vortexed vigorously. To the suspen-
sion was added DHPC in D2O (0.5 M, 150 l for q = 4 or 495 L for
lL for q = 4 bicelles or
l
l
l
3. Conclusions
q = 0.5) and the mixture was subjected to heating (60 °C) and cool-
ing (0 °C) cycles and vigorous vortexing until becoming transpar-
ent. The total lipid concentration (bSM + DHPC) was 500 mM for
both q = 4 and 0.5 bicelles. After both 31P NMR and dynamic light
scattering experiments were completed, the lipid solutions were
In this study, we demonstrated for the first time that SM and
DHPC form both oriented and isotropic bicelles, and in particular,
isotropic bicelles are found to be relatively tolerant to concentra-
tion and temperature changes. Then, using isotropic bicelles, we

T. Yamaguchi et al. / Bioorg. Med. Chem. 20 (2012) 270–278
277
3
diluted to 100 mM by adding H2O, and 31P NMR and dynamic light
Supplementary) was used for calculating JH3/H4 of SM. The C3–
C4 bond was rotated through 360° in 30° degree increments, and
at each point structure was optimized at the B3LYP/6-31G(d,p) le-
vel of theory with the C3–C4 bond rotation fixed. Spin–spin cou-
pling constants were calculated with GIAO method using B3LYP/
6-31G(d,p) basis set.27 The resultant coupling constants were fitted
to a general Karplus Eq. (1) with a dihedral angle (h) and a param-
eter (/) that reflects the phase shift between the cosine modula-
tion and the dihedral angle.
scattering experiments were carried out again.
2
For H NMR measurements,100-d2-SSM (4.8 mg, 6.5
l
mol) and
mol) were dissolved in methanol,
dried in vacuo overnight, and mixed with DHPC (6.6 mol) in
H2O (13.2 L). To the mixture was added H2O (200 l), and the
purified SSM (14.5 mg, 19.8
l
l
l
l
resultant suspension was vortexed vigorously until becoming
transparent, lyophilized, and rehydrated with deuterium-depleted
water (57
500 mM.
lL). The final concentration of lipid (SSM + DHPC) was
3JH=H ¼ Acos2ðh þ /Þ þ Bcosðh þ /Þ þ C
ð1Þ
For 1H NMR measurements, purified SSM (10.3 mg, 14.1
l
mol)
was dissolved in methanol, dried in vacuo overnight, and mixed
with DHPC (28.2 mol) in D2O (282 L). The mixture was vortexed
vigorously until becoming transparent, lyophilized, and rehydrated
with H2O and D2O (380 and 42 L, respectively). The final concen-
The curve fitting was carried out using Origin 6.1 software. The
resultant parameters in Eq. (1) are A = 6.95 Hz, B = ꢂ0.34 Hz,
C = 1.84 Hz, and / = 4.78° (Fig. S8, Supplementary).
l
l
l
tration of lipid (SSM + DHPC) was 100 mM.
4.7. Structure calculation
4.4. NMR measurements
The initial partial structure of SM was constructed based on
NOE and coupling constant data. The structure was minimized
using MMFFs force field in MacroModel version 8.6.
1H and 31P NMR spectra were obtained on a JEOL ECA-500
(500 MHz) spectrometer. JH/H values were extracted from 1D 1H
3
NMR and 2D exclusive COSY (E.COSY) spectra.30–32 1H chemical
shifts were referenced to 3-(trimethylsilyl)-1-propanesulfonic acid
sodium salt (DSS). The digital resolution for the 1D 1H NMR spec-
trum was 0.076 Hz/point. The FID data of E.COSY experiment were
acquired with 24 scans per inclement of 4096 (F2) ꢀ 256 (F1) ma-
trix and processed with 2 and 4 times zero-filling for F1 and F2,
respectively. To extract spin coupling constants from the E.COSY
spectrum, a final digital resolution of 0.076 Hz/point in F2 was
achieved by inverse Fourier transformation, 8 times zero-filling,
and back transformation of selected traces. NOESY experiments
were acquired with 12 scans per inclement of 2048 (F2) ꢀ 256
(F1) matrix and with 30 ms of mixing time. All the 1D and 2D 1H
spectra were recorded at 37 °C.
Acknowledgments
We are grateful to Dr. Yuichi Umegawa and Mr. Mototsugu Doi,
Osaka University, for their help in NMR measurements. This work
was supported by Grant-In-Aids for JSPS Fellows, for Scientific
Research (B) (No. 20310132) and (S) (No. 18101010), and by
SUNBOR grant from Suntory Institute for Bioorganic Research,
Japan. T.Y. expresses special thanks for the Global Center of
Excellence (COE) Program ‘Global Education and Research Center
for Bio-Environmental Chemistry’ of Osaka University.
Supplementary data
31P NMR spectra were acquired with a digital resolution of
4.9 Hz/point. 31P chemical shifts were referenced to 85% phospho-
ric acid. An exponential line broadening of 25 Hz was applied to
the FID before Fourier transformation for both the oriented and
isotropic bicelles.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2011.11.001.
References and notes
2H NMR experiments were carried out on a JEOL ECA-400
(400 MHz) spectrometer. 2H NMR spectra were recorded at
61.4 MHz using a standard quadrupole-echo pulse sequence.33
1. Simons, K.; Ikonen, E. Nature 1997, 387, 569.
2. Anderson, R. G. W.; Jacobson, K. Science 2002, 296, 1821.
3. van der Goot, F. G.; Harder, T. Semin. Immunol. 2001, 13, 89.
4. Liao, Z.; Cimakasky, L. M.; Hampton, R.; Nguyen, D. H.; Hildreth, J. E. K. AIDS Res.
Hum. Retrov. 2001, 17, 1009.
The 90° pulse width was 0.1 ls, interpulse delay was 30 ls, acqui-
5. Mañes, S.; del Real, G.; Lacalle, R. A.; Lucas, P.; Gómez-Moutón, C.; Sánchez-
Palomino, S.; Delgado, R.; Alcamí, J.; Mira, E.; Martínez-A, C. EMBO Rep. 2000, 1,
190.
6. Puri, A.; Hug, P.; Jernigan, K.; Barchi, J.; Kim, H.-Y.; Hamilton, J.; Wiels, J.;
Murray, G. J.; Blumenthal, R. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 14435.
7. Bruzik, K. S. Biochim. Biophys. Acta 1988, 939, 315.
8. Talbott, C. M.; Vorobyov, I.; Borchman, D.; Taylor, K. G.; DuPré, D. B.; Yappert,
M. C. Biochim. Biophys. Acta 2000, 1467, 326.
sition time was 16.38 ms, and recycle delay was 0.5 s. The sweep
width was 500 kHz, and delay between the second 90° pulse and
data acquisition was 20 ls. The number of scans was 24000. An
exponential line broadening of 1000 Hz was applied to the FID be-
fore Fourier transformation.
9. Sanders, C. R.; Hare, B. J.; Howard, K. P.; Prestegard, J. H. Prog. Nucl. Magn. Reson.
Spectrosc. 1994, 26, 421.
4.5. Dynamic light scattering
10. Sasaki, H.; Araki, M.; Fukuzawa, S.; Tachibana, K. Bioorg. Med. Chem. Lett. 2003,
13, 3583.
11. Dürr, U. H. N.; Yamamoto, K.; Im, S.-C.; Waskell, L.; Ramamoorthy, A. J. Am.
Chem. Soc. 2007, 129, 6670.
12. Matsumori, N.; Morooka, A.; Murata, M. J. Am. Chem. Soc. 2007, 129, 14989.
13. Matsumori, N.; Murata, M. Nat. Prod. Rep. 2010, 27, 1480.
14. Loudet, C.; Manet, S.; Gineste, S.; Oda, R.; Achard, M.-F.; Dufourc, E. J. Biophys. J.
2007, 92, 3949.
15. Frazier, M. L.; Wright, J. R.; Pokorny, A.; Almeida, P. F. F. Biophys. J. 2007, 92,
2422.
16. Uddin, M. N.; Morrow, M. R. Langmuir 2010, 26, 12104.
17. Ottiger, M.; Bax, A. J. Biomol. NMR 1998, 12, 361.
18. Sasaki, H.; Fukuzawa, S.; Kikuchi, J.; Yokoyama, S.; Hirota, H.; Tachibana, K.
Langmuir 2003, 19, 9841.
19. Schmidt, C. F.; Barenholz, Y.; Thompson, T. E. Biochemistry 1977, 16, 2649.
20. Henderson, T. O.; Glonek, T.; Myers, T. C. Biochemistry 1974, 13, 623.
21. Haasnoot, C. A. G.; De Leeuw, F. A. A. M.; Altona, C. Tetrahedron 1980, 36, 2783.
22. (a) Wang, A. C.; Bax, A. J. Am. Chem. Soc. 1996, 118, 2483; (b) Dreef, C. E.; Elie, C.
J. J.; Hoogerhout, P.; van der Marel, G. A.; van Boom, J. H. Tetrahedron Lett. 1988,
29, 6513.
The sizes of bicelles were evaluated by dynamic light scattering
at 20–50 °C. The concentration of bicelles was 500 or 100 mM. All
histograms were obtained on a HORIBA LB-550 dynamic light scat-
tering particle size analyzer.
3
4.6. Dihedral angle analysis using JH/H
In this study, dihedral angles for three bonds in SM, C2–C3, C3–
3
C4, and C2–N, were analyzed using JH/H values. Reported Karplus
relations were adopted for the rotations of the C2–C321 and C2–
N22 bonds. On the other hand, because no Karplus type equation
has been reported for C(sp2)–C(sp3) bond, Karplus relation for the
C3–C4 bond was derived from spin coupling constants calculated
by GIAO method on the Gaussian 09W software.34 A truncated
model including the interfacial region of SM (as shown in Fig. S8,

278
T. Yamaguchi et al. / Bioorg. Med. Chem. 20 (2012) 270–278
23. Yamamoto, T.; Hasegawa, H.; Hakogi, T.; Katsumura, S. Org. Lett. 2006, 8, 5569.
24. Bystrov, V. F. Prog. Nucl. Magn. Reson. Spectrosc. 1976, 10, 41.
25. Deng, W.; Cheeseman, J. R.; Frisch, M. J. J. Chem. Theory Comput. 2006, 2, 1028.
26. Lutnæs, O. B.; Ruden, T. A.; Helgaker, T. Magn. Reson. Chem. 2004, 42, 117.
27. Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785.
28. Jungalwala, F. B.; Hayssen, V.; Pasquini, J. M.; McCluer, R. H. J. Lipid Res. 1979,
20, 579.
30. Griesinger, C.; Sørensen, O. W.; Ernst, R. R. J. Am. Chem. Soc 1985, 107, 6394.
31. Griesinger, C.; Sørensen, O. W.; Ernst, R. R. J. Chem. Phys. 1986, 85, 6837.
32. Griesinger, C.; Sørensen, O. W.; Ernst, R. R. J. Magn. Reson. 1987, 75, 474.
33. Davis, J. H.; Jeffrey, K. R.; Bloom, M.; Valic, M. I.; Higgs, T. P. Chem. Phys. Lett.
1976, 42, 390.
34. Frisch, M. J. et al. Gaussian 09, Revision A.02, Gaussian, Inc., Wallingford CT,
2009.
29. Mäler, L.; Gräslund, A. Methods Mol. Biol. 2009, 480, 129.