Conformations of spermine in adenosine triphosphate complex: The structural basis for weak bimolecular interactions of major cellular electrolytes

Article abstract of DOI:10.1002/chem.201002759

Selectively ²H- and ¹³C-labeled spermines (SPM) were efficiently synthesized and analyzed by NMR spectroscopy to determine the spin-spin coupling constants for six conformationally relevant bonds. SPM that is composed of three alkyl moieties, a butanylene, and two propanylene chains undergoes a conformational change when interacting with multivalent anions (e.g., adenosine triphosphate (ATP), ATP-Mg²⁺, and tripolyphosphate). Upon interaction with ATP, the C-C bonds, which affect the distance between the neighboring pairs of ammonium groups (i.e., N1/N5 and N5/N5′), increase the population of gauche rotamers by 17-20 % relative to those in the 4 HCl salt of SPM. However, the trend in increments of the gauche conformers for the SPM-ATP complex profoundly differs from that of the spermidine (SPD)-ATP complex. This implies that SPM may preferentially recognize the adenyl group of ATP rather than the tripolyphosphate moiety. This may account for the higher affinity of SPM to ATP-Mg²⁺ than with that of SPD, which chiefly interacts with β- and I-phosphates and is easily replaced by Mg ²⁺. These results may provide a clue for the further understanding of the structural basis of polyamine biological functions. Spermines and spin: Spermines (SPM) labeled selectively with ²H and ¹³C were used to determine the spin-spin coupling constants for six conformationally relevant bonds. SPM revealed diverse conformational changes upon interaction with adenosine triphosphate (ATP), ATP-Mg²⁺, and tripolyphosphate (TPP, see graph). Copyright

Full text of DOI:10.1002/chem.201002759

DOI: 10.1002/chem.201002759
Conformations of Spermine in Adenosine Triphosphate Complex: The
Structural Basis for Weak Bimolecular Interactions of Major Cellular
Electrolytes
Keisuke Maruyoshi, Toshiyuki Yamaguchi, Tetsuo Demura, Nobuaki Matsumori,
Tohru Oishi, and Michio Murata*[a]
Abstract: Selectively 2H- and 13C-la-
beled spermines (SPM) were efficiently
synthesized and analyzed by NMR
spectroscopy to determine the spin–
spin coupling constants for six confor-
mationally relevant bonds. SPM that is
composed of three alkyl moieties, a
butanylene, and two propanylene
ATP, the C C bonds, which affect the
distance between the neighboring pairs
of ammonium groups (i.e., N1/N5 and
N5/N5’), increase the population of
gauche rotamers by 17–20% relative to
those in the 4HCl salt of SPM. Howev-
er, the trend in increments of the
gauche conformers for the SPM–ATP
complex profoundly differs from that
of the spermidine (SPD)–ATP com-
plex. This implies that SPM may pref-
erentially recognize the adenyl group
of ATP rather than the tripolyphos-
phate moiety. This may account for the
higher affinity of SPM to ATP–Mg2+
than with that of SPD, which chiefly in-
teracts with b- and g-phosphates and is
easily replaced by Mg2+. These results
may provide a clue for the further un-
derstanding of the structural basis of
polyamine biological functions.
À
chains undergoes
a conformational
change when interacting with multiva-
lent anions (e.g., adenosine triphos-
phate (ATP), ATP–Mg2+, and tripoly-
phosphate). Upon interaction with
Keywords: adenosine triphosphate ·
conformation analysis · NMR spec-
troscopy · polyamines · spermines
Introduction
Polyamines such as spermine (SPM, 1) and spermidine
(SPD, 2) play essential roles in diverse biological events.
The distribution and functions of these polyamines are
known to be largely similar.[3a] Earlier titration experiments
for the corresponding aqueous solutions revealed variations
in the affinity to nucleotides as a function of the degree of
phosphorylation and the species of nucleotide base exami-
ned.[3b] Furthermore, the question of why biological poly-
amines are mainly composed of three compounds (i.e.,
putrescine (PUT), spermidine, and spermine) remains unan-
swered. With respect to the stimulation of protein synthesis,
three polyamines demonstrate a similar effect though their
effective concentrations are different.[4] Additionally, some
proteins that specifically recognize PUT, SPD, or SPM are
known to be involved in biological functions.[5,6] The differ-
ent roles of the polyamines should be derived from their
structural differences. Therefore, conformation studies of
SPM may provide a clue to answer this question. Poly-
amines, in particular SPM, form a ternary complex with
ATP–Mg2+ that can affect phosphorylation of proteins. It is
Molecular recognition between ligands and their corre-
sponding receptors often exhibit markedly high affinities,
which can result in nanomolar or sub-nanomolar dissocia-
tion constants, as exemplified by receptor interactions with
hormones, drugs, secondary messengers, and toxins. In addi-
tion to these strict molecular recognitions, weak interactions
between biomolecules often play crucial roles, such as those
observed in neurotransmitters reception.[1] However, molec-
ular basis responsible for these weak interactions remains
largely unknown since the relevant compounds are flexible
and undergo rapid association and dissociation. Due to the
dynamic nature of these interactions, the structural elucida-
tion of a given 3D ligand–receptor complex is extremely dif-
ficult. As a result, well-defined conformations of highly flex-
ible linear biomolecules remain largely an unmet objec-
tive.[2]
reported that SPM predominantly interacts with the b- and
[a] Dr. K. Maruyoshi,+ T. Yamaguchi,+ T. Demura, Prof. N. Matsumori,
Prof. T. Oishi, Prof. M. Murata
[7]
g-phosphates of ATP in the presence of Mg2+
.
The
Department of Chemistry, Graduate School of Science
Osaka University, Machikaneyama
Toyonaka, Osaka 560-0043 (Japan)
Fax : (+81)66850-5774
E-mail: murata@chem.sci.osaka-u.ac.jp
ATPase activity of PotA, a protein component of a bacterial
uptake system of SPD,[8] was greatly enhanced by SPM pos-
sibly due to formation of a ternary complex of SPM–ATP–
Mg2+. The activity of protein kinase A was also stimulated
about twofold by SPM. The ternary complex, therefore, may
play an important role in signal transduction during the pro-
liferation stage of eukaryotic cells. To gain a better under-
[+] These authors equally contributed to this work.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201002759.
4788
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 4788 – 4795

FULL PAPER
Table 1. Populations of gauche rotamers of SPM in complexation with ATP.
standing of the functional dif-
ferences among polyamines, we
examined in this study the con-
formation change of SPM in
the presence and absence of
ATP and compared the results
with those of SPD.
Labeled SPM
Bond (torsion angle)
Jobsd.
(H2,H3)
Counterion
3J [Hz] gauche [%] Increment[a] [%]
3J
3J
G
4HCl
9.4
8.5
7.4
7.4
10.1
9.5
8.3
8.3
2.33
2.35
2.34
2.36
2.57
2.51
2.55
2.55
10.1
9.4
8.5
18
27
38
38
11
17
29
29
26
26
26
27
34
32
33
33
11
18
28
28
24
24
25
27
À
SPM-2,3-d12 (3)
C2 C3 (N1/C4)
pH 3.3 (ATP2À
pH 5.6 (ATP3À
pH 7.3 (ATP4À
4HCl
)
)
)
9
20
20
À
SPM-3,4-d12 (4)
SPM-6-13C-d8 (5)
C3 C4 (C2/N5)
A
pH 3.3 (ATP2À
pH 5.6 (ATP3À
pH 7.3 (ATP4À
4HCl
)
)
)
6
18
18
3J
3J
N
À
C4 N5 (C3/C6)
Results
pH 3.3 (ATP2À
pH 5.6 (ATP3À
pH 7.3 (ATP4À
4HCl
)
)
)
0
0
1
2
Molecular design of H-labeled
spermines: Molecular probes 3– SPM-4-13C-d8 (6)
8 were designed to examine the
N5 C6 (C4/C7)
À
pH 3.3 (ATP2À
pH 5.6 (ATP3À
pH 7.3 (ATP4À
4HCl
)
)
)
À2
À1
À1
spin–spin coupling constants for
the six conformationally rele-
vant bonds of a C2-symmetric
SPM molecule. With the eryth-
ro-deuterated SPMs 3, 4, 7, and
8, the rotational conformation
3J
A
À
SPM-6,7-d14 (7)
C6 C7 (N5/C8)
pH 3.3 (ATP2À
pH 5.6 (ATP3À
pH 7.3 (ATP4À
4HCl
)
)
)
7
17
17
8.5
SPM-7-13C-d2 (8)
C7 C7’ (C6/C6’)
3J
(H7,H7’)
10.3
10.3
10.3
9.9
À
pH 3.3 (ATP2À
pH 5.6 (ATP3À
pH 7.3 (ATP4À
)
)
)
0
1
3
À
À
À
around C2 C3, C3 C4, C6 C7,
À
and C7 C7’ can be elucidated
by the measurement of the re-
[a] Increment of the gauche population from that of the 4HCl salt of SPM.
spective 1H–1H coupling con-
stant;[9] additional 13C labeling
at C7 in 8 was necessary to resolve H7 and H7’ signals in
titration experiments, which agrees well with the reported
data.[13] The value indicates that 99% of SPM is bound to
ATP under the conditions of the NMR spectroscopic experi-
ments described later. Thus the spin coupling constants in
Table 1 should be regarded as those of the SPM–ATP com-
plex. Next, the binding affinity of the polydeuterated SPM
to ATP was estimated by NMR spectroscopic titrations to
ensure that the deuterium substitutions did not influence
1
the H NMR spectra. The remaining 13C-labeled SPMs 5 and
À
6 were prepared to analyze the rotamer about the C4 N5
3
À
and N5 C6 bonds, respectively, by measuring J
N
pling. Synthetic routes were similar to those employed in
the synthesis of SPD as we have reported previously.[9]
Briefly, erythro-deuteration was effected by a cis-selective
deuteration of the a,b-unsaturated d-lactone. A successive
Curtius rearrangement provided the propanylene moiety,
whereas a reduction/azidation/hydrogenACTHNUTRGNEaNUG tion sequence fur-
nished the butanylene moiety. Coupling between the two
parts was carried out using the Fukuyama method.[10] 13C la-
beling was introduced by SN2 displacement with 13CNÀ or a
Wittig olefination (Scheme 1; see the Supporting Informa-
tion for details).
Binding affinity and stoichiometry of labeled SPMs to ATP:
The stoichiometry of the SPM–ATP complex was deter-
mined to be one to one by examination of Jobꢁs plot (Fig-
ure S1 in the Supporting Information).[11] The Stokes radius,
d, was calculated from the experimental data of the diffu-
sion coefficients (D values) by using the Stokes–Einstein
equation (Supporting Information).[12] The Stokes radius of
the 4HCl salt of SPM (1.21 nm) appeared to be smaller
than the length of SPM in the extended conformation
(ꢀ1.6 nm). The Stokes radius of SPM in the presence of
ATP is slightly larger than that of SPM. These results rule
out the formation of larger multivalent complexes such as
(SPM)2–(ATP)2 or (SPM)3–(ATP)3, thereby inferring their
one-to-one relationship.
The association constant (Ka) of nondeuterated SPM with
Scheme 1. Labeled spermines. [a] These compounds are racemic. [*] Po-
sition is labeled with 13C.
ATP was determined to be 1722mÀ1 by NMR spectroscopic
Chem. Eur. J. 2011, 17, 4788 – 4795
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4789

M. Murata et al.
the complexation of SPM with ATP. The association con-
stant Ka between SPM-6,7-d14 (7) and ATP was estimated to
be 1670mÀ1 by the nonlinear least-square fit of the SPM H7
chemical shift (Figure S2 in the Supporting Information).
This data supports the notion that deuterium incorporation
does not appreciably affect the stability of the SPM–ATP
complex.
When compared with the HCl salt, the 3J
ACTHUNGTERNNU(G H,H) values of
the ATP salts notably decreased for deuterated SPMs 3, 4,
and 7. Regarding the ionization state of ATP, the pKa values
in H2O were reported at 4.00 and 6.50 under conditions sim-
ilar to the present experiments.[14] By converting pKa from
H2O to D2O solutions according to a reported method,[15]
these values become 4.25 and 6.94, respectively, thereby in-
dicating that ATP is largely divalent at pH 3.3, trivalent at
pH 5.6, and tetravalent at pH 7.3. At neutral pH, the 3J-
Determination of the spin–spin coupling constants for SPM
as the HCl, ATP, ATP–Mg2+, and tripolyphosphate (TPP)
ACHTUNGTREN(NUGN H,H) value was the smallest, which clearly indicates that
3
salts: The spin–spin coupling constants, J
N
SPM undergoes conformational changes that serve to in-
crease the bent conformation (equal to gauche-rich) by 17–
20%. These conformational alterations are further observed
to be pH-dependent, an effect that is primarily due to the
number of anionic sites on ATP; 3, 4, and 5 exhibited higher
gauche populations as the pH was raised from 3.3 to 5.6.
sured for the labeled SPMs. The SPMs as HCl salts
(Figure 1, series A) and those of SPM in the presence of
ATP were subjected to NMR spectroscopic measurements
at various pH values (Figure 1, series B–D). For the labeled
3
SPMs 3, 4, 7, and 8, the J
N
À
with an accuracy greater than 0.1 Hz despite the fact that
the relevant 1H signals were somewhat broadened as a
result of the association–dissociation equilibrium, for which
Unlike the other C C bonds, the gauche enhancement for
the C7/C7’ rotamers was small in the presence of ATP
(Table 1). For analogues 5 and 6, the gauche ratios were de-
termined by measuring the J
3
3
spectral simulations were carried out to obtain accurate J-
N
3
G
N
eronuclear multiple bond correlation (HMBC) experiments;
as shown in Figure 2; the HMBC cross-peak intensities that
were obtained by varying the correlation time are in agree-
salts of SPM ranged between 9.4–10.1 Hz, which revealed
that the dihedral angles of N1/C4, C2/N5, N5/C8, and C6/
C6’ predominantly adopt the anti orientation (Table 1).
ment with the theoretical curves calculated for 3J
ACHTUNGTRENNUNG(C,H)
Figure 1. Partial 1H NMR spectra (500 MHz) of the labeled SPMs (3, 4, 7, and 8) at 408C. NMR spectroscopic experiments were carried out with a
25 mm solution of the 4HCl salt of deuterated SPM at pH 7.3 in D2O and combined with: A) no additive, B) a 100 mm ATP solution at pH 3.3, C) a
100 mm ATP solution at pH 5.6, or D) a 100 mm ATP solution at pH 7.3. *J values of these broad peaks with an asterisk were determined by spectral
simulations (gNMR) since the peak-to-peak distances of broad signals often led to smaller coupling constants.
4790
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 4788 – 4795

Spermine Conformations
FULL PAPER
3
Figure 2. Interval-dependent changes in intensity of HMBC cross peaks due to JACHTUNGTRNEUNG
(C,H) values of 13C-labeled SPMs (5 and 6). NMR spectroscopic experi-
ments were carried out: A) 25 mm of the 4HCl salt of labeled SPM 5 at pH 7.3 in D2O; B) 25 mm of 5 in the presence of 100 mm ATP at pH 7.3;
C) 25 mm of 4HCl salt of 6 at pH 7.3 in D2O; and D) 25 mm of 6 in the presence of 100 mm ATP at pH 7.3. See the Supporting Information for details of
the NMR spectroscopic experiments.
values. As is the case with SPD, the gauche populations with
complex with ATP with higher affinity than SPM, the con-
formational changes of SPM are somewhat attenuated
(Table 2).
À
À
À
respect to the C N bonds (i.e., C4 N5 and N5 C6) are not
significantly affected by ATP (Table 1).
3
We next measured J
N
and ATP–Mg2+ for the C2 C3, C3 C4, C6 C7, and C7 C7’
bonds by using the same methodology since the other two
bonds that showed no significant differences upon interac-
tion with ATP were deduced to be unaffected as discussed
later. In the presence of Mg2+, which is known to form a
À
À
À
À
Discussion
Conformation of SPM in the presence of ATP: Before dis-
cussing the SPM–ATP complex, we should examine the con-
formation of the 4HCl salt of
SPM. The population of the
Table 2. Populations of anti/gauche rotamers of SPM in the presence of Mg2+ (100 mm).
À
anti rotamer about each C C
Labeled SPM
Bond (torsion angle)
Jobsd.
Counterion
3J [Hz] gauche [%] Increment[a] [%]
bond of SPM is approximately
82–89% (Table 1), which indi-
cates that the anti conformer of
SPM is more populated than
that of the alkanes; this is most
likely due to the electrostatic
repulsion among the four am-
monium groups. A similar anti-
dominant conformation has
been observed for SPD.[9]
3J
3J
3J
(H2,H3) pH 3.3 (ATP2À
)
)
)
)
)
)
)
)
)
)
)
)
8.7
8.7
8.8
9.7
9.4
9.5
9.5
9.2
9.1
25
25
24
15
18
17
17
21
22
24
25
25
7
7
6
4
7
6
6
10
11
0
À
SPM-2,3-d12 (3)
C2 C3 (N1/C4)
pH 5.6 (ATP3À
pH 7.3 (ATP4À
(H3,H4) pH 3.3 (ATP2À
pH 5.6 (ATP3À
À
SPM-3,4-d12 (4)
SPM-6,7-d14 (7)
SPM-7-13C-d2 (8)
C3 C4 (C2/N5)
pH 7.3 (ATP4À
À
C6 C7 (N5/C8)
A
pH 5.6 (ATP3À
pH 7.3 (ATP4À
À
C7 C7’ (C6/C6’)
3J
(H7,H7’) pH 3.3 (ATP2À
10.3
10.2
10.1
pH 5.6 (ATP3À
pH 7.3 (ATP4À
1
1
In the presence of ATP, the
[a] Increment of the gauche population from that of the 4HCl salt of SPM.
1H–1H spin coupling constants
Chem. Eur. J. 2011, 17, 4788 – 4795
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4791

M. Murata et al.
of SPM were successfully determined for the specific com-
plex that is thought to play a physiological role in ATP hy-
drolysis.[7] When interacting with ATP, the gauche conform-
The SPM–ATP interaction at different ionization states
was examined by determining the gauche populations at var-
ious pH values. The optimized structures of ATP with the
net charges of À2, À3, and À4 have been deduced previous-
ly (see Figure 4).[16,17] The populations of rotamers about
À
À
À
ers markedly increase for the C2 C3, C3 C4, and C6 C7
bonds; the populations are roughly doubled or tripled
(Table 1). These data suggest that SPM tends to assume a
bent conformation in the ATP complex, which is also the
case with SPD.[9] However, with respect to the remaining
À
each of the C C bonds did not significantly change between
pH 5.6 and 7.3 in the presence of ATP (Table 1). At pH 7.3,
the g-phosphate becomes divalent, thus leading to tetra-
À
À
À
bonds (i.e., C4 N5, N5 C6, and C7 C7’) the gauche/anti
ratios are virtually unchanged. This can be explained by the
attenuation of the electrostatic repulsion by ATP. The re-
spective distances between the nitrogen atoms in the propa-
ACHTUNGTRENvNUGN alent ATP. These unchanged gauche populations indicate
that SPM has very weak interaction with the g-phosphate
terminus. These assertions are also additionally supported
by the pH-dependent change in affinity of SPM with ATP
obtained from the current NMR spectroscopic titration ex-
periments; the Ka values at pH 3.3, 5.6, and 7.3 are 912,
1611, and 1722mÀ1, respectively.
À
À
nylene and butanACHTUNGTRENNUNGylene moieties (i.e., N1 N5 and N5 N5’)
are about 0.50 and 0.63 nm in the extended conformation,
which is larger than the distance between the neighboring
oxygen atoms in the tripolyphosphate portion of ATP (0.41–
0.43 nm). Thus, SPM needs to adopt the conformation that
arranges each cationic ammonium group in close proximity
to the anionic site of ATP (as observed for SPD).[9] On the
other hand, the rotation of the C4 N5, N5 C6, or C7 C7’
bond does not affect the distance of neighboring ammonium
groups to the same extent as that of the C2 C3, C3 C4, or
The present findings are in agreement with the previous
report by Meksuriyen et al. in that the conformational
changes induced by SPM were observed only for the a-phos-
phate of ATP and not for the b- and g-phosphate in the ab-
sence of Mg2+ as revealed by 31P NMR spectroscopy.[7] Our
unpublished data indicated that the Ka value of SPM with
cytosine triphosphate (CTP) is appreciably less than that of
SPM with ATP, whereas the Ka value between SPD and
CTP is nearly equivalent to that between SPD and ATP
(data not shown). It is implied that SPM may predominantly
interact with the adenosine moiety and/or the a-phosphate
group in the absence of Mg2+. The present results may shed
light on the differing physiological roles between SPD and
SPM. As depicted in Figure 3, SPM exhibited an enhanced
population of gauche conformers relative to that of SPD in
the presence of ATP, which corroborates the higher affinity
À
À
À
À
À
À
À
C6 C7 bond; this is analogous to that of the C3 C4 bond of
SPD.[9] The gauche conformer with respect to the C7 C7’
À
bond reduces the distance between N5 and N5’ by 0.05 nm,
which is much smaller than the reduction of the distance by
0.11 nm with the rotation about C6 C7 (or C6’ C7’). Rota-
tion of C4 N5 or N5 C6 evidently produces no change in
the N1/N5 or N5/N5’ distance.
A different trend in the coupling constants was observed
for the SPM complex with TPP (Figure 3), which clearly in-
dicates that the butanylene moiety bends more deeply to fa-
À
À
À
À
À
À
of SPM to ATP. Specifically, the C2 C3 and C3 C4 bonds,
À
which correspond to the C8 C9
À
and C7 C8 bonds of SPD, re-
spectively, revealed
a signifi-
cantly higher proportion of the
gauche rotamers in the pres-
ence of ATP. This is not promi-
nent in the SPM–TPP complex
(Figure 3). These results again
imply that SPM more preferen-
tially recognizes the adenyl
group of ATP rather than the
g-phosphate moiety. Another
notable difference between
SPD and SPM is found in con-
formation of the TPP complex.
Whereas the three relevant
bonds in SPD essentially indi-
cated the same populations be-
tween the ATP and TPP com-
plexes, SPM revealed a substan-
Figure 3. Increments of the gauche conformers in a) SPM–ATP, b) SPM–TPP, c) SPM–ATP–Mg2+ ternary com-
plex, d) SPD–ATP, e) SPD–TPP, and f) SPD–ATP–Mg2+ ternary complex compared with those in the 4HCl
salt of SPM or 3HCl salt of SPD. All the data were obtained at pH 7.3. Detailed data for a)–f) are provided in
Table 1, Table S1 in the Supporting Information, Table 2, and our previous report.[9]
cilitate interaction with TPP because of its longer alkyl
chain. The difference from those of the SPM–ATP complex
in Figure 3A clearly demonstrates that SPM, unlike SPD,
recognizes not only the triphosphate group of ATP but also
the remaining portion of ATP.
tially different trend (Figure 3). The discrepancy may be ex-
plained by the notion that close interaction of the adenine
moiety in ATP with the peripheral N1 amino group of SPM
ACHTUNGTRENNUNG
À
À
encourages the C2 C3 or C3 C4 bond to more deeply bend
(Figure 4C). In the SPM–TPP salt, in which only three out
4792
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 4788 – 4795

Spermine Conformations
FULL PAPER
the terminal N1/N1’ amino groups is not involved in the salt
formation with TPP (Figure 4D). These interactions may
cause an increased bent conformation in the butanylene
À
moiety (C6 C7) in comparison with the propanylene (i.e.,
À
À
C2 C3/C3 C4) in the SPM–TPP complex.
Conformation of SPM in SPM–ATP–Mg2+ ternary
complex: SPM is known to form a ternary complex SPM–
ATP–Mg2+ with an ATP–Mg2+ salt under physiological con-
ditions,[7] and it is thought to play a significant role in ATP-
dependent events. Hence, we examined the conformation of
the SPM–ATP–Mg2+ ternary complex using the same strat-
egy. Under the conditions of NMR spectroscopic measure-
ments with relatively high concentrations of these three
electrolytes for Table 2, the NMR spectroscopic data could
largely be attributed to the ternary complex since it was re-
ported that the presence of Mg2+ does not greatly reduce
the affinity of spermine to ATP, and vice versa.[7]
To deduce the conformational change of SPM upon inter-
action with ATP–Mg2+, the relevant J values (Table 2) were
measured under the conditions in which 85% of SPM was
deduced to be complexed with ATP–Mg2+ on the basis of
the Ka value of 71.0mÀ1 at pH 7.3 (Figure 5). No significant
Figure 5. The association constants (Ka) between SPM and ATP–Mg2+ in
D2O at pH 7.3 and 408C were estimated with the nonlinear least-squares
fit of the chemical shift of the H7 signal of SPM (Ka =71.0mÀ1).
Figure 4. Hypothetical depictions of SPM molecules that account for the
pH-dependent conformational change. Three of the possible conforma-
tions are present at pH 3.3, 5.6, and 7.3. The conformation of ATP was
obtained by the force-field calculation using the optimized potential for
À
conformational change about each C C bond was observed
between pH 5.6 and 7.3 in the presence and absence of
Mg2+. Moreover, increments of the gauche conformers were
smaller than those observed in the absence of Mg2+ (Table 2
and Figure 3). However, an 11% increase was observed for
liquid simulations (OPLS; B=butanylene part, P=propanylACTHNUGTRNEUNGene part).
A) SPM mainly interacts with the triphosphate group of ATP at pH 3.3.
B) The propanylene portion of SPM recognizes the adenyl moiety while
bearing the weak interaction of the terminal g-phosphate at pH 5.6.
C) SPM–ATP interaction at pH 7.3 is essentially the same as that at
pH 5.6 (B) although the charge of the g-phosphate becomes divalent.
D) SPM shapes interaction with TPP4À. The conformation of TPP was
obtained by the molecular modeling (MM) calculation. The relative posi-
tion of SPM and ATP (or TPP) derived from the simulation is not based
on experimental data.
À
the C6 C7 bond, which may suggest that the SPM–ATP–
Mg2+ complex possesses a different SPM–ATP interaction
than that of the SPM–ATP complex. A similar trend in
À
which the C6 C7 bond predominantly adopted the gauche
conformation was observed for the complex between SPM
and TPP (Figure 3). This indicates that SPM preferentially
interacts with the tripolyphosphate moiety, particularly with
the a- and b-phosphate groups of ATP in the presence of
Mg2+. These results are roughly consistent with the previous
of the four amino groups in SPM are involved in electrostat-
ic interactions, the internal N5/N5’ amino groups can fully
form the salt with the phosphates in TPP, whereas one of
Chem. Eur. J. 2011, 17, 4788 – 4795
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4793

M. Murata et al.
observation that Mg2+ chiefly interacts with the b-, g-phos-
phate group.[18,19] In contrast, SPD tends to bind to the g-
phosphate part as reported previously (Scheme 2). In this
physiological conditions and could be important in other
molecular interactions with biological polyanions, which
may partly account for the physiologically differential roles
among polyamines, spermine, spermidine, and putrescine.
Experimental Section
Preparation of labeled SPMs: Labeled SPMs 3–8 were prepared by three
key steps: a) erythro-selective hydrogenation of a a,b-unsaturated lactone
or lactam, b) Curtius rearrangement to provide a primary amino group,
and c) N-alkylation by using the method of Fukuyama et al.[10] Details of
the synthesis are provided in the Supporting Information and our previ-
ous report.[9]
Measurements of stoichiometry and association constants in the SPM–
ATP complex: The Jobꢁs plot was performed in D2O at 408C using a
JEOL GSX-500 500 MHz spectrometer. The total concentration of SPM
plus ATP was maintained at 10 mm. The pH value was adjusted to 7.3
with deuterium chloride and sodium deuteroxide. The chemical shifts
were recorded from 3-(trimethysiliy)-1-propanesulfonic acid sodium salt
(DSS). The DdꢂnSPM values (Dd: the change in the chemical shift of H7
of SPM induced by addition of ATP; nSPM: molar fraction of SPM,
Scheme 2. Simplified models of the SPD–ATP and SPM–ATP complexes
in the absence of Mg2+
.
[SPM]/ACHTNUTGRENUNG([ATP]+[SPM])) were monitored as a function of nSPM. The self-
diffusion coefficient, D, was estimated by the pulsed-field-gradient spin
echo (PFGSE) method[21] in D2O at 408C using a Varian Unity Inova
750 MHz spectrometer. The concentrations of SPM and ATP were 25
and 100 mm, respectively. The pH value was adjusted to 7.3 with deuteri-
um chloride and sodium deuteroxide. The titration experiments were per-
formed in D2O at 408C with constant SPM concentrations of 25 mm and
ATP concentrations of approximately 6.25, 9.38, 12.5, 18.75, 25, 37.5, and
50 mm. The spectra were obtained using a JEOL GSX-500 500 MHz
spectrometer. The pH value was adjusted to 7.3 with deuterium chloride
and sodium deuteroxide. The chemical shifts were recorded from DSS.
The downfield shift of the H7 signal of SPM was monitored as a function
of the ATP concentration. The data were fitted to a theoretical titration
curve to obtain the association constants by using Origin 6.1 software
provided by OriginLab Corporation.
case, SPD should be readily repelled by Mg2+ to avoid an
unfavorable charge–charge interaction. This difference in
the binding mode may account for the higher affinity of
SPM to the ATP–Mg2+ complex relative to that of SPD. To
further examine the conformation of SPM in ATP or ATP–
Mg2+ complexes, molecular modeling should be necessary.
Elucidation of 3D structures of these highly flexible mole-
cules, however, is still very difficult since conformation of
SPM is very dynamic and depends heavily upon hydration
and also the presence of other ions such as K+ and ClÀ; for
example, even upon complexation with structurally firm
DNA duplex, SPD is reported to be very flexible and forms
no defined complexes.[20]
Method for determining spin–spin coupling constants: Samples of 3–8
were dissolved in D2O. The concentrations of SPM tetrachloride salt,
ATP dipotassium salt, MgCl2, and TPP pentapotassium salt were 25, 100,
100, and 100 mm, respectively. The pH value was adjusted to 3.3, 5.6, or
7.3 with deuterium chloride and sodium deuteroxide. 3J
ACTHNUGRTENUNG(H,H) values
were extracted from 1D 1H NMR spectra. NMR spectra were obtained
using a JEOL GSX-500 500 MHz spectrometer and a JEOL ECA-500
500 MHz spectrometer. The chemical shifts were recorded from DSS.
Conclusion
The digital resolution for the 1H NMR spectrum was 0.076 HzpointÀ1
.
2
Using the selectively H- and 13C-labeled SPMs, we succeed-
Hence, measurement of coupling constants can be carried out with an ac-
ed in determining the spin–spin coupling constants for six
conformationally relevant bonds. SPM revealed diverse con-
curacy of Æ0.1 Hz. 3J
ACHTUNGTRENUN(NG C,H) values were measured with HMBC experi-
ments[21] using a JEOL ECA-500 500 MHz spectrometer. The FID data
were acquired with 80 scans per inclement of a 4096 (F2)ꢂ16 (F1)
matrix. The temperature of the probe was kept at 408C. The intensities
of HMBC cross peaks were monitored as a function HMBC interval
from 0 to 400 ms in approximately 30 ms steps while keeping constant ex-
formational changes upon interaction with ATP, ATP/Mg2+
,
and tripolyphosphate. The greater increments of the gauche
conformers in the SPM–ATP complex than those in the
SPM–tripolyphosphate complex imply that SPM more pref-
erentially recognizes the adenyl group of ATP rather than
the polyphosphate moiety in the absence of Mg2+
(Scheme 2). On the other hand, SPM may chiefly interact
perimental time, and fitted to a theoretical curve to obtain the 3J
value by using Origin 6.1 software. The each error of a measured J
A
3
(C,H)
value was less than 0.1 Hz.
Determination of anti/gauche populations and conformation energy: The
with the tripolyphosphate of ATP in the presence of Mg2+
.
population ratio between the anti and gauche conformers were deter-
3
mined on the basis of J
N
The conformations reported here reflect the time-average
images of an SPM–ATP complex. However, these conforma-
tional changes of SPM may provide insight into the func-
tional structure responsible for the weak interactions under
culated as reported previously.[9] Determination of the gauche rotamers
3
based on the J
(C,H) values was carried out in a similar manner by using
the static anti and gauche values. Details of the analysis are provided in
the Supporting Information and our previous report.[9]
4794
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 4788 – 4795

Spermine Conformations
FULL PAPER
[8] K. Igarashi, K. Kashiwagi, Biochem. J. 1999, 344, 633–642.
Acknowledgements
[9] K. Maruyoshi, K. Nonaka, T. Sagane, T. Demura, T. Yamaguchi, N.
Matsumori, T. Oishi, M. Murata, Chem. Eur. J. 2009, 15, 1618–1626.
[10] a) T. Fukuyama, C.-K. Jow, M. Cheung, Tetrahedron Lett. 1995, 36,
6373–6374; b) T. Fukuyama, M. Cheung, C.-K. Jow, Y. Hidai, T.
Kan, Tetrahedron Lett. 1997, 38, 5831–5834; c) T. Fukuyama, M.
Cheung, T. Kan, Synlett 1999, 1301–1303.
We are grateful to Prof. Kazuei Igarashi, Chiba University for encourag-
ing us to initiate polyamine research. This work was supported by
Grants-in-Aid for Scientific Research (A) (no. 15201048) and (S) (no.
18101010), and for Young Scientists (A) (no. 17681027) from MEXT,
Japan, and by the Naito Foundation.
[11] P. Job, Ann. Chim. Phys. (Paris) 1928, 9, 113–203.
[12] E. O. Stejskal, J. E. Tanner, J. Chem. Phys. 1965, 42, 288–292.
[13] D. Lꢃthi, D. Gꢃnzel, J. A. S. McGuigan, Exp. Physiol. 1999, 84, 231–
252.
[14] R. A. Alberty, R. M. Smith, R. M. Bock, J. Biol. Chem. 1951, 193,
425–434.
[15] A. Krezel, W. Bal, J. Inorg. Biochem. 2004, 98, 161–166.
[16] Y. Huse, Y. Iitaka, Acta Crystallogr. B 1969, 25, 498–509.
[17] a) O. Kennard, N. W. Isaacs, J. C. Coppola, A. J. Kirby, S. Warren,
W. D. S. Motherwell, D. G. Watson, D. L. Wampler, D. H. Chenery,
A. C. Larson, K. A. Kerr, L. Riva di Sanseverino, Nature 1970, 225,
333–336; b) O. Kennard, N. W. Isaacs, W. D. S. Motherwell, J. C.
Coppola, D. L. Wampler, A. C. Larson, D. G. Watson, Proc. R. Soc.
London, Ser. A 1971, 325, 401–436.
[1] For example, U. Quast, M. I. Schimerlik, M. A. Raftery, Biochem-
istry 1979, 18, 1891–1901.
[2] a) R. Brasseur, J.-M. Ruysschaert, P. Chatelain, Biochim. Biophys.
Acta, Biomembr. 1985, 815, 341–350; b) S. W. Wilson, W. Cui, F.
Guarnieri, Data Handl. Sci. Technol. 1995, 15, 351–367; c) G.
Barone, D. Duca, A. Silvestri, L. Gomez-Paloma, R. Riccio, G. Bi-
fulco, Chem. Eur. J. 2002, 8, 3240–3245; d) similar conformational
studies have been carried out for synthetic polymers:R. G. Snyder, J.
Chem. Phys. 1967, 47, 1316–1360; e) A. J. Hopfinger, Conformation-
al Properties of Macromolecules, Academic Press, New York, 1973,
Chapter 5, p. 297.
[3] a) S. S. Cohen, Introduction to the polyamines, Prentice Hall, New
Joursey, 1971; b) C. Nakai, W. Glinsmann, Biochemistry 1977, 16,
5636–5641.
[4] R. G. Martin, B. N. Ames, Proc. Natl. Acad. Sci. USA 1962, 48,
2171–2178.
[5] T. Liang, G. Mezzetti, C. Chen, S. Liao, Biochim. Biophys. Acta,
Gen. Subj. 1978, 542, 430–441.
[6] E. C. Wolff, M. H. Park, Yeast 1999, 15, 43–50.
[7] D. Meksuriyen, T. Fukuchi-Shimogori, H. Tomitori, K. Kashiwagi, T.
Toida, T. Imanari, G. Kawai, K. Igarashi, J. Biol. Chem. 1998, 273,
30939–30944.
[18] M. Cohn, T. R. Hughes, J. Biol. Chem. 1962, 237, 176–181.
[19] V. L. Pecoraro, J. D. Hermes, W. W. Cleland, Biochemistry 1984, 23,
5262–5271.
[20] N. Korolev, A. P. Lyubartsev, A. Laaksonen, L. Nordenskiold, Bio-
phys. J. 2002, 82, 2860–2875.
[21] T. Yamaguchi, K. Maruyoshi, N. Matsumori, M. Murata, Chem. Lett.
2008, 37, 1172–1173.
Received: September 27, 2010
Revised: December 25, 2010
Published online: March 8, 2011
Chem. Eur. J. 2011, 17, 4788 – 4795
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4795