Download hydrogen-bonding effect on 15n nmr chemical shifts of the glycine

Journal of Molecular Structure, 240 (1990) 19-29
ElsevIer SCIence Pubhshers B V , Amsterdam
19
HYDROGEN-BONDING EFFECT ON 15N NMR CHEMICAL
SHIFTS OF THE GLYCINE RESIDUE OF OLIGOPEPTIDES IN
THE SOLID STATE AS STUDIED BY HIGH-RESOLUTION
SOLID-STATE NMR SPECTROSCOPY
SHIGEKI KUROKI, SHINJI ANDO, ISAO ANDO
Department of Polymer Chemistry, Tokyo Institute of Technology, Ookayama, Meguro-ku,
152, Tokyo (Japan)
AKIRA SHOJI and TAKUO OZAKI
Department of Industnal Chemistry, College of Technology, Gunma Unwersity, TenJm-cho,
KiryU, 376, Gunma (Japan)
GRAHAM A WEBB
Department of Chemistry, Unwersay of Surrey, Gwldford, Surrey (Gt Bntam)
(ReceIved 19 February 1990)
ABSTRACT
HIgh-resolutIOn I5N NMR spectra of a vanety ofsohd ohgopeptIdes (X-Gly-Gly) contammg
glycme resIdues have been measured, of whIch the crystal structures had already been determmed
by X-ray dIffractIOn The expenmental '5 N chemIcal shIfts of the glycme reSIdues were plotted
agamst the N"'O hydrogen bond length and the N-H bond length m the :-C=O··· H-N; type
hydrogen bond form, respectively It was found that the decrease of the N - H bond length leads to
a hnear mcrease m ION shIeldmg, but there IS no clear relatIOnshIp between the I5N chemIcal shIfts
and the N· .. 0 separatIOn Further, I5N chemIcal shIft calculations were carned out usmg a model
compound, by the FPT -INDO method, m order to further understand the nature of the hydrogen
bond The calculated results reasonably explam the expenmental ones
INTRODUCTION
Recently, high-resolution 15N NMR spectroscopy in the sohd state has been
mcreasingly applied to the investigation of polypeptides, protems and biopolymers [1-12]. In a previous paper [13], we demonstrated that the isotropic
15N chemical shifts of a number of homopolypeptides in the solid state, as
determined by the cross polarization-magic angle spinning (CP -MAS) method,
are significantly displaced according to their particular conformations such as
the a-helix and fJ-sheet forms. Moreover, it was found that the 15N chemical
shift dIfference between the a-helix andfJ-sheet forms for such homopolypeptides in the solid state varies by as much as 1.2-10.0 ppm, depending mamly
0022-2860/90/$0350
© 1990 - ElsevIer SCIence Pubhshers B V
20
on the nature of the amino acid residue [13] Such a large chemical shift dIfference may come from changes m the hydrogen bond through changes in the
dihedral angles (rp, !If).
We have also measured 13C CP-MAS NMR spectra of some oligopeptides
contammg glycine residues in the solid state [14], in order to obtam mformation about the relationshIp between the carbonyl carbon chemIcal shift and
hydrogen bond length. It was found that the 13C sIgnals ofthe carbonyl carbons
in the ~C=O··· H-N( type hydrogen bond form are deshieided wIth a decrease
in the hydrogen bond length but those in the ~C=O··· H-N+E type hydrogen
bond form are shielded with a decrease in the hydrogen bond length Quantum
chemIcal calculations of the 1,C shiel dings for model compounds were performed and found to reproduce reasonably the experimental results, taking
into account the hydrogen bond and conformatIOnal effects [14]. In this work,
in order to obtain and accumulate further knowledge of the hydrogen bonding
in peptides, we attempt to measure 15N CP-MAS NMR spectra of the oligopeptIdes containing glycine residues in the solId state and to clarify the origm
of the relatIOnship between the 15N chemical shift and the manner of the hydrogen bond. Further, m an attempt to obtain a deep msight mto the nature of
the hydrogen bond, we calculate the 15N shieidings and tensor components of
the glycine amide nitrogens by employmg quantum-chemIcal methods.
EXPERIMENTAL SECTION
Matenals
A series of oligopeptides containing glycine resIdues, except for sarcosylglycylglycme (Sar-Gly-Gly), were purchased from Sigma Co. and were recrystallized according to the same procedures as those used in the X-ray dIffractIOn
studies on them. Sar-Gly-Gly was synthesIzed by stepwise elongatIOn of Nhydroxysuccinimide active esters and amino acids [15]. The N -terminal ofthe
active ester was protected by the o-mtrophenylsulfenyl (Nps-) group. This
sample was purified and recrystallized from aqueous solution. Glycylglycine
nitrate (Gly-Gly· HN0 3 ) was obtained by slow evaporatIOn from an equimolar
mixture of glycylglycine and nitric acid m water. Glycylglycine monohydrochloride monohydrate (Gly-Gly· HCI· H 20) was obtained by slow evaporation
from an equimolar mixture of glycylglycine and hydrogen chlonde in water.
Measurements
15N CP-MAS NMR spectra were recorded at room temperature and 27.25
MHz with a JEOL GSX-270 NMR spectrometer equipped wIth a CP-MAS
accessory. Field strength of the IH decoupling was 1.2 mT. The contact time
was 5 ms, and the repetition time was 10 s. Spectral width and data points were
21
20 kHz and 8k, respectIvely. Samples were placed in a cylindrical rotor and
spun at 4-5 kHz. Spectra were usually accumulated 100-600 times to achieve
a reasonable signal-to-noise ratio. 15N chemical shIfts were calibrated mdIrectly through external glycine- 15 N (6 = 11.59 ppm; line wIdth = 17Hz) relatIve to saturated 15NH4N03 (6=0 ppm) solution in H 2 0.
Theoret~cal
calculatwn
We employed the fimte perturbation theory (FPT) wIthin the INDO framework for calculating the 15N shIeldmgs. The FPT INDO theory has the advantage of permitting the calculation of the paramagnetic term wIthout reqmrmg
the explicit wavefunctIOns of the excited states and the one-electron excitation
energies, whIch are difficult to obtam wIth high accuracy by the usual semIempIrical MO approxImations. This approach reproduces reasonably the expenmental 13 C chemIcal shifts of L-alanme residues in peptides [16]. According to the FPT INDO framework [17-19], (Jdap(A) (diamagnetic term) and
(Jp(tP(A) (paramagnetic term) are expressed by
a,p=x,Y,Z
where the gauge ongin of the vector potential IS set at the posItIOn of nucleus
A The vectors r v and r A are the position vectors of the electron consIdered
from the nucleus of the atom containing the atomIC orbital (AO) X,, and from
the nucleus A, respectively. P J1V (B) and PI'" (0) are the elements of the density
matnx with and witlJ.out the perturbation due to the magnetic field, B, respectively. The 15N shielding calculatIOn was carned out by a similar procedure to
that reported previously for a 13C shielding calculation [13]. In the calculation,
we adoptedN-acetyl-N' -methylglycme amide as a model compound. The bond
lengths and bond angles proposed by Momany et al. [20] were used. A HITAC
M780H computer at the Computer Center of the Tokyo Institute of Technology was used for the calculation The N - H bond length and N· . ·0 hydrogen
bond length, optimIzed by an ab imtio STO-3G MO calculation, were determmed for N-methylacetamide as a model compound. The calculatIOns were
performed on a HITAC M680H computer at the Computer Center of the Institute for Moleculer Science, Okazaki.
22
RESULTS AND DISCUSSION
15N NMR chem~cal shifts of glycme amide mtrogens of pept~des m the solld
state
A 27.25 MHz 15N CP-MAS NMR spectrum ofL-alanylglycylglycine (L-AlaGly-Gly) in the solid state is shown as a typical example in Fig. 1. 15N CPMAS spectra of the remaining samples were also obtained with sImilar resolutions. SIgnals were also assigned wIth reference to prevIOUS 15N CP-MAS
NMR and solution state 15N NMR data [21].
We use selected glycylglycine (X-Gly-Gly) sequence ohgopepbdes in this
work. It is well known that the 15N chemical shifts of pep tides vary with the
amino acid sequence [22]. For example, 15N chemical shift of the glycine residue depends upon the ammo acid residue linked to the N-termmal ofthe glycine residue. Consequently, we use the glycylglycine sequence ohgopeptides, in
order to neglect the sequence-dependent 15N chemical shifts.
All the 15N chemical shift values of oligopeptides determined from the observed spectra are tabulated in Table 1, together with the geometrical parameters obtained by X-ray dIffraction studies. Some of the geometrical parameters were calculated by usmg the unit cell parameters and fractIOnal coordmates
given in the literature [23-31]. The 15N chemical shifts and the geometncal
parameters of tert-butyloxycarbonyl-glycylglycylglycme-benzylester (Boc-G lyGly-Gly-OBzl) are inferred from the results of Hlyama et al. [31].
Figure 2 shows the plot of the observed 15N chemIcal sh1fts ofGly NH against
the hydrogen bond length N· .. 0 (R N 0)' However, it is found that there is
no clear relationship between RN 0 and 15N chemical shifts. This is dIfferent
Ala GliGly2
Flg 1 A typlcal 27 25 MHz !5N CP -MAS NMR spectrum of L-alanylglycylglycme m the crystallme state
23
TABLE 1
15N chemIcal shIft of glycme resIdue amIde mtrogens for ohgopeptJdes contaInmg glycme as determmed
by 15N CP-MAS NMR (± 02 ppm from 15NH4N03) and the geometrIcal parameters of X-Gly-Gly
peptIdes
15N
chemIcal
shIft", t5
(ppm)
Sample
GlyGly*OH
GlyGlY*'H 2 0 HCI
GlyGly*OH' HN0 3
VaiGlyGly*OH
ProGlyGly*OH
AlaGlyGly*OH
SarGlyGly*OH
TyrGlyGly*OH
BocGlyGlyGly*OBzl
9995
9095
8971
9443
8944
8855
8900
9230
9275 c
GeometrIcal parameters
DIhedral angle (deg)
¢
If!
1571
-796
1656
-1466
-1105
1709
-854
-1039
-778
107
38
1769
-43
1751
1752
-177 5
-1526
-1784
N·
H"'O
N-H
(A)
(A)
N-H ·0
(deg)
Ref
(A)
0
294
330 b
312
305
289
300
306
288
292
197
230
238
219
207
219
223
213
210
102
079
076
093
085
084
085
086
092
158
162
165
152
165
160
167
144
157
23
24
25
26
27
28
29
30
31
"These values are for the Gly resIdues wIth an asterIsk bThe glycIne resIdue partIcIpates In an H
type hydrogen bond cThls value IS converted to the 15NH4N01 reference from ref 31
0
0
-0
~90
'CI
0
0
Go
L
U
0
0
Z
0
'"
0
95
0
28
29
30
31
32
33
N 0 length(A)
FIg 2 Plots of the observed 15N shleldmg of oilgopeptIdes m the crystallme state agamst the
N· . ·0 hydrogen bond length
from the case of 13C chemical shifts of the carbonyl carbons assocIated with
the hydrogen bond which we reported previously [14] where the 13C sIgnals of
the carbonyl carbons are linearly desh18lded WIth a decrease in RN o.
Figure 3 shows the plot of the observed 15N chemical shifts (JobsN) of the
24
~
85
2
::C
:f-
g
0'
to 90
0.
0.
\
o
0
0
\0,
,
,
0'\0
E
,
if)
-0
u
~
.J::
\
0\,
,
95
u
,
\0
:f-
06
07
08
09
10
11
N- H lergth(A)
FIg 3 Plots of the observed
N-H bond length
15N
shIeldmg of ohgopeptldes m the
Cf)
stalhne state agamst the
glycme resIdue in X-Gly-Gly against the N-H bond length (R N - H ) assocIated
WIth the hydrogen bond. It is found that there IS a clear relatIOnship between
these parameters and the decrease of RN - H leads to a linear increase in shIelding. The expression for this relationshIp is
Jobs N
= 39 32R N _H
+ 57.73
Such a trend is very dIfferent from that obtamed from 1JC NMR. Amide 15N
chemIcal shifts are closely related to the length of the N-H bond but are not
related to the N· . ·0 hydrogen bond length. This implies that the ION chemical
shift value gIves useful information about the N-H bond length m the hydrogen bond. It seems that the hydrogen bond angle (L N - H .. ·0) IS also related
to the 15N chemIcal shift
Ab m~tw MO calculatwn relatwnsh~p between the N-H bond length and the
N···O hydrogen bond length
The structure of the two hydrogen-bonded N-methyl acetamides used as a
model system is shown in Fig. 4. At first, the geometrical parameters of an Nmethyl acetamide molecule were optimized usmg the ab ImtIO STO-3G MO
method Next, for two hydrogen-bonded N-methyl acetamides, the bond length
N - H was optimized as a function of the hydrogen bond length N· . ·0 FIgure
5 shows the relationshIp between the mmimized bond length N-H and the
hydrogen bond length N· . ·0, as determined by ab mitio MO calculatIOns
It IS shown that at RN 0 < 2.97 A, an mcrease of RN 0 leads to an mcrease
25
~
H
H
\
N-C----H
/
I
H _C'--- /
\
C
H
H
II
?
:
H
H
/
H-C
I~
I
H
/N""",\
C-H
H C
II
o
/
H
FIg 4 Molecular structure of the two hydrogen-bonded N-methylacetamldes used as model
compounds
10200
N-H
"4
:<:
g
~
10100
:J:
I
Z
10000
24
26
N
28
30
o length (A)
32
34
FIg 5 Plots of the calculated N-H bond length agamst the hydrogen bond length (N·· 0) obtamed usmg the ab InItIO STO-3G MO method
of RN- H • However, at RN 0> 2 97 A, an Increase of RN 0 leads to a decrease
of RN-H • In the samples used in thIS work, the RN 0 values are between 2.85
and 3.30 A. Therefore, it can be said that the R N - H values decrease wIth an
Increase of RN o.
On the other hand, X-ray dIffractIOn studIes, have shown that the R N - H values decrease with an Increase of the RN 0 values [32] From the above results,
It can be SaId that there is an apparent relationship between the hydrogen bond
length RN 0 and the bond length RN- H •
26
15N
shieldmg calculatwns
FIgures 6 and 7 show the calculated isotropIc shieldings (alSO) and the paramagnetic terms of the tensor components (all> a2b a3 .3, from downfield to
upfield) ofGly NH using the model compoundN-acetyl-N' methylglycine (Fig.
8). The calculated values are all expressed in parts per millIon (ppm) with an
OpposIte sign to those in Table 1. Note that the negative sign for the calculated
E -300
(l
~
C
.9 -305
~
(3
~
-310
~
§, -315
f2
;'§
-320
-325
N- H
length(A)
FIg 6 Plots of the calculated 15N shIeldmg agamst the N-H bond length from the FPT INDO
method
E
0.
6
-490
-500
-510
E0.
E0.
':'250
-:140
tS-260
t5-150
-270
160
-280
-170
-290
180
-300
-190
0.
2-
~
0.
-310
08
09
10
11
N-H length (A)
08
09
10
11
N-H length (A)
08
09
10
11
N-H length (Ai
FIg 7 Plots of the calculated 15N shIeldmg tensor components agamst the N - H bond length from
the FPT INDO method
27
FIg 8 Molecular structure of N-acetyl-N' -methylglycme amIde used as the model compound
FIg 9 OnentatlOn of the prmclpal axes of the
mtrogen as determmed m the lIterature [31]
15N
shleldmg tensors of the glycme resIdue amIde
shieldmg denotes deshieldmg, which is similar to the positive sign of the experimental chemical shift values. A shIelding value, or tensor component, is
usually represented as the sum of the diamagnetic and the paramagnetic terms.
However, the amsotropic behavIOur of the shielding tensor can be predominantly explained by the paramagnetIc term, smce the diamagnetic term is
isotropic.
Figure 6 shows the R N - H dependence of the calculated isotropic 15N shIelding
(a,SO) of Gly NH. It IS shown that a decrease of R N - H leads to a very large
mcrease of a,SO; for example, a decrease of 0.4 A m R N - H leads to a calculated
shieldmg mcrease of about 30 ppm. This agrees wIth the observed results. ThIs
ImplIes that the observed shielding increase is a consequence of changes m the
electromc structure through a decrease of the N - H bond length. Figure 7 shows
the R N - H dependence of the calculated principal values of the 15N shielding of
Gly NH. It IS shown that an decrease of R N - H leads to an increase in shielding
of all> a 22 and a33, respectively. The magmtudes for changes of the shieldmg
occur m the order a22> a 33 > all' The dIrection of the prmcipal axes of the Gly
NH 15N shielding tensor components, as determmed by the NMR study of a
Boc-Gly-Gly-Gly-OBzl smgle crystal [31], are shown m FIg. 9. The all component lies approximately along the N-H bond, and the a 33 component lies
approxImately along the N-C' bond. The a22 component IS aligned m the dI-
28
rectlOn perpendIcular to the peptide plane. From this, it can be Said that the
a22 and a 33 components, rather than the all component, are sensitive to R N - H
changes. This may be due to the fact that the a22 component lies approximately
along the direction of the nitrogen lone-pair electrons and the electron density
is very high in this direction. The a33 component lies approxImately along the
N-C' bond into which lone-paIr electrons of the nitrogen atom transfer, and
so the bond order becomes very high. Consequently, the a22 and a33 components are sensitive to changes in R N - H •
Finally, we can draw the following conclusions. The observed 15N shIeldmgs
of amide nitrogens increase linearly wIth a decrease of the N-H bond length
associated with the hydrogen bond; this can be justified by quantum chemical
calculatIOns. The 15N shieldmg is applicable as a means for obtaining direct
information about the nature of the hydrogen bond in the solId state, m addibon to the 13C shielding of the carbonyl carbons m the hydrogen bonds
REFERENCES
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
G A Webb and M WitanowskI, Proc IndIan Acad SCI (Chern ScI) ,94 (1985) 241
G Harblson,J HerzfeldandRJ Gnffm,J Am Chern Soc, 103 (1981) 4752
T A Cross, J A DlVerdl and S T, Opella, J Am Chern Soc, 104 (1982) 1759
T A Cross, M H Frey and S J Opella, J Am Chern Soc, 105 (983) 7471
H G Forster, D Muller and H R Kncheldorf,lnt J BIOI Macromol, 5 (1983) 101
T H Huang, W W Achovchm, R G Gnffm and C M Dobson, BIOchemlstry, 23 (1984)
5933
EO Stejskal, J Schaefer and R A McKay, J Magn Reson, 57 (1984) 471
C N Matthews, R Ludlcky. J Schaefer, EO Stejskal and R A McKay, Ongms LIfe, 14
(1984) 243
T A Cross and S T, Opella, J Mol BIOI, 182 (1985) 367
B S ChOl, J E Roberts, J N S Evans and M F Roberts, BIOchemIstry, 25 (1986) 557
J Schaefer, J R Garbo\\', G S Jacob, T M Forest and G E WIlson, Jr, BIOchem BIOphys
Res Commun, 137 {1986) 736
P L Stewart, K G Valentme and S T Opella, J Magn Reson, 71 (987) 45
(a) A ShOJI, T OzakI, T FUJlto, K Deguchl and I Ando, Macromolecules, 20 (1987) 2441
(b) A ShOJI, T Ozaki, T FUJltO, K Deguchl, SAndo and I Ando. Macromolecules, 22
(1989) 2860
SAndo. I Ando, A ShOJI and T OzakI, J Am Chern Soc. 110 (1988) 3380
(a) I., Zervas, D Borovas and E Gazls, J Am Chern Soc, 85 (1963) 3660
(b) E Wunsch, Synthese von Peptlden, Tell 1 and 2, Georg Thieme Verlag Stuttgart, 1974
(c) N IzumIYa, T Kato, M Ohno and H Aoyagl, PeptIde SyntheSIS. Maruzen, Tokyo. 1974
lAndo, H SaIto, R Tabeta. A ShOJI and T OzakI, Macromolecules, 17 (1984) 457
R DltchfIeld. D P MIller and.J A Pople, J Chern Phys. 54 (1971) 4186
P D Elhss, G E MaCiel and J M Mciver. Jr , J Am Chern Soc, 94 (1972) 4069
I Ando and G A Webb, Theory of NMR Parameters, AcademIC Press, London, 1983
FA Momany, R F McGUIre, J F Yan and H A Scheraga,.J Phys Chern, 75 (1971) 2286
M Wnanowsk!. L Stefamak and G A Webb, Annual Reports on NMR Spectroscopy. Vol
18. AcademIC Press, London, 1986
H R Kncheldorf,Org Magn Reson, 15 (2) (1981) 162
29
23
24
25
26
27
28
29
30
31
32
A B Blswas, E W Hughes, B D Sharma and J N WIlson, Acta Crystallogr , Sect B, 24
(1968) 40
T F Koetzle, W C Hamliton and R Parthasarathy, Acta Crystallogr, Sect B.28 (1972)
2083
S N Rao and R Parthasarathy, Acta Crystallogr , Sect B, 29 (1973) 2379
V L Murah and E Subramaman.lnt J Peptlde Protem Res, 27 (1986) 472
V Lahta. E Subramaman and R Parthasarathy,lnt J Peptlde Protem Res. 27 (1986) 223
V Lahtha. E Subramaman and J Bordener, IndIan J Pure Appl Phys. 23 (1985) 506
JP Glusker,HL Carrell,HM Berman,B GallenandRM Peck,J Am Chern Soc,99(2)
(1977) 595
W M Carson and M L Hackert, Acta Crystallogr , Sect B, 34 (1978) 1275
Y HlYama, C NlU, J V SIlverton. A Bavoso and D A TorchIa, J Am Chern Soc, 110
(1988) 2378
R Taylor, 0 Kennard and W VersIchel, Acta Crystallogr , Sect B.40 (1984) 280