Solution conformation of a pectin fragment disaccharide using molecular modelling and nuclear magnetic resonance Soizic Cros, Catherine Herv~ du Penhoat, Nadia Bouchemal, Hamid Ohassan Laboratoire de Chimie, ENS, 24 rue Lhomond, 75231 Paris cedex 05, France
Anne Imberty LSO-CNRS, Facult~ des Sciences et Techniques, 2 rue de la Houssini~re, 44072 Nantes cedex 03, France
and Serge P~rez* lngknierie Mol~culaire, INRA, BP 527, 44026 Nantes cedex 03, France
(Received 12 March 1992; revised 20 July 1992) In the present study, the conformational behaviour of methylated pectic disaccharide 4-O-ct-o-galactopyranurosyl 1-O-methyl-ct-o-galactopyranuronic 6,6'-dimethyl diester 1 has been completely characterized through combined n.m.r, and molecular modelling studies. The 1 H J H n.O.e, across the glycosidic bond was measured by both steady-state and transient 1D and 2D experiments. In parallel, the complete conformational analysis of the disaccharide has been achieved with the M M3 moIcular mechanics method. The conformation of the pyranose ring is confirmed by the excellent agreement between the experimental and calculated intracyclic scalar coupling constants. The iso-energy contours displayed on the 'relaxed' map indicate an important flexibility about the glycosidic linkage. There is no significant influence of the methoxyl group on the conformational behaviour of the disaccharide. The theoretical n.m.r, data were calculated taking into account all the accessible conformations and using the averaging methods appropriate for slow internal motions. 3Jc_n coupling constants were calculated using an equation suitable for C-O-C-H segments. The agreement between experimental and theoretical data is excellent. Within the potential energy surface calculated for the disaccharide, several conformers can be identified. When these conformations are extrapolated to a regular polymer structure, they generate pectins with right- and left-handed chirality along with a two-fold helix. These different types of helical structure are the result of small changes in conformation, without any drastic variation of the fibre repeat. Keywords: Pectin fragment; molecular modelling; conformation;n.m.r.
Introduction Pectins are a family of polysaccharides which are a major constituent of the cell wall of many higher plants. They contribute to the firmness and structure of plant tissues both as part of the primary cell wall and as the main middle lamella component involved in intercellular adhesion. The prototypical polysaccharide is based on a linear (1,4)-linked ~-D-galacturonan backbone (pectic acid) and may be partially methyl esterified (pectinic acid). This backbone is also interrupted by (1,2)-linked C~-L-rhamnopyranose and carries a proportion of lateral chains containing neutral sugars. Pectins have been used extensively as gelling agents over a long period of time, and less widely, as texturizers, emulsifiers, thickeners and stabilizers. The reasons why such a biopolymer, unlike others, develops a jelly in the presence of water, sugar and acid are still not elucidated. Nevertheless, the implications of structural details such as the degree of methoxylation, the distribution of * To whom correspondenceshould be addressed. 0141-8130/92/060313-08 © 1992 Butterworth-HeinemannLimited
rhamnopyranoses, the molecular weight of the polymer on some of the functional properties of pectins have been clearly demonstrated (for review, see ref. 1). An understanding of the mechanism of pectin gelation and how such a process is related to the molecular structure would make it possible to systematically control and manipulate the gel-rheological conditions through minute and controlled changes of molecular features. The characterization of the molecular features of pectins, in the solid state, in the gel state and in solution, has not been studied in as much detail as other polysaccharides. This may be explained by the occurrence of structural defects (which depend on the plant source and the conditions of preparation) and the concomitant lack of stereoregularity in pectins. The polyelectrolyte character of the chain in its acidic form along with the conformational transitions induced by monovalent and divalent ions complicate further the structural characterizations. Fibre diffraction analyses suggested that pectic acid and sodium pectate have similar three-fold right-handed structures 2'3. However, the occurrence of a left-handed single-stranded helical
Int. J. Biol. Macromol., 1992, Vol. 14, December
313
Conformation of pectin frayment disaccharide: S. Cros et al. structure has been also suggested 4. Structures deduced from solid-state measurements do not necessarily represent the hydrated and diffuse state of the gel network. Conformational analysis of disaccharides is a prerequisite to gaining useful information for modelling three-dimensional crystal structures of polysaccharides based on the independent residue approximation 5. Also, it helps to gain some insight into characteristics of the chain in its disordered state, such as radius of gyration, persistence length or the occurrence of local helical regions that might be appropriate for further ordering 6. In the present study, the conformational behaviour of methylated pectic disaccharide 4-O-a-D-galactopyranurosyl 1-O-methyl e-D-galactopyranuronic 6,6'-dimethyl diester 1 (see Figure 1) has been completely characterized through combined n.m.r, and molecular modelling studies. Separation of the anomers was possible in the case of these methylated derivatives and the spectral dispersion of the n.m.r, spectra considerably enhanced. Moreover, due to a much lower gyromagnetic moment, the relaxation rate from 2H is only 6% of that of 1H and this contribution can be generally neglected in an analysis of relaxation data of deuterated compounds. The present investigation complements a recent report 7 on the conformational analysis of the (1,4)-linked a-I>galacturono disaccharide 2.
J \ • C-1 c-~
Materials 4-O-c~-I>galactopyranurosyl 1-O-methyl-d 3 e-i>galactopyranuronic 6,6'-dimethyl-d 6 diester, 1, was prepared as follows. A total of 212.6mg ( ~ 0 . 6 m m o t ) of 4-O-c~-Dgalactopyranurosyl galactopyranuronic acid, 2, were refluxed overnight in 15 ml of MeOH-d4 in the presence of a trace of HC1 (obtained from 25 #1 of acetylchloride according to a literature procedureS). The solvent was evaporated under reduced pressure and the residue chromatographed over a S i O 2 column (Merck 9385 Kieselgel 60 - 5.5 x 20cm) using a 4 / 2 / 3 / 1 mixture of chloroform/acetone/methanol/water as the eluent. A minor fraction contained 8.5 mg of I while the major fractions yielded a mixture of products (220 mg) which were not identified. Product 1 was lyophilized in D 2 0 (99.96%) twice to remove residual water, degased and sealed in a 5 mm n.m.r, tube under argon.
Nomenclature The recommendations and symbols proposed by the Commission on Nomenclature 9 are used throughout this paper. A schematic drawing of the disaccharide, along with the labelling of the atoms, is given in Figure I. The relative orientation of a pair of contiguous residues about the glycosidic linkage is described by a set of two torsional angles: = O-5'_C-1' O-1'_C-4 tp = C-1'_O-1'_C-4_C-5 The magnitude of the valence angle z ( C - I ' _ O - I ' C-4) at the glycosidic bridge is also an important variable. As for the methyl-esterified group (see Figure 1 for the atomic nomenclature used), its orientation is described by two torsional angles X = O-5_C-5 C-6_O6 and o9 = C-5 C-6_O-6 C-7.
314
Int. J. Biol. Macromol., 1992, Vol. 14, December
6o
C-1'
C-5'
C-2"
\c,z>~ /
1
O
1
Experimental
~
2
Figure 1 Schematic representation of the methyl-esterified galacturonate disaccharide 1 and the acid disaccharide 2. The labelling of the torsion angles and some atom names of interest have been indicated on a diagram of 1
N,m.r.
1H-(400.13 MHz) and 13C_n.m.r. ( 100.6 MHz) spectra were recorded with a Bruker instrument operating in the Fourier-transform mode at 296 K. Both the 1H and 13C chemical shifts of 1 have been previously assigned ~° from homo- and heteronuclear correlation spectroscopy, respectively. The 3JH_I_Iwere extracted from a spectrum with a digital resolution of 0.12 Hz/pt. Two sets of steady-state n.O.e, difference spectra were obtained with irradiation times and pulse intervals greater than five times the longest TI for the sugar signals. 512 difference FIDs were accumulated and multiplied with a 1 Hz line broadening factor prior to Fouriertransform to reduce noise. Values from the two data sets were averaged. Both 13C and 1HT~ measurements were acquired with the inversion-recovery sequence (180-t-90-FID) and relaxation times were calculated with the Bruker 7"1 routine. In the case of selective 1H relaxation times, the duration of the soft 180 ° pulse, which was performed with the Dante sequence, was 20 ms. The decoupler offset was set to the frequency of the centre of either the H-4 or the H-I'/H-1 multiplet. In this latter case the resonances of these two protons overlap to form a triplet with the low-field peak corresponding to the H-I' doublet while the high-field transitions are part of the H-1 doublet. Thus, this selective 7"1measurement allowed simultaneous measurement of the diagonal elements of the relaxation matrix for the H-I' and H-1 protons through the recovery curves of the two outer branches of the triplet. Phase-sensitive NOESY ~~ were acquired with mixing times of 0 and 1 s. A 20 ms variable delay was introduced
Conformation of pectin fragment disaccharide: S. Cros et al. at the beginning of the mixing time in order to suppress J-peak transfer. The recycle time was set to five times the longest T1 to ensure that the normalized NOESY volume matrix would be symmetrical. A total of 512 x 1K data matrices were obtained and zero-filled to 1K x 1K. Prior to Fourier transformation the first data file was halved to reduce ta ridges, and n/2-shifted sine-squared weighting functions were applied t2. NOESY crosspeak intensities were evaluated from the summed o91 subspectra contributing to a specific signal. The values of the normalized aij and ai~ elements were averaged before back-transformation to the relaxation matrix ~3 with in-house software. Separation of the H-I' and H-1 contributions to the diagonal volumes for both 0 and 1 s mixing times was governed by two criteria: no more than a 15% deviation from the mean value of the zero mixing time diagonal elements was tolerated and a 2% fit to the inverse of the selective T~ after back-transformation was sought. This H - I ' / H - 1 ratio was also used in all the calculations of experimental relaxation values. The n.O.e, buildup for the H - I ' / H - 4 interaction was also measured with the (180sel-90-FID) sequence with the decoupler offset alternatively on the H-I' frequency or in a region where there were no signals. The variation in intensity for the signal of the H-4 proton was monitored by the resulting difference spectra with mixing times of 0.12, 0.16, 0.2 or 0.24 s. 512 difference FIDs were recorded.
Computational methods Molecular modelling. The starting geometry of the disaccharide was derived from the database of threedimensional structures, M O N O B A N K 14. Geometry optimizations were performed using the molcular mechanics program MM315'16. The total energy takes into account the stretching, bending, stretch-bending, torsional and dipolar contributions as well as van der Waals interactions. This force field, which has been compared with semiempirical methods 17, contains a correction for the anomeric effect without further parametrization 18. The MM3 force field has been extensively used for the study of carbohydrate structures 18-22. Partial atomic charges were not used since MM3 calculates dipole-dipole interaction energy from stored bond moments. The dielectric constant (e) was set to 4.0 for all calculations, as found most appropriate from studies of the Dglucopyranose ring in comparison with crystal structures 18. The block diagonal minimization method, with a default convergence criterion of 3.6cal/mol, was used for grid-point optimization. The preferred orientation of the methyl-ester exocyclic group was assessed with the DRIVER option of the MM3 program ~5'16, for both the monosaccharide and disaccharide 1. The torsion angles X and co were gridded by 20 ° intervals on the whole angular range. The lowest energy conformation was then used as starting point for calculating the relaxed map of the disaccharide about the q~ and u? glycosidic torsion angles. Like the other atoms, those belonging to the methyl ester exocyclic group were allowed to adjust. The relaxed map was computed using a 20 ° increment step for qb and ~ angles for the whole angular range. For each point of this grid, the optimized atomic coordinates and the corresponding energy were stored for further calculations of the n.m.r, observables.
Calculations of theoretical n.m.r, parameters. Averaging is required to correctly predict properties for a population
of conformations. Using the Boltzmann distribution, the relative population of each conformer is given by the expression: Pi - exp( - E i / k T ) / ~" exp( - Ei/kT ) For a n.m.r, parameter M depending in a non-linear way on an associated structural parameter S, the average value can be computed by
( M ) = ~,Pif(Si) Averaged coupling constant (3Jc_H) across the glycosidic bonds were calculated using the equation appropriate for C _ O _ C H segment 23. Equations for the calculation of the several (3JH_H) of the rings were established according to a known method 24. For the NOESY conditions, the ensemble average cross-relaxation matrix ( R ) was calculated from the ( r - 6 ) values following previously published procedures 13. The spectral density functions used in these calculations are appropriate for a molecule undergoing isotropic tumbling. It is also assumed that the internal motion is slow compared to the overall motion. For steady-state conditions, the ensemble average %n.O.e. can be calculated by solving a series of simultaneous equations 15'26. In order to obtain the best fit for the diagonal elements, Rii, of the relaxation matrix, a leakage factor of 2 0 % was included in the calculations.
Helical parameters. When subjected to the constraints imposed by the helical symmetry of macromolecular chains, equivalent monomeric units should occupy equivalent positions about the molecular axis. This is achieved when the (qb, ~g) angles are the same at every linkage; the secondary structure is described in terms of a set of helical parameters: (n, h), n being the number of residues per turn of the helix, and h being the translation of the corresponding residue along the helix axis. The chirality of the helix is defined by the sign of n. Arbitrarily, a right-handed helix will correspond to positive values of n; conversely, negative values of n will be designated left-handed helices. Whenever the values h = 0 or n = 2 are intercrossed, the screw sense of the helix changes to the opposite sign. These parameters are calculated for all combinations arising from rotations about the glycosidic linkage following an algorithm reported previously27; the corresponding iso-n and iso-h contours are plotted superimposed on the potential energy surfaces.
Results and discussion Molecular modelling The orientation of the methyl-ester exocyclic group has been established as a function of the torsion angles co and Z (Figure 2). For the monosaccharide, the trans orientation (180 °) of the torsion angle co is always preferred. The Z torsion angle does not exhibit a clear conformational preference. For this reason, the orientation of the methyl-ester group was studied for the disaccharide. In this study, the exocyclic group of the reducing end was selected, for different conformations of the glycosidic bond. The minimum with Z = 140° and o) = 180 ° is the lowest energy one. In all cases, the co angle retains its 180° favoured conformation. The second angle presents two possible orientations ()~ = 0 ° and Z = 140°) . The conformation (co = 180 °, g = 140°) is always the lowest energy one (see Figure 2). This conformation was retained
Int. J. Biol. Macromol., 1992, Vol. 14, December
315
Conformation of pectin fragment disaccharide: S. Cros et al. 360 t
'~
(80, 140) and (120, 160). The most stable conformer is I, conformers II and III being 0.8 kcal/lnol and 2.6 kcal/mol, respectively. The relaxed map of the acidic disaccharide 2 has been calculated, it is not shown here because it does not exhibit any significant changes with respect to that corresponding to 1. These two maps look like the one computed previously for the acidic disaccharide 2 using the M M 2 C A R B program 7. The only difference between the two investigations is the prediction of a fourth conformer ( 0 = 8 0 ° and q~= 300 ° ) by the M M 2 C A R B force field, being 5 kcal/mol higher than the most stable conformer. From the present study, we found that such a conformation would have an energy of approximately 9 k c a l / m o l in relative energy which explains why it is not represented here. Therefore, the minute differences are mainly due to slightly different force fields.
\
300t 240-
CO 180-
"o
120
-120
-60
0
60
120
180
240
Z Figure 2 Iso-energy map of the orientation of the methyl-ester exocydic group as a function of the 09 and Z torsion angles. The exocyclic group is the one at reducing end of the disaccharide. For each set of ~o and Z values, the whole geometry of the disaccharide has been optimized. Isoenergy contours are drawn with interpolation of 1 kcal/mol above the minimum of the map
N.m.r. data Both 1H and 13C chemical shifts of the precursor of 1 have been assigned. Methylation s h i f t s 7 - 2 8 of +7.99 and - 0.44 p.p.m, are observed for the C- 1 and H- 1 signals of 1 in agreement with reported values (Table •)29. A correlation time of 1.2 × 10- los for the overall motion was obtained from the average 13C T1 of the methine carbons and a p r o t o n - c a r b o n distance of 0.11 nm with the equation,
1
h272cTzHf
Zc
3~c TI -- 10r~_H\l----" + (COB-- COC)2Z~+ 1 + CO2Z2 240
6Zc ) + 1 + (con + coc)2z2
200" Table 1 400.13MHz 1H- and 13C-n.m.r. chemical shift data for methyl-esterified galacturonic acid dimer 1 in D20 (6nH°D = 4.80 p.p.m, and 6c_v = 100.93 p.p.m.)
160-
,t,
6n
120-
H-1 H-2 H-3 H-4 H-5
80-
40-
0
I
0
40
'
I
I
80
120
'
I
160
'
I
200
H-I' H-2' H-3' H-4' H-5'
4.96 3.72 3.92 4.33 5.09
C-1 99.99 C-I' C-2 67.98 C-2' C-3 68.34 C-3' C-4 79.39 C-4' C-5 70.28 @5' C-6, @6' 171.86, 170.95
100.93 68.11 68.99 70.38 71.70
=
240
0 Figure 3 'Relaxed' potential energy surface of the methylesterified galacturonate dimer as a function of the • and q~ torsion angles. Isoenergy contours are drawn with interpolation of 1 kcal/mol above the minimum of the map
as the starting conformation of the methyl-ester exocyclic group in the study of parent disaccharide. The relaxed map of disaccharide ! computed as a function of the qb and qu torsion angles is given in Figure 3. There is only one main low energy region, rather extended along both the • and qJ axis. Within the 8 kcal/mol limit, three energy minima (I III) are found on the relaxed map with qb and qJ values of (80, 100),
316
4.95 3.84 4.00 4.44 4.70
6C
Int. J. Biol. Macromol., 1992, Vol. 14, December
Table 2 400.13MHz observed and calculated 3JH. H and 3Jc. H coupling constants of methyl-esterified galacturonic acid 1. The calculated values were obtained by averaging over the whole relaxed map
Proton
JExp
JCalc
JH I'-H 2' Ju Z-H 3' JH 3'-H 4' JH 4'H-5' JH m-2 Jn 2H-3 JH 3H-4 JH_nH-5 JH I'-C 4 Jc I'-H 4
4.0 10.5 3.4 1.5 3.8 10.5 3.4 0.7 3.7a 4.9"
3.9 9.6 3.1 1.1 3.8 9.8 3.0 1.1 3.4 4.7
a Ref. 7
Conformation o f pectin f r a g m e n t disaccharide: S. Cros et al.
Experimental 3JH_H v a l u e s a r e g i v e n i n Table 2 a l o n g with the corresponding data which have been calculated for the 4C 1 conformation of the galactopyranuronosyl r i n g a c c o r d i n g t o H a a s n o o t et al. 24. T h e s m a l l m e a n d e v i a t i o n b e t w e e n t h e t w o sets o f p a r a m e t e r s , 0 . 4 H z , i m p l i e s t h e c l a s s i c 4C1 f o r m f o r t h e r i n g s o f 1. S t e a d y s t a t e % n.O.e, m e a s u r e d f o r t h e d i s a c c h a r i d e 1 a t 400.13 M H z a r e r e p o r t e d i n Table 3. A m o n g t h e 12 p r o t o n p a i r s w h i c h y i e l d e d a n.O.e, e n h a n c e m e n t , t w o correspond to inter-residue interactions. The H-I'/H-4 p a i r g i v e s a s t r o n g n.O.e, effect ( 1 7 % ) w h i l e t h e H - 5 ' / H - 2 g i v e s a w e a k effect ( 4 % ) . T h e s e r e s u l t s a r e in q u a l i t a t i v e a g r e e m e n t w i t h t h e n.O.e, d a t a o b t a i n e d o n c o m p o u n d 2 a t 300.13 M H z 7.
The NOESY spectrum ofgalacturonic acid disaccharide 1 is d i s p l a y e d i n Figure 4. T h i s p h a s e - s e n s i t i v e N O E S Y spectrum acquired with a mixing time of 1 s contains the c r o s s - p e a k s c o r r e s p o n d i n g t o all t h e i n t e r a c t i o n s m e a s u r e d in the steady-state spectra. Some additional weak inter-residue NOESY peaks can be observed.
Solution conformation o f the disaccharide The inter-residue conformation can be attained through b o t h t h r e e - b o n d p r o t o n c a r b o n scalar coupling, 3Jc_H, across the glycosyl linkage and inter-residue crossrelaxation rates. These heteronuclear coupling constants h a v e b e e n m e a s u r e d i n t h e c a s e o f 2 7. T h e s e d a t a c a n b e compared with the ones calculated for disaccharide 1
3 Steady-state n.O.e, values % observed (italic) a n d calculated for the methyl-esterified galacturonic acid 1. T h e calculated values were o b t a i n e d by averaging over the whole of the relaxed m a p
Table
Saturated spins
H-I'
H-2'
(23.6) 19.7 (2.5) 4.1 -
H-I' -
(11.2) 12.2 -
H-2' H-3'
-
H-4'
H-3'
(4.3) 3.3
(8.0)
-
8.3
. .
.
.
H-3
. . (17.4) 15.9
H-4
. .
.
-
-
-
(4.4) 4.9
. .
.
.
5.8
-
-
-
-
-
-
-
-
(3.6) 3.1 (7.0) 8.3 (9.4) 11.8 (8.5) 11.3
.
.
.
(18.7) 20.8
. .
-
-
(4.1)
. .
.
.
.
-
-
. .
. .
.
-
-
(7.0) 15.4
. .
.
. .
-
-
-
-
. .
.
.
-
. .
H-5
.
.
-
H-2
H-5
13.8
10.9
.
H-4
(16.8) .
. (8.6) 7.6 (9.3)
12.3
.
H-3
.
.
(11.5)
H-2
. .
-
(7.7)
H-1
.
. . . (11.8) 13.0 -
-
-
H-1
. .
.
12.6
-
H-5'
. .
H-5'
H-4'
.
.
(5.2) 7.4
(11.0) 12.0 (9.6) 10.9 -
Table 4 Relaxation matrix elements o b s e r v e & (italic) a n d calculated for the methyl esterified galacturonic acid 1: The calculated values were o b t a i n e d by averaging over the whole of the relaxed m a p Protons
H-I'
H-2'
H-Y
H-4'
H-5'
H-1
H-2
H-3
H-4
H-5
H-I'
(0.898) 0.974 -
(0.126) 0.117 (0.552) 0,594 . .
(0.018) 0.007 (0.019) 0.026 (0,764) 0.749 -
(0.002) 0.002 (0.014) 0.008 (0.102) 0.102 (0.807) 0.726 -
(0.007) 0.009 (0,016) 0.006 (0.087) 0.076 (0.080) 0.098 (0.784) 0.815
-
-
(0) 0,001 (0.001) 0.000 (0.00t) 0.001 (0.001) 0.000 (0.002) 0.001 (0.514) 0.508
(0,003) 0.003 (0,016) 0.001 (0.026) 0.012 (0.013) 0.001 (0.041) 0.040 (0.102) 0.106 (0.680) 0.690
(0.009) 0.003 (0,0t7) 0.000 (0) 0.001 (0) 0.000 (0.001) 0.003 (0.022) 0.007 (0.009) 0.026 (0.809) 0.777
(0.164) 0.153 (0.012) 0,003 (0.002) 0.002 (0.001) 0.001 (0.007) 0.011 (0,002) 0.002 (0.019) 0.008 (0.104) 0.102 (1.267) 1.113
(0.009) 0.006 (0.010) 0,000 (0) 0.001 (0) 0.000 (0.001) 0.001 (0.006) 0.010 (0.015) 0.006 (0.106) 0.096 (0.117) 0.093 (0.745) 0.737
H-2' H-3' H-4'
-
H-5'
-
H-1
.
.
H-2
.
.
.
H-3 .
.
.
.
.
. .
.
. .
.
.
.
. .
. .
.
.
.
n-4 H-5
.
.
.
. . .
. .
.
.
.
. .
.
. .
a The NOESY peaks intensities were obtained following ref. 37
I n t . J. Biol. M a c r o m o l . ,
1992, Vol. 14, D e c e m b e r
317
Conformation of pectin fragment disaccharide: S. Cros et al. with proper averaging over all the conformations of the relaxed map. As shown in Table 2, the mean deviation between the experimental values and those calculated for the averaged structure is excellent, < 5%. Steady state %n.O.e. were averaged from all the conformations of the relaxed map. In Table 3 these calculated values are compared with the experimental ones. Calculated values remain in the majority of cases larger than experimental ones. It appears that, although in at least 10% of the cases partial saturation of neighbouring resonances was observed, incomplete saturation was often a problem. When taking in account all the values (23 n.O.e, data) the standard mean deviation
is 27%. This value drops to 19% if the n.O.e, oberved for H-2 when H-1 is saturated (together with H-I') is omitted. These mean deviations are almost identical for the intra-residue n.O.e. (19 data) and the inter-residue n.O.e. (four data). With the exception of the H - l / H - 2 pair, the ensemble averaged n.O.e, values are in good agreement with those derived from experiments. As for the NOESY experiments, the comparison between observed and ensemble averaged values has been evaluated for the relaxation matrix ( Table 4). Comparison of relaxation matrix (or NOESY intensities) is preferred to comparison of calculated and estimated proton distances 3°. When taking into account all matrix elements
5 5' 4
4'
\ I
5.0
I
41~
~l.
4~8
I"
4.2
3J.
~.0
PPH
o
o
(I
0
t
o
~
0o
O
t4
o ¢00
.q
V4,1'
.~
¢
:
,¢ 0!
r !
,) I
~ " - - -
-
-
-
~
:
j'
--
'
iV1"4 O
~ O
"-
Figure 4 Phase-sensitive NOESY spectrum of esterefied galacturonic acid dimer 1 recorded with 1 s mixing time in D20. The cross-peaks corresponding to the H-I'/H-4 interaction are labelled and the 1 D spectrum is given above
318
Int. J. Biol. Macromol., 1992, Vol. 14, December
Conformation of pectin fragment disaccharide: S. Cros et al. with observed values greater than 2% (i.e. stemming from measurable N O E S Y volumes, 22 data) the mean deviation is 11.8 %. F o r both galactopyranuronosyl rings, the N O E S Y experiments were compatible with the classic 4C 1 form. As regards the inter-residue data, the mean deviation is 21% (three data > 2%). However, it should be noted that in the case of the two strongest interactions, H - I ' / H - 4 (0.164) and H - 2 / H - 5 ' (0.041), the data are in excellent agreement with the modelled values, 0.153 and 0.040 respectively. The latter H - I ' / H - 4 cross-relaxation rate was measured by three methods: a combination of steady-state n.O.e, values and selective relaxation rates (0.205s-1), N O E S Y spectra (0.164s-1), and n.O.e. buildup (0.201s-~). It should be noted that the data obtained in the latter approach indicated that the n.O.e. buildup was not fast compared to the selective relaxation rate. Carver 31 has recently summarized the treatment appropriate for extracting internuclear distances from relaxation data of carbohydrates in the presence of internal motion. In the case of internal motion that is slow compared to the overall motion the approach described above is justified. However, when the internal movements are fast compared with the molcular tumbling, appropriate spectral density functions 32'33 are required. Such rapid motion has been described recently for torsion about the glycosidic linkage of sucrose 34. In the intermediate case, where internal motion is on a timescale comparable with the overall tumbling, the full dipolar interaction (both angular and distance) would have to be evaluated from the dynamics trajectory. A reliable protocol for evaluating the rate of internal motion in carbohydrates for all of these cases is being studied.
..... .- ..~"... ~'2;;;;7"::::::!::~7::... ........ . . " ' " ......... "..;.....--:::.:.:.:?.'.!!:ii:!!::... ." ..' . ................~"--.~':~;;::~i:-..
240 .
. ~
20o .
. . .
16o-.
...-~i.."i ...........
"................ ':.::!~i!!~!!!~..
........ .......
~ \\~.i
i :::..i!i; .....
8o ~i;.i~)., -..
"-
~:i::::...
ii.?..~ ~
0
0
A
40
80
B
120
160
C
200
240
D
Helical conformation of the polysaccharide In agreement with the experimental evidence provided by the n.m.r, data, the conformational analysis of the disaccharide in solution established that a large number of conformations are available for the molecule. The potential energy surface of Fiyure 3 is therefore representative of the conformational behaviour of the disaccharide unit, and may be used to model some features associated with the parent polysaccharide. Among these, the description of the stable helical conformation of a stereoregular polymer is straightforward. This has been done, using a rigid residue approach. The iso-n contours corresponding to n = 2, n = 3, n - - - 3 and n = 4 along with the iso-h contour have been superimposed on the low energy region (Figure 5). This shows that the m a p is divided into regions of left-handed and right-handed chirality. The chiral transition occurs through conformations of q~, W yielding values of n = 2. Since the iso-h in the low energy regions have the highest values, a tendency towards m a x i m u m extension is exhibited. Four stable conformers, generating integral helices can be found on the potential energy surfaces (A: n=-3, h = 0 . 4 4 n m ; B: n = 2 , h = 0 . 4 3 n m ; C: n = 3 , h = 0 . 4 4 n m and D: n = 4 , h = 0 . 4 4 n m ) . These four helices which all exhibit maximum extension are displayed in Figure 5. Helices A, B and C have glycosidic torsion angles in close correspondence with energy minima: III, II and I, respectively. These different models can be compared with experimental data obtained on both acidic and esterified pectins. Fibre diffraction performed on N a + and Ca 2+ pectate gels yielded models 2,3 of three-fold, right-
Figure 5 Iso-n and iso-h contours for the methyl-esterified galacturonic acid dimer 1 superimposed on the low energy region of the potential energy map. Four low energy conformations which would generate integral helices have been indicated by a, and labelled A (n = - 3, h = 0.44 nm), B (n = 2, h = 0.43 nm), C (n = 3, h = 0.44 nm) and D (n = 4, h = 0.44 nm). These four helices have been represented using projection parallel and orthogonal to their axis
handed helices repeating in 1.3 nm (n = 3, h = 0.43 nm). On the basis of circular dichroism, calcium stoichiometry and competitive inhibition, a two-fold helix having a repeat of 0.870 nm (n = 2, h = 0.435 nm) was proposed for calcium gel 35'36. More recently, pectins polygalacturonate single chain images were visualized through electron microscopy; the occurrence of a left-handed helix with a 1.3 nm period (n = - 3, h = 0.43 nm) was proposed 4. The results of our calculations indicate that each type of helical conformation seems to be almost equally preferred. The calculated energy differences indicate that interconversion between low energy conformers are feasible. It is also clear that neighbouring of ions, solvent or other macromolecules can easily induce a conformational change. The most striking features are the facts that these changes in conformation can occur without any noticeable variation of the fibre repeat.
Int. J. Biol. Macromol., 1992, Vol. 14, December
319
Conformation of pectin fragment disaccharide: S. Cros et al.
Conclusion The present work has been performed on the dimeric segment of methoxyl pectins and pectins. The conformational behaviour of the methylated pectic disaccharide has been characterized through combined n.m.r, and molecular modelling studies. The agreement reached between observed and calculated data is a satisfactory test of the validity of the conformational energy surface which may be used to safely model some of the structural features of the polymers. The relevance of the stable conformers with respect to helical structures of pectins indicates that both right- and left-handed single-stranded helices may occur, along with a two fold-helix. These different types of helical structure occur for small conformational changes without any drastic variation of the fibre repeat. Of a certain interest is the conclusion that there is no significant influence of the metho×yl group on the conformational behaviour of both the disaccharide and the polysaccharide. Therefore, it may be concluded that the well known influence of methoxylation on functional properties of pectin may be arising from intermolecular origins.
9 10 11 12 13 14 15 16 t7 18 19 20 21 22 23
Acknowledgements
24
The authors are indebted to D r Kevin Hicks (USDA, Philadelphia) who kindly provided c o m p o u n d 2.
25
References 1
8
Walter, R. (Ed.) 'The Chemistry and Technology of Pectin', Academic Press, London, 1991 Walkingshaw, M.D.andArnott, S.J.Mol.Biol. 1981,153, 1075 Walkingshaw, M. D. and Arnott, S. d. Mol. Biol. 1981,153, 1055 Ruben, G. C. and Bokelman, G. H. in 'Pectin's polygalactouronique single sugar chain helix visualized in Pt/C Replicas', San Francisco Press, California, 1987, p 966 Brant, D. A. Q. Rev. Biophys. 1976, 2, 232 P6rez, S. in 'Methods in Enzymology', Vol. 203, Academic Press, New York, 1991, p 510 Hricovini, M., Bystricky, S. and Malovikova, A. Carbohydr. Res. 1991, 220, 23 Gmunder, J. Helv. Chim. Acta 1953, 36, 2021
320
Int. J. Biol. Macromol., 1992, Vol. 14, December
2
3 4 5 6 7
IUPAC-IUB Commission on Biochemical Nomenclature. Arch. Biochem. Biophys. 1971, 145, 405
26 27 28 29 30 31 32 33 34 35 36 37
Ohassan, H. DEA dissertation, Paris VII, 1989 Bodenhauseu, G., Kogler, H. and Ernst, R. R. J. Maon. Resort. 1984, 58, 370 Neuhaus, D. and Williamson, M. in 'The Nuclear Overhauser Effect in Structural and Conformational Analysis' VCH Publishers, 1989, p 292 Olejniczak, E. T., Gampe, R. T. and Fesik, S. W. J. Magn. Reson. 1986, 67, 28 P6rez, S. and Delage, M. M. Carbohydr. Res. 1991, 212, 253 Allinger, N. L., Yuh, Y. H. and Lii, J.-H. J. Am. Chem. Soe. 1989, 111, 8551 Allinger, N. L., Yuh, Y. H. and Lii, J.-H. J. Am. Chem. Soe. 1990, 112, 8293 Gundertofte, K., Palm, J., Petterson, I. and Stamvik, J. Comput. Chem. 1991, 12, 200 French, A. D., Rowland, R. S. and Allinger, N. L. in 'Computer Modeling of Carbohydrate Molecules', Am. Soc., Washington, DC, 1990, p 120 French, A. D. Carbohydr. Res. 1989, 188, 206 French, A. D. in 'Cellulose and Wood-Chemistry and Technology', Wiley, New York, 1989, p 103 French, A. D. and Brady, J. W. in 'Computer Modeling of Carbohydrate Molecules', Am. Soc., Washington, DC, 1990, p 1 Dowd, M. K., Reilly, P. J. and French, A. D. J. Comp. Chem. 1991, 13, 102 Tvaroska, I., Hricovini, M. and Petrakova, E. Carbohydr. Res. 1989, 189, 359 Haasnoot, C. A. G., De Leeuw, F. A. A. M. and Altona, C. Tetrahedron 1980, 36, 2783 Noggle, J. H. and Schirmer, R. E. in 'The Nuclear Overhauser Effect', Academic Press, New York, 1971 Cumming, D. A., Carver, J. P. Biochemistry 1987, 26, 6664 Gagnaire, D., P&ez, S. and Tran, V. Carbohydr. Res. 1980, 78, 89 Rinaudo, M., Ravanat, G. and Vincendon, M. Macromol. Chem. 1980, 121, 1059 Perlin, A. S. and Casu, B. in 'The Polysaccharides', Academic Press, London, 1982, p 133 Clore, G. M. and Gronenborn, A. M. J. Magn. Res. 1989, 84, 398 Carver, J. 'Current Opinion in Structural Biology', 1991, 1, 716 Lipari, G. and Szabo, A. J. Am. Chem. Soc. 1982, 104, 4546 Kessler, H., Griesinger, C., Lautz, J., Miiller, A., Van Gunsteren, W. and Berendsen, H. J. C. Am. J. Chem. Soc. 1988, 110, 3393 Poppe, L. and van Halbeek, H. Am. J. Chem. Soc. 1992,114, 1092 Morris, E. R., Powell, D. A., Gidley, M. J. and Rees, D. A. J. Mol. Biol. 1981, 155, 507 Powell, D. A, Morris, E. R., Gidley, M. J. and Rees, D. A. J. Mol. Biol. 1982, 155, 517 Breg, J., Kroon-Batenburg, L. M. J., Strecker, G., Montreuil, J. and Vliegenthart, J. F. G. Eur. J. Biochem. 1989, 178, 727
Anne Imberty LSO-CNRS, Facult~ des Sciences et Techniques, 2 rue de la Houssini~re, 44072 Nantes cedex 03, France
and Serge P~rez* lngknierie Mol~culaire, INRA, BP 527, 44026 Nantes cedex 03, France
(Received 12 March 1992; revised 20 July 1992) In the present study, the conformational behaviour of methylated pectic disaccharide 4-O-ct-o-galactopyranurosyl 1-O-methyl-ct-o-galactopyranuronic 6,6'-dimethyl diester 1 has been completely characterized through combined n.m.r, and molecular modelling studies. The 1 H J H n.O.e, across the glycosidic bond was measured by both steady-state and transient 1D and 2D experiments. In parallel, the complete conformational analysis of the disaccharide has been achieved with the M M3 moIcular mechanics method. The conformation of the pyranose ring is confirmed by the excellent agreement between the experimental and calculated intracyclic scalar coupling constants. The iso-energy contours displayed on the 'relaxed' map indicate an important flexibility about the glycosidic linkage. There is no significant influence of the methoxyl group on the conformational behaviour of the disaccharide. The theoretical n.m.r, data were calculated taking into account all the accessible conformations and using the averaging methods appropriate for slow internal motions. 3Jc_n coupling constants were calculated using an equation suitable for C-O-C-H segments. The agreement between experimental and theoretical data is excellent. Within the potential energy surface calculated for the disaccharide, several conformers can be identified. When these conformations are extrapolated to a regular polymer structure, they generate pectins with right- and left-handed chirality along with a two-fold helix. These different types of helical structure are the result of small changes in conformation, without any drastic variation of the fibre repeat. Keywords: Pectin fragment; molecular modelling; conformation;n.m.r.
Introduction Pectins are a family of polysaccharides which are a major constituent of the cell wall of many higher plants. They contribute to the firmness and structure of plant tissues both as part of the primary cell wall and as the main middle lamella component involved in intercellular adhesion. The prototypical polysaccharide is based on a linear (1,4)-linked ~-D-galacturonan backbone (pectic acid) and may be partially methyl esterified (pectinic acid). This backbone is also interrupted by (1,2)-linked C~-L-rhamnopyranose and carries a proportion of lateral chains containing neutral sugars. Pectins have been used extensively as gelling agents over a long period of time, and less widely, as texturizers, emulsifiers, thickeners and stabilizers. The reasons why such a biopolymer, unlike others, develops a jelly in the presence of water, sugar and acid are still not elucidated. Nevertheless, the implications of structural details such as the degree of methoxylation, the distribution of * To whom correspondenceshould be addressed. 0141-8130/92/060313-08 © 1992 Butterworth-HeinemannLimited
rhamnopyranoses, the molecular weight of the polymer on some of the functional properties of pectins have been clearly demonstrated (for review, see ref. 1). An understanding of the mechanism of pectin gelation and how such a process is related to the molecular structure would make it possible to systematically control and manipulate the gel-rheological conditions through minute and controlled changes of molecular features. The characterization of the molecular features of pectins, in the solid state, in the gel state and in solution, has not been studied in as much detail as other polysaccharides. This may be explained by the occurrence of structural defects (which depend on the plant source and the conditions of preparation) and the concomitant lack of stereoregularity in pectins. The polyelectrolyte character of the chain in its acidic form along with the conformational transitions induced by monovalent and divalent ions complicate further the structural characterizations. Fibre diffraction analyses suggested that pectic acid and sodium pectate have similar three-fold right-handed structures 2'3. However, the occurrence of a left-handed single-stranded helical
Int. J. Biol. Macromol., 1992, Vol. 14, December
313
Conformation of pectin frayment disaccharide: S. Cros et al. structure has been also suggested 4. Structures deduced from solid-state measurements do not necessarily represent the hydrated and diffuse state of the gel network. Conformational analysis of disaccharides is a prerequisite to gaining useful information for modelling three-dimensional crystal structures of polysaccharides based on the independent residue approximation 5. Also, it helps to gain some insight into characteristics of the chain in its disordered state, such as radius of gyration, persistence length or the occurrence of local helical regions that might be appropriate for further ordering 6. In the present study, the conformational behaviour of methylated pectic disaccharide 4-O-a-D-galactopyranurosyl 1-O-methyl e-D-galactopyranuronic 6,6'-dimethyl diester 1 (see Figure 1) has been completely characterized through combined n.m.r, and molecular modelling studies. Separation of the anomers was possible in the case of these methylated derivatives and the spectral dispersion of the n.m.r, spectra considerably enhanced. Moreover, due to a much lower gyromagnetic moment, the relaxation rate from 2H is only 6% of that of 1H and this contribution can be generally neglected in an analysis of relaxation data of deuterated compounds. The present investigation complements a recent report 7 on the conformational analysis of the (1,4)-linked a-I>galacturono disaccharide 2.
J \ • C-1 c-~
Materials 4-O-c~-I>galactopyranurosyl 1-O-methyl-d 3 e-i>galactopyranuronic 6,6'-dimethyl-d 6 diester, 1, was prepared as follows. A total of 212.6mg ( ~ 0 . 6 m m o t ) of 4-O-c~-Dgalactopyranurosyl galactopyranuronic acid, 2, were refluxed overnight in 15 ml of MeOH-d4 in the presence of a trace of HC1 (obtained from 25 #1 of acetylchloride according to a literature procedureS). The solvent was evaporated under reduced pressure and the residue chromatographed over a S i O 2 column (Merck 9385 Kieselgel 60 - 5.5 x 20cm) using a 4 / 2 / 3 / 1 mixture of chloroform/acetone/methanol/water as the eluent. A minor fraction contained 8.5 mg of I while the major fractions yielded a mixture of products (220 mg) which were not identified. Product 1 was lyophilized in D 2 0 (99.96%) twice to remove residual water, degased and sealed in a 5 mm n.m.r, tube under argon.
Nomenclature The recommendations and symbols proposed by the Commission on Nomenclature 9 are used throughout this paper. A schematic drawing of the disaccharide, along with the labelling of the atoms, is given in Figure I. The relative orientation of a pair of contiguous residues about the glycosidic linkage is described by a set of two torsional angles: = O-5'_C-1' O-1'_C-4 tp = C-1'_O-1'_C-4_C-5 The magnitude of the valence angle z ( C - I ' _ O - I ' C-4) at the glycosidic bridge is also an important variable. As for the methyl-esterified group (see Figure 1 for the atomic nomenclature used), its orientation is described by two torsional angles X = O-5_C-5 C-6_O6 and o9 = C-5 C-6_O-6 C-7.
314
Int. J. Biol. Macromol., 1992, Vol. 14, December
6o
C-1'
C-5'
C-2"
\c,z>~ /
1
O
1
Experimental
~
2
Figure 1 Schematic representation of the methyl-esterified galacturonate disaccharide 1 and the acid disaccharide 2. The labelling of the torsion angles and some atom names of interest have been indicated on a diagram of 1
N,m.r.
1H-(400.13 MHz) and 13C_n.m.r. ( 100.6 MHz) spectra were recorded with a Bruker instrument operating in the Fourier-transform mode at 296 K. Both the 1H and 13C chemical shifts of 1 have been previously assigned ~° from homo- and heteronuclear correlation spectroscopy, respectively. The 3JH_I_Iwere extracted from a spectrum with a digital resolution of 0.12 Hz/pt. Two sets of steady-state n.O.e, difference spectra were obtained with irradiation times and pulse intervals greater than five times the longest TI for the sugar signals. 512 difference FIDs were accumulated and multiplied with a 1 Hz line broadening factor prior to Fouriertransform to reduce noise. Values from the two data sets were averaged. Both 13C and 1HT~ measurements were acquired with the inversion-recovery sequence (180-t-90-FID) and relaxation times were calculated with the Bruker 7"1 routine. In the case of selective 1H relaxation times, the duration of the soft 180 ° pulse, which was performed with the Dante sequence, was 20 ms. The decoupler offset was set to the frequency of the centre of either the H-4 or the H-I'/H-1 multiplet. In this latter case the resonances of these two protons overlap to form a triplet with the low-field peak corresponding to the H-I' doublet while the high-field transitions are part of the H-1 doublet. Thus, this selective 7"1measurement allowed simultaneous measurement of the diagonal elements of the relaxation matrix for the H-I' and H-1 protons through the recovery curves of the two outer branches of the triplet. Phase-sensitive NOESY ~~ were acquired with mixing times of 0 and 1 s. A 20 ms variable delay was introduced
Conformation of pectin fragment disaccharide: S. Cros et al. at the beginning of the mixing time in order to suppress J-peak transfer. The recycle time was set to five times the longest T1 to ensure that the normalized NOESY volume matrix would be symmetrical. A total of 512 x 1K data matrices were obtained and zero-filled to 1K x 1K. Prior to Fourier transformation the first data file was halved to reduce ta ridges, and n/2-shifted sine-squared weighting functions were applied t2. NOESY crosspeak intensities were evaluated from the summed o91 subspectra contributing to a specific signal. The values of the normalized aij and ai~ elements were averaged before back-transformation to the relaxation matrix ~3 with in-house software. Separation of the H-I' and H-1 contributions to the diagonal volumes for both 0 and 1 s mixing times was governed by two criteria: no more than a 15% deviation from the mean value of the zero mixing time diagonal elements was tolerated and a 2% fit to the inverse of the selective T~ after back-transformation was sought. This H - I ' / H - 1 ratio was also used in all the calculations of experimental relaxation values. The n.O.e, buildup for the H - I ' / H - 4 interaction was also measured with the (180sel-90-FID) sequence with the decoupler offset alternatively on the H-I' frequency or in a region where there were no signals. The variation in intensity for the signal of the H-4 proton was monitored by the resulting difference spectra with mixing times of 0.12, 0.16, 0.2 or 0.24 s. 512 difference FIDs were recorded.
Computational methods Molecular modelling. The starting geometry of the disaccharide was derived from the database of threedimensional structures, M O N O B A N K 14. Geometry optimizations were performed using the molcular mechanics program MM315'16. The total energy takes into account the stretching, bending, stretch-bending, torsional and dipolar contributions as well as van der Waals interactions. This force field, which has been compared with semiempirical methods 17, contains a correction for the anomeric effect without further parametrization 18. The MM3 force field has been extensively used for the study of carbohydrate structures 18-22. Partial atomic charges were not used since MM3 calculates dipole-dipole interaction energy from stored bond moments. The dielectric constant (e) was set to 4.0 for all calculations, as found most appropriate from studies of the Dglucopyranose ring in comparison with crystal structures 18. The block diagonal minimization method, with a default convergence criterion of 3.6cal/mol, was used for grid-point optimization. The preferred orientation of the methyl-ester exocyclic group was assessed with the DRIVER option of the MM3 program ~5'16, for both the monosaccharide and disaccharide 1. The torsion angles X and co were gridded by 20 ° intervals on the whole angular range. The lowest energy conformation was then used as starting point for calculating the relaxed map of the disaccharide about the q~ and u? glycosidic torsion angles. Like the other atoms, those belonging to the methyl ester exocyclic group were allowed to adjust. The relaxed map was computed using a 20 ° increment step for qb and ~ angles for the whole angular range. For each point of this grid, the optimized atomic coordinates and the corresponding energy were stored for further calculations of the n.m.r, observables.
Calculations of theoretical n.m.r, parameters. Averaging is required to correctly predict properties for a population
of conformations. Using the Boltzmann distribution, the relative population of each conformer is given by the expression: Pi - exp( - E i / k T ) / ~" exp( - Ei/kT ) For a n.m.r, parameter M depending in a non-linear way on an associated structural parameter S, the average value can be computed by
( M ) = ~,Pif(Si) Averaged coupling constant (3Jc_H) across the glycosidic bonds were calculated using the equation appropriate for C _ O _ C H segment 23. Equations for the calculation of the several (3JH_H) of the rings were established according to a known method 24. For the NOESY conditions, the ensemble average cross-relaxation matrix ( R ) was calculated from the ( r - 6 ) values following previously published procedures 13. The spectral density functions used in these calculations are appropriate for a molecule undergoing isotropic tumbling. It is also assumed that the internal motion is slow compared to the overall motion. For steady-state conditions, the ensemble average %n.O.e. can be calculated by solving a series of simultaneous equations 15'26. In order to obtain the best fit for the diagonal elements, Rii, of the relaxation matrix, a leakage factor of 2 0 % was included in the calculations.
Helical parameters. When subjected to the constraints imposed by the helical symmetry of macromolecular chains, equivalent monomeric units should occupy equivalent positions about the molecular axis. This is achieved when the (qb, ~g) angles are the same at every linkage; the secondary structure is described in terms of a set of helical parameters: (n, h), n being the number of residues per turn of the helix, and h being the translation of the corresponding residue along the helix axis. The chirality of the helix is defined by the sign of n. Arbitrarily, a right-handed helix will correspond to positive values of n; conversely, negative values of n will be designated left-handed helices. Whenever the values h = 0 or n = 2 are intercrossed, the screw sense of the helix changes to the opposite sign. These parameters are calculated for all combinations arising from rotations about the glycosidic linkage following an algorithm reported previously27; the corresponding iso-n and iso-h contours are plotted superimposed on the potential energy surfaces.
Results and discussion Molecular modelling The orientation of the methyl-ester exocyclic group has been established as a function of the torsion angles co and Z (Figure 2). For the monosaccharide, the trans orientation (180 °) of the torsion angle co is always preferred. The Z torsion angle does not exhibit a clear conformational preference. For this reason, the orientation of the methyl-ester group was studied for the disaccharide. In this study, the exocyclic group of the reducing end was selected, for different conformations of the glycosidic bond. The minimum with Z = 140° and o) = 180 ° is the lowest energy one. In all cases, the co angle retains its 180° favoured conformation. The second angle presents two possible orientations ()~ = 0 ° and Z = 140°) . The conformation (co = 180 °, g = 140°) is always the lowest energy one (see Figure 2). This conformation was retained
Int. J. Biol. Macromol., 1992, Vol. 14, December
315
Conformation of pectin fragment disaccharide: S. Cros et al. 360 t
'~
(80, 140) and (120, 160). The most stable conformer is I, conformers II and III being 0.8 kcal/lnol and 2.6 kcal/mol, respectively. The relaxed map of the acidic disaccharide 2 has been calculated, it is not shown here because it does not exhibit any significant changes with respect to that corresponding to 1. These two maps look like the one computed previously for the acidic disaccharide 2 using the M M 2 C A R B program 7. The only difference between the two investigations is the prediction of a fourth conformer ( 0 = 8 0 ° and q~= 300 ° ) by the M M 2 C A R B force field, being 5 kcal/mol higher than the most stable conformer. From the present study, we found that such a conformation would have an energy of approximately 9 k c a l / m o l in relative energy which explains why it is not represented here. Therefore, the minute differences are mainly due to slightly different force fields.
\
300t 240-
CO 180-
"o
120
-120
-60
0
60
120
180
240
Z Figure 2 Iso-energy map of the orientation of the methyl-ester exocydic group as a function of the 09 and Z torsion angles. The exocyclic group is the one at reducing end of the disaccharide. For each set of ~o and Z values, the whole geometry of the disaccharide has been optimized. Isoenergy contours are drawn with interpolation of 1 kcal/mol above the minimum of the map
N.m.r. data Both 1H and 13C chemical shifts of the precursor of 1 have been assigned. Methylation s h i f t s 7 - 2 8 of +7.99 and - 0.44 p.p.m, are observed for the C- 1 and H- 1 signals of 1 in agreement with reported values (Table •)29. A correlation time of 1.2 × 10- los for the overall motion was obtained from the average 13C T1 of the methine carbons and a p r o t o n - c a r b o n distance of 0.11 nm with the equation,
1
h272cTzHf
Zc
3~c TI -- 10r~_H\l----" + (COB-- COC)2Z~+ 1 + CO2Z2 240
6Zc ) + 1 + (con + coc)2z2
200" Table 1 400.13MHz 1H- and 13C-n.m.r. chemical shift data for methyl-esterified galacturonic acid dimer 1 in D20 (6nH°D = 4.80 p.p.m, and 6c_v = 100.93 p.p.m.)
160-
,t,
6n
120-
H-1 H-2 H-3 H-4 H-5
80-
40-
0
I
0
40
'
I
I
80
120
'
I
160
'
I
200
H-I' H-2' H-3' H-4' H-5'
4.96 3.72 3.92 4.33 5.09
C-1 99.99 C-I' C-2 67.98 C-2' C-3 68.34 C-3' C-4 79.39 C-4' C-5 70.28 @5' C-6, @6' 171.86, 170.95
100.93 68.11 68.99 70.38 71.70
=
240
0 Figure 3 'Relaxed' potential energy surface of the methylesterified galacturonate dimer as a function of the • and q~ torsion angles. Isoenergy contours are drawn with interpolation of 1 kcal/mol above the minimum of the map
as the starting conformation of the methyl-ester exocyclic group in the study of parent disaccharide. The relaxed map of disaccharide ! computed as a function of the qb and qu torsion angles is given in Figure 3. There is only one main low energy region, rather extended along both the • and qJ axis. Within the 8 kcal/mol limit, three energy minima (I III) are found on the relaxed map with qb and qJ values of (80, 100),
316
4.95 3.84 4.00 4.44 4.70
6C
Int. J. Biol. Macromol., 1992, Vol. 14, December
Table 2 400.13MHz observed and calculated 3JH. H and 3Jc. H coupling constants of methyl-esterified galacturonic acid 1. The calculated values were obtained by averaging over the whole relaxed map
Proton
JExp
JCalc
JH I'-H 2' Ju Z-H 3' JH 3'-H 4' JH 4'H-5' JH m-2 Jn 2H-3 JH 3H-4 JH_nH-5 JH I'-C 4 Jc I'-H 4
4.0 10.5 3.4 1.5 3.8 10.5 3.4 0.7 3.7a 4.9"
3.9 9.6 3.1 1.1 3.8 9.8 3.0 1.1 3.4 4.7
a Ref. 7
Conformation o f pectin f r a g m e n t disaccharide: S. Cros et al.
Experimental 3JH_H v a l u e s a r e g i v e n i n Table 2 a l o n g with the corresponding data which have been calculated for the 4C 1 conformation of the galactopyranuronosyl r i n g a c c o r d i n g t o H a a s n o o t et al. 24. T h e s m a l l m e a n d e v i a t i o n b e t w e e n t h e t w o sets o f p a r a m e t e r s , 0 . 4 H z , i m p l i e s t h e c l a s s i c 4C1 f o r m f o r t h e r i n g s o f 1. S t e a d y s t a t e % n.O.e, m e a s u r e d f o r t h e d i s a c c h a r i d e 1 a t 400.13 M H z a r e r e p o r t e d i n Table 3. A m o n g t h e 12 p r o t o n p a i r s w h i c h y i e l d e d a n.O.e, e n h a n c e m e n t , t w o correspond to inter-residue interactions. The H-I'/H-4 p a i r g i v e s a s t r o n g n.O.e, effect ( 1 7 % ) w h i l e t h e H - 5 ' / H - 2 g i v e s a w e a k effect ( 4 % ) . T h e s e r e s u l t s a r e in q u a l i t a t i v e a g r e e m e n t w i t h t h e n.O.e, d a t a o b t a i n e d o n c o m p o u n d 2 a t 300.13 M H z 7.
The NOESY spectrum ofgalacturonic acid disaccharide 1 is d i s p l a y e d i n Figure 4. T h i s p h a s e - s e n s i t i v e N O E S Y spectrum acquired with a mixing time of 1 s contains the c r o s s - p e a k s c o r r e s p o n d i n g t o all t h e i n t e r a c t i o n s m e a s u r e d in the steady-state spectra. Some additional weak inter-residue NOESY peaks can be observed.
Solution conformation o f the disaccharide The inter-residue conformation can be attained through b o t h t h r e e - b o n d p r o t o n c a r b o n scalar coupling, 3Jc_H, across the glycosyl linkage and inter-residue crossrelaxation rates. These heteronuclear coupling constants h a v e b e e n m e a s u r e d i n t h e c a s e o f 2 7. T h e s e d a t a c a n b e compared with the ones calculated for disaccharide 1
3 Steady-state n.O.e, values % observed (italic) a n d calculated for the methyl-esterified galacturonic acid 1. T h e calculated values were o b t a i n e d by averaging over the whole of the relaxed m a p
Table
Saturated spins
H-I'
H-2'
(23.6) 19.7 (2.5) 4.1 -
H-I' -
(11.2) 12.2 -
H-2' H-3'
-
H-4'
H-3'
(4.3) 3.3
(8.0)
-
8.3
. .
.
.
H-3
. . (17.4) 15.9
H-4
. .
.
-
-
-
(4.4) 4.9
. .
.
.
5.8
-
-
-
-
-
-
-
-
(3.6) 3.1 (7.0) 8.3 (9.4) 11.8 (8.5) 11.3
.
.
.
(18.7) 20.8
. .
-
-
(4.1)
. .
.
.
.
-
-
. .
. .
.
-
-
(7.0) 15.4
. .
.
. .
-
-
-
-
. .
.
.
-
. .
H-5
.
.
-
H-2
H-5
13.8
10.9
.
H-4
(16.8) .
. (8.6) 7.6 (9.3)
12.3
.
H-3
.
.
(11.5)
H-2
. .
-
(7.7)
H-1
.
. . . (11.8) 13.0 -
-
-
H-1
. .
.
12.6
-
H-5'
. .
H-5'
H-4'
.
.
(5.2) 7.4
(11.0) 12.0 (9.6) 10.9 -
Table 4 Relaxation matrix elements o b s e r v e & (italic) a n d calculated for the methyl esterified galacturonic acid 1: The calculated values were o b t a i n e d by averaging over the whole of the relaxed m a p Protons
H-I'
H-2'
H-Y
H-4'
H-5'
H-1
H-2
H-3
H-4
H-5
H-I'
(0.898) 0.974 -
(0.126) 0.117 (0.552) 0,594 . .
(0.018) 0.007 (0.019) 0.026 (0,764) 0.749 -
(0.002) 0.002 (0.014) 0.008 (0.102) 0.102 (0.807) 0.726 -
(0.007) 0.009 (0,016) 0.006 (0.087) 0.076 (0.080) 0.098 (0.784) 0.815
-
-
(0) 0,001 (0.001) 0.000 (0.00t) 0.001 (0.001) 0.000 (0.002) 0.001 (0.514) 0.508
(0,003) 0.003 (0,016) 0.001 (0.026) 0.012 (0.013) 0.001 (0.041) 0.040 (0.102) 0.106 (0.680) 0.690
(0.009) 0.003 (0,0t7) 0.000 (0) 0.001 (0) 0.000 (0.001) 0.003 (0.022) 0.007 (0.009) 0.026 (0.809) 0.777
(0.164) 0.153 (0.012) 0,003 (0.002) 0.002 (0.001) 0.001 (0.007) 0.011 (0,002) 0.002 (0.019) 0.008 (0.104) 0.102 (1.267) 1.113
(0.009) 0.006 (0.010) 0,000 (0) 0.001 (0) 0.000 (0.001) 0.001 (0.006) 0.010 (0.015) 0.006 (0.106) 0.096 (0.117) 0.093 (0.745) 0.737
H-2' H-3' H-4'
-
H-5'
-
H-1
.
.
H-2
.
.
.
H-3 .
.
.
.
.
. .
.
. .
.
.
.
. .
. .
.
.
.
n-4 H-5
.
.
.
. . .
. .
.
.
.
. .
.
. .
a The NOESY peaks intensities were obtained following ref. 37
I n t . J. Biol. M a c r o m o l . ,
1992, Vol. 14, D e c e m b e r
317
Conformation of pectin fragment disaccharide: S. Cros et al. with proper averaging over all the conformations of the relaxed map. As shown in Table 2, the mean deviation between the experimental values and those calculated for the averaged structure is excellent, < 5%. Steady state %n.O.e. were averaged from all the conformations of the relaxed map. In Table 3 these calculated values are compared with the experimental ones. Calculated values remain in the majority of cases larger than experimental ones. It appears that, although in at least 10% of the cases partial saturation of neighbouring resonances was observed, incomplete saturation was often a problem. When taking in account all the values (23 n.O.e, data) the standard mean deviation
is 27%. This value drops to 19% if the n.O.e, oberved for H-2 when H-1 is saturated (together with H-I') is omitted. These mean deviations are almost identical for the intra-residue n.O.e. (19 data) and the inter-residue n.O.e. (four data). With the exception of the H - l / H - 2 pair, the ensemble averaged n.O.e, values are in good agreement with those derived from experiments. As for the NOESY experiments, the comparison between observed and ensemble averaged values has been evaluated for the relaxation matrix ( Table 4). Comparison of relaxation matrix (or NOESY intensities) is preferred to comparison of calculated and estimated proton distances 3°. When taking into account all matrix elements
5 5' 4
4'
\ I
5.0
I
41~
~l.
4~8
I"
4.2
3J.
~.0
PPH
o
o
(I
0
t
o
~
0o
O
t4
o ¢00
.q
V4,1'
.~
¢
:
,¢ 0!
r !
,) I
~ " - - -
-
-
-
~
:
j'
--
'
iV1"4 O
~ O
"-
Figure 4 Phase-sensitive NOESY spectrum of esterefied galacturonic acid dimer 1 recorded with 1 s mixing time in D20. The cross-peaks corresponding to the H-I'/H-4 interaction are labelled and the 1 D spectrum is given above
318
Int. J. Biol. Macromol., 1992, Vol. 14, December
Conformation of pectin fragment disaccharide: S. Cros et al. with observed values greater than 2% (i.e. stemming from measurable N O E S Y volumes, 22 data) the mean deviation is 11.8 %. F o r both galactopyranuronosyl rings, the N O E S Y experiments were compatible with the classic 4C 1 form. As regards the inter-residue data, the mean deviation is 21% (three data > 2%). However, it should be noted that in the case of the two strongest interactions, H - I ' / H - 4 (0.164) and H - 2 / H - 5 ' (0.041), the data are in excellent agreement with the modelled values, 0.153 and 0.040 respectively. The latter H - I ' / H - 4 cross-relaxation rate was measured by three methods: a combination of steady-state n.O.e, values and selective relaxation rates (0.205s-1), N O E S Y spectra (0.164s-1), and n.O.e. buildup (0.201s-~). It should be noted that the data obtained in the latter approach indicated that the n.O.e. buildup was not fast compared to the selective relaxation rate. Carver 31 has recently summarized the treatment appropriate for extracting internuclear distances from relaxation data of carbohydrates in the presence of internal motion. In the case of internal motion that is slow compared to the overall motion the approach described above is justified. However, when the internal movements are fast compared with the molcular tumbling, appropriate spectral density functions 32'33 are required. Such rapid motion has been described recently for torsion about the glycosidic linkage of sucrose 34. In the intermediate case, where internal motion is on a timescale comparable with the overall tumbling, the full dipolar interaction (both angular and distance) would have to be evaluated from the dynamics trajectory. A reliable protocol for evaluating the rate of internal motion in carbohydrates for all of these cases is being studied.
..... .- ..~"... ~'2;;;;7"::::::!::~7::... ........ . . " ' " ......... "..;.....--:::.:.:.:?.'.!!:ii:!!::... ." ..' . ................~"--.~':~;;::~i:-..
240 .
. ~
20o .
. . .
16o-.
...-~i.."i ...........
"................ ':.::!~i!!~!!!~..
........ .......
~ \\~.i
i :::..i!i; .....
8o ~i;.i~)., -..
"-
~:i::::...
ii.?..~ ~
0
0
A
40
80
B
120
160
C
200
240
D
Helical conformation of the polysaccharide In agreement with the experimental evidence provided by the n.m.r, data, the conformational analysis of the disaccharide in solution established that a large number of conformations are available for the molecule. The potential energy surface of Fiyure 3 is therefore representative of the conformational behaviour of the disaccharide unit, and may be used to model some features associated with the parent polysaccharide. Among these, the description of the stable helical conformation of a stereoregular polymer is straightforward. This has been done, using a rigid residue approach. The iso-n contours corresponding to n = 2, n = 3, n - - - 3 and n = 4 along with the iso-h contour have been superimposed on the low energy region (Figure 5). This shows that the m a p is divided into regions of left-handed and right-handed chirality. The chiral transition occurs through conformations of q~, W yielding values of n = 2. Since the iso-h in the low energy regions have the highest values, a tendency towards m a x i m u m extension is exhibited. Four stable conformers, generating integral helices can be found on the potential energy surfaces (A: n=-3, h = 0 . 4 4 n m ; B: n = 2 , h = 0 . 4 3 n m ; C: n = 3 , h = 0 . 4 4 n m and D: n = 4 , h = 0 . 4 4 n m ) . These four helices which all exhibit maximum extension are displayed in Figure 5. Helices A, B and C have glycosidic torsion angles in close correspondence with energy minima: III, II and I, respectively. These different models can be compared with experimental data obtained on both acidic and esterified pectins. Fibre diffraction performed on N a + and Ca 2+ pectate gels yielded models 2,3 of three-fold, right-
Figure 5 Iso-n and iso-h contours for the methyl-esterified galacturonic acid dimer 1 superimposed on the low energy region of the potential energy map. Four low energy conformations which would generate integral helices have been indicated by a, and labelled A (n = - 3, h = 0.44 nm), B (n = 2, h = 0.43 nm), C (n = 3, h = 0.44 nm) and D (n = 4, h = 0.44 nm). These four helices have been represented using projection parallel and orthogonal to their axis
handed helices repeating in 1.3 nm (n = 3, h = 0.43 nm). On the basis of circular dichroism, calcium stoichiometry and competitive inhibition, a two-fold helix having a repeat of 0.870 nm (n = 2, h = 0.435 nm) was proposed for calcium gel 35'36. More recently, pectins polygalacturonate single chain images were visualized through electron microscopy; the occurrence of a left-handed helix with a 1.3 nm period (n = - 3, h = 0.43 nm) was proposed 4. The results of our calculations indicate that each type of helical conformation seems to be almost equally preferred. The calculated energy differences indicate that interconversion between low energy conformers are feasible. It is also clear that neighbouring of ions, solvent or other macromolecules can easily induce a conformational change. The most striking features are the facts that these changes in conformation can occur without any noticeable variation of the fibre repeat.
Int. J. Biol. Macromol., 1992, Vol. 14, December
319
Conformation of pectin fragment disaccharide: S. Cros et al.
Conclusion The present work has been performed on the dimeric segment of methoxyl pectins and pectins. The conformational behaviour of the methylated pectic disaccharide has been characterized through combined n.m.r, and molecular modelling studies. The agreement reached between observed and calculated data is a satisfactory test of the validity of the conformational energy surface which may be used to safely model some of the structural features of the polymers. The relevance of the stable conformers with respect to helical structures of pectins indicates that both right- and left-handed single-stranded helices may occur, along with a two fold-helix. These different types of helical structure occur for small conformational changes without any drastic variation of the fibre repeat. Of a certain interest is the conclusion that there is no significant influence of the metho×yl group on the conformational behaviour of both the disaccharide and the polysaccharide. Therefore, it may be concluded that the well known influence of methoxylation on functional properties of pectin may be arising from intermolecular origins.
9 10 11 12 13 14 15 16 t7 18 19 20 21 22 23
Acknowledgements
24
The authors are indebted to D r Kevin Hicks (USDA, Philadelphia) who kindly provided c o m p o u n d 2.
25
References 1
8
Walter, R. (Ed.) 'The Chemistry and Technology of Pectin', Academic Press, London, 1991 Walkingshaw, M.D.andArnott, S.J.Mol.Biol. 1981,153, 1075 Walkingshaw, M. D. and Arnott, S. d. Mol. Biol. 1981,153, 1055 Ruben, G. C. and Bokelman, G. H. in 'Pectin's polygalactouronique single sugar chain helix visualized in Pt/C Replicas', San Francisco Press, California, 1987, p 966 Brant, D. A. Q. Rev. Biophys. 1976, 2, 232 P6rez, S. in 'Methods in Enzymology', Vol. 203, Academic Press, New York, 1991, p 510 Hricovini, M., Bystricky, S. and Malovikova, A. Carbohydr. Res. 1991, 220, 23 Gmunder, J. Helv. Chim. Acta 1953, 36, 2021
320
Int. J. Biol. Macromol., 1992, Vol. 14, December
2
3 4 5 6 7
IUPAC-IUB Commission on Biochemical Nomenclature. Arch. Biochem. Biophys. 1971, 145, 405
26 27 28 29 30 31 32 33 34 35 36 37
Ohassan, H. DEA dissertation, Paris VII, 1989 Bodenhauseu, G., Kogler, H. and Ernst, R. R. J. Maon. Resort. 1984, 58, 370 Neuhaus, D. and Williamson, M. in 'The Nuclear Overhauser Effect in Structural and Conformational Analysis' VCH Publishers, 1989, p 292 Olejniczak, E. T., Gampe, R. T. and Fesik, S. W. J. Magn. Reson. 1986, 67, 28 P6rez, S. and Delage, M. M. Carbohydr. Res. 1991, 212, 253 Allinger, N. L., Yuh, Y. H. and Lii, J.-H. J. Am. Chem. Soe. 1989, 111, 8551 Allinger, N. L., Yuh, Y. H. and Lii, J.-H. J. Am. Chem. Soe. 1990, 112, 8293 Gundertofte, K., Palm, J., Petterson, I. and Stamvik, J. Comput. Chem. 1991, 12, 200 French, A. D., Rowland, R. S. and Allinger, N. L. in 'Computer Modeling of Carbohydrate Molecules', Am. Soc., Washington, DC, 1990, p 120 French, A. D. Carbohydr. Res. 1989, 188, 206 French, A. D. in 'Cellulose and Wood-Chemistry and Technology', Wiley, New York, 1989, p 103 French, A. D. and Brady, J. W. in 'Computer Modeling of Carbohydrate Molecules', Am. Soc., Washington, DC, 1990, p 1 Dowd, M. K., Reilly, P. J. and French, A. D. J. Comp. Chem. 1991, 13, 102 Tvaroska, I., Hricovini, M. and Petrakova, E. Carbohydr. Res. 1989, 189, 359 Haasnoot, C. A. G., De Leeuw, F. A. A. M. and Altona, C. Tetrahedron 1980, 36, 2783 Noggle, J. H. and Schirmer, R. E. in 'The Nuclear Overhauser Effect', Academic Press, New York, 1971 Cumming, D. A., Carver, J. P. Biochemistry 1987, 26, 6664 Gagnaire, D., P&ez, S. and Tran, V. Carbohydr. Res. 1980, 78, 89 Rinaudo, M., Ravanat, G. and Vincendon, M. Macromol. Chem. 1980, 121, 1059 Perlin, A. S. and Casu, B. in 'The Polysaccharides', Academic Press, London, 1982, p 133 Clore, G. M. and Gronenborn, A. M. J. Magn. Res. 1989, 84, 398 Carver, J. 'Current Opinion in Structural Biology', 1991, 1, 716 Lipari, G. and Szabo, A. J. Am. Chem. Soc. 1982, 104, 4546 Kessler, H., Griesinger, C., Lautz, J., Miiller, A., Van Gunsteren, W. and Berendsen, H. J. C. Am. J. Chem. Soc. 1988, 110, 3393 Poppe, L. and van Halbeek, H. Am. J. Chem. Soc. 1992,114, 1092 Morris, E. R., Powell, D. A., Gidley, M. J. and Rees, D. A. J. Mol. Biol. 1981, 155, 507 Powell, D. A, Morris, E. R., Gidley, M. J. and Rees, D. A. J. Mol. Biol. 1982, 155, 517 Breg, J., Kroon-Batenburg, L. M. J., Strecker, G., Montreuil, J. and Vliegenthart, J. F. G. Eur. J. Biochem. 1989, 178, 727
