Application of two-dimensional NMR spectroscopy and molecular dynamics simulations to the conformational analysis of oligosaccharides corresponding to the cell-wall polysaccharide of Streptococcus group A

Application of two-dimensional N M R spectroscopy and molecular dynamics simulations to the conformational analysis of oligosaccharides corresponding to the cell-wall polysaccharide of Streptococcus group A U w e C. Kreis, Vikram V a r m a and B. M a r i o Pinto* Department of Chemistry, Simon Fraser University, Burnaby, British Columbia, Canada, VSA IS6 Received 19 August 1994; revised 4 November 1994

This paper describes the use of a protocol for conformational analysis of oligosaccharide structures related to the cell-wall polysaccharide of Streptococcus group A. The polysaccharide features a branched structure with an L-rhamnopyranose (Rhap) backbone consisting of alternating ~-(1 --.2) and ~-(1 --*3) links and g-N-acetylglucosamine (GlcpNAc) residues/~-(1-,3)-connected to alternating rhamnose rings: I

A I--~2)-a-L-Rhap-( B' I---~3)-c~-L-Rhap-( A' B -ot-L-Rhap-( 1---~2)-a-L-Rhap-( 1---~3)-l [ (I--->3) I (I--)3) [],-D-GlcpNAc I$-D-GlcpNAc C' C n

Oligomers consisting of three to six residues have been synthesized and nuclear magnetic resonance (NMR) assignments have been made. The protocol for conformational analysis of the solution structure of these oligosaccharides involves experimental and theoretical methods. Two-dimensional NMR spectroscopy methods (TOCSY, ROESY and NOESY) are utilized to obtain chemical shift data and proton-proton distances. These distances are used as constraints in 100ps molecular dynamics simulations in water using QUANTA and CHARMm. In addition, the dynamics simulations are performed without constraints. ROE build-up curves are computed from the averaged structures of the molecular dynamics simulations using the CROSREL program and compared with the experimental curves. Thus, a refinement of the initial structure may be obtained. The ~-(1 -,2) and the/%(1 --,3) links are unambiguously defined by the observed ROE cross peaks between the A-B',A'-B and C-B,C'-B' residues, respectively. The branch-point of the trisaccharide CBA' is conformationally well-defined. Assignment of the conformation of the B-A linkage (~-(1--.3)) was problematic due to TOCSY relay, but could be solved by NOESY and T-ROESY techniques. A conformational model for the polysaccharide is proposed.

Keywords: Streptococcusgroup A oligosaccharide antigen; conformational analysis;combined NM R spectroscopic and molecular dynamics protocol The organism Streptococcuspyogenescauses skin infections, pneumonia, streptococcal pharyngitis, and toxic shocklike syndrome 1'2. If untreated or improperly treated, streptococcal infections can lead to rheumatic fever and heart-valve damage 1'2. The close relationship between * To whom correspondence should be addressed

0141-8130/95/$09.50 © 1995 ElsevierScienceB.V. All rights reserved

Streptococcus infection and

rheumatic diseases is thought to be the result of an auto-immune response - an action led by antibodies produced during an immune response to streptococcal infection cross-reacting with heart-valve tissue and other host tissues. Antibodies directed against both the surface M protein and the group-specific cell-wall carbohydrate have been implicated in the

Int. J. Biol. Macromol. Volume 1 7 Number 3-4 1995

117

Conformational analysis of Streptococcus group A cell-wall polysaccharide-related oligomers: U. C. Kreis et at.

OPt

auto-immune reaction and have been shown to react with cardiac tissue 3 v. The Streptococcus group A cell-wall polysaccharide is comprised of a backbone of poly-e-rhamnopyranosyl units connected by alternating ~-L-(l--*3) and C~-L-(l--*2) linkages, to which N-acetyl-/~-o-glucosamine residues are attached at the 3-position of the rhamnose backboneS'9:

HO~"~ HO

c

O

B A'

NHAc O

OH

HO OH 1

OPt A 1-~2)-~-L-Rhap-( B' I ---)3)-ct-L-Rhap-( A' 1---~2)-o~-L-Rhap-( 13 1---)3)-cc-L-Rhap-( I (1--.3) I (1-->3) I~-D-GlcpNAc ~-D-GlcpNAc C' C

l

Experimental N M R measurements Each oligosaccharide (5 mg) was dissolved in 99.98% D 2 0 and lyophilized five times to exchange the hydroxyl protons with deuterium. It was then redissolved in 99.98% D 2 0 and subjected to five freeze thaw cycles to remove dissolved oxygen. The tubes were sealed under vacuum. Most N M R experiments were performed using a Bruker AMX-400. Samples were run spinning at temperatures ranging from 288-294K. Typically, each two-dimensional experiment consisted of 512 experiments, each 2 K in size, using time-proportional phase increments to provide quadrature detection in F1. The sweep width in the ROESY spectra T M was 3 3.5ppm (1200 1300Hz) giving a maximal digital resolution of 0.6 Hz/pt in FI. In order to minimize the effects of folded peaks, the filter was set equal to the sweep width. The transmitter power was attenuated so that a high-power n pulse was usually between 5 6 #s. For ROESY and TOCSY 19 experiments, the spin lock power was attenuated to give n pulses of ! 50 and 30/~s, respectively. T-ROESY experiments were performed using the pulse sequence described in the literature z°. A (n/2) pulse was followed by a composite spin lock of n-xnx pulses in order to eliminate unwanted TOCSY effects. The free induction decays (FIDs) were zero-filled to a 4 K (F2) × 1 K (F1) data set, providing a resolution

Int. J. Biol. Macromol. Volume 17 Number 3-4 1995

OH

HO'~ B 0 0 OH A' HO O NHAc O

A

Ho~O OHOO B

n

In an attempt to identify fragments of the group A cell-wall polysaccharide that will elicit an immune response that targets the organism specifically without triggering harmful sequelae, we have synthesized several oligosaccharide subunits of the polysaccharide ~°-~4. These compounds have been used to prepare and characterize mono- and polyclonal anti-streptococcal antibodies that can be used as immunological probes of cross-reactive epitopes on the organism and host tissues 15'16. In order to delineate further the nature of the antibody-antigen interactions implicated in anti-streptococcal responses, knowledge about hapten topology is required. To this end, our synthetic oligosaccharides have been subjected to initial nuclear magnetic resonance (NMR) analysis ~° a4. In this paper, we report the detailed conformational analysis of some of the key compounds, namely trisaccharide 1 [A'-~C)B], tetrasaccharide 2 [A'-(C)B-A], pentasaccharide 3 [C'B' A'-(C)B], hexasaccharide 4 [C'-B' A'-(C)B-A] and hexasaccharide 5 [A-(C')B' A'-(C)B] (Figure 1), by a combined N M R spectroscopic and molecular dynamics protocol. The information is expected to guide further studies on the design ofimmunodiagnostics and vaccines.

118

O

PrO

0

NHAc 0

HOO H

2

H

A' H

HO

OHHO O~oHO

3

HO#

B'

NHAc,

HO OH 0

pro

OH

OPr

A

HO'~

B

0

HO O 0

O

HOO~I ~ NHAc C

H~/H

0

OH

A'

NHAc O

OH 0

A'

B'

HOOH H O ~

O ~ ~ HO~)

O

O

OH

B'

H

O

4 .ojf<

H

OH

H

HOOI4 5

HO OH

Pr =

CH2CH2CH3

Figure I Oligosaccharidescorrespondingto fragmentsof the cell-wall polysaccharideof Streptococcus group A of 0.3 Hz/pt in F2 and 1.2 Hz/pt in F1, and processed using an exponential line-broadening of 0.3 in the F2 dimension and a sine-squared apodization function with a shift of 2 in Ft. Integration was performed after automatic baseline correction in both dimensions. Measurement of the three-bond coupling constants (3Jc.) of the hexasaccharide fi was done by inverse two-dimensional 13C 1H correlated spectroscopy on a 600 MHz Bruker AMX spectrometer at the Carlsberg Laboratory in Copenhagen, Denmark, using a mixing time (125 ms) optimized for detection of the three-bond coupling constants with a 4 K x 512 data set (digital resolution = 1.2 Hz/pt). In cases where the T1 noise of the residual H 2 0 interfered with the signal, the coupling constants were obtained from a one-dimensional coupled x3C spectrum, acquired on a Bruker AM500 spectrometer, also at the Carlsberg Laboratory. The one-dimensional data sets were acquired with a digital resolution of 0.2 Hz/pt.

Computations Calculations were performed on Silicon Graphics Personal Iris 4D25 and 4D35 workstations and an SGI 380 Powerseries computer. All the molecular mechanics calculations and dynamics simulations were performed using CHARMm2221 23 interfaced with Q U A N T A version 3.3. For each of the oligosaccharides, the interglycosidic linkages were minimized individually

Conformational analysis of Streptococcus group A ceil-waft polysaccharide-related oligomers: U. C. Kreis et al. by molecular mechanics calculations using C H A R M m , subjecting each conformer to a full optimization. Initial p r o t o n - p r o t o n distances were obtained from R O E enhancements using the two-spin approximation and assuming isotropic tumbling according to the following equation: //0-0\1/6

rij=ro [ ~ \tTij/

(1)

where r o is a ruler distance, i.e. the distance between 1H and 2H within the same rhamnose ring or between 1H and 3H in the N-acetylglucosamine ring, and ao/aij is the ratio of the corresponding cross-relaxation rates. The uncertainty of the cross-peak intensities was of the order of + 1 5 % . Distances calculated from different data sets showed consequent variations of + 5%. The deviations for the upper and lower bounds of the distance constraints were of the order of 5-10%. These distances were used to create a starting conformation for the molecular dynamics simulations using CHARMm222a-2a. For the dynamics simulations in water, the oligosaccharides were minimized using the adopted basis Newton-Raphson algorithm with the constraints, centred in a water droplet of 30 A diameter utilizing the TIP3 model, and minimized further until the root-mean-square derivative was <0.01 kcal m o l - 1 A - 1. An additional water image of 30 A radius was added. For all dynamics calculations, a 10 ps heating period and a 10ps equilibration period preceded the production dynamics. The dynamics simulations were performed with these distances as constraints as well as without constraints for 100 ps each. A time step of 1 fs was used, and the plotting of the frames was adjusted to create 1250 frames for each dynamics run using the Verlet algorithm. The average distances rij between all the nonexchangeable protons were calculated by in-house programs from coordinates of all the frames of the dynamics simulation using equation (2):

1~ 1 rtj'avg

r/

Table ! tH-NMR data (ppm) for compounds 1-5 in D20 at room temperature Ring proton

l

4

5

4.74 3.97 3.77 3.51 3.72 1.25

5.14 4.01 3.74 3.39 3.63 1.21

4.77 4.13 3.80 3.46 3.68 1.25 4.63 3.69 3.50 3.40 3.40 3.88 3.72 5.14 4.04 3.81 3.50 3.71 1.24

5.07 4.26 3.94 3.49 3.78 1.26 4.68 3.70 3.51 3.41 3.42 3.90 3.73 5.15 4.05 3.81 3.50 3.67 1.23

4.77 4.13 3.81 3.45 3.68 1.25 4.63 3.69 3.50 3.40 3.40 3.87 3.72 5.15 4.03 3.81 3.52 3.71 1.24

I B' 2B'

5.00 4.24

5.00 4.24

5.08 4.25

3B'

3.88

3.89

3.95

4B' 5B' 6B' 1c' 2c' 3c' 4c' 5c' 6c's 6C'R

3.47 3.77 1.26 4.68 3.72 3.52 3.40 3.40 3.88 3.72

3.48 3.79 1.26 4.68 3.72 3.53 3.42 3.42 3.88 3.72

3.50 3.77 1.27 4.70 3.71 3.51 3.42 3.42 3.87 3.72

1A 2A 3A 4A 5A 6A 1B 2B 3B 4B 5B 6B lC 2c 3c 4c 5c 6cs 6CR 1A' 2A' 3A' 4A' 5A' 6A'

2

3

4.74 3.96 3.77 3.50 3.71 1.25 4.78 4.10 3.80 3.46 3.67 1.22 4.64 3.69 3.50 3.41 3.40 3.87 3.71 5.11 3.99 3.73 3.39 3.67 1.22

5.09 4.25 3.94 3.49 3.77 1.25 4.70 3.70 3.51 3.42 3.42 3.89 3.74 5.14 3.99 3.74 3.39 3.63 1.22

(2)

r6

where n is the number of frames. ROESY build-up curves were then calculated using the C R O S R E L program 24. Multi-spin and non-relaxation effects in ROESY spectra are accounted for in the full relaxation matrix calculations performed by CROSREL, permitting quantitative analysis and the computation of theoretical ROE build-up curves. The scale was adjusted by the slope of the experimental versus theoretical diagonal peak intensities.

Results and discussion The synthesis of the oligosaccharides presented in Figure 1 and the assignment of their 1H and 13C-NMR spectra have been described previously 1° 14. A summary of the ~H chemical shift assignments is given in Table 1, and the ~3C chemical shifts are shown in Table 2. The differences in chemical shifts yield information about the conformation and have been used for the conformational analysis of saccharides with reasonable s u c c e s s Z 5 28. Correlation between the chemical shifts and the conformation of these compounds can best be accomplished by comparison with a parent compound.

Since it represents the minimal repeating unit, the trisaccharide 1 was chosen as the parent so that the A3 values in Figure 2 represent the differences between chemical shifts of the ring protons of the various oligosaccharides 2-5 and the trisaccharide 1. Negative A6 values correspond to shielding, while positive values correspond to deshielding of the corresponding hydrogen atom in the oligosaccharide. The major differences in chemical shifts for both 1H and 13C are congruent with the different substitution patterns of the individual compounds. The following discussion details this analysis for the proton spectra. The negative A6 values of H1 of the A-ring of the tetrasaccharide 2 and the hexasaccharide 4 result because this is the terminal residue in these compounds. The anomeric portion of an A-ring within the chain could be close in space to O1 and 0 5 of the C-ring, which would result in deshielding and provide an explanation for the observed differences. Similarly, the B-rings of the tetrasaccharide 2 and the hexasaccharide 4 show rather large differences in chemical shifts compared to the trisaccharide 1, the pentasaccharide 3 and the hexasaccharide 5 in which the B-ring is the terminal residue. The B'-ring of pentasaccharide 3 and of both hexasaccharides 4 and 5 is part of the chain, hence the

Int. J. Biol. Macromol, Volume 17 Number 3-4 1995

119

Conformational analysis of Streptococcus group A cell-wall polysaccharide-related oligomers." U. C. Kreis et al. Ring A

Ring A'

0.15 1 0.1 0

t,o <1 0.05

-0.1 -0.2 -0.3

-0.05 J

-0.4

1A

2A

3A

4A

5A

6A

1A'

2A'

3A'

Ring B

4A'

5A'

6A'

5B'

6B'

Ring B' 0.35

0.35 0.25 ~

l

~

0.25 t~

0.15 0.05

0.15 0.05

-0.05

-0.05 1B

2B

3B

4B

5B

6B

IB'

2B'

3B'

Ring C

4B'

Ring C'

o J,, 0.05

0.05

0

0

-0.05

"

"

;

*

-0.05

1c

2c

3c

4c

5c

6cs 6CR

• Tetra • Penta • Hexal [] Hexa2

1C'

2C'

3C'

4C'

5C' 6C'S 6C'R

2 3 4 5

Figure 2 Differencesin 1H chemicalshifts of compounds 2-5 compared to those of trisaccharide 1 differences in chemical shifts. The C-rings exhibit only small differences as the substitution patterns are the same for all C- and C'-rings. The observed differences in 1H and 13C chemical shifts are a strong indicator of a similar conformation of the branch point in all examined structures because all observed differences are congruent with differences in substitution patterns. Due to the fact that ~o~c ~- 1 for most of the compounds studied here, N O E effects are small and ROESY spectra rather than N O E S Y spectra were recorded. Integrations of the cross peaks observed in the ROESY spectra were converted into p r o t o n - p r o t o n distances after corrections for offset and H a r t m a n n - H a h n effects 29. The problem of TOCSY transfer was addressed by examination of additional N O E S Y and T-ROESY experiments 2°. Based on these experimental data, we devised a protocol for

120

Int. J. Biol. Macromol. Volume 17 Number 3-4 1995

conformational analysis which is depicted in Figure 3. This protocol will be detailed for the hexasaccharide 5; analogous data for the other compounds are available from the authors as supplementary material. The first series of steps in this ,protocol results in the elucidation of p r o t o n - p r o t o n distances from the integrated cross-peak volumes. As these volumes are obtained from ROESY experiments, they need to be corrected for H a r t m a n n - H a h n and offset effects. Figure 4 depicts an expansion of the ROESY spectrum of hexasaccharide 5 which includes assignment of the relevant cross peaks. The spectrum shows features similar to those of the related structures 1 4 (see supplementary material), and, although the hexasaccharide 5 would seem to be more sterically crowded than the other structures, the observable cross peaks suggest that

Conformational analysis of Streptococcus group A celt-wall l~olysaccharide.related oligomers: U. C. Kreis et al.

Perform Roesy Experiment

!

Input Experimental Distance

c>

I Minimize Structure with Constraints Integrate Cross Peak Volumes

I

Run Molecular Dynamics Simulation with Constraints

I

I Run Molecular I Dynamics Simulation without Constraints

I

Correct Volumes for | Hartmann Hahn effects

Calculate Distances

l--

CalculateDistancesAverage I

Calculate Average Distances

Calculate ROESY Intensities using CROSREL

Calculate ROESY Intensities using CROSREL

| ~

[

Figure 3 Flowchart for the protocol of conformational analysis 1C-3B

1C-2B

°

Ic'3c ~

t

~'~ ~t~.4~,~C~~ 3 ' 1C'-2W

~ +m.0. ~.,.

.

!

C~.la~3~,,0t~, * IC'-3B'

~ C +1

• b

.. ' ,~e~O°0

4.7 1C'-3C'

1C'-5C'

4.8

-~

1B-2B

IB-3B

,I

-4.6

1B-5A' 1B-HPr(a)

1B-HPr(b)

-4.9 FI (ppm) 1B'-2B'

1B'-3A, 1B'-3A' 1B'-5A' 1B'-5A

1B'-2A' *0

-5.0 IB'~4A'

1B'-4A

*!

1A-2A

1A-2C'

, , 0

-5.1

Q~

0t 1A-2~' I

4.3

4.2

411

1A,.2A' 1 4.0

1A'-2C 3'.9

F2 Figure 4 Expansion of the ROESY spectrum of hexasaccharide5

3'.8 3'.7 (ppm)

3'.6

315

314

Int. J. Biol. Macromol. Volume 17 Number 3-4 1995 121

Conformational analysis of Streptococcus group A cell-wall polysaccharide-related oligomers." U. C. Kreis et al. Table 2

13C-NMR data (ppm) for compounds 1 5 in D20 at room temperature Ring carbon

IA 2A 3A 4A 5A 6A IB 2B 3B 4B 5B 6B 1C 2C 3C 4C 5C 6C 1A' 2A' 3A' 4A' 5A' 6A' 1B' 2B' 3B' 4B' 5B' 6B' IC' 2C' 3C' 4C' 5C' 6C'

1

2

101.2 79.6 82.7 74.8 71.7 19.4

102.4 72.5 80.3 74.9 71.5 19.3 104.4 79.0 82.4 74.6 72.1 19.3

105.3 58.7 76.6 72.9 78.6 63.7 104.4 72.8 72.8 73.9 71.9 19.4

105.2 58.7 76.7 72.9 78.6 63.8 103.8 72.9 72.8 74.0 71.9 19.4

3

4

5

101.2 79.3 82.8 73.6 72.0 19.4

102.3 72.8 80.4 74.6 71.5 19.3 103.8 78.9 82.5 74.1 72.2 19.5

104.3 72.6 72.7 74.7 71.8 19.3 101.1 79.3 82.7 74.0 71.6 19.3

105.3 58.6 76.6 72.6 78.5 63.6 104.1 72.6 79.7 74.3 72.1 19.5 104.4 72.6 82.7 74.0 72.1 19.3 105.4 58.4 76.5 72.6 78.4 63.4

105.2 58.7 76.7 72.6 78.6 63.7 104.1 72.6 79.9 74.3 72.4 19.4 104.5 72.8 82.8 73.7 72.1 19.6 105.4 58.6 76.5 72.6 78.5 63.5

105.3 58.5 76.5 72.5 78.4 63.5 104.0 72.5 79.3 74.4 72.0 19.2 103.6 78.9 82.4 73.8 72.0 19.5 105.2 58.5 76.6 72.5 78.4 63.5

the protons I A and I A' are very close, the cross peaks are well separated, permitting u n a m b i g u o u s assignment and integration. An exception is the I A - 2 C ' and I A ' - 2 C cross peaks which had to be integrated together. Their low intensities made a judgement of the relative magnitudes impossible and rendered it necessary to divide the integrals evenly in order to obtain an estimate of these distances. The cross peaks were integrated and corrected for H a r t m a n n H a h n and offset effects according to equations given by Bax z9. After applying these corrections to the integrals, the p r o t o n - p r o t o n distances were calculated according to equation (I). The p r o t o n - p r o t o n contacts observed and the distances obtained are displayed in Figure 6 and the constraints derived from them are shown in Table 3. Each linkage is well defined by two inter-glycosidic R O E interactions. In addition, the branching point is determined by additional inter-ring cross peaks between H 1 A ' - H 2 C and H I A H2C'. These constraints were used to set up a conformation for each c o m p o u n d solvated in a droplet of T I P 3 water with a diameter of 30A, which was then minimized using CHARMm222~ z3. An additional constraint of the ® angle (H2 C2 N - H ) of the N-acetyl g r o u p had to be introduced to reflect the preference of this substituent towards the trans position as observed in X-ray structures of N-acetylglucosamine 32 and its dimer N,N'-diacetyl-chitobiose ~3.

IC.5CI IC-3B

#r

t

there are no obvious compensations in terms of conformation. The corresponding F I slices in the spectrum of hexasaccharide 5 are presented in Figure 5 and the strongest inter-residue R O E s are observed between protons across the glycosidic linkages. The characteristic 1B-5A' cross peak that appears in all the ~-(1 ~2)-linked c o m p o u n d s is present in both the A ' - B and A - B ' linkages of hexasaccharide 5. The same cross peaks were observed in previous studies of the a-methyl glycoside of ct-(1--*2)-linked dirhamnoside 3°'31. In none of the c o m p o u n d s in this study could a cross peak between the two anomeric protons 1 A ' - I B be observed. The single ~-(1--*3) linkage joining the B' and the A' rings displays the strong 1B'-3A' cross peak. As with the other compounds, 1B'-2A' and 1B'~4A' peaks are also present, which makes it difficult to discern which peak is an R O E and which is due to residual T O C S Y effects. This question was addressed by a pulse sequence proposed by H w a n g and Shaka 2° that eliminates T O C S Y effects in R O E S Y experiments, herein referred to as T-ROESY. A T - R O E S Y spectrum of the test candidate, pentasaccharide 3, was acquired, and the only cross peaks visible from the anomeric proton 1B are those to H3 and H4 of ring A, thereby resolving this issue. The result was further c o r r o b o r a t e d by a N O E S Y experiment. The degree of overlap in the spectrum of hexasaccharide 5 is significant. However, although the chemical shifts of

122

Int. J. Biol. Macromol. Volume 17 Number 3-4 1995

IC-3

1C'-5C

[ i

IA'.2B :

5.2

5'At

* Residual T O C S Y

4.11

4;6

4~4-

4.2

IA'-2A'

- 4.0

',

3'.8

3'.~ . . . . . . ~.,~

peaks to the H2, and H 6 of the GlcpNAc rings

Figure 5 Fl slicesfrom the R O E S Y spectrum of hexasaccharidc S. The Fl slice through the chemical shift of the IA proton shows residual peaks from the IA' cross peaks and vice versa due to the close chemical shiftsof the two protons

Conforrnational analysis of Streptococcus group A cell-wall polysaccharide-related oligomers: U. C. Kreis et al. 180

180

120

120 60

60

o

qJ

0 -60

.120

-120

-180

-180

-180 -120

-00

0

60

120

180

.

-180

180

180

120

120

-1~

-~

o

60

120

180

-120

-60

0

60

120

180

60

o

0

-120

-IlO

-60

-180

-180

-180

-120

-60

0

60

120

180

-180

II Figure 7 Dynamics trajectories for the ~-(1---,2) linkage, i.e. A-B' (I) and A'-B (II) of hexasaccharide 5 from 100 ps simulations in water; constrained on the left, unconstrained on the right 180

180

120

120

Figure 6 Proton-proton contacts derived from the ROESY spectrum of hexasaccharide 5

60

60

o Table 3 Constraints input for the molecular dynamics simulation of hexasaccharide 5 Proton pair

Constrained distance (,~)

Upper limit

Lower limit

HIB HPr(a) H 1B-H Pr(b) H 1B-H5A' HIA' H2B H 1A'-H2C H1B'-H3A' H 1B'-H4A' H 1B'-H5A H 1A-H2B' H 1A-H2C' H1C H2B H1C-H3B H1C' H2B' H1C'-H3B'

2.79 2.44 2.64 2.22 3.06 2.38 2.77 2.59 2.30 3.05 2.95 2.18 2.85 2.20

2.94 2.69 2.84 2.42 3.26 2.23 2.97 2.74 2.45 3.35 3.30 2.33 3.05 2.35

2.69 2.34 2.54 2.12 2.91 2.28 2.57 2.44 2.20 2.90 2.65 2.08 2.65 2.10

'P

-60

-60

-120

-120

-180 -180

-180 -120

-60

0

60

120

180

-180

-120

-60

0

60

120

180

Figure 8 Dynamics trajectories for the ~t-(l~3) linkage, i.e. B'-A' of hexasaccharide 5 from lOOps simulations in water; constrained on the left, unconstrained on the right 180

180 ,

120 60

Wo °~ -1~

-120

-180

-180 -180

Molecular dynamics simulations were performed at 3 0 0 K for lOOps, after 10ps of heating and 10ps of equilibration, with and without the constraints. The coordinates were plotted every 80 steps with a time-step of 1 fs, thus creating 1250 frames. Again, the results will be discussed using the hexasaccharide 5 as an example. Figures 7-9 present the trajectories of the intersaccharidic torsion angles • and • which are defined as follows: = H 1- C 1- O 1- C x and • = C 1- O 1 - C x - H x where x = 2 or 3 depending on the linkage. As expected, the unconstrained dynamics simulations display more flexibility, although this effect is not very pronounced, thereby indicating a fairly rigid structure. In no case was a transition to another minimum energy conformation observed, suggesting that the

q

o

-120

-~

0

60

120

-180

180

-l~

-60

0

60

120

180

-120

-60

0

60

120

180

I

q'

180

IH

120

120

60

6O

0 -60 -120 -180

-180 -180 -120

-60

0

N

120

180

.180

II

Figure 9 Dynamics trajectories for the fl-(l~3) linkage, i.e. C-B (I) and C'-B' (II) of hexasaccharide 5 from 100 ps simulations in water; constrained on the left, unconstrained on the right

Int. J. Biol. Macromol. Volume 17 Number :3~ 1995

123

Conformational analysis of Streptococcus group A cell-wall polysaccharide-related oligomers. U. C. Kreis et al. link of hexasaccharide 5 and of the B' A' link of hexasaccharide 4 are low in comparison because the minimum values for both angles are negative, i.e. - 2 0 and - 2 4 : , respectively. The angles in Table 5 display a higher degree of variability, and unusual deviations such as the tp angles of the A ' - B link of c o m p o u n d 3 and of the B-A link of tetrasaccharide 2 can again be explained by the negative minimum values of the angles. The next step in the protocol (see Figure 3) involves calculation of averaged distances from the constrained as well as the unconstrained molecular dynamics simulations and calculation of R O E S Y intensities using the C R O S R E L p r o g r a m 24. The p r o t o n p r o t o n distances were averaged over all dynamics frames according to equation (2). Two serious limitations are the lack of experimentally determined correlation times and the assumption of isotropic re-orientation. Due to limited quantities of sample available, the measurement of 13C T 1 values was impossible at this stage. In order to obtain values for r, and the leakage rate R e, C R O S R E L permits a grid search in which these variables are incremented and compared to experimental intensities 24. Three independent grid searches were performed by including different sets of experimental cross peaks for comparison. First, only intra-ring cross peaks with defined distances such as H 1 - H 2 of the rhamnose residues and H 1 - H 3 or H I - H 5 of the N-acetylglucosamine residues were used for fitting. The drawback here is that the deviations in the experimental R O E build-up curves are quite large, rendering these curves far from g o o d for fitting. Second, inter-ring cross peaks across the interglycosidic linkages were employed. F o r these, the experimental R O E build-up data are well defined but the distances are dependent on the conformation. The resulting r, values for inter- and intra-ring build-up are quite different and range from 50 to l t00ps. Because of these large deviations, a third grid search was performed in which intensities from inter- and intra-ring cross peaks were included in the fit. In addition, the minimal and maximal

starting conformation represents a stable low-energy structure for each compound. The variation of both interglycosidic angles is similar for all linkages with a scatter of 30 to 50 °. The values of the q) and u? angles for both the c~-(l--*2) linkages (Figure 7) range from - 2 0 to 71 °, their average values being 11 and 24 ° for qb and 18 and 24 ° for u?. This conformation is in agreement with the observed R O E s about this linkage as well as with the absence of the R O E between I A ' - I B . F o r the unconstrained dynamics simulation, the values are 12 and 45 ° for • and 28 and 37 ° for tF. In the ct-(l-*3) linkage, the average • angle is 45 ° in the constrained and 31 ~ in the unconstrained simulation; the values for the • angle are 38 and 37 °, respectively. This is consistent with the observed R O E s between 1 B - 3 A and 1B~,A and the finding that the 1B-2A and I B - 5 A cross peaks are artefacts. In addition, an R O E is observed between protons 2A and 5B that corroborates the stated angles. The same discussion is valid for the B' A' link. The strongest peak observed in the R O E S Y spectrum for both the fl-(1-*3) linkages, i.e. C - B and C' B', is the I C - 3 B cross peak. The other R O E defining this linkage is the 1C-2B peak. Both are consistent with the averaged calculated angles of 49°,39 ° and 43°,47 ° for the (I) angles and - 3 3 ° , - 3 9 ° and - C , - 3 2 ° for the • angles from the constrained and unconstrained dynamics simulations, respectively. A final piece of evidence that supports the existence of the described conformation about the branch point A ' - B C is the inter-residue cross peak ! A ' - 2 C . The restricted flexibility about this branched unit appears to be a well-defined conformational feature and is likely to dictate a crucial element of the group A Streptococcus polysaccharide epitope 13.16. The average angles of all the oligosaccharides 1-5 from the constrained dynamics simulations are shown in Table 4, whereas those from the unconstrained simulations are shown in Table 5. The majority of the values are consistent and only the • angles of the A' B

Table 4 Comparison of the average qb and • angles derived from constrained molecular dynamics simulations of compounds 1 5 ~-(1-,3)

~-11--,2) A'-B Compound TriTetraPentaHexalHexa2-

1 2 3 4 5

A B'

•

•

~

34 53 26 46 11

20 23 19 31 18

B-A ~

24

fl-(l --,3) B' A'

~

~

~

37

26

50

31

55 9 45

24

C B

C' B'

~

@

~

@

16 50 38

37 31 56 42 49

37 -40 23 41 - 33

22 26 39

43 27 39

Table 5 Comparison of the average • and • angles derived from unconstrained molecular dynamics simulations of compounds 1-5

~-(1--,2)

~t-(1-~3)

A' B

Compound TriTetraPentaHexalHexa2-

124

1 2 3 4 5

A-B'

@

•

69 66 39 76 12

31 41 - 19 25 28

q)

45

B-A

~

B'-A'

~

~

55

-7

12

41

37

Int. J. Biol. Macromol. Volume 17 Number 3-4 1995

fl-(1---*3)

@

75 36 31

C-B

C'-B'

q'

q~

hu

14 32 37

36 46 50 47 43

--3 -33 -53 -31 -6

75 77 47

-23 --22 32

Conformational analysis of Streptococcus group A cell-wall polysaccharide-related oligomers: U. C. Kreis et al. values for zc and R L had to be chosen so that they came close to the values that would be expected for these types of compound. The % and R L values from the latter search are shown in Table 6. The ~ values thus obtained range from 160 to 750ps for compounds 1-5. The NOESY spectrum of the pentasaccharide 3 shows a negative N O E which corresponds to a correlation time of >400 ps. The R L values range from 0 to 10 s -~, the maximum value provided for the grid searches. The variation in ~¢ and R L values produced by the grid search with CROSREL seems to indicate that the physical meaning of these data should be treated with caution and that they should instead be perceived as variables to produce a good fit between calculated and experimental data. It is difficult to estimate the error introduced by the assumption of a single z~. CROSREL has some provisions for the inclusion of anisotropic motion but this involves calculation of the z¢ value from long dynamics simulations. This did not seem feasible at present. In addition, it has been shown that the anisotropic zc factors did not produce an appreciable improvement in the fit between experimental and calculated build-up curves 24. Using the r~ and RL values shown in Table 6, ROE build-up curves were calculated using CROSREL. The results will again be presented and discussed using hexasaccharide 5 as an example. The calculated data had to be scaled in order to have the same magnitude as the experimental data. The scaling factor was obtained from the slope achieved by plotting the calculated ROE intensities of the diagonal peaks versus the experimental data. Figure 10 shows the experimental and calculated build-up curves for intra-ring, inter-glycosidic and inter-residue ROE intensities for compound 5. The shape

Table 6

Correlationtimeszc,leakagerates RL, and Rwfactorsobtained from CROSREL grid searches calibrated with intra-ring and interglycosidic ROEs of compounds 1-5 Dynamics (constrained)

Compound

Trisaccharide Tetrasaccharide Pentasaccharide Hexasaccharide Hexasaccharide

Table

1 2 3 4 5

Dynamics (no constraints)

re (ps)

RL (s- 1)

Rw

Zc (ps)

RL (s- 1)

R.

160 380 650 610 655

10 10 0 0 0

0.25 0.51 0.62 0.44 0.67

160 400 650 530 750

9 10 7.5 10 0

0.68 0.45 0.89 0.46 0.80

and relative intensity for most of the build-up curves calculated from the constrained dynamics simulation compare reasonably well with the experimental curves. For most of the inter-ring interactions, the computed data are smaller, i.e. more negative, than the experimental values. This fact is explained by the differences between the r c values in Table 6 that were used in this calculation and the zc values obtained from the grid search to fit the inter-ring interactions specifically, i.e. a value of 655 ps versus 345 ps. As anticipated, the deviation between calculated and experimental data is somewhat larger for the unconstrained dynamics simulation. Nevertheless, the agreement between the experimental and computed ROE build-up curves is good enough to suggest that the averaged structure from the constrained molecular dynamics simulation represents a valid model for the structure of hexasaccharide 5 in solution. It is obvious that inter-residue ROE contacts are crucial for validation of the averaged structure as a model for the conformation of the oligosaccharides 1-5 in solution. Figures 11 and 12 present the experimental and calculated ROE build-up curves of the constrained dynamics simulations for inter-glycosidic contacts, i.e. the H 1 - H 2 interaction for the ct-(l~2) linkage and the H1-H3 interaction for the ~-(1 ~ 3 ) and fl-(1 --*3) linkages, and inter-residue contacts for the remaining compounds 1~1. The z¢ values and leakage rates shown in Table 6 were used for the CROSREL calculations. The ROE build-up curves calculated using CROSREL are only qualitative. Most of the experimentally observed trends are reproduced successfully and some of the calculated curves compare well with the experimental graphs. The findings are similar to those obtained for hexasaccharide 5. CROSREL fails to reproduce the observed ROESY-TOCSY-type peaks, i.e. transfer of magnetization through a system ijk, where the protons i andj exhibit an ROE and protonsj and k are J-coupled. As a final point of comparison, the 3Jcn values of the hexasaccharide 5 were measured by both direct observation and inverse-detected 13C-1H correlations, optimized for long-range coupling 34. The results of these experiments are shown in Table 7 along with the 3 J c n values calculated from the dynamics trajectories using the averaged structure, according to Karplus-type equations proposed for 3Jcocn coupling 35. Most of the calculated coupling constants show good correspondence with the observed values, especially with those values derived from the constrained dynamics simulations. Some of the values are indicative of large variations in 3 J c r t

7 Experimentaland calculated35 3Jot values (Hz) for hexasaccharide5

Linkage

Contact

Experimental value

Dynamics (constrained)

Dynamics (unconstrained)

A' B

H1A'-C2B (~) C1A' H2B (W)

4.5 5.1

4.3 4.6

4.9 3.8

C B

H1C-C3B C1C H3B

(~) (W)

4.5 -

3.3 3.7

1.0 3.1

B'-A'

H 1B'-C3A' (~) C1B'-H3A' (hv)

5.0 5.3

2.7 3.9

1.8 4.5

A-B'

H1A-C2B ((I)) C1A-H2B' (ho)

4.0 5.1

2.9 4.6

1.9 3.8

C'-B'

H 1C'423B' (~) C1C'-H3B' (W)

4.5 -

3.7 3.1

4.3 3.7

Int. J. Biol. Macromol. Volume 17 Number 3-4 1995

125

Conformational analysis of Streptococcus group A cell-wall polysaccharide-related oligomers. U. C, Kreis et al,

----I~

5elB.2B

]

5cclB-2B

•

.;elC'-5C'

]~

_gcrlC'.SC'

Ji

-4

-6

= "~

5e1C'-5C' ]

-S

-6 -10 c -12 .14 -10 -16 -12

, 0

0.1

0.2

0.3

0.4

-18

0.5

0

0.1

T i m e (see)

0.2

0.3

0.4

0.5

Time (see)

II ----II-- S.lA'-2, 0

[]

~'-8

L

I---*

~

J

+5.1A'-2,

ScclA'.2B

4

[]

~u1A'-2B

5¢1A-2B'

0

A

5e1A-2B'

O

5culC-3B

I

O

5cctC-3B [

~

'

~~"~A

-8

i

-12

~

-12

-16

~

-16

,.

,

-20 -20 -24 -24

-28 0

0.1

0.2

0.3

0.4

0.5

0

0.1

T i m e (see)

0.2

0.3

0.4

0.5

0.4

0.5

T i m e (see)

III 0

A

SelC'-2B'

A

$cclC'-2B'

i

~ .4

[.6

ffi

-2

-6

-8

-10

, 0

0.1

0.2 Time

0.3

0.4

0.5

(sec)

-10 0

0.1

0.2 Time

0.3 (sec)

Figure 10 Selected experimental ('e') and calculated ROE build-up curves from constrained ('cc'; left column) and unconstrained ('cu': right column) dynamics simulations of hexasaccharide 5: (1) intra-ring contacts, (11) inter-glycosidic contacts, and (III) other inter-residue contacts

126

Int. J. Biol. Macromol. Volume 17 Number 3 4 1995

Conformational analysis of Streptococcus group A cell-wall polysaccharide-related oligomers." U. C. Kreis et al.

I

II .2 ~IelA'-2B

K

[]

.4

IcclA'-2B

-8

-6

"Q

-8

'~ -10

1

~ -12

~

-14

-16

~

2cclC-3B

-18

-20

-24 0.4

0

-16

-22

0.3

2cclB-3A

~ -14

-20 0.2

2elB-3A

-12

-18

0.1

• ~2elC-3B

~ -1o ~

2cclA'-2B

/~t

-6

~IelC-3B

e~

2e1A'-2B ]

A

0.5

0.1

Time (sec)

0.2 Time

0.3

0.4

0.5

(aec)

IV

III ~3elA'-2B

~4elA'-2B

[]

3cclA'-2B

[]

4cclA'-2B

•

3elB'-3A'

A

4elB'-3A'

/~t

4cclB'-3A'

-2

0

3cclC-3B

|

-4

-8 g

-6

-10

-12 -14

-8

-16 0

0.1

0.2 Time

0.3

0.4

0.5

(see)

0

0.1

0.2 Time

0.3

0.4

0.5

(see)

Figure 11 Experimental ('e') and calculated inter-glycosidic ROE build-up curves from the constrained dynamics simulations ('cc') of (I) trisaccharide l, (II) tetrasaccharide2, (III) pentasaccharide3, and (IV) hexasaccharide4

values associated with even small changes in the • and angles. The computational results for the five compounds that form the basis of this examination are consistent with each other. In addition, the calculated ROE build-up curves are in reasonable agreement with the experimental ROE intensities. We thus propose a model for the solution structure of a polysaccharide which has the following values for the inter-glycosidic torsion angles, obtained by averaging the angles in Table 4. For the ct-(1--*2) link between the rhamnose rings, A'-B or A-B', we propose a • angle of 32 ° and a W angle of 23 °. The B A link, i.e. ~-Rhap-(1--*3)-Rhap, has a • angle of 40 ° and a W angle of 32 °. Finally, the fl-(1--*3) link between the Nacetylglucosamine (ring C) and the rhamnose residue (ring B) exhibits ~ and W angles of 38 and - 3 5 °, respectively. The • angles of all the linkages are in a +gauche orientation in accord with the exo-anomeric effect24. The W angles connecting two rhamnose residues dictate +gauche conformations whereas the W angle between the N-acetylglucosamine and rhamnose rings dictates a - g a u c h e orientation.

An oligomer with 24 residues was constructed from the average * and W angles obtained from the constrained dynamics simulation of all oligosaccharides 1-5, which, if the extrapolation of conformational properties is valid, should approximate part of the structure of the polymer. The 24-mer thus constructed and displayed in Figure 13 forms a well-defined helix with an internal diameter of ~ 13 A and a pitch of -,~ 10 ,~ (measured between several pairs of arbitrarily chosen atoms across the helix), with the GlcpNAc residues aligned on the periphery of the helix. To check the validity of this model, an analysis of the chemical shift differences between the Streptococcus group A polysaccharide and the oligosaccharides was performed (Figure 14). The native polysaccharide has been characterized by two-dimensional N M R methods that have yielded complete assignment of the 1H-NMR spectrumS'9; the chemical shifts are reproduced for easy reference in Table 8. In order to account for the differences in the spectra due to the different reference compounds and temperature (55°C) used, the difference in the chemical shift between proton 1B' of hexasaccharide 5 and proton 1B of the polysaccharide was added to the

Int. J. Biol. Macromol. Volume 17 Number 3 - 4 1995

127

Conformational ana/ysis of Streptococcus group A cell-wall polysaccharide-related oligomers: U. C. Kreis et al.

I

II

0

o

,

-2 -4

-4

-6

.= lelB-5A' lcclB-5A'

-10

_AA

IelC-2B

A

lcclC.2B

~

o .= -10

IelA'-2C

I

-12

J

IcclA'-2C -12

-14 0

0.1

0.2

Time

0.3

0.4

0.5

0.1

0.2 Time

(sec)

III

0.3

0.4

0.5

0.4

0.5

(sec)

IV

0

-2

# ~.

3elC'-2B'

A

3cclC'.2B'

-4

~3elB-5A' 3cclB-SA' 0

0.1

0.2 Time

0.3

0.4

0.5

(sec)

0

0.1

0.2 Time

0.3 (see)

Figure 12 Experimental ('e') and calculated inter-residue ROE build-up curves from the constrained dynamics simulations ('cc') of (I) trisaccharide 1, (II) tetrasaccharide 2, (III) pentasaccharide 3, and (IV) hexasaccharide 4

F i g u r e 13 24-mer constructed from the average qb and tp angles obtained from the constrained molecular dynamics simulations of I-5: (a) side view; (b) top view

128

Int. J. Biol. Macromol. Volume 17 Number 3-4 1995

Conformational analysis of Streptococcus group A ceil-wall polysaccharide-related oligomers." U. C. Kreis et al. Table 8 1H-NMR chemical shifts (ppm) of the Streptococcus group A cell-wall polysaccharide 9 Unit 1H

B

A

C

1 2 3 4 5 6 6R

5.09 4.27 4.00 3.52 3.81 1.31 -

5.17 4.07 3.85 3.56 3.73 1.27 -

4.75 3.72 3.57 3.46 3.46 3.94 3.77

chemical shift of all the p r o t o n s of the o l i g o s a c c h a r i d e s 1-5. This p r o t o n was chosen as a reference because it seemed to represent the same e n v i r o n m e n t as I B in the polysaccharide. Because of the necessary a d j u s t m e n t , only chemical shift differences greater t h a n 0.05 p p m were considered to be significant. A n o t h e r factor to be t a k e n into a c c o u n t is the a s s i g n m e n t of the p o l y s a c c h a r i d e s p e c t r u m using s t a n d a r d C O S Y a n d relayed R O E S Y experiments, which, in the case of highly o v e r l a p p e d regions, will n o t p r o v i d e the a c c u r a c y afforded by the T O C S Y experiments e m p l o y e d in the present study. T h e chemical shifts of the p r o t o n s of c o m p o u n d s 1-5 were

Ring A

Ring B 0.4

0.5 0.4 0.3

0.2

0.2

0.!

1 ._

0.1

0 1A

2A

3A

4A

o

I_j 5A

.o.1 6A

IB

2B

3B

Ring C

5B

6B

5A'

6A°

Ring A'

0.2

0.2

0.1

~ 0.1

0

0 1C

2C

3C

4C

5C

6CS

6CR

1A'

2A'

3A'

Ring B'

4A'

Ring C' 0.1

0.2

0.1

0

4B

~

[ ' ~ 1B' 2B'

L. 3B'

~

L. 5B'H~ 6B'I~

-O.1

4B'

Tri 1 • Tetra w Penta [] H e x a l • Hexa2

0

1C'

2C'

3C'

4C'

5C'

6C'S 6C'R

•

2 3 4 5

Figure 14 Differences in the 1H chemical shifts between the ring protons of the polysaccharide and those in compounds 1-5

Int. J. Biol. Macromol. Volume 17 Number 3-4 1995

129

Conformational analysis of Streptococcus group A cell-wall polysaccharide-related oligomers: U. C. Kreis et al. s u b t r a c t e d from the c o r r e s p o n d i n g values for the polysaccharide so that positive k 6 values c o r r e s p o n d to an upfield shift in the oligosaccharides whereas negative values indicate a downfield shift. Analysis of the Aa values reveals that, besides the expected shielding of the a n o m e r i c p r o t o n s of specific residues due to the absence of a sugar aglycon, i.e. H 1 A in c o m p o u n d s 2 a n d 4 a n d H 1B in c o m p o u n d s I, 3 a n d 5, there are few p r o t o n s with significant A6 values. These include H2, H3 and H5 of the r h a m n o s e ring B in !, 2 a n d 5, i.e. those w i t h o u t the A-substituent, which are on the face of the ring in close p r o x i m i t y to A. Similar e x p l a n a t i o n s can be given for the chemical shift differences of H 4 A ' and H2B. O n l y the A6 values of some 1C a n d 3B' p r o t o n s ( ~ 0.1 ppm) are n o t easily explicable. Overall, the analysis tends to imply a similarity between the c o n f o r m a t i o n of the oligosaccharides e x a m i n e d a n d that of the native polysaccharide.

Conclusions The N M R spectra of the oligosaccharides 1-5 (Figure 1), subunits of the cell-wall p o l y s a c c h a r i d e of S t r e p t o c o c c u s g r o u p A, show m a n y similar features as revealed by the analysis of chemical shift differences (Figure 2). This leads to the conclusion that the c o m p o u n d s share similar c o n f o r m a t i o n a l properties. The presence of characteristic cross peaks in the R O E S Y spectra of all the c o m p o u n d s e x a m i n e d confirms this statement. There are three distinct types of linkage present in these oligosaccharides, i.e. an :~-(1--+2) link from the A- to the B-ring a n d an ~-(1--+3) link between the B- a n d the A-ring c o m p r i s i n g a rhamnopyranosyl backbone to which N-acetylglucosamine residues (C-rings) are a t t a c h e d via a fi-(1--,3) link to the B-residues. F o r each of the linkages, several true R O E contacts were o b s e r v e d (Figure 6), thus defining the c o n f o r m a t i o n a l p r o p e r t i e s of these links quite well. I n t e g r a t i o n of the R O E S Y cross peaks leads to p r o t o n p r o t o n distances which serve as initial constraints (Table 3) in molecular dynamics simulations of the oligosaccharides 1 5 in the solvated state using C H A R M m 2 2 . In total, 1250 frames of the 100 ps c o n s t r a i n e d a n d u n c o n s t r a i n e d d y n a m i c s simulations at 300 K were averaged to give the following values for the angles. All ~ angles are + ,qauche, i.e. 32, 40 a n d 38 ~: for the A - B , B A and C B links, respectively, in a c c o r d a n c e with expectations based on the e x o - a n o m e r i c effect. The same + 9 a u c h e c o n f o r m a t i o n is found for the • angles of b o t h the ~-(1 -+2) (23 '~) and ~-(1--+3) (32 °) linkages connecting the r h a m n o p y r a n o s y l units that c o m p r i s e the b a c k b o n e . The W angle of the fi-(1-+3) link between the C-ring and the B-ring is -gauche ( - 35). The p r o p o s e d m o d e l of c o n f o r m a t i o n is further c o r r o b o r a t e d by the a v e r a g e d p r o t o n - p r o t o n distances calculated from the frames of the d y n a m i c s simulations and the R O E intensities c o m p u t e d using the p r o g r a m C R O S R E L . The m i n i m a l 1H chemical shift differences observed between the ring p r o t o n s of the oligosaccharides 1-5 and the p o l y s a c c h a r i d e also validate the m o d e l of conformation.

130

Int. J. Biol. Macromol. Volume 17 Number 3-4 1995

Acknowledgements The authors are grateful to B.R. Leeflang and L.M.J. KroonB a t e n b u r g for the p r o g r a m C R O S R E L and for support. B.M.P. is also grateful to K. Bock for use of the N M R facilities at C a r l s b e r g L a b o r a t o r i e s , C o p e n h a g e n , D e n m a r k , and to B.O. Petersen for recording the spectra. F i n a n c i a l s u p p o r t was p r o v i d e d by the H e a r t and Stroke F o u n d a t i o n of British C o l u m b i a and the Yukon.

References 1

Bisno. A.L. in "Principles and Practice of Infectious Diseases {Eds G.L. Mandell, R.G. Douglas and E.J. Bennet) 2nd edn. Wiley'. New York, 1985, pp 1133 1142 2 Fischeni, V.A. Sci. Am. 1991.59 65 3 Cunningham M.W., Antone, S.M. Guliza, J.M., McManus, B.M., Fischetti, V.A. and Gauntt, C.J. Proc. Natl Acad. Sei. USA 1992. 89, 1320 1324 4 Fischetti, V.A. Clin. Microbiol. Rev. 1989, 2, 285-314 5 Dale, J.B. and Beachey, E.H.J. Exp. Med. 1985, 162, 583 591 6 Goldstein, I., Rebeyrotte, P., Parlebas, J. and Halpern, B. Nature 1968, 219, 866-868 7 Goldstein. 1., Scebat, L., Renais, J., Hadjinsky, P. and Dutartre, ,l. Israel J. Med. Sei. 1983, 19, 483M.90 8 Coligan, J.E., Kindt, T.J. and Krause, R.M. lmmunoehemislry 1978. 15, 755 760 9 Huang, D.H., Krishna, N.R. and Pritchard, D.G. Carbohydr. Res. 1986, 155, 193-199 10 Reimer,K.B. and Pinto, B.M.J. Chem. Soc. Perkin 7)'ans l 1988, 2103-2111 11 Andrews,J.S. and Pinto, B.M.J. Chem. Soe. Perkin Trans I 1990, 1785~ 1792 12 Pinto. B.M., Reimer, K.B. and Tixidre, A. Carbohy&'. Res. 1991, 210, 199-.219 13 Reimer, K.B., Harris. S.L. Varma V. and Pinto, B.M. Carbohrdr. Re.v t992, 228, 399 414 14 Marino-Albernas, J.-R., Harris, S.L., Varma, V. and Pinto, B.M. Carhohydr. Res. 1993, 245, 245-257 15 Reimer, K.B., Gidney, M.A.J., Bundle. D.R. and Pinto, B.M. Carhohvdr. Res. 1992, 232, 131 142 t6 Pinto, B.M. in 'Carbohydrate Antigens" (Eds P.J. Garegg and A.A. Lindberg) ACS Symposium Series 519, American Chemical Society,, Washington DC, 1993, pp 112-131 17 Bothner-By, A.A., Stephens, R.L., Lee, J.-M., Warren, CD. and Jeanloz, R.W.J. Am. Chem. Soe. 1984, 106, 811 813 18 Bax, A. and Davis, D.B.J. Ma.qn. Reson. 1985, 63, 207-213 19 Braunschweiler, L. and Ernst, R.R.J. Magn. Reson. 1983, 53, 521 -528 20 Hwang, T.-L. and Shaka, A.J..I. Am. Chem. Soc. 1992. !14, 3157 3159 21 Brooks, B.R., Bruccoleri, R.E., Olafson, B.D., States, D.J., Swaminathan, S. and Karplus, M. J. Comput. Chem. 1983, 4, 187-217 22 Momany, F.A., Klimkowski. V.G. and Sch/ifer, L. J. ('omput. ('hem. 1990, 11,654-662 23 Momany, F.A. and Rone, R..I. Comput. Chem. 1992, 13, 888-900 24 Leeflang. B.R. and Kroon-Batenburg, L.M.J.J. Biomol. NMR 1992, 2, 495 518 25 Lemieux.R.U. and Koto, S. Tetrahedron 1974, 30, 1933 1944 26 Shaskov. A.S. Lipkind, G,M., Knirel, Y.A. and Kochetkov, N.K. Maim. Res. Chem. 1988, 26, 735 27 Dais, P. and Perlin, A.S. Carbohydr. Res. 1986, 146, 177 191 28 De Bruyn, A. J. Carbohydr. Chem. 1991, 10, 159 180 29 Bax, A. J. Magn. Reson. 1988, 77, 134-147 30 Bock, K., Josephson, S. and Bundle, D.R.J. Chem. Soc. Perkin Trans 2 1982, 59-70 3t Widmalm, G., Byrd, R.A. and Egan. W. Carbohydr. Res. 1992, 229, 195 211 32 Mo. F. and Jensen, L.H. Aeta Crystallo,qr. 1975, B31, 2867 2873 33 Mo, F. Aeta Chem. Stand. 1979, A33, 207 218 34 Maudsley, A.A. and Ernst, R.R. Chem. Phys. Left. 1977, 50, 368 372 35 Tvaroska, I., Hricovini, M. and Petrakova, E. Carhohydr. Re.~. 1989, 189, 35%362