- Email: [email protected]
Structure and chromatographic properties of carbohydrates—III
Tetrahedron.
1962. Vol.
18. pp. 1149 to 1154.
STRUCTURE GAS-LIQUID
Pergamon
Press Ltd.
Printed
in Northern
AND CHROMATOGRAPHIC OF CARBOHYDRATES-III*
PARTITION
MOBILITIES
Ireland
PROPERTIES
OF PYRANOSE
DERIVATIVES
R. J. FERRIER~ Lawrence Radiation
Laboratory, University of California, Berkeley, California
(Received 5 April 1962) Abstract-The gas-liquid chromatographic mobilities of the isomeric tetra-0-trimethylsilylpentoses and methyl tri-0-trimethylsilylpentopyranosides have been compared and interpreted in terms of molecular conformations. Striking differences are noted between the order of elution of trimethylsilylated isomers and polyacetyl derivatives of carbohydrates.
liquid-liquid and gas-liquid chromatographic characteristics of sugars and their derivatives result not only from their fundamental properties, e.g. molecular weight and degree of substitution, but from subtler factors dependent upon stereochemical relationships of the component groupings within the molecules. It is thus a simple matter to resolve many isomeric substances using chromatographic techniques. Little attention has been paid to the factors which decide the order in which isomers travel on liquid-liquid paper chromatograms, although correlations between stereochemical features and mobilities have been noted .2j3 It is becoming evident that corresponding correlations can be made relating retention times on gas-liquid chromatograms and the structures of substituted carbohydrates. Although the physical basis of these correlations is not understood at present it is hoped that their study will help with the interpretation of other gas chromatographic data and lead to a more fundamental understanding of the phenomena involved in the application of this technique in sugar chemistry. Gunner et aL4 have convincingly shown that in the polyol acetate series the isomers with the greatest number of ester groupings on one side of the planar zig-zag carbon backbone have the highest column retention times, and that, if two isomeric compounds have the same maximum number of substituents on one side then the member having those groupings closest together will move more slowly. We have independently observed5 that the introduction of axial ester groupings round the pyranose ring in tetra-0-acetylpentoses increases retention times but that at Co, an axial substituent has the opposite effect. These two sets of results can be rationalized by noting that in both the cyclic and acyclic series the molecules in which non-bonded steric interactions are at a maximum have the highest retention times, but this may not be the fundamental connecting factor. Indeed the exceptional behaviour of compounds THE
* Part II. J. Chromatography in the Press. 1 Permanent address. Chemistry Department, Birkbeck College, London. * F. A. Isherwood and M. A. Jermyn, Biochem. J. 48,515 (1951). s R. J. Ferrier and W. G. Overend, Quart. Rev. 13, 265 (1959). 4 S. W. Gunner, J. K. N. Jones and M. B. Perry, Canud. J. Chem. 39, 1892 (1961). 6 R. J. Ferrier, Chem. & Znd. 831 (1961). 1149
1150
R. J. FERRIER
which have axial substituents at Co, and the differences in the orders in which acetylated and trimethylsilylated derivatives of carbohydrates are eluted (see below) lead us to believe that polar factors are particularly important in determining retention times. Although the methyl ethers of sugars have received most study by this technique little attention has been given to the order in which they are eluted. Bishop and Coopefl have shown however that with methyl tri- and tetra-O-methylglucopyranosides the a-isomers are consistently eluted after their anomers, but the results of Kircher7 and those obtained in this laboratory indicate that rationalization within the fully methylated pyranose series will be complex. We now report the results of an examination of some trimethylsilyl ethers of carbohydrates. Hedgley and Overends have prepared volatile trisyl ethers of several non-reducing sugar derivatives and have shown by vapour phase chromatography that free aldoses give mixed products on treatment with trimethylsilyl chloride in pyridine solution. In the case of mutarotated D-xylose we have demonstrated that four products are obtained, and that the major two (each 47 %, minor pair each 3 % of total) are the anomeric pyranose ethers .9 Trisylation of the crystalline aldose (a-pyranose form) gave products containing 80 % tetra-O-trimethylsilyl-a-D-xylopyranose. The retention times relative to tetra-0-trimethylsilyl-/?-D-xylopyranose (R,) given (Table 1) for the trisyl ethers of the D-pentotopyranoses have been measured following the assumption that the major product obtained on etherification of each of the crystalline aldopentoses is the pyranose derivative (PJ having the same anomeric configuration as the parent sugar and that the major two components (P1 and P.J of the mixtures obtained by trisylation of the mutarotated sugars are the two pyranose forms. In the cases of arabinose and lyxose these assumptions are substantiated by noting the large proportions of the P, components in the mixtures (see Fig. l), and by comparison of the retention times of the P, and PZ pairs with those of the corresponding pairs of trisylated methyl pentopyranosides (see Table 1). Poor resolution was obtained with the ethers of ribose so an uncertainty in the P, retention time must be assumed. This is indicated in Table 1. The relative retention times of the trisilylated methyl pentopyranosides are also given in Table 1. Methyl /?-lyxoside and the ribosides were not available but their retention times were determined by indirect methods. D-Lyxose was reacted under reflux with methanolic hydrogen chloride till equilibrium was established when the ratio of the a- and B-pyranosides (assuming no contribution to the total optical rotation by furanosides) was determined polarimetrically to be 89 : 11. Trisylation of the mixture gave two chromatographically resolvable compounds present in the ratio 9 : 1. The major product corresponded in retention time to methyl tri-o-trimethylsilyl-a-D-lyxopyranoside and the minor was assumed to be the anomer. From its optical rotation and reaction with periodate the equilibrium mixture of methylribosides was shown to contain predominantly the j3-pyranoside. The major peak in the chromatogram obtained from the ethers of this mixture was assumed therefore to correspond with this isomer and the second peak (poorly resolved) was thought to be the apyranoside. The possible error involved in this last assumption is indicated in Table 1. BC. T. Bishop and F. P. Cooper, Canad. J. Chem. 38,388 (1960). ’ H. W. Kircher, Analyt. Chem. 32, 1103 (1960). BE. J. Hedgley and W. G. Overend, Chem. & Znd. 378 (1960). g R. J. Ferrier and M. F. Singleton preceding paper.
Structure and chromatographic
properties
a-
D - xylose (a - pyrome)
:I I!
a-
*.
!’ :
\
1151
of carbohydrates-III
&_______
D- lyxase
(,&pyronose)
fl,____ D -aroblnose (P-pyranose)
D-obese (,B- pyranose) /3-
.________k b_________J I
I I I I
5\
‘mm__
I
0
II
5
IO
1
I5
11
20
25
’
30
”
35
40
“1
45
50
55 Time,
I(
0
5
1
IO
II
I5
20
11
25
30
11
35
40
1
45
I
50
min
FIG. 1. Separations of the trimethylsilyl ethers of the pentoses. Products from crystalline sugars - - - - - ; products from mutarotated mixtures -. The ring structures of the crystalline aldoses are parenthesized, and the peaks assigned to the pyranose derivatives are indicated. TABLE1. RELA-DVERETEN~ONAMESOF D-PENTOPYRANOSE DERIVA~WES Tetra-acetates6 (relative to tetra-0-acetyl B-xylose)
Pentose
Tetratrisyl ethers. Tritrisyl methylglycosides (Relative to tetra-0-trimethylsilyl B-xylose, Rr) -
D-Ribose D-Arabinose b-Xylose D-Lyxose
CL B a B c( B c( B
1.23 1.09 1.07 0.90 0.85 l*OO 0.95 1.10
O-48” 052 0.49 0.41 0.73 1.00 0.39 053
0*3Y 0.43 040 0.38 0.69 0.77 0.37 0.42
o These figures may contain negative errors of about 0.04 TABLE2. RELATWERETENTION AMESOF PEN~OL DERIVATIVES
Pentitol Ribitol Arabitol Xylitol
Penta-acetate4 (Relative to per&-O-a&y1 arabitol) 0.92 1.00 1.18
Penta trimethylsilyl ethers (Relative to penta-O(Relative to tetra-Otrimethylsilyl arabitol) trimethylsilyl j3-xylose, R,) 1.06 1.00 0.91
0.57 0.54 0.49
1152
R. J. FERRER
The relative retention times of the penta-0-trimethylsilylpentitols were determined ‘and are given in Table 2. The following points emerge from Tables 1 and 2 (the regularities noted in the acetate series have been discussed previously).5 1. (a) The order of elution of the pentitol ethers is the opposite of that of the corresponding acetates and (b), the separation between xylitol and arabitol is greater than that between arabitol and ribitol in both series. 2. In the tetra-trimethylsilylpentose ether series the derivatives of /I-lxyose (Cl), ,!I-ribose (Cl) and cr-arabinose (lC), which each have one bulky axial grouping (at CM, CO) and CM respectively) in the indicated conformations, have strikingly similar retention times (Rr O-53, O-52, O-49) which are much lower than that of the all-equatorial /?-xylose ether (l*OO),but greater than those of their anomers (0.39, O-48, 0.41). These last three have two axial groupings in either chair conformation. The a-xylopyranose compound has a retention time (O-73) quite different from, and higher than, those of the other ethers with one axial substituent. 3. The tri-0-trimethylsilyl glycosides can be similarly classifled into (a), the allequatorial #I-xylose derivative (R, O-77), (b), the ethers with one trimethylsilyl group axial at Cc2), C(a) or Co, (O-42, O-43, O-40), (c), compounds having two axial groupings (O-37, 0.39, O-38) and (d), methyl tri-O-trimethylsilyl-a-Dxylopyranoside (O-69). 4. Differences between the retention times of anomers is significantly smaller in the glycoside ether series than in the case of the fully trisylated ethers, so that inversion of the methoxy substituent at Co, has, as would be expected, less effect on retention times than inversion of the larger trimethylsilyloxy group. With the cyclic compounds attention has been confined to the pentopyranoses since these were most likely to reveal regularities. In derivatives of /?-xylopyranose, since the lone pairs of electrons on the ring oxygen have approximately the steric requirements of two protons,lO one might expect that changing any one of the four equatorial substituents to an axial position would have the same effect on retention times. (In particular similar effects would be expected by change of configuration at Co, or Co, and at Co, or Cuu since the line joining the mid points of the C&---C,,,
: m I
0
x0
:
x0
: :
ox
ox
i
FIG. 2.
andC044.t~~ bonds (Fig. 2) approximates to a geometrical two-fold axis of symmetry, and therefore corresponding substituents on these pairs of carbon atoms have similar steric environments). lo N. Baggett, J. S. Brimacombe, A. B. Foster, M. Stacey and D. H. Whiffen, J. Chem. Sot. 2574(1960).
Structure and chromatographic
properties
of carbohydrates-III
1153
In fact, equatorial to axial inversions at Cfa), Cc,, or Cc,, have very similar effects both the acetate and the tetra-0-trimethylsilyl series although (for reasons unknown) these operate in opposite directions in the two classes. The presence of the ring oxygen atom is assumed to be responsible for the anomaly which arises when a similar change takes place at Co,, but whether its effect is a direct electronic one which influences the character of the Co, grouping or whether dipolar interaction between this grouping and dipole of the ring oxygen is the cause is not clear. Because of the symmetry element mentioned above the measured effect of equatorial to axial inversion at Co, can be resolved into two factors (a) a factor resulting in
TABLE 3. STEREOCHEMICALRELATIONSHPSANDCHROMATOGRAPHJCBEHAVIOUR
Effects on retention times Relationship
of substituents
Staggered on either side of zigzag polyol chain One axial group at Cpj, Ccsl or Co in pyranose ring One axial group at Cu, in pyranose ring
Two axial groupings in pyranose ring
0 Relative to compounds
Acetyl esters
Tetra trimethylsilyl ethers
Decreasea
Increasea
Increaseb
Decreaseb
Decreaseb [Increase due to axial nature of c (11 substituent (factor A) < decrease due to effect of neighbouring ring oxygen (factor B)] Dependent on conformationb (whether C(,,-OAc is axial or equatiorial)
with substituents
together on one
Modified decreaseb due to A >Increase
[Decrease due to B]
Greater decreaseb than that caused by one axial group.
side of the chain.
b Relative to the correspondingall-equatorialcompounds. from the group’s resultant axial nature (and equal to that found for the other axial substituents; A in Table 3) and (b), a modifying factor (B) due to the proximity of the ring oxygen atom. In the case of the tetra-O-acetates5 factor B opposes and is greater than factor A while in the tetra-0-trisyl series where electronic factors would be expected to have relatively less influence, B again opposes but now is less than A. There exists therefore the anomalous situation in which the same changes in the steric interrelationships between the acetoxy and trimethylsilyloxy groups in substituted polyols and pyranose rings influence gas-liquid chromatographic behaviour in entirely opposite ways. This is most likely to be resultant upon the largely different polarities of these two groupings. Similar detailed comparisons between the relative chromatographic properties of trisylated isomers and methylated sugars have not been carried out but marked differences would be expected since Bishop and Cooper11 find that the order of elution of the methyl tri-0-methylxylosides is ,3-pyranoside, a-pyranoside, b-furanoside, a-furanoside, while in the trimethylsilyl series the /L pyranose ether has longest retention time and the two minor products of the trimethylsilylation of xylose (believed to be the furanose derivatives) are eluted first. (Fig. 1.) The different effects noted are summarized in Table 3. I1C. T. Bishop and F. P. Cooper, Canad.J.
Chem. 40,224 (1962).
1154
R. J. Fzrous~
The physical properties of the systems which relate stereochemical changes to differences in retention times are not understood at present but experiments designed to help an understanding are in progress. EXPERIMENTAL The gas-liquid chromatography was carried out using the “Aerograph” gas chromatograph (Wilkins Instrument and Research, Inc.) at 180” on a column (0.6 cm I.D.) packed with Apiezon L grease (20 ‘A) on fire brick. The flow rate (Helium) was 45 m.l/min and the relative retention times were determined using internal standards of mixed a- and b-tetra-0-trimethylsilylxylopyranoses or where convenient penta-0-trimethylsilylribitol. They were reproducible to hO.01. The trimethylsilylation reactions were carried out as described in the preceding paper and the ethers were injected into the chromatograph without the aid of a solvent. Acknowledgements-The work described in this paper was sponsored by the U.S. Atomic Energy Commission and was carried out during tenure of a N.A.T.O. Fellowship awarded by D.S.I.R. The author thanks Professor Melvin Calvin for his interest and for providing laboratory facilities, Dr. N. R. Williams for a sample of methyl a-o-arabinopyranoside, and Dr. J. C. P. Schwarz for helpful criticism and comment. The Governors of Birkbeck College are thanked for granting leave of absence.
1962. Vol.
18. pp. 1149 to 1154.
STRUCTURE GAS-LIQUID
Pergamon
Press Ltd.
Printed
in Northern
AND CHROMATOGRAPHIC OF CARBOHYDRATES-III*
PARTITION
MOBILITIES
Ireland
PROPERTIES
OF PYRANOSE
DERIVATIVES
R. J. FERRIER~ Lawrence Radiation
Laboratory, University of California, Berkeley, California
(Received 5 April 1962) Abstract-The gas-liquid chromatographic mobilities of the isomeric tetra-0-trimethylsilylpentoses and methyl tri-0-trimethylsilylpentopyranosides have been compared and interpreted in terms of molecular conformations. Striking differences are noted between the order of elution of trimethylsilylated isomers and polyacetyl derivatives of carbohydrates.
liquid-liquid and gas-liquid chromatographic characteristics of sugars and their derivatives result not only from their fundamental properties, e.g. molecular weight and degree of substitution, but from subtler factors dependent upon stereochemical relationships of the component groupings within the molecules. It is thus a simple matter to resolve many isomeric substances using chromatographic techniques. Little attention has been paid to the factors which decide the order in which isomers travel on liquid-liquid paper chromatograms, although correlations between stereochemical features and mobilities have been noted .2j3 It is becoming evident that corresponding correlations can be made relating retention times on gas-liquid chromatograms and the structures of substituted carbohydrates. Although the physical basis of these correlations is not understood at present it is hoped that their study will help with the interpretation of other gas chromatographic data and lead to a more fundamental understanding of the phenomena involved in the application of this technique in sugar chemistry. Gunner et aL4 have convincingly shown that in the polyol acetate series the isomers with the greatest number of ester groupings on one side of the planar zig-zag carbon backbone have the highest column retention times, and that, if two isomeric compounds have the same maximum number of substituents on one side then the member having those groupings closest together will move more slowly. We have independently observed5 that the introduction of axial ester groupings round the pyranose ring in tetra-0-acetylpentoses increases retention times but that at Co, an axial substituent has the opposite effect. These two sets of results can be rationalized by noting that in both the cyclic and acyclic series the molecules in which non-bonded steric interactions are at a maximum have the highest retention times, but this may not be the fundamental connecting factor. Indeed the exceptional behaviour of compounds THE
* Part II. J. Chromatography in the Press. 1 Permanent address. Chemistry Department, Birkbeck College, London. * F. A. Isherwood and M. A. Jermyn, Biochem. J. 48,515 (1951). s R. J. Ferrier and W. G. Overend, Quart. Rev. 13, 265 (1959). 4 S. W. Gunner, J. K. N. Jones and M. B. Perry, Canud. J. Chem. 39, 1892 (1961). 6 R. J. Ferrier, Chem. & Znd. 831 (1961). 1149
1150
R. J. FERRIER
which have axial substituents at Co, and the differences in the orders in which acetylated and trimethylsilylated derivatives of carbohydrates are eluted (see below) lead us to believe that polar factors are particularly important in determining retention times. Although the methyl ethers of sugars have received most study by this technique little attention has been given to the order in which they are eluted. Bishop and Coopefl have shown however that with methyl tri- and tetra-O-methylglucopyranosides the a-isomers are consistently eluted after their anomers, but the results of Kircher7 and those obtained in this laboratory indicate that rationalization within the fully methylated pyranose series will be complex. We now report the results of an examination of some trimethylsilyl ethers of carbohydrates. Hedgley and Overends have prepared volatile trisyl ethers of several non-reducing sugar derivatives and have shown by vapour phase chromatography that free aldoses give mixed products on treatment with trimethylsilyl chloride in pyridine solution. In the case of mutarotated D-xylose we have demonstrated that four products are obtained, and that the major two (each 47 %, minor pair each 3 % of total) are the anomeric pyranose ethers .9 Trisylation of the crystalline aldose (a-pyranose form) gave products containing 80 % tetra-O-trimethylsilyl-a-D-xylopyranose. The retention times relative to tetra-0-trimethylsilyl-/?-D-xylopyranose (R,) given (Table 1) for the trisyl ethers of the D-pentotopyranoses have been measured following the assumption that the major product obtained on etherification of each of the crystalline aldopentoses is the pyranose derivative (PJ having the same anomeric configuration as the parent sugar and that the major two components (P1 and P.J of the mixtures obtained by trisylation of the mutarotated sugars are the two pyranose forms. In the cases of arabinose and lyxose these assumptions are substantiated by noting the large proportions of the P, components in the mixtures (see Fig. l), and by comparison of the retention times of the P, and PZ pairs with those of the corresponding pairs of trisylated methyl pentopyranosides (see Table 1). Poor resolution was obtained with the ethers of ribose so an uncertainty in the P, retention time must be assumed. This is indicated in Table 1. The relative retention times of the trisilylated methyl pentopyranosides are also given in Table 1. Methyl /?-lyxoside and the ribosides were not available but their retention times were determined by indirect methods. D-Lyxose was reacted under reflux with methanolic hydrogen chloride till equilibrium was established when the ratio of the a- and B-pyranosides (assuming no contribution to the total optical rotation by furanosides) was determined polarimetrically to be 89 : 11. Trisylation of the mixture gave two chromatographically resolvable compounds present in the ratio 9 : 1. The major product corresponded in retention time to methyl tri-o-trimethylsilyl-a-D-lyxopyranoside and the minor was assumed to be the anomer. From its optical rotation and reaction with periodate the equilibrium mixture of methylribosides was shown to contain predominantly the j3-pyranoside. The major peak in the chromatogram obtained from the ethers of this mixture was assumed therefore to correspond with this isomer and the second peak (poorly resolved) was thought to be the apyranoside. The possible error involved in this last assumption is indicated in Table 1. BC. T. Bishop and F. P. Cooper, Canad. J. Chem. 38,388 (1960). ’ H. W. Kircher, Analyt. Chem. 32, 1103 (1960). BE. J. Hedgley and W. G. Overend, Chem. & Znd. 378 (1960). g R. J. Ferrier and M. F. Singleton preceding paper.
Structure and chromatographic
properties
a-
D - xylose (a - pyrome)
:I I!
a-
*.
!’ :
\
1151
of carbohydrates-III
&_______
D- lyxase
(,&pyronose)
fl,____ D -aroblnose (P-pyranose)
D-obese (,B- pyranose) /3-
.________k b_________J I
I I I I
5\
‘mm__
I
0
II
5
IO
1
I5
11
20
25
’
30
”
35
40
“1
45
50
55 Time,
I(
0
5
1
IO
II
I5
20
11
25
30
11
35
40
1
45
I
50
min
FIG. 1. Separations of the trimethylsilyl ethers of the pentoses. Products from crystalline sugars - - - - - ; products from mutarotated mixtures -. The ring structures of the crystalline aldoses are parenthesized, and the peaks assigned to the pyranose derivatives are indicated. TABLE1. RELA-DVERETEN~ONAMESOF D-PENTOPYRANOSE DERIVA~WES Tetra-acetates6 (relative to tetra-0-acetyl B-xylose)
Pentose
Tetratrisyl ethers. Tritrisyl methylglycosides (Relative to tetra-0-trimethylsilyl B-xylose, Rr) -
D-Ribose D-Arabinose b-Xylose D-Lyxose
CL B a B c( B c( B
1.23 1.09 1.07 0.90 0.85 l*OO 0.95 1.10
O-48” 052 0.49 0.41 0.73 1.00 0.39 053
0*3Y 0.43 040 0.38 0.69 0.77 0.37 0.42
o These figures may contain negative errors of about 0.04 TABLE2. RELATWERETENTION AMESOF PEN~OL DERIVATIVES
Pentitol Ribitol Arabitol Xylitol
Penta-acetate4 (Relative to per&-O-a&y1 arabitol) 0.92 1.00 1.18
Penta trimethylsilyl ethers (Relative to penta-O(Relative to tetra-Otrimethylsilyl arabitol) trimethylsilyl j3-xylose, R,) 1.06 1.00 0.91
0.57 0.54 0.49
1152
R. J. FERRER
The relative retention times of the penta-0-trimethylsilylpentitols were determined ‘and are given in Table 2. The following points emerge from Tables 1 and 2 (the regularities noted in the acetate series have been discussed previously).5 1. (a) The order of elution of the pentitol ethers is the opposite of that of the corresponding acetates and (b), the separation between xylitol and arabitol is greater than that between arabitol and ribitol in both series. 2. In the tetra-trimethylsilylpentose ether series the derivatives of /I-lxyose (Cl), ,!I-ribose (Cl) and cr-arabinose (lC), which each have one bulky axial grouping (at CM, CO) and CM respectively) in the indicated conformations, have strikingly similar retention times (Rr O-53, O-52, O-49) which are much lower than that of the all-equatorial /?-xylose ether (l*OO),but greater than those of their anomers (0.39, O-48, 0.41). These last three have two axial groupings in either chair conformation. The a-xylopyranose compound has a retention time (O-73) quite different from, and higher than, those of the other ethers with one axial substituent. 3. The tri-0-trimethylsilyl glycosides can be similarly classifled into (a), the allequatorial #I-xylose derivative (R, O-77), (b), the ethers with one trimethylsilyl group axial at Cc2), C(a) or Co, (O-42, O-43, O-40), (c), compounds having two axial groupings (O-37, 0.39, O-38) and (d), methyl tri-O-trimethylsilyl-a-Dxylopyranoside (O-69). 4. Differences between the retention times of anomers is significantly smaller in the glycoside ether series than in the case of the fully trisylated ethers, so that inversion of the methoxy substituent at Co, has, as would be expected, less effect on retention times than inversion of the larger trimethylsilyloxy group. With the cyclic compounds attention has been confined to the pentopyranoses since these were most likely to reveal regularities. In derivatives of /?-xylopyranose, since the lone pairs of electrons on the ring oxygen have approximately the steric requirements of two protons,lO one might expect that changing any one of the four equatorial substituents to an axial position would have the same effect on retention times. (In particular similar effects would be expected by change of configuration at Co, or Co, and at Co, or Cuu since the line joining the mid points of the C&---C,,,
: m I
0
x0
:
x0
: :
ox
ox
i
FIG. 2.
andC044.t~~ bonds (Fig. 2) approximates to a geometrical two-fold axis of symmetry, and therefore corresponding substituents on these pairs of carbon atoms have similar steric environments). lo N. Baggett, J. S. Brimacombe, A. B. Foster, M. Stacey and D. H. Whiffen, J. Chem. Sot. 2574(1960).
Structure and chromatographic
properties
of carbohydrates-III
1153
In fact, equatorial to axial inversions at Cfa), Cc,, or Cc,, have very similar effects both the acetate and the tetra-0-trimethylsilyl series although (for reasons unknown) these operate in opposite directions in the two classes. The presence of the ring oxygen atom is assumed to be responsible for the anomaly which arises when a similar change takes place at Co,, but whether its effect is a direct electronic one which influences the character of the Co, grouping or whether dipolar interaction between this grouping and dipole of the ring oxygen is the cause is not clear. Because of the symmetry element mentioned above the measured effect of equatorial to axial inversion at Co, can be resolved into two factors (a) a factor resulting in
TABLE 3. STEREOCHEMICALRELATIONSHPSANDCHROMATOGRAPHJCBEHAVIOUR
Effects on retention times Relationship
of substituents
Staggered on either side of zigzag polyol chain One axial group at Cpj, Ccsl or Co in pyranose ring One axial group at Cu, in pyranose ring
Two axial groupings in pyranose ring
0 Relative to compounds
Acetyl esters
Tetra trimethylsilyl ethers
Decreasea
Increasea
Increaseb
Decreaseb
Decreaseb [Increase due to axial nature of c (11 substituent (factor A) < decrease due to effect of neighbouring ring oxygen (factor B)] Dependent on conformationb (whether C(,,-OAc is axial or equatiorial)
with substituents
together on one
Modified decreaseb due to A >Increase
[Decrease due to B]
Greater decreaseb than that caused by one axial group.
side of the chain.
b Relative to the correspondingall-equatorialcompounds. from the group’s resultant axial nature (and equal to that found for the other axial substituents; A in Table 3) and (b), a modifying factor (B) due to the proximity of the ring oxygen atom. In the case of the tetra-O-acetates5 factor B opposes and is greater than factor A while in the tetra-0-trisyl series where electronic factors would be expected to have relatively less influence, B again opposes but now is less than A. There exists therefore the anomalous situation in which the same changes in the steric interrelationships between the acetoxy and trimethylsilyloxy groups in substituted polyols and pyranose rings influence gas-liquid chromatographic behaviour in entirely opposite ways. This is most likely to be resultant upon the largely different polarities of these two groupings. Similar detailed comparisons between the relative chromatographic properties of trisylated isomers and methylated sugars have not been carried out but marked differences would be expected since Bishop and Cooper11 find that the order of elution of the methyl tri-0-methylxylosides is ,3-pyranoside, a-pyranoside, b-furanoside, a-furanoside, while in the trimethylsilyl series the /L pyranose ether has longest retention time and the two minor products of the trimethylsilylation of xylose (believed to be the furanose derivatives) are eluted first. (Fig. 1.) The different effects noted are summarized in Table 3. I1C. T. Bishop and F. P. Cooper, Canad.J.
Chem. 40,224 (1962).
1154
R. J. Fzrous~
The physical properties of the systems which relate stereochemical changes to differences in retention times are not understood at present but experiments designed to help an understanding are in progress. EXPERIMENTAL The gas-liquid chromatography was carried out using the “Aerograph” gas chromatograph (Wilkins Instrument and Research, Inc.) at 180” on a column (0.6 cm I.D.) packed with Apiezon L grease (20 ‘A) on fire brick. The flow rate (Helium) was 45 m.l/min and the relative retention times were determined using internal standards of mixed a- and b-tetra-0-trimethylsilylxylopyranoses or where convenient penta-0-trimethylsilylribitol. They were reproducible to hO.01. The trimethylsilylation reactions were carried out as described in the preceding paper and the ethers were injected into the chromatograph without the aid of a solvent. Acknowledgements-The work described in this paper was sponsored by the U.S. Atomic Energy Commission and was carried out during tenure of a N.A.T.O. Fellowship awarded by D.S.I.R. The author thanks Professor Melvin Calvin for his interest and for providing laboratory facilities, Dr. N. R. Williams for a sample of methyl a-o-arabinopyranoside, and Dr. J. C. P. Schwarz for helpful criticism and comment. The Governors of Birkbeck College are thanked for granting leave of absence.