Biomacromolecules 2001, 2, 965-969
965
Regioselective Functionalization of Starch: Synthesis and 1H NMR Characterization of 6-O-Silyl Ethers| Katrin Petzold,† Lars Einfeldt,‡ Wolfgang Gu¨nther,† Armin Stein,§ and Dieter Klemm*,† Institute of Organic and Macromolecular Chemistry, Friedrich Schiller University of Jena, Humboldtstrasse 10, D-07743 Jena, Germany Received March 26, 2001; Revised Manuscript Received June 27, 2001
A high regioselective 6-O silylation of starch by using thexyldimethylchlorosilane (TDSCl, chlorodimethyl(2,3-dimethylbut-2-yl)silane) as bulky silylating agent in the reaction system N-methylpyrrolidone (NMP)/ ammonia was carried out and investigated. The control of the degree of substitution (DSSi), the control of the regioselectivity, and the control of the reaction pathway are described in detail. After peracetylation of the silyl ethers of starch, the distribution of the silyl and acetyl substituents was characterized not only in the anhydroglucose repeating units (AGU) but also in the nonreducing end groups (NEG) by means of multidimensional 1H NMR techniques. In both cases, the silyl substituents were detected exclusively in the 6-O position, and the acetyl groups in the 2-O and 3-O positions of the AGU and in the 2-O, 3-O, and 4-O positions of the NEG, respectively. The described 6-O-thexyldimethylsilyl (TDS) units are potentially protecting groups of the primary OH position of starches. Introduction In the case of partially functionalized polysaccharides, the distribution of the introduced groups as well as of the remaining OH groups within the anhydroglucose units (AGU) and along the polymer chains can exert a strong influence on the product properties. To evaluate this effect and to stimulate investigations on polysaccharide structures in solution as well as in the controlled formation of supramolecular architectures, it is imperative to make materials available wich appreciably differ from the functional group distribution usually encountered in commercial polysaccharides. The same holds true for the synthesis of analogues and mimics of natural polysaccharides for the purpose of structure/activity investigations or activity modification. Therefore, synthetic methods have to be developed which generate new products with regioselective patterns of functionalization as well as analytical tools for their structure characterization. Up to now, in the field of polysaccharide regiochemistry, serveral review articles and book contributions have been published. The pioneering work in this area has been performed by Horton1 and Yalpani.2 Recent papers describe the regiochemistry of cellulose3-7 and chitin.8,9 In the case of starch, the first examples of regioselective functionalization describe 6-O tritylation,10 6-O acylation,11,12 6-deoxy6-bromation,12 and 2-O acylation.13 The major structural feature of the most important polysaccharides cellulose and starch is the 1,4-linkage of their * To whom correspondence may be addressed. E-mail: c9kldi@ rz.uni-jena.de. † University of Jena. ‡ Present address: Frommannstr. 3, D-07745 Jena, Germany. § Present address: Klostermattenstr. 6a, D-79341 Kenzingen, Germany. | Dedicated to Prof. Dr. E.-G. Ja ¨ ger on the occasion of his 65th birthday.
anhydroglucose units (AGU). The main differences in their chemical reactivities are based on the β-glycosyl linkage in cellulose and on the R-glycosyl linkage in starch. Additionally, in the case of starch the well-known complex structure14s a mixture of nonbranched amylose and high-branched amylopectin (Scheme 1)smay be important and is part of the present work on polysaccharide silylation which uses starch with 28% amylose (w/w) as a typical example. Previous investigations on the silylation of starch were carried out mainly with trimethylsilyl (TMS) reagents. Trimethylsilyl starches and trimethylsilylamylose with the degree of substitution (DSSi) values between 0.75 and 3 were obtained by silylation with hexamethyldisilazane in the presence of solvents like pyridine,11 formamide,15 or liquid ammonia.16 Tris-O-TMS-starch and -amylose were prepared by treatment of the polymers with trimethylchlorosilane/pyridine17,18 or with molten N-trimethylsilylacetamide.19 With respect to a regioselective silylation, bulky silylating agents as thexyldimethylchlorosilane (TDSCl, chlorodimethyl(2,3-dimethylbut-2-yl)silane), or tert-butyldimethylchlorosilane (TBDMSCl) are of importance. In the case of TBDSMCl, a procedure for preparing TBDMS-amylose in N,N-dimethylformamide (DMF)/imidazole with DSSi of 1.62 is described as the first example for regioselective silylation at the 6-O and 2-O positions of amylose.20 The original silicon-containing reagents can be changed by the base used. In the case of amines, chlorosilanes lead to the corresponding silylamines. These intermediates formed in situ may have a specific influence on the regioselectivity of silylation.21, 22 Besides the structure of the reagent, the reaction media play a dominant role in controlling the pathway of polysac-
10.1021/bm010067u CCC: $20.00 © 2001 American Chemical Society Published on Web 08/18/2001
966
Biomacromolecules, Vol. 2, No. 3, 2001
Petzold et al.
Scheme 1. Cluster Model of Amylopectin with Nonreducing End Groups (NEG), Anhydroglucose Units (AGU), Branch Point, and Reducing End Groups
charide silylation. It is well-known that the dispersity of the polymers has a great influence on the structure of the functionalized derivatives. In our own investigations on cellulose silylation, an exclusive functionalization of the 6-OH-groups with DS values up to 1.0 was obtained despite the heterogeneous starting conditions in N-methylpyrrolidone (NMP)/ammonia.21,23 In homogeneous mediasin N,N-dimethylacetamide (DMA)/LiCl/pyridinesa lower 6-O regioselectivity of the cellulose silylation was observed. Therefore, TDS-celluloses with DS values of about 1.0 contain a remarkable amount of 2,6-di-O-TDS units.21 The silylation of oligomer R-glucanes with TBDSMCl and TDSCl, which depends on the reaction system, is described in the field of cyclodextrins. A complete 6-O silylation takes place when using pyridine as a solvent. 2,6-Di-O silylation could be obtained in a mixture of DMF with bases like pyridine or imidazole.24-26 Resulting from these facts, still open questions concerning starch silylation are as follows: Is there the chance to silylate starch with the complex structure regioselectively? Is it possible to synthesize 6-O-silyl ethers of starch suitable for cellulose silylation in the reaction system NMP/ammonia? What is the influence of the end groups of the high-branched amylopectin? Is 1H NMR spectroscopy useful for a detailed structure characterization of the products? From this point of view, the present paper describes our results of the first starch silylation regiocontrolled by bulky reagents and the dispersity in the reaction medium using NMP/ammonia. On the basis of NMR results, the pattern of substitutions of the AGU and the nonreducing end groups (NEG) could be detected separately.
Experimental Section Materials. As the starting polymer, potato starch (amylose content 28%) from Merck was used. It was dried under vacuum over potassium hydroxide at 100 °C and 0.1 Torr for 1 h. N-Methylpyrrolidone (NMP) (water content < 0.02%) and thexyldimethylchlorosilane (TDSCl) from ABCR (Karlsruhe, Germany) were used as available. All syntheses were carried out under anhydrous conditions. Measurements. The one- and two-dimensional NMR spectra were measured on a 400 MHz spectrometer (DRX 400, Bruker) with Bruker standard pulse programms and processed with Bruker software package XWINNMR in CDCl3 at 50 °C. The COSY experiment (1H/1H correlated spectrum) was performed with a double quantum filter for an effective suppression of singlets. The DSSi (EA, elemental analysis) results were calculated from silicon-contents estimated gravimetrically.27 Then 2 mL of 98% H2SO4 (4 °C) were added to 80-100 mg of the silyl starches in a platinum crucible. After 24 h at 20 °C, the mixture was heated to remove the H2SO4 completely. The SiO2 formed was taken as a measure for the Si content of the polymer. The DSSi (NMR) values were calculated from 1H NMR spectra of the peracetylated TDS starches. Silylation of Starch (1) in N-Methylpyrrolidone (NMP)/ Ammonia. A suspension of starch in NMP (12-19% (w/ w)) was saturated with gaseous ammonia at -20 °C. After the mixture was stirred for 1.5 h at -20 °C, a solution of TDSCl (1.6-4.0 mol/mol AGU) in NMP (50% (w/w)) was added drop by drop. The polymer precipitated after addition of 50% of the TDSCl/NMP solution. The mixture was kept at -20 °C for 1 and 2 h (see Table 1), respectively, and
Biomacromolecules, Vol. 2, No. 3, 2001 967
Regioselective Functionalization of Starch Table 1. Degree of Substitution (DSSi) of TDS Starches (2) Obtained with Different Molar Equivalents of TDSCl starch TDSCl silyl ether (mol/mol AGU) 2a 2b
1.6 4.0
DSSi reaction conditions -20 °C, 1 h; 20 °C, 24 hc -20 °C, 2 h; 20 °C, 24 h
EAa NMRb 1.0 1.0
1.0 1.0
a DS calculated from elemental analysis (EA) using Si-content. b DS Si Si calculated from 1H NMR spectra after peracetylation. c See Experimental Section.
Scheme 2. Silylation of Starch with Bulky Thexyldimethylchlorosilane (TDSCl) in N-methylpyrrolidone (NMP)/Ammonia Figure 1. 1H- (i) and 1D-TOCSY (ii) NMR spectra of peracetylated TDS starch (2a). Selective excitation at 5.19 ppm shows the NEG structure [2,3,4-O-Ac-6-O-TDS] in part ii.
then slowly heated to room temperature and stirred for 24 h at 20 °C. Then ammonia was removed under reduced pressure, and then silylated polymers 2 precipitated in phosphate buffer (pH 7), which were washed with water and methanol and dried up to 100 °C under vacuum. For characterization of the polymers, see Table 1. Acetylation of 6-O-TDS Starches (2) in Tetrahydrofuran (THF)/Pyridine. First, 10 mmol of starch silyl ether was dissolved under stirring in THF (7% (w/w)) and pyridine (220 mmol). Then, 10 mmol 4-N,N-(dimethylamino)pyridine and 100 mmol acetic anhydride were added. The mixture was heated at 50 °C for 7 h. After cooling to room temperature the product was precipitated in an aqueous NaHCO3 solution (2% (w/w)), washed with water, and dried up to 100 °C under vacuum. Results and Discussion Starch (1) dissolves in NMP saturated with ammonia at -20 °C, which forms a highly viscous solution. A phase separation could be observed after addition of about 50% of the silylating agent (TDSCl/NMP solution), which was independent of the molar ratio of the reagent (1.6-4.0 mol/ mol AGU). It was very difficult to stop the silylation and to isolate the products at that point due to an ongoing silylation during workup procedures. Nevertheless, we assume that the phase separation was caused by precipitation of the partially silylated starch. During the ongoing reaction, the silylated polymer dissolved in the reaction medium and formed a highly viscous solution. The TDS starches (2), DSSi ) 1.0, were isolated as described (cf. Experimental Section, Scheme 2, Table 1). They are soluble in aprotic solvents such as tetrahydrofuran, chloroform, NMP, or DMF. In a previous work,21 we investigated the behavior of TDSCl in a solution of ammonia in NMP. The corresponding thexyldimethylsilylamine was formed. This intermediate shows a sufficient silylating activity demonstrated by the reaction of cellulose dissolved in DMA/LiCl after isolation and by characterization of the silylamine.
All free OH groups were acetylated to analyze the substitution pattern of the TDS starches 2a and 2b by 1H NMR spectroscopy. The 1H NMR spectrum of the peracetylated TDS starch 2a shows uniformly structured proton signals of the AGU (Figure 1), whose assignment could be done easily by a COSY experiment (Figure 2, Table 2). All signals of the methine and methylene protons of the AGU are detectable by means of the cross-peaks (Figure 2). As a result, a completely regioselective silylation of the 6-OH groups in the AGU of the peracetylated TDS starch 2a, DSSi ) 1.0, could be observed on one hand; the uniformly low field shifts of H-2 (4.66 ppm) and H-3 (5.34 ppm) are typical of the acetylation of the two secondary OH groups of the AGU. On the other hand, the methylene protons of the AGU have also shown typically high field shifts of H-6a (4.01 ppm) and H-6b (3.77 ppm) due to the silylation of the 6-OH group (Figure 1, Table 2). The same DSSi and the same regioselective substitution pattern were detected after peracetylation of the TDS starch 2b, DSSi ) 1.0, when higher amounts of the silylating agent (4.0 mol/mol AGU) were used. That means, the silylation of starch in NMP/ammonia with TDSCl leads to the 6-O silylated product with high selectivity, independent of the excess of the silylating agent. If the cut surface in the COSY spectrum is placed lower, a very small content of units different from the main AGU [2,3-O-Ac-6-O-TDS]n will be detectable. But these signals cannot interpreted on the basis of their very low intensity in the 1H NMR spectrum (Figure 1). Only the relatively strong cross-peak [H-2′, H-3′] at 3.48, 5.32 ppm indicates a possible trace of 2-O-silylation (Figure 2). The same holds true for the strong cross-peak [H-6a′, H-6b′] at 4.45, 4.27 ppm, indicating a very small amount of peracetylated AGU caused by residual nonsilylated units (Figure 2). Furthermore, very low signals with small line widths at 5.19 (t) ppm and 4.77 (dd) ppm were identified in the 1H NMR spectrum of peracetylated TDS starch (2a). However, these signals form very strong cross-peaks in the COSY spectrum and correlate with the chemical shifts of the methine and methylene protons of the NEG of the functionalized amylopectin part of peracetylated TDS starch (2a). An isolated 1H NMR spectrum (1D-TOCSY) of the silylated and peracetylated NEG can be obtained by selective excita-
968
Biomacromolecules, Vol. 2, No. 3, 2001
Petzold et al.
Figure 2. [1H,1H]-COSY NMR spectrum of peracetylated TDS starch (2a): assigned cross-peaks of the AGU [2,3-O-Ac-6-O-TDS]n and the NEG [2,3,4-O-Ac-6-O-TDS] (methine and methylene protons are underlined). Table 2. 1H NMR Data of TDS Starch (2a) after Peracetylationa structural units of peracetylated starch silyl ethers (2a) AGU NEGb a
[2,3-OAc-6-OTDS]n [2,3,4-OAc-6-OTDS]
δ (ppm) H-1
H-2
H-3
H-4
H-5
H-6a
H-6b
5.27(d) 5.30(d)
4.66(dd) 4.77(dd)
5.34(t) 5.31(t)
3.82 5.19(t)
3.72 3.88
4.01 3.70
3.77 3.62
All spectra were recorded in CDCl3 at 50 °C. b Coupling constants J (Hz): J12 ) 3.4, J23 ) 10.2, J34 ) 10.1, J45 ) 9.3.
tion at 5.19 (t) ppm (Figure 1). The structure of the 6-Osilylated NEG could be exclusively detected. The fine spectral resolution in the 1D-TOCSY spectrum also enables the calculation of coupling constants in the NEG [2,3,4-OAc-6-O-TDS] (Table 2). That means, the silylation of secondary OH groups has never taken place. The assignment of the proton signals in the NEG of the peracetylated TDS starch (2a) can also be done by using the cross-peaks in the COSY spectrum (Figure 2). The typical downfield shifts of the methine protons H-2 (4.77 ppm), H-3 (5.31 ppm), and H-4 (5.19 ppm), are due to the acetylation of the corresponding OH groups. The high field shifts of H-6a (3.70 ppm) and H-6b (3.62 ppm) prove 100% regioselectivity by the functionalization of the primary OH groups of the NEG with TDS units (Figure 1). The branched glycosyl units could not be detected independently under the conditions described. In summary, under the reaction conditions described, the silylation of starch in NMP/ammonia exclusively proceeds at the 6-OH groups both within the glucan chains (AGU)
and at the chain ends (NEG). In contrast to the results in the reaction medium DMSO/pyridine,28 a continued silylation at the 2-OH groups is not detectable. In the case of the silylation of starch in a DMSO solution containing 2.0 mol/ mol TDSCl pyridine, the silylation proceeds with 0.6-6.0 mol/mol AGU TDSCl at room temperature. As a result of these systematic investigations on the silylation of starch with DMSO as a solvent, pyridine as a base, and TDSCl as the silylating agent, TDS starches with DSSi values between 0.6 and 1.8 have been obtained. The reaction proceeds homogeneously up to the DSSi of 0.6. With ongoing silylation, the polymers form a separate phase incorporating the silylating agent to form TDS starches with DSSi values higher than 1.0. At first, the primary OH groups were silylated up to completation by applying moderate temperatures (T < 25 °C). Subsequently, the 2-OH groups of the AGU and the 3-OH groups of the NEG were silylated. In the case of TDS starch with DSSi ) 1.1susing 1.2 mol TDSCl/mol AGUsabout 10% of the secondary OH groups were silylated.28 It is assumed that the reason for this different
Regioselective Functionalization of Starch
reaction behavior is the phase state of the silylated starch in the respective reaction medium. The TDS starch formed dissolves again in the system NMP/ammonia during the reaction. Evidently, silylation of the secondary OH groups is blocked by solvation effects. On the other hand, the TDS starches precipitated in the system DMSO/pyridine forming a separate phase together with the reactive silylating agent, which is active in further silylation of the secondary OH groups. The regioselectivity of thexyldimethylsilylation of starch in NMP/ammonia is in agreement with the functionalization behavior in the case of silyl celluloses.21,23 Both cellulose and starch are presented in a highly activated form at the beginning of the reaction and dissolve during silylation. In the case of cellulose, the maximum of DSSi ) 1.0 is not exceeded either in this reaction medium with a high excess of TDSCl. The peracetylated 6-O-TDS starches (2) gave well-resolved 1H NMR spectra. Longer transversal relaxation times and smaller signal widths allow two-dimensional NMR spectroscopy to be used as shown in Figure 2. This is of importance in an exact structural analysis as well as in the determination of the DS values of the introduced ester or ether groups. Conclusion The silylation of starch with TDSCl in NMP/ammonia represents a highly regioselective polymeranalogous reaction forming 6-O-TDS starches (2), DSSi ) 1.0, independent of an excess of the silylating agent. For the first time, this regioselectivity of starch silylation could be proved with the help of 1H NMR spectroscopy by using multidimensional techniques after peracetylation of all remaining OH groups along the glucan chain (AGU) and in the chain ends (NEG). In contrast to the starch silylation in DMSO/pyridine,28 the silylation in the reaction medium NMP/ammonia is still stopped after complete functionalization of the primary OH groups both in the AGU and in the NEG even with an excess of TDSCl. If the described silyl ethers are used as protected polysaccharides, new pathways for the preparation of regioselectively
Biomacromolecules, Vol. 2, No. 3, 2001 969
functionalized starch esters and ethers will be opened. As an example, the 2,3-di-O-methyl ether of starch will be described in a further paper. References and Notes (1) Horton, D. New DeVelopements in Industrial Polysaccharides, Proceedings of the Conference, 1984; Crescenzi, V., Dea, T. C. M., Stivala, S. S., Eds.; Gordon & Breach: New York, 1985; p 173. (2) Yalpani, M. Tetrahedron 1985, 41, 2957. (3) Nakatsubo, F. In Wood Cellulose Chemistry, 2nd ed.; Hon, D. N.S., Shiraishi, N., Eds.; Marcel Dekker, Inc.: New York, 2001; p 627. (4) Klemm, D.; Stein, A.; Heinze, T.; Philipp, B.; Wagenknecht, W. In Polymeric Materials Encyclopedia; Salamone, J. C.; Ed.; CRC Press: Boca Raton, New York, London, Tokyo, 1996; Volume 2, 1043. (5) Klemm, D.; Philipp, B.; Heinze, T.; Heinze, U.; Wagenknecht, W. ComprehensiVe Cellulose Chemistry, 1st ed.; Wiley-VCH: Weinheim, Germany, 1998; Volume 2, pp 262-302. (6) Philipp, B.; Klemm, D.; Heinze, U. Polym. News 1999, 24, 305. (7) Klemm, D.; Einfeldt, L. Macromol. Symp. 2001, 163, 35. (8) Kurita, K.; Nishimura, S.-I.; Ishi, S.; Kohgo, O.; Munakata, T.; Tomito, K.; Kobayshi, U. Front. Biomed. Biotechnol. 1993, 1, 218. (9) Kurita, K.; Hirakawa, M.; Mori, T.; Nishiyama, J. AdV. Chitin Sci. 1997, 2, 355. (10) Whistler, R. L.; Hirase, S. J. Org. Chem. 1961, 26, 4600. (11) Horton, D.; Lehmann, J. Carbohydr. Res. 1978, 61, 553. (12) Cimecioglu, A. L.; Ball, D. H.; Kaplan, D. L.; Huang, S. H. Mater. Res. Soc. Symp. Proc. 1994, 330, 7. (13) Klohr, E.; Koch, W.; Klemm, D.; Dicke, R. EP 0001035135A1, 1999. (14) Whistler, R. L.; BeMiller, J. E.; Paschall, E. F. In Starch: Chemistry and Technology, 2nd ed.; Academic Press: New York, 1984. (15) Harmon, R. E.; De, K. K.; Gupta, S. K. Starch/Sta¨ rke 1973, 25, 429. (16) Wagner, T.; Mormann, W. DE 4309297 A1 1994; Chem. Abstr. 1995, 122, 2426717. (17) Keilich, G.; Tihlarik, K.; Husemann, E. Makromol. Chem. 1968, 120, 87. (18) Kerr, R. W.; Hobbs, K. C. Ind. Eng. Chem. 1953, 45, 2542. (19) Bredereck, K.; Strunk, K.; Menrad, H. Makromol. Chem. 1969, 126, 139. (20) Mischnik, P.; Lange, M.; Gohdes, M.; Stein, A.; Petzold, K. Carbohydr. Res. 1995, 277, 179. (21) Koschella, A.; Klemm, D. Macromol. Symp. 1997, 120, 115. (22) Petzold, K. Dissertation, 1996, University of Jena, Germany. (23) Stein, A.; Klemm, D. Papier 1995, 49, 732. (24) Takeo, K.; Uemura, K.; Mitoh, H. J. Carbohydr. Chem. 1988, 7, 293. (25) Fu¨gedi, P. Carbohydr. Res. 1989, 192, 366. (26) Coleman, A. W.; Zhang, P.; Ling, C.-C.; Parrot-Lopez, H.; Galons, H. Carbohydr. Res. 1992, 224, 307. (27) Klemm, D.; Philipp, B.; Heinze, T.; Heinze, U.; Wagenknecht, W. ComprehensiVe Cellulose Chemistry, 1st ed.; Wiley-VCH: Weinheim, Germany, 1998; Volume 1, p 174. (28) Einfeldt, L.; Petzold, K.; Gu¨nther, W.; Stein, A.; Kussler, M. Macromol Chem. Phys., submitted for publication.
BM010067U
