Synthesis of 13C-Perlabeled Cellulose with more ... - ACS Publications

Sep 16, 2009 - Dissolution of cellulose, its stepwise mechanism and its molecular basics, have been in the research focus for decades and are still un...
1 downloads 16 Views 245KB Size
Download PDF
Biomacromolecules 2009, 10, 2817–2822

2817

Synthesis of 13C-Perlabeled Cellulose with more than 99% Isotopic Enrichment by a Cationic Ring-Opening Polymerization Approach Christian Adelwo¨hrer,† Toshiyuki Takano,‡ Fumiaki Nakatsubo,‡ and Thomas Rosenau*,† Department of Chemistry, University of Natural Resources and Applied Life Sciences Vienna, Muthgasse 18, A-1190 Vienna, Austria, and Graduate School of Agriculture, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan Received June 11, 2009; Revised Manuscript Received August 17, 2009 13

C-Perlabeled cellulose was obtained in a seven-step approach from 13C6-labeled D-glucose with a cationic ringopening polymerization as the key step. Isopropylidene protection, benzylation of the remaining free 3-O-position and subsequent deprotection afforded 3-O-benzyl-13C6-glucose (2). Regioselective bis-pivaloylation followed by subsequent ortho-esterification provided the precursor for the cationic ring-opening polymerization, 3-O-benzyl13 C6-glucopyranose 1,2,4-orthopivalate (4). The actual polymerization step gave a stereo- and regioregular 13Cperlabeled (1f4)-β-glucopyranan derivative, which was deprotected into fully labeled 13C-cellulose, as the cellulose II allomorph with a DP of 40, in an overall 28% yield. All reaction steps were optimized beforehand with nonlabeled compounds toward high yields and high reproducibility and the final compound was comprehensively analytically characterized.

Introduction Dissolution of cellulose, its stepwise mechanism and its molecular basics, have been in the research focus for decades and are still unsolved problems in polysaccharide chemistry. With regard to this subject there is not much more known in the pertinent literature than the standard statement that a cellulose solvent must break the strong hydrogen bond network of cellulose and must establish new H bonds to the polymer. The detailed, stepwise process of dissolution and the defined features that render a compound or a mixture a good solvent for cellulose are still unknown. A suitable and promising technique to tackle the dissolution problem on a molecular level is NMR spectroscopy, both in solution and solid state. This technique, in principle, is able to monitor the initial phases of cellulose dissolution, such as adsorption of the solvent at the polymer, attack at specific polymer sites, and swelling processes. The detection of hydrogen bonds and their changes is possible by detecting protons “via” the neighboring carbons (in a formal reversal of the well-known HMQC or HMBC experiments although with completely different pulse sequences and techniques). This approach requires perdeuterated cellulose solvents on one hand, and solutes with a degree of isotopic enrichment in 13C as high as possible. In previous work we have synthesized the 13Cperlabeled cellulosic model compound methyl 4′-O-methyl-β1 D-cellobioside-13C12 with more than 99% isotopic enrichment and studied its solvent interactions and dissolution.2 Also perlabeled cellulose solvents, such as N-methylmorpholine N-oxide (NMMO-d11 and NMMO-15N-d11),3,4 N,N-dimethylacetamide (DMAc-d9 and DMAc-15N-d9), as well as the ionic liquids 1-ethyl-3-methylimidazolium acetate (EMIM-OAc-d14) and 1-butyl-3-methylimidazolium acetate (BMIM-OAc-d18)5 * To whom correspondence should be addressed. E-mail: thomas.rosenau@ boku.ac.at. † University of Natural Resources and Applied Life Sciences Vienna. ‡ Kyoto University.

were provided. These synthesis efforts were accompanied by structural studies with an emphasis on allomorphism2 and the hydrogen bond systems of cellulosic model compounds.6 Continuing this line of studies, the step from “small” model compounds to the polymer cellulose had to be made, once more with the prerequisite of a maximum degree of 13C-enrichment. For bacterial cellulose enriched in 13C, the degree of enrichment (max 65%) proved to be insufficient for studying hydrogen bond changed upon swelling and dissolution in detail, so that we changed to a synthesis approach that is described in the following, being well aware that only the cellulose II allomorph (“regenerated” or “mercerized” cellulose) can be obtained this way, but not the cellulose I allomorph (“natural” cellulose). However, by tuning the NMR studies with the fully (>99%) labeled, synthetic cellulose II material, we hope to gain sufficiently refined techniques that are also applicable to the bacterial cellulose (cellulose I allomorph) material with lower 13C content.

Experimental Section General. Commercial chemicals were of the highest grade available and were used without further purification. Reagent-grade solvents were used for all extractions and workup procedures. Distilled water was used for all aqueous extractions and for all aqueous solutions. n-Hexane, diethyl ether, and ethyl acetate used in chromatography were distilled before use. All reactions involving nonaqueous conditions were conducted in oven-dried (140 °C, overnight) or flame-dried glassware under a nitrogen atmosphere. TLC was performed using Merck silica gel 60 F254 precoated plates. Flash chromatography was performed using Wako silica gel (C-200) (70-150 µm particle size). All products were purified to homogeneity by TLC/GCMS analysis. The use of brine refers to saturated NaCl(aq). All given yields refer to isolated, pure products. Melting points, determined on a Kofler-type micro hot stage with Reichert-Biovar microscope, are uncorrected. Elemental analyses were performed at the Microanalytical Laboratory of the Institute of Physical Chemistry at the University of Vienna. 1 H NMR spectra were recorded at 300.13 MHz (400.13 MHz, respectively) for 1H and at 75.47 MHz (100 MHz, respectively) for

10.1021/bm9006612 CCC: $40.75  2009 American Chemical Society Published on Web 09/16/2009

2818 13

Biomacromolecules, Vol. 10, No. 10, 2009

C NMR in CDCl3 as the solvent if not otherwise stated. Chemical shifts, relative to TMS as internal standard, are given in δ values, and coupling constants are given in Hz. 13C peaks were assigned by means of APT, HMQC, HMBC, and INADEQUATE spectra. The abbreviations “bn” and “piv” denote benzyl and pivaloyl moieties, respectively. The DPn data were determined by GPC analyses of the cellulose acetate with a Shimadzu LC-10 system, equipped with Shodex columns (K802, K802.5, and K805) and a Shimadzu refractive index detector (RID-10A), in CHCl3 at the flow rate (1.0 mL/min) at 40 °C. Calibration curves were obtained by using polystyrene standards (Shodex). X-ray diagrams (XRD) were recorded with a Rigaku RINT 2200 V. Determination of the MW data of nonderivatized cellulose was performed in DMAc/LiCl with a system based on multiangle laser light scattering (MALLS), as described previously in detail.7,8 1,2:5,6-Di-O-isopropylidene-r-D-glucofuranose-13C6 (1). D-Glucose13 C6 (2.00 g, 10.54 mmol) was stirred in dry acetone (50 mL) with CuSO4 (5 g) and a catalytic amount of sulfuric acid for 24 h. The solution was neutralized with concentrated aqueous ammonia and filtered. The remainder was washed thoroughly with acetone and dichloromethane. The combined organic phases were concentrated to about 20 mL in vacuo, and water (100 mL) was added. The mixture was stirred vigorously for 1 min and repeatedly extracted with dichloromethane to obtain the crude diacetonide 1. The aqueous phase was also concentrated and dried in vacuo to recover unconverted D-glucose-13C6 and monoacetonide. After purification of the crude diacetonide by recrystallization, 74% of pure compound 1 was obtained. The recovered 13C6-glucose monoacetonide and starting material were dried carefully before the reaction procedure was repeated until no starting material was recovered in the aqueous phase. Combining the purified amounts of 1, an overall yield of 2.53 g (90%) of 13C6-glucose diacetonide was obtained as a white powder: mp ) 110-112 °C. 1H NMR: δ 1.32, 1.37, 1.45, 1.50 (4 × s, 4 × 3H, CH3), 2.77 (s, br, 1H, OH), 3.99 (m, 1H, H-6a), 4.06 (m, 1H, H-4), 4.17 (m, 1H, H-6b), 4.27-4.40 (m, 2H, H-3, H-5), 4.53 (m, 1H, H-2), 5.94 (q, 3JH,H ) 3.84 Hz, 1JC,H ) 183.10 Hz,2JC,H ) 5.42 Hz, 1H, H-1). 13C NMR: δ 25.4, 26.4, 27.0, 27.1 (4 × s, CH3), 67.9 (dd, 1JC,C ) 34.0 Hz, 2JC,C ) 1.7 Hz, 13C-6), 73.6 (t, 1JC,C ) 36.9 Hz, 13C-3), 75.6 (td, 1JC,C ) 34.0 Hz, 2JC,C ) 4.0 Hz, 13C-5), 81.3 (t (br), 1JC,C ) 36.9 Hz, 13C-4), 85.3 (t (br), 1JC,C ) 34.0 Hz, 13C-2), 105.5 (td, 1JC,C ) 33.4 Hz, 2JC,C ) 4.0 Hz, 13C-1), 109.6 (s, C(CH3)2), 111.8 (s, C(CH3)2). Anal. Calcd for 13 C612.01C6H20O6 (266.28): C, 54.12; H, 7.57. Found: C, 54.32; H, 7.61. 3-O-Benzyl-D-glucopyranose-13C6 (2). Sodium hydride (60% suspension in paraffin, 1.23 g, 30.73 mmol) was added to dry DMF (70 mL) under a nitrogen atmosphere. The mixture was cooled to 0 °C and 13C-glucose diacetonide 1 (6.82 g, 25.61 mmol) dissolved in dry DMF (30 mL) was added over a period of 30 min. Then the reaction mixture was allowed to warm up to r.t. and was stirred for another 30 min. The flask was cooled down again to 0 °C and benzyl bromide (3.65 mL, 30.73 mmol) dissolved in dry DMF (30 mL) was added over a period of 30 min under constant stirring. The reaction mixture was warmed to room temperature and stirred for 3 h. After careful quenching with ice/water and stirring for 2 h, the mixture was repeatedly extracted with dichloromethane. The extracts were dried over NaSO4, concentrated, and purified by chromatography on silica gel (EtOAc/ n-hexane, v/v ) 1:5) to obtain 3-O-benzyl-13C6-glucose diacetonide (8.24 g, 90%) as a colorless oil. This intermediate was dissolved in water/ethanol (24 mL, v/v ) 3:1), acidic ion-exchange resin (DowexH+, 3.3 g) was added, and the mixture was refluxed for 5 h. After neutralization (NaOMe), the resin was filtered off and was washed with EtOAc and water extensively. The combined solvents were removed by lyophilization to provide a yellowish powder, which was recrystallized five times from dry ethyl acetate to give 5.72 g (90%) of 3-Obenzyl-D-glucopyranose-13C6 (2). The high number of recrystallization steps was required to ensure optimum yields in the subsequent regioselelctive pivaloylation: mp ) 123-125 °C. 1H NMR (D2O): δ 3.32 (t, 1H, H-2), 3.41 - 3.55 (m, 2H, H-5, H-3), 3.58 - 3.77 (m, 2H, H-6), 3.87 (m, 1H, H-4), 4.64 (d, 1H, H-1β), 4.74 and 4.80 (2 × s,

Adelwo¨hrer et al. 2H, CH2 of benzyl), 5.22 (m, 1H, H-1R), 7.36-7.50 (m, 5H, HAr). 13C NMR (D2O): δ 60.3 and 60.9 (d, 1JC,C ) 13.3 Hz, 13C-6), 62.4-64.4 (m, 13C-2), 68.9-72.2 (m, 13C-4), 75.2 and 75.4 (2 × s, CH2 in benzyl), 76.0 and 74.1 (t, 1JC,C ) 42.6 Hz, 13C-5), 84.0 and 81.5 (t, 1JC,C ) 42.0 Hz, 13C-3), 92.3 (d, 1JC,C ) 46.3 Hz, 13C-1R), 96.0 (d, 1JC,C ) 45.5 Hz, 13 C-1β), 128.9, 128.7, 128.4 (3 × s, HCAr), 137.6 (s, CAr). Anal. Calcd for 13C612.01C7H18O6 (276.28): C, 56.51; H, 6.57. Found: C, 56.81; H, 6.52. 3-O-Benzyl-2,6-di-O-pivaloyl-glucopyranose-13C6 (3). 3-O-Benzyl13 C6-glucose (2, 1.00 g, 3.72 mmol), Bu2SnO (2.03 g, 8.15 mmol), and 1.5 g of activated (150 °C, 4 h in high vacuum), powdered molecular sieves 4 Å were combined in a 100 mL flask and kept under vacuum in a desiccator overnight. Anhydrous toluene (30 mL) was added by a syringe, and the reaction mixture was stirred for 11 min at 105 °C. After cooling to r.t., pyridine (0.8 mL, freshly distilled) was added. Pivaloyl chloride (1.01 mL) was dissolved in anhydrous toluene (10 mL) and added continuously (syringe) at -15 °C within 30 min. The reaction mixture was stirred another 2.5 h at this temperature, MeOH (0.5 mL) was added, and the toluene was evaporated in vacuo (r.t.). The remainder was filtered through a short silica gel column, eluting first with dichloromethane, then with ethyl acetate/n-hexane (v/v ) 2:1), and purified by column chromatography (75 g silica gel), eluting first with dichloromethane, then with ethyl acetate/n-hexane (v/v ) 5:1) to obtain 1.31 g (81%) of 3 as a colorless oil. 1H NMR: δ 1.16-1.28 (s, br, 18H, CH3), 2.72 (d, 1H, OH-4), 3.17 (d, 1H, OH-1), 3.54 (m, 1H, H-4), 3.95 (m, 1H, H-3), 4.03 (m, H1, H-5), 4.32 and 4.38 (m, 2H, H-6), 4.71 (m, 1H, H-2), 5.35 (m, 1H, H-1), 5.19 and 5.62 (2 × s, 2H, CH2 in benzyl), 7.22-7.37 (m, 5H, HAr). 13C NMR: δ 27.1 and 27.2 (2 × s, CH3 in piv), 39.0 and 38.9 (2 × s, C in piv), 62.4 and 62.8 (d, 1JC,C ) 19.1 Hz, 13C-6), 69.1-70.2 (m, 13C-4), 75.2-73.2 (m, 13C-5), 78.9 (d, 1JC,C ) 40.2 Hz, 13C-2), 75.3 (t, CH2 in benzyl), 81.6 and 80.6 (t, br, 1JC,C ) 40.2 Hz, 13C-3), 90.3 and 96.0 (d, 1 JC,C ) 43.3 Hz, 13C-1), 127.7, 128.1, and 128.7 (s, HCAr), 137.4 (s, CAr), 178.2 and 178.4 (2 × s, COO). Anal. Calcd for 13C612.01C17H34O8 (444.51): C, 62.14; H, 7.71. Found: C, 62.37; H, 7.82. 3-O-Benzyl-6-O-pivaloyl-glucose-13C6-1,2,4-orthopivalate (4). Compound 3 (1.40 g, 3.19 mmol) was dissolved in carefully dried n-hexane/ dichloromethane (v/v ) 2:1, 30 mL) and triethyl amine (2 equiv, 0.89 mL, 6.38 mmol) as well as benzenesulfonyl chloride (1.1 equiv, 0.45 mL, 3.51 mmol) were added. The reaction mixture was slowly stirred overnight at room temperature, quickly transferred directly on a silica column (50 g of silica gel), and eluted with n-hexane/dichloromethane (v/v ) 2:1) to obtain 4 (1.08 g, 81%) as a colorless solid, which was recrystallized from dry n-hexane to afford white crystals: mp ) 72-74 °C. 1H NMR: δ 1.03 (s, 9H, CH3), 1.23 (s, 9H, Piv-CH3), 3.95 (m, 1H, H-4), 4.26 (m, 1H, H-3), 4.31-4.41 (m, 2H, H-6), 4.39-4.43 (m, 1H, H-2), 4.50 (m, 1H, H-5), 4.64 and 4.63 (s, 2H, CH2 in benzyl), 5.76 (m, 1H, H-1), 7.26-7.38 (m, 5H, HAr). 13C NMR: δ 25.1 (CH3 in piv), 27.4 (CH3 in piv), 35.7 (C in piv), 38.8 (C in piv), 64.4 (d, 1JC,C ) 44.9 Hz, 13C-6), 71.2-72.8 (m, 13C-4, 13C-2), 75.0-76.0 (m, 13C-5, 13 C-3), 75.3 (t, CH2-Ar), 97.8 (d, 1JC,C ) 34.6 Hz, 13C-1), 123.6 (s, C-CH3), 127.7, 128.1, 128.7 (s, HCAr), 137.4 (s, CAr), 179.4 (s, COO). Anal. Calcd for 13C612.01C17H32O7 (426.51): C, 64.76; H, 7.56. Found: C, 64.81; H, 7.35. 3-O-Benzyl-2,6-di-O-pivaloyl-(1f4)-β-D-(13C6-glucan) (5). Compound 4 (200 mg, 456 µmol) was dried in a glass ampule under high vacuum overnight. Dichloromethane (0.2 mL) was distilled from CaH2, degassed by freeze/thaw cycles three times, and transferred to the polymerization ampule in a vacuum line. BF3 · Et2O (2.9 µL, 5 mol %) was added via a syringe, and the ampule was sealed off and agitated overnight. The content partly solidified during that time. After opening the ampule, chloroform (10 mL) was added and was washed sequentially with saturated aqueous NaHCO3 solution, water, and brine. After drying and filtration, the solvent was concentrated in vacuo to give the substituted glucan polymer in quantitative yield. 1H NMR (CDCl3): δ 0.98 and 1.03 (CH3 in piv), 3.44 (m, H-3), 3.52-3.56 (m, H-4), 3.60-3.63 (m, H5), 3.80-3.81 (m, H-6a), 4.06 (m, H-6b), 4.24 (s, br,

Synthesis of

13

C-Perlabeled Cellulose

H-1), 4.40 and 4.96 (2 × s, CH2 in benzyl), 4.82 (m, H-2), 7.13-7.30 (HAr). 13C NMR (CDCl3): δ 26.9 and 27.0 (CH3 in piv), 35.7 and 35.8 (2 × s, C in piv), 64.4 (d, 1JC,C ) 45.5 Hz, 13C-6), 71.5-73.6 (m, 13 C-4, 13C-2), 75.4 (t, CH2 in benzyl), 76.2-77.4 (m, 13C-5), 80.6 (t, 1 JC,C ) 39.2 Hz, 13C-3), 100.1 (d, 1JC,C ) 48.3 Hz, 13C-1), 126.7, 128.0, 128.7 (s, HCAr), 137.4 (s, CAr), 179.2 (s, COO). Anal. Calcd for (13C612.01C17H32O7)n: C, 64.76; H, 7.56. Found: C, 64.68; H, 7.72. Cellulose-13C (7), (1f4)-β-D-Glucopyranan-13C6. 3-O-Benzyl-2,6di-O-pivaloyl-(1f4)-β-D-glucan-13C6 (5, 200 mg, 456 µmol) was dissolved in THF/AcOH (5 mL, v/v ) 1:1), placed in an autoclave containing Pd(OH)2 on charcoal (10%, 200 mg), and hydrogenated at 50 °C for 24 h under a H2 pressure of 0.4 MPa. The reaction mixture was filtered through Celite, which was washed thoroughly with THF/ AcOH (v/v ) 1:1). The organic phases were combined, and the solvents were removed in vacuo under repetitive addition of ethanol to obtain 150 mg (100%) of debenzylated polymer (6) as a white powder, which was insoluble in water and ethanol: alcohol insoluble polymer as a white powder. The 2,6-di-O-pivaloyl-13C-glucan (6; 150 mg, 456 µmol) was dissolved in THF/MeOH (30 mL, v/v ) 10:1), and NaOMe/MeOH (28%, 0.4 mL, 912 µmol) was added. The mixture was slowly stirred overnight at 50 °C to afford a white suspension. After neutralization with 0.1 M HCl, solvents were removed in vacuo and water (50 mL) was added. The precipitate was removed by centrifugation, washed several times with water and MeOH, and finally dried in vacuo to obtain 37 mg (58%) of cellulose-13C (7) as a white powder. From the combined aqueous and methanolic washings, 22.3 mg (35%) of cellooligosaccharides (6 < DP < 12) were obtained. For the solid state NMR spectrum of 7, see Figure 3, a detailed analysis of the NMR spectroscopic features will follow in due course in combination with the swelling/dissolution studies. Anal. Calcd for (13C6H10O5)n: C, 42.86; H, 6.00. Found: C, 42.96; H, 6.09.

Results and Discussion With the cationic ring-opening polymerization of appropriately substituted orthopivalate derivatives of glucose, it became possible to synthesize cellulose chemically9 with a relatively narrow molecular weight distribution at a predefined degree of polymerization (DP). Such chemical synthesis of cellulose seems rather demanding with regard to stereo- and regio-specific control, let alone the yield issue which becomes increasingly important in the case of the highly costly multinucleus isotopically enriched compounds. Even with a suitable 1,2,4orthopivalate precursor in hand, at least four possible linkages, (1f4)-R, (1f4)-β-, (1f2)-R-, (1f2)-β, are conceivable. A possible mixing of these variants in the glucan chains would additionally complicate the situation. The problem of stereoand regiocontrol has been solved by introduction of suitable substituents into the polymerization precursor,10,11 and 3-Obenzyl-glucose-1,2,4-orthopivalate was shown to be the optimum starting compound to obtain a stereo- and regioregularly polymerized β-(1f4)-linked glucan (cellulose) derivative that only awaited deprotection to eventually afford nonsubstituted cellulose. In the present attempt to synthesize 13C-perlabeled cellulose, D-glucose-13C6 appeared to be the only reasonable, in terms of costs and synthetic efforts, starting material. To provide amounts of the target compound sufficient for the structural studies, about 500 mg, every single step had to be as high-yielding as possible and highly reproducible. In addition, all perlabeled byproducts had to be recovered and recycled to reduce losses. The minimization of yield penalties was not only dictated by some idealistic feelings of synthetic elegance, but also simply by the down-to-earth fact of extremely high costs of the perlabeled compounds. After testing several alternatives with nonlabeled reagents, we focused on the synthesis shown in Scheme 1,

Biomacromolecules, Vol. 10, No. 10, 2009

2819

Scheme 1. Synthesis of the Polymerization Precursor, the 1,2,4-Orthopivalate Derivative 4, 13C6-Perlabeled at the Glucose Corea

a Note: the pyran ring in compound 4 is not present as 4C1 chair, but is shown in this conformation for the reason of clarity.

because each individual step in the sequence could be optimized to 90% yield and above. The commercially available 13C6-D-glucose was converted into its diacetonide 1 in near-quantitative yields according to standard procedures.12 In contrast to the literature, we failed to obtain a satisfying yield in one reaction step, so that the given high yields refer to the sum of three reaction cycles, each one using nonconverted educt and monoacetonide byproduct of the preceding step as the new starting material. Subsequent benzylation of the remaining free 3-OH group of diacetonide 1, followed by removal of the isopropylidene protecting groups provided 3-O-benzylglucose (2) as colorless crystals after repeated recrystallization (5-6 times) from EtOAc. Acidic ionexchange resin (H+, Dowex-50W) gave the best results as acidic catalyst in this deprotection step, as tested in numerous trials, in addition to the benefit of its easy removal. The repetitive crystallization was crucial for a satisfying outcome of the subsequent regioselective pivaloylation step; simple lyophilization or vacuum-drying of precipitated 3-O-benzylglucose (2) turned out to be insufficient. Reaction of 2 to the dibutyltinoxide complex under thoroughly water-free conditions gave a bright yellow reaction mixture and good yields (88%), whereas traces of water caused a brown discoloration of the mixture or even the appearance of a second, oily brownish phase, accompanied by a significant yield loss (40-60%). For the subsequent actual acylation step, several promoter/reagent couples were tested, with the simplest one, pyridine/pivaloyl chloride, giving the best yields of 2,6-bis-pivalate 3. A strongly dehydrating reagent was required to convert this intermediate into an orthopivalate as a suitable “monomer” for a cellulose-building polymerization step. Because orthopivalate 413 was highly sensitive toward traces of acids (and thus also water and moisture), a mixture of benzenesulfonyl chloride with 2 equiv of triethyl amine was used. A binary mixture of hexane and dichloromethane was used as the solvent and also further on as the eluant in the subsequent chromatographic purification step. It is imperative for the subsequent polymerization step to remove byproduct from benzenesulfonyl chloride perfectly. The use of alternative solvents followed by the concentration of the reaction mixture and mixing with a different chromatographic eluant caused formation of medium to large amounts of an intractable, oily residue, possibly consisting of products of hydrolysis and uncontrolled oligomerization of the orthopivalate. The thoroughly recrystallized product was sufficiently pure in this form to allow for prolonged storage at -20 °C. As 13C-perlabeled compounds are not routinely used, some peculiarities with regard to analytical characterization may be

2820

Biomacromolecules, Vol. 10, No. 10, 2009

Adelwo¨hrer et al.

Figure 1. Comparison of the 1H NMR spectra of diacetone glucose (top) and diacetone glucose-13C6 (bottom) showing the additional heteronuclear (H,C) couplings caused by the presence of >99% 13C.

highlighted in the following. The NMR spectra of the various glucose-13C6 derivatives are rendered far more complex by the additional heteronuclear (H-C) and homonuclear couplings (C-C). In 13C NMR spectra (1H decoupled), C-1 and C-6 appeared as doublets, while the other carbons give triplet or multiplet resonances. One example of 1H NMR spectra is given in Figure 1: while common diacetone glucose (upper trace) produces a relatively simple spectrum with the anomeric proton appearing as doublet at 5.92 ppm, the spectrum of diacetone glucose-13C6 (1) becomes more crowded (lower traces), with the anomeric proton being split into two signals by the heteronuclear coupling, with a 1JH,C coupling constant of 183 Hz. Additional long-range couplings (>1JH,C) produce the multiplet (broad) appearance of the resonances. The chemical shifts of the protons can be retrieved by comparison with the spectra of the respective nonlabeled compounds in combination with C,H-heterocorrelation experiments. Nevertheless, full signal assignment of the fully 13C6-labeled compounds (cf. Experimental Section) is more challenging than for the counterparts with natural isotopic content. The 13C NMR spectra somewhat “behave like proton spectra” in a sense that the carbon resonances are not singlets, but show multiplicities due to the presence of homonuclear C,C-couplings over 1-4 bonds, and in a sense that nearly 100% of NMRactive nuclei are present as compared to commonly about 1%, with inherent gains in spectral intensity. The direct couplings 1 J ranged between 35-45 Hz for carbons within the glucopyranose ring and between 12-19 Hz outside the ring (C6). Geminal couplings were around 4-6 Hz, and higher C,C couplings generally below 1 Hz. Due to the highest degree of isotopic enrichment, also the spectral intensity was comparable to what is usually known from proton measurements. With only four scans, a full quality carbon spectrum is obtained, and INADEQUATE experiments easily allow assigning the carbons in a very short period of time. Also, with regard to microanaly-

sis, the effect of the 13C-enrichment has to be considered. While introduction of just one 13C would hardly influence the microanalytical result, the isotopic effect in the case of 13Cperlabeled glucose and its derivatives and in the case of 13Cperlabeled cellulose is quite significant: the values of elemental composition are noticeably changed as compared to the corresponding compounds with natural isotopic abundance, and have to be considered accordingly. Polymerization of the 1,2,4-orthopivalate 4 into protected glucan 5 was a matter of extensive optimization (Scheme 2). With regard to the mechanism, it is a regio- and stereoselective ring-opening polymerization process. Under optimized conditions, the reaction was carried out in an ampule sealed under vacuum after multiple freeze/thaw cycles for degassing to achieve absolute water- and oxygen-free conditions. The solvent dichloromethane and BF3 diethylether complex as the catalyst in concentration of (5 mol %) were indispensable. Containing the monomer 4 in rather high concentration (100 mg monomer/ 0.1 mL), the reaction mixture gelled after some hours, and the Scheme 2. Synthesis of 13C-Perlabeled Cellulose (7) by Cationic Ring-Opening Polymerization of the Glucose 1,2,4-Orthopivalate Precursor 4, Followed by Deprotection Stepsa

a Note: the pyran ring in compound 4 is not present as 4C1 chair, but is shown in this conformation for the reason of clarity.

Synthesis of

13

C-Perlabeled Cellulose

Biomacromolecules, Vol. 10, No. 10, 2009

was determined by GPC as 38. Results obtained for the acetate (eluant CDCl3, calibration with polystyrene standards) and the nonderivatized cellulose (eluant DMAc/LiCl, absolute determination based on multiangle laser light scattering [MALLS]) were in complete agreement. According to the literature,10 first the pivaloyl protecting groups of 3-O-benzyl-2,6-di-O-pivaloyl glucan (5) are removed and then the benzyl groups. However, as depivaloylation progresses, the polymer starts to precipitate and the solids must be acetylated in pyridine/acetic anhydride to redissolve the polymer before further depivaloylation is performed. The yields decrease significantly with every deprotection/dissolution cycle, and the polymer becomes increasingly difficult to separate from chromophores formed. Due to the high losses during this reaction sequence - from 200 mg of monomer only 10 mg of acetylated polymer were obtained - a more suitable route had to be found for the 13C-material. Compared to the 3-O-benzyl2,6-di-O-acetyl polymer, which was readily debenzylated under ambient hydrogen pressure after 4 h, the deprotection of the glucan 5 required more severe conditions: at a H2 pressure of 0.4 MPa, in the presence of palladium hydroxide on carbon at 50 °C, debenzylation was complete after 24 h. Neither the common Pd/C catalyst nor ambient hydrogen pressure are appropriate. The debenzylated glucan 6 remains fully soluble in organic solvents due to the pivaloyl groups and can readily be washed through Celite without losing any material. GPC measurements indicated that the hydrogenation process did not cause any degradation of the polymer. Finally, the pivaloylated polymer was deprotected under Zemple´n-type conditions. A reaction temperature of 50 °C was a good compromise between precipitation of incompletely deacylated material at lower temperatures and pronounced depolymerization at reflux temperature. The precipitated polymer was collected by centrifugation and was washed thoroughly with water and MeOH. The yield of 58% of 13C-perlabeled cellulose in this step seems rather disappointing at a first glance, but 35% of the material is obtained in the form of cellooligosaccharides between 6 and 12 glucose units from the washings and is thus not lost. Glucans in the range between 12 and 30 were not observed. The yield penalty due to the removal of lower molecular weight oligosaccharides in this last purification step seemed unavoidable to obtained material of low polydispersity, that is, a narrow molecular weight distribution. The analytical data in Figure 3 confirm the presence of the cellulose II allomorph.

Figure 2. 1H-decoupled 13C NMR spectrum of 3-O-benzyl-2,6-di-Opivaloyl-β-1,4-glucan (5).

polymerization process was completed after 10 h. Aqueous workup provided colorless crystals, crystallization being induced and assisted by sonication in an ultrasonic bath for few seconds. It should be pointed out again that the selection of protecting groups as in orthopivalate 4 is crucial for the regio- and stereoselective polymerization, otherwise, (1f2)-linkages and R/β-mixtures become prominent. The absence of water and oxygen during the polymerization determines largely the degree of polymerization (DPn) of the polymer, which reached about 40. The 13C NMR spectrum of 3-benzyl- and 2,6-O-dipivaloylprotected glucan 5 is given in Figure 2, which displays the “typical” resonances of the glucopyranan backbone of a cellulose derivative, keeping in mind the signal splitting by the homonuclear couplings. The 1H NMR spectrum of polymer 5 becomes intricate due to the C-H couplings, including long-range C-H interactions. The anomeric proton resonates at 4.24 ppm (1H), the anomeric carbon at 100.1 ppm. The specific rotation was negative ([R]D ) -4°), all data sustaining the β-linkage of the glucopyranose unit. The methylene protons of the 3-O-benzyl moieties showed a strong diastereotopic splitting, resonating at 4.40 and 4.96 ppm, which can also be attributable to a steric fixation of the benzyl group. The DPn of the resulting polymer

Figure 3. X-ray diffractogram (left) and

13

C CPMAS NMR spectrum of

2821

13

C-perlabeled cellulose II (7) obtained by chemical synthesis (right).

2822

Biomacromolecules, Vol. 10, No. 10, 2009

Conclusion

Adelwo¨hrer et al.

References and Notes 13

The presented approach was able to provide C-perlabeled cellulose (7) as the cellulose II allomorph in 30% overall yield starting from D-glucose-13C6, a cationic ring-opening polymerization being the key step in the sequence. Comprehensive optimization was required to ensure high reproducibility and optimum yield in every step. The material will be used for studies of cellulose swelling and dissolution, mainly in combination with liquid and solid state NMR techniques, which will be reported in due course. About 500 mg of labeled material, as obtained by repetitive synthesis according to Schemes 1 and 2, will be sufficient to meet the needs of these experiments, in addition the material can nearly be fully recovered and reused as swelling and dissolution are fully reversible. Acknowledgment. The financial support by the Austrian “Fonds zur Fo¨rderung der wissenschaftlichen Forschung”, project P-17426, by the Austrian Christian-Doppler Research Society (CD lab “Advanced cellulose chemistry and analytics”) and by the Japanese Society for the Promotion of Science (JSPS) is gratefully acknowledged. C.A. is indebted to Dr. Hiroshi Kamitakahara, Graduate School of Agriculture, Kyoto University, for his continuing support during this project.

(1) Yoneda, Y.; Kawada, T.; Rosenau, T.; Kosma, P. Carbohydr. Res. 2005, 340, 2428–2435. (2) Malz, F.; Ja¨ger, C.; Yoneda, Y.; Kosma, P.; Rosenau, T. Carbohydr. Res. 2007, 342, 65–70. (3) Adelwo¨hrer, C.; Yoneda, Y.; Nakatsubo, F.; Rosenau, T. J. Labeled Comp. Radiopharm. 2008, 51, 28–32. (4) Adelwo¨hrer, C.; Yoneda, Y.; Takano, T.; Nakatsubo, F.; Rosenau, T. Cellulose 2008, 16, 139–150. (5) Yoneda, Y.; Ebner, G.; Takano, T.; Nakatsubo, F.; Potthast, A.; Rosenau, T. J. Labeled Comp. Radiopharm. 2009, 52, 223-226. (6) Yoneda, Y.; Mereiter, K.; Ja¨ger, C.; Brecker, L.; Kosma, P.; Rosenau, T.; French, A. J. Am. Chem. Soc. 2008, 130, 16678–16690. (7) Ro¨hrling, J.; Potthast, A.; Rosenau, T.; Lange, T.; Ebner, G.; Sixta, H.; Kosma, P. Biomacromolecules 2002, 3, 959–968. (8) Bohrn, R.; Potthast; Schiehser, S.; Rosenau, T.; Sixta, H.; Kosma, P. Biomacromolecules 2006, 7, 1743–1750. (9) Nakatsubo, F.; Kamitakahara, H.; Hori, M. J. Am. Chem. Soc. 1996, 118, 1677–1681. (10) Kamitakahara, H.; Hori, M.; Nakatsubo, F. Macromolecules 1996, 29, 6126–6131. (11) Hori, M.; Kamitakahara, H.; Nakatsubo, F. Macromolecules 1997, 30, 2891–2896. (12) Agrofoglio, L. A.; Jacquinet, J. C.; Lancelot, G. Tetrahedron Lett. 1997, 38, 1411–1412. (13) Karakawa, M.; Nakatsubo, F. Carbohydr. Res. 2002, 337, 951–954.

BM9006612