Light-controlled enzymatic synthesis of γ-CD using a recyclable azobenzene template

Cyclodextrins (CDs) are important molecular hosts for hydrophobic guests in water and extensively employed in the pharmaceutical, food and cosmetic industries to encapsulate drugs, flavours and aromas. Compared with α- and β-CD, the wide-scale use of γ-CD is currently limited due to costly production processes. We show how the yield of γ-CD in the enzymatic synthesis of CDs can be increased 5-fold by adding a tetra-ortho-isopropoxy-substituted azobenzene template irradiated at 625 nm (to obtain the cis-(Z)-isomer) to direct the synthesis. Following the enzymatic reaction, the template can then be readily recovered from the product mixture for use in subsequent reaction cycles. Heating induces thermal cis-(Z) to trans-(E) relaxation and consequent dissociation from γ-CD whereupon the template can then be precipitated by acidification. For this study we designed and synthesised a set of three water-soluble azobenzene templates with different ortho-substituents and characterised their photoswitching behaviour using UV/vis and NMR spectroscopy. The templates were tested in cyclodextrin glucanotransferase-mediated dynamic combinatorial libraries (DCLs) of cyclodextrins while irradiating at different wavelengths to control the cis/trans ratios. To rationalise the behaviour of the DCLs, NMR titrations were carried out to investigate the binding interactions between α-, β- and γ-CD and the cis and trans isomers of each template.

A stock solution of CGTase derived from Bacillus macerans was kindly gifted to our group by Amano Enzyme, Inc., Nagoya, Japan, and stored at 5 °C. According to specifications from the supplier, the stock solution contains approximately 20% glycerol, which was removed by performing a 160-fold solvent exchange with water using a Pall MicroSep Advance Centrifugal Device (0.5-5 ml) with a 10 kDa Omega Membrane according to the procedure described by the manufacturer. The final volume after solvent exchange was kept constant and the enzyme activity was tested. [S1] The glycerolfree stock solution of CGTase was used in all experiments except the preparative scale enzymatic synthesis of g-CD, where the commercial enzyme solution was used without purification.
All reactions were monitored either by reversed-phased ultra-performance liquid chromatography mass spectrometry (RP-UPLC-MS) or by thin-layer chromatography (TLC) using Merck aluminium sheets covered with silica (C60). Visualisation of the TLC plates was conducted under UV light or by staining with a developing agent. All new compounds were characterised using NMR, RP-UPLC-MS (ESI), HRMS (ESI), and melting point where appropriate. NMR samples were measured in capped standard 5 mm borosilicate glass NMR tubes. Assignments of 1 H NMR spectra were achieved by the use of standard 2D NMR techniques: 1 H-1 H COSY, 1 H-13 C HSQC and 1 H-13 C HMBC.

S1.2 Instruments
HPLC chromatographic analysis was performed on a Thermo Scientific Dionex UltiMate 3000 HPLC (ultra-high pressure) system with a Waters Acquity UPLC BEH Amide 1.7 μm 2.1 × 150 mm column maintained at 30 °C. The system was equipped with an autosampler which was maintained at 20 °C. Injection volumes were 10 μL. For detection, the chromatographic system was connected to an Agilent Technologies 1260 Infinity ELSD, operating at evaporator and nebulizer temperatures of 90 and 70 °C, respectively, and an N2 gas flow of 1.0 L/min. NMR spectra were acquired on a Bruker Avance III 400 MHz NMR spectrometer equipped with Prodigy broadband observe (BBO) probe, except for 1 H NMR spectra acquired for the NMR titrations, which were acquired on a Bruker Avance III 400 MHz NMR spectrometer equipped with a BBO SmartProbe. Chemical shifts (δ) are quoted in ppm and coupling constants (J) are presented in Hz. NMR spectra were referenced against residual solvent peaks. Measurements were performed at 298 K. When quantification of relative concentrations was required, 1 H NMR spectra were acquired using a delay of 10 seconds between pulses.
Analytical RP-UPLC-MS (ESI) analysis was performed on an S2 Waters AQUITY RP-UPLC system equipped with a diode array detector using a Thermo Accucore C18 column 2.6 µm, 2.1 x 50 mm; column temp: 50 °C; flow: 0.6 mL/min. Solvent A: 0.1% formic acid in water. Solvent B: 0.1% formic acid in MeCN. Gradient for short run: 5% B to 100% B in 2.4 min., hold 0.1 min., total run time 2.6 min. Gradient for long run: 5% B to 100% B in 3 min., hold 0.1 min., total run time 5 min. The LC system was coupled to an SQD mass spectrometer operating in both positive and negative electrospray modes. The temperature for all recordings was approximately 20 °C.
UV-vis absorption spectra were measured with a double beam Specord 210 Plus spectrometer from analytik jena. All measurements were performed in a 3.5 mL Hellma Quartz cuvette (labelled QS 282) with a 10 mm path length at room temperature.
ESI-HRMS were performed on a SolariX ESI FTMS spectrometer with dithranol as matrix. External calibration of the spectrometer was conducted with sodium trifluoroacetate cluster ions.
Melting points were measured on a Stuart SMP 30 melting point apparatus and are uncorrected. A Heraeus Biofuge Pico centrifuge was used for the preparation of samples for HPLC analysis. An Eppendorf Centrifuge 5810R was used during the synthesis of the mAzo and ipAzo templates, and for the preparative scale synthesis of g-CD. For experiments involving light irradiation, all the LEDs were of the brand ThorLabs and driven by an adjustable power supply (0-1.2 A), and they were operated at their highest permitted power input (as defined by the supplier). Unless stated otherwise, the LEDs were operated at 1000 mA for the nominal wavelengths of 625 nm, 530 nm, and 470 nm, and for the nominal wavelength of 365 nm the LEDs were operated at 700 mA. S4

Synthesis of compound 6
The procedure for the preparation of compound 6 was adapted from protocols reported by Wu et al. [S3] Procedure: 6-Bromohexanoic acid (1.99 g, 10.2 mmol), 4-phenyldiazenylaniline (2.05 g, 10.4 mmol), 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDC) (1.94 g, 10.1 mmol), and 4-dimethylaminopyridine (DMAP) (0.247 g, 2.02 mmol) were dissolved in CH2Cl2 (470 mL) resulting in an orange solution. The mixture was stirred at room temperature for 4 days. The solvent was removed in vacuo and the solid crude was purified by flash column chromatography (diameter: 7.5 cm, isocratic eluent: CH2Cl2) to yield the product as an orange solid (2.28 g, 6.09 mmol, 60%  H8). Data in agreement with previously published data. [S3] Synthesis of compound 4 The experimental procedure was inspired by procedures reported by York et al. [S4] and Erichsen et al. [S1] Procedure: To a solution of diethyl 3-oxoglutarate (20.03 g, 0.10 mol) in absolute ethanol (350 mL) with activated molecular sieves (12.36 g, 3 Å) was added ammonium acetate (24 g, 0.38 mol). The S5 formed suspension was stirred for 22 hours and then NaBH3CN (7.16 g, 0.11 mol) and HCl in EtOH/EtOAc * (~ 1 M, 600 mL) were added, resulting in a suspension with pH ~ 3. The suspension was stirred for 1.5 hours, then filtered and the resulting filtrate was concentrated in vacuo. The resulting oil was dissolved in water (200 mL) and K2CO3 (20 g) was added. Then aqueous NaOH (2 M, 200 mL) was added to adjust the pH to > 13 and the aqueous phase was extracted with ethyl acetate (5 × 200 mL). The combined organic extracts were washed with brine (3 × 160 mL) and dried over Na2SO4, filtered and concentrated in vacuo to leave a crude yellow oil. The crude oil was purified by flash column chromatography (diameter: 8.5 cm, length: 15 cm, ~ 7.5 g of crude oil, isocratic eluent: MeOH : CH2Cl2, 5 : 95) to yield the product as a yellow oil (7.09 g, 0.034 mol, 35%  [S1,S4] * Note: The ~1 M solution of HCl in EtOH/EtOAc was made by the slow addition of acetyl chloride (60 mL) to a cooled (0 °C) solution of absolute ethanol (800 mL, from a newly opened bottle). The reaction was stirred at 0 °C for one hour and then at room temperature for two hours.
The reaction was left to stir for 16 hours at room temperature. The formed precipitate was collected, dissolved in ethyl acetate and concentrated in vacuo. The crude product was purified by flash column chromatography (diameter: 3 cm, gradient elution from 2%-6% methanol in ethyl acetate). The product was obtained as an orange solid (793 mg, 1.84 mmol, 55%

Synthesis of template ipAzo
Procedure: Compound 13 (140 mg, 0.131 mmol) was dissolved in THF (5 mL) and aqueous NaOH (1 M, 5 mL) was added. The two-phase solution was stirred at room temperature for 2 hours. The reaction mixture was concentrated in vacuo until approximately 2 mL remained, then transferred to a centrifuge tube by use of aqueous NaOH (1 M, ∼ 2 mL). The pH was adjusted to 1 by addition of aqueous HCl (6 M, ∼ 5 mL), and the formed precipitate was collected and washed with water (2 × 5 mL) upon centrifugation. After drying in vacuo the product was obtained as a red solid (

S3. UV-vis experiments
All template solutions (50 µM) were prepared in 100 mM sodium phosphate buffered water at pH 7.5 and stored in darkness. In all experiments a reference of air was applied and an absorbance spectrum of the phosphate buffer was subtracted from the sample absorption spectra. At least 1.5 mL of desired solution was added to the cuvette to ensure that the light beam passed though the solution. All UV-vis spectra were measured in the range 220−700 nm with a scan rate of 10 nm/s and a ∆λ of 1 nm. A slit width of 1 nm was applied and lamp change was set at 320 nm.
For photoswitching experiments a general method was applied in all experiments: First, an absorption spectrum was measured before any light treatment. Secondly, the samples were irradiated with light of specific wavelengths until no more change was observed in the absorption spectrum. The cuvette containing the samples were placed at a fixed distance from the light source in a customised box to be shielded from ambient light. The samples were moved directly from the box to the instrument to minimise interference from ambient light, and the measurements were performed instantly.

S4. NMR experiments
All samples for NMR experiments were prepared in 100 mM phosphate buffered D2O, pH 7.5. All spectra were acquired at 298 K. Samples were stored in darkness prior to analysis unless otherwise stated. In all irradiation experiments the NMR sample was placed in a customised box with a fixed length from the light source. When handling the NMR samples the tubes were wrapped in foil to avoid irradiation from ambient light. The cis/trans ratios were determined from integrals of peaks that were assigned exclusively to the cis-or trans-isomers of the templates.

S4.1 Determination of photostationary states
Photostationary states of the templates were determined at 10 mM for each template. The NMR samples were stored in darkness at 30 °C at least 15 hours prior to analyses. The photostationary states were obtained by irradiating the NMR samples with light of the specific wavelengths until no more change was observed in the obtained NMR spectra.

S4.2 Aggregation studies
To investigate the aggregation of the templates, a dilution series of the templates were analysed using 1 H NMR spectroscopy. The samples were investigated before light treatment, since it would be expected that the trans-isomers would have the strongest tendency to aggregate. 10 mM stock solutions of each template (Azo, mAzo, or ipAzo) in 100 mM phosphate buffered D2O, pH 7.5 were prepared and analysed by 1 H NMR spectroscopy. By diluting with 100 mM phosphate buffered D2O, pH 7.5, a dilution series down to 0.05 mM was acquired for each template.

S4.3 1 H NMR titrations
For each titration a 'host' solution and a 'guest' solution were prepared. A 0.1 mM host solution was prepared by dissolving the desired template Azo, mAzo, or ipAzo in 100 mM phosphate buffered D2O, pH 7.5. The guest solution was prepared by dissolving the desired amount of guest (α-, β-, or γ-CD) in the 0.1 mM host solution. 500 µL host solution was transferred to an NMR tube. For the titrations of the trans-isomers, the NMR sample and guest solution were stored in darkness at 30 °C for at least 15 hours prior to a titration experiment. For the titrations to investigate binding by the cisisomer, the NMR sample was irradiated with light at appropriate wavelengths prior to the titration experiment and for five minutes after each addition of guest solution prior to acquisition. 1 H NMR spectra were acquired at increasing concentrations of cyclodextrin guest.
For titration experiments with one host the change in chemical shift of selected proton signals were plotted against the corresponding guest concentrations and the binding isotherm was fitted nonlinearly to a 1:1 binding model, as described in the equation below. [S7] For relevant titration experiments with both the trans-and cis-hosts in solution the chemical shift changes of a cis-proton was plotted against the chemical shift changes of a trans-proton. When a binding constant, Ka, and the maximum chemical shift change, ∆δa,max, for a trans-host-guest complex are known (from the titration performed in the dark), the binding constants for the corresponding cishost-guest complex can be determined by fitting the data to the equation below. [S8] When applying this method uncertainties determined for the Kb and ∆δb,max values for the cis-isomer were compounded with the fitting errors on the values determined for the trans-isomer, since the latter values were used to calculate the former.
For all fitting procedures the data analysis software OriginPro (Version 2021. OriginLab Corporation, Northampton, MA, USA) was applied. Figure S11: 1 H NMR titration of trans-Azo with α-CD, β-CD and γ-CD. a) trans-Azo with assigned proton numbers. Partial 1 H NMR spectra and fitted 1:1 binding isotherms based on ∆δobs of the indicated protons upon addition of b) α-CD, c) β-CD, and d) γ-CD. Conditions: the titrations were performed with 0.1 mM template solutions in 100 mM phosphate buffer in D2O, pH 7.5, at 298 K. Figure S12: 1 H NMR titration of trans-mAzo with α-CD, β-CD and γ-CD. a) trans-mAzo template with assigned proton numbers. Partial 1 H NMR spectra of trans-mAzo and fitted 1:1 binding isotherms based on observed chemical shift change for the marked protons upon addition of b) α-CD, c) β-CD, and d) γ-CD. Conditions as in Figure S11. Figure S13: 1 H NMR titration of trans-ipAzo with α-CD, β-CD and γ-CD. a) trans-ipAzo template with assigned proton numbers. Partial 1 H NMR spectra of trans-ipAzo and where relevant fitted 1:1 binding isotherms based on observed chemical shift change for proton H2 upon addition of b) α-CD, c) β-CD, and d) γ-CD. The binding affinity between tras-ipAzo and γ-CD was too weak to quantify. Conditions as in Figure S11. Figure S14: 1 H NMR titrations to determine binding constants for cis-Azo with α-CD, β-CD and γ-CD. a) cis-Azo template with assigned proton numbers. b) Partial 1 H NMR spectra of a mixture of cis and trans Azo and fitted plot of Δδcis versus Δδtrans based on observed chemical shift changes of the marked signals upon addition of α-CD. Partial 1 H NMR spectra of cis-Azo and fitted 1:1 binding isotherms based on Δδobs for the right side of the marked signal corresponding to the cis-protons H3 and H4 upon addition of c) β-CD and d) γ-CD. Conditions as in Figure S11. Figure S15: 1 H NMR titrations to determine binding constants for cis-mAzo with α-CD, β-CD and γ-CD. a) cis-mAzo template with assigned proton numbers. Partial 1 H NMR spectra of a mixture of cis-and trans-mAzo and fitted plots of Δδcis versus Δδtrans based on observed chemical shift changes for the marked signals upon addition of b) α-CD and c) β-CD. d) Partial 1 H NMR spectra of a mixture of cis-and trans-mAzo upon addition of γ-CD. Binding was too weak to quantify. Conditions as in Figure S11. Figure S16: 1 H NMR titrations to determine binding constants for cis-ipAzo with α-CD, β-CD and γ-CD. a) cis-ipAzo template with assigned proton numbers. Partial 1 H NMR spectra of a mixture of cis-and trans-ipAzo and fitted plot of Δδcis versus Δδtrans based on observed chemical shift changes of the marked signals upon addition of b) α-CD and c) β-CD. d) Partial 1 H NMR spectra of a mixture of cis-and trans-ipAzo and fitted 1:1 binding isotherms based on the observed chemical shift change for H19 upon addition of γ-CD. The binding interaction between γ-CD with trans-ipAzo was judged to be so weak that competitive binding could be disregarded when determining the association constant for the interaction between γ-CD and cis-ipAzo. Conditions as in Figure S11, except the titration with γ-CD which was performed with 0.5 mM template solution.

S5. CGTase-mediated dynamic combinatorial libraries of cyclodextrins
All reactions were set up in clear 2 mL glass vials. For libraries under irradiation with light, the vials were placed in a fitted chamber in the exact same distance from the light source to ensure consistency. The libraries were irradiated with light at all times during the studies, except when taking out samples for analysis. The libraries were irradiated before starting the library for a period of time that ensured full conversion to photostationary state (30 mins at 365 nm for Azo, overnight at 625 nm for mAzo and 90 mins at 625 nm for ipAzo). The libraries performed in darkness were stored in darkness a minimum of 15 hours before starting the library and were kept in darkness at all time during the studies, except when taking out samples for analysis. The spin-filtered CGTase enzyme stock solution was used with addition of 65 µL enzyme mixture per mL library reaction mixture in all cases to ensure the desired activity. All libraries were performed at room temperature.
The reference libraries (untemplated) were prepared from a solution of 10 mg/mL α-CD in 100 mM sodium phosphate buffer at pH 7.5. Libraries were started by addition of CGTase stock solution. The templated libraries were prepared from a solution of 10 mg/mL α-CD with the desired template(s) at 10 mM in 100 mM sodium phosphate buffer at pH 7.5. Libraries were started by addition of CGTase stock solution.

S5.1 Sampling and analysis method
The libraries were analysed over time by taking out 3 µL samples from the library which were added to a 120 µL quenching mixture consisting of 10 mM ammonium chloride in acetonitile/water (3:1) with 1% TFA. The 123 µL mixture was centrifuged at 10000 rpm for 4 minutes in a Heraeus Biofuge Pico. After centrifugation, the top 110 µL fraction were transferred to an HPLC insert in a HPLC vial capped with screw caps equipped with PTFE lined rubber septa for injection in the HPLC-ELSD system. An injection volume of 10 µL was used in all cases.
The eluent system was water (A) and acetonitrile (B), both with 0.1% formic acid. For separation, a linear gradient program running from 25 to 45% eluent A over 8.00 minutes with a flow rate of 0.60 mL/min was used, after which the column was washed for three minutes with 100% eluent A. Before each injection, at least five column volumes of the starting eluent were applied to condition the column. The resulting chromatograms were analysed and the peak area was converted to a concentration (by weight) through calculations based on calibrations performed individually for the short linear α-1,4-glucans glucose (G1), maltose (G2), and maltotriose (G3) as well as α-CD, β-CD, and γ-CD by analysing stock solutions using the above-mentioned chromatographic method. The concentrations of the longer linear α-1,4-glucans (G4−G8) were calculated based on the G3 calibration. The calibrations were based on µg injected in the 0.018−3.66 µg range, and the resulting response curves were fitted non-linearly (using OriginPro 2019b from OriginLab Corp.) to a simple power equation.
where M is the injected mass of the compound, A is the area under the peak in the chromatogram, while the fitted parameters k and p are referred to as the coefficient and the exponent, respectively. By knowing the exact dilution factor of the sampled aliquot and the injection volume used on the S32 chromatographic equipment, the calculated injected mass of a compound can easily be converted to a concentration by weight in the actual reaction mixture.
The values of k and p obtained from these fits, along with the errors on the fits and the resulting adjusted R 2 -values, are listed in Table S4 below.  Figure S17: Top: distribution of α-, β-, and γ-CD and total CD yield over 5 hours for the untemplated DCL performed in darkness (left) and under irradiation with at 365 nm (right). Bottom: distribution of α-, β-, and γ-CD and total CD yield over 5 hours for the DCL templated with Azo performed in darkness (left) and under irradiation with UV light at 365 nm (right). Conditions: the DCLs were performed using 10 mg/mL α-CD with and without 10 mM template in 100 mM phosphate buffer at pH 7.5 and at room temperature. For reactions performed in darkness the reaction mixture was stored overnight at 30 °C in the dark prior to starting the reaction by addition of enzyme. For the irradiated reactions, the reaction mixture was pre-irradiated to obtain the photostationary state of the template before starting the reaction by addition of enzyme. Figure S18: Distribution of α-, β-, and γ-CD and total CD yield over 6 hours for the DCLs templated with mAzo performed in darkness (left) and under irradiation with red light at 625 nm. Conditions as in Figure S17. Figure S19: Distribution of α-, β-, and γ-CD and total CD yield over 5 hours for the DCLs templated with ipAzo performed in darkness (left) and under irradiation with red light at 625 nm. Conditions as in Figure S17. Figure S20: Left: distribution of α-, β-, and γ-CD and total CD yield for a DCL templated with Azo started in darkness (97% trans-Azo), then irradiated at 365 nm (to promote trans to cis isomerisation), and finally irradiated with blue light at 470 nm (to promote cis to trans isomerisation). Right: distribution of α-, β-, and γ-CD and total CD yield for a DCL templated with ipAzo started in darkness (91% trans-ipAzo), then irradiated at 625 nm (to promote trans to cis isomerisation), and finally irradiated with at 470 nm (to promote cis to trans isomerisation). Conditions as in Figure  S17.

S5.3 Template recovery and preparative scale templated enzymatic synthesis of g-CD
Template recovery from enzyme-mediated synthesis of g-CD monitored by in situ NMR spectroscopy A solution of a-CD (10 mg/mL) and ipAzo (10 mM) in sodium phosphate buffered (100 mM, pH 7.5) D2O was prepared in a colorless 2 mL glass vial with a screw-cap and irradiated with light at 625 nm for 15 hours. CGTase enzyme stock solution (65 µL/mL) was added and the reaction was kept under red light irradiation. To monitor the reaction by ¹H NMR spectroscopy, the reaction mixture was transferred to a standard 5 mm borosilicate glass NMR tube. Water suppression was applied using presaturation to obtain useful spectra, since addition of the enzyme stock solution introduces a significant amount of non-deuterated water. After acquisition, the reaction mixture was transferred back to the glass vial and irradiated with red light. After 5 hours, the enzyme was denatured by installing the NMR tube in a specially fitted aluminium block thermostatically heated to 95 °C for 15 minutes. After cooling to room temperature, an ¹H NMR spectrum was acquired. The mixture was transferred to a centrifuge tube and centrifuged at 10000 rpm for 8 minutes in a Heraeus Biofuge Pico to remove enzyme. The supernatant was transferred to another centrifuge tube with 100 µL buffer and 1% (v/v) TFA was added leading to immediate precipitation of a red solid. After centrifugation as just described, the filtrate was transferred to an NMR tube and analysed. The precipitate was suspended in MilliQ water (0.5 mL), centrifuged, the water was decanted off and the precipitate was dried in vacuo. The precipitate was redissolved in sodium phosphate buffered (100 mM, pH 7.5) D2O for NMR analysis. Signals corresponding to a-CD, b-CD, and g-CD was assigned using reference spectra of a-CD, b-CD, and g-CD in sodium phosphate buffered (100 mM, pH 7.5) D2O. Additionally, the supernatant and precipitate were analysed by HPLC (sample preparation as described in section S5.1). The chromatograms are shown below in Figure S21. Preparative scale templated enzymatic synthesis of g-CD A stock solution of a-CD (11 mg/mL) in sodium phosphate buffered water (100 mM, pH 7.5) was prepared. ipAzo (22.0 mg, 0.023 mmol) was dissolved in the a-CD stock solution (2099 µL) to obtain a concentration of 11 mM. This solution was diluted with 94.5 µL buffer and divided equally between two colourless 2 mL glass screw-cap vials. The two vials were irradiated with light at 625 nm for a least 2 hours. With continuous irradiation the enzymatic reaction was started by addition of unfiltered commercial CGTase stock solution (115.5 µL) which was divided equally into the glass vials. In the reaction mixtures a final concentrations were 10 mM ipAzo, 10 mg/mL a-CD concentration and 50 µL CGTase stock solution per mL reaction mixture was obtained. After 5 hours, the enzyme was denatured by installing the glass vials in a specially fitted aluminium block thermostatically heated to 95 °C for 15 minutes. After cooling to room temperature, the reaction mixtures were transferred to a 15 mL centrifuge tube and centrifuged at 10000 rpm for 8 minutes in an Eppendorf Centrifuge 5810R to remove enzyme. The supernatant was transferred to another centrifuge tube with 200 µL buffer and 1% (v/v) TFA was added leading to immediate precipitation of the template. The mixture was centrifuged and the supernatant containing glucans was collected by decanting. The sediment was washed with MilliQ water (1 mL), centrifuged again, the supernatant removed, and the sediment dried in vacuo. The thus isolated template was used again in another cycle of the same reaction, ensuring the same final concentrations in the reaction mixture. After 5 reaction cycles ipAzo was recovered (20.4 mg, 0.021 mmol, 93%, high purity confirmed by 1 H NMR, Figure S23). The collected S36 supernatants containing the glucans from the 5 cycles were combined and concentrated to approximately 1 mL by blowing nitrogen over the surface. The solution was filtered through a syringe filter and injected on a HILIC type HPLC column (XBridge BEH Prep OBD Amide 5 µm from Waters, 19Î150 mm) using a Büchi Pure C-850 FlashPrep chromatography system equipped with an ELS detector. Gradient elution from 25% water in acetonitrile to 35% water in acetonitrile over 10 minutes and then from 35% acetonitrile in water to 45% acetonitrile in water over 35 minutes was used. Fractions eluting from 36 to 39 minutes contained only g-CD, and these were combined, concentrated in vacuo and lyophilised to yield g-CD (36.2 mg, 32% yield of glucan starting material). The reaction cycles and the isolation of the g-CD were analysed by HPLC and a shown in Figure S22.