Supplemental information Peters et al.
Preparation and measurement of pullulan fragments
Pullulan is a water soluble, viscous polysaccharide consisting of basic unit of three α-1,4 linked glucose molecules (one maltotriose-unit) that is repeatedly polymerized by an α-1,6 linkage on the terminal glucose, resulting in a stair-step structure.
Pullulan (PI20, Hayashibara Co, Ltd, Okayama, Japan; molecular mass of 200 kDa according to the supplier) was split into smaller fragments using the food-grade enzyme pullulanase (EC 3.2.1.41). This endo-glucanase splits α-1,6 linkages between the maltotriose-units (Nakamura, 1984). By increasing the duration of exposure to the enzyme, more linkages are broken, and smaller fragments are produced with shorter chain lengths. Therefore, 6.5 kg of pullulan was dissolved in 58.5 L water, 106.6 gram sodium acetate was added and the pH was adjusted to 5.2 with 6N HCl. After this 87 gram Promozyme 400L (Novozymes A/S, Bagsvaerd, Denmark) was added and hydrolysis was allowed to proceed at approximately 25°C. Three batches were prepared, and hydrolysis was stopped after 0, 2 and 6 hours by heating at 142°C for 14 seconds at a flux of 150 kg per hour.
After production, the molecular mass of the pullulan fragments was determined by gel filtration. Gel filtration was performed in a HPLC set-up using a TSKgel G4000SW column (Toso Haas, Japan). The column was equilibrated with 10mM sodium acetate buffer pH 5.8 and calibrated using Shodex pullulan standards (Showa Denko, Munich, Germany) P10 (MW=11.8kDa), P20 (MW=22.8kDa), P50 (MW=47.3kDa), P100 (MW=112kDa), P200 (MW=212kDa) and P400 (MW=404kDa). The sample concentration was 1 mg/ml, the sample volume was 20 µL and the flow rate was 0.9 ml/min. Detection was done with a refractive index meter.
In vitro digestibility of pullulan fragments
To test the rate and extent of hydrolysis of the pullulan fragments in vitro, the pullulan fragments exposed to the brush border enzyme amyloglucosidase (EC 3.2.1.3), as the linkages in Pullulan in the human digestive tract can only be split by brush border enzymes and not by amylase. To 14 ml of a solution containing 2% (w/w) pullulan (fragments) 1 ml Aspergillus niger amyloglucosidase solution (11.1 mg/ml 0.1M sodium acetate buffer pH5.2) was added and the mixture was incubated under constant shaking at 37°C. Samples (0.5ml) were drawn from the mixture to monitor hydrolysis. The enzyme reaction was stopped by immediately transferring the samples to glass tubes that were placed in a 100°C water bath and heated for 3 minutes. Samples were diluted 1:10 with water. Glucose release was measured using Enzytec Glucose-D kit (Scil Diagnostics, Germany).
Characterization of pullulan fragments and drinks
By treating the glucose polymer pullulan (200kDa) with the enzyme pullulanase, fragments with an average molecular mass of 22.6 kDa (called medium-chain) were produced. As hypothesized, a clear relationship between the in vitro rate of glucose release and the fragment size is shown in Figure 1.
Figure 1. In vitro hydrolysis of maltodextrin (MDX), medium-chain (MCP) and long-chain (LCP) pullulan (duplicate measurements).
The medium-chain pullulan fragments showed slow and almost complete hydrolysis within a time frame of 2 hours, while the long-chain pullulan was hardly hydrolyzed and the maltodextrin rapidly hydrolyzed. After 2 hours, 77 and 23% of the medium- and long-chain pullulan, respectively, were digested by brush border enzymes, providing in vitro evidence of the slow-carb behavior of the medium-chain pullulan, and the resistant (fibre-like) behavior of the long-chain version.
Viscosity of the test drinks
Viscosity of the test drinks was measured with a TA-AR1000 viscometer (TA Instruments, Leatherhead, UK) over shear rates from 0.03 to 1 000 s-1. A cone and plate geometry was used, where the cone had a diameter of 40 mm, a truncation of 2°, and the gap set to 70 μm. The viscosity was measured at 7°C (the temperature at which the drinks were served to the subjects).
At a shear rate of 50 s-1 (approximating shear in mouth) viscosities were 0.15, 0.14 and 0.27 Pa∙s for the control, the medium- and long-chain pullulan, respectively.
Viscosity of the long-chain pullulan was slightly greater than the other materials, which were largely similar. However, these differences in viscosity are far smaller than what would be expected to have any effects on appetite by itself. Data from Marciani et al. (2000) suggest that an 80 times higher viscosity is needed for a significant effect of viscosity on appetite.
Marciani L, Gowland PA, Spiller RC, Manoj P, Moore RJ, Young P, et al. (2000). Gastric response to increased meal viscosity assessed by echo- planar magnetic resonance imaging in humans. J Nutr130, 122-127.
Nakamura S (1984). Pullulan. JSynthetic Org Chem Japan42, 584-588.