Towards a better understanding of amylose isolation and production, semicrystalline aggregation and potential pharmaceutical use

Amylose (AM) is an almost linear glucose polymer. Together with amylopectin (AP), it makes up the bulk of granular starch. The AM content in regular starches ranges from 20 to 35%. It greatly influences the functionality of starch in aqueous systems. In fully gelatinized starch dispersions, AM crystallization plays a major role in gelation. In many food preparation protocols, AM network formation starts immediately after starch gelatinization. Connected cylindrical objects and ultimately fractal structures are formed by aggregation of AM and the outer branches of AP. These structural transitions are exploited to provide texture to different food products. AM crystallization has also been used to produce type III enzyme resistant starch (RS), a type of starch that resists digestion, is fermentable in the colon and has several potential health benefits.

AM crystallization depends on factors such as its average degree of polymerization (DP), its concentration in the suspension, the temperature and time. Its concentration determines the type of crystal arrangement which it adopts. While in diluted [typically <10.0% (w/v)] aqueous systems AM arranges into lamellar platelets, spherulitic crystal arrangements are observed in concentrated [typically >10.0% (w/v)] systems. These crystal architectures resist digestion by pancreatic α-amylase. However, they are fermentable by the microbiota in the human colon.

Despite the relevance of AM as a standalone polymer, the underlying mechanisms of its aggregation have mainly been investigated in starch systems, which by their own nature also contain AP. Against this background, this doctoral dissertation aimed to provide a better understanding of the production or isolation of AM and its physico-chemical properties in order to provide a basis for exploiting AM crystallinity in the development of potential applications in and beyond the food industry.

A method for producing pure AM on laboratory scale was needed for studying the mechanisms of AM aggregation. Thus, the first part of this dissertation focused on producing or isolating AM. Three main in vitro approaches were considered for obtaining AM: enzymatic synthesis, AM leaching, and AM complexation following starch dispersion. However, the production or isolation of AM is not a simple task. The properties (i.e., purity, average DP and polydispersity) of isolated AM are influenced by the experimental conditions in each methodology. Aqueous leaching allows isolating AM on large scale and involves heating a starch suspension above the starch gelatinization temperature. Different factors influence leaching of AM, including the leaching temperature (LT) and starch concentration. A response surface analysis with a face centered central composite design was implemented to study the effect of maize (Zea mays L.) starch concentration [3.0–7.0% (w/v)] and LT (70–90 ˚C) on aqueous leaching of AM as a way to optimize the conditions for obtaining the highest yield of long chain AM [number average DP (DPn) ranging from 860 to 930] and highest purity. Second order empirical models were fitted via the least squares approach. Negligible terms were removed using backwards model reduction. Negligible lack of fit terms were obtained for the responses total leached carbohydrate and DPn. The optimization was complemented with a desirability test using the purity of the extracts. As optimization targets, maximum leachate yields, DPn ≈ 900, and purity > 95% were set. Contour plots and prediction profilers were obtained and can be used by others for tailor made production of leachates. LT had the most significant effect as yields and DPn increased with temperature at the expense of purity. Purity was highly compromised when treatments were at temperatures exceeding 85 ˚C. This was reflected in the high DPn values (> 1,500) which suggested the presence of AP material. When using 3.0% (w/v) maize starch suspension at an LT of 81 °C, the largest yield (15.0%, starch basis) of high DPn AM chains (DPn ≈ 900) and less than 3.3% of non-AM material were obtained.

The effect of starch crystallinity on the aqueous leaching of AM was also studied. Starch crystal stability was altered via annealing. Leaching was studied in a 60-90 °C temperature range. The leachate yield, average DP and purity were related to the extent of melting of the starch crystals at the LT as determined via differential scanning calorimetry (DSC). Annealing increased the AP crystal stability and hence the remaining crystallinity at a given LT. Negligible AM leaching occurred at temperatures below those of the annealing dependent onset of melting. Leaching thus benefited from partial melting. Similar AM leachates were obtained when the extent of starch melting was below 80%. Loss of more than 95% of the melting enthalpy resulted in higher leachate average DP at the expense of purity. As the crystallinity of annealed starches at a given LT was higher than that of the native starches, the purity of leachates obtained from such starches was higher. Although no residual AP crystals remained at 90 °C, annealed starches subjected to leaching at such temperature still yielded AM extracts in higher yields and of higher purity than did native starch. More effective leaching in this case may be due to annealing-induced strengthened AP-AP interactions and AM disentanglement from AP.

The second part of this work focused on the semicrystalline aggregation of AM. While previous studies had elucidated the role of AM average DP on its aggregation in diluted aqueous systems, no reports have elaborated on the crystallization of AM in concentrated systems. Here, AM samples with different weight average DP (DPw) were produced and subjected to a heating-cooling-heating process. Since AM crystals only melt at temperatures exceeding 100 °C, high-pressure devices were used to analyze the hydrated samples. High (DPw = 830), mid (DPw = 340), and low (DPw = 60) DP AM aqueous dispersions [25.0% (w/v)] were first heated to 180 ˚C to produce fully dissolved aqueous solutions of AM. During subsequent cooling to ambient temperature and re-heating to 180 ˚C, their thermal and structural transitions were studied by DSC and X-ray diffraction at small (SAXS) and wide (WAXD) angles. During cooling, spherulitic crystal aggregates were formed the sizes of which decreased in the order mid DP > low DP > high DP AM. The crystallization events also depended on AM DP. Mid DP AM crystals were formed at high temperatures and its exothermic transition peaked at 74 ˚C. Those from low and high DP AM were formed below 60 °C with peaks at 37 and 42 °C respectively. A second (small) fraction was visible for low DP AM and it appeared close to the observed high temperature transition for mid DP. During the subsequent heating, mid DP AM crystals were the most stable and their melting signal peaked at 156 °C. Low DP AM crystals melted in two broad temperature ranges with peaks at 104 °C (large fraction) and 150 °C (small fraction). Those from high DP AM melted in a similar range as the low temperature signal for low DP AM.

Time-resolved SAXS and WAXD measurements were implemented for the first time to reveal the nanostructural transitions of AM during the formation and disappearing of semicrystalline spherulites from low and mid DP AM. WAXD measurements revealed B-type AM crystals besides amorphous material, regardless of the temperature. Changes in the crystallinity index occurred in the temperature ranges where DSC revealed exo or endothermic transitions. Inter-crystallite interference was found in SAXS for low DP AM while this was not the case for mid DP AM. Mid DP AM spherulites were classified as open (no interference) while those of low DP AM as compact (with interference). Open spherulites had a lower internal crystallinity (below 20 %) than the compact ones (up to 80 %) but were able to fill the space completely at the end of cooling. Open spherulites from mid DP AM started from the crystallization of AM within a homogeneous liquid phase. This forms AM depletion zones around the crystals at a high crystallization rate. At lower temperatures spherulites grow and new ones are created at a lower rate. Here the liquid phase can follow the pace and can homogenize concomitantly. In the case of low DP AM spherulites, open spherulites similar to those from mid DP AM are first formed at high temperatures. During further cooling, the system separates into AM-rich and –poor liquid phases. Spherulites migrate to the AM-rich zone and further AM crystallization occurs rapidly turning the aggregates into compact spherulites. For high DP AM it was proposed that chain entanglement might inhibit the formation of large and ordered aggregates at high temperatures. Instead, liquid-liquid phase separation is favored and the crystallization of AM into small, disordered spherulites takes place in the AM-rich zone.

In a third and final part, the crystallinity of AM was exploited in the development of a potential pharmaceutical application. Type III RS from debranched cassava (Manihot esculenta Crantz) starch was produced by favoring AM crystallization in a hydrothermal treatment. A thermostable crystal fraction was formed. Indeed, type III RS crystal melting was only observed at temperatures exceeding 120 °C. RS levels measured via in vitro digestion increased from 36.6 (in the starting material) to 95.1% (in the final type III RS product). RS levels were positively correlated to the degree of crystallinity. Type III RS was highly fermentable in vitro by human fecal microflora and resulted in a significant production of short chains fatty acids. Acetate levels were much more increased than those of propionate and butyrate when compared to their levels noted for a fecal blank. A granulate of type III RS [60 - 70 % (w/w)] and ethyl cellulose [40 - 30 % (w/w)] was used to coat tablets of 5-aminosalicylic acid (5-ASA) using a compression-coating approach. Coating protected 5-ASA during its transit through a model for the gastrointestinal tract. No release was observed in gastric medium and negligible release was observed after 2h in intestinal medium containing pancreatic amylase. This material may be of more interest than polymers such as cellulose due to a (potential) therapeutic effect exerted by increased short chain fatty acid levels as a result of fermentation. This study thus also provided the basis for applications of AM crystals in the pharmaceutical industry.