Document [original]

RESEARCH ARTICLE

www.starch-journal.com

High Amylose Corn Starch Gels – A Molecular Investigation

of the Network Constituting Polymers

Marco Ulbrich,* Fanni Scholz, and Eckhard Flöter

The relationship between microstructural features and mechanical strength of

aqueous starch gels is investigated. The gels are prepared by systematic

variation of the factors starch type (high amylose [AM] corn starches [HACS]

having about 50 and 70% w/w AM; HACS-50/-70), starch concentration (6

and 9% w/w) and storage time (1 and 14 days). The gel matrices are

separated by means of a centrifugation method, and two phases are obtained

account for the dissolved starch (liquid phase) and the network constituting

starch (swollen phase). Subsequently, the latter starch fraction is partially

digested using two amylases (AMY) and in combination with pullulanase

(PUL) [𝜶-AMY, 𝜶-AMY-PUL, 𝜷-AMY, 𝜷-AMY-PUL], too. The starch samples

(dissolved and network constituting) and the degradation products

(polysaccharides after partial enzymatic hydrolysis) are characterized

molecularly by means of size exclusion chromatography (SEC)-techniques,

and amounts and molecular properties of speciﬁc fractions and molecule

segments, respectively, are determined. A clear correlation between the

speciﬁc involvement and contribution of the starch polymers (e.g., absolute

amount, state, function) in the gel network and the mechanical gel strength is

found. Particularly, the starch’s AM content and the polymer concentration of

the gel are evident as controlling factors in developing the gel’s ﬁrmness.

1. Introduction

Starch is a versatile polysaccharide with many food and non-

food applications. It normally consists of two diﬀerent structure

fractions, AM and amylopectin (AP). The polymers are based

on 𝛼-D-glycosidic linked anhydroglucose units (AGU), at which

AM is the largely non-branched (𝛼-1,4-glycosidic linked AGU;

Mwabout 105–106g∙mol−1) and AP the highly branched frac-

tion (𝛼-1,4-glycosidic linked AGU building the linear molecule

sections, and additional 𝛼-1,6-glycosidic linkages forming the

M. Ulbrich, F. Scholz, E. Flöter

Department of Food Technology and Food Chemistry

Chair of Food Process Engineering

Technische Universität Berlin

Oﬃce ACK3, Ackerstraße 76, Berlin 13355, Germany

E-mail: [email protected]

The ORCID identiﬁcation number(s) for the author(s) of this article

can be found under https://doi.org/10.1002/star.202200032

This is an open access article under the terms of the Creative Commons

Attribution License, which permits use, distribution and reproduction in

any medium, provided the original work is properly cited.

DOI: 10.1002/star.202200032

molecule branches; Mwabout 106–109

g∙mol−1). The existence of an interme-

diate structure fraction (IM), which can-

not be clearly assigned to both AM and

AP, is commonly accepted for diﬀerent

starches (normal CS,[1] HACS,[2,3] diﬀerent

CS[4,5] pea.[6,7] The AM/AP ratio of common

starches is about 20/80 to 30/70,[8,9] and pea

starches have AM contents higher than 30%

w/w.[10] However, so-called starch geno-

types exhibit AM contents, which are highly

diﬀerent compared to normal starches (e.g.,

waxy or high AM varieties[11,12]). This inher-

ent special feature, in particular in terms

of the molecular composition of the starch,

remarkably impacts the speciﬁcity of dif-

ferent important techno-/functional proper-

ties like hot paste viscosity and achievable

mechanical gel strength.

The starch polymers are synthetized in

the form of semicrystalline granules with al-

ternating amorphous and crystalline struc-

tures. The fact of the supramolecular struc-

ture (granular state) accounts for the re-

quirement of a disintegration process (e.g.,

pressure cooking) to convert the polymers

in the presence of water to an aqueous starch paste, which is a

prerequisite for most applications. Particularly, the utilization as

a gelling agent requires a very high degree of disintegration of the

starch polymers (ﬁrst step), i.e., the complete loss of the starch’s

granular integrity, since the gelation is based mostly on retrogra-

dation of dissolved starch polysaccharides, a partial rearrange-

ment of polymer chain sections (preferentially of AM) upon cool-

ing (second step).[13] The micro-/macro structures and associated

gel characteristics (e.g., achievable mechanical strength) depend

on several and partly interacting factors like starch source,[14,15]

AM/AP-ratio[16,17] and molecular composition as well as polymer

concentration,[14,18] solution state,[15] or the presence of lipids.[19]

Moreover, with retrogradation, the mechanical properties are es-

sentially controlled by the storage conditions.[16] The develop-

ment of the gel matrix, including rigidity, molecular re-order,

and increasing crystallinity is necessarily ascribed to AM (irre-

versible, short-term changes) and also the relative amount of

AM available,[20] respectively, and long-term changes of the gel

strength (enhancement) are attributed to reordering and on-

going crystallization processes mainly within the AP fraction

(reversible).[21,22]

Altogether, crystallization of the AM fraction plays the major

role in the sol-to-gel transition and formation of a mechani-

cally stable gel structure,[21] which makes particularly high AM

2200032 (1 of 10)

www.advancedsciencenews.com www.starch-journal.com

genotypes interesting for research involving the study of detailed

gel buildup. AM gel structures show a dynamic nature (non-

equilibrium), i.e., the diﬀusion-controlled aggregation processes

govern the mechanical properties of the hydrated network.[21]

Besides the formation of an elastic gel-network, the gelation is

characterized by a development of opacity generally attributed

to further chain aggregation, resulting in crystalline areas de-

tectable by X-ray diﬀraction (XRD) experiments (slow process).

The network formation of AM can be regarded as initiated by

a phase-separation (demixing, fast process) in the homoge-

nous sol, which yields to polymer-rich regions interspersed

with polymer-deﬁcient regions. The enhancement of local AM

concentration in the polymer-rich regions enables interchain-

associations; the partial crystallization within the polymer-rich

regions occurs at a much slower rate than the demixing.[21]

However, nucleation and limited growth of rod-shaped micro-

crystals occurring promptly during network formation were

assumed elsewhere.[23] The speciﬁc cross-linking involves the

cooperative molecular interaction of many residues from each

participating AM chain and the formation of double-helices

and interhelix interactions, respectively. These conformationally

ordered segments show a B-type crystalline structure (XRD,

immobile “solid-like” structures). They are interconnected by

possibly more mobile amorphous single-chain segments with a

conformation similar to those present in solution and “elastically

active.”[24] In general, growth and coarsening of AM gels may

depend on molecular size of involved polymers, concentration of

the system, and the overall gelation conditions.[25,26] The macro-

molecular organization within the 3-D microstructure of AM gels

was described by Leloup et al.[27] (continuous model), including

interconnecting network strands, with the crystalline portion

(crystallites) embedded in an amorphous matrix. The molecule

chain segments constituting the crystallites are disposed oblique

to the microﬁber axis building the inﬁnite 3-D network. The net-

work strands, which probably consist of continuous associated

blocks and involved double helices, are linked to others by loops

of amorphous amylose segments extent into the pores.[27]

The gel strength is an important application characteristic of

starch when used as a gelation agent. The achieved mechani-

cal gel ﬁrmness is generally adjusted by many factors such as

starch type and the diﬀerent speciﬁc molecular properties (e.g.,

AM content), the starch concentration in the aqueous system

and the storage conditions. However, strictly speaking, the gel

strength is supposed to be essentially dependent directly on the

macromolecular organization, i.e., possibly the portion of crys-

talline structures. The absolute amounts (portion) and molec-

ular properties (chain length, molecule branches) as well as

the state (crystalline, amorphous) of molecule segments con-

stituting the polymer-rich region of the gel matrix seem to

govern the ﬁrmness in a complex way. Aiming to specify the

(detailed) macromolecular contribution to the gel microstruc-

ture and a correlation to the mechanical strength, an elabo-

rate method was used, including i.a. gel preparation (disintegra-

tion by pressure cooking; varied parameters starch [HACS-50/-

70], starch concentration [6 and 9% w/w] and storage time [1

and 14 d]), separation of the gel matrix (polymer-deﬁcient and

polymer-rich phase, centrifugation), speciﬁc enzymatic diges-

tion of the network constituting phase (𝛼-AMY, 𝛼-AMY-PUL, 𝛽-

AMY, 𝛽-AMY-PUL), and molecular characterization of the prod-

ucts (SEC-techniques). This is the second of a series of three

publications.

2. Experimental Section

2.1. Starch Genotypes

Commercial native HACS genotypes (HYLON V [HACS-50] and

HYLON VII [HACS-70], Ingredion Germany GmbH, Hamburg,

Germany) were used for the gel preparation. The speciﬁcations

indicated are supplier information (HACS-50: 55% w/w AM

[nominal], 1.0% w/w total fat, 0.5% w/w protein, ≤0.1% w/w ash;

HACS-70: 73.8% w/w AM [colorimetric method], 0.58% w/w pro-

tein). The dry matter contents were 88.46% w/w (HACS-50) and

88.36% w/w (HACS-70).[28] Deionized water was used for all ex-

periments.

2.2. Enzymes

Diﬀerent enzymes were used for the examination, an 𝛼-AMY

(powder, Optizym A 16 126 [21.000 U∙g−1], SternEnzym GmbH

& Co. KG, Ahrensburg, Germany), a 𝛽-AMY (solution, Secura

[5000 BAMU∙g−1], Novozymes A/S, Bagsvaerd, Denmark), and

a PUL (solution, PromozymeD2 [200 U∙mL−1], Novozymes A/S,

Bagsvaerd, Denmark). A solution of the 𝛼-AMY was prepared by

dispersion of the powder in water (addition of 9 g water to 1 g

enzyme; freshly prepared before each experiment).

2.3. Gel Preparation

Dispersions of the starches (HACS-50/-70) with diﬀerent con-

centrations (6 and 9% w/w) were prepared according to Ulbrich

and Flöter[29] based on pressure cooking and subsequent high-

shear treatment. Gels were casted (containers: 30.0 mm diame-

ter, 20.0 mm height) and stored at 5.5±1.5°C for two diﬀerent

durations (1 and 14 days).[30]

2.4. Gel Characterization

The methodical approach for the gel characterization including

gel preparation, investigation of the mechanical strength, sepa-

ration of the gel matrix phases, and molecular characterization

of the gel network constituting starch before and after speciﬁc

enzymatic digestion is represented schematically in Figure 1.

2.4.1. Mechanical Gel Strength

The examination was made according to the description

elsewhere[30] with modiﬁcations. After storage, a fresh and planar

surface was realized by cutting the gels. The mechanical strength

of the gel matrix was determined by compression using a tex-

ture analyzer (Test Control II, Z1.0, 1kN, Zwick/Roell, Ulm, Ger-

many) equipped with a cylindrical penetration probe (diameter

25.4 mm). The peak force [N] of the ﬁrst penetration was taken

as the gel strength. The experiments were carried out in triple

determination; and the arithmetic average and the correspond-

ing standard deviation were calculated.

2200032 (2 of 10)

www.advancedsciencenews.com www.starch-journal.com

Figure 1. Schematic representation of the methodical approach of the starch gel preparation (A), the separation of the starch gel matrix into diﬀerent

constituent phases (B; networked polymers and dissolved polymers), and the isolation of enzyme resistant polymer residues from the gel network

involved starch fraction (C).

2.4.2. Separation of the Gel Matrix and Molecular Characterization

of the Phases

Preparation of Phases: In order to analyze the starch gels

molecularly by means of SEC-techniques (SEC-MALS-DRI, SEC-

cal-DRI), the samples were processed according to Ulbrich

et al.[31] The gel matrix was diluted 1:1 (w/w) with water and

subjected to a high-shear treatment (11 000 min−1, 30 s) with an

Ultra-Turrax T25 (IKA-Werke GmbH & Co. KG, Staufen, Ger-

many). Afterwards, an aliquot of the sample was centrifuged

(10 000 min-1, 15 min; Biofuge 28RS, Heraeus, Hanau, Ger-

many) for separation. The supernatant (SUP) was decanted (liq-

uid phase; dissolved starch), and an aliquot stabilized by di-

lution 1:10 (w/w) in preheated DMSO. The sediment (SED)

was blended with twice the amount of water (2500 min−1,30s;

mini shaker/vortex mixer, MS2, IKA-Werke GmbH & Co. KG,

Staufen, Germany), and centrifuged again (10 000 min-1,15min).

The washed SED was obtained (swollen phase; network involved

starch), and the SUP was discarded. Afterwards, 40 mL water was

added to the SED and premixed (mini shaker/vortex mixer, 5 s) in

order to completely obtain the SED out of the centrifuge cup. The

sample (swollen phase) was subjected to a high-shear treatment

(8000 min−1, 30 s), and an aliquot of the homogenized sample

was stabilized by dilution 1:10 (w/w) in preheated DMSO, and

heated at 90°C for 48 h under continuous stirring for complete

dissolving. The stabilized solutions (liquid phase, swollen phase)

were passed through 5 μm PTFE ﬁlters (Carl Roth GmbH &

Co. KG, Karlsruhe, Germany) before analysis (SEC-MALS-DRI;

2.4.2.2). All calculations of the relative and absolute amounts

were based on known concentrations of the gels, known dilu-

tion steps and amounts of phases obtained (e.g., after centrifu-

gation, weighing) as well as determination of the concentration

by means of SEC-techniques (detection of carbohydrate concen-

tration).

Separation Technique: The molecular characterization of the

solutions was carried out by means of SEC-MALS-DRI. The sep-

aration was executed with an SEC-3010 module (WGE Dr. Bu-

res GmbH & Co. KG, Dallgow-Döberitz, Germany) including

degasser, pump and auto sampler connected to a MALS detec-

tor and a diﬀerential refractive index detector (DRI). The MALS

detector was a Bi-MwA (Brookhaven Instruments Corporation,

Holtsville, NY, USA) ﬁtted with a diode laser operating at 𝜆=

635 nm and equipped with seven detectors at angles ranging

from 35°to 145°. The DRI was a SEC-3010 RI detector operating

at 𝜆=620 nm. Three columns in a row were used: AppliChrom

ABOA DMSO-Phil-P-100 (100-2500 Da), P-350 (5-1500 kDa), and

P-600 (20 to >20 000 kDa) (Applichrom, Oranienburg, Germany).

The samples were eluted with degassed DMSO (Carl Roth GmbH

& Co. KG, Karlsruhe, Germany) containing 0.1 M NaNO3at a

ﬂow rate of 0.5 mL∙min−1and a temperature of 70°C. During

the sample run on the SEC-MALS-DRI system (single determi-

nation), the data from the MALS and DRI detectors were col-

lected and processed using ParSEC Enhanced V5.61 chromatog-

raphy software to give the concentration of the eluted solution

and MM at each retention volume (Mi). The basis for the molec-

ular characterization by means of SEC-MALS-DRI has been de-

scribed elsewhere.[32,33]

The separation system was additionally calibrated (SEC-cal-

DRI) using a set of 10 pullulan standards as well as glucose with a

MM range between 180 and 805 000 g∙mol−1(PSS Polymer Stan-

dards Service GmbH, Mainz, Germany). The standards were dis-

solved in DMSO (2.5 mg∙mL−1w/v) and gently stirred 24 h at

80°C. The standard solutions were measured and the elution vol-

ume at the position of the peak maximum was used as the ref-

erence for the particular Miand the calculation of the calibration

curve. The calibration related to the DP was calculated from the

Midivided by 162. The weight average DP (DPw) was calculated

from the Mwdivided by 162.

2200032 (3 of 10)

www.advancedsciencenews.com www.starch-journal.com

Figure 2. SEC-chromatograms of Mal and Glc (A), and the SEC-

chromatogram of the gel structure (SED) of HACS-50 (9% starch concen-

tration, 14 days storage time) after enzymatic digestion with 𝛼-AMY (orig-

inal), the corresponding chromatogram ﬁt, and the chromatograms ob-

tained by means of mathematical peak separation (PeakFit) correspond-

ing to diﬀerent molecule fractions (F1-F3; B).

2.4.3. Partial Enzymatic Digestion of the Network-Involved Starch

and Molecular Characterization

A mass of 10 g of the homogenized sample (swollen phase;

2.4.2.1) was taken, and the respective volume of the speciﬁc en-

zyme solution added (𝛼-AMY: 500 μL𝛼-AMY solution, 𝛼-AMY-

PUL: 500 μL𝛼-AMY solution and 200 μL PUL solution, 𝛽-AMY:

160 μL𝛽-AMY solution, 𝛽-AMY-PUL: 160 μL𝛽-AMY solution and

200 μL PUL solution). Hydrolysis was performed at 40°Cfor

45 min while continuously stirring, and the dispersion subse-

quently heated at 95°C for 20 min for enzyme deactivation.

The solution was stabilized by dilution 1:10 (w/w) in preheated

DMSO, and heated at 90°C for 48 h under continuous stirring

for complete dissolving. The stabilized solutions were passed

through a 5 μm PTFE ﬁlter before analysis (SEC-MALS-DRI and

SEC-cal-DRI; 2.4.2.2).

The SEC-chromatograms of the enzymatically digested

swollen phase (𝛼-AMY/-PUL, 𝛽-AMY/-PUL) were advanced

analyzed using peak separation and analysis software PeakFit

Version 4.12 as described elsewhere.[34] According to Figure 2

(exemplarily shown for HACS-50/9% w/w starch concentra-

tion/14 days storage time/digestion with 𝛼-AMY), single peaks

(chromatograms) representing diﬀerent fractions were identi-

ﬁed. The SEC-chromatogram originating from the respective

enzyme solution/formulation/preparation was subtracted,[28]

and the relative chromatogram area of each separated fraction

was taken for the calculation of the relative amount/portion.

The values of Mwand DPwwere calculated by means of the

correspondent separated chromatogram and the MM curve (ﬁt)

from the MALS-detector (SEC-MALS-DRI; F1 [HMM fraction])

or the standard calibration curve (SEC-cal-DRI; F2 [LMM frac-

tion]), respectively, according to the description elsewhere.[28]

The portion of F3, which corresponds to mostly Mal/Glc,[28] was

taken as the degradable fraction.

2.5. Statistical Analysis

The impact of the diﬀerent parameters (starch type [HACS-50/-

70], starch concentration [6 and 9% w/w], storage time [1 and

14 days]) on diﬀerent properties was investigated statistically us-

ing Statgraphics Plus 5.0 software (ﬁrst experimental design: gel

strength and properties of gel network involved starch [absolute

amount and Mw], second experimental design: absolute amounts

of resistant [F1 and F2] and degradable fractions [F3] after speciﬁc

enzymatic digestion [𝛼-AMY, 𝛼-AMY-PUL, 𝛽-AMY, and 𝛽-AMY-

PUL]). The values for the probability of error (p-value) were listed

in the ANOVA tables (analysis of variance; Tables 2 and 3). With

ap-value less than 0.05, the factor investigated had a statistically

signiﬁcant eﬀect with a 95.0% level of signiﬁcance (indicated by

boldface type in the ANOVA tables).

3. Results and Discussion

3.1. Gel Network Constituting Starch

The starch gels were separated by centrifugation aiming to obtain

a gel network constituting starch fraction (swollen phase, super-

lattice) and the starch fraction dissolved in the liquid phase (en-

closed, not molecularly bound, or entangled). Figure 3 shows the

weighted SEC-chromatograms (chromatogram area corresponds

to the quantitative contribution within the respective gel phase)

of the isolated starch fractions. Most starch polymers were found

to be involved in the gel network (gel structure, SED; ≥95% w/w),

and just a marginal portion was dissolved in the liquid phase (sol-

uble polymers, SUP; ≤5% w/w). Since the predominant part of

the starch used was network involved, the molecular composition

of this fraction was very similar compared to the starting mate-

rial. The Mwof HACS-50 and HACS-70 is about 7.54∙106and

3.63∙106g∙mol−1, respectively.[28] In the case of HACS-50, partic-

ularly the starch concentration of the gel impacted the Mwof the

network involved starch systematically, which was lower at 6%

w/w (about 6.7∙106g∙mol−1) and marginally higher at 9% w/w

(about 7.65∙106g∙mol−1). For HACS-70, Mwof the network in-

volved starch isolated was slightly lower (2.8…3.3∙106g∙mol−1)

compared to the initial starch (starting material used). Since

the carbohydrate concentration was very low, the Mwof the re-

spective dissolved starch fraction could not be calculated via the

SEC-techniques. However, both phases when weighted should

– theoretically – yield the same total molecular composition as

the initial starch, which is eﬀectively the case estimated from

the measurements. Within the experimental design (starch va-

rieties, concentration and storage time), the gel network con-

stituting starch broadly equates the initial starch both quanti-

tatively and qualitatively. This fact is, however, absolutely inde-

pendent of the nature of the detailed involvement on a molec-

ular and supramolecular level, namely, e.g., crystalline or amor-

phous state, which is assumed to impact the functional properties

directly.

2200032 (4 of 10)

www.advancedsciencenews.com www.starch-journal.com

Figure 3. Weighted SEC-chromatograms of the starch involved in the gel network (SED) and starch polymers dissolved (SUP). The chromatogram

areas correspond to the quantitative contribution within the respective gel phase. Gels were prepared by variation of the starch type (HACS-50/-70), the

concentration (6 and 9% w/w) and the storage time (1 and 14 days).

Table 1. Enzymatic treatment (𝛼-/𝛽-AMY/-PUL), molecular fraction obtained (F1-F3, Mwrange), and SEC-technique applied for detailed molecular

characterization (MALS detection, conventional calibration).

Fractions obtained F1 F2 F3

HMMa) LMMb) Mal/Glcc)

[g∙mol−1][g∙mol−1][g∙mol−1]

Enzymes

𝛼-AMY 130.000…432.000 8.780…16.900 ●d)

𝛼-AMY-PUL – 8730…11.800 ●

𝛽-AMY 429.000…1.35∙10616.500…27.700 ●

𝛽-AMY-PUL – 15.600…21.700 ●

enzyme combinations SEC-MALS-DRI SEC-cal-DRI SEC-cal-DRI

a) High molar mass fraction; b) Low molar mass fraction; c) Fraction largely assigned to maltose and glucose, containing other structures is possible; d) Peak position

determined accords to maltose.

3.2. Detailed Involvement of Molecule Fraction/Fragments within

the Gel Network Fine Structure

As shown and discussed in Section 3.1 (see also Figure 3), most

of the starch was involved in the gel network (≥95% w/w), and

just a slight portion was dissolved in the liquid phase (≤5%

w/w). For analytical purposes, the network constituting starch

fraction was digested enzymatically using diﬀerent enzymes (𝛼-

AMY, 𝛼-AMY-PUL, 𝛽-AMY, and 𝛽-AMY-PUL) and characterized

chromatographically aiming for information in particular about

the starch’s state. Diﬀerent molecular fractions (F1, F2, and F3)

were identiﬁed and classiﬁed according to the MM level (Table 1),

where F1 (HMM fraction) and F2 (LMM fraction) are molecule

remnants not completely degraded to Mal/Glc, and F3 the prod-

uct from (nearly) complete hydrolysis to Mal/Glc. In particular,

the release of isomaltose is also possible for the digestion with 𝛼-

AMY. The respective relative portions of the diﬀerent degradation

products are summarized in Figure 4.

The partial resistance to complete molecular cleavage (F1

and F2) is mainly attributed to the (inherent) molecular struc-

ture (1; linear [𝛼-1,4-glycosidic linked AGU] or branched [𝛼-1,4-

glycosidic linked AGU and 𝛼-1,6-glycosidic linkages]), the speci-

ﬁcity of the applied enzyme/enzyme combination (2; endo-/exo-

/debranching enzyme), the participation or existence of molecule

segments in highly ordered structures (3; double helical struc-

tures and crystallization [retrogradation], AM-lipid-complexes

[resistant to amylolytic enzyme hydrolysis][35]) protecting the re-

spective molecule sections from/to enzymatic cleavage of the gly-

cosidic linkages,[36,37] and the interrelation of 1–3. In contrast, F3

is released from molecule segments, which are accessible and

not resistant to the respective enzymatic attack (not protected by

both molecule ﬁne structure and supramolecular structure).

The HMM fraction F1 was one of the products owing to di-

gestion of the gels with 𝛼-and𝛽-AMY, respectively (Table 1).

The origin is most likely the IM/AP fraction exclusively, and

the relative portions ranged about 10–15% (𝛼-AMY) and 5–20%

(𝛽-AMY), respectively. The Mwrange of F1 (𝛼-AMY) of about

130.000…430.000 g∙mol−1was signiﬁcantly higher compared to a

fraction ascribed to 𝛼-LDs in a previous study with 90.000 g∙mol-

1 (HACS-50) and 70.000 g∙mol−1(HACS-70).[28] This indicates

2200032 (5 of 10)

www.advancedsciencenews.com www.starch-journal.com

Figure 4. Relative portion of diﬀerent molecule fractions (F1: HMM fraction, F2: LMM fraction, and F3: Mal/Glc resulting from complete hydrolysis)

derived after speciﬁc enzymatic digestion (𝛼-AMY, 𝛼-AMY-PUL, 𝛽-AMY, and 𝛽-AMY-PUL) of the starch gel structure.

a speciﬁc and enhanced protection of the respective polymers

to enzymatic hydrolysis due to the involvement in the gel ma-

trix. A diﬀerent relationship was found for the HMM fraction

F1 after digestion with 𝛽-AMY. Compared to the respective 𝛽-

LDs (HACS-50: 800.000 g∙mol−1, HACS-70: 900.000 g∙mol−1),[28]

the Mwof F1 was basically in the same range or slightly

lower (430.000…1.35∙106g∙mol−1, one exception). Besides F1, the

degradation of the gel network with 𝛼-/𝛽-AMY resulted in a LMM

fraction F2 with a Mwrange of 8.780…16.900 g∙mol−1(𝛼-AMY)

and 16.500…27.700 g∙mol−1(𝛽-AMY), representing molecule sec-

tions which are supposed to be largely constituent of crystalline

regions within the gel network. The origin is presumably mostly

the AM fraction, however, contribution of the branched molecule

fractions isn’t impossible and even likely, respectively.

The LMM fraction F2 after digestion with the enzyme com-

binations (𝛼-AMY-PUL and 𝛽-AMY-PUL) represents the only

enzyme resistant polymer fraction of the gel network involved

starch. The Mwrange is 8.730…11.800 g∙mol−1(𝛼-AMY-PUL)

and 15.600…21.700 g∙mol−1(𝛽-AMY-PUL), respectively, and in

particular the lower limits accord well with that of the F2 after

𝛼-/𝛽-AMY treatment discussed before. F2, owing to digestion

with 𝛼-AMY-PUL, ultimately represents the fraction of the

polymer sections existent in the crystalline form or completely

embedded in a crystalline block and hence are protected. The

DPwwas calculated to be between 54 and 73, which accords

basically well with literature data,[38] and the relative portion was

about 29…34% (HACS-50) and 40…43% (HACS-70) of the total

starch (Figure 4), which is remarkable. In contrast, F2 released

after 𝛽-AMY-PUL digestion consists of longer chains on average

(DPw96…134). Presumably, each chain is preserved by both the

partial involvement in crystalline sections by itself (protected by

supramolecular structure) and the limitation due to the fact that

𝛽-AMY is an exo-enzyme cleaving exclusively 𝛼-1,4-glycosidic

linkages (every second) from the non-reducing end of the chain

(protected by the enzyme’s speciﬁcity). The relative portion of

F2 (𝛽-AMY-PUL) was about 43…46% (HACS-50) and 55…60%

(HACS-70) of the starch (Figure 4), which is in average about

13% (HACS-50) and 16% (HACS-70) higher when compared

to degradation with 𝛼-AMY-PUL. The diﬀerences regarding

the relative portions (F2 of 𝛼-AMY-PUL as well as 𝛽-AMY-PUL)

2200032 (6 of 10)

www.advancedsciencenews.com www.starch-journal.com

Table 2. ANOVA of the impacts starch type (HACS-50/-70), starch concen-

tration (6 and 9% w/w), and storage time (1 and 14 days) have on gel

strength and starch in the gel network (boldface type: factor investigated

had a statistically signiﬁcant eﬀect with a 95.0% level of signiﬁcance).

Impact Starch in gel network

Gel strength Amount Mw

Starch type 0.0089 0.4788 0.0001

Starch concentration 0.0003 0.0000 0.1775

Storage time 0.0647 0.1748 0.6899

between HACS-50 and HACS-70 are deeply rooted in the varying

AM/AP-ratio. The higher AM content of the starch increases the

relative portion of polymers or rather polymer segments, which

are involved in crystalline structures.

F3 quantitatively represents the gel network involved starch

and starch polymer chain sections, respectively, which were un-

restrictedly accessible for the enzymes (𝛼-AMY, 𝛼-AMY-PUL, 𝛽-

AMY, and 𝛽-AMY-PUL) to complete hydrolysis to Mal/Glc. This

fraction was calculated to be about 40–68% (HACS-50) and 33–

57% (HACS-70), respectively, and represents the minimum of

the starch within the gel network existing in the amorphous state.

3.3. Absolute Portions/Amounts of Diﬀerent Molecule Fractions

and Correlation to Gel Strength

The relative composition of the gel network involved starch and

the state assigned to it were discussed in the above Section 3.2.

However, the absolute content of the dissimilar existing starch

(state) is more important, since a direct correlation to the result-

ing mechanical gel strength was assumed.

Tables 2 and 3summarize the data from the statistical anal-

ysis (ANOVA), and Figure 5 shows the gel strength and the ab-

solute portions/contents of the enzyme resistant (F1 and F2) as

well as the corresponding degradable starch (F3). Starch type

and starch concentration are statistically signiﬁcant impacts on

the gel strength (ANOVA, Table 2 [gel strength], Figure 5A). In-

creasing AM content of the starch and increasing starch concen-

tration in the gel enhance the mechanical strength remarkably.

Higher storage time also increased gel ﬁrmness, but the impact

was not found to be statistically signiﬁcant at a 95% conﬁdence

level. Since nearly the total amount of the starch used for the

gel was bound to the network, independent of the starch con-

centration adjusted in the system (6 or 9% w/w), the impact was

highly signiﬁcant (ANOVA, Table 2 [amount]). However, starch

type (which includes the diﬀerent AM contents of the genotypes)

and storage time did not impact the amount of starch bound to

the gel network. The fact that almost all starch was bound to the

network explains that the molecular composition of the gel net-

work bound starch was exclusively controlled by the starch type

(HACS-50/-70), which is obvious from the ANOVA (Table 2 [Mw])

and Figure 3 (Section 3.1).

The impact of the gel’s starch concentration on the mechan-

ical strength is explicit (Figure 5A); the gel strength increases

remarkably up to about 40 (HACS-50) and 50 N (HACS-70), re-

spectively, due to enhancement from 6 to 9% w/w. However, the

simultaneously strong inﬂuence of the starch type (varying AM

content) indicates the diﬀerent eﬀect of both non-branched (AM)

and branched (IM and AP fraction) on contributing to the devel-

opment of a supramolecular structure and corresponding ﬁrm-

ness. Moreover, a respective impact of the IM fraction,[28] which

exists most likely in both starch varieties, seems self-evident. Di-

gestion of the gel matrix with 𝛼-AMY caused about 1.0 % w/w

(absolute) of the HMM fraction F1 (Figure 5B), independently

on the parameters varied (ANOVA, Table 3). In contrast, the

amount of the resistant HMM fraction F1 after degradation with

𝛽-AMY was impacted with statistical signiﬁcance by the starch

type (ANOVA, Table 3). The lower AM content of the HACS-50

resulted in a higher portion of resistant branched structures (Fig-

ure 5B). Since F1, which is allocated to structures similar to 𝛼-/𝛽-

LDs, does not exist after digestion in combination with PUL (𝛼-

AMY-PUL, 𝛽-AMY-PUL), the fraction is supposed to be present

in the amorphous state, acting probably as a ﬁller retaining

water.

The portion of the LMM fraction F2 was found between about

2.0 and 4.0% w/w (absolute, digestion with 𝛼-AMY/-PUL) and

between about 2.5 and 5.0% w/w (absolute, digestion with 𝛽-

AMY/-PUL, Figure 5C), respectively. Both starch type and starch

concentration were statistically signiﬁcant factors in the respec-

tive resistant portions (Table 3). Increasing AM content of the

starch as well as increasing starch concentration of the gel en-

hanced the portions. F2 isolated after digestion with 𝛼-/𝛽-AMY-

PUL, but particularly after 𝛼-AMY-PUL treatment, accounts for

the molecule fraction largely present in the double helical and

the crystalline state. A direct correlation between the absolute

amount of the LMM fraction F2 and the gel strength is obvi-

ous from Figure 5A and C. The origin of the resistant fraction

F2 is assumed to be largely the AM, since the signiﬁcant involve-

ment of (unbranched) chain segments with DPwhigher than 50

(Table 1) based on AP would be impossible. The participation of

the IM fraction[4] is questionable because of the branched nature.

The amounts of F2 (Figure 5C) are in all cases not higher than

the (theoretical) accounted for AM in the respective gel systems,

which are also indicated in Figure 5C.

The amount of the degradable fraction F3, which was deter-

mined to about 2.5-5.5% w/w (𝛼-AMY/-PUL) and about 2.0-5.0%

w/w (𝛽-AMY/-PUL), respectively, is dependent on both the starch

type and the starch concentration of the gel (Figure 5D). The fac-

tors are statistically signiﬁcant (Table 3, exception: starch type/𝛽-

AMY). On the one hand, fraction F3 decreases with an increasing

amount of AM in the system, since the unbranched AM is more

protected from the enzymatic degradation due to large scale in-

volvement in the double helical and crystalline areas as discussed

in the previous section. Simultaneously, the content of branched

polymers (IM, AP), which are more exposed to complete degra-

dation to Mal/Glc due to the amorphous state and associated ac-

cessibility, is lower. On the other hand, the increased amount of

starch in the gel system (enhancement of the starch concentra-

tion) increases the amount of degradable material (Figure 5D).

Compared to the treatment with 𝛼-/𝛽-AMY, the amount of F3 in-

creased generally when digested using the enzyme combination

(𝛼-/𝛽-AMY-PUL). That increase of F3 (𝛼-/𝛽-AMY-PUL) basically

accounts for the complete disappearance of F1, the digestion of

the branched fraction associated with 𝛼-/𝛽-LDs as discussed be-

fore. Moreover, the relationship of F2 and F3 is very interesting

when comparing the degradation based on 𝛼-/𝛽-AMY. The higher

2200032 (7 of 10)

www.advancedsciencenews.com www.starch-journal.com

Table 3. ANOVA of the impacts starch type (HACS-50/-70), starch concentration (6 and 9% w/w), and storage time (1 and 14 days) on absolute amounts

of resistant (F1 and F2) and degradable fractions (F3) after speciﬁc enzymatic digestion (𝛼-AMY, 𝛼-AMY-PUL, 𝛽-AMY, and 𝛽-AMY-PUL) of the starch gel

structure (boldface type: factor investigated had a statistically signiﬁcant eﬀect with a 95.0% level of signiﬁcance).

Enzyme 𝛼-AMY 𝛽-AMY

Fraction resistant degraded resistant degraded

F1 (HMMa) )F2(LMM

b) )F3(Mal/Glc

c) ) F1 (HMM) F2 (LMM) F3 (Mal/Glc)

Impact

Starch type 0.8237 0.0380 0.0079 0.0253 0.0064 0.0520

Starch concentration 0.1743 0.0117 0.0010 0.0583 0.0014 0.0018

Storage time 0.7496 0.8626 0.7561 0.5499 0.5081 0.8782

Enzyme 𝛼-AMY-PUL 𝛽-AMY-PUL

Fraction resistant Degraded resistant degraded

F2 (LMM) F3 (Mal/Glc) F2 (LMM) F3 (Mal/Glc)

Impact

Starch type 0.0005 0.0004 0.0037 0.0032

Starch concentration 0.0001 0.0000 0.0004 0.0010

Storage time 0.0527 0.1114 0.7605 0.9870

a) High molar mass fraction; b) Low molar mass fraction; c) Fraction largely assigned to maltose and glucose, containing other structures is possible.

Figure 5. Gel strength (A) dependent on starch type (HACS-50/-70), starch concentration (6 and 9% w/w), and storage time (1 and 14 days), and

absolute contents of diﬀerent resistant molecule fractions (B and C: HMM F1 and LMM F2) as well as the degraded fraction (D: F3, completely digested

to Mal/Glc) derived after speciﬁc enzymatic digestion of the gel matrix (𝛼-AMY, 𝛼-AMY-PUL, 𝛽-AMY, and 𝛽-AMY-PUL). The calculative amounts of AM

(indicated in C) and AP/IM (indicated in D) in the gel are included (determination in Ulbrich et al., 2021).

2200032 (8 of 10)

www.advancedsciencenews.com www.starch-journal.com

Figure 6. Starch gel composition (schematic).

amount of enzyme resistant fraction F2 results in a lower amount

of F3, and vice versa.

3.4. Gel Microstructure – Contribution of the Starch

Polysaccharides within the Gel Matrix

Aqueous dispersions of the starch genotypes (HACS-50/-70;

detailed molecular investigation in a previous study[28]) with

diﬀerent concentrations (6 and 9% w/w) were prepared by pres-

sure cooking and subsequent high-shear treatment of the paste

(Figure 6A). Any granular remnants within the aqueous system

were largely precluded by the procedure. The gels were casted

and stored (1 and 14 days). For analytical purposes, the gel matrix

(Figure 6B) was separated by centrifugation into a liquid and a

swollen phase (Figure 6C), where about 5% of the starch were

proved to be dissolved in the liquid phase (SUP; low polysac-

charide concentration, polymer-deﬁcient), and about 95 % were

found to exist as a highly hydrated starch based matrix (SED;

high polysaccharide concentration, polymer-rich) (Figure 6C

and D). Since the starch used, independent of the varied param-

eters, was almost completely bound to the matrix (or entangled;

existent in the swollen phase), the molecular composition of

the latter accorded well with that of the initial starch, i.e., both

branched (AP/IM fraction) as well as largely non-branched (AM)

polymer fractions are eﬀectively matrix-constituting.

Within the gel matrix (network), the branched polymer frac-

tion (AP/IM) appears to exist entirely in the amorphous state (ex-

posed to enzymatic digestion), most likely connected via inter-

molecular associations to the microscopic network somehow or

other or embedded within. In contrast, the largely non-branched

structure fraction (AM) most probably constitutes the network

strands (ﬁlament, microﬁber), which are of semicrystalline char-

acter containing crystalline blocks (preserved from enzymatic di-

gestion) (Figure 6E). These crystalline sections are estimated to

be about 20–26 nm in length,[27] since the molecule segments

inside the crystallites possess a DPwbetween 54 and 73 (double

helices or B-type crystalline, 6 AGU per helical turn [left-handed,

6-fold helices], pitch per helical turn about 2.10 nm[39]). A di-

rect relationship between the absolute portion of the enzyme re-

sistant carbohydrate fraction (particularly when digested with 𝛼-

AMY-PUL) and the gel strength is evident. In contrast, the dis-

solved starch within the liquid phase, which is not associated with

the network, is most likely irrelevant concerning gel ﬁrmness. A

foam-like structure of the gel is conceivable. In addition to a wide

variety of reasons, the comparatively high degree of retrograda-

tion caused the high turbidity of the gel (light transmittance 0.0%,

results not ampliﬁed).

4. Conclusions

A very complex method for the microstructural examination

of starch gels was applied and advanced, and the methodical

approach was found to be suitable regarding the study’s speciﬁc

issue. The starch polymer fraction included within the swollen

gel phase – constituting the actual matrix network – was ana-

lyzed in detail, particularly with respect to the state of diﬀerent

existing segments (enzymatically degradable/resistant; amor-

phous/crystalline) and the molecular properties. An adaptive

classiﬁcation was conducted, and by means of calculation of

weighted amounts, both the relative involvement of the starch

polysaccharides within the organization levels of the starch’s net-

work and the absolute amounts of distinct fractions within the

aqueous starch gel were considered. The correlation of the ap-

parent gel strength to quantities of speciﬁc network-constituting

starch polysaccharide structures is impressive. The increase of

the gel ﬁrmness from 10 (HACS-50, 6% w/w, 1 d) to about 80 N

(HACS-70, 9% w/w, 14 d) is accompanied by an enhancement

of the AP/IM fraction (branched, amorphous) from about 3

to 4% w/w and concurrently, of the double helical/crystalline

2200032 (9 of 10)

www.advancedsciencenews.com www.starch-journal.com

portion from about 2 to 4% w/w in the gel microstructure, which

illustrates the dependence of functional properties on molecular

parameters and the superior structure (supramolecular) of the

network. The latter are governed by the inherent molecular com-

position of the starch, the concentration as well as the storage

conditions. The results of the study provide a new insight in the

starch gels microstructure on a molecular and supramolecular

level, and can support the speciﬁc industrial application of starch

products as a gelling agent.

Acknowledgements

This research was ﬁnancially supported by the German Federal Ministry

for Economic Aﬀairs and Energy (BMWi) within the scope of the Indus-

trial Collective Research (IGF) of the German Federation of Industrial Re-

search Associations (AiF; research project 20248 N). The authors would

like to thank Ingredion Germany GmbH for providing the starches, and

SternEnzym GmbH & Co. KG as well as Novozymes A/S for providing the

enzymes. The authors also thank Ms. Donna Hastings for checking the

manuscript linguistically.

Conﬂict of Interest

The authors declare no conﬂict of interest.

Author Contributions

M.U. contributed to Conceptualization, Methodology, Software, Valida-

tion, Writing - Original Draft, Writing - Review - Editing, Visualization, Su-

pervision, Project administration. F.S. contributed to Data curation, Valida-

tion, Formal analysis. E.F. contributed to Head of department, Resources.

Data Availability Statement

Research data are not shared.

Keywords

high amylose corn starch, network involved polymers, size exclusion chro-

matography, speciﬁc enzymatic digestion, starch gels

Received: January 27, 2022

Revised: February 21, 2022

Published online: March 20, 2022

[1] S. You, S.-T. Lim, Cereal Chem. 2000,77, 303.

[2] C. Takeda, Y. Takeda, S. Hizukuri, Carbohydr. Res. 1993,246, 273.

[3] T. Baba, Y. Arai, Agric. Biol. Chem. 1984,48, 1763.

[4] Y.-J. Wang, P. White, L. Pollak, J. Jane, Cereal Chem. 1993,70, 171.

[5] Y.-C. Shi, T. Capitani, P. Trzasko, R. Jeﬀcoat, J. Cereal Sci. 1998,27,

289.

[6] E. Bertoft, Z. Qin, R. Manelius, Starch/Staerke 1993,45, 420.

[7] P. Colonna, C. Mercier, Carbohydr. Res. 1984,126, 233.

[8] M. Ulbrich, T. Beresnewa-Seekamp, W. Walther, E. Flöter,

Starch/Staerke 2016,68, 399.

[9] M. Ulbrich, C. Natan, E. Flöter, Starch/Staerke 2014,66, 903.

[10] N. U. Haase, Plant Breed 1993,111, 325.

[11] J. Jane, Y. Y. Chen, L. F. Lee, A. E. McPherson, K. S. Wong, M. Ra-

dosavljevic, T. Kasemsuwan, Cereal Chem. 1999,76, 629.

[12] Y. Song, J. Jane, Carbohydr. Polym. 2004,41, 365.

[13] T. Harada, A. Harada, in Polysaccharide association structures in food

(Ed. R. H. Walter), Marcel Dekker, NY, 1998,Ch.3.

[14] C. J. A. M. Keetels, T. van Vliet, P. Walstra, Food Hydrocolloids 1996,

10, 343.

[15] C. J. A. M. Keetels, T. van Vliet, P. Walstra, Food Hydrocolloids 1996,

10, 363.

[16] C. J. A. M. Keetels, T. van Vliet, P. Walstra, Food Hydrocolloids 1996,

10, 355.

[17] V. M. Leloup, P. Colonna, A. Buleon, Journal of Cereal Science 1991,

13,1.

[18] S. G. Ring, Starch/Staerke 1985,37, 80.

[19] C. G. Biliaderis, J. R. Tonogai, J. Agric. Food Chem. 1991,39, 833.

[20] P. D. Orford, S. G. Ring, V. Carroll, M. J. Miles, V. J. Morris, J. Sci. Food

Agric. 1987,39, 169.

[21] C. G. Biliaderis, in Polysaccharide association structures in food (Ed. R.

H. Walter), Marcel Dekker, NY, 1998,Ch.4.

[22] M. J. Miles, V. J. Morris, P. D. Orford, S. G. Ring, Carbohydr. Res. 1985,

135, 271.

[23] A. H. Clark, S. B. Ross-Murphy, Br.Polym.J.1985,17, 164.

[24] M. J. Gidley, Macromolecules 1989,22, 351.

[25] A. H. Clark, M. J. Gidley, R. K. Richardson, S. B. Ross-Murphy, Macro-

molecules 1989,22, 346.

[26] V. J. Morris, Trends Food Sci. Technol. 1990,1,2.

[27] V. M. Leloup, P. Colonna, S. G. Ring, K. Roberts, B. Wells, Carbohydr.

Polym. 1992,18, 189.

[28] M. Ulbrich, F. Scholz, E. Flöter, Starch/Staerke 2022,

star.202100303.R1.

[29] M. Ulbrich, E. Flöter, Starch/Staerke 2020,73, 2000145.

[30] M. Ulbrich, E. Flöter, Starch/Staerke 2020,73, 2000146.

[31] M. Ulbrich, T. A. Terstegen, E. Flöter, Starch/Staerke 2018,71,

1800080.

[32] P. J. Wyatt, Anal. Chim. Acta 1993,272,1.

[33] S. Podzimek, L. Scattering, Size Exclusion Chromatography and Asym-

metric Flow Field Flow Fractionation, Powerful Tools for the Character-

ization of Polymers, Proteins and Nanoparticles, John Wiley & Sons,

Hoboken, NJ 2011.

[34] M. Ulbrich, V. Lampl, E. Flöter, Starch/Staerke 2016,68,885.

[35] J. Hasjim, Y. Ai, J.-l. Jane, in Resistant Starch: Sources, Applications

and Health Beneﬁts,.

[36] H. Zhang, Z. Jin, Carbohydr. Polym. 2011,86, 1610.

[37] R. C. Eerlingen, J. A. Delcour, J. Cereal Sci. 1995,22, 129.

[38] J. F. Robyt, in Starch: Chemistry and technology Enzymes and their ac-

tion on starch (Eds. J. BeMiller, R. Whistler), Elsevier Inc., Oxford,

2009,Ch.7.

[39] A. Imberty, S. Perez, Biopolymers 1988,27, 1205.

2200032 (10 of 10)