Extrusion Processing of Pure Chokeberry (Aronia melanocarpa) Pomace: Impact on Dietary Fiber Profile and Bioactive Compounds

The partial substitution of starch with dietary fiber (DF) in extruded ready-to-eat texturized (RTE) cereals has been suggested as a strategy to reduce the high glycemic index of these food products. Here, we study the impact of extrusion processing on pure chokeberry (Aronia melanocarpa) pomace powder (CPP) rich in DF and polyphenols (PP) focusing on the content and profile of the DF fractions, stability of PP, and techno-functional properties of the extrudates. Using a co-rotating twin-screw extruder, different screw speeds were applied to CPP with different water contents (cw), which resulted in specific mechanical energies (SME) in the range of 145–222 Whkg−1 and material temperatures (TM) in the range of 123–155 °C. High molecular weight soluble DF contents slightly increase with increasing thermomechanical stress up to 16.1 ± 0.8 g/100 g dm as compared to CPP (11.5 ± 1.2 g/100 g dm), but total DF (TDF) contents (58.6 ± 0.8 g/100 g dm) did not change. DF structural analysis revealed extrusion-based changes in the portions of pectic polysaccharides (type I rhamnogalacturonan) in the soluble and insoluble DF fractions. Contents of thermolabile anthocyanins decrease linearly with SME and temperature from 1.80 ± 0.09 g/100 g dm in CPP to 0.24 ± 0.06 g/100 g dm (222 Whkg−1, 155 °C), but phenolic acids and flavonoids appear to be largely unaffected. Resulting techno-functional (water absorption and water solubility) and physical properties related to the sensory characteristics (expansion, hardness, and color) of pure CPP extrudates support the expectation that granulated CPP extrudates may be a suitable food ingredient rich in DF and PP.


Introduction
The major component of ready-to-eat texturized (RTE) cereal products, for example, breakfast cereals and snacks, is starch, an easily digestible carbohydrate making these products high glycemic index food. Consumption of high glycemic index food is considered to play an important role in the development of many metabolic disorders, for example, obesity, type 2 diabetes, and cardiovascular diseases [1]. A high intake of dietary fiber (DF) may reduce these risks [2][3][4][5]. Thus, partial or complete substitution of starch with DF-rich materials in RTE cereals is a promising approach towards improving the nutritional quality of these starch-based products [6].

Extrusion Processing
All extrusion trials were carried out using a co-rotating twin-screw extruder (ZSK 26 Mc, Coperion, Stuttgart, Germany) with a barrel length to screw diameter ratio (L/D) of 29 and a screw diameter of 25.5 mm. The extruder barrel consists of seven sections. At the first section, CPP is fed with a constant solid mass flow of 9 and 8 kgh −1 using a gravimetrically controlled feeder (DDW-DDSR40, Brabender, Duisburg, Germany). In addition, in the first section, a piston-membrane pump (KM 25, Alldos, Pfinztal, Germany) added water (1 kgh −1 and 2 kgh −1 ) to keep the total mass flow constant (10 kgh −1 ). The addition of water resulted in water contents (c w ) of 13% and 23%, respectively. The extruded material left the extruder by passing a circular die of 3 mm diameter and 10 mm length. All trials were run at screw speeds (n) of 200, 400, 600, and 800 min −1 and samples were taken upon achieving steady state conditions. The barrels 2-7 could be heated and cooled separately and temperatures were adjusted to usual values of 40, 60, 80, 100, 100, 100 • C (T B 100 • C). The material temperature (T M ) was measured at the die entrance using a thermocouple (type J, Ahlborn, Holzkirchen, Germany). Specific mechanical energy (SME) (Whkg −1 ) was calculated by the following Equation (1) (1) where n and n max are the actual and maximum screw speed (1800 min −1 ), respectively; M d and M d,unload are the actual and idle torque (%), respectively;ṁ represents the total mass flow (kgh −1 ); and P max represents the maximum engine power (40 kW). All experiments performed are given in Table 1. Table 1. Experiments performed. Extrusion parameters (n, screw speed (min −1 ); T B , barrel temperature ( • C); c w , water content (%); SME, specific mechanical energy (Whkg −1 ); T M , material temperature ( • C); total mass flow for all trials was 10 kgh −1 ). Each trial was carried out twice. Immediately after extrusion, the samples were equilibrated at 40 • C for 15 min (Heraeus UT6200, Hanau, Germany), packaged in vacuum bags, stored at −80 • C, and analyzed as described in the following sections. Samples A-D were used to describe the effects of thermomechanical stress. The impact of two different c w at constant n and T B was evaluated comparing A1 and A2 as well as B1/B2, C1/C2, and D1/D2. Samples were also analyzed for dietary fiber (DF) and polyphenols (PP) contents and profiles, as well as for sugar contents in order to evaluate the effects of n and T B at c w 13% (A1, B1, C1, and D1) and of c w at highest n 800 min −1 (D1 and D2).

Dietary Fiber Isolation
Preparative isolation of IDF and HMW-SDF fractions of CPP and extruded materials (A1, D1, and D2) was performed, as described earlier [27]. Briefly, mortar ground powder (1 g) was suspended in sodium phosphate buffer and subsequently incubated with thermostable α-amylase, protease, and amyloglucosidase. IDF was separated by centrifugation (4696× g, 10 min, Multifuge X1, Thermo Fisher Scientific, Schwerte, Germany) and washed with water, 99.5% ethanol, and acetone. Using the supernatant from previous centrifugation, HMW-SDF fractions were precipitated from 80% ethanol, separated by centrifugation, and washed using 80% aqueous ethanol, 99.5% ethanol, and acetone. Both fiber fractions were vacuum dried at 60 • C (Vacutherm, Thermo Fisher Scientific, Schwerte, Germany) and stored at 20 • C until analysis. In contrast to the method described earlier [27], corrections for residual proteins or ash were not performed.

Monosaccharide Constituents
In order to hydrolyze IDF and HMW-SDF polysaccharides sulfuric acid hydrolysis [28] and methanolysis [29] were carried out, respectively. The monomer composition was determined by high-performance anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD). The analysis was performed on a Dionex ICS-5000 system equipped with a CarboPac PA-20 column (6.5 µm, 150 × 3 mm, Thermo Fisher Scientific, Schwerte, Germany) at 25 • C. A ternary gradient was applied, as described by Wefers et al. [30] with ultra-pure water (A), 0.1 M sodium hydroxide (B), and 0.1 M sodium hydroxide with 0.2 M sodium acetate (C) at a flow rate of 0.4 mL/min. Results are expressed as mol%.

Monosaccharide Linkage Patterns
Methylation analysis was performed, as described [30,31]. Briefly, IDF and HMW-SDF fractions were swollen in dimethyl sulfoxide, methylated twice using sodium hydroxide and methyl iodide, hydrolyzed using 2 M trifluoroacetic acid, reduced by adding sodium borodeuteride in 2 M aqueous ammonia, and finally acetylated by using methylimidazole as catalyst and acetic anhydride. The partially methylated alditol acetates (PMAA) were analyzed by GC-MS for identification and by GC-FID for relative quantification, using molar response factors described by Sweet et al. [32].

Analysis of Sugar Contents
The sugar contents were analyzed twice in LMW-SDF fractions from CPP and extruded materials (A1, B1, D1, and D2) as prepared according to AOAC 2011.25 [25], as described in Section 2.3.1. Commercial enzyme test kits (R Biopharm AG, Darmstadt, Germany) were used.

Residual Moisture Content
Karl Fischer titration (Titroline alfa, Schott Instruments GmbH, Mainz, Germany) was used to determine the residual moisture content of CPP and extruded samples (A1, B1, D1, and D2) six times each [35]. CPP as well as all extruded samples were milled using a coffee mill (M55, Petra Electric, Ense, Germany), and then sieved to particle size between 0.07 and 0.14 mm. Sieved powders were dried in a vacuum dryer (VT 5042 EK, Heraeus, Hanau, Germany) at 40 • C and 8 mbar. The water solubility index (WSI) and water absorption index (WAI) were determined according to Anderson [36] with slight modifications. Sieved powders (0.5 g) were added to 19.5 g of demineralized water, and the suspensions were mixed for 1 min (REAX top, Heidolph Instruments GmbH & Co. KG, Schwabach, Germany). The samples were shaken at 200 min −1 (Orbital shaker Incubator SI 50, Stuart, Staffordshire, United Kingdom) for 24 h at room temperature (about 25 • C) and centrifuged at 4600× g for 50 min at 20 • C (Rotanta 460 R, Andreas Hettich GmbH & Co. KG, Tuttlingen, Germany). After separation, the wet precipitates were weighed directly, whereas the supernatants were dried for 72 h at 80 • C (T-6060 Heraeus, Hanau, Germany) before weighing. WAI and WSI were calculated according to the following Equations (2) and (3): Analyses were performed in triplicate.

Expansion Indices
The sectional expansion index (SEI) and longitudinal expansion index (LEI) are parameters to describe the expansion of extruded products. The SEI is defined as the ratio of the cross-sections of the expanded extrudate and the die (Equation (4)) as follows: The die diameter d die was 3 mm for all extrusion trials. The diameter of extrudates d ext was determined 6 times by using a caliper 24 h after extrusion.
The LEI is defined as ratio of the extrudate velocity v ext after the die exit and the melt velocity inside the die v die . Extrudate samples were taken manually for a period of 3 s and the final lengths were measured. The LEI was calculated according to Equation (5) using Equations (6) and (7) as follows: where l is the measured length of the extrudate, t the time for sampling (3 s),ṁ the total mass flow (10 kgh −1 ), A die the area of the die (7.07 m 2 ), and ρ die the density of matrix (1400 kgm −3 ) [37].

Hardness
Hardness of the product was measured by using a texture analyzer (Z2.5 TS, Zwick-Roell, Ulm, Germany). A Kramer shear cell with one blade was used for testing (settings of pretest speed of 0.1 mmin −1 , test speed of 0.01 mmin −1 , test distance 6 mm, and pre-force 0.2 N). Hardness was determined four times.

Color
The color of the sieved and dried, samples was measured by a spectral photometer (CM 700d, Konica Minolta, Marunouchi, Tokyo, Japan). The powder was placed on a white sheet. The color was determined three times.

Statistics
Structural data of isolated DF (monosaccharide constituents and linkage patterns, arabinan profiling) are reported as mean ± range/2; all other data are presented as mean ± standard deviation (SD). ANOVA followed by Holm-Sidak test was performed to determine statistical significance (p < 0.05) between groups using SigmaPlot software (version 13.0, Systat Software GmbH, Erkrath, Germany). Pearson's correlation coefficient (r and p values) were reported to describe the linear correlation between techno-functional and sensory relevant physical properties and SME or T M , respectively.

Extrusion Processing
Specific mechanical energy (SME) and material temperature (T M ) give information on the thermomechanical stress profiles achieved during extrusion processing and are commonly used to evaluate the impact of processing on nutritional and techno-functional attributes of extrudates. As expected, an increase in screw speed (n) from 200 to 800 min −1 led to an increase in SME by 35% for a water content (c w ) of 13% ( Figure 1A). Increasing the c w to 23% (total mass flow was kept constant at 10 kgh −1 ) at different n results in decreases in SME by 8 to 18% (Table 1) due to a reduction in the matrix viscosity and subsequently lower shear stresses [38]. Figure 1B shows the T M at the die entrance, shortly before exiting the extruder, for both c w 13% and 23%. The lowest T M of 123 • C is measured at c w 23% and the lowest n of 200 min −1 . Increasing n and reducing c w increases T M up to 155 • C. T M was always significantly higher than T B , which was kept constant at 100 • C, indicating a high viscous dissipation energy resulting from mechanical stresses.

Thermomechanical Stability of Dietary Fiber
Data about the impact of extrusion processing on the dietary fiber (DF) structure of pure fruit pomaces are scarce. However, both Aronia melanocarpa and Malus domestica (apple) bear the same botanical fruit type (pome) and both are members of the Rosaceae subtribe Malinae. Therefore, the results are compared to extruded pure apple pomace wherever possible.

Dietary Fiber Contents
Individual DF (insoluble DF, IDF; high molecular weight soluble DF, HMW-SDF; and low molecular weight soluble DF, LMW-SDF) as well as total DF (TDF) contents of chokeberry pomace powder (CPP) before and after extrusion processing are shown in Table 2. Table 2. Dietary fiber (DF) contents of chokeberry pomace powder (CPP) before and after extrusion processing (g/100 g dm, mean ± SD, n = 2) at water contents of 13% (A1, B1, and D1) and 23% (D2), respectively. SME, specific mechanical energy (Whkg −1 ); T M , material temperature ( • C); IDF, insoluble DF; HMW-SDF, high molecular weight soluble DF; LMW-SDF, low molecular weight soluble DF; TDF, total DF. The contents of TDF and LMW-SDF remain unchanged irrespective of the SME applied and the materials c w . Likewise, no significant changes in the contents of IDF, the main fiber fraction, are identified. Yet, at the highest SME (D1, 222 ± 10 Whkg −1 ) the IDF content tends to be slightly reduced by about 5%. In contrast, at the highest SME (D1), the HMW-SDF contents are enhanced with a significant increase of about 40%. This reflects the solubilization of proportions of IDF arabinans which results in higher amounts of soluble arabinans in HMW-SDF (see below). The SDF/IDF ratio is shifted from 0.32:1 in CPP to 0.42:1 under D1 conditions. An increasing solubilization of DF with increasing SME was also reported for the extrusion of pure apple pomace [39]. A higher c w (D2, 23%) reduces SME and T M ( Table 2) but has no impact on the HMW-SDF content as compared with extrusion at lower c w (D1, 13%). Different from the results of this study, the IDF contents increased largely during defined thermomechanical treatment in a closed cavity rheometer after long-time treatment (20 min) at 140 • C and 160 • C (shear rate 50 s −1 , c w 12%) [20].

Sample
The nearly unmodified IDF contents after extrusion at all applied conditions are explained by the very short residence time of the material within the extruder, in the range of seconds, which is one of the advantages of short-time extrusion at high temperatures [40,41].
The knowledge gained here about the impact of extrusion processing of pure chokeberry pomace on the structure and contents of its DF is new, as literature data on the extrusion of pure fruit pomace is scarce. Many papers describe the impact of extrusion on the DF of fruit pomace and starch blends. However, in such studies it is difficult to distinguish between the impact of extrusion processing on starch and on the non-starch polysaccharides from pomace. In this study, the extrusion of pure CPP was investigated focused on the possible impact of processing on the content and profile of the HMW-SDF fraction [13][14][15] which is associated with beneficial health-related effects [2]. Extrusion results in an increase in this DF fraction and extrudates contain up to 60% of TDF. These results suggest that, with regard to possible nutritional benefits, pure CCP extrudates could be used as DF-rich food ingredients for fiber enrichment. Furthermore, it was of scientific interest to validate our previous data on the effects of thermal and mechanical stress investigated using a closed cavity rheometer [20].

Changes in Monosaccharide Composition
Extrusion processing causes only minor modifications in the monosaccharide composition of the IDF fraction of CPP ( Figure 2A). As shown earlier [20], the monosaccharide composition of IDF from CPP consists mostly of glucose, the majority originating from cellulose and xylose, mostly originating from hemicelluloses. Less dominant monomers such as arabinose, galactose, and galacturonic acid originate from insoluble pectic polysaccharides. Only the portions of monosaccharides from insoluble pectins decrease slightly from 26.1 mol% to 18.1-19.0 mol% in total during extrusion irrespective of c w and SME applied. The monosaccharide com-position of HMW-SDF from CPP shows more distinct modifications after extrusion processing ( Figure 2B). Arabinose, galactose, and rhamnose portions that can be assigned to rhamnogalacturonan type I with its specific sidechains (arabinans and galactans, see also Section 3.2.3) rise markedly from 39.2 mol% to 53.7-54.1 mol% in total. Again, these effects seem to be independent of SME and c w . Overall, due to thermomechanical processing, ratios of pectic sidechain monomers decrease in the IDF, whereas they increase in the HMW-SDF, indicating a solubilization of pectins rather than a degradation of pectic polysaccharides from IDF. However, both processes, solubilization of polysaccharides and degradation of specific structural elements of these polysaccharides, are likely to occur to some extent concurrently. Similar findings were obtained by Hwang et al. [39] who described that extrusion processing of apple pomace reduced arabinose and galactose ratios in insoluble polysaccharide fractions. The authors assumed that extrusion cleaves preferentially arabinogalactan sidechains from pectins [39]. Unfortunately, no additional data on the neutral sugar composition of soluble polysaccharide fractions or polysaccharide linkage types of extrusion processed apple pomace polysaccharides were presented.

Arabinan Structural Details
Our prior work [20] showed that chokeberry pomace DF arabinans are largely based on the structural elements A-2a, A-5b, and A-4a oligosaccharides, as shown in Table 4 and Figure 3. These structures represent linear and O3-branched arabinose units, whereas A-4b and A-5c, the latter exclusively in IDF, contain O2-branched arabinose units. A-6a and A-7b demonstrate more highly branched arabinans structures [33]. Extrusion processing does not result in major changes in HMW-SDF. The IDF arabinan motives show some minor modifications. The A-5c and A-4b motives are no longer detectable after extrusion processing almost irrespective of SME applied. However, in general, data from the arabinan profiling approach indicate that pectic polymer solubilization and modifications are not a consequence of overall specific structural changes with the exception, that O2 branches appear to be slightly less stable than O3 branches. Whether or not this hypothesis is correct needs to be clarified in future studies. Table 3. Partially methylated alditol acetates (PMAA) from insoluble dietary fiber (IDF) and high molecular weight soluble dietary fiber (HMW-SDF) fractions isolated from chokeberry pomace powder (CPP) before and after extrusion processing 1 (mol%, range/2, n = 2).

Sugar Contents
Increasing the SME (A1 < B1 < D2 < D1) reduces initial glucose and fructose contents of CPP by about 15-25% and about 13-27%, respectively (Table 5). This indicates the possible formation of Maillard glycation products, as previously suggested [42]. Sensory properties of extruded products are based to some extend on these compounds [43,44]. Some authors also ascribed beneficial health effects of Maillard products [45]. Sucrose contents remain unchanged during extrusion processing. The data presented are in accordance with data on thermal and mechanical treatment in a closed cavity rheometer, as published earlier [20]. Table 5. Glucose, fructose, and sucrose contents of chokeberry pomace powder (CPP) before and after extrusion processing (g/100 g dm, mean ± SD, n = 2) at water contents of 13% (A1, B1, and D1) and 23% (D2), respectively. SME, specific mechanical energy (Whkg −1 ); T M , material temperature ( • C).
Means with different superscript capital letters within the same column differ significantly (p < 0.05).
Finally, extrusion processing reduces the content of polyphenols (sum of monomeric anthocyanins, phenolic acids, and flavonols, TPP-HPLC) to about 31% to 48% at the different SME applied, whereas the measured total polyphenols (TPP) content, in contrast, is enhanced to 113-133% (Table 8). Table 8. Polyphenols contents of chokeberry pomace powder (CPP) and samples after extrusion processing (g/100 g dm, mean ± SD, n = 2) at water contents of 13% (A1, B1, and D1) and 23% (D2), respectively. SME, specific mechanical energy (Whkg −1 ); T M , material temperature ( • C); TPP, total polyphenols contents analyzed by the Folin-Ciocalteu test; TPP-HPLC, total polyphenols contents calculated as the sum of monomeric anthocyanins, phenolic acids, and flavonols. It has to be kept in mind that the Folin-Ciocalteu assay is not specific only for phenolic compounds [34], but other reducing compounds present in the sample, for example, reducing compounds resulting from the Maillard reaction are recorded at the same time [8,49]. Moreover, increased TPP contents may also indicate a higher extractability of proanthocyanidins, which have been described as the main class of TPP in chokeberry pomace [17,23]. Oligo-and polymeric proanthocyanidins have been shown to adsorb to apple cell-wall polysaccharides via hydrogen bonds with a high affinity in particular to pectic polysaccharides, in particular arabinans, and high T M during thermomechanical treatment may enhance their extractability [52,53]. Overall, these findings confirm the results of our previous study on the stability of polyphenols when applying defined thermal and mechanical stress in a closed cavity rheometer [20]. Combined with known health-related properties of chokeberry polyphenols [22], these data suggest that chokeberry pomace powder is a valuable source of polyphenols for the fortification of extruded food.

Techno-Functional and Sensory Relevant Physical Properties
The effect of extrusion processing on techno-functional (water solubility and water absorption) and physical properties related to the sensory characteristics (sectional and longitudinal expansion, hardness, and color) is shown in Table 9.
values can be used to describe the extent to which the fiber structure (on a macroscopic and/or molecular level) was modified by extrusion processing. The WAI of untreated CPP (2.92 ± 0.08) is comparable to blueberry pomace (2.66 ± 0.05) but lower than apple pomace (5.60 ± 0.08) and higher than starch (1.15 ± 0.09) [9]. Through extrusion, the WAI is reduced to about 2 without an impact of n and c w applied (Table 9). No linear correlation between WAI and SME or T M were found. The WSI of CPP extruded at SME 145 ± 6 Whkg −1 and c w 13% (A1) is in the range of the untreated CPP and increases almost linearly (r = 0.88, p = 0.004) with increasing SME by about 4% at the highest SME (D1) ( Table 9). There is very little literature dealing with the extrusion of unblended pomace in a twin-screw extruder. For the extrusion of apple pomace, which is also a member of the Rosaceae plant family, similar results have been reported. The WSI value of raw apple pomace was in the same range as CPP and increased with increasing thermomechanical treatment (SME, T M ). However, the increase was more pronounced than for chokeberry [54]. The increase in the WSI is in accordance with increased HMW-SDF contents detected (Table 2). However, it has to be kept in mind that sugar contents are reduced during processing (Table 5). Enhanced c w during processing (D2, 23%) at SME 190 Whkg −1 slightly reduces the WSI.

Expansion
The expansion of extruded cereal products is described by sectional and longitudinal expansion indices, SEI and LEI, respectively. SEI values between 0.85 and 1.95 are measured for all CPP extrusion trials. A slight reduction is observed with increasing n, SME (r = −0.83, p = 0.010) and T M (r = −0.74, p = 0.034). Mostly, additional water increases SEI, except for A2. In contrast, SEI values between 1 and 15 are to be expected for pure starch [55], whereas SEI values of starch-based materials typically range between 10 and 20 [56]. The LEI increases with n as expected. Furthermore, an almost linear correlation to SME (r = 0.92, p = 0.001) was observed. However, T M (r = 0.80, p = 0.016) was less correlated with the LEI values. Increasing c w resulted in slightly lower LEI. As described above (Figure 2), CPP composition and structure change slightly with varying process parameter. Therefore, no large changes in expansion based on structural changes were expected and the slight changes in expansion observed are explained by the influence of process parameters.

Hardness
The extruded samples are similar in hardness (Table 9). There is only a slight reduction in hardness with increasing n from 200 to 800 min −1 observed for both water contents (A1, B1, C1, D1 and A2, B2, C2, D2). Briefly, the higher the SME and T M , the softer the extrudates ((SME) r = −0.81, p = 0.014 and (T M ) r = −0.82, p = 0.012). In general, CPP extrudates are much softer than common RTE cereals consisting of starch-based materials with a typical hardness of about 25 Nm −2 [57]. Due to comparable hardness and longitudinal, as well as sectional expansion, no change in porosity is observed ( Figure 5).

Conclusions
No significant changes in total DF contents occur during extrusion processing of pure CPP in a co-rotating twin-screw extruder applying SME in the range of 145-222 Whk −1 , typical for the production of RTE cereals. Although the composition of the individual DF fractions (IDF and HMW-SDF) is modified during extrusion, we assume that these modifications are mostly based on solubilization of the pectic rhamnogalacturonan type I polymers; degradation of specific structural elements of cell wall polysaccharides seems to be less pronounced. Sugar contents are reduced with increasing SME. Anthocyanin contents decrease linearly with SME, whereas contents of phenolic acids and flavonols remain unaffected. On the basis of these data, after extrusion processing CPP remains a valuable source of both, DF and PP, and has the potential to partially substitute starch in RTE cereals, for example, crispy breakfast cereals and snacks. As expected, the technofunctional and sensory relevant physical properties of pure CPP extrudates, analyzed and presented here, are different as compared with common starch-based RTE cereals. However, due to a moderate crispness and hardness, as well as an appealing color, together with a surprisingly overall appealing visual impression and pleasant taste (as evaluated by the working group and lab colleagues), granulated CPP extrudates may potentially be used as a food ingredient for DF and PP enrichment after further sensory analysis confirmed the assumption.