Chestnut flour addition in commercial gluten-free bread: A shelf-life

Study

1. Introduction
Celiac disease (CD), an immune-mediated enteropathy caused by the ingestion of gluten in
genetically susceptible individuals, is one of the most common lifelong disorders. At present,
the only available treatment for CD is a strict gluten free diet. The estimated prevalence of
this disease is about 1% of the general population, and it affects persons of any age, race, and
ethnic group (Fasano &Catassi, 2012). Recently, the market of gluten-free foods grew not
only for the increment of CD incidence; consumers consider glutenfree foods appealing
because they perceive these products as healthy, although no scientific evidences are still
published about (Capriles &Ar^eas, 2014). Among gluten-free foods, bread is the most
important. The development and/or improvement of glutenfree bread appear as a big
challenge of food technology in view of the unique role of gluten in yeast-leavened baked
goods and in bread-making process. The absence of gluten is well known to show a great
influence on dough rheology also leading to bread with crumbling texture, poor colour, not
satisfying taste and low specific volume (Houben H€ochst€otter, &Becker, 2012). Bread is
alsoperishable; its integrity begins to deteriorate immediately after baking due to the chemical
and physical changes that occur during the well-known staling process (Gray &Bemiller,
2003). Glutenfree breads are reported to show a short shelf-life, probably as a consequence of
the lack of the viscoelastic network formed by gluten that is responsible for slowing down the
movement of water (Houben, H€ochst€otter, &Becker, 2012). In the last years, different
challenges (e.g. different gluten-free flours and starches, new additives, novel technologies)
have been developed to overcome these problems, as reported in two interesting reviews
recently published (Segura &Rosell, 2015 and references therein cited; Capriles &Ar^eas,
2014 and references therein cited). In this context, the point currently most debated in
literature is the improvement of the nutritional value of gluten-free breads by adding
ingredients with a high nutritional value, as the gluten-free dietary pattern is often
characterised by an excessive consumption of fats and reduced intake of complex
carbohydrates, dietary fibre, vitamins and minerals (Pellegrini &Agostoni, 2015; Segura
&Rosell, 2011).
The nutritional quality of gluten-free breads could be improved by incorporating nutrient-
dense alternative flours and/or ingredientswith the nutritional purpose of increasing fibres'
content above all but also nutrients and phytochemicals such as g fruit- and vegetable-based
ingredients (Capriles &Ar^eas, 2014). Fitting in with this outlook, in the last years chestnut
flour received more and more attention due to its nutritional and health benefits both onwheat
and gluten-free bread. Chestnut flour contains high qualityproteins with essential amino acids
(4e7 g/100 g), dietary fibre (4e10 g/100 g), low amount of fat (2e4 g/100 g) and also vitamin
E, vitamin B group, potassium, phosphorous, and magnesium (Sacchetti, Pinnavaia,
Guidolin, &Dalla Rosa, 2004). Dall’Asta et al. (2013) reported that wheat breads enriched
with the addition of chestnut flour presented an increased quality from both organoleptic
(more complex flavour, darker colour and more heterogeneous crumb) and nutritional (higher
antioxidant capacity and fibre content) points of view. In addition, chestnut flour added
breads showed a delay in the staling process, confirming the feasibility of producing bread
with improved nutritional and qualitative characteristics, not only just after baking but also
during shelf-life (Rinaldi, Paciulli, Dall'Asta, Cirlini, &Chiavaro, 2015). Regarding GF
breads, Demirkesen, Mert, Sumnu, and Sahin (2010) studied the effects of different levels of
addition on a rice-based gluten-free formulation reporting that elevated amounts of chestnut
flour led to some deterioration in quality parameters. The same authors published several
papers on chestnut flour addition in GF breads prepared on lab scale and mainly about the
effects of cooking technique (Demirkesen, Sumnu, &Sahin, 2013a and 2013b). The use of
commercial GF bread formulation was not considered. However, outcomes of these studies
suggested that the replacement of rice flour with chestnut flour represents a promising way to
enhance nutritional values of GF breads and, very interesting, to potentially retard staling of
these kinds of bread (Demirkesen, Campanella, Sumnu, Sahin, &Hamaker, 2014). Thus, the
aim of the present work was to evaluate the effects of chestnut flour addition on technological
and nutritional properties of two common commercial GF bread formulations with complex
recipes and already optimized from food industry for bread-making performances.
2. Materials and methods
2.1. Samples
Two common commercial (Biaglut, Latina, Italy) gluten free bread mixtures were used and
ingredients, as reported on labels, were as follows: corn starch (43.5 g/100 g), potato starch
(40.0 g/ 100 g), skimmed milk (6.5 g/100 g), destrose (4.5 g/100 g), cellulose (2.0 g/100 g),
guar gum (1.8 g/100 g), hydroxypropylmethylcellulose (1.7 g/100 g) for the first mixture
(named F1gf) and corn starch (43.5 g/100 g), rice flour (40.0 g/100 g), lupine proteins (6.5
g/100 g), destrose (4.5 g/100 g), hydroxypropylmethylcellulose (2.0 g/100 g), vegetable fiber
(2.0 g/100 g), salt (1.5 g/100 g) for the second (named F2gf), The proximate composition
(g/100 g d.m.) of the two mixtures were: F1gf: carbohydrate 85.2., fibers 6.0, protein 4.4, fat
0.2; F2gf: carbohydrates 84.3, fibers 7.1, protein 4.4, fat 0.8. A chestnut flour (C) was
obtained as previously reported (Dall’Asta et al., 2013) from four cultivars (Ampollana,
Perticaccia, Leccardina and Gursona) from the Ceno Valley (Parma, Italy) and so used for the
enrichment It showedavailable carbohydrates, protein, fat and dietary fiber contents of77.4
g/100 g d.m., 6.0 g/100 g d.m., 4.6 g/100 g d.m., and 12.0 g/ 100 g d.m., respectively. The
commercial mixtures were used to prepare two control breads, coded as M1 from F1gf, and
M2 from F2gf, respectively. Further two breads were prepared with a ratio of 200 g/kg
C/F1gf (sample named as M1C) and of 100 g/kg C/F2gf(sample named as M2C),
respectively. These chestnut flour ratioswere selected based on previous results (Rinaldi et
al., 2015) and preliminary experimentations.
2.2. Bread-making and storage
Breads were prepared with the following formulation on flour basis: M1 and M1C: flour (1
kg), water (880 g), sunflower oil (50 g), yeast (50 g) and salt (20 g); M2 and M2C: flour (1
kg), water (900 g), sunflower oil (50 g), yeast (50 g) and salt (20 g). A domestic bread maker
machine (Moulinex, Groupe Seb Italia S.p.A., Milano, Italy) was used for breadmaking, with
the rapid program: stirring þ kneading, þ rising, 40 min; baking, 45 min at 210 _C. The
breads were then cooled at room temperature. The loaves were packaged in alcohol-sprayed
sealed air-tight plastic bags and stored in a 25 _C temperature-controlled chamber in the dark
(ISCO 9000, Milan, Italy). Samples were analysed at 0, 1, and 3days post-production. Three
loaves were used for the characterization of breads at each storage time for a total of 9 loaves
for eachbread type. 2.3. Chemical composition Protein content was determined by the
Kjeldhal method using 1 g of ground sample and a correction factor of 5.7, as previously
reported (Dall’Asta et al., 2013). Fat content was determined utilizing a Soxhlet extractor
(Velp Scientifica, Monza-Brianza, Italy) on 5 g of ground sample and diethyl ether as
solvent. The moisture content (%) within the bread loaves was evaluated following the
AACC standard method, 44-15.02 (AACC, 2000). The crust, undercrust layer, and central
crumb were examined at each shelf-life time for each bread type. All bread samples were also
analysed for starch (AOAC method 996.11) and total, insoluble, soluble dietary fiber (AOAC
method 2009.01 and 2011.25). Three loaves of each bread type were analysed at day 0 of
storage.
2.4. Crumb grain characteristic and specific bulk volume
Crumb grain was evaluated by means of a digital image analysis system, as reported
previously (Dall’Asta et al., 2013). The number of pores (expressed as percentage of the total
number) was obtained according to pre-selected dimensional classes based on their area:
class-1: 0.01e0.049 mm2; class-2: 0.05e0.99 mm2; class-3: 1e4.99 mm2; class-4: >5 mm2 .
Crust thickness (mm) was measured by means of the size function of the same image analysis
software. Specific bulk volume of breadswas determined according to the AACC Approved
Method 10-05.01 (AACC, 2000) andexpressed as the volume/weight ratio of cooked bread
(cm3/g).
2.5. Physical analysis
Texture analysis was performed on crust and crumb using aTA.XT2 Texture Analyzer
equipped with a 25 kg load cell (StableMicro Systems, Godalming, UK). Crust hardness was
measured on five preselected points by means of a puncture test using a 3 mm diameter
stainless steel probe and a test speed of 2 mm/s. Maximum peak force (N) from the
penetration curve was taken as crust hardness. Crumb evaluation was carried out on ten cube
of 20 _ 20 _ 20 mm extracted from two central slices. A TPA test wasperformed with a 35
mm diameter cylindrical aluminium probe by means of a double compression with a speed of
1 mm/s up to the 50% of the original sample height. The textural parameters considered were
hardness (maximum peak force of the first compression cycle, N), cohesiveness (ratio of
positive force area during the second compression to that during the first compression area,
dimensionless), resilience (area during the withdrawal of the penetration, divided by the area
of the first penetration, dimensionless), and chewiness (product of hardness x cohesiveness x
springiness, N) (Bourne, 1978). Colour was determined on ten preselected locations of the
crust and crumb of each bread loaf by means of a Minolta Colorimeter (CM 2600d, Minolta
Co., Osaka, Japan) equipped with a standard illuminant D65 and a 10_ position of the
standard observer. L* (lightness), a* (redness) and b* (yellowness) were quantified on each
sample using the Spectramagic software (Ver. 3.6). The individual colour differences of each
breadat 1 and 3 days of shelf-life were evaluated with respect to thecolour at time 0 using DE
calculation.
2.6. Nutritional parameters
The antioxidant capacitywas determined by means of DPPH test on flours and breads
(Dall’Asta et al., 2013). The radical scavenging activity was calculated as follows: I% ¼
[(Abs0 e Abs1)/Abs0] x100, where Abs0 was the absorbance of the blank and Abs1 was
theabsorbance of the sample. The antioxidant capacity value (mmol/L Trolox/g) of samples
was obtained from the calibration curve calculated measuring the absorbance at 517 nm of
Trolox methanolic solutions at different concentrations. Breads were tested in vitro to
determine the rate of starch hydrolysis; in vitro digestions were performed following the
method described by Zaupa et al. (2014). The in vitro gastrointestinal model was set up as
follows: about 8 g of sample was suspended in phosphate buffer (20 mmol/L) and incubated
at 37 _C stepwise with human saliva, for 2 min at pH 6.9, and porcine pepsin (2500 U), for 2
h at pH 2.0e2.5.
Each sample was then transferred into 20 cm dialysis tubing strips (12 000 Da molecular
weight cut off) with 100 mg of pancreatin from porcine pancreas (3xUSP), sealed with plastic
clamps, and incubated for 5 h at pH 6.9 into 1000 mL sealed containers containing 500 mL of
phosphate buffer. Two aliquots (0.5 mL) from the dialyzed solution were removed for
analysis at time 0, every 15 min during the first hour and every 30 min until 5 h digestion.
The aliquots were used to determine the number of glucose monomers of the permeated
fragments. To this purpose, each aliquot sample was hydrolyzed using 20 ml of 0.5 g/100 g
amyloglucosidase solution(200 U) at pH 5.6 and the glucose concentration was
determinedwith a glucose analyzer (2900 Biochemisty Analyzer, YSI Inc., Yellow Springs,
USA). All analyses were performed in triplicate. 2.7. Statistical analysis SPSS (Version 22.0,
SPSS Inc., Chicago, USA) statistical software was used to identify differences during shelf-
life through one-way analysis of variance (ANOVA) and least significant difference (LSD)
test at a 95% confidence level (p <0.05). Comparisons between breads made with the same
gluten-free mixture, with or without chestnut flour at each storage time were performed by
means of a Student t-test at a 95% and 99% confidence level (p <0.05 and p <0.01).
3. Results and discussion
3.1. Specific bulk volume and crumb grain characteristics
Bulk volumes of control gluten-free breads were 3.6 ± 0.2 and 5.1 ± 0.3 cm3 l_1 for M1 and
M2, respectively. The presence of lupine proteins in M2 probably contributed to a
significantly higher volume compared to M1, as previously reported by Ziobro, Witczak,
Juszczak, and Korus (2013). M2 recipe presented also little higher amount of water but no
significant differences were observed between breads moisture content at time 0 (below
discussed). Thus, the difference in the recipe could affect the volume but cooking process
removed the difference. M1C and M2C exhibited significant lower volume values (3.1 and
2.6 cm3 g_1, respectively) due to the addition of chestnut flour, as previously observed by
Demirkesenet al. (2010). These authors attributed the low volumes in gluten free breads
prepared with chestnut flour to the rigid and compact structure of the fibrous chestnut flour
dough. Collar, Santos, and Rosell (2007) affirmed that the fiber content could reduce the
expansion of the gas cells also in gluten free products. In addition, the relatively high sugar
content of the chestnut flour was found to hinder or reduce the starch gelatinization during
baking leading to low specific volume (Demirkesen et al., 2010). Bulk volumes did not
significantly change during shelf-life for all breads. Crust thickness was not significant
different between M1 and M1C (1.5 ± 0.1 vs. 1.8 ± 0.1 mm) breads. On the other hand, M2C
crust resulted significantly thicker than M2 showing an almost twofold thickness (2.0 ± 0.1 vs
1.1 ± 0.1 mm). Crumb grain characteristic of all samples are reported in Fig. 1A (M1 and
M1C) and 1B (M2 and M2C). At time 0, significant differences were observed for M1 and
M1C breads: the addition of chestnut flour (M1C) caused a significant increase of the pores
of the greatest dimensions (class 3 and 4) in comparison with M1 breads that showed a
significantly higher number of little holes (class 1 and 2). In addition, the presence of a higher
amount of cells of large sizes in M1C than M1 could explain the observed significantly
lower development of the final bread, due to the collapse of the internal structure after
baking. Similarly Mariotti, Pagani, and Lucisano (2013) found a significantly higher
alveolate area to total area ratio on two gluten-free mixtures enriched with buckwheat in
association to the lower development of the finalbread (loaf collapse) and the lower softness
of the crumb. It ispossible that during the early stage of baking, the crumb pores undergo a
rapid expansion, as a result of both increased CO2 production and steam formation: a failure
in the matrix increases the tendency for pores to coalesce and possibly collapse. Also
Dall’Asta et al. (2013) found a higher number of pores belonging to the major dimensional
classes in soft wheat breads enriched with the same level of chestnut flour. On the other hand
this finding is in disagree with data presented by Demirkesen et al. (2013) who reported a
decrease of the gas retention capacity of dough and a decrease in expansion of the gas cells
due to chestnut fibres. The low percentage of chestnut flour added (20 g/100 g), compared to
the 30 g/100 g suggested by Demirkesen et al. (2010), may have probably limited the effect
of the fibre on gas expansion. During storage, both M1 and M1C presented similar changes.
Pores belonging to class 2 significantly increased while pores of class 1 significantly
decreased and no variations were observed for the remaining classes, in agreement with
Rinaldi et al. (2015). Thus, the addition of chestnut flour to M1 mix did not affect crumb
grain characteristic changes during shelf-life. M2 and M2C breads (Fig. 1B) significant
differed at time 0 days, too, but in this case the addition of chestnut flour caused a significant
increase only of the pores of class 3 (1.0e5.0 mm2). During the three days of storage, M2C
samples exhibited a similar behaviour to that observed for M1 and M1C, while M2 did not
showany significant difference for class 1 and 2 and a decrease in class 3. Probably the
highest percentage oflupine proteins of M2 than M2C, which acted as stabilizer, emulsifier,
water and lipid binder, was enough to prevent or retard the water migration and cell walls
width reduction, as consequence, inaccordance with Kohajdov_a, Karovicova, and Schmidt
(2011).
3.2. Moisture content

Moisture content profiles of all breads are reported in Fig. 2AeD. At day 0, no significant
differences were observed between the samples not enriched and the chestnut enriched breads
in all the portion of the loaves. The only exception was the crust of M2C that resulted
significantly wetter than the control. During the shelf life, all breads showed no significant
differences in crumb moisture content with the exception of M1 sample, which presented a
significant decrease at day 3. Regarding near-crust data, M1 showed a decrease (p <0.05)
from day 0e1 and a subsequent increase at the end of storage. This significant decrease was
probably due to the water migration to the crust that increased its moisture content, while the
increase at day 3 was due to the water absorbed from the crumb (Fig. 2A) even if no
differences were observed in this value during storage probably because of the height of the
slices that masked the slow water migration. The trend of M1C near-crust did not present this
behaviour; only a significant increase (p <0.05) was observed at day 1 and no further changes
until day 3. On the contrary, M2 and M2C maintained constant near-crust moisturecontents
during shelf life (Fig. 2C and D). Crust of all samplesshowed an increase in moisture content,
as expected. M1C showed a faster migration from crumb to near-crust and crust (Fig. 2B)
compared to M1 probably due to the coarser crumb grain and to the presence of a higher
amount of cells of large sizes that allowed water migrating easier. M2 crust instead didn't
show significant changes at day 1 with an increase (p <0.05) at day 3, while M2C
continuously increased moisture content at each day of analysis. Finally, the addition of
chestnut flour did not significantly change moisture migration of gluten-free breads during
storage probably because of the low amount of addition, which scarcely affected the effect of
fibres on water dynamics.

3.3. Texture analysis

Crust hardness of all samples is shown in Fig. 3.M2 presented an higher crust hardness than
M1, due to the lower water content and the higher crust thickness (Fig.2). The addition of
chestnut flour did not cause significant changes at time 0 days both for M1 and M2, in
relation with the similar crust thickness. During the shelf-life, the two enriched mixes showed
an opposite trend; the crust of M1C exhibited a significant decrease of hardness at time 1 and
an increase at time 3 with values significantly lower compared to M1, which did not show
any change. On the other hand, breads prepared with F2gf formulation and without any
enrichment exhibited a significant decrease of crust hardness already at time 1 of storage.
M2C crust became softer only at the end of shelf-life instead, because of the thicker crust that
could have retarded the effect of thewater migration. Changes in crust texture and moisture
content were not strictly correlated during shelf life, as already observed in a previous study
(Rinaldi et al., 2015). Data obtained from TPA test are reported in Table 1. Generally, for
both GF mixes the addition of chestnut flour caused a highly significant increase (p <0.01) in
crumb hardness with a greater extent for M2. The observed increase in crumb hardness was
probably due to fibre content brought by chestnut flour and to the significantly lower volume
observed for M1C and M2C breads in accordance with Demirkesen et al. (2010). During
storage, both breads added with chestnut flour presented a higher staling rate, expressed as
percentage increase of hardness, at time 1 day compared to GF mixes: 85.9 and 63.6% crumb
hardness increasing were observed for M1C and M2C, while 62.5 and 31.4% were obtained
for M1 and M2, respectively. This observation is in disagreement with Demirkesen et al.
(2014) but in that case the authors reported a significant increase in specific bulk volume,
conversely to the present paper. On the contrary, stalingsamples: thus, the addition of
chestnut flour seemed to be not able to retard staling that, in gluten-free breads which lack
gluten and have low volume, is faster than in standard breads (Demirkesen et al., 2014).
Cohesiveness was an indicator of the internal cohesion of the material: generally, breads with
low cohesiveness are susceptible to fracture and crumble (Onyango, Mutungi, Unbehend,
&Lindhuaer, 2010) and are not desirable. No significant differenceswere observed for M1
and M1C breads just after baking for thisparameter but M1C presented significantly lower (p
<0.01) values than M1 at both times of storage indicating a faster staling and a product with
lower quality. Probably, fibres of chestnut flour were not able to prevent staling and tendency
to crumble, as previously reported by Sabanis, Lebesi, and Tzia (2009) in gluten-free breads
due to their water binding capacity and to the possible hydrogen bonding with starch.
Conversely, cohesiveness values of M2C were significantly lower than M2 at every day of
analysis. It could be hypothesized that the interaction between chestnut starch and lupine
proteins negatively modified the cohesiveness of bread, because of the faster staling,
counteracting the positive effect of lupine protein in gluten-free breads (Segura &Rosell,
2012). Resilience values showed the same trend in accordance with cohesiveness, as it has
been reported that the reduction in resilience characterizes loss of elasticity and tendency to
crumble (Onyango et al., 2010). Both M1 and M2 samples did not show any differences with
enriched breads at time 0 but chestnut added samples exhibited significantly lower values (p
<0.01) during shelf-life. Finally, chewiness values of the two mixes (M1 and M2) were very
similar at all times of analysis but the effects of chestnutaddition were different: M1C breads
presented significantly higher chewiness than M1 only at day 0, while M2C samples
exhibited higher (p <0.01) values at all days of storage. Chewiness gives an indication of the
energy required to masticate a solid food and therefore the time required masticating a bread
piece prior to swallow resulted the highest for M2C samples, due to the interaction between
F2gf ingredients and the chestnut flour. On the contrary, low chewiness values, as observed
for both M1 and M2, mean easy break of the bread in the mouth (Hager et al., 2012).
3.4. Colour analysis
Crust colorimetric values of tested breads are reported in Table 2. M1 crust colour resulted
darker (more red and more yellow) compared to M2 probably due to the presence of
skimmed milk in the first formulation, as previously reported by Gallagher, Gormley, and
Arendt (2003) for gluten free breads. The addition of chestnut flour caused a highly
significantly (p <0.01) darkening effect on both M1C and M2C bread crust, in accordance
with previous observations of other papers (Dall'Asta et al., 2013): higher values of a* and b*
parameters and lower of L*, as consequence (Table 2), were observed in M1C compared to
M1. Darkening of the crust was reported to be desirable in gluten-free breads, as theytend to
have a lighter crust colour that sometimes appear artificial (Gallagher et al., 2003). During
shelf-life, only M1C crust showed significant variations for L* and b* values: the former
presented a significant decrease while the latter showed an increase thus resulting in a
globally darker crust probably due towater migration. Crumb colour data were also shown in
Table 2.M1 crumb presented higher L* and lower a* and b* values compared to M2. This
was probably related to the different ingredients present in the glutenfreemix formulations:
M1 crumb was lighter because of potatostarch and presented slightly negative a* values due
to the high presence of corn starch, as previously observed by Segura and Rosell (2013). In
addition, lupine proteins, present only in M2 mix, have been reported to give the bread a
more yellow colour (higher b* values), as consequence (Kohajdov_a et al., 2011). The
chestnut flour addition in both mixtures caused a highly significant decrease (p <0.01) of L*
values and an increase of a* and b* (Table 2), as expected, with a great extent for M1C due
to the higher percentage added. During shelf life M1 crumb retained better the colour than the
M2 ones with stable L* and b* decreasing only at day 3, while M2 presented L* increasing
and b* decreasing since the day 1, a* decreasing was instead observed from the day 1 for
both the loaves. The shift to paler tones was also observed by Rinaldi et al. (2015) during a
short shelf life of wheat breads, in relation to starch crystallization. Adding chestnut flour to
the formulations resulted in not relevant colour changes for M1C during shelf life, while for
M2C other than stable L*, faster b* and slower a* decreasing were observed. The higher
content of chestnut flour added in M1C formulation, better preserved the bread discoloration
during shelf life. Finally, DE values ranged from 1.4 to 2.4 (Table 2). M2 samples presented
significantly greater DE values at day 1 in comparison to M2C ones, while no significant
differences were instead recorded between M1 and M1C. In general, similar behaviours were
observed among the studied breads.
3.5. Nutritional parameters
Breads were analysed for the composition in macroconstituents; data, on dry matter basis, are
reported in Table 3. Concerning the lipid fraction, the addition of chestnut flour caused a
slight but significant increase in fat for M1C breads, and protein for M2C in comparison to
control samples. On the contrary, the addition of chestnut flours did not significantly affect
the starch content for both gluten free mixtures. Concerning the total fibre content, the
amount found in M1C is significantly higher than M1 (Table 3) and differences are related to
both the soluble and insoluble fractions. Conversely, chestnut flour addition to the
formulation M2 in percentages of 10 g/100 g did not affect significantly the fibre content. In
general, the fibre content of the four considered breads is good, and 20 g/100 g of chestnut
flour addition demonstrated to be sufficient to improve the nutritional balance of gluten-free
breads (Foschia, Peressini, Sensidoni, &Brennan, 2013; European commission, 2006). As
chestnut flours is an important source of antioxidant compounds, mainly containing ellagic
acid (Dall’Asta et al., 2013; Morrone et al., 2015), its addition to the GF mixture used for
bread formulation may enhance the antioxidant capacity and the stability along the shelf-life.
Furthermore, the baking step significantly affected the content of antioxidant compounds in
the final products, by allowing thermal degradation and chemical browning due to the
Maillard reaction (Michalska, Amigo- Benavent, Zielinski, &del Catillo, 2008). As chestnut
flour is naturallyrich in reducing sugars (Morrone et al., 2015), this may boostthe formation
of Maillard-type antioxidant products. As reported in Fig. 4, both MC1 and MC2 presented a
significantly higher antioxidant content compared to M1 and M2. Over the observation time,
M1 and M2 showed a comparable trend, indicating that the flour mixtures used for
formulation were similar in terms of antioxidant capacity. On the other hand, the TEAC value
for M1C was significantly higher than MC2 at t0 (p <0.000), on account of the higher content
of chestnut flour in the formulation. This is reflected also in the behaviour recorded over the
observation time: while MC1 seemed to be rather stable along the shelf-life, the antioxidant
capacity in MC2 decreased with a comparable trend as monitored for M2, although at higher
absolute values. The data obtained within this study confirm thus that the addition of chestnut
flours to the formulation of GF products may increase both the nutritional value in terms of
antioxidant load and the stability along the shelflife towards oxidative degradation. To better
characterize the nutritional properties of the breads prepared within this study, the starch
digestibilitywasassessed over 3 h by enzymatical hydrolysis. Curves expressing the
percentage of hydrolysed starch permeated through the dialysis tube during a 3 h in vitro
digestion are reported in Fig. 5. Histograms represent the total areas under the curves (AUC)
of digested starch over 3 h for each bread. After 3h hydrolysis, the digested starch fraction of
the total starch was: 70.2 ± 10.1% for M1, 72.2 ± 1.3% for M1C, 66.0 ± 10.4% for M2, 73.6
± 3.0% for M2C; the AUC values were 7857.4 ± 557.9, 7393.5 ± 532.3, 8158.8 ± 611.9,
8044.0 ± 555.0 (mg min dL_1), respectively. The differences in digested starch fraction
among breads were not significantly different. This result was expected since nutritional
composition was similar among breads (Table 3); in fact, Segura and Rosell (2011) did not
observe differences in total starch hydrolysis fraction, among gluten-free breads, despite the
wide range in macronutrient composition.

4. Conclusions

This study revealed that the addition of chestnut flour to two common gluten free mixtures
influenced some of the characteristics of breads just after baking and during storage. The
different ingredients of the gluten free mixtures led from the beginning to dissimilar loaves:
M1 showed lower volume, harder and lighter crumb, softer and darker crust than M2. The
addition of chestnut flour led to volume reduction, different crumb grain characteristic with
larger alveoli and hardened the crumb texture just after baking and during the shelf life for all
the breads, particularly forM2 breads. The crust hardness appeared to be not significantly
influenced by the chestnut addition in both breads. However, breads obtained with the two
mixtures showed different moisture dynamics, independently from the chestnut flour
enrichment. Improved colour was observed after the addition of the chestnut flour, resulting
darker and more stable along the shelf-life. The improved antioxidant activity measured in
both the chestnutenriched breads was retained during the entire shelf-life, particularlyfor
M1C. In addition, the enrichment with 20 g/100 g of chestnut flour (M1C) caused a
significant increase in soluble, insoluble and total fibre content; on the contrary, 10 g/100 g of
chestnut flour (M2C) didn't lead to significant differences. No differences in starch
digestibility were observed in all the samples. In conclusion, the addition of chestnut flour to
commercial glutenfree mixtures appeared to be promising for commercial mixture F1gf
representing an interesting starting point for further studies for improving the with the aim to
reduce the defects in the final products.