Effect of Fertilization Regime of Common Wheat (Triticum aestivum) on Flour Quality and Shelf-Life of PDO Tuscan Bread

The shelf-life of bread is influenced by flour components, such as starch, composed of amylose and amylopectin. The aim was to test the effect of different balances of N (45, 90, 135 kg/ha) and P (48, 96 kg/ha) fertilizers on the flour characteristics and consequently the shelf-life of PDO Tuscan bread, stored in different modified atmosphere packaging (Ar, N2, Air). The amylose and phytochemical compounds were increased by N and decreased by the addition of P, but excessive doses of N (135 kg/ha) had a negative effect on flour quality. In the bread, the study highlighted the tendency of N2 and Ar, as storage filler gases, to reduce water loss, slow down the staling process, and prolong shelf-life. However, the most significant influence on shelf-life was related to the different fertilizations of wheat. In fact, when N was present in equal dose to P (90/96 or 45/48 kg/ha) or slightly higher (90/48 kg/ha), the bread tended to last longer over time. Instead, when these ratios were unbalanced in favor of N (135/48 or 135/96 kg/ha) and in favor of P (45/96 kg/ha), the shelf-life decreased considerably.


Introduction
Wheat-based products are the most consumed foods all over the world, especially in the Mediterranean area [1]. In particular, the cultural tradition of breadmaking found its origins in Italy, where around 3.2 million tons of bread are produced and consumed per year [2]. The expected world population increase by 2050 and the economic globalization lay the foundations for reinvesting in bread production, looking not only to enhance quality and quantity production, but also the shelf-life of bread [3,4].
Based on the history and culture of the area, bread in Tuscany has developed into a staple food, like in many other Italian regions [5]. In 2013, Tuscan bread received the protected denomination of origin (PDO) by the Commission of the European Community [4]. The product specification requires specific flour and dough characteristics that are the sum of local varieties and environmental conditions of the origin area [5]. PDO Tuscan bread exhibits a considerably higher level of taste complexity than common commercial white bread because sourdough is used as a leavening agent [6,7], even without any salt added in the formulation. Additionally, the low pH and high levels of lactic and acetic acids in the influenced the flour features and consequently the shelf-life of the bread stored in modified atmosphere packaging (Ar, N 2 , air).

Experimental Field
Field experiments were carried out for two consecutive growing seasons from September 2018 to August 2020 under rainfed conditions in Pienza, Tuscany, Italy (42.986569 • N, 11.763888 • E, 330 m a.s.l.). The 0-0.3 m soil layer was silty clay loam (Aquic Haplustepts, fine, mixed, mesic), sub-alkaline (pH 8.1), and contained 13.8 g/kg of total organic carbon, 1.2 g/kg of total nitrogen, and 7.6 mg/kg of available phosphorus.
The treatments consisted of factorial combinations of two phosphorus (P) levels (48 and 96 kg/ha), three nitrogen (N) levels (45, 90 and 135 kg N/ha), and four common wheat varieties (Triticum aestivum) allowed in the mix for PDO Tuscan bread production, for a total of 24 treatments. The four bread varieties comprised three registered dwarf varieties (namely Panda, Bolero, and Bologna) and one old, non-dwarf landraces (namely Verna). The experiment field was arranged in a strip-plot design with three replicate blocks per year ( Figure S1). Egyptian clover (Trifolium alexandrinum, L.) was the previous crop in both growing seasons. In both seasons, the fields were plowed and then disk harrowed in late October. Then, phosphorous (triple superphosphate; P 2 O 5 : 46%) was distributed homogeneously on the soil surface and incorporated by disk harrowing at a 5 cm depth. Common wheat seeds were sown in December 2018 and 2019 with an inter row distance of 13 cm. The total dose of nitrogen was distributed in three applications: 20% at sowing by broadcasting urea (N: 46%), 50% by broadcasting ammonium nitrate (N: 26%) at tillering, and 50% by broadcasting urea (N: 46%) at stem elongation. Common wheat from each treatment was harvested separately using a plot combine-harvester equipped with Trimble GPS sensors. For each treatment, 5 kg of harvested wheat kernel were sampled for preparing the flour mix to be used for the analyses and the breadmaking trials. Weather conditions were monitored by consulting data acquired from a weather station located in the field where the experiment was being conducted. During the growing and production seasons (November-June), cumulative rainfall (mm) and cumulative growing degree days (DD) were calculated and analyzed. The cumulative degree days value was calculated daily as the difference between the average daily temperature and the base temperature considered useful for growth and development of wheat. For the characterization of the entire crop cycle, a thermal threshold of 4 • C was taken into consideration as reported by Saiyed et al. (2009) [38]. The analysis of the thermometric trend showed that compared to an average thermal of 2036 DD, the first year of tests (2018-2019) was substantially in line, while in the second year (2019-2020), there was a positive anomaly of 76 DD mainly due to a mild winter (Figure 1a). Conversely, in the maturation phase, there were more days with maximum temperatures above 28 • C in the 2018-2019 season, with June 2019 recording 18 hot days, compared to 8 days in the second year ( Figure 1b).
Rainfall during the 2018-2019 production season was 431 mm, and 515 mm in the 2019-2020 season (Figure 2a). Analyzing the distribution of rainfall, the first year showed a much higher number of rainy days than the second did ( Figure 2b). However, a monthly analysis showed a slight winter drought in the first year of trials (2018-2019) and excessive rainfall in June 2020 that affected the final stage of wheat ripening. The final stage of cereal ripening plays a key role in starch accumulation and protein translocation. In fact, very high temperatures in June are often the cause of a sudden senescence of the plant with interruption of starch synthesis and accumulation [38].

Characterizzation of Flours
According to the specifications for PDO Tuscan bread production, the flour was obtained by a mix of four varieties of common wheat divided for the six combination   A commercial mill (Industry-Combi, Waldner Biotech, Lienz, Austria) was used for the milling process at the Department of Agricultural, Food, and Environment (DAFE) at the University of Pisa.
The chemical composition and the technological features of flours were determined as previously reported [4] according to the methods accepted by the International Organization for Standardization Moreover, the flour was characterized from a phytochemical point of view (total polyphenols, total flavonoids, and anti-radical activity). In particular, an 80% methanol solution was used to perform a solid/liquid extraction (ratio 1/20 w/v) from 0.5 g of a fresh flour sample, sonicating the mixture for 30 min. All the extracts were subsequently centrifuged (15 min, 3500 rpm), filtered on a syringe filter (0.45 µm), and stored at 4 • C for the immediate analysis.
The Folin-Ciocalteu colorimetric method was applied for total polyphenols spectrophotometry as previously reported [39], expressing the results as milligrams of gallic acid equivalents (GAE) per kilogram of dry matter (dm). The total flavonoids were estimated according to the procedure reported by Bianchi et al. (2023) [40], comparing the measures to a standard curve of catechin, and the results were reported as milligrams of catechin equivalents (CE) per kilogram of dm. Using the free radical methods (FRAP [40], ABTS [41], and DPPH [40]), the anti-radical activity of the extracts was determined. According to different standard curves of Trolox (range: 0-2.0 mM for the FRAP H, 0.2-1.5 mM range for ABTS, and 0-200 µmol/L for the DPP), the results were expressed as micromoles of Trolox equivalents (TE) per gram of dm.

Breadmaking Process
The Consortium for the protection of PDO Tuscan Bread procured the sourdough used in the study. Consecutive back slopping was used for preserving the sourdough in order to maintain its acidifying and leavening properties [42]. Breadmaking was carried out from a pre-ferment leavening agent, according to the method of "biga". According to the production specification for PDO Tuscan bread, for all the experimental runs, sourdough biga was prepared by mixing a strong wheat flour type 0 (56% w/w) and sterile water (33% w/w), and (11% w/w) then leaving to ferment for 18 h at 20 • C. For each experimental run, the specific formulation was produced with 32% of water, 16% of leavening agent (biga), and 52% of flour selected among the six different treatments (M1-M6) described in Section 2.2. The first leavening lasted for 90 min at 26 ± 1 • C, then, the dough was broken and shaped and left for a further 2.5 h at 35 ± 1 • C (second leavening). Finally, the loaves were baked at 220 • C for 45 min.

Bread Shelf-Life Assessment
After baking, the bread loaves were cooled for 2 h at room temperature (23 ± 1 • C), then sliced with an automatic slicing machine to a 20 mm thickness. Each slice was packed individually in plastic bags (two plastic layers, outer nylon layer, Food Saver, Moncalieri, Torino, Italy), by an industrial packaging machine (Lavezzini 450 GAS, Fiorenzuola d'Arda, Piacenza, Italy). Twenty loaves (1 kg each) for the six treatments were prepared, 500 slices each were packed separately in three different MAPs (Ar (100%), N 2 (100%), air (100%)). Each pack was stored at a controlled temperature (23 • C) during the whole observation period.
In this study, the samples are represented by bread cut into slices, not by the whole loaf, as the result is a much larger exposed surface that determines a more rapid evolution of the parameters. The choice of slices was made in order to be able to study the process in an accelerated-shelf-life way, as reported in a previous study [4].
Four sliced samples of each storage condition were opened each day and subjected to the following analysis in order to evaluate the bread shelf-life as a function of the flours and storage conditions. Slices were weighed daily for each experimental run to determine the weight loss brought on by the evaporation of water from the slices during storage; the value was shown as a percentage decrease from the starting value.
Water activity (a w ) was measured by a HygroPalm HP23-AW-A equipment (Rotronic AG, Bassersdorf, Switzerland). The results were calculated as a percentage reduction in water activity compared to the initial value.
As reported by Bianchi et al. (2022) [4], the softness of the crumb was determined by a penetrometer PNR-12 (Anton Paar, Rivoli, Italy), and the results were expressed as a percentage reduction in softness compared to the initial value.
All the samples were checked daily for the presence of mold; each experimental run was stopped when 3% of the samples showed mold spoilage.
Finally, the sensory profiles of the bread samples were evaluated by a panel of eight trained judges (aged between 23 and 60 years) of the Department of Agriculture, Food, and Environment Sciences of the University of Pisa. The tasting was carried out according to a previously developed protocol [43], and the overall hedonic index of bread was calculated as reported by Bianchi et al. (2022) [4]. The research obtained the approval of the Ethics Committee of the University of Pisa (protocol no. 0088081/2020).

Statistical Analysis
All the physical chemical parameters were evaluated in quadruplicate. One-way ANOVA (CoStat, Version 6.451, CoHort Software, Pacific Grove, CA, USA) was used to assess the significance of the difference between the samples, and Tukey's HSD test (p ≤ 0.05) was used for the separation of the samples.
On the parameters of flour quality, two-way ANOVA was also used to determine the effect of the two different factors (year or treatments or the combination) on the parameters of flour quality.
The trend of the shelf-life parameters over time (decrease of weight, water activity, and softness) and the linear regression were elaborated with the JMP software package version 17 (SAS Institute, Cary, NC, USA).
Big Sensory Soft 2.0 software (ver. 2018) analyzed the findings of the sensory analysis. Panelists and samples were the factors in the two-way ANOVA used to assess the sensory data [4].

Flour Quality Characterization
Among the analyses carried out on the various types of flours related to two successive years (2019 and 2020), the most significant parameters related to further bread spoilage attitude, mainly related to the staling process and mold development, are reported below and further discussed (Tables 1 and 2). All data showed almost the same trend as a function of fertilization rate, regardless of the crop season analyzed.  Further, as reported in Table S1, the most significant differences among the flour samples were determined by treatment, followed by year crop season, while only a few parameters were significantly affected by the combination "treatment × year".
According to Rekowski et al. (2019) [25], higher protein content in flour determines a reduction in the loss of free water during staling, because protein's water retention power allows it to gradually counteract the starch recrystallization process. A greater amount of protein and starch were induced by the highest nitrogen fertilization rate (flours M5 and M6), followed by that with M3 and M4 flours, while the lowest values were reported for M1 and M2 flours.
As the sugar content increases, the gelatinization temperature increases too, while the rate of the starch retrogradation decreases [16]. In the experimental conditions, flours M1, M3, and M4 had the highest sum of total sugars compared to that of the other flours. The concentration of gluten, both dry and wet, is fundamental to evaluating the breadmaking properties of flour [10]. The higher its value, the more the flour is suitable for this purpose. M5 and M6 flours showed the highest gluten concentration among others, while M2 flour showed the lowest one, thus suggesting that the increase in N fertilization had a positive effect on this parameter [30].
According to Schirmer et al. (2013) [11], amylose/amylopectin content is directly proportional to starch retrogradation rate, as amylose tends to recrystallize much faster than amylopectin. Looking at the results obtained in the two years (Tables 1 and 2), it is clear that the flours M1, M3, and M4 contained statistically significantly lower amylose content than the others and therefore could show a reduced tendency of staling [10,11].
Furthermore, phenolic compounds have an inhibitory activity towards the development of microorganisms; therefore, a high concentration of them determines a delayed appearance of fungal bodies on the bread. In addition, as reported in [40], the combination with sourdough leavening allows for their increase in the produced bread, thanks to the ability to increased availability and bio-accessibility of this phytochemical compound. Flours M1, M3, and M4 had a statistically higher quantity of phenolic compounds (both total polyphenols and total flavonoids) and higher antioxidant power compared with those of the other flours (Tables 1 and 2).
Taken together, the data showed that the six flour mixes could be divided into two main groups showing both a different chemical composition and potentially different technological features mainly related to bread spoilage, with flours M5 and M6 being clearly different from flours M1 and M2, regardless of the crop season analyzed.

Trend of Weight Loss, Water Activity, and Softness
Water migration consists of the redistribution of free water molecules due to the moisture gradient between different areas of the product, mainly from the crumb to the crust, and this migration contributes strongly to the phenomenon of staling. During storage, the water initially included in the gel fraction of starch is gradually released because of the starch recrystallization process, thus driving further crumb softness loss typical of bread staling.
To evaluate the shelf-life trend during the storage period as a function of storage atmosphere composition, the linear regression of weight loss ( Figure S2a,b) together with the water activity decrease (Figure S3a,b) was calculated. The slopes of the regression lines and R 2 are reported in Table 3 for the two-year observation period. Table 3. Slope and the coefficient of determination (R 2 ) for the linear regression of the decrease of weight and the decrease of water activity (a w ) in the two years (2019 and 2020).

Sample
Year 2019 Year 2020 According to what was reported about the flour's features as a function of fertilization rate, both weight loss and water activity decrease showed the same trend in all the experimental conditions tested, regardless of the crop season analyzed.
As expected [4], given the flour mix, significantly faster water loss was detected when the bread was stored in air (Table 3).
Furthermore, given the storage atmosphere, the flour composition represented the main factor in determining the crumb softness reduction rate (Figure 3a,b), with a lower one measured for breads produced with flours M1, M3, and M4, while flours M2, M5, and M6 significantly reduced the bread's shelf-life. Furthermore, the technological characteristics of the flour induced by the different fertilization regimes appeared to be more important than the storage atmosphere in determining the bread's shelf-life. Indeed, breads produced with flours M1, M3, and M4 stored in air showed the same trend as when breads produced with flours M2, M5, and M6 were stored in an inert atmosphere (100% Ar or 100% N 2 ).

Mold Appearance
Molds determine the decay of the organoleptic quality that derives from the production of off-flavors in terms of alcohols and esters, which cause unpleasant hints and the appearance of spots with an abnormal color on the surface of the product.
As shown in Figure 4a

Sensory Evaluation of Bread
Breads produced with the different flours were assessed before packaging to determine their sensory profile at t = 0 ( Figure S4a,b), and flours M1, M3, and M4 produced the breads with the best sensory profiles.
During the whole observation period, further panel tests were performed daily on the breads stored in different gas atmospheres until the first molds were observed on the slice surface of the sample stored in the worst conditions.
As expected, a decrease in overall pleasantness during storage was observed due to the staling phenomenon, where the breads produced with flours M1, M3, and M4 were still acceptable (HI > 6) at the end of the trials regardless of the storage atmosphere, thus confirming the main effect of flour composition discussed above.
When gas composition was assumed as the main effect (Figure 5a,b), the best sensory profile was observed for breads stored in 100% Ar, probably due to its superior antioxidant power that better preserved aromatic molecules.

Conclusions
The study confirmed that N 2 and Ar could reduce water loss, slow down the staling process, and allowed the bread to not only last longer but also to better maintain its initial characteristics, compared to those with air. However, a more important effect is linked to differences in fertilizations of the wheat flour. In fact, the breads using the flours M1, M3, and M4 in air had the same results as those of the flours M2, M5, and M6 in a protective atmosphere of Ar and N 2 .
During two different crop seasons with different climatic behaviors, the main parameters related to flour quality and composition were mainly affected by the fertilization approach, followed by crop season, while only a few parameters were significantly affected by the combination "treatment × year".
In fact, in N/P the fertilization, when N was present in a dose equal to that of P (90/96, 45/48 and 90/48 kg/ha) or in a higher but not excessive dose (90/48 kg/ha), the bread obtained tended to have better characteristics and to last longer over time. When the N/P ratios were unbalanced in favor of nitrogen (135/48 and 135/96 kg/ha) or in favor of phosphorus (45/96 kg/ha), the shelf-life and also the chemical-physical and sensory characteristics strongly decreased.
In conclusion, the flours obtained with less abundant fertilizations of N and P had better results, representing an opportunity for producers to save by administering lower doses of fertilizer. At the same time, there would be an environmental benefit, as less energy and chemical inputs would be used.