Analysis of microstructure, starch properties, and volatilomicse reveals the effect of nitrogen fertilizer on cooking quality of proso millet

Abstract

The microstructure, starch and protein structures, cooking properties, and flavor substances of proso millet (PM) grains under high nitrogen fertilizer were analyzed, and their relationships were investigated. In terms of morphology, high nitrogen application caused shrinkage of the honeycomb pores of chyme flour and reduced the water absorption capacity, leading to increased difficulty in grain pasting. High nitrogen level results in a lumpy material and high crystallinity, which improves the stability of starch and affects the water absorption and swelling ratio of the PM during cooking. Normally, the aroma of PM porridge is mainly fruity and floral burnt, slightly woody, and earthy. The addition of nitrogen fertilizer has effect on the three main substances esters, terpenes, and aldehydes that determine the aroma of the cooked PM. Therefore, nitrogen fertilizer affects the cooking quality of PM by altering the structure of starch and protein as well as flavor substances.

Highlights

• •Morphology reflect the stability of starch and its interaction with protein.
• •Proso millet cooking quality is related to protein and starch structure change.
• •Nitrogen fertilizer mainly affects ester, terpene and aldehydes metabolites.
• •High nitrogen level increased the difficulty of grains gelatinization and cooking.

1. Introduction

In the Chinese “Book of Poetry: Odes of Wei”, there is a phrase: “Large rat, large rat, Eat no more millet we grow.” The millet in the poem is proso millet (Panicum miliaceum L.), which was the staple food of the people in the pre-Qin period. Proso millet (PM) originated from the Yellow River Valley in northern China and was once the predmoinant grain crop grown in China (Wang, Li, et al., 2023; Wang, Wu, et al., 2023). It is a drought-, salt-, and alkali-tolerant crop adapted to arid or semi-arid areas and is widely distributed in the northern region of China (Yuan et al., 2021). PM has a low glycemic index and contains a variety of nutritional biomacromolecules (60 %–70 % starch, 8 %–14 % protein) and phytochemicals (phenolic compounds, active peptides, carotenoids, and fatty acids), which support its many bioactivities, including immunoregulation and hypoglycemic and antioxidant properties (Das et al., 2019). It is expected that PM will become a more common ingredient in pancakes, porridges, nutritious powders, and other millet-based food items because of the rising customer demand for healthier options.

After hulling, waxy PM or “Huangmi”, which is low in amylose, is used to make holiday cakes, while non-waxy PM (high amylose) is eaten as steamed grain or as porridge (Yang et al., 2019). Therefore, cooking quality is an important factor affecting the processing and utilization of PM. The quality of cooked PM is a multifaceted attribute, influenced by the grain's inherent characteristics such as starch and protein composition. Studies have shown that non-waxy PM contained higher resistant starch and lower rapidly digestible starch than waxy PM; compared to cooked PM with low amylose content, cooked PM with high amylose content had more soluble solids and lower light absorption value (Yang et al., 2018). It is widely believed that the hardness of cooked rice is positively correlated with protein levels. During rice cooking, protein competes with starch for water and prevents starch granules from expanding by generating disulfide bonds. Meanwhile, after heating, the protein forms a gel matrix, which increases starch integrity and reduces viscosity while increasing hardness in cooked rice (Shi et al., 2023). Aroma serves as a key determinant of the quality of cooked PM, with volatile metabolites being the primary source of its formation. Boiled foxtail millet produces ‘cooked rice’ aromas, which are attributed to dienals such as (E,E)-2,4-decadienal (Bi et al., 2019). There is comparatively little known about the flavor characteristics of cooked PM, as the focus of most of the research on PM is the chemistry of the fresh grain.

Nitrogen is one of the most important nutrients for crop development. The application of high nitrogen fertilizer can greatly increase crop yields but also adversely affect the cooking quality of the crops. Studies have shown that protein and amino acid contents increase with the nitrogen application rate (Borchers & Pieler, 2010). Amylose content and large granules of starch decreased with increasing nitrogen level, thereby resulting in high swelling power, water solubility and gelatinization enthalpy of rice starch (Zhu et al., 2016). In another study (Jiang et al., 2022), researchers demonstrated that panicle nitrogen fertilizer degraded the eating quality of japonica rice, increasing its hardness and deteriorating its pasting properties. However, studies have shown that the reactions of crop starch to nitrogen fertilizer vary with the variety (Wang, Li, et al., 2023; Wang, Wu, et al., 2023). The breakdown and peak viscosities of giant embryo rice decrease with increasing nitrogen content, whereas Koshihikari shows the opposite trend (Hu et al., 2022). However, the impact of high nitrogen fertilizer application in the crop field on PM cooking behavior has not yet been investigated.

The focus of the majority of studies on PM chemistry has been the effect of nitrogen fertilizer on PM starch characteristics (Wang, Li, et al., 2023; Wang, Wu, et al., 2023); the morphological changes that occur during cooking have not been well researched. Therefore, to better understand the changes in PM cooking quality under high nitrogen fertilizer levels, it is necessary to analyze the morphology and structure of PM proteins and starch during cooking. Accordingly, we analyzed non-waxy and waxy PM varieties to explore the relationships between starch and protein morphology, structure, cooking characteristics, and flavor under high nitrogen levels. The results of this study will aid in the production of high-quality PM, the processing of PM starch, and the enhancement of flavor of cooked PM.

2. Materials and methods

2.1. Experimental sites

The field study was carried out at an experimental station at Northwest A&F University, Yulin, Shaanxi Province, China (109.7°E, 38.3°N) during the PM growing season (June–October) in 2022. The soil parameters of the 0–20 cm layer were as follows: pH 8.59, 6.12 g∙kg⁻¹ soil organic matter, 0.25 g∙kg⁻¹ total N, 19.80 mg∙kg⁻¹ available phosphorus, and 161.67 mg∙kg⁻¹ available potassium. A waxy PM variety, Zhengninghongnian (W139, with 1.99 % amylose content), and a non-waxy PM variety, NeiMi 5 (N297, with 21.86 % amylose content) were used in this study. The experiment was conducted using a split-plot design and was replicated three times. The treatments included two nitrogen levels (0 and 270 kg N/ha, representing N0 and N3, respectively) in the main plots and 2 PM varieties in the subplots. Before planting, P (100 kg∙ha⁻¹) and K (75 kg∙ha⁻¹) were applied as basic fertilizers, followed by N (urea) application. Field management followed the recommended standard practices in this area.

2.2. Sample preparation

Mature PM grains were placed in kraft paper bags for moisture equilibration. When the moisture content reached 12 %–14 %, the grains were hulled by experimental hulling (SY88-TH, Korea) and then ground and sieved to determine the total protein, total starch, crude fat, and fatty acid contents. PM grains (2 g) were placed in a beaker and washed for 20 s, after which the excess water was removed. Distilled water (20 mL) was added to beaker, which was then placed in sauce-pan with a lid, and heated in an induction cooker (C21-WT2118, Midea, China) at 2000 W for 40 min to obtain the PM porridge. The PM grains were removed at the following timepoints: 10 min (T1), 20 min (T2), 30 min (T3), and 40 min (T4). The cooked grains were immediately submerged in freezing water to stop the cooking process (Yang et al., 2019). Some of the grains were used for morphological observations, and the remainder were freeze-dried and ground for structural analysis.

2.3. Chemical composition of PM grains

A previously published method (Zhou et al., 2020) was followed to measure the starch absorbance at 510 nm using a multifunctional enzyme marker (Multiskan GO, Thermo Fisher, USA). The nitrogen content of PM was determined using the Kjeldahl method and the result multiplied by a conversion factor of 6.25 to calculate the PM total protein content. The crude fat content of the PM was determined following the standard method GB 5009.6–2016. Detailed determination methods are described in Supplementary Material 1. The fatty acid content was determined using an Agilent 7820 gas chromatography system (Agilent Technologies, USA). Technical support was provided by Sanshu Biotech. Co., Ltd. (Shanghai, China).

2.4. Cooking quality analysis of PM grains

To analyze the cooking quality (pH, soluble solids, light transmittance, iodine blue value, expansion ratio, and water absorption ratio), cooked grains were removed from the grain-water mixture. The pH of PM porridge was measured using a PHS-3C pH meter (INESA Instrument Company, Shanghai, China). The soup was centrifuged using a low-speed centrifuge (SF-TDL-5 A, Shanghai Fichar Analytical Instruments Co., Ltd., China) at 4000 ×g for 5 min after water was added to increase the volume to 45 mL. A drying oven was used to dry 10 mL of supernatant at 105 °C until the weight of the soup-soluble solid did not change. The light transmittance of the supernatant was measured at 660 nm using an ultraviolet spectrophotometer (Blue Star B, Lab Tech Ltd., China). 1 mL of the supernatant was mixed with 5 mL of 0.5 mol/L HCL, followed by 1 mL of iodine solution. The solution was then added to 100 mL, and the iodine blue value was measured at 660 nm. The grains were cleaned with filter paper to remove surface water and weighed, and the volume was calculated using the volume displacement technique by draining water (Yang et al., 2018).

2.5. Morphology analysis of cooked PM grains

The changes in morphology of the grains at different stages of cooking was recorded using an entity microscope (SZX16, Olympus, Japan). The number of grains with altered morphology was counted at each cooking stage. PM flour was spread over a double-sided carbon adhesive tape on an aluminum plate before being vacuum-sprayed with gold. Images were obtained with a scanning electron microscope (Hitachi S-3400 N, Hitachi, Japan) set at 5.0 kV accelerating voltage and 500 and 1000× magnification (Shi et al., 2023).

2.6. Structure analysis of cooked PM grains

Structural changes in starch and protein during PM cooking were analyzed using Fourier transform infrared spectroscopy (FTIR) (Nicolet iS10, Thermo Fisher Scientific, MA, USA). The FTIR spectrum was obtained from 4000 to 400 cm⁻¹ at a resolution of 4 cm⁻¹ across 32 scans. The spectra between 1200 and 800 cm⁻¹ were deconvoluted using OMNIC 8.2 software, and absorbance levels at 995, 1022, and 1045 cm⁻¹ were recorded in order to determine the short-range order structure of the starch. The secondary structure of the protein was examined using the amide I bands (1600 cm⁻¹–1700 cm⁻¹). Used OMNIC and PeakFit V4.12 software for baseline correction, smoothing, and deconvolution (Wang et al., 2021).

2.7. In vitro digestibility determination of starch in cooked PM grains

The in vitro digestibility was determined according to a previously published method (Chen et al., 2021). Briefly, 0.1 g of cooked PM flour was mixed with 5 mL of 0.01 mol/L NaOH and 15 mL of 0.1 mol/L phosphate standard buffer (pH = 6.8) while continuously stirring, after which 2500 U of α-amylase, 300 U of amyloglucosidase, and 200 U of trypsin were added, and the reaction vessel set in a water bath at 37 °C for 2 h. After 0, 20, and 120 min, the digestion solution (0.5 mL) was removed, and anhydrous ethanol (0.5 mL) was added to remove the enzyme. The glucose content of the digestive fluid was determined using DNS colorimetry. The hydrolyzed glucose content of starch at different time points was recorded as FG, G₂₀ and G₁₂₀. The values of rapidly digestible starch (RDS), slowly digestible starch (SDS), and resistant starch (RS) were calculated as follows:

$$\mathrm{RDS}\left(\%\right)=\frac{\left({\mathrm{G}}_{20}-\mathrm{FG}\right)\times 0.9}{\mathrm{TS}}\times 100$$

$$\mathrm{SDS}\left(\%\right)=\frac{\left({\mathrm{G}}_{120}-{\mathrm{G}}_{20}\right)\times 0.9}{\mathrm{TS}}\times 100$$

$$\mathrm{RS}\left(\%\right)=\frac{\mathrm{TS}-\mathrm{RDS}-\mathrm{SDS}}{\mathrm{TS}}\times 100$$

where TS represent total starch content of sample.

2.8. Volatilomics analysis of cooked PM grains

Headspace-solid phase microextraction-gas chromatography–mass spectrometry (HS-SPME-GC–MS) was used to analyze the volatile profiles of the PM porridge (Li et al., 2021). The PM porridge was ground using liquid nitrogen and homogenized. Saturated NaCl solution and 20 μL of 10 μg/mL Benzaldehyde-d6 were added to a vial containing 1 mL of sample. The samples were shaken at a constant temperature of 60 °C for 5 min. The volatiles were then extracted from the PM porridge by headspace solid-phase microextraction using Divinylbenzene/Carbon Wide Range/Polydimethylsiloxane (DVB/CWR/PDMS) fibers for 15 min. Before the extraction process, the DVB/CWR/PDMS fibers were activated at 250 °C for 5 min in a Fiber Conditioning Station. Samples were extracted using HS-SPME for GC–MS analysis.

The GC–MS was set up with the following parameters. The injection port temperature was 250 °C, and split-less injection was employed. A DB-5MS column (30 m × 0.25 mm × 0.25 μm, Agilent J&W Scientific, Folsom, CA, USA) was used for separation. The temperature program was as follows: 40 °C for 3.5 min; increased to 100 °C at 10 °C/min; increased to 180 °C at 7 °C/min; increased to 280 °C at 25 °C/min and held for 5 min.

2.9. Statistical analysis

Metware Cloud (https://cloud.metware.cn), a free online data analysis tool, was used to perform principal component analysis (PCA) and orthogonal partial least squares discriminant analysis (OPLS-DA). Volatile differential metabolites between groups were determined by variable importance in the projection (VIP)>1 and |log₂FC| ≥ 1. The OPLS-DA data, which also included scores and permutation plots, were used to extract the VIP values. One-way analysis of variance (ANOVA) and Duncan's test were used in SPSS 23.0, to detect significant differences between means (p < 0.05). Graphs were created using Origin 2022 and Adobe Illustrator 2021.

3. Results and discussion

3.1. Effect of nitrogen fertilizer on chemical composition of PM grains

Significant differences were found in the total protein, starch, and fat content of PM grains under nitrogen fertilization (Table S1). Under nitrogen treatment, the total protein content increased significantly, whereas the amylose content decreased. Changes in crude fat content showed different patterns between the two varieties, decreasing in the waxy variety (W139) and increasing in the non-waxy variety (N297). In a previous study, it was shown that as the amount of N fertilizer increases, the content of total protein and amino acids of cooked grain increased, whereas total starch and amylose content decreased (Singh et al., 2011), which is consistent with the results of our research. Nitrogen application increased nitrate reductase and glutamine synthase activities and promoted protein synthesis of PM grains. In cereals, carbon and nitrogen metabolism utilize similar ATP and carbon skeletons, which coincide with an increase in the rate of protein synthesis. Therefore, increased protein synthesis results in lower starch content (Wang, Li, et al., 2023; Wang, Wu, et al., 2023). High-protein rice may be difficult for heat and moisture to penetrate, which inhibits starch granules from gelatinizing and further degrades the quality of the food (Zhou et al., 2020). Furthermore, protein degradation provides additional flavor precursors, such as free amino acids, which could help create aromas during PM cooking. Lipids interact with starch molecules to form starch-lipid complexes, thus affecting the viscosity, texture, and flavor of rice (Zhu et al., 2020). Similarly, in our study, nitrogen fertilization reduced the fat content of the PM variety W139 and increased that of variety N297, which may explain the differences in relative odor activity value (rOAV) values of the PM porridge. Eight fatty acids were detected in the PM grains, and the unsaturated fatty acid content exceeded 78 % (Table 1). Li et al. (2021) reported that the unsaturated fatty acid content was positively correlated with aroma and overall quality of rice. Increasing the unsaturated fatty acid content can significantly improve the eating quality of rice. Therefore, it is reasonable to speculate that nitrogen fertilizer promoted an increase in the unsaturated fatty acid content of N297, which led to the upregulation of the aroma metabolites of N297 and an increase in the rOAV value. In summary, the regulation of the main components of PM grains by nitrogen fertilizer may be an important factor in the formation of differences in porridge aroma.

Table 1.

Effect of nitrogen fertilizer on fatty acid content of PM grains ^a.

Treatment	Unsaturated fatty acid				Saturated fatty acid
	Oleic acid (C18:1n9c)	Linoleic acid (C18:2n6c)	Arachidic acid (C20:0)	Linolenic acid (C18:3n3)	Palmitic acid (C16:0)	Stearic acid (C18:0)	Behenic acid (C22:0)	Erucic acid (C22:1n9)
W139N0	19.51 ± 0.06c	57.22 ± 0.04b	0.60 ± 0.01b	1.57 ± 0.02a	12.20 ± 0.04b	8.16 ± 0.02b	0.35 ± 0.01b	0.39 ± 0.01c
W139N3	19.37 ± 0.04d	56.05 ± 0.09c	0.67 ± 0.01a	1.51 ± 0.01b	12.43 ± 0.02a	8.73 ± 0.03a	0.39 ± 0.02a	0.86 ± 0.01b
N297N0	21.85 ± 0.01b	57.27 ± 0.02b	0.58 ± 0.01c	1.06 ± 0.01d	9.84 ± 0.01c	8.13 ± 0.01b	0.33 ± 0.01c	0.94 ± 0.01a
N297N3	22.48 ± 0.07a	58.71 ± 0.04a	0.58 ± 0.01c	1.13 ± 0.01c	9.23 ± 0.02d	7.40 ± 0.01c	0.32 ± 0.01d	0.16 ± 0.01d

^aData are means ± standard deviation, n = 3. Values in the same column with different letters are significantly different (p < 0.05).

3.2. Effect of nitrogen fertilizer on cooking quality of PM grains

The cooking quality of PM grains were demonstrated in Fig.1. The acidic substances in the PM soup mainly originate from fatty acids produced by fat (Liu et al., 2020). Thus, the pH of the PM soup of the N297 variety with high fatty acid content was greater than that of the W139 variety. When the pH of soup is close to neutral, the starch granules are more likely to absorb water and gelatinize, thereby forming a suitable porridge quality for consumption (Yang et al., 2018). The light transmittance, iodine blue values, and soluble solids are closely related to the leaching characteristics of amylose and the structural features of the grain (Yang et al., 2019). The light transmittance and iodine blue values of the N297 variety were greater than the corresponding values from the W139 variety, which can be attributed to the higher amylose content of non-waxy PM. Furthermore, nitrogen fertilizer significantly reduced soluble solids, probably because nitrogen fertilizer maintains the network structure of proteins, and starch granules do not easily detach, resulting in lower cooking losses (Tang et al., 2019). Water absorption and expansion ratio are important parameters for measuring the cooking quality. The higher the water absorption, the more significant the improvement in sensory quality of the grain (Yang et al., 2018). Study found that after nitrogen application, water absorption and expansion ratio increased in W139 and significantly decreased in N297. Water immersion has been shown to expand grain volume. In the case of the PM analyzed in the present study, the N297 variety processed using nitrogen had poor water absorption capacity; therefore, the volume expansion rate was reduced. Furthermore, the greater expansion ratio may directly lead to improved palatability and, ultimately, improved edible quality of the porridge (Asimi et al., 2023).

Fig. 1.

Effect of nitrogen fertilizer on cooking properties of proso millet.

3.3. Effect of nitrogen fertilizer on morphological changes of PM grains during cooking

Images of the PM grains were obtained during the cooking process at 10 min intervals. As shown in Fig. 2A, the grain morphology of W139 and N297 changed by varying degrees during cooking. Cooked PM grains are classified into four morphological types: undeformed (I), slightly deformed (II), greatly deformed (III), and completely deformed (IV). Interestingly, the two varieties presented four completely different forms, with the W139 variety defined as implicit, and the N297 variety defined as open. During cooking, the percentage of cooked W139 variety grains with small splits gradually increased, whereas the proportion of grains with large splits and broken shapes remained low (Fig. S1). These results indicated that the W139 variety was more resistant to cooking than was the N297 variety. The PM grains were squeezed using a glass plate at 10-min interval to observe the morphological changes (Fig. 2B). After cooking for 20 min, only the innermost part of the grains remained ungelatinized at N0, whereas a more opaque section emerged at N3. The proportions of morphology types after cooking are shown in Fig. S1. Compared to the N0 treatment, the N3 treatment reduced the percentage of III and IV morphologies at the T3 and T4 stages, indicating a decrease in the degree of grain gelatinization, which was consistent with the results of pressing morphology (Sun et al., 2023). At a given cooking time, the grains in the N3 treatment were cooked to a lesser degree, which suggests that nitrogen fertilization leads to longer cooking times in PM. Because protein and starch are the most abundant compounds in PM, the increased cooking time caused by nitrogen fertilizer application might be attributed to protein and starch changes (Xiong et al., 2022).

Fig. 2.

Characteristics of morphological changes of proso millet grains during cooking under nitrogen fertilizer application. (A) Four grain morphologies. (B) Morphology of grains pressed between two slides.

The SEM images revealed that the starch in the PM flour resides mostly in the form of agglomerates, possibly because of the gelatinization of starch during the cooking process (Fig. 3). This conclusion is consistent with the findings of a previous report (Silva et al., 2013), in which it was found that undamaged starch granules in the dough were completely enlarged in cooked noodles. Starch was almost gelatinized after the T2 stage. Compared to the W139 variety, the N297 variety showed a higher level of gelatinization. Specifically, the N297 variety showed an extensive and porous network structure, which was associated with the leaching of starch after the appropriate heat treatment (red arrow). The starch was detached from the protein surroundings, and the mixture of paste and protein formed a porous network structure. In addition, the presence of porous network structure was associated with the escape of water during freeze-drying (Chen et al., 2019). The N297 variety had a higher water absorption ratio and better gelatinization properties (Fig. 1, Fig. 2), which may be attributed to more water molecules penetrating and remaining within the starch particles. Lv et al. (2023) found that foxtail millet varieties with higher quantities of pores but smaller pore sizes may have low edible qualities. The number and size of the holes in the paste may be related to the palatability and homogeneity of the porridge. Compared with the N0 treatment group, the grains at N3 in the T2 stage had a small amount of ungelatinized starch granules (red box), and there was some granular matter around the large amyloplasts (Zhu et al., 2019). Starch begins to gelatinize as the cooking time increases, and more amyloplasts become smooth and flat. Similar results were reported in a previous study (Pan et al., 2017). At the N3 level, the grains still had a significantly flat appearance, indicating that the protein and starch were more stable when combined, making it difficult to cook and mechanically destroy.

Fig. 3.

Micromorphology of proso millet flour during cooking under nitrogen fertilizer application. (A and C) 500×. (B and D) 1000×.

3.4. Effect of nitrogen fertilizer on structural properties of PM grains during cooking

FTIR spectra has been widely used to study the short-range ordered structure of cereal and plant seed starches (Lv et al., 2023; Shi et al., 2023; Wang, Li, et al., 2023; Wang, Wu, et al., 2023). As shown in Fig. 4A, no new peaks were observed in the FTIR spectra of PM following nitrogen treatment, and the peak locations were not significantly altered. Peaks at 1047, 1022, and 995 cm⁻¹ indicated the organized, amorphous, and helical structures, respectively, of hydrated carbohydrates (mostly starch) (Wang, Li, et al., 2023; Wang, Wu, et al., 2023). Absorbance ratios of 1047/1022 and 1022/995 cm⁻¹ typically characterize the degree of short-range order of starch (Wang et al., 2019). With increased cooking time, the short-range order structure of PM was significantly reduced, but N3 exhibited a higher short-range order structure. No ordered structure was detected after T1, and the starch-pasting structure was disrupted (Table S2). Previous research found that the 1047/1022 cm⁻¹ of starch reduced as cooking time increased (Zhu et al., 2020), which is consistent with our findings. The secondary structures of protein are thought to be related to the amide I area (1700–1600 cm⁻¹) (Wang et al., 2021). We classified 1646–1664 cm⁻¹ as α-helix, 1664–1681 cm⁻¹ as β-turn, 1637–1645 cm⁻¹ as random coils, 1613–1637 cm⁻¹ and 1681–1695 cm⁻¹ as β-sheet. As shown in Fig. 4B, the secondary structure of the protein, with the most prevalent secondary structure was the β-sheet. This was supported by earlier research that found that the β-sheet protein secondary structure was the most prevalent in a large number of monocotyledonous and dicotyledonous plants (Ellepola et al., 2005). With an increase in cooking time, the protein secondary structure at the N3 level changed less than that at the N0 level, indicating that the protein secondary structure was more easily affected by temperature under low nitrogen fertilizer application. N3 treatment conditions induce greater α-helical and random coil contents in the protein, thereby improving its thermal stability through the stabilizing effect of α-helical structures (Shi et al., 2023). The T4 stage was characterized by a decrease in α-helix and β-sheets content and an increase in β-turns and random coils content in PM protein compared to the T1 stage. It is well acknowledged that the presence of α-helices and β-sheets is positively associated with the orderliness of proteins, whereas the presence of β-turns and random coils is associated with disorganized structures (Wang et al., 2021).

Fig. 4.

Fourier transform infrared spectra (A) and secondary structure changes (B) of proso millet proteins during cooking under nitrogen fertilizer application.

3.5. Effect of nitrogen fertilizer on in vitro digestibility characteristics of PM grains

The content of the cooked PM rapidly digestible starch (RDS), slowly digestible starch (SDS), and resistant starch (RS) is shown in Table 2. The changes were similar for all treatments with increasing cooking time, with an increase in the RDS content and a decrease in the RS content. Uncooked PM grains contained markedly lower RDS and SDS and higher RS than did the cooked PM grains because the grains disrupted the structure of the starch granules during cooking, which facilitates the enzyme-substrate contact/binding (Chung et al., 2012). At the T2 and T4 stages, the N3 treatment of both varieties resulted in significantly higher RDS content than did the N0 treatment, whereas RS content was significantly lower than in the N0 treatment. The rate and extent of starch digestion are determined by many factors including botanical source, surface characters, fine structure of starch polymers and interactions between starch and other components. Previous studies have shown that nitrogen fertilization promotes starch biosynthesis of PM, favoring the synthesis of amylopectin with shorter chain lengths (Wang, Li, et al., 2023; Wang, Wu, et al., 2023), which enhances starch accessibility to enzymatic attack and consequently elevates RDS content (Dhital et al., 2015). Concurrently, nitrogen application increased grain protein content, and heat-induced protein aggregation creates a disulfide-linked network, limiting water penetration and confining gelatinization primarily to granule surfaces (Wu et al., 2023). This surface gelatinization enhances surface enzyme access (increasing RDS), while restricting water access to the granule core, inhibiting full gelatinization and reducing RS formation.

Table 2.

Effect of nitrogen fertilizer on in vitro digestibility of PM during cooking ^a.

Treatment	RDS (%)			SDS (%)			RS (%)
	T b	T2 b	T4 b	T	T2	T4	T	T2	T4
W139N0	42.41 ± 1.14c	49.37 ± 0.81d	59.89 ± 0.98c	27.86 ± 0.96a	23.91 ± 1.42a	25.11 ± 1.19b	30.07 ± 1.07b	28.73 ± 0.62b	15.33 ± 0.40b
W139N3	49.05 ± 1.6b	72.56 ± 1.26a	76.32 ± 1.33a	17.63 ± 0.45c	8.40 ± 0.94b	18.40 ± 0.31c	34.31 ± 0.30a	18.37 ± 0.49d	6.61 ± 1.33d
N297N0	52.92 ± 0.39a	61.16 ± 0.66c	45.88 ± 0.16d	11.49 ± 0.75d	8.62 ± 0.42b	35.70 ± 0.72a	35.58 ± 0.51a	30.22 ± 1.03a	18.41 ± 0.72a
N297N3	48.82 ± 0.58b	65.55 ± 0.58b	66.68 ± 0.58b	24.46 ± 0.47b	9.27 ± 0.20b	13.57 ± 0.09d	26.39 ± 0.58c	25.18 ± 0.71c	10.42 ± 0.50c

^aData are means ± standard deviation, n = 3. Values in the same column with different letters are significantly different (p < 0.05).

^bT, T2, and T4 represent the grains cooked for 0 min, 20 min, and 40 min, respectively.

3.6. Effect of nitrogen fertilizer on volatile metabolites of PM porridge

3.6.1. Volatile compounds profiling in PM porridge

Through GC–MS, a total of 231 volatile compounds, classified into 16 categories, were identified in the PM porridge (Fig. S2 A). Heterocyclic compounds accounted for 19.05 % of the total volatile compounds, followed by esters, terpenoids, hydrocarbons, alcohols, aldehydes, and aromatics, accounting for more than 68 % of the volatile compound profile. To investigate the variations in nitrogen treatment, PCA was applied to both qualitative and quantitative data collected from the four samples. The first two principal components (PC1 and PC2) explained 80.19 % of the data variance. The principal component score plot and heatmap cluster are shown in Fig. S2 B and C. Samples from the different nitrogen fertilizer treatment groups were distinct from one another, reflecting the large differences in volatile compounds between the different nitrogen fertilizer treatment groups. The sample replicates were clustered together, indicating good method stability and reliable reproducibility. The OPLS-DA method combines partial least squares discriminant analysis (PLS-DA) and orthogonal signal correction (OSC) to screen variances by deleting unrelated variances. A permutation test (n = 200) was conducted to validate the model. According to the model built using OPLS-DA, the Q² of each comparison group was >0.5, confirming that the OPLS-DA comparison model of the four groups was non-overfitting and reliable (Fig. S3).

3.6.2. Differential volatile compounds of PM porridge with different nitrogen treatments

Based on the variable influence on projection (VIP) of the OPLS-DA model, the volatile differential metabolites were screened with VIP ≥ 1 and |Log2FC| ≥ 1. The results are shown in Fig. 5. In W139N3, 10 volatile differential metabolites were upregulated (compared with W139N0) and 122 volatile differential metabolites were downregulated. In the case of N297N3, 131 volatile differential metabolites were upregulated and two volatile differential metabolites were downregulated (compared with N297N0). The top 10 upregulated differentially expressed metabolites were mainly terpenoids and hydrocarbons, while the down-regulated metabolites were mainly aldehydes and acids. It can be seen that nitrogen application is an effective cultivation measure to increase terpenoids in PM grain. Terpenoids are crucial not only for aroma and flavor, but also for health. Researchers have recently focused on their beneficial bioactive components. Hydrocarbons are mainly derived from the oxidation and decomposition of lipids. The high content of hydrocarbons had an important contribution to the overall harmony of smell (Liu et al., 2024).

Fig. 5.

Volatile differential metabolite volcano plots and TopFC distribution.

Using enrichment pathway analysis, the available KEGG IDs of the differentially expressed metabolites were searched and examined. The differential metabolites in cooked PM were assigned to eight metabolic pathways (Fig. 6): biosynthesis of secondary metabolites, monoterpenoid biosynthesis, benzoxazinoid biosynthesis, sesquiterpenoid and triterpenoid biosynthesis, vitamin B6 metabolism, limonene and pinene degradation, metabolic pathways, and phenylpropanoid biosynthesis. In the present study, volatile differential metabolites and enriched metabolic pathways of nitrogen fertilizer-regulated PM porridge were mainly terpenoids and esters, with monoterpenes and limonene as important aroma components.

Fig. 6.

The pathway analysis of volatile differential metabolites.

3.6.3. Identification of key volatile compounds in PM porridge

The odor of a food sample cannot be described only by the presence of a single aromatic component. Thus, the odor activity value (OAV) was used to assess the contribution of different volatile chemicals to the overall odor of the PM porridge. It has been suggested that compounds with OAV values >1 are known to possess important flavor characteristics (Guo et al., 2021). In Table S2 are shown the threshold values and odor descriptors of the 25 rOAVs of the PM porridge that were >1 overall. Of the 25 compounds most had fruity, floral, and woody odors. Based on the results of this study, 5-ethyl-3-hydroxy-4-methyl, 2-nonenal, 2-thiophenemethanethiol, furaneol, 2-methylisoborneol and 3-mercaptohexyl acetate were found to be the key aroma compounds contributing the most to PM porridge aroma. These compounds had low thresholds and were detected in PM porridge samples with very high OAV values. The main aroma molecules responsible for the aroma of japonica rice porridge are hexanal, octanal, nonanal, decanal, (E)-2-nonenal, naphthalene, (E, E)-2, 4-decadienal, butylated hydroxytoluene, 1-dodecanol, and 2-pentadecanone (Hu et al., 2020). The most important heterocyclic component in rice, 2-pentylfuran, is the main factor that differentiates between aromatic and non-aromatic rice types (Mi et al., 2023). It is noteworthy that nitrogen fertilizer reduced most of the volatile metabolite odors in the W139 variety, whereas the odor values were elevated in the N297variety. Research on rice has shown that higher levels of total soil nitrogen lead to higher levels of 1-proline, a precursor of 2-AP, as well as richer aromas in traditional Chinese regional aromatic rice (Yang et al., 2012).

3.6.4. Differential metabolite flavoromics analysis

In each differential comparison group of metabolites, the 10 sensory flavors with the highest number of annotations were selected for radar plotting. As shown in Fig. 7A, the W139N3_ vs _W139N0 group flavor was green, sweet, fruity, and fatty, and contained the following annotated metabolites: four alcohols, 14 esters, eight aldehydes, four terpenes, four heterocyclic compounds, one ketone, one nitrogen-containing compound, and one acid. The N297N3_vs_N297N0 group flavor was predominantly green, sweet, fruity, and woody and included three alcohols, 15 esters, six aldehydes, eight terpenes, six heterocyclic compounds, two ketones, one sulfur-containing compound, and one acid. Flavouromic analysis of the differential metabolites revealed that nitrogen fertilization mainly affected three major substances, namely esters, terpenoids, and aldehydes, which in turn led to differences in the odor of the PM porridge. Esters are mainly responsible for the aroma of fruits and tea, which may be an important factor in the fruity flavor of PM porridge (Wang & Ha, 2013). Aldehydes are thought to be mainly produced via lipid oxidation and decomposition, contributing the most to the overall flavor among all categories because of their relatively low odor threshold. Another study has also shown that aldehydes had a strong aroma, generally with light fragrance, fruit fragrance, etc., and even an important part of the aroma of steamed bread (Liu et al., 2024).

Fig. 7.

Radar plot (A) and flavoromics sankey plots (B) of sensory flavor profile analysis of volatile differential metabolites.

The upregulation or downregulation of the differential metabolites of each sensory flavor can be seen more clearly in the Sankey diagram (Fig. 7B). Nitrogen fertilizer downregulated flavor-producing metabolites (only four metabolites were upregulated) in the W139 variety and upregulated all flavor-producing metabolites in the N297 variety, which was in general agreement with our results for crude fat and unsaturated fatty acid content. Furthermore, Li et al. (2023) reported that volatiles are mainly generated from the oxidation of fatty acids during cooking, Unsaturated fatty acids produce the most volatile substances and contribute the most to aroma. Based on the above two points, nitrogen fertilization may have increased the fat content and unsaturated fatty acid content of the N297 variety, accelerated lipid oxidation, and promoted aroma formation.

4. Conclusion

Nitrogen fertilizer significantly decreased the total starch and amylose contents and increased the protein and amino acid contents. Compared to the non-waxy variety (N297), the waxy variety (W139) had a higher starch crystallinity and a low amylose and unsaturated fatty acid content, which resulted in the PM porridge presenting high digestibility and low iodine blue value, and a long cooking time of the grains. In terms of morphology at the microscopic level, high‑nitrogen fertilizer application caused shrinkage of the honeycomb pores of chyme flour and reduced the water absorption capacity, leading to increased difficulty in grain pasting and cooking. High nitrogen content results in a lumpy material and high crystallinity, which improves the stability of starch and affects the water absorption and swelling ratio of PM grains during cooking, thus affecting the homogeneity and texture of the PM porridge. The aroma of PM porridge is mainly fruity and floral burnt, slightly woody, and earthy; nitrogen fertilizer mainly regulates the three main substances, esters, terpenes, and aldehydes, leading to differences in aroma. High nitrogen application led to an increase in protein and fat content, changed the structure of protein and starch during cooking, and ultimately affected the PM aroma and cooking quality.

CRediT authorship contribution statement

Honglu Wang: Writing – original draft, Conceptualization. Yahong Xiong: Formal analysis, Data curation. Yining Zhang: Methodology, Investigation. Yernaz Yermekov: Software. Jinfeng Gao: Visualization, Validation. Baili Feng: Writing – review & editing, Project administration, Funding acquisition.

Declaration of competing interest

This manuscript has not been published or presented elsewhere in part or in entirety and is not under consideration by another journal. We have read and understood your journal's policies, and we believe that neither the manuscript nor the study violates any of these. There are no conflicts of interest to declare.

Appendix A. Supplementary data

Supplementary material

Data availability

Data will be made available on request.