Starch Properties, Nutrients Profiles, In Vitro Ruminal Fermentation and Molecular Structure of Corn Processed in Different Ways

Abstract

Processing will improve the digestion of corn by ruminant animals. The objectives of this study were to investigate the effects of processing methods (grinding, G; steam flaking, SF; extrusion, E) on the starch properties, nutrient profiles, in vitro ruminal fermentation and molecular structure of corn. Compared with G, SF and E increased (p < 0.05) the starch content, starch gelatinization, ruminal gas production (GP, 0.5–32 h), propionic acid, starch degradability (SD), the area and height of carbohydrate peaks, and decreased (p < 0.05) starch crystallinity, content of crude protein, neutral detergent fiber and acid detergent fiber, ruminal NH₃-N, the area and height of amide I and II, α-helix, and β-sheet. The total VFA (24 h, 48 h) tended to be increased by SF and E (p < 0.10). The carbohydrate peak area and height were positively (p < 0.05) correlated with GP (1–24 h) and SD. The protein molecular absorption intensity was negatively correlated with SD (p < 0.05). The change in starch properties, GP (1–24 h) and molecular structure caused by E was greater than SF (p < 0.05). These results indicated that the higher starch gelatinization and lower starch crystallinity of E corn, induced by the high temperature and pressure, enabled more fermentation and digestion in the artificial rumen. The carbohydrate and protein molecular structures were correlated with the nutritional characteristics of corn.

1. Introduction

The feed costs constitute up to 70% of the costs of animal production [1]. Increasing the utilization of feed is very important to improve the animal production efficiency and benefit. Corn represents the most important energy source in the diet of ruminant animals, usually accounting for about 60% of the concentrated feed formula [2]. The energy of corn mainly comes from starch, which accounts for about 70% of the weight of corn [3]. The starch of corn exists in endosperm. The endosperm texture, the structure of starch granules and the interactions with proteins affect the starch fermentation in the rumen [4]. Processing of the corn will destroy the structure of starch granules and improve the utilization. Previous research showed that a simple mechanical process such as grinding (G) could break up the pericarp and reduce the granular size, partially exposing the endosperm [2], increasing the surface area of the endosperm for rumen enzymes [5], and improving the digestibility [6]. However, the starch–protein matrix undestroyed by G may be a physicochemical impairment to starch digestion by rumen microorganisms [7]. In addition to the alteration of granular size and surface area by G, thermal processing such as steam flaking (SF) and extrusion (E) modifies the crystalline structure of the starch granules under the combined action of shear force, heat and pressure [2], promotes starch gelatinization [2], facilitates the rumen microbe and enzyme access to the starch granules [8], and increases the digestibility [9] and energy utilization [10] for the ruminant animal. Although SF and E follow a similar general pattern where the application of heat, pressure and shear forces are involved in all of them, they are different processing techniques. SF usually refers to heat treatment with steam (80–120 °C for 37–67 min, moisture 16–21%) to make the corn swell and soften, followed by pressing it into 0.5–2 mm thick flakes, while E provides higher temperature (100–180 °C) and pressure (3–10 MPa) in a shorter time. The two kinds of thermally processed corn are currently used in diets for young ruminant animals. However, none of the available studies made a clear comparison of the effects of SF and E on the starch properties, nutrient profiles and rumen fermentation of the corn. The only comparative study [11] between the two types of thermally processed corn using an in vitro gas test showed that the gas production (GP), organic matter digestibility and metabolizable energy of steam-flaked corn (SFC) were lower than those of extruded corn (EC). Limited research has been conducted in this area with lambs.

Attenuated total reflectance Fourier-transformed infrared (ATR-FT/IR) molecular spectroscopy is a technique that can reveal the molecular structure of biological tissues with ultra-high spatial resolution [12]. This method has been applied successfully to evaluate the carbohydrate or protein molecular structure of faba [13], canola seed [14], oats [15], barley [16], carinata meal [17], and alfalfa hay [18]. The results revealed correlations between the molecular structure characteristics and nutrient degradation in the rumen of consumers. However, few comparative studies of the molecular structure of differently processed corn have been reported. Xu et al. [19] observed that SFC had greater molecular absorbance intensities of most carbohydrate biopolymers, lower protein-associated molecular spectral intensities compared to untreated corn and the carbohydrate and protein spectral intensities were associated with the ruminal degradation of dry matter (DM) and starch.

The objectives of this study were to investigate the effects of three corn-processing methods on the starch properties, nutrient profiles, in vitro rumen fermentation, molecular structures of carbohydrates and proteins; and to analyze correlations between the molecular structure and rumen fermentation.

2. Materials and Methods

2.1. Corn Processing and Treatments

A completely randomized design with 3 treatments and 3 replicates for each treatment were used. Corn was obtained from Xianyang Fengruida Biotechnology Co., Ltd. (Xianyang, China) and processed in 3 ways (G, SF, and E). Ground corn (GC) was milled using a farm grinder (SP132-75M, FAMSUN Group Co., Ltd., Yangzhou, China) equipped with a 3 mm screen and run at a speed of 2950 RPM. The SF process (30T/D, Shandong Guanfeng Machinery Co., Ltd., Jining, China) involved steaming (100–110 °C) the whole grain (50 min) to raise the moisture level to 21–23%, after which the steamed grain was passed through a preheated roller to produce a density of about 190–200 g/L and thickness of 0.5 mm. During the E process, the corn was extruded through a single screw extruder (Y-32, Jiangsu Muyang Group Co., Ltd., Yangzhou, China) with a screw diameter of 25 mm, length–diameter ratio of 10:1, spindle speed of 400 RPM, extrusion temperature of 130–140 °C, and diameter of 3 mm. For each processing method, feed samples were collected every 5 min, 3 times to obtain 3 replicate samples. The three kinds of processed corn came from the same raw material and thus avoided possible basic differences caused by different varieties.

2.2. Chemical Analysis

The DM content was determined by oven-drying a subsample at 100 °C [20]. Ether extract (EE) contents were measured using methods 945.18 [20]. Total nitrogen (TN) content was determined using the Kjeldahl method [20] on a Kjeltec Auto 8400 (Kjel-Foss, Ramnes, Norway). Neutral detergent fiber (NDF) and acid detergent fiber (ADF) were measured following Van Soest et al. [21]. Starch content was measured using a starch content assay kit [9] purchased from Jiancheng Bioengineering Institute (Nanjing, China). The gelatinization of starch was determined according to the procedure (AACC method 76-31.01; K-SDAM, 09/2018) [22]. The starch sample was analyzed with an X-ray diffractometer (D8 ADVANCE, Bruker Co., Karlsruhe, Germany) and relative crystallinity was calculated as the ratio of the area of the peaks to the total area under the curve using the DIFFREAC.EVA V3.1 software [23].

2.3. In Vitro Ruminal Fermentation

Rumen fluid was obtained from 3 Hu sheep with similar body weights (22 ± 0.5 kg). The animals were fed a diet [24] containing forage and concentrate (50:50) twice a day (8:00 and 17:00). In vitro gas production tests were conducted following Menke et al. [25]. The main device was a 100 mL calibrated glass syringe (Haeberle, Lonsee, Germany). The syringe was filled with 30 mL of culture medium, which consisted of 10 mL of rumen fluid and 20 mL of buffer solution [26]. The medium was incubated in a water bath at 39 °C, with 6 replicates for each treatment. The amount of produced gas was recorded at 0, 0.5, 1, 1.5, 2, 4, 8, 12, 24, 32, 40, and 48 h of incubation. Cumulative gas production data was fitted to the model of Orskov and McDonald (1979) [27] as follows:

GP = a + b (1 − e ^−ct)

(1)

where: a is rapid GP (mL), b is slow GP (mL), (a + b) is potential (mL), c is rate constant of slow GP (%/h), t is time (0.5, 1, 1.5, 2, 4, 8, 12, 24, 32, 40, 48 h) since commencement of incubation (h).

The residuals in the glass tubes at 24 (3 syringes) and 48 h (3 syringes) were collected to determine the volatile fatty acid (VFA), NH₃-N and starch degradability (SD). VFA was determined by gas chromatography (6890N, Agilent, Santa Clara, CA, USA) according to Wu et al. [28]. NH₃-N was analyzed using the colorimetric method described by Rhine et al. [29]. SD was calculated using the following equation:

SD = (starch content before fermentation − starch content after fermentation)/starch content before fermentation.

(2)

2.4. Vibrational Molecular Spectroscopy

The vibrational molecular spectroscopy analysis was performed at Lanzhou Institute of Chemical Physics, Chinese Academy of Sciences (Lanzhou, China). Spectral profiles were obtained using an ATR-FT/IR molecular spectroscopy device (Perkin Elmer, Waltham, MA, USA) equipped with a universal ATR sampling accessory. The samples were finely ground and pressed uniformly against the diamond surface, and the mid-IR spectra were recorded at a resolution of 4000 to 600/cm at 1/cm [30]. Each treatment had 3 replicates and each replicate was measured 3 times.

2.5. Univariate Molecular Spectral Analysis

Molecular spectral features were analyzed using OMNIC 8 software (Spectra Tech, Madison, WI, USA). The following carbohydrate biopolymer-related functional group-associated spectral regions [19] were analyzed: (1) total carbohydrate (CHO) biopolymer spectral peak (spectral baseline ca. 1187 to 946 cm⁻¹); (2) first, second, and third major sub-peaks within the CHO fingerprint region with spectral baseline ca. 1187 to 1131 cm⁻¹, ca. 1131 to 1066 cm⁻¹, and ca. 1066 to 946 cm⁻¹, respectively. The following protein-associated spectral features were analyzed: (1) protein structural amide I peak (spectral baseline ca. 1717 to 1575 cm⁻¹) and amide II peak (spectral baseline ca. 1575 to 1485 cm⁻¹), and (2) secondary protein structures, including α-helix (centers at ca. 1658 cm⁻¹) and β-sheet (centers at ca. 1652 cm⁻¹). The latter were collected by the second derivative function and Fourier self-deconvolution method in OMNIC 8 software [31].

2.6. Multivariate Molecular Spectral Analyses

Multivariate analysis of the total CHO spectral region (ca. 1187 to 946 cm⁻¹) and protein structural spectral region (ca. 1717 to 1485 cm⁻¹) were applied to evaluate structural differences caused by G, SF, and E using Statistica 12.0 software (StatSoft Inc., Tulsa, OK, USA) [32].

2.7. Data Analysis

The GLM program of SAS 9.4 (SAS Institute, Inc., Cary, NC, USA) [33] was used to analyze starch properties, nutrient profiles, ruminal fermentation and absorbance intensities of CHO and protein. The model used in the analysis was as follows:

Y = μ + Txi + εij

(3)

where Y is the variable, μ is the population mean for the variable, Txi is the effect of treatment, and εij is the experimental error. Significance was designated as p < 0.05 with a trend being between p ≥ 0.05 and p < 0.10. A Duncan’s multiple range test was used to rank group means when p < 0.05. The relationships between molecular spectral features and ruminal fermentation parameters were assessed by correlation analysis using the PROC CORR of SAS 9.4 with the Spearman option after the normality test. Prism software (version 6.01) was used to estimate gas production according to models proposed by Ørskov and McDonald (1979) [27].

3. Results

3.1. Starch Properties

The gelatinization and crystallinity of the starch are shown in Figure 1. The highest percentage of starch gelatinization (91.08%) was observed in EC, followed by SFC (67.29%) and then GC (11.01%, p < 0.05). E and SF increased starch gelatinization by up to 727.25 and 511.17 percent respectively, while the starch crystallinity of EC was the lowest (5.45%), followed by SFC (20.27%) and then GC (31.45%, p < 0.05). E and SF decreased starch crystallinity by up to 82.67% and 35.55% respectively.

3.2. Nutrient Profiles

Compared with G, the thermal processing increased the starch content of corn (p < 0.001) (

Table 1. Nutrient profiles of corn processed in different ways (g/kg DM).

Items	Ground Corn	Steam-Flaked Corn	Extruded Corn	SEM	p-Value
Starch	635.42 b	714.32 a	715.66 a	0.970	<0.001
Neutral detergent fiber	131.15 a	120.22 b	112.11 b	0.231	<0.001
Acid detergent fiber	43.35 a	24.82 b	27.51 b	0.214	<0.001
Crude protein	94.64 a	90.11 c	93.63 b	0.025	<0.001
Ether extract	37.13 a	12.28 c	21.44 b	0.057	<0.001

^{a, b, c} Means with different superscripts within the same line differ significantly (p < 0.05); SEM, standard error of the mean.

), and there was no difference (p > 0.05) between SFC and EC. The NDF and ADF contents of GC were higher than those of SFC and EC (p < 0.001). The CP and EE contents differed (p < 0.001) between the three types of processed corn and they were the lowest in SFC, followed by EC and then GC.

3.3. In Vitro Ruminal Fermentation Parameters

3.3.1. Gas Production (GP) and GP Parameters

The cumulative GP from 1 h to 24 h was the highest in EC, followed by SFC and then GC (p < 0.001) (

Table 2. In vitro ruminal fermentation parameters of corn processed in different ways.

Items	Ground Corn	Steam-Flaked Corn	Extruded Corn	SEM	p-Value
Gas production (mL/g DM)
0.5 h	4.16 b	5.66 a	6.50 a	0.302	0.010
1 h	5.08 c	8.16 b	11.50 a	0.569	0.002
1.5 h	5.66 c	11.75 b	19.08 a	0.914	<0.001
2 h	6.83 c	17.00 b	31.83 a	1.306	<0.001
4 h	17.58 c	45.91 b	85.91 a	4.390	<0.001
8 h	38.33 c	90.66 b	190.83 a	8.387	<0.001
12 h	100.41 c	169.25 b	219.58 a	8.675	<0.001
24 h	270.66 c	300.16 b	338.66 a	6.944	0.002
32 h	315.91 b	341.41 a	359.25 a	6.983	0.010
40 h	356.75	359.75	365.58	7.518	0.728
48 h	359.66	380.00	388.33	9.783	0.196
Gap production parameters
a (ml)	−25.18	−22.07	−26.11	4.563	0.628
b (ml)	613.82 a	466.88 b	417.53 b	26.482	=0.001
a + b (ml)	588.58 a	444.81 b	391.43 b	28.083	=0.001
c (%/h)	0.02 c	0.04 b	0.08 a	0.004	=0.001
VFA, 24 h (mmol/L)
Total VFA	19.21	20.65	23.74	0.544	0.095
Acetic acid	13.03	14.48	16.19	0.682	0.210
Propionic acid	4.50 c	5.35 b	6.33 a	0.333	0.045
Butyric acid	0.71	0.71	0.78	0.093	0.953
Acetic acid/propionic acid	2.89	2.71	2.69	0.405	0.102
VFA, 48 h (mmol/L)
Total VFA	19.92	21.92	25.19	0.250	0.079
Acetic acid	13.51	14.56	16.81	0.712	0.135
Propionic acid	5.34	5.79	6.40	0.261	0.169
Butyric acid	0.78	0.92	1.06	0.141	0.747
Acetic acid/propionic acid	2.52	2.53	2.63	0.107	0.201
NH3-N, 24 h (mg/L)	11.29 a	7.57 b	7.25 b	0.360	0.001
NH3-N, 48 h (mg/L)	16.96 a	12.89 b	11.19 b	0.610	0.001
Starch degradability, 24 h (%)	33.10 c	86.72 b	94.06 a	0.403	<0.001
Starch degradability, 48 h (%)	67.72 c	94.77 b	95.88 a	0.233	<0.001

^{a, b, c} Means with different superscripts within the same line differ significantly (p < 0.05); SEM, standard error of the mean; a: rapid gas production; b: slow gas production; c: rate constant of slow gas production; (a + b): gas production potential.

). GP, at 0.5 and 32 h, was higher in thermally processed corn than GC (p < 0.001) but there was no difference (p > 0.05) between SFC and EC. GP after 32 h incubation (40 h and 48 h) was not different among the three treatments (p > 0.05).

The c of E was the highest, followed by SFC and then GC (p = 0.001) (Table 2). The b and (a + b) of GC were decreased by SF and E (p = 0.001), and no difference (p > 0.05) between SFC and EC was observed. There was no difference for the value of a between the three treatments (p > 0.05).

3.3.2. Ruminal VFA, NH3-N and Starch Disappearance

Corn processing affected the propionic acid (24 h) (p < 0.05) (Table 2). It was the greatest in EC, followed by SFC and then GC. Thermal processing decreased (p = 0.001) NH₃-N and increased (p < 0.001) SD. The SD at 24 h and 48 h of EC (94.06%, 95.88%) was the greatest, followed by SFC (86.72%, 94.77%) and then GC (33.10%, 66.72%). The NH₃-N was greater in GC than SFC and EC. Thermal processing tended to improve the total VFA at 24 h (p = 0.095) and 48 h (p = 0.079). Other parameters were not affected by the treatments (p > 0.05).

3.4. Carbohydrate and Protein Structure Spectral Profiles

The carbohydrate total spectral peak areas, spectral peak height and area of CHO first, second and third peak of EC were the highest, followed by SFC and then GC (p < 0.001) (

Table 3. Molecular spectral features of carbohydrates and proteins of corn processed in different ways (absorbance).

Items	Wave, cm−1	Ground Corn	Steam-Flaked Corn	Extruded Corn	SEM	p-Value
Carbohydrate molecular spectral features
Total area	1187-946	13.70 c	18.16 b	22.55 a	0.802	<0.001
CHO 1st peak area	1187-1131	1.74 c	2.34 b	2.70 a	0.094	<0.001
CHO 2nd peak area	1131-1066	2.92 c	3.92 b	4.54 a	0.168	<0.001
CHO 3rd peak area	1066-946	9.03 c	11.89 b	15.30 a	0.541	<0.001
CHO 1st peak height	1148	0.04 c	0.05 b	0.07 a	0.002	<0.001
CHO 2nd peak height	1078	0.05 c	0.07 b	0.09 a	0.003	<0.001
CHO 3rd peak height	996	0.09 c	0.12 b	0.17 a	0.006	<0.001
Protein molecular spectral features
Amide I area	1717-1575	3.74 a	2.34 c	2.97 b	0.355	<0.001
Amide II area	1575-1485	2.09 a	1.30 c	1.51 b	0.204	<0.001
Amide I height	1646	0.03 a	0.02 c	0.02 b	0.003	<0.001
Amide II height	1533	0.02 a	0.01 c	0.01 b	0.002	<0.001
Secondary structure
α-helix height	1652	0.033 a	0.021 c	0.026 b	0.003	<0.001
β-sheet height	1658	0.034 a	0.021 c	0.026 b	0.003	<0.001

^{a, b, c} Means with different superscripts within the same line differ significantly (p < 0.05); SEM, standard error of the mean.

). The thermal processing of corn decreased (p < 0.001) the spectral peak heights and areas of protein structural Amides I and II and secondary protein structure α-helix and β-sheet. It was the lowest in the SFC, followed by EC, and then GC.

The results of multivariate analysis are shown in Figure 2. Principal component analysis divided the absorption peak regions of total CHO and protein molecules into three independent parts, which indicated that there was significant difference in the absorption peaks of CHO and protein structures between GC, SFC and EC.

3.5. Correlation between the Molecular Structure and In Vitro Gas Production and Starch Degradability

As shown in

Table 4. Correlation between molecular structure and in vitro gas production and starch degradability of corn processed in different ways.

Items	Total CHO Peak Area	CHO 1st Peak Area	CHO 2nd Peak Area	CHO 3rd Peak Area	CHO 1st Peak Height	CHO 2nd Peak Height	CHO 3rd Peak Height	Amide I Area	Amide II Area	Amide I Height	Amide II Height	α-Helix Height	β-Sheet Height
Gas production
0.5 h	0.934 *	0.946 *	0.946 *	0.924 *	0.917 *	0.926 *	0.866 *	−0.547	−0.654	−0.547	−0.595	−0.573	−0.588
1 h	0.838 *	0.802 *	0.807 *	0.846 *	0.855 *	0.854 *	0.863 *	−0.380	−0.532	−0.403	−0.459	−0.415	−0.422
1.5 h	0.928 *	0.900 *	0.903 *	0.934 *	0.938 *	0.938 *	0.937 *	−0.409	−0.561	−0.434	−0.490	−0.446	−0.457
2 h	0.935 *	0.908 *	0.900 *	0.942 *	0.953 *	0.948 *	0.964 *	−0.376	−0.543	−0.411	−0.459	−0.414	−0.426
3 h	0.913 *	0.893 *	0.874 *	0.922 *	0.936 *	0.929 *	0.960 *	−0.348	−0.519	−0.39	−0.431	−0.385	−0.395
4 h	0.916 *	0.891 *	0.874 *	0.925 *	0.939 *	0.930 *	0.967 *	−0.301	−0.475	−0.345	−0.388	−0.338	−0.349
8 h	0.937 *	0.922 *	0.906 *	0.942 *	0.954 *	0.950 *	0.971 *	−0.398	−0.564	−0.434	−0.479	−0.433	−0.443
12 h	0.945 *	0.925 *	0.918 *	0.950 *	0.959 *	0.956 *	0.978 *	−0.399	−0.563	−0.434	−0.486	−0.435	−0.445
24 h	0.848 *	0.803 *	0.817 *	0.859 *	0.866 *	0.863 *	0.872 *	−0.246	−0.398	−0.268	−0.330	−0.283	−0.290
32 h	0.699 *	0.641	0.666	0.712 *	0.718 *	0.714 *	0.715 *	−0.104	−0.231	−0.120	−0.178	−0.140	−0.145
40 h	0.567	0.513	0.539	0.580	0.585	0.583	0.575	−0.069	−0.172	−0.070	−0.128	−0.097	−0.098
48 h	0.503	0.463	0.488	0.510	0.514	0.513	0.498	−0.053	−0.131	−0.048	−0.099	−0.076	−0.075
Starch degradability
24 h	0.854 *	0.905 *	0.884 *	0.834 *	0.832 *	0.846 *	0.830 *	−0.808 *	−0.900 *	−0.818 *	−0.853 *	−0.823 *	−0.828 *
48 h	0.788 *	0.857 *	0.835 *	0.761 *	0.756 *	0.773 *	0.750 *	−0.873 *	−0.938 *	−0.873 *	−0.905 *	−0.880 *	−0.882 *

* p < 0.05.

, the CHO spectral peak area, area and height of the first, second, and third CHO peak were positively correlated with GP before 32 h and SD (24 and 48 h) (p < 0.05). The heights and areas of amide I and II, and α-helix and β-sheet height were negatively correlated with SD (24 and 48 h) (p < 0.05).

4. Discussion

4.1. Starch Properties

Starch properties can affect digestion by the animal. Starch is easily hydrolyzed by amylase after gelatinization [4]. The gelatinization of starch during the corn processing is equivalent to the pre-digestion of starch in vitro, which can improve the digestion of feed in vivo [34]. As expected, thermal processing (S and E) promoted the starch gelatinization compared with the simple mechanical processing (G). This could be caused by the disruption of the starch granular structure by the high temperature and pressure. Previous studies obtained similar results when comparing SFC with pelleted corn [35] and SFC with GC [2]. We also observed that the alteration by E was significantly greater than that by SF. This indicated that the degree of destruction of the structure of the starch granules by E was larger than that by SF, and E may be a more effective strategy for increasing starch gelatinization. This could presumably relate to the higher temperature, higher mechanical shear rate and shorter treatment time of E compared with SF [1]. Starch crystallinity can be affected by gelatinization. During the starch gelatinization process, high temperature and pressure induce the breaking of the chemical bonds between molecules, allowing the hydrogen bonding sites to engage more water, which causes a loss of crystallinity [1]. The starch crystallinity was the lowest for EC, followed by SFC and then GC in our study, supporting the above points. Little or no previous work has been done to compare the effect of SF and E on the starch properties of corn.

4.2. Nutrient Profiles

In this experiment, the higher measured starch of SFC and EC indicated that thermal processing could improve the starch content more than by simple mechanical process, possibly because the degree of gelatinization of starch was increased by thermal processing, which resulted in more starch acid hydrolyzing into glucose. Ma et al. [36], Xu et al. [19] and Rahimi et al. [2] also reported that SF increased the starch content measured in corn. The decreased content of CP in EC and SFC could be due to the destruction of the protein matrix. The decrease of EE in SFC and EC might be due to the formation of a complex between fat and gelatinized starch during heat processing, which would mean that fat could not be extracted by organic solvents [37]. Similarly, Cornoa et al. [38] reported lower CP and EE in SFC compared with untreated corn. Amornthewaphat et al. [39] observed that the extrusion process significantly decreased the EE of corn. However, Rahimi et al. [2] found no difference in the EE content between SFC and GC. We also found that the decreasing effect of SF on CP and EE content was greater than that of E. It has been shown that feed processing could cause the breakage of cell wall components [40]. The lower NDF, ADF of SFC and EC suggested that the degree of damage to the cell wall of corn caused by SF and E was greater than G and would be beneficial for the utilization of carbohydrates by the consumer. These results agreed with those reported by Amornthewaphat et al. [39], who found that the crude fiber of corn was decreased by E.

4.3. In Vitro Fermentation

GP is an important index reflecting the fermentation extent of feed in the rumen [26]. The increased GP before 32 h of incubation for the EC and SFC in comparison with GC indicated that SF and E were more beneficial to the fermentation of organic matter in the rumen than G. This could be caused by the high starch and gelatinized starch content caused by E and SF. Starch is the easily fermentable carbohydrate in the rumen and starch gelatinization promotes its exposure to microorganisms and thus leads to increased GP [11]. Similar results had been reported by Rahimi et al. [2] who observed that SFC produced more GP than untreated corn. DePeters et al. [41] reported that SFC increased the GP during the initial hours of incubation. However, Karami et al. [11] observed different results, and the GP of untreated corn at 24 and 48 h of incubation were greater than SFC and EC. The discrepancies between the results could be related to differences in the fermentation system, processing techniques or conditions used during thermal processing. The greater GP of EC than SFC before 24 h of incubation in this study showed that E was more effective in promoting the fermentation of corn in vitro. This also contributed to the difference in starch gelatinization caused by the two processes. However, Karami et al. [11] reported that the GP of SFC was significantly higher than EC at most incubation periods. The results of the GP parameters showed that E and SF promoted the rate constant of slow gas production of corn. Similar results were reported by Rahimi et al. [2]. The negative value of a indicated that all three types of corn had a GP lag effect. The decrease of b and (a + b) for SFC and EC is difficult to explain. More studies need to be carried out to find the reason.

In the current study, the higher total VFA of EC and SFC were consistent with those of GP. The results might be attributable to the greater starch content and greater starch digestion in the rumen which could supply more fermentable substrate for ruminal fermentation [42]. VFA is the main fermentation product of carbohydrate, mainly including acetic acid, propionic acid and butyric acid, and is the main energy source for ruminants. Propionic acid is the main VFA that supplies greater energy on a molar basis than acetic acid and butyric acid [43]. Propionic acid is the precursor of endogenous glucose synthesis in ruminants. The increase in propionic acid in the rumen is beneficial to glucose synthesis and energy deposition and thus to the fattening of the meat animal. The highest propionic acid at 24 h for EC showed that the EC may be the best for the fattening of lambs or calves. This is likely due to the highest starch content and starch gelatinization caused in the processing of EC. A high starch diet tends to ferment toward propionic acid [44]. However, the lack of difference for propionic acid at 48 h between the three treatments is likely due to the similar content of organic matter of the three kinds of processed corn which came from the same raw material. Makizadeh et al. [10] also observed higher total ruminal VFA concentrations and greater molar percentages of propionic acid in calves fed SFC starter compared with those fed GC starter during both pre- and post-weaning periods. However, Shabi et al. [45] reported that ruminal VFA concentrations were similar in the EC diet as compared with the GC diet. Our results also indicated that EC was better than SFC in promoting in vitro ruminal fermentation.

Ammonia nitrogen is the decomposition product of nitrogen-containing substances in feed and is the main nitrogen source for rumen microbial protein synthesis [10]. In this study, the decrease of protein content in the two types of thermally processed corn may be the first reason leading to the low NH₃-N in the artificial rumen. In addition, the high starch and gelatinized starch content in these two types of processed corns increased the easily degradable energy and promoted the utilization of nitrogen by rumen microorganisms, which further decreased the ammonia nitrogen content in the rumen. Similarly, Casper et al. [46] and Makizadeh et al. [10] observed that SF significantly reduced the ruminal NH₃-N concentrations of corn. A decrease in ruminal ammonia was observed when cows were fed the EC diets compared with GC diets [45]. Our results also showed that SF and E had no different effects on in vitro NH₃-N concentrations.

Starch digestion was positively related to the percentage of starch gelatinization [47]. The higher SD (24 and 48 h) of SFC and EC in the present study indicated that SF and E were more beneficial to the degradation of starch in the rumen than G, possibly due to the disruption of the germ layer and protein matrix, increasing the availability of starch molecules for degradation by amylase, destruction of amylase inhibitors, fragmentation of cell walls and to increase the gelatinization of starch during processing [1]. This also could be caused by the starch swelling during SF and E because starch swelling is the preliminary stage of enzymatic degradation in the gastrointestinal tract of animals. Xu et al. [19] and Rahimi et al. [2] also reported higher SD for SFC than for untreated corn and GC. In addition, the SD of EC was greater than that of SFC, which indicated that E was more effective for starch digestion in the rumen than SF. This change is consistent with starch gelatinization and crystallinity.

The rumen fluid was obtained from sheep and in vitro fermentation results showed that corn processing could affect its fermentation and digestion in the rumen of sheep, and E may be the best way of processing corn for fattening lambs, followed by SF and then G.

4.4. Carbohydrate and Protein Molecular Structure

The evaluation of the nutritional value of feed is usually carried out by “wet” chemical analysis and animal testing. Traditional wet chemical analysis is based on the chemical reaction of substances to determine content, and it usually destroys the inherent molecular structure of the feed [14]. Animal tests are time-consuming, laborious, and costly. Attenuated total reflectance Fourier-transformed infrared (ATR-FT/IR) molecular spectroscopy is a direct, rapid, non-destructive, and non-invasive bioanalysis technique [12].

Each biomolecule has its own specific molecular structure, and its unique conformation produces a specific infrared spectrum [48]. The spectral region of carbohydrates is mainly caused by C-O vibration and C-O-H deformation [49]. In proteins, the functional groups are peptide bonds (-CO-NH-), which can be divided into two types according to their composition: amide I and amide II. The amide I bond is composed of C=O stretching vibrations (80%) and C-N stretching vibrations. The amide I band is used to identify secondary structures (α-helix and β-sheet) of proteins [50]. The amide II band is mainly composed of N-H bending vibration (60%) and C-N stretching vibration (40%) [51].

In this experiment, the carbohydrate total CHO spectral peak areas, spectral peak height, and area of the first, second, and third CHO peaks of EC were the highest, followed by SFC and then GC, which implied that EC had the highest carbohydrate molecular absorption. Spectral peak height and area of the CHO third peak was mainly associated with starch enrichment [52], the change of which between the three treatments was consistent with the starch content change in this study. Xu et al. [19] reported similar results in a comparison of SFC and untreated corn.

The lower spectral peak heights and areas of amides I, II, α-helix, and β-sheet of SFC and EC indicated that, compared with G, high temperature and pressure processing had a greater effect on the protein molecular structure of corn. We also observed that the effect of SF was greater than E. This is consistent with the change of CP content of three treatments. Xu et al. [19] (with untreated corn and SFC), Peng et al. [53] (with camellia seeds), and Samadi and Yu [54] (with soybean) reported similar results.

To our knowledge, no study has been carried out yet to explore the structural-chemical features (bonding and functional group characteristics) of EC within cellular dimensions, therefore no comparison could be made in the present study.

Previous studies showed that ATR-FT/IR molecular spectroscopy was more sensitive than wet chemical analysis [19,55]. Our results confirmed that theory. We found that the spectral structures of carbohydrate and protein molecules differed between SFC and EC, whereas no difference was detected for the content of starch, NDF and ADF between SFC and EC.

The molecular structure of feed is related to its nutritional value and fermentation characteristics [56]. Xu et al. [19] reported that a larger area of the first, second, and third CHO peaks indicated higher degradation of starch in the rumen. Our results confirmed the theory. We observed that the molecular structure of carbohydrate was positively correlated with the ruminal fermentation and SD of corn. Higher peak height and area of the CHO spectral region indicated faster carbohydrate fermentation in the rumen (especially in the early stage), more SD, and greater GP. In contrast, higher height and area of amide I and amide II and higher α-helix and β-sheet height indicated lower SD. Xu et al. [19] obtained similar results with untreated corn and SFC. However, Chen et al. [57] found no correlation between the total spectral peak area of CHO and DM degradation for DDGS. Further study is required to define the relationship between molecular structure and the nutritional value of feed.

One important objective of the current study was to make a clear comparison of the effects of SF and E on the rumen fermentation and the nutritive value of corn. The outcomes showed a significant difference between the two thermal processes for starch properties, nutrients profiles, in vitro rumen fermentation, carbohydrates and protein molecular structure. Further study is required to define the effects of these two ways of thermally processing corn on the diets of the ruminant animal.

Compared with G, the processing costs of SF and E are high. At present, SFC and EC are used in the starter feed of ruminant animals, which is generally used in the period of young ruminants and not throughout the whole fattening period. The increase in the processing cost of SFC and EC does not result in an uneconomical cost–benefit ratio in ruminant production. Further comparative animal testing is required to define the cost–benefit ratio of the three processing methods of corn in the diets of ruminant animals.

5. Conclusions

The results indicated that the higher starch gelatinization and lower starch crystallinity of thermal processing induced by the high temperature and pressure enabled more fermentation and digestion in the artificial rumen. Most changes caused by E were greater than those by SF. The carbohydrate and protein molecular structures were correlated with the nutritional characteristics of corn. For ruminant animals, especially in their early growth stage, E is probably the most effective processing method for corn.