Nutritional values of cottonseed meal from different sources fed to gestating and non-pregnant sows

Abstract

This study set out to determine the apparent total tract digestibility (ATTD) of the nutrients and energy in six cottonseed meal (CSM) feedstuffs fed to pregnant and non-pregnant sows. The six types of CSM were: two expelled CSMs with crude protein (CP) levels of 40.67% and 44.64%, and four solvent-extracted CSMs with CP levels of 45.18%, 51.16%, 56.44%, and 59.63%. Fourteen gestating sows (at the fourth parity with body weights of 220.6 ± 18.4 kg at days 30 of gestation) and 14 non-pregnant sows (after the third parity with body weights of 219 ± 14.6 kg) were assigned to a replicated 7 × 3 Youden square design with seven diets and three periods. The seven diets included an entirely corn-based diet and six diets each containing 20.0% of the six CSMs tested. Each period included a 5-d acclimation to the experimental diets, followed by a 5-d period during which urine and feces were collected. Significant differences were found among the six CSM diets, regardless of reproductive stage, regarding 1) the ATTD of neutral detergent fiber (NDF) (P < 0.05) and 2) the ATTD of dry matter (DM), organic matter (OM), and CP and the gross energy (GE) (P < 0.01). Non-pregnant sows had a greater ATTD of OM and CP (P < 0.01) compared with gestating sows. The digestible energy (DE) and metabolizable energy (ME) of the six CSM samples ranged from 12.48 to 17.15 MJ/kg and 11.35 to 15.88 MJ/kg, respectively, for non-pregnant sows, and from 12.86 to 16.41 MJ/kg and 12.43 to 14.72 MJ/kg, respectively, for gestating sows. However, the DE, ME, and ME:DE ratios of each CSM were similar between gestating and non-pregnant sows. DE and ME were negatively correlated with NDF and ADF, respectively, but were positively corrected with CP level (P < 0.01). Collectively, the DE, ME, and nutrient digestibility of CSM varied greatly according to the chemical compositions, and CSMs with higher protein and lower fiber levels had greater DE and ME levels.

Better knowledge of the nutritional value of cottonseed meals would allow more precise diet formulations, decrease feed costs, and stimulate its use as a high-quality protein source for sows.

Introduction

Cottonseed meal (CSM) is a by-product of cottonseed processing and provides a superior source of protein (Cheng et al., 2020). Global cottonseed production is forecast to reach about 43.8 million metric tons by 2021/2022 (Statista, 2022 accessed on 9 February 2022), and the use of CSM as a protein source in swine production has attracted growing attention due to the shortage of protein feed. In recent years, studies have been conducted on the nutritive value of CSM as feed for nursery pigs (Wang et al., 2019) and growing pigs (Ma et al., 2018; da Silva et al., 2021). Growing pigs and sows were assigned the same effective feed energy requirement values by the NRC (2012), which may not be accurate (Le Goff and Noblet, 2001; Lowell et al., 2015; Casas and Stein, 2017). The ability of swine to digest the nutrients in a given diet may be impacted by reproductive stage, variations in body weight, and feeding level (Stein et al., 2001; Casas and Stein, 2017). Additionally, gestating sows may have greater energy utilization than the empty sows due to the pregnancy anabolism (Miller et al., 2019). However, the nutritive value of CSM for both gestating and non-pregnant sows remains uncertain.

Different processing methods can influence the nutritive value of CSM (Ma et al., 2018; Satankar et al., 2021). The nutritive value of the sample CSM varied greatly according to the source of supply, and showed differences in chemical composition such as the levels of free gossypol (FG), tannin, phytic acid, fiber, and other components (Sarwar et al., 2012; Ma et al., 2018). FG can decrease the availability of amino acids in CSM because of its affinity for proteins and amino acids (Nicholson, 2012), and negatively affects the functioning of the reproductive tract of pigs (Randel et al., 1992; Wani and Nazir, 2022). To accurately formulate diets for sows, the nutritive value of CSM must be evaluated using sows as the experimental model. This study therefore aimed to determine the nutritional value of CSM in sows at various reproductive stages, and also to determine the relationship between chemical composition and available energy and nitrogen utilization of each CSM tested.

Materials and Methods

The protocols used in this study were reviewed and approved by the Institutional Animal Care and Use Committee of Sichuan Agricultural University (SICAU 20210038).

Preparation of CSM samples

Six CSM samples were collected from different areas of Xinjiang province and processed by the Tycoon Co. Ltd. (Xinjiang, China) using different processing methods (Table 1). Two CSMs were prepared using an expulsion process (CSM1and CSM2), three were prepared using a pre-press and solvent-extraction process (CSM3, CSM4, and CSM5), and one was derived by direct solvent-extraction (CSM6). Summaries of the processing methods of each sample are given in Table 1.

Table 1.

Sources of cottonseed meal

Ingredients1	Processing techniques	Origin in China
CSM1	Expelled	North Xinjiang
CSM2	Expelled	South Xinjiang
CSM3	Pre-press and solvent-extracted	North Xinjiang
CSM4	Pre-press and solvent-extracted	North Xinjiang
CSM5	Pre-press and solvent-extracted	North Xinjiang
CSM6	Direct solvent-extracted	South Xinjiang

¹ Ingredients: The six CSM (cottonseed meal) samples are numbered from low to high content of crude protein level.

Animals and experimental design

Fourteen gestating Landrace × Yorkshire crossbred (LY) sows (parity three; 220.6 ± 18.4 kg, at day 30 of gestation) and 14 non-pregnant LY sows (219 ± 14.6 kg, after the lactation period of the third parity) were assigned to a replicated 7 × 3 Youden square design with seven diets and three periods. The diets comprised a corn-based diet and six corn/CSM diets, each containing 20.0% of the CSM to be tested (Table 2). Before beginning the metabolic experiment, a chemical analysis of the CSM samples was carried out (Tables 3 and 4). The composition and nutrient levels of the experimental diets are shown in Table 5. We prepared two batches of the various diets, one to feed to the gestating sows and the other for the non-pregnant sows.

Table 2.

Composition of the experimental diets (on an as-fed basis, %)

Ingredients	Basal diet	Test diet
Corn	96.37	76.59
CSM1	0.00	20.00
Dicalcium phosphate	1.52	1.26
Limestone	0.96	1.00
Choline chloride (50%)	0.25	0.25
NaCl	0.40	0.40
Vitamin-mineral premix2	0.50	0.50
Total	100.00	100.00

¹CSM, cottonseed meal.

²Premix provided per kg of complete diet: 6,000 IU of vitamin A; 2,000 IU of vitamin D3; 80 IU of vitamin E; 3.8 mg of Vitamin K; 2.0 mg of vitamin B1; 6.0 mg of riboflavin; 4.0 mg of Vitamin B6; 0.02 mg of Vitamin B12; 26.0 mg of niacin; 18.0 mg of pantothenic acid; 3.2 mg of folic acid; 0.4 mg of biotin; 100 mg of Fe; 20 mg of Cu; 100 mg of Zn; 25 mg of Mn; 0.4 mg of I; and 0.30 mg of Se.

Table 3.

Composition of the six cottonseed meal samples (%, on an as-fed basis)

Items1	CSM2						Mean	Maximum	Minimum	CV%
	CSM1	CSM2	CSM3	CSM4	CSM5	CSM6
DM3	90.52	87.36	88.67	89.38	88.50	95.25	89.94	95.25	87.36	3.11
GE, MJ/kg	17.31	16.69	16.84	17.51	17.42	18.58	17.39	18.58	16.69	3.84
CP3	40.67	44.64	45.18	51.16	56.44	59.63	49.62	59.63	40.67	14.91
EE	1.43	1.04	0.85	1.14	2.00	1.64	1.35	2.00	0.85	31.48
CF	15.92	11.07	14.98	10.02	4.93	3.23	10.03	15.92	3.23	51.35
NDF3	28.76	24.20	25.56	20.27	11.31	5.37	19.24	28.76	5.37	47.14
ADF3	17.94	13.94	13.46	12.48	2.79	1.62	10.37	17.94	1.62	63.69
Ash	5.86	6.23	5.81	6.48	7.02	7.09	6.42	7.09	5.81	8.64
Carbohydrates
TDF	45.00	32.10	34.30	32.10	36.50	22.20	33.70	45.00	22.20	21.94
SDF	5.30	4.33	4.38	5.22	5.03	4.43	4.78	5.30	4.33	9.41
IDF	39.70	27.80	29.90	26.90	31.50	17.80	28.93	39.70	17.80	24.58
Antinutritional factors
Free gossypol, mg/kg3	1077.75	546.85	501.95	535.15	632.30	266.15	593.36	1077.75	266.15	45.04

¹ ADF, acid detergent fiber; CF, crude fiber; CP, crude protein; DM, dry matter; EE, ether extract; GE, gross energy; IDF, insoluble dietary fiber; NDF, neutral detergent fiber; SDF, soluble dietary fiber; TDF, total dietary fiber.

² CSM (cottonseed meal) samples are described in Table 1.

³ These data were also presented in a companion study for a better illustration of CSM.

Table 4.

Mineral and vitamin concentrations in the six cottonseed meal samples (mg/kg, on an as-fed basis)

Items	CSM1						Mean	Maximum	Minimum	CV%
	CSM1	CSM2	CSM3	CSM4	CSM5	CSM6
Macro minerals
Calcium, %	0.23	0.24	0.21	0.20	0.24	0.24	0.23	0.24	0.20	7.73
Total phosphorus, %	1.01	1.15	1.08	1.23	1.35	1.35	1.20	1.35	1.01	11.76
Potassium, %	1.48	1.44	1.47	1.59	1.40	1.59	1.50	1.59	1.40	5.26
Sodium, %	ND2	ND	ND	ND	ND	ND	ND	ND	ND	ND
Sulfur, %	0.05	0.07	0.07	0.06	0.08	0.06	0.07	0.08	0.04	16.14
Micro minerals
Copper	9.96	11.95	10.79	10.70	11.35	11.70	11.08	11.95	9.96	6.63
Iron	81.00	110.5	119.5	105.0	101.5	120.0	106.25	120.0	81.00	13.60
Manganese	12.60	12.30	11.65	11.75	12.15	12.10	12.09	12.60	11.65	2.91
Selenium	0.03	0.04	0.06	0.04	0.07	0.03	0.05	0.07	0.03	36.51
Zinc	42.10	53.90	43.90	46.05	52.60	53.65	48.70	53.90	42.10	10.88
Vitamin
Thiamine	1.60	2.89	2.52	3.72	4.39	5.10	3.37	5.10	1.60	38.10
Riboflavin	0.18	0.26	0.37	0.58	0.98	0.62	0.50	0.98	0.18	58.72
Choline	1613.00	2770.00	3221.00	2656.50	2605.00	3958.50	2804.00	3958.50	1613.00	27.57
β-Carotene	0.04	0.08	0.05	0.04	0.07	0.05	0.05	0.08	0.04	29.88

¹ CSM (cottonseed meal) samples are described in Table 1.

² ND, not detected, the minimum detection level of sodium was 500 mg/kg.

Table 5.

Nutrient composition of the experimental diets (%, on an as-fed basis)

	Non-pregnant sows diets							Gestating sows diets
	CSM1 diets							CSM diets
Item2	CSM1	CSM2	CSM3	CSM4	CSM5	CSM6	Basal diet3	CSM1	CSM2	CSM3	CSM4	CSM5	CSM6	Basal diet
DM	86.80	86.38	86.56	86.71	86.53	87.28	86.03	85.84	85.51	85.25	86.33	85.81	85.68	86.03
CP	15.01	15.90	15.87	16.92	18.48	18.97	8.12	13.13	14.08	15.10	15.18	16.11	16.84	6.97
EE	2.84	2.35	2.34	2.34	2.56	2.41	2.18	1.87	1.80	1.69	1.96	2.13	1.44	2.03
NDF	13.10	11.68	12.34	11.84	8.65	7.79	8.30	12.11	10.47	11.48	10.55	8.49	7.96	8.05
Ash	4.27	4.24	4.06	4.37	4.28	4.37	3.33	4.01	4.00	4.12	4.19	4.40	4.39	3.28
GE, MJ/kg	15.47	15.47	15.49	15.51	15.59	15.72	15.23	15.27	14.91	15.04	15.49	15.13	15.28	15.14

¹ CSM (cottonseed meal) samples are described in Table 1.

² ADF, acid detergent fiber CF, crude fiber; CP, crude protein; DM, dry matter; EE, ether extract; GE, gross energy; NDF, neutral detergent fiber

³ Basal diet: a corn-based diet shown in Table 2.

Both gestating and non-pregnant sows were fed a total of 3.0 kg/d of the respective experimental diets, and the feeding level energy was estimated using equation 1.5 × 100 kcal ME per kg BW^0.75, where ME is metabolizable energy and BW is the body weight (Velayudhan et al., 2019). Two equal-sized meals were offered at 0800 hours and 1500 hours. Water was provided ad libitum via a drinking nipple. The room temperature of the metabolic test room was maintained at 20 ± 2 °C.

Sample collection

The metabolic trial comprised a 5-d acclimation period in a metabolic cage (2.10 m × 0.97 m × 1.20 m), followed by collection of feces and urine over the next five consecutive days. A color marker (ferric oxide) was added to the morning meal on days 6 and 11. Fecal collection started and stopped with the first and second appearances of the colored feces. After collection and weighing of the samples, 1 mL of 10% HCl was added for every 10 g of fecal sample, and the samples were then immediately kept at −20 °C to prevent microbial fermentation and nitrogen loss. We implemented procedures to avoid feces and urine contamination, following the methods described in our previous study (Yang et al., 2021). Urine collection was initiated at 0900 hours on day 6 and ceased at 0900 hours on day 11. Urine was collected in a bucket containing 50 mL 6 N HCL (a preservative), emptied once daily, and the weight of the urine was recorded. Then, 20% subsamples of the urine samples were stored at −20 °C (Lowell et al., 2015). Feed refusals and spillages were collected, dried, and weighed daily. At the end of the experiment, urine and feces samples were thawed and mixed, respectively. Finally, a portion of the feces samples was dried and crushed, along with part of the urine samples, and sent to the laboratory for analysis (Yang et al., 2021).

Chemical analysis and calculations

All of the CSM diets and feces samples were analyzed for dry matter (DM), ash, ether extract (EE), and minerals (AOAC, 2007). Crude fiber (CF), neutral detergent fiber (NDF), and acid detergent fiber (ADF) levels were determined using ­filter bags and a fiber analyzer (Fiber Analyzer, Ankom ­Technology, Macedon, NY, USA) following a modification of the procedure described by Van Soest et al. (1991). The gross energy (GE) of the CSM samples, diets, feces, and urine were determined using an automatic isoperibol oxygen bomb calorimeter (6400 Isoperibol Calorimeter, Parr Instrument Co., Moline, IL, USA). Total (N) content of the samples was determined using the micro Kjeldahl method (AOAC 2007) and crude protein (CP) was calculated as N × 6.25. Total dietary fiber (TDF), insoluble dietary fiber (IDF), and soluble dietary fiber (SDF) were analyzed using a combination of enzymatic and gravimetric procedures (Zhuo et al., 2020). The FG content was determined using the method described by de Cássia Romero et al. (2020). The digestible energy (DE) and ME values, and the apparent total tract digestibility (ATTD) of GE, DM, organic matter (OM), CP, and NDF of the six CSMs were calculated using the difference method. The digestibility of the component in the test ingredient ($\mathrm{D}\mathrm{t}\mathrm{i}$) is calculated as follows, on a dry matter basis (Adeola, 2001):

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{D}\mathrm{t}\mathrm{i},\mathrm{\%}=100\times \frac{\left[(\mathrm{D}\mathrm{t}\times \mathrm{D}\mathrm{t}\mathrm{p})\text{-}(\mathrm{D}\mathrm{b}\times \mathrm{D}\mathrm{b}\mathrm{p})\right]}{\mathrm{D}\mathrm{t}\mathrm{i}\mathrm{p}}}\end{array}}$$

where $Dt$ is the digestibility (%) of the component in the total diet (basal diet plus the test feedstuff); $Db$ is the digestibility (%) of the component in the basal diet; $Dbp$ is the proportion (%) of the component in the total diet contributed by the basal diet; $Dtip$ is the proportion (%) of the component in the total diet contributed by the test feedstuff; $Dtp$ is the sum of $Dbp$ and $Dtip$.

Statistical analysis

We used PROC BOXPLOT and UNIVARIATE in SAS 9.4 (SAS Institute Inc., Cary, NC, USA) to confirm the outliers and normality of the data. The data on digestibility were analyzed using PROC MIXED, with sow as the experimental unit. The model included the fixed effect of CSM or diet, and the random effects of sows, reproductive stage, and period. The LSMEANS statement was used to calculate the mean least squares for each treatment. The significance of differences among major effects was tested using Tukey’s multiple range test. Differences were considered significant if P < 0.05, and tendencies were confirmed if 0.05 ≤ P < 0.10.

The relationship between energy content and chemical composition was analyzed using PROC CORR in SAS. Using the stepwise regression procedure (PROC REG) in SAS (Kaps and Lamberson, 2017), we developed DE and ME prediction equations for gestating and non-pregnant sows. The R², P-value, root mean square error (RMSE), and Akaike’s information criterion (AIC) were used to identify the best-fit equations. Equations with the greatest R² and the least RMSE were considered to be the best fit.

Results

Chemical compositions of the CSM samples

The chemical compositions of the six CSM samples are shown in Tables 3 and 4. Among them, CSM 1 had the greatest CF, NDF, and ADF contents and the lowest CP content. CSM 5 and CSM 6 had greater CP (56% to 60%) and ash (> 7.0%) contents, while their CF, NDF, and ADF contents were the lowest. The GE content showed no significant variation, while the coefficients of variation (CV) of TDF, SDF, EE, CF, NDF, ADF, thiamine, riboflavin, choline, and β-carotene values exceeded 10% (Tables 3 and 4). The FG content varied greatly, ranging from 266.15 mg/kg in CSM1 to 1077.75 mg/kg in CSM6 (on an as-fed basis), while the FG contents of CSM 2, CSM 3, CSM 4, and CSM 5 were similar and ranged from 501.95 to 632.30 mg/kg.

Energy content and energy digestibility

By calculating the nutrient content of the diets (Table 5) and the nutrient levels excreted in the feces and urine, the energy balance and N balance could be assessed, and are shown in Table 6. Sows fed with CSM substitution diets excreted more energy in the feces compared with sows fed the basal diet (P < 0.01). No significant differences among the seven diets were observed in the GE intake and energy excreted in the urine. Gestating sows had lower (P < 0.01) levels of energy excreted in the urine and greater net GE utilization compared with non-pregnant sows (P < 0.01). Energy excreted in the feces was greater in non-pregnant sows than in gestating sows. The basal diet had greater retained GE levels (P < 0.01) and net GE utilization than the other diets, in both gestating and non-pregnant sows.

Table 6.

Energy and N balances for sows fed on the basal diet and cottonseed meal diets

	Energy balance (MJ/d)						N balance (g/d)
Item	DMI (kg/d)	GE intake	GE in feces	GE in urine	GE retained	GE net utilization3, %	N intake	N in feces	N in urine	N retained	N digested4	Net N utilization, %	N retention5, %
Non-pregnant sows
Basal diet1	2.58	45.70	3.52c	0.75	41.27ab	90.32ab	38.95	4.89c	15.86b	17.98b	34.06f	46.14	52.93
CSM12	2.60	45.80	5.72a	1.68	38.41c	83.85c	72.04	10.58ab	27.72ab	33.74ab	61.46e	46.83	54.95
CSM2	2.59	46.40	6.16a	1.81	38.43c	82.83c	76.30	10.86a	29.09ab	33.04ab	65.44d	47.70	55.59
CSM3	2.60	46.47	5.82a	0.82	39.23bc	84.43bc	76.16	10.68a	29.07ab	36.42a	65.48d	47.81	55.44
CSM4	2.60	46.54	5.46a	1.88	39.39bc	84.65bc	81.22	10.34ab	37.26a	33.62ab	70.88c	41.40	47.42
CSM5	2.60	46.78	4.35b	1.60	40.84ab	87.30ab	88.69	9.80ab	29.37ab	49.52a	78.89b	55.84	62.60
CSM6	2.62	47.16	4.13bc	1.57	41.46a	87.92a	91.04	9.00b	40.43a	41.61a	82.04a	45.70	50.74
Mean	2.60	46.41	5.02	1.44	39.86	85.90	74.91	9.46	29.86	35.13	65.46	47.28	52.93
SEM	–	–	0.06	0.09	0.14	0.30	–	0.14	1.32	1.48	0.14	1.94	2.18
Gestating sows
Basal diet	2.58	45.43	3.16c	0.80	41.50a	91.35a	33.45	3.92c	14.36b	15.18c	29.54f	45.38	51.44
CSM1	2.58	45.80	5.72a	0.74	39.35cd	85.92d	63.05	9.83ab	21.03ab	32.19ab	53.22e	51.05	60.42
CSM2	2.57	44.72	5.35b	0.84	38.43e	85.93d	67.56	9.72b	28.79a	28.90b	57.84d	42.78	50.06
CSM3	2.56	45.12	5.82a	0.70	38.85de	86.51cd	72.46	10.47ab	33.16a	29.07b	62.19c	40.12	46.72
CSM4	2.59	46.49	5.55a	1.14	39.80bc	85.61d	72.87	10.76a	29.00a	33.27ab	62.29c	41.40	53.45
CSM5	2.57	45.41	4.15b	1.13	40.13b	88.38b	77.31	9.70b	29.37a	38.24ab	67.60b	49.47	56.58
CSM6	2.57	45.83	4.46b	1.03	40.26b	87.84bc	80.82	9.59b	30.03a	41.75a	71.23a	51.66	58.23
Mean	2.57	45.54	4.86	0.91	39.76	87.36	66.79	9.16	26.53	31.26	57.69	46.63	51.44
SEM6	–	–	0.06	0.09	0.14	0.30	–	0.15	1.32	1.49	0.14	1.94	2.18
P-value
Non-pregnant sows	–	–	<0.01	0.08	<0.01	<0.01	–	<0.01	<0.01	<0.01	<0.01	0.65	0.67
Gestating sows	–	–	<0.01	0.05	<0.01	<0.01	–	<0.01	<0.01	<0.01	<0.01	0.29	0.33
Phase	–	–	<0.05	<0.01	0.59	<0.01	–	0.07	0.06	0.03	<0.01	0.78	0.92

¹Basal diet: a corn-based diet.

²The six CSM (cottonseed meal) samples are described in Table 1.

³Net utilization, % = retained/intake × 100

⁴N digested = N intake − N in feces

⁵N retention, % = retained N/digested N × 100

⁶SEM, standard error of means. Means are the least square means (n = 6).

^{a, b, c, d, e, f} Means within the same row with different superscript letters are significantly different (P < 0.05).

The DE, ME, ME:DE ratio, and ATTD of the GE values of the six CSM samples are shown in Table 7. In gestating sows, the DE, ME, and ATTD of the GE values (on a DM basis) varied from 13.11 to 16.71 MJ/kg, 12.43 to 14.72 MJ/kg, and 68.39% to 83.43%, respectively. In non-pregnant sows, the DE, ME, and ATTD of the GE values (on a DM basis) varied from 12.48 to 17.14 MJ/kg, 11.30 to 15.88 MJ/kg, and 65.26% to 87.89%, respectively. However, CSM5 and CSM6 had greater DE, ME, and ATTD of GE values, regardless of reproductive stage (P < 0.01). With the exception of the ME:DE ratio, these values differed significantly among the various CSM samples in gestating (P < 0.01) and non-pregnant sows (P < 0.01). Gestating sows showed different ME:DE ratios when fed with different diets; CSM1 and CSM3 had greater ME:DE ratios than CSM4 and CSM5 (P < 0.05).

Table 7.

Comparative nutrients and energy digestibility in cottonseed meal diets fed to sows (on a dry matter basis)

	Non-pregnant sows								Gestating sows
	CSM1 diets								CSM diets								P-value
Item2	CSM1	CSM2	CSM3	CSM4	CSM5	CSM6	Mean	SEM	CSM1	CSM2	CSM3	CSM4	CSM5	CSM6	Mean	SEM	Non-pregnant	Gestation	Phase
Energy values, MJ/kg
DE	12.48c	12.76c	13.37bc	14.61b	16.78a	17.14a	14.52	0.13	13.11b	13.38b	12.87b	13.91b	16.41a	16.71a	14.40	0.14	<0.01	<0.01	0.48
ME	11.30b	11.49b	11.83b	13.03b	15.88a	15.60a	13.19	0.17	12.57b	12.43b	12.44b	12.67b	14.68a	14.72a	13.24	0.17	<0.01	<0.01	0.79
ME:DE, %	92.91	92.06	92.51	89.28	94.43	90.97	92.02	0.74	95.99ab	94.48abc	96.60a	91.13bc	89.52c	91.72bc	93.24	0.75	0.52	<0.05	0.26
Digestibility coefficient, %
ATTD of DM	62.18d	61.46d	66.96cd	67.65bc	78.72ab	83.65a	70.60	0.71	65.43b	66.24b	63.10b	67.06b	78.26a	83.88a	70.61	0.73	<0.01	<0.01	0.98
ATTD of GE	65.26c	66.78c	70.39bc	74.54b	85.27a	87.89a	75.02	0.66	68.39b	70.02b	67.75b	70.98b	83.36a	83.43a	73.99	0.69	<0.01	<0.01	0.22
ATTD of OM	65.30c	66.03b	70.48bc	74.49b	85.68a	88.31a	75.05	0.74	67.27b	69.27b	66.44b	69.62b	82.43a	80.06a	72.52	0.76	<0.01	<0.01	<0.05
ATTD of CP	84.07d	84.97cd	85.34cd	87.53bc	90.16ab	91.89a	87.33	0.29	81.29c	83.65bc	84.04b	84.03b	86.96a	88.05a	84.67	0.30	<0.01	<0.01	<0.01
ATTD of NDF	37.20ab	31.15b	47.33ab	56.07ab	53.05ab	72.78a	49.66	3.77	37.68ab	29.19b	37.95ab	50.72ab	48.80ab	62.12a	44.62	3.81	<0.05	<0.05	0.27

¹ CSM (cottonseed meal) samples are described in Table 1.

² ATTD, apparent total tract digestibility; CP, crude protein; DM, dry matter; GE, gross energy; ME:DE = ME:DE ratio. NDF, neutral detergent fiber; OM, organic matter.

^{a, b, c, d, e, f} Means within the same row with different superscript letters are significantly different (P < 0.05).

Nitrogen balance and ATTD of nutrients

The N intake, N in feces, N in urine, N retained, N digested, and net N utilization of the basal and the various CSM diets are shown in Table 6. There were no significant differences in net N utilization among the basal and CSM diets, for both gestating and non-pregnant sows. Non-pregnant sows had greater amounts of digested N (P < 0.01) and retained N Compared with gestating sows (P < 0.05). The N in both the urine and feces differed significantly between the different diets (P < 0.01), but there were no significant differences by reproductive stage. Interestingly, a tendency was observed for greater N levels in both the urine (P = 0.06) and feces (P = 0.07) in non-pregnant sows than in gestating sows. The N retention ratio, calculated as retained N/digested N, were not affected by diets and reproductive phases (P > 0.05).

As shown in Table 7, for both gestating and non-pregnant sows, the ATTD of DM, OM, and CP were greater (P < 0.01) in CSM5 and CSM 6 than in the other CSM samples. In addition, non-pregnant sows had a greater ATTD of OM (P < 0.05) and CP (P < 0.01) than gestating sows. The greatest ATTD of NDF was found in CSM6, while the lowest was found in CSM2, in both gestating and non-pregnant sows. However, no difference was observed in the ATTD of NDF between gestating and non-pregnant sows.

Correlations and predictive equations

Table 8 shows the correlations between chemical composition and the DE and ME values of the six CSM samples at different reproductive stages. The CP content was negatively correlated with CF (r = −0.94; P < 0.01), NDF (r = −0.93; P < 0.01), and ADF (r = −0.94; P < 0.01) contents. The GE content was positively correlated with the CP content (r = 0.86; P < 0.05) and negatively correlated with the CF content (r = −0.82; P < 0.05). The DE value was negatively correlated with the CF (r = −0.95; P < 0.01), NDF (r = −0.97; P < 0.01) and ADF (r = −0.98; P < 0.01) contents and positively correlated with the CP content (r = 0.94; P < 0.01), GE content (r = 0.83; P < 0.05), and ME (r = 0.97; P < 0.01) value in non-pregnant sows. For gestating sows, the DE value was negatively correlated with the CF (r = −0.97; P < 0.01), NDF (r = −0.99; P < 0.01), and ADF (r = −0.98; P < 0.01) contents and positively correlated with the CP content (r = 0.94; P < 0.01), GE content (r = 0.83; P < 0.05), and ME (r = 0.97; P < 0.01) value.

Table 8.

Correlation coefficients between chemical characteristics and energy values of the experimental cottonseed meal samples (n = 6)¹ for sows

Item1	GE	CP	EE	CF	NDF	ADF	Ash	FG	TDF	SDF	IDF	DEN	MEN	DEG	MEG
GE	1
CP	0.86*	1
EE	0.71	0.61	1
CF	−0.82*	−0.94**	−0.67	1
NDF	−0.77	−0.93**	−0.65	0.97**	1
ADF	−0.74	−0.94**	−0.68	0.95**	0.98**	1
Ash	0.86*	0.91*	0.71	−0.9*	−0.8	−0.83*	1
FG	−0.31	−0.67	0.11	0.63	0.66	0.64	−0.46	1
TDF	−0.35	−0.64	0.03	0.7	0.73	0.65	−0.44	0.94**	1
SDF	0.3	−0.16	0.56	0.26	0.32	0.32	<0.01	0.73	0.72	1
IDF	−0.39	−0.65	0.02	0.71	0.74	0.66	−0.46	0.94**	1**	0.69	1
DEN	0.83*	0.94**	0.78	−0.95**	−0.97**	−0.98**	0.86*	−0.52	−0.55	−0.15	−0.56	1
MEN	0.79	0.93**	0.78	−0.86*	−0.89*	−0.95**	0.84*	−0.46	−0.42	−0.08	−0.43	0.97**	1
DEG	0.83*	0.94**	0.72	−0.97**	−0.99**	−0.98**	0.83*	−0.58	−0.65	−0.2	−0.66	0.99**	0.92**	1
MEG	0.78	0.85*	0.78	−0.94**	−0.97**	−0.93**	0.77	−0.46	−0.59	−0.18	−0.6	0.95**	0.85*	0.97**	1

¹ADF, acid detergent fiber; CF, crude fiber; CP, crude protein; DE_G, digestibility energy in gestating sows; DE_N, digestibility energy in non-pregnant sows; EE, ether extract; FG, free gossypol; GE, gross energy; IDF, insoluble dietary fiber; ME_G, metabolizable energy in gestating sows; ME_N, metabolizable energy in non-pregnant sows; NDF, neutral detergent fiber; SDF, soluble dietary fiber; TDF, total dietary fiber.

^* P < 0.05, ** P < 0.01.

The DE and ME prediction equations for the six CSM samples fed to gestating and non-pregnant sows are shown in Table 9. Based on the correlations, the DE, NDF, and ADF predictors accurately predicted the DE and ME equations. The best-fit equations for DE and ME, in both gestating and non-pregnant sows were: DE_N [MJ/kg DM] = 17.33 – 0.26 × ADF (R² = 0.97) and ME_N [MJ/kg DM] = 0.89 × DE_N + 0.46 (R² = 0.94); DE_G [MJ/kg DM] = 18.27 – 0.17 × NDF (R² = 0.98) and ME_G [MJ/kg DM] = 1.29 × DE_G + 0.82 (R² = 0.94).

Table 9.

Stepwise regression equations for predicting the DE and ME contents of cottonseed meal fed to gestating and non-pregnant sows

Item1	Prediction equations 2	R 2	RMSE	AIC	P-value
Non-pregnant sows
1	DE [MJ/kg DM] = 17.33 – 0.26 × ADF	0.97	0.41	–9.26	P < 0.01
2	DE [MJ/kg DM] = 0.24 + 0.25 × CP	0.89	0.74	–2.02	P < 0.01
3	DE [MJ/kg DM] = 0.99 GE – 0.29 × CF – 1.59	0.91	0.92	–1.16	P < 0.05
4	ME [MJ/kg DM] = 0.89 × DE + 0.46	0.94	0.51	–6.58	P < 0.01
5	ME [MJ/kg DM] = 16.50 – 0.16 × NDF	0.79	0.93	0.66	P < 0.05
6	ME [MJ/kg DM] = 15.81 – 0.23 × ADF	0.91	0.62	–4.18	P < 0.01
Gestating sows
1	DE [MJ/kg DM] = 18.27 – 0.17 × NDF	0.98	0.24	–15.74	P < 0.01
2	DE [MJ/kg DM] = 2.57 + 0.22 × CP	0.88	0.66	–3.33	P < 0.01
3	DE [MJ/kg DM] = 0.61 GE – 0.26 × CF + 5.71	0.94	0.56	–4.99	P < 0.05
4	ME [MJ/kg DM] = 1.29 × DE + 0.82	0.94	0.39	– 9.79	P < 0.01
5	ME [MJ/kg DM] = 16.32 – 0.14× NDF	0.93	0.42	– 8.76	P < 0.01
6	ME [MJ/kg DM] = 15.48 – 0.18 × ADF	0.86	0.62	– 4.19	P < 0.01

¹ ADF, acid detergent fiber; AIC, Akaike information criterion; CF, crude fiber; CP, crude protein; CSM, cottonseed meal; DE, digestible energy; GE, gross energy; ME, metabolizable energy; NDF, neutral detergent fiber; RMSE, root mean square error.

² The value of the energy and chemical composition in the equations as dry matter basis.

Discussion

In this study, the chemical composition of the CSM samples was generally quite different. The main reason for this is that the CSM samples were collected from various parts of southern and northern Xinjiang province, China, where the climatic conditions (soil, sunlight, and rainfall rate) for cotton cultivation varied significantly. For example, northern Xinjiang has a temperate continental arid and semi-arid climate, with the average annual temperature is −4 to 9 °C and the annual precipitation 150 to 200 mm. While southern Xinjiang has a warm temperate continental arid climate, and the average annual temperature is 7 to 14 °C and the annual precipitation is 25 to 100 mm. In addition, variations in cottonseed types, fertilization methods, processing techniques, and other factors also contributed to varied chemical compositions (Cheng et al., 2020; Świątkiewicz et al., 2016; Tanksley, 1981). The large CV of CF, ADF, and NDF may be due to the different amounts of cottonseed hulls and cotton linters retained during processing (Stein et al., 2006; Cheng et al., 2020). The results showed that the CP content increased when the CF, ADF, and NDF contents decreased, and these findings were consistent with previous studies (Li et al., 2012).

Although the CV of EE of the six varieties of CSM was large, the mean EE concentration was quite low, and similar to the findings of previous studies (Li et al., 2012; Ma et al., 2018), although lower than that reported by the NRC (2012). These results can be explained in that more lipids are extracted by solvent or mechanical extrusion processes, a common phenomenon in both canola and soybean meal processing (Adewole et al., 2016). Three different processing techniques were used to generate the CSMs tested in this study: extrusion-expulsion, pre-press solvent-extraction, and direct solvent-extraction. It has been reported that FG content is greater in direct solvent-extraction of CSM compared with either pre-press solvent-extraction or extrusion (Knabe et al., 1979), but this effect was not observed in our study. Different solvents were utilized in the extraction processes, and hexane can remove FG more effectively than acetone (Sultana, 2012). The amount of FG in CSM can also be affected by the heat treatment regime and, as temperature rises, FG forms complexes with other chemical components more easily (Gadelha et al., 2014). The calcium and total phosphorus contents were consistent with those previously reported (NRC, 2012; Ma et al., 2018). Different soil mineral contents, sources, storage conditions, and processing techniques may be responsible for the variations in carbohydrates, thiamine, riboflavin, choline, β-carotene, and micro-minerals (Thomas et al., 1998; Tanksley, 2017). The GE value of the CSM samples was within the range found in previous studies (Li et al., 2012; Ma et al., 2019), but was lower than that reported by the NRC (2012). Energy concentration will rise as CF content decreases and CP content increases (Hynes et al., 2016). The current findings showed the largest differences in DE and ME between the different CSM samples, ranging from a DE of 12.48 MJ/kg and ME of 11.30 MJ/kg for the CSM1 sample to a DE of 17.14 MJ/kg and ME of 15.60 MJ/kg for the CSM6 sample in non-pregnant sows. This great variation in available energy between different samples shows that we cannot ignore variations among the ingredients when formulating feedstuffs.

According to our recent research, energy utilization is greater in late-gestation than in mid-gestation sows (Wang et al., 2022), but no difference has been found between mid-gestation and non-pregnant sows in terms of efficiency of energy utilization for maintenance (Close et al., 1985) and the findings of this study were in agreement with this. The mean DE and ME values in gestating and non-pregnant sows were 14.40 MJ/kg and 14.52 MJ/kg and 13.24 MJ/kg and 13.19 MJ/kg, respectively, which is consistent with the results of Dong et al. (2020), but greater than in other studies of pigs in the growing or finishing stages (Li et al., 2012; NRC, 2012; Rodríguez et al., 2013). It is generally believed that sows have a more developed gastrointestinal tract than growing gilts, with a greater capacity for hindgut fermentation to harvest more energy from generally indigestible ingredients (Casas and Stein, 2017).

Considering the dietary ingredients, CSM with high fiber and low CP contents had lower DE and ME values. ­Previously published studies have shown that the DE and ME values of CSM fed to gestating sows were lower than those of soybean meal (Wang et al., 2022), and similar to those of canola meal (Lowell et al., 2015). The ATTD of GE decreased with increasing CF, NDF, and ADF content, which was consistent with previous reports. Pigs do not have fiber-digesting enzymes (Lindberg, 2014), so as cell wall components in the feed increase, indigestible components (such as lignin and non-starch polysaccharides) and nutrient encapsulation also increase, reducing energy utilization (Gutierrez et al., 2016; Jaworski and Stein, 2017). Adult sows have been reported to lose more energy through the urine compared to growing pigs, the difference in body weight and feeding level being a crucial indicator of the ME/DE ratio (Shi and Noblet, 1993; Le Goff and Noblet, 2001).

In this study, the net utilization of GE (87.36%) was greater in gestating sows than the non-pregnant sows (85.90%), and this might be attributed to the fact that sows at pregnancy usually retained greater energy than the non-pregnant sows (Miller et al., 2019). However, there was no difference in the N retention ratios (retained N/digested N) between gestating and non-pregnant sows. Anabolic pregnancy usually occurs late in gestation (Miller et al., 2019; Yang et al., 2022), and a recent report found that late gestating sows have greater N utilization than mid-gestation sows (Yang et al., 2022). However, our metabolic trial started at 30 d of gestation and ceased before 80 d of gestation and covered a pregnancy phase not known for obvious anabolic characteristics regarding energy utilization.

Regardless of reproductive stage, in this study CSM diets with lower dietary fiber had lower fecal excretion and higher net energy utilization. In restricted-feeding pregnant sows, an increase in dietary fiber caused a change in the speed of the movement of digesta through the digestive tract (de Leeuw et al., 2008). It also caused an increase in fecal excretion, lowering the digestibility of nutrients and energy gain (Pedersen et al., 2007), and our findings confirmed this. Furthermore, higher fiber contents in CSM diets during pregnancy have been found to decrease urinary energy excretion (Yang et al., 2021). The energy lost in the urine depends on the urine N ratio (Noblet et al., 2022). This is probably because of the endogenous protein loss and the transfer of urinary N to fecal microbial proteins increase in response to dietary fiber (Le Gall et al., 2009; Yang et al., 2021). In addition, we cannot ignore the fact that the calculation of ME should include the energy loss from gas production (primarily as methane), which has been shown to increase with increasing CF content (Ramonet et al., 2000; Bach Knudsen et al., 2016).

Corn was used as the basal diet in this study, and since its N and fiber contents were relatively low, it is expected the CSM supplemented diets had increased N excretion and N digestion. A negative correlation has been demonstrated between the fiber content of feedstuffs and the digestibility of protein in pigs (Pastuszewska et al., 1993; Fan et al., 2001; Wilfart et al., 2007). It is reported that greater fiber content may increase the water absorbing capacity of the digesta (Montagne et al., 2003), and facilitate the movement of the digesta in the digestive tract (Chasse et al., 2021). The resulting short retention time in the small intestine would decrease the digestibility of N. Meanwhile, dietary fiber stimulates microbial activity in the gut, which would increase the excretion of microbial protein, and thus decrease the apparent digestibility of N (Yang et al., 2021).

In this study, the ATTD of DM, OM, GE, and CP decreased linearly with rising CF, NDF, and ADF contents, in both pregnant and non-pregnant sows, as previously observed (Wilfart et al., 2007). However, the ATTD of OM and CP during gestation were significantly lower in gestating sows than in non-pregnant sows. It has been observed that the relative abundance of intestinal microbiota increased significantly during gestation, and this would result in greater uptake of N for microbial protein synthesis, and lower the ATTD of N (Yang et al., 2021). This result can be explained by the increased abundance of the microbiota in gestating sows. Fiber affects the microbiota through fermentation, which might also decrease N digestibility in the hindgut (Le Goff and Noblet, 2001; Shang et al., 2021). Interestingly, changes in the intestinal microbiota (Ji et al., 2019) and hormone levels occur when a sow transitions from non-pregnant to gestation (Langendijk, 2021), and these changes will affect nutrient utilization (Niu et al., 2019). However, whether changes in the microbiota are involved in the utilization of CSM by sows still awaits to need further investigation.

The mean ATTD of NDF in gestating and non-pregnant sows was 44.62% and 49.66%, respectively, which was greater than the 13.78% observed in growing pigs (Dong et al., 2020). According to Lowell et al. (2015) protein feed with low NDF content tends to have higher NDF digestibility, which is consistent with this study. This effect might depend on the chemical properties of the feed, processing method, and animal factors such as their health, physiology, and body weight (Noblet et al., 2022). The digestion of dietary fiber by pigs is influenced by their hindgut microbiota (Niu et al., 2019; Pu et al., 2020) and although increased fiber levels could enrich the diversity of the distal gut microbiota (Jarrett and Ashworth, 2018), the specific relationships between dietary fiber, microbiota, and nutrient utilization in sows remain unclear and require further research.

However, there is a limitation in the present study. Because the CSM samples had different CP contents ranging from 40.67% to 59.63%, and all test diets included similar levels of CP (20%) of CSM. This resulted in different CP levels of tested diet, ranging from 15.01% to 18.97 %. Indeed, if tested diets were formulated to contain similar levels of CP, then it would cause greater differences in levels of other chemical components, such as crude fiber, neutral detergent fiber, acid detergent fiber, and total dietary fiber, which is known as a critical factor influencing nutrient digestibility. Under such conditions, we evaluated the available energy and amino acids by including similar levels of ingredients in they have diet, despite they havinf varied chemical compositions. Fan et al. formulated six cornstarch-based diets containing graded levels of dietary CP (4, 8, 12, 16, 20, and 24% CP, respectively) using soybean meal as the only source of amino acids, and found that the apparent digestibility of ileal amino acid and dry matter was increased sharply from 4% to 16% as the dietary CP content, and then gradually reached their plateaus as dietary CP content increased from 16% to 24% (Fan et al., 1994). If this is also true for sows, then the varied contents of dietary CP (ranged from 15.01% to 18.97 %) would result in minimal effects on the evaluation of nutritive values of CSM in the present study. However, whether varied dietary CP content of tested diets on the energy and nitrogen efficiency for gestating and non-pregnant sows awaits to need further investigation.

Conclusion

The chemical compositions of the CSM samples tested varied greatly according to their various sources and processing technologies and contributed to significant differences in their energy value and nitrogen utilization. In general, DE and ME were positively associated with CP content, but negatively associated with fiber content. The DE and ME values of the CSM samples in the present study averaged 14.52 MJ/kg and 13.19 MJ/kg, respectively, for non-pregnant sows, and 14.40 MJ/kg and 13.24 MJ/kg, respectively, for pregnant sows. The current findings provide the basis for precise diet formulation for sows using feeds containing CSM.

Glossary

Abbreviations

ADF

acid detergent fiber

AIC

Akaike’s information criterion

ATTD

apparent total tract digestibility

BW

body weight

CF

crude fiber

CP

crude protein

CSM

cottonseed meal

CV

coefficient of variation

DE

digestible energy

DM

dry matter

DMI

dry matter intake

EE

ether extract

FG

free gossypol

GE

gross energy

IDF

insoluble dietary fiber

ME

metabolizable energy

N

nitrogen

NDF

neutral detergent fiber

OM

organic matter

RMSE

root mean square error

SDF

soluble dietary fiber

TDF

total dietary fiber

Conflict of Interest Statement

The authors declare that there are no conflicts of interest.