1. Introduction
Feed cost accounts for over 60% of the total cost in animal husbandry, profoundly impacting production efficiency. The primary objective of feed formulation is to ensure that farm animals receive an optimal supply of nutrients, with a particular emphasis on the adequacy of various amino acids with appropriate proportions. Research has firmly established that balanced amino acid profiles are essential for promoting muscle protein deposition and maintaining intestinal health in farm animals, highlighting their indispensable significance in nutritional management [
1,
2]. Traditional feed formulations often increase protein-dense ingredients to boost amino acid content, but this can result in imbalance, inefficiencies, and resource wastage [
3]. Low-protein diets balanced for essential amino acids (EAAs), supplemented with synthetic crystalline amino acids, offer a more precise approach to meeting the nutritional requirements of livestock and poultry [
4]. This innovative strategy enables a 30% reduction in the consumption of soybean meal, a conventional high-protein ingredient in feed formulations [
5]. However, recent studies and practical experiences in swine production have revealed the instability of the performance of low-protein diets [
6,
7].
Comparative analysis across studies indicates that this instability predominantly occurs in low-protein diets that extensively incorporate unconventional feed ingredients, such as rapeseed meal and wheat middling [
6,
8,
9]. This may be attributed to the fact that while the dietary supplementation with synthetic amino acids meets the requirements for digestible EAAs, the supply of digestible non-essential amino acids (NEAAs) remains inadequate. Specifically, the ileal digestibility of protein in conventional ingredients such as corn and soybean meal ranges from approximately 84% to 88%, whereas unconventional feed ingredients such as soybean hulls, rice bran meal, and whey powder exhibit substantially greater variability in digestibility, ranging from 30% to 95% [
10,
11]. Consequently, at the determined optimal crude protein (CP) levels, diets incorporating diversified feed ingredients may demonstrate significant variability in bioavailable protein content, potentially resulting in insufficient NEAA nitrogen substrates [
12]. In other words, CP, as a traditional standard for characterizing dietary nitrogen content, fails to reflect protein bioavailability. More precise evaluation methods need to be established [
13].
Digestible protein (DP), referring to the protein digested and absorbed prior to the terminal ileum, enables effective fulfillment of pigs’ nutritional requirements for bioavailable protein in diets with diverse ingredient compositions when incorporated into feed formulation systems [
14]. This study aimed to determine the DP requirements for 80–110 kg pigs. In Experiment 1, five diets were formulated with the same standardized ileal digestible (SID) EAAs but incremental DP levels. Subsequently, with growth performance serving as the primary evaluation criterion, linear regression modeling was employed to ascertain the DP requirement for finishing pigs. Based on these findings, Experiment 2 was conducted with diets of different DP contents and ingredient types to validate the practical applicability of diversified diets with optimal DP content on indicators including growth performance, carcass traits, and meat quality in pigs. Furthermore, given that protein serves as the fundamental substrate for hepatic metabolic processes and myofibrillar protein synthesis, Experiment 2 systematically quantified transcriptomic variations in hepatic and longissimus dorsi tissues under different DP intakes and dietary compositions using RNA sequencing. The present study experimentally determined the DP requirements for 80–110 kg pigs.
2. Materials and Methods
2.1. Ethical Statement
This study complied with the Chinese guidelines for experimental protocols and animal welfare. All animal procedures were approved by the Animal Care Committee of Beijing University of Agriculture (Protocol No. AW22804202-1-1, Beijing, China). Experimental pigs and materials were provided by the Swine Research Unit of Beijing Zhongyu Breeding Pig Co., Ltd. (Pingquan, China).
2.2. Experiment 1: Determination of Digestible Protein Requirements
This experiment aimed to establish the optimal DP content in low-protein diets for finishing pigs. A total of 450 crossbred pigs (25% Duroc × 25% Yorkshire × 50% Landrace; initial body weight: 81.79 ± 2.34 kg) were assigned to one of five dietary treatments based on initial body weight, with 7 male and 8 female pigs per pen and six pens per treatment. Dietary treatments were determined as graded DP levels (8.82%, 9.43%, 10.04%, 10.65%, and 11.26%) based on standardized ileal digestibility of protein in feed ingredients from NRC (2012) [
10] and Nutrient Requirements of Pigs in China (2020) [
11]. Eight amino acids were supplemented to balance the essential amino acid levels across treatment groups, ensuring that differences in dietary DP content primarily reflected varia-tions in NEAAs. All other nutrients were standardized across dietary treatments while meeting NRC (2012) [
10] swine nutrient requirements. The experimental diversified formulations utilized 11 kinds of feed ingredients with varying standardized ileal digestibility coefficients [
10], all commonly employed in swine husbandry to enhance the practical applicability of the present study (
Table 1).
Pigs were housed in environmentally controlled facilities with concrete-floored pens and provided ad libitum access to feed and water. The experiment lasted 28 days, with individual body weight and pen feed intake recorded on days 1, 14, and 28 to calculate average daily gain (ADG), average daily feed intake, and feed conversion ratio (FCR). On day 28 at 06:00, after a night of fasting, blood samples were aseptically collected from the anterior vena cava of one pig per pen (selected based on proximity to average pen body weight). Serum was separated by centrifugation at 3000× g for 10 min (4 °C) and stored at −80 °C. Total RNA was extracted using TRIzol reagent (Invitrogen, CA, USA).
2.3. Experiment 2: Verification of Digestible Protein Requirements
This experiment validated the practical applicability of the DP requirement determined in Experiment 1. A total of 270 crossbred pigs (Duroc × Yorkshire × Landrace; initial body weight: 80.51 ± 2.66 kg) were assigned to three dietary treatments in a randomized complete block design: a high-protein traditional diet (HPTD; 13.52% CP and 11.26% DP), a low-protein traditional diet (LPTD; 11.72% CP and 9.55% DP), and a low-protein diversified diet (LPDD; 11.77% CP and 9.55% DP). The HPTD was formulated to reflect the protein levels and ingredient compositions commonly employed in commercial swine production, ensuring its representativeness of industry standards. Diets differed in DP content and ingredient composition, formulated as described in Experiment 1 (
Table 2). The 28-day experiment maintained identical housing, feeding protocols, performance data collection, and blood sample collection as Experiment 1.
On day 29, one medium-weight pig per pen was euthanized by electrical stunning for carcass trait measurements. Backfat thickness at five anatomical sites (thickest shoulder point, 6th–7th rib, 10th rib, last rib, and waist joint) was measured using digital calipers. Carcass dimensions were determined with a measuring tape; specifically, carcass length was defined as the distance from the midline of the pubic symphysis to the depression at the first cervical vertebra, carcass straight length was defined as the distance from the anterior midline of the pubic symphysis to the first cervical vertebra depression, and carcass diagonal length was defined as the distance from the pubic symphysis midline to the junction of the first rib and sternum. Liver apex samples (~5 g) and longissimus dorsi muscle (10th–11th rib region, ~5 g) were collected, flash-frozen in liquid nitrogen, and stored at −80 °C for transcriptomic analysis.
2.4. Feed Chemical Analysis
Feed samples were analyzed in duplicate. Dry matter (930.15; UFE 500, Memmert GmbH, Swabach, Germany), CP (954.01; 8200, FOSS Analytical A/S, Hillerød, Denmark), calcium (984.04), and phosphorus (965.17) were quantified following [
15,
16] protocols. Acid detergent fiber and neutral detergent fiber were determined using a fiber analyzer (200i, ANKOM Technology, NY, USA) [
17].
2.5. Plasma Biochemical Analysis
Plasma biochemical parameters, including alanine aminotransferase (WB-ALT-IFCC; 0–1000 U/L), aspartate aminotransferase (WB-AST-IFCC; 0–1000 U/L), glucose (WB-GLU-GOD; 0–30 mmol/L), total protein (WB-TP-BCG; 0–120 g/L), albumin (WB-ALB-BCG; 0–60 g/L), urea nitrogen (WB-BUN-URE; 0–50 mmol/L), total cholesterol (WB-TC-CHOD; 0–20 mmol/L), total antioxidant capacity (WB-TAOC-ABTS; 0–5 mmol/L), total amino acids (TAAs; WB-TAA-NIN; 0–500 μmol/mL), and non-esterified fatty acids (WB-NEFA-ACOD; 0–2.0 mmol/L), were measured using commercial kits (Beijing Winter Song Boye Biotechnology Co., Ltd., Beijing, China) and an automatic biochemical analyzer (BS-420, Mindray, Shanghai, China), in accordance with the manufacturer’s protocols.
2.6. Meat Quality Assessment
Postmortem pH was measured at 45 min and 24 h using a portable pH meter (Matthäus pH Star, MATTHAUS, Stuttgart, Germany). The pH meter was calibrated with pH 4.01 and 7.00 buffers, and each sample was measured in triplicate. Meat color (CIE L *, a *, b *) was assessed in duplicate using a CR-410 colorimeter (Konica Minolta, Japan). The colorimeter was calibrated using a white standard tile, with three replicate measurements per sample to ensure accuracy. Drip loss (%) was determined by suspending meat cubes (1 × 1 × 2 cm
3) at 4° C for 24 h. Shear force (Warner–Bratzler device; TA.XT Plus, Surrey, UK) and cooking loss (70 °C water bath for 20 min) were measured as described in [
18,
19]. Color and marbling scores were evaluated on fresh longissimus dorsi using National Pork Producers Council standards (1.0 = pale, 6.0 = dark purplish-red [
20]).
2.7. Transcriptomic Profiling and Functional Enrichment
Total RNA was extracted with TRIzol, treated with DNase I (TaKaRa, Tokyo, Japan), and quality-checked (Agilent 2100 Bioanalyzer; NanoDrop, DE, USA). Libraries (n = 5 per group) were prepared from high-quality RNA (OD260/280: 1.8–2.2; RIN ≥ 6.5) using the TruSeq RNA Kit (Illumina, CA, USA) and sequenced on an Illumina HiSeq X Ten/NovaSeq 6000 (2 × 150 bp). Differentially expressed genes (DEGs) were identified using DESeq2/DEGseq/EdgeR (|log2FC| > 1, FDR < 0.05). Functional enrichment (KEGG) was analyzed via KOBAS (FDR-corrected p ≤ 0.05).
2.8. Statistical Analysis
All data were analyzed using SAS 9.4 (SAS Institute Inc., Cary, NC, USA). Prior to analysis, normality of distributions was verified using the Shapiro–Wilk test (UNIVARIATE procedure), and homogeneity of variances was assessed via Levene’s test. For Experiments 1 and 2, the statistical model was defined as follows:
where
Yij represents the dependent variable,
μ is the overall mean,
Ti is the fixed effect of dietary treatment (i = 5 in Experiment 1; i = 3 in Experiment 2), and
eij is the random error term with pen (j = 6 replicates per treatment) as the experimental unit.
For growth performance and biochemical parameters, GLM was used, followed by Tukey’s HSD test for post-hoc multiple comparisons. Orthogonal polynomial contrasts were applied to evaluate linear and quadratic effects of dietary DP levels, with polynomial coefficients constructed based on treatment gradients. Optimal DP levels were estimated using broken-line regression and quadratic regression models, with goodness-of-fit assessed by coefficient of determination (R2). Specifically, the broken-line model was structured as follows:
where X
0 denotes the break point, and a and b are regression coefficients.
For transcriptomic data, DEGs were identified using DESeq2 (v1.34.0) with thresholds of |log2FC| > 1 and FDR < 0.05. Functional enrichment analysis of DEGs was performed via the KyoKEGG database using the KOBAS tool (FDR-corrected p ≤ 0.05). PCA was conducted to visualize transcriptomic variation across groups.
All results are presented as mean values ± standard error of the mean (SEM). Statistical significance was defined at p ≤ 0.05, and highly significant effects were denoted at p ≤ 0.01.
4. Discussion
Low-protein EAA-balanced diets have been demonstrated to reduce the use of protein-rich feed ingredients, thereby alleviating metabolic burdens on animals and lowering feed costs [
21]. However, the optimal dietary protein level remains a matter of debate. In previous research and practical applications, CP has been adopted as a metric to characterize nitrogen levels in diets [
22]. Typically, practitioners conduct gradient experiments with varying CP levels and employ growth performance indicators, combined with regression models, to determine CP requirements for farm animals. This approach is feasible for diets primarily composed of corn and soybean meal. However, in diversified feed formulations, the evaluation of dietary nitrogen content necessitates moving beyond CP and emphasizing the importance of non-essential amino acids.
NEAAs are defined as those synthesized endogenously by animals, with their production contingent upon the absorption of sufficient nitrogen substrates in the gut [
23]. The digestibility of nitrogen varies markedly across feed ingredients, such as 84% for soybean meal, 65% for rapeseed meal, and 94% for brown rice [
10,
11]. Consequently, diets with identical CP levels but differing ingredient compositions exhibit significant variability in DP content. While synthetic amino acids are commonly supplemented to meet EAA requirements, insufficient DP compromises the substrate availability for endogenous NEAA synthesis [
24]. This discrepancy renders CP requirements determined for one diet type inapplicable to others, underscoring the need for DP as a novel standard in formulating diversified feeds. Based on previous research, this study innovatively set up graded diets with DP replacing CP and used the broken-line model and quadratic curve model to determine that the DP requirement for 80–110 kg pigs is 9.55%. Subsequently, Experiment 2 validated the reliability of the above conclusion in diversified diets.
In Experiment 1, five diets with graded DP levels were formulated using 11 feed ingredients, with all other nutrients satisfying the pigs’ requirements. Notably, the present study omitted comparative analysis between DP and CP metrics, as the detailed introduction of the digestible amino acid system and its advantages over the total amino acid system in accurately characterizing availability have been thoroughly reviewed [
25]. The DP-to-CP ratio across all treatment groups was within 80.7% to 81.5%. As DP increased to 9.55%, ADG and FCR exhibited linear improvements, consistent with prior studies highlighting the critical role of adequate nitrogen substrates in supporting EAA and NEAA synthesis [
26,
27]. We found that pigs achieved improved growth performance when the dietary DP level reached 9.43%, and the ADG did not incur a significant continuous increase when the DP level exceeded 10.04%. Additionally, considering the goodness of fit of the curves, we excluded the conclusions from the quadratic curve model (10.41–10.45%). The observed plateau beyond 10.4% DP suggests diminishing returns, likely due to excess nitrogen being catabolized into urea rather than being utilized for protein synthesis. This aligns with the ideal protein concept, wherein surplus amino acids undergo deamination and excretion, increasing metabolic burden without enhancing growth [
28]. Our findings corroborate [
29], which reported that exceeding optimal protein levels in swine diets elevates urea nitrogen excretion and reduces nitrogen retention.
The absence of significant differences in growth performance among the HPTD, LPTD, and LPDD groups validates the 9.55% DP threshold. Notably, the LPDD group, formulated with unconventional ingredients such as wheat middlings, sorghum, and rapeseed meal, achieved performance comparable to corn–soybean meal-based diets. This challenges traditional reliance on high-protein ingredients and demonstrates the feasibility of diversified feed formulations when DP requirements are met. Such insights are particularly valuable in regions where corn and soybean meal are scarce or costly, enabling the substitution with locally available alternatives without compromising productivity. Additionally, accurately meeting the DP requirement of pigs can also reduce the intake of excessive protein, which is of great significance for reducing nitrogen pollution excretion.
PUN and TAA levels emerged as sensitive indicators of dietary protein utilization [
30,
31]. As DP increased, TAA levels rose linearly, while PUN remained stable at 8.82–10.04% DP but surged significantly at 10.65% and 11.26% DP. This reflects that when the DP requirement is met, excessive protein supplementation may induce enhanced amino acid catabolism and increased metabolic burden [
32]. The quadratic response of total protein and globulin levels (peaking at 9.43–10.04% DP) suggests an optimal range for hepatic protein synthesis and immune function. Beyond this range, excess nitrogen may redirect metabolic resources toward detoxification pathways, as evidenced by the downregulation of lysosomal and renin–angiotensin system genes in HPTD liver tissues. These findings resonate with the resource allocation theory, where nutrient oversupply shifts metabolic priorities from growth to maintenance and stress responses [
33]. The stability of albumin levels across treatments contrasts with globulin variability, indicating preferential albumin synthesis even under nitrogen-limiting conditions. This may reflect its critical roles in maintaining colloid osmotic pressure and fatty acid transport, functions vital for survival [
34].
The lack of significant differences in backfat thickness, carcass dimensions, or meat quality parameters among dietary groups reinforces the adequacy of 9.55% DP for maintaining carcass yield and meat integrity. These results dispel earlier concerns that low-protein diets might impair fat deposition or muscle development due to insufficient substrates [
35]. Instead, balanced amino acid profiles—rather than CP content—emerge as the primary determinant of carcass composition, consistent with [
36], who demonstrated that low-protein diets with synthetic amino acids preserve lean meat percentage in finishing pigs. Stable postmortem pH, drip loss, and shear force across treatments further indicate that protein adequacy—not source—dictates meat quality. However, transcriptomic analysis revealed subtle hepatic differences: HPTD pigs exhibited upregulation of inflammation-related pathways (e.g., phagosome and Epstein–Barr virus infection), potentially linked to immune activation from excess nitrogen metabolites [
37]. Conversely, muscle tissues showed minimal transcriptional changes, underscoring their metabolic conservatism compared to the liver.
The liver’s transcriptional response to dietary protein offers vital insights into systemic adaptation mechanisms. Notably, the downregulation of lysosomal and renin–angiotensin system genes in HPTD pigs is indicative of suppressed autophagy and altered blood pressure regulation, likely resulting from the cellular stress induced by excessive nitrogen processing [
4,
38]. This aligns with previous research demonstrating that high dietary protein can overwhelm the liver’s capacity to efficiently metabolize surplus nitrogen, leading to the activation of stress-related pathways. Conversely, the upregulation of complement and coagulation cascades in LPTD and LPDD groups suggests a state of heightened immune preparedness. This could be attributed to the metabolic benefits derived from optimal protein utilization, which may enhance the immune system’s responsiveness [
39,
40]. The pronounced differences between hepatic and muscular transcriptomes underscore the tissue-specific nature of dietary protein responses. Unlike the liver, which actively adjusts nitrogen metabolism and detoxification processes, skeletal muscle focuses on anabolic activities and remains transcriptionally stable unless subjected to severe nutrient deficiency [
41,
42]. The minimal DEGs observed in muscle tissues confirm that a 9.55% DP diet effectively supports myofibrillar protein synthesis without inducing stress responses. This tissue-specific metabolic partitioning is crucial for maintaining growth performance under moderate protein restriction, provided there is a balanced supply of EAAs and NEAAs.