Interactive effects of dietary crude protein reduction and resistant starch inclusion on growth performance and cecal microbial fermentation in broiler chickens

TL;DR

This study shows how reducing dietary protein and adding resistant starch affects chicken growth and gut microbes.

Contribution

The study reveals how resistant starch modulates the effects of reduced dietary protein on broiler growth and cecal fermentation.

Findings

Reduced dietary protein combined with resistant starch alters growth performance and cecal fatty acid profiles in broilers.

Resistant starch inclusion level modulates the response of cecal microbiota and fermentation to protein reduction.

Lower protein diets shift microbial populations toward fiber-fermenting species, impacting short-chain fatty acid production.

Abstract

This study investigated the interaction between dietary crude protein (CP) reduction and supplemental raw potato starch (RPS) on growth performance, nutrient digestibility, cecal short-chain fatty acids (SCFA) profiles, and cecal microbiota in broiler chickens. On day 7, birds were allocated to a 2 × 3 factorial arrangement in a randomized complete block design with eight replicates (22 birds per pen). Six dietary treatments comprised three CP levels (standard diet or reductions of 15 or 30 g/kg CP) and two RPS inclusion levels (20 or 40 g/kg) during both grower (day 7–21) and finisher (day 21–35) stages. The main effects of RPS level, as well as the linear and quadratic effects of dietary CP and their interactions, were evaluated using orthogonal polynomial contrasts. An interaction (P < 0.05) between RPS level and the quadratic effect of dietary CP levels was observed for weight gain and feed intake during the experimental period. Birds fed 20 g/kg RPS showed linear reductions in weight gain (WG) and feed intake (FI) as dietary CP decreased, whereas those fed 40 g/kg RPS showed a quadratic response, with the highest performance at the intermediate CP level. Cecal concentrations of acetate, butyrate, straight-chain fatty acids, and total SCFA on day 35 showed interactions (P < 0.05) between RPS level and dietary CP, following trends similar to growth performance. Isobutyrate and branched-chain fatty acids exhibited an interaction (P < 0.05) between RPS level and linear CP reduction, with linear decreases observed in birds fed 40 g/kg RPS but not in those fed 20 g/kg RPS. Irrespective of RPS level, decreasing dietary CP reduced the abundance of protein-fermenting microbes and enriched fiber-fermenting microbes (P < 0.05). Correlation analysis showed that straight-chain fatty acids were positively correlated with WG and negatively correlated with feed conversion ratio, whereas branched-chain fatty acids were negatively correlated with WG and FI. In conclusion, growth performance responses to reduced-protein diets in broilers can be modulated by dietary RPS inclusion level through shifts in cecal microbiota and fermentation patterns.

Figures

Significant cecal microbiome markers of 35-day-old broiler chickens fed experimental diets at the genus level, identified by Linear Discriminant Analysis (LDA) Effect Size (LEfSe). Different colors indicate different dietary treatments (n = 8). Experimental diets were separated into two stages: grower (day 7 to 21) and finisher (day 21 to 35). Raw potato starch was supplemented at 20 or 40 g/kg in experimental diets, crude protein concentrations of experimental diets were reduced from 200 g/kg to 170 g/kg for grower and 180 g/kg to 150 g/kg for finisher, respectively. Abbreviations: RS, resist

Full text

Introduction

Dietary fiber is traditionally regarded as an anti-nutritional factor in poultry diets because it consists mainly of non-starch polysaccharides and lignin, which are not digested by endogenous enzymes (Jha and Mishra, 2021). As a result, dietary fiber was considered a diluent that reduced nutrient utilization. However, this view has shifted as accumulating evidence indicates that moderate inclusion of dietary fiber can exert beneficial physiological effects (Mateos et al., 2012; Zhang et al., 2023). Dietary insoluble fibers stimulate gizzard development and digestive enzyme secretion, prolong digesta retention time, and escape digestion in the small intestine, thereby providing fermentable substrates for microbial activity in the hindgut (Mateos et al., 2012). Resistant starch (RS) has been increasingly recognized as a functional carbohydrate similar to dietary fiber. The RS escapes enzymatic digestion in the small intestine, and most forms exhibit properties similar to insoluble fiber by promoting digestive organ development and undergoing microbial fermentation in the hindgut, which results in the production of metabolites beneficial to the host (Tan et al., 2021; Oluseyifunmi et al., 2024). Recent work from our research group demonstrated that graded inclusion of RS, including banana starch, raw potato starch (RPS), and high-amylose cornstarch, increased ileal oligosaccharide concentrations and cecal short-chain fatty acid (SCFA) production in a dose-dependent manner (Oluseyifunmi et al., 2024). These findings suggest that RS provides adequate substrates for cecal microbes to produce beneficial fermentation products in broiler chickens (Oluseyifunmi and Olukosi, 2025a).

Reduced-crude protein (CP) diets have been extensively investigated as a nutritional strategy to improve dietary protein utilization, reduce nitrogen excretion, and limit indigestible protein flow to the hindgut, thereby minimizing protein fermentation (Wang et al., 2018; Ajao and Olukosi, 2024; Cho et al., 2024). Htoo et al. (2007) reported a gradual reduction in protein-fermentation metabolites, such as branched-chain fatty acids (BCFA), as dietary CP levels decreased. Yoon et al. (2025) demonstrated that reduced-protein corn–soybean meal (SBM) diets produced fewer toxic metabolites than reduced-protein diets containing canola meal or corn distillers’ dried grains with solubles, due to the lower proportion of indigestible protein in the SBM. Excessive protein fermentation in the hindgut leads to the production of toxic compounds such as ammonia and amines, which can compromise gut health (Wang et al., 2018).

Taken together, the present study aimed to address the knowledge gap regarding the combined effects of reduced dietary CP concentrations and supplemental RS on growth performance, digesta pH, nutrient digestibility, relative gene expression, and cecal SCFA profiles and microbiota in broiler chickens. We hypothesized that simultaneous reduction of dietary CP and increased RS inclusion would suppress protein fermentation while shifting microbial activity toward carbohydrate fermentation by reducing undigested protein and supplying fermentable substrates to the hindgut microbiota.

Materials and methods

All animal experimental procedures were approved by the Institutional Animal Care and Use Committee at the University of Georgia.

Experimental design, birds, and housing

A total of 1,056 male by-product broiler chicks were obtained from a Cobb hatchery (Cleveland, GA). From day 0 to 7, birds were fed a common corn–SBM–based starter diet formulated to meet the nutrient recommendations of Cobb 500 broilers (Cobb, 2022). On day 7, chicks were assigned to a 2 × 3 factorial treatment arrangement using a randomized complete block design based on body weight (BW). Each treatment had eight replicate pens with 22 birds per pen, and allocations were performed using the Experimental Animal Allotment Program (Kim and Lindemann, 2007).

All birds were housed in a single environmentally controlled room in floor pens (1 m × 1 m) bedded with clean wood shavings. The expected stock density was 40 kg/m^2^ at the end of the experiment. Each pen was equipped with a single feeder and nipple drinkers, which were adjusted according to the birds’ height throughout the experimental period. Temperature, lighting, and ventilation were automatically controlled in accordance with the Cobb (2021). The body weight of birds and the feed leftovers were recorded on days 7, 21, and 35. Mortality was monitored daily, and the number of dead birds was used to adjust calculations of feed intake (FI) and feed conversion ratio (FCR).

Dietary treatments

Six experimental diets were formulated for the grower (day 7 to 21) and finisher (day 21 to 35) phases (Tables 1 and 2, respectively). Diets were arranged as a 2 × 3 factorial with two levels of raw potato starch inclusion (20 or 40 g/kg) and three dietary CP concentrations. The different CP diets included a standard corn–soybean meal–based diet (200 and 180 g/kg CP for grower and finisher phases, respectively), and two reduced-protein diets containing 15 or 30 g/kg less CP than the standard diet at each phase. All diets within each phase were formulated to contain similar metabolizable energy concentrations. Standardized ileal digestible AA concentrations were formulated to meet Cobb 500 recommendations. Digestible glycine equivalents (Gly_equi_) were maintained above 11.0 g/kg for both grower and finisher diets. Dietary electrolyte balance was kept at approximately 245 and 235 mEq/kg for the grower and finisher phases, respectively. Titanium dioxide was included in grower diets at 3 g/kg as an indigestible index to determine apparent jejunal and ileal digestibility of nutrients. All diets were supplemented with phytase (Quantum Blue; AB Vista, Marlborough, UK) at 0.1 g/kg, supplying 500 FTU/kg. One phytase unit (FTU) was defined as the amount of enzyme required to release 1 µmol of inorganic phosphorus per minute from sodium phytate at 37 °C and pH 5.5. The starter diet was provided as crumbles, whereas grower and finisher diets were pelleted. The analyzed nutrient compositions of the grower and finisher diets are presented in Table 3.Table 1. Ingredient and calculated nutrient compositions of starter and grower diets (as-fed basis, g/kg)1^,^2.Table 1 dummy alt textItemsStarter dietGrower dietsRaw potato starch inclusion, g/kg2040Dietary crude protein, g/kg200185170200185170Ingredient compositionsCorn591.6615.4648.9686.6588.2621.3660.0Soybean meal350.0307.8271.0233.0313.0276.0238.0Raw potato starch–20.020.020.040.040.040.0Soybean oil25.025.019.013.027.022.015.0Dicalcium phosphate17.57.67.98.27.67.98.3Limestone5.58.38.38.48.28.38.3L-Lys-HCl0.71.12.53.81.12.43.8DL-Met2.01.61.82.01.61.82.0L-Thr0.30.20.81.40.20.81.4L-Val–––0.7––0.7L-Arg––0.51.8–0.51.7L-Cys–0.71.01.20.81.01.3L-Gly––2.12.1–2.12.1L-Ile–––0.4––0.4L-Ser––2.62.6–2.62.6Titanium dioxide–3.03.03.03.03.03.0NaHCO_3_2.02.02.02.02.02.02.0NaCl2.83.53.53.53.53.53.5Vitamin premix31.01.01.01.01.01.01.0Trace mineral premix40.80.80.80.80.80.80.8KCO_3_–1.22.53.71.22.23.3Phytase (Quantum Blue)0.10.10.10.10.10.10.1Choline Chloride0.70.70.70.70.70.70.7Calculated nutrients and energyCrude protein220200185170200185170Metabolizable energy, kcal/kg2,915295429512957295029532953Total Ca9.68.08.08.08.08.08.0Total P56.95.04.94.84.94.84.8Available P55.84.04.04.04.04.04.0Dietary electrolyte balance, mEq/kg246244247249245244243SID Gly7.14.96.55.85.06.55.9SID Ser9.18.310.39.78.410.39.7SID Gly + Ser16.213.316.815.513.416.915.6SID Gly_equi_613.610.913.812.811.013.912.81SID, standardized ileal digestible; Gly_equi_, glycine equivalents.2Birds fed a common starter diet from day 0 to 7. Experimental phases were divided into two different stages: grower (day 7 to 21) and finisher (day 21 to 35).3Vitamin premix supplemented the following per kg of diets: vitamin A, 3,527 IU; vitamin D_3_, 1,400 IU; vitamin E, 19.4 IU; niacin, 20.28 mg; D-pantothenic acid, 5.47 mg; riboflavin, 3.53 mg; vitamin B_6_, 1.46 mg; menadione, 1.10 mg; thiamin 0.97 mg; folic acid 0.57 mg; biotin, 0.08 mg; vitamin B_12_, 0.01 mg.4Trace mineral premix supplemented the following per kg of diets: iodine, 0.75 mg; manganese, 100.5 mg; copper, 3 mg; iron, 19.7 mg; zinc, 80.3 mg; selenium, 0.3 mg; calcium, 24 mg.5The total calcium and available phosphorus concentration was calculated based on the matrix value for phytase at 500 FTU/kg of feed (Quantum Blue, AB Vista, Marlborough, UK).6Gly_equi_ was calculated using the following equation: concentration of Gly + 0.714 × concentration of Ser.Table 2. Ingredient and calculated nutrient compositions of finisher diets (as-fed basis, g/kg)1^,^2.Table 2 dummy alt textItemsFinisher dietsRaw potato starch inclusion, g/kg2040Dietary crude protein, g/kg180165150180165150Ingredient compositionsCorn650.7685.8720.6614.8646.6683.6Soybean meal260.0223.0185.0266.0229.0192.0Raw potato starch20.020.020.040.040.040.0Soybean oil40.032.028.050.045.038.0Dicalcium phosphate6.56.87.06.56.87.0Limestone8.78.78.88.78.78.8L-Lys-HCl1.52.84.21.42.84.1DL-Met1.51.71.91.51.81.9L-Thr0.20.81.40.20.81.5L-Trp–0.10.3–0.10.3L-Val–0.31.0–0.31.1L-Arg–1.22.5–1.22.4L-Cys0.81.01.20.81.11.3L-Gly–2.12.1–2.12.1L-Ile––0.8––0.7L-Ser–2.62.6–2.62.6NaHCO_3_2.02.02.02.02.02.0NaCl3.53.53.53.53.53.5Vitamin premix31.01.01.01.01.01.0Trace mineral premix40.80.80.80.80.80.8KCO_3_2.03.04.52.03.04.5Phytase (Quantum Blue)0.10.10.10.10.10.1Choline Chloride0.70.70.70.70.70.7Calculated nutrients and energyCrude protein180165150180165150Metabolizable energy, kcal/kg303430223039301530193020Total Ca57.47.47.47.47.47.4Total P4.64.54.44.54.44.3Available P53.73.73.73.73.73.7Dietary electrolyte balance, mEq/kg235233239235234239SID Gly4.25.75.14.35.85.2SID Ser7.59.58.87.59.58.8SID Gly + Ser11.715.113.911.815.214.0SID Gly_equi_69.512.411.49.612.511.51SID, standardized ileal digestible; Gly_equi_, glycine equivalents.2Birds fed a common starter diet from day 0 to 7. Experimental phases were divided into two different stages: grower (day 7 to 21) and finisher (day 21 to 35).3Vitamin premix supplemented the following per kg of diets: vitamin A, 3,527 IU; vitamin D_3_, 1,400 IU; vitamin E, 19.4 IU; niacin, 20.28 mg; D-pantothenic acid, 5.47 mg; riboflavin, 3.53 mg; vitamin B_6_, 1.46 mg; menadione, 1.10 mg; thiamin 0.97 mg; folic acid 0.57 mg; biotin, 0.08 mg; vitamin B_12_, 0.01 mg.4Trace mineral premix supplemented the following per kg of diets: iodine, 0.75 mg; manganese, 100.5 mg; copper, 3 mg; iron, 19.7 mg; zinc, 80.3 mg; selenium, 0.3 mg; calcium, 24 mg.5The total calcium and available phosphorus concentration was calculated based on the matrix value for phytase at 500 FTU/kg of feed (Quantum Blue, AB Vista, Marlborough, UK).6Gly_equi_ was calculated using the following equation: concentration of Gly + 0.714 × concentration of Ser.Table 3. Analyzed nutrient compositions of the experimental diets (as-fed basis, g/kg).Table 3 dummy alt textItemsGrower dietsFinisher dietsRaw potato starch inclusion, g/kg20402040Dietary crude protein, g/kg200185170200185170180165150180165150Total starch330.6347.3364.3329.6360.0374.0336.9376.2406.2374.0372.5384.7Resistant starch9.613.314.621.619.320.112.826.314.038.840.732.6Crude protein201.4195.9182.0195.2189.5175.2175.5172.3169.3179.0172.7153.9Indispensable amino acids Arg13.812.512.513.312.512.211.611.512.112.111.211.2 His5.75.14.75.45.04.64.94.54.55.04.44.0 Ile9.68.68.09.28.47.97.77.37.78.27.17.0 Leu18.417.015.517.616.415.415.415.215.316.314.413.0 Lys12.712.312.212.212.112.111.111.211.811.610.911.2 Met4.94.74.54.64.54.54.44.74.74.74.74.4 Phe10.79.68.710.39.58.78.88.38.29.38.17.2 Thr13.812.512.513.312.512.211.611.512.112.111.211.2 Trp2.22.01.72.12.01.81.92.01.92.11.91.9 Val10.59.39.09.99.18.88.58.48.99.18.18.3Dispensable amino acids Ala10.69.99.110.19.49.09.28.99.09.68.47.7 Asp21.018.816.720.318.516.717.515.815.218.115.413.8 Cys3.83.93.83.93.73.83.63.94.14.23.83.8 Glu38.635.231.536.834.031.332.730.730.434.629.526.6 Gly8.89.89.28.59.58.87.48.88.77.78.76.1 Pro11.811.110.211.310.610.110.310.110.411.19.89.0 Ser8.89.89.28.59.79.17.78.79.08.78.98.3 Tyr7.06.25.86.86.45.86.05.75.46.25.64.7 Gly + Ser17.619.618.417.019.217.915.117.517.716.417.614.4 Gly_equi_1^,^215.116.815.814.616.415.312.915.015.113.915.112.01Gly_equi_, glycine equivalents.2Gly_equi_ was calculated using the following equation: concentration of Gly + 0.714 × concentration of Ser.

Sample collection

On day 21 of age, six birds per pen were randomly selected and euthanized using carbon dioxide. Jejunal, ileal, and cecal digesta were collected from one bird per pen for pH determination. Jejunal and ileal digesta from the remaining five birds were pooled to form a single sample per pen for the determination of apparent digestibility of dry matter (DM), nitrogen (N), AA, and starch. The distal two-thirds of the ileum was excised and gently flushed with distilled water. Cecal contents from three birds per pen were pooled for analysis of SCFA concentrations and protein content. Jejunum, liver, and ceca tissues from one bird per pen were collected for quantitative real-time reverse transcriptase polymerase chain reaction (qRT-PCR) analysis.

On day 35, three birds per pen were randomly selected and euthanized using carbon dioxide. Cecal contents from two birds were collected for SCFA analysis, and whole ceca were obtained for microbiota analysis. Jejunum, liver, and ceca tissues from the remaining bird were collected for qRT-PCR analysis. Samples designated for qRT-PCR and microbiota analyses were rinsed with phosphate-buffered saline, snap-frozen in liquid N, and stored at −80 °C until further analysis. Digesta samples for digestibility and SCFA analyses were stored at −20 °C.

Digesta pH

Digesta samples were kept on ice and analyzed within 12 h of collection. Samples were diluted at a 1:9 ratio (1 g of digesta to 9 mL distilled water), vortexed, and homogenized using a magnetic stirrer for 20 min. Digesta pH was measured using a sterile glass pH electrode (Thermo Scientific, Beverly, MA).

Chemical analyses and calculations

Experimental diets and ileal digesta samples were ground to pass through a 0.5 mm screen prior to chemical analysis. The DM was measured by oven-drying samples at 105 °C for 24 h according to Method 934.01 (AOAC, 2016). Titanium concentrations were determined as described by Short et al. (1996). The N content was measured using a combustion analyzer (LECO Corp., St. Joseph, MI). The AA analysis was conducted following acid hydrolysis with 6 N hydrochloric acid for 24 h at 110 °C under a N atmosphere (Method 982.30 E[a]; AOAC, 2016). Performic acid oxidation at 0–5 °C for 16 h was used to quantify total sulfur amino acids (Method 982.30 E[b]; AOAC, 2016). Tryptophan was determined following alkaline hydrolysis with barium hydroxide at 110 °C for 20 h. Quantification was performed using high-performance liquid chromatography after post-column derivatization (Method 982.30 E[a, b, and c]; AOAC, 2016). Total and resistant starch contents were analyzed using commercial assay kits (a-amylase/amyloglucosidase) (K-TSTA-100A; Megazyme, Bray, Ireland) and Resistant Starch assay kit (K-RAPRS; Megazyme, Bray, Ireland) according to the Methods 996.11 and 2002.02 (AOAC, 2016), respectively.

All calculations are presented on a dry matter basis. Coefficients of the apparent jejunal and ileal digestibility of nutrients of interest in the experimental diets were calculated using the following equation:

$$[eqn]$$

where Ti_i_ and Ti_o_ (g/kg) represent concentrations of titanium in the input and output, respectively, and Nutr_i_ and Nutr_o_ (g/kg) represent concentrations of nutrients in the input and output, respectively.

Quantitative real-time reverse transcriptase PCR analysis

Approximately 100 mg of jejunum, liver, and ceca tissues were homogenized in 1 mL of QIAzol Lysis Reagent (Qiagen, Valencia, CA) using 0.5 mm diameter zirconia/silica beads. Homogenization was performed for 60 s at 50 × g using a bead beater (Biospec Products Inc., Bartlesville, OK). Total RNA was extracted following the manufacturer’s protocol and resuspended in 250 μL of HyPure^TM^ Molecular Biology Grade Water (Cytiva, Marlborough, MA). RNA concentration and purity were assessed using a Nanodrop^TM^ Eight Spectrophotometer (Thermo Scientific, Waltham, MA). Complementary DNA was synthesized using a High-Capacity Reverse Transcription kit (Applied Biosystems, Foster City, CA). The qRT-PCR was performed using a QuantStudio 3 Real-Time PCR System (Applied Biosystems, Foster City, CA) with iTaq^TM^ Universal SYBR® Green Supermix (Bio-Rad Laboratories Inc., Waltham, MA). Glyceraldehyde-3-phosphate dehydrogenase served as the reference gene. Relative mRNA expression was calculated using the 2^−ΔΔCt^ method (Livak and Schmittgen, 2001), with the standard protein diet with 20 g/kg of RPS serving as the reference treatment. Primer sequences for all target genes are listed in Table S1.

Short-chain fatty acids analysis

Short-chain fatty acid concentrations were analyzed following the method described by Lourenco et al. (2020). Approximately 1 g of cecal digesta was diluted with distilled water (1:3, weight/volume) and centrifuged at 10,000 × g for 10 min. Subsequently, 1 mL of supernatant was mixed with 0.2 mL of a 25% metaphosphoric acid solution and stored at −20 °C overnight. After thawing, samples were centrifuged again and the supernatant mixed with an internal standard (2-ethylbutyric acid at 4.55 g/L) at a 5:1 ratio (sample:internal standard). The mixture was diluted with ethyl acetate at a 1:2 ratio, vortexed, and allowed to settle for 5 min. The upper layer was collected and analyzed using gas chromatography (Shimadzu GC-2010 plus; Shimadzu Co., Tokyo, Japan), equipped with a flame ionization detector and a capillary column (Zebron ZB-FFAP; 30 m × 0.32 mm × 0.25 μm; Phenomenex Inc., Torrance, CA). The injection volume was 1.0 μL, with helium used as the carrier gas. The oven temperature was initially set at 110 °C, gradually increased to 200 °C, while the injector and detector temperatures were kept at 250 °C and 350 °C, respectively. The SCFA concentrations were determined by comparing peak heights of the samples to those of external standards.

16S rRNA amplicon sequencing and analysis

Total genomic DNA was extracted from approximately 200 mg of frozen cecal digesta using the QIAamp® Fast DNA Stool Mini Kit (Qiagen Inc., MD) following the manufacturer’s protocol. Library preparation, quality control, and sequencing of the V4 hypervariable region of the 16S rRNA gene were performed. Raw paired-end FASTQ reads were processed using Quantitative Insights Into Microbial Ecology 2 (QIIME2) version 2024.10 (Bolyen et al., 2019). Primer sequences were detected using Cutadapt and confirmed to match the V4 region (515F/806R). Subsequently, primer trimming was performed using default mismatch tolerance settings and wildcard matching. Trimmed reads were denoised using DADA2 without additional truncation to retain maximum overlap for V4 merging. DADA2 was used to filter low-quality reads, infer exact amplicon sequence variants (ASV), remove chimeras, and produce an ASV frequency table, the ASV nucleotide sequences, and denoising statistics. Representative ASV sequences were aligned, masked to remove highly variable regions, and used to construct both rooted and unrooted phylogenetic trees via FastTree. Taxonomic assignment was performed using a Naïve Bayes classifier trained on the Greengenes2 V4 backbone (version 2024.09; McDonald et al., 2024). Low-abundance ASVs, features with fewer than 10 total reads or detected in fewer than two samples, were filtered to improve downstream analyses.

Alpha and beta diversity metrics were calculated using the core-metrics-phylogenetic pipeline. Samples were rarefied to 20,000 sequences per sample according to the sequencing depth distribution observed in the feature table summary. Alpha rarefaction curves approached a plateau at approximately 20,000 sequencing depth, indicating adequate sampling coverage for subsequent diversity analyses. Alpha diversity indices, including Shannon index, Pielou’s evenness, Faith’s phylogenetic diversity (PD), and observed features (ASVs) were obtained. Phylogenetic (weighted and unweighted UniFrac) and non-phylogenetic (Bray–Curtis, Jaccard) beta diversity distances were computed. Group differences were assessed using the PERMANOVA with 999 permutations. Significant genus microbiome markers were identified using Linear Discriminant Analysis (LDA) Effect Size (LEfSe) analysis through the MicrobiomeMaker and Phyloseq packages in RStudio (R version 4.4.1; Posit PBC, Boston, MA). The significant thresholds were set to 0.05 for the Wilcoxon and Kruskal-Wallis tests, and 3.0 for the LDA score.

Statistical analyses

Data normality was tested using the Shapiro–Wilk test in the UNIVARIATE procedure in SAS (SAS Inst., Cary, NC). Statistical analyses were performed using the MIXED procedure in SAS. Orthogonal polynomial contrasts were used to assess the effects of (1) RS inclusion, (2-3) linear and quadratic effects of dietary CP concentration, (4) interaction of RS level and linear effect of dietary CP concentration, and (5) interaction of RS level and quadratic effect of dietary CP concentration.

Pearson correlation coefficients were calculated using the CORR procedure in SAS. Least squares means were generated using the LSMEANS statement. The pen was the experimental unit, and statistical significance was declared at 0.05.

Results

Growth performance and digesta pH

During the grower phase (day 7 to 21), the weight gain (WG) and BW at day 21 declined linearly (P < 0.05) as dietary CP concentrations decreased (Table 4). There was an interaction (P = 0.015) between RPS inclusion level and the linear effect of dietary CP reduction in the FCR. Diets supplemented with 20 g/kg RPS were not different in FCR, whereas those containing 40 g/kg RPS showed a linear increase in FCR as dietary CP concentration decreased. Birds fed 40 g/kg RPS diets had a lower (P = 0.001) FCR than those fed 20 g/kg RPS diets.Table 4. Growth performance of broilers fed experimental diets during the experimental period from day 7 to 351^,^2.Table 4 dummy alt textGrower (day 7 to 21)Finisher (day 21 to 35)Overall (day 7 to 35)ItemsIBWWGFIFCRFBWWGFIFCRFBWWGFIFCRRaw potato starch inclusion, g/kgCP level20Standard protein203106313741.2921266156525261.6162832262939001.48415 g/kg reduction202105313571.2881255147325151.7162728252638721.53530 g/kg reduction204104913591.2961254145324731.7152706250238321.53540Standard protein204107813541.2571282138723131.6732669246436671.48915 g/kg reduction204107813641.2661282151525411.6782796259239041.50530 g/kg reduction203104013451.2941243137523911.7402618241537371.549Pooled SE (n = 8)2.07.912.50.00678.944.351.40.030145.746.056.30.0137Raw potato starch, g/kg20203105513641.2921259149725041.6832755255238681.51840204106513541.2721269142624151.6972694249037691.514Pooled SE (n = 24)1.24.67.20.00395.228.832.00.018929.129.534.50.0081CP levelStandard protein204107113641.2741274147624191.6442750254737841.48615 g/kg reduction204106513601.2771269149425281.6972762255938881.52030 g/kg reduction204104513521.2951248141424321.7282662245837841.542Pooled SE (n = 16)1.45.68.80.00486.333.337.80.022234.034.441.10.0098P-values3Resistant starch0.6350.1410.3780.0010.1590.0410.0330.5490.0940.0900.0330.737CP linear0.9000.0020.3510.0040.0060.1400.8010.0060.0500.0490.989<0.001CP quadratic0.8900.2640.8460.2020.3350.1820.0220.6620.1430.1420.0340.616Resistant starch × CP linear0.5740.1370.8010.0150.1500.2280.1900.5890.3920.3790.2140.749Resistant starch × CP quadratic0.5570.1140.2700.7410.1270.0230.0510.1190.0140.0150.0440.1031CP, crude protein; SE, standard error; IBW, initial body weight; WG, weight gain; FI, feed intake; FCR, feed conversion ratio; FBW, final body weight.2Experimental diets were separated into two stages: grower (day 7 to 21) and finisher (day 21 to 35). Raw potato starch was supplemented at 20 or 40 g/kg in experimental diets, crude protein concentrations of experimental diets were reduced from 200 g/kg (standard protein) to 170 g/kg (30 g/kg reduction) for grower and 180 g/kg (standard protein) to 150 g/kg (30 g/kg reduction) for finisher, respectively.3CP linear, linear effect of CP; CP quadratic, quadratic effect of CP.

In the finisher phase (days 21 to 35), WG and BW at day 35 showed an interaction (P = 0.023 and P = 0.014, respectively) between RPS level and the quadratic effect of dietary CP concentration. Birds receiving 40 g/kg RPS diets showed a quadratic response, whereas WG and BW at day 35 in birds fed diets containing 20 g/kg RPS declined linearly with decreasing CP concentrations. The FI and WG were lower in birds fed 40 g/kg RPS compared with those fed 20 g/kg RPS (P < 0.05). Feed intake responded quadratically to CP reduction (P = 0.022), peaking at 165 g/kg CP. In contrast, FCR increased linearly as dietary CP decreased (P = 0.006).

For the overall experimental period (day 7 to 35), WG and FI were affected by an interaction (P = 0.015 and P = 0.044, respectively) between RPS level and the quadratic effect of dietary CP concentration. Birds fed 20 g/kg RPS exhibited linear reductions in WG and FI as dietary CP decreased, whereas those fed 40 g/kg RPS diets showed a quadratic response. Feed intake was higher (P = 0.033) in birds fed 20 g/kg RPS diets than in those fed 40 g/kg of RPS. Regardless of RPS inclusion level, FCR increased linearly (P < 0.01) with decreasing dietary CP concentration.

No interaction effects were detected for digesta pH on days 21 and 35 of age (Table 5). Cecal pH on day 21 increased linearly (P = 0.03) as dietary CP concentration decreased. On day 35, birds fed 40 g/kg RPS showed lower (P = 0.014) jejunum pH compared to the diet having 20 g/kg RPS. Jejunum pH decreased linearly (P = 0.019) with dietary CP reduction.Table 5. Digesta pH of broilers fed experimental diets on day 21 and 35 of age1^,^2.Table 5 dummy alt textDay 21Day 35ItemsJejunumIleumCecaJejunumIleumCecaRaw potato starch inclusion, g/kgCP level20Standard protein6.626.675.896.767.058.0615 g/kg reduction6.536.866.216.747.318.2230 g/kg reduction6.666.726.416.627.588.3240Standard protein6.516.595.906.606.698.2215 g/kg reduction6.636.856.236.637.038.1930 g/kg reduction6.656.926.526.237.488.14Pooled SE (n = 8)0.0510.1840.2530.1060.4060.110Raw potato starch, g/kg206.606.756.176.717.318.20406.606.796.226.487.078.19Pooled SE (n = 24)0.0340.1090.1460.0610.2340.064CP levelStandard protein6.566.635.906.686.878.1415 g/kg reduction6.586.856.226.687.178.2130 g/kg reduction6.666.826.476.427.538.23Pooled SE (n = 16)0.0390.1340.1790.0750.2870.078P-values3Resistant starch0.8740.8210.8260.0140.4630.852CP Linear0.0710.3200.0300.0190.1130.411CP Quadratic0.4680.4290.8640.1700.9370.846Resistant starch × CP Linear0.3230.4440.8540.2630.7540.123Resistant starch × CP Quadratic0.0910.8260.9270.3700.9390.8931CP, crude protein; SE, standard error.2Experimental diets were separated into two stages: grower (day 7 to 21) and finisher (day 21 to 35). Raw potato starch was supplemented at 20 or 40 g/kg in experimental diets, crude protein concentrations of experimental diets were reduced from 200 g/kg (standard protein) to 170 g/kg (30 g/kg reduction) for grower and 180 g/kg (standard protein) to 150 g/kg (30 g/kg reduction) for finisher, respectively.3CP linear, linear effect of CP; CP quadratic, quadratic effect of CP.

Coefficients of apparent jejunal and ileal digestibility of nutrients

On day 21, no interactions were observed between RPS inclusion and dietary CP concentration for apparent jejunal or ileal digestibility of DM, N, and starch (Table 6). The coefficient of apparent jejunal DM digestibility increased quadratically (P = 0.035) with decreasing dietary CP concentrations. Coefficients of apparent jejunal and ileal digestibility of starch increased linearly (P < 0.01) as CP concentration decreased.Table 6. Coefficients of apparent digestibility of dry matter, nitrogen, and starch of jejunal and ileum in broilers fed experimental diets on day 21 of age1^,^2.Table 6 dummy alt textJejunal digestibilityIleal digestibilityItemsDry matterNitrogenStarchDry matterNitrogenStarchRaw potato starch inclusion, g/kgCP level, g/kg202000.4300.5000.5560.7160.8130.8991850.5650.6080.6740.7410.8230.9161700.5660.5860.6910.7330.8010.914402000.4960.5660.6000.7390.8150.9011850.5530.5860.6590.7410.8190.9101700.5650.5810.6790.7460.8190.926Pooled SE (n = 8)0.02670.02680.02820.01390.01010.0073Raw potato starch inclusion20 g/kg0.5200.5650.6400.7300.8120.91040 g/kg0.5380.5780.6460.7420.8180.913Pooled SE (n = 24)0.01870.01590.01810.00890.00680.0050CP level, g/kg2000.4630.5330.5780.7280.8140.9001850.5590.5970.6660.7410.8210.9131700.5660.5840.6850.7390.8100.920Pooled SE (n = 16)0.02100.01900.02110.01040.00780.0056P-values3Resistant starch0.3660.5560.8040.2650.4780.590CP Linear<0.0010.069<0.0010.3690.6880.004CP Quadratic0.0350.1060.1400.4940.2820.587Resistant starch × CP Linear0.1580.2030.2960.7400.4230.452Resistant starch × CP Quadratic0.2750.2760.5090.4280.3970.2351CP, crude protein; SE, standard error.2Experimental diets were separated into two stages: grower (day 7 to 21) and finisher (day 21 to 35). Raw potato starch was supplemented at 20 or 40 g/kg in experimental diets, crude protein concentrations of experimental diets were reduced from 200 g/kg (standard protein) to 170 g/kg (30 g/kg reduction) for grower and 180 g/kg (standard protein) to 150 g/kg (30 g/kg reduction) for finisher, respectively.3CP linear, linear effect of CP; CP quadratic, quadratic effect of CP.

An interaction between RPS inclusion and the linear effect of dietary CP concentration was detected (P = 0.040) for apparent ileal digestibility of Met (Table 7). Birds fed diets containing 40 g/kg RPS showed a linear increase in Met digestibility with CP reduction, whereas no change was observed in those fed 20 g/kg RPS. The coefficient of apparent ileal digestibility of Trp decreased linearly (P = 0.049) with decreasing dietary CP concentrations. Apparent ileal digestibility of Gly and Ser quadratically increased (P < 0.05) as dietary CP concentrations decreased (Table 8).Table 7. Coefficients of apparent ileal digestibility of indispensable amino acids in broilers fed experimental diets on day 21 of age1^,^2.Table 7 dummy alt textItemsArgHisIleLeuLysMetPheThrTrpValRaw potato starch inclusion, g/kgCP level, g/kg202000.8930.8330.8160.8250.8500.8990.8330.8930.8460.8061850.9000.8360.8230.8360.8630.9100.8390.9000.8590.8091700.8880.8110.7960.8080.8510.8910.8060.8880.8140.789402000.8910.8280.8130.8260.8430.8950.8330.8910.8360.8001850.9000.8340.8200.8340.8590.9000.8360.9000.8430.8051700.8990.8290.8180.8330.8710.9110.8330.8990.8400.809Pooled SE (n = 8)0.00660.00890.01000.00950.00870.00650.00900.00660.00940.0104Raw potato starch inclusion, g/kg200.8930.8270.8120.8230.8550.9000.8260.8930.8400.801400.8970.8300.8170.8310.8580.9020.8340.8970.8400.805Pooled SE (n = 24)0.00450.00590.00670.00640.00580.00470.00610.00450.00550.0070CP level, g/kg2000.8920.8300.8140.8260.8460.8970.8330.8920.8410.8031850.9000.8350.8210.8350.8610.9050.8380.9000.8510.8071700.8930.8200.8070.8200.8610.9010.8190.8930.8270.799Pooled SE (n = 16)0.00510.00680.00770.00730.00670.00520.00700.00510.00670.0079P-values3Resistant starch0.4990.6250.5060.2650.6570.6490.2360.4991.0000.667CP linear0.8360.2350.4160.5150.0690.4360.1110.8360.1360.645CP quadratic0.1560.1710.1870.1090.3270.2250.1050.1560.0490.471Resistant starch × CP linear0.3030.1820.1790.1740.0940.0400.1110.3030.0620.172Resistant starch × CP quadratic0.6320.5450.4810.2990.4740.0680.2690.6320.1440.5181CP, crude protein; SE, standard error.2Experimental diets were separated into two stages: grower (day 7 to 21) and finisher (day 21 to 35). Raw potato starch was supplemented at 20 or 40 g/kg in experimental diets, crude protein concentrations of experimental diets were reduced from 200 g/kg (standard protein) to 170 g/kg (30 g/kg reduction) for grower and 180 g/kg (standard protein) to 150 g/kg (30 g/kg reduction) for finisher, respectively.3CP linear, linear effect of CP; CP quadratic, quadratic effect of CP.Table 8. Coefficients of apparent ileal digestibility of dispensable amino acids in broilers fed experimental diets on day 21 of age1^,^2.Table 8 dummy alt textItemsAlaAspCysGluGlyProSerTyrRaw potato starch inclusion, g/kgCP level, g/kg202000.8150.8060.7200.8700.7680.8050.8050.8311850.8280.8130.7510.8800.8150.8190.8460.8311700.7980.7830.7400.8590.7890.7860.8290.806402000.8130.8030.7240.8710.7660.8050.8090.8341850.8200.8080.7380.8780.8090.8100.8390.8351700.8250.7990.7530.8730.8000.8080.8440.831Pooled SE (n = 8)0.00980.01060.01480.00730.01180.01020.00940.0096Raw potato starch inclusion, g/kg200.8130.8000.7370.8700.7900.8030.8270.823400.8190.8030.7380.8740.7920.8080.8300.833Pooled SE (n = 24)0.00670.00720.00990.00490.00780.00660.00580.0061CP level, g/kg2000.8140.8040.7220.8710.7670.8050.8070.8331850.8240.8100.7440.8790.8120.8140.8430.8331700.8110.7910.7460.8660.7940.7970.8360.819Pooled SE (n = 16)0.00760.00820.01130.00560.00900.00770.00690.0072P-values3Resistant starch0.4240.7510.9400.4420.8880.5940.6150.166CP linear0.7790.1600.0800.4520.0160.3970.0030.136CP quadratic0.1500.1410.3840.0700.0020.1100.0120.343Resistant starch × CP linear0.0980.3040.7480.3480.5670.2700.5390.220Resistant starch × CP quadratic0.1990.5020.3560.3850.5520.2460.2900.5261CP, crude protein; SE, standard error.2Experimental diets were separated into two stages: grower (day 7 to 21) and finisher (day 21 to 35). Raw potato starch was supplemented at 20 or 40 g/kg in experimental diets, crude protein concentrations of experimental diets were reduced from 200 g/kg (standard protein) to 170 g/kg (30 g/kg reduction) for grower and 180 g/kg (standard protein) to 150 g/kg (30 g/kg reduction) for finisher, respectively.3CP linear, linear effect of CP; CP quadratic, quadratic effect of CP.

Relative expressions of nutrient transporter, tight junction protein, and hepatic protein metabolism genes

On day 21, no interaction effects were observed for the relative expression of nutrient transporter, tight junction protein, or hepatic protein metabolism genes (Table 9). The relative expression of glucose transporter 1 (GLUT1) decreased linearly (P = 0.05) as dietary CP concentration was reduced. Tight junction protein-related genes, including occludin (OCLDN) and claudin 1 (CLDN1), showed linear or quadratic decreases (P = 0.036 or P = 0.045, respectively) with dietary CP reduction. Birds fed 20 g/kg RPS diets exhibited higher (P = 0.037) relative expression of mammalian target of rapamycin (mTOR) compared with those fed 40 g/kg RPS diets (Table 10). Expression of hepatic protein metabolism genes [mTOR, ribosomal protein S6 kinase beta-1 (RPS6KB1), and eukaryotic translation initiation factor 4E-binding protein 1 (EIF4EBP1)] of birds linearly increased (P < 0.05) as dietary CP concentrations decreased.Table 9. Relative gene expressions of nutrient transporters and tight junction proteins in jejunum tissue of broilers on day 21 of age1^,^2.Table 9 dummy alt textNutrient transportersTight junction proteinsItemsPEPT1EAAT3GLUT1GLUT2B^0,+^ATSGLT1CAT2y^+^LAT1OCLDNCLDN1JAM2**ZO1Raw potato starch inclusion, g/kgCP level, g/kg202001.0001.0001.0001.0001.0001.0001.0001.0001.0001.0001.0001.0001851.0460.9910.9111.1010.8831.0180.7540.8940.9330.5420.9741.0091700.7120.8860.7300.9300.6840.8290.6330.8550.7630.5190.7980.949402001.0231.1691.1130.9831.1770.9681.5341.0241.1502.1611.0151.1901850.8380.9960.7350.9640.8350.9160.6100.8250.8920.4730.7190.9101701.0551.1480.8030.8891.1750.9961.3451.0490.9911.2521.0041.106Pooled SE (n = 8)0.16060.13180.14250.10940.14730.11620.37610.20890.10480.41030.16250.1086Raw potato starch inclusion, g/kg200.9190.9590.8801.0100.8560.9490.7960.9160.8990.6870.9240.986400.9721.1040.8840.9451.0620.9601.1630.9661.0111.2950.9131.069Pooled SE (n = 24)0.11590.07610.08230.07150.08870.07270.23080.12060.07420.24620.10170.0703CP level, g/kg2001.0111.0851.0560.9921.0890.9841.2671.0121.0751.5801.0071.0951850.9420.9930.8231.0330.8590.9670.6820.8590.9130.5070.8470.9591700.8831.0170.7670.9090.9300.9120.9890.9520.8770.8850.9011.027Pooled SE (n = 16)0.12860.09320.10080.08260.10640.08570.27440.14770.08290.29590.11980.0816P-values3Resistant starch0.6380.1830.9780.4390.0880.9030.2240.7740.1390.0720.9280.324CP linear0.3530.6100.0500.4220.2770.5220.4500.7760.0360.0930.4990.509CP quadratic0.9650.6180.4880.3550.2370.8480.1660.5020.4250.0450.4280.254Resistant starch × CP linear0.2470.7250.8900.9070.2830.3780.8080.6860.6680.5980.5410.874Resistant starch × CP quadratic0.1070.3630.2960.5440.1350.3850.2320.6240.1510.1530.1820.1311PEPT1, peptides transporter 1; EAAT3, excitatory amino acid transporter 3; GLUT1, glucose transporter 1; GLUT2, glucose transporter 2; B^0,+^AT, b (0,+)-type amino acid transporter 1; SGLT1, sodium/glucose cotransporter 1; CAT2, cationic amino acid transporter 2; y^+^LAT1, Y+L amino acid transporter 1; OCLDN, occludin; CLDN1, claudin 1; JAM2, junctional adhesion molecule 2; ZO1, zonula occludens 1; CP, crude protein; SE, standard error.2Experimental diets were separated into two stages: grower (day 7 to 21) and finisher (day 21 to 35). Raw potato starch was supplemented at 20 or 40 g/kg in experimental diets, crude protein concentrations of experimental diets were reduced from 200 g/kg (standard protein) to 170 g/kg (30 g/kg reduction) for grower and 180 g/kg (standard protein) to 150 g/kg (30 g/kg reduction) for finisher, respectively.3CP linear, linear effect of CP; CP quadratic, quadratic effect of CP.Table 10. Relative gene expression of hepatic protein metabolism and cecal fatty acid transporters of broilers on day 21 of age1^,^2.Table 10 dummy alt textHepatic protein metabolismCecal fatty acid transportersItemsmTORRPS6KB1EIF4EBP1FBXO32TRIM36Eef2MCT1MCT4FFAR2**FFAR3Raw potato starch inclusion, g/kgCP level, g/kg202001.0001.0001.0001.0001.0001.0001.0001.0001.0001.0001851.1361.1211.2091.5871.6181.1601.0060.9171.8432.0051701.6401.4751.5121.7391.7511.7110.9011.3561.8161.508402000.7000.8100.9811.0901.1471.0500.8830.9270.8020.9351850.9841.1441.2891.2821.3451.5100.6780.5280.5800.5081701.1021.2631.7531.3571.4651.4891.9211.7532.7323.973Pooled SE (n = 8)0.22310.22130.29690.31240.33240.36020.37070.34070.86241.0775Raw potato starch inclusion, g/kg201.2591.1991.2401.4421.4561.2900.9691.0911.5531.504400.9291.0721.3411.2431.3191.3501.1611.0691.3721.805Pooled SE (n = 24)0.16300.17060.22050.20320.22210.27000.21740.19670.55670.6728CP level, g/kg2000.8500.9050.9911.0451.0741.0250.9420.9640.9010.9681851.0601.1321.2491.4341.4821.3350.8420.7231.2121.2571701.3711.3691.6331.5481.6081.6001.4111.5542.2742.740Pooled SE (n = 16)0.17990.18460.24190.23530.25420.29510.26420.24090.66730.7936P-values3Resistant starch0.0370.3750.6150.4080.5830.8040.5390.9400.7940.731CP linear0.0080.0110.0120.0920.0870.0570.2110.0910.1050.095CP quadratic0.7570.9730.7680.5880.5940.9290.3290.0920.6170.533Resistant starch × CP linear0.5260.9480.5970.4220.4790.6440.1320.4950.5030.228Resistant starch × CP quadratic0.4150.4590.9410.7540.7010.3940.2570.3810.2840.1641mTOR, mammalian target of rapamycin; RPS6KB1, ribosomal protein S6 kinase beta-1; EIF4EBP1, eukaryotic translation initiation factor 4E-binding protein 1; FBXO32, F-box protein 32; TRIM36, tripartite motif containing 36; Eef2, eukaryotic elongation factor 2; MCT1, monocarboxylate transporter 1; MCT4, monocarboxylate transporter 4; FFAR2, free fatty acid receptor 2; FFAR3, free fatty acid receptor 3; CP, crude protein; SE, standard error.2Experimental diets were separated into two stages: grower (day 7 to 21) and finisher (day 21 to 35). Raw potato starch was supplemented at 20 or 40 g/kg in experimental diets, crude protein concentrations of experimental diets were reduced from 200 g/kg (standard protein) to 170 g/kg (30 g/kg reduction) for grower and 180 g/kg (standard protein) to 150 g/kg (30 g/kg reduction) for finisher, respectively.3CP linear, linear effect of CP; CP quadratic, quadratic effect of CP.

On day 35, an interaction between RPS inclusion level and dietary CP concentrations was observed only for the relative expression of zonula occludens 1 (ZO1; Table 11). Birds fed diets containing 40 g/kg RPS showed greater (P = 0.038) expression of peptide transporter 1 (PEPT1) than the diets having 20 g/kg RPS. Expression of PEPT1 and b^(0,+)^-type amino acid transporter 1 (B^0,+^AT) increased linearly (P < 0.05) as dietary CP concentration decreased, whereas cationic amino acid transporter 2 (CAT2) and Y^+^L amino acid transporter 1 (y^+^LAT1) showed quadratic decreases (P < 0.05) with decreasing CP concentrations. Expression of OCLDN and CLDN1 increased linearly with decreasing CP concentrations, while junctional adhesion molecule 2 (JAM2) increased quadratically (P = 0.043). There was an interaction for ZO1 (P = 0.018); ZO1 expression increased linearly with CP reduction only in birds fed 40 g/kg RPS diets, whereas the relative expression of ZO1 was unchanged in response to CP reduction in birds fed the diets having 20 g/kg RPS. There were no differences in expressions of hepatic protein metabolism and fatty acid transporter genes between dietary treatments (Table 12).Table 11. Relative gene expressions of nutrient transporters and tight junction proteins in jejunum tissue of broilers on day 35 of age1^,^2.Table 11 dummy alt textNutrient transportersTight junction proteinsItemsPEPT1EAAT3GLUT1GLUT2B^0,+^ATSGLT1CAT2y^+^LAT1OCLDNCLDN1JAM2ZO1Raw potato starch inclusion, g/kgCP level, g/kg201801.0001.0001.0001.0001.0001.0001.0001.0001.0001.0001.0001.0001651.1560.9380.6951.1461.0261.0110.5450.9021.1240.5440.7450.7351501.4841.0442.0111.1321.4281.0981.4191.5271.2341.5791.2881.247401801.1900.9080.6031.3270.9991.1260.6560.6311.0150.8210.5950.6141651.4880.8830.6271.0441.1330.8990.5510.5471.1290.4570.5790.7671502.1211.1621.0871.2851.3031.1692.9062.1711.4573.6552.0261.869Pooled SE (n = 8)0.23220.10920.48930.16910.16460.15790.48790.28580.10340.78600.32330.2120Raw potato starch inclusion, g/kg201.2130.9941.2351.0931.1511.0360.9881.1431.1201.0411.0110.994401.6000.9840.7721.2191.1451.0651.3711.1161.2011.6451.0671.083Pooled SE (n = 24)0.14760.07490.30990.10000.10340.12040.30890.19410.07240.47480.20080.1321CP level, g/kg1801.0950.9540.8011.1631.0001.0630.8280.8151.0080.9110.7970.8071651.3220.9110.6611.0951.0790.9550.5480.7251.1270.5000.6620.7511501.8031.1031.5491.2091.3651.1342.1621.8491.3462.6171.6571.558Pooled SE (n = 16)0.17270.08480.36310.11960.12160.13080.36210.21850.08130.56880.23740.1560P-values3Resistant starch0.0380.9020.2300.3720.9620.7830.3180.9030.2790.3420.8280.595CP linear0.0030.1340.1160.7940.0260.5750.007<0.0010.0010.0330.0090.001CP quadratic0.5090.1710.2090.5410.4520.1960.0240.0120.5270.0650.0430.019Resistant starch × CP linear0.3160.2890.5740.6170.6950.8310.0560.0650.2590.1510.0740.018Resistant starch × CP quadratic0.8340.6910.4650.2530.5370.3370.4850.2860.4710.4410.5400.8081PEPT1, peptides transporter 1; EAAT3, excitatory amino acid transporter 3; GLUT1, glucose transporter 1; GLUT2, glucose transporter 2; B^0,+^AT, b (0,+)-type amino acid transporter 1; SGLT1, sodium/glucose cotransporter 1; CAT2, cationic amino acid transporter 2; y^+^LAT1, Y+L amino acid transporter 1; OCLDN, occludin; CLDN1, claudin 1; JAM2, junctional adhesion molecule 2; ZO1, zonula occludens 1; CP, crude protein; SE, standard error.2Experimental diets were separated into two stages: grower (day 7 to 21) and finisher (day 21 to 35). Raw potato starch was supplemented at 20 or 40 g/kg in experimental diets, crude protein concentrations of experimental diets were reduced from 200 g/kg (standard protein) to 170 g/kg (30 g/kg reduction) for grower and 180 g/kg (standard protein) to 150 g/kg (30 g/kg reduction) for finisher, respectively.3CP linear, linear effect of CP; CP quadratic, quadratic effect of CP.Table 12. Relative gene expression of hepatic protein metabolism and cecal fatty acid transporters of broilers on day 35 of age1^,^2.Table 12 dummy alt textHepatic protein metabolismCecal fatty acid transportersItemsmTORRPS6KB1EIF4EBP1FBXO32TRIM36MCT1MCT4FFAR2FFAR3Raw potato starch inclusion, g/kgCP level, g/kg201801.0001.0001.0001.0001.0001.0001.0001.0001.0001651.1621.2071.0360.8090.8772.8561.9661.0101.0221500.4070.5300.6151.0341.1274.4141.9502.2752.165401800.4310.5070.7871.0481.2851.2911.5697.2664.7571650.3030.3750.4240.7010.8862.6962.1882.1382.1531500.5460.4860.5370.7050.8222.8423.3207.8844.237Pooled SE (n = 8)0.36320.38360.36050.48790.54241.67541.05473.62852.1406Raw potato starch inclusion, g/kg200.8560.9130.8830.9481.0012.7571.6391.4281.396400.4270.4560.5830.8180.9972.2762.3595.7633.716Pooled SE (n = 24)0.25470.27780.24280.31290.34830.96730.58302.18021.3452CP level, g/kg1800.7160.7530.8941.0241.1431.1451.2854.1332.8781650.7330.7910.7300.7550.8812.7762.0771.5741.5871500.4770.5080.5760.8690.9743.6282.6355.0803.201Pooled SE (n = 16)0.28570.30770.27690.36460.40561.18470.72212.61851.5820P-values3Resistant starch0.1060.0930.2670.7300.9930.7270.3820.1440.172CP linear0.4560.4540.3370.7380.7430.1460.1860.7910.875CP quadratic0.6220.5710.9870.6320.6900.7900.8920.3310.417Resistant starch × CP linear0.2720.4930.8370.6830.5660.5810.6920.9270.682Resistant starch × CP quadratic0.2480.3220.4150.9680.9830.8690.6660.4400.6171mTOR, mammalian target of rapamycin; RPS6KB1, ribosomal protein S6 kinase beta-1; EIF4EBP1, eukaryotic translation initiation factor 4E-binding protein 1; FBXO32, F-box protein 32; TRIM36, tripartite motif containing 36; MCT1, monocarboxylate transporter 1; MCT4, monocarboxylate transporter 4; FFAR2, free fatty acid receptor 2; FFAR3, free fatty acid receptor 3; CP, crude protein; SE, standard error.2Experimental diets were separated into two stages: grower (day 7 to 21) and finisher (day 21 to 35). Raw potato starch was supplemented at 20 or 40 g/kg in experimental diets, crude protein concentrations of experimental diets were reduced from 200 g/kg (standard protein) to 170 g/kg (30 g/kg reduction) for grower and 180 g/kg (standard protein) to 150 g/kg (30 g/kg reduction) for finisher, respectively.3CP linear, linear effect of CP; CP quadratic, quadratic effect of CP.

Cecal short-chain fatty acid compositions and correlation

On day 21 of age, there were no significant interactions for cecal SCFA or protein concentrations (Table 13). Concentrations of unbranched-chain fatty acids (UBCFA) and total SCFA decreased linearly (P < 0.05) as dietary CP concentrations were reduced.Table 13. Short-chain fatty acid profiles (SCFA; mM/mL) and protein concentration (mg/mL) in the ceca of 21-day old broilers receiving the experimental diets1^,^2.Table 13 dummy alt textShort-chain fatty acid profilesCecal proteinItemsAcetatePropionateIsobutyrateButyrateIsovalerateValerateUBCFABCFATotal SCFARaw potato starch inclusion, g/kgCP level, g/kg202001117.570.57231.90.6111.4271521.18315343.31851127.700.64428.10.7071.5241491.35115140.11701006.000.47728.70.5021.2741360.97913745.6402001089.120.55428.20.5821.5961471.13514841.61851106.360.57227.70.6431.2441451.21514643.51701046.700.49329.70.5161.3491411.00814241.2Pooled SE (n = 8)4.31.8870.10261.770.09580.23145.50.19545.51.92Raw potato starch inclusion, g/kg201087.090.56529.60.6071.4081461.17114743.0401077.390.53928.50.5801.3961441.11914642.1Pooled SE (n = 24)2.81.0980.06071.250.05610.13363.80.11533.81.11CP level, g/kg2001108.340.56330.00.5971.5121491.15915142.51851117.030.60827.90.6751.3841471.28314941.81701026.350.48529.20.5091.3111390.99414043.4Pooled SE (n = 16)3.21.3400.07351.400.06820.16364.30.13974.31.36P-values3Resistant starch0.8500.8460.7630.4090.7340.9500.7300.7450.7260.569CP linear0.0580.2960.4480.5780.3620.3910.0320.3970.0320.637CP quadratic0.1340.8460.3440.2040.1460.8930.4510.2250.4310.510Resistant starch × CP linear0.3760.8220.8710.1310.8220.8400.2680.8450.2730.475Resistant starch × CP quadratic0.7760.4550.6870.7470.7350.3210.6370.7060.6320.0571UBCFA, unbranched-chain fatty acids; BCFA, branched-chain fatty acids; CP, crude protein; SE, standard error.2Experimental diets were separated into two stages: grower (day 7 to 21) and finisher (day 21 to 35). Raw potato starch was supplemented at 20 or 40 g/kg in experimental diets, crude protein concentrations of experimental diets were reduced from 200 g/kg (standard protein) to 170 g/kg (30 g/kg reduction) for grower and 180 g/kg (standard protein) to 150 g/kg (30 g/kg reduction) for finisher, respectively.3CP linear, linear effect of CP; CP quadratic, quadratic effect of CP.

On day 35, acetate, butyrate, UBCFA, and total SCFA were affected by interactions (P < 0.05) between RPS level and the quadratic effect of dietary CP concentration (Table 14). Birds fed diets having 20 g/kg RPS showed linear decreases with decreasing dietary CP concentration, whereas those fed diets having 40 g/kg RPS decreased quadratically, with the highest value observed at 165 g/kg CP diet. Isobutyrate and branched-chain fatty acids (BCFA) showed interactions (P < 0.05) between RPS level and the linear effect of dietary CP concentration. As dietary CP decreased, birds fed diets containing 40 g/kg RPS showed linear reductions in isobutyrate and BCFA, whereas no change was observed in those fed 20 g/kg RPS. Regardless of RPS level, propionate, isovalerate, and valerate concentrations declined linearly (P < 0.05) with CP reduction. Cecal protein concentration on day 35 showed an interaction effect (P = 0.029) between RPS level and the linear effect of dietary CP concentration. Cecal protein was not affected by dietary CP in the birds fed diets containing 20 g/kg RPS, whereas cecal protein increased linearly in the birds fed the diets containing 40 g/kg RPS as dietary CP concentration decreased.Table 14. Short-chain fatty acid profiles (SCAF; mM/mL) and protein concentration (mg/mL) in the ceca of 35-day old broilers receiving the experimental diets1^,^2.Table 14 dummy alt textShort-chain fatty acid profilesCecal proteinItemsAcetatePropionateIsobutyrateButyrateIsovalerateValerateUBCFABCFATotal SCFARaw potato starch inclusion, g/kgCP level, g/kg201801099.511.08924.11.0561.9041442.14514732.4165998.481.18818.91.2411.7811282.42913029.5150837.351.05915.71.0251.5771082.08411034.1401809710.891.47120.41.4111.9351302.88213328.416510710.531.30822.61.2671.9511422.57514433.0150756.230.92313.41.0291.604961.9529837.1Pooled SE (n = 8)4.40.90.09411.710.12780.13386.30.21736.41.67Raw potato starch inclusion, g/kg20978.451.11219.61.1071.7541272.21912932.040939.221.23418.81.2361.8301232.47012532.8Pooled SE (n = 24)2.80.60.07041.130.10170.10804.20.16974.31.10CP level, g/kg18010310.201.28022.31.2331.9201372.51414030.41651039.501.24820.71.2541.8661352.50213731.2150796.790.99114.61.0271.5911022.01810435.6Pooled SE (n = 16)3.30.70.07701.300.10880.11504.80.18284.91.26P-values3Resistant starch0.2310.2240.0590.5480.1050.3420.4000.0730.4290.525CP linear<0.001<0.0010.001<0.0010.0360.002<0.0010.005<0.0010.002CP quadratic0.0030.1350.0990.1010.1410.1940.0060.1100.0050.192Resistant starch × CP linear0.7170.1100.0020.6840.0720.9880.8770.0130.9350.029Resistant starch × CP quadratic0.0170.1550.9810.0190.3570.4030.0120.5900.0120.1441UBCFA, unbranched-chain fatty acids; BCFA, branched-chain fatty acids; CP, crude protein; SE, standard error.2Experimental diets were separated into two stages: grower (day 7 to 21) and finisher (day 21 to 35). Raw potato starch was supplemented at 20 or 40 g/kg in experimental diets, crude protein concentrations of experimental diets were reduced from 200 g/kg (standard protein) to 170 g/kg (30 g/kg reduction) for grower and 180 g/kg (standard protein) to 150 g/kg (30 g/kg reduction) for finisher, respectively.3CP linear, linear effect of CP; CP quadratic, quadratic effect of CP.

Pearson correlation analysis revealed that UBCFA concentrations were positively correlated with WG and negatively correlated with FCR on day 35 (P < 0.05; Table 15). In contrast, BCFA concentrations were negatively correlated with WG, FI, and cecal protein (P < 0.05).Table 15. Pearson correlation coefficients between cecal short-chain fatty acids (mM/mL), and cumulative growth performance (day 7 to 35) and cecal protein (mg/mL) on day 35 of broilers fed experimental diets1^,^2.Table 15 dummy alt textItemWeight gainFeed intakeFeed conversion ratioCecal proteinUnbranched-chain fatty acids0.290.15−0.36−0.18Branched-chain fatty acids−0.30*−0.38^⁎⁎^−0.02−0.45^⁎⁎^Total short-chain fatty acids0.280.14-0.35−0.201*, P < 0.05; **, P < 0.01.2Experimental diets were separated into two stages: grower (day 7 to 21) and finisher (day 21 to 35). Raw potato starch was supplemented at 20 or 40 g/kg in experimental diets, crude protein concentrations of experimental diets were reduced from 200 g/kg (standard protein) to 170 g/kg (30 g/kg reduction) for grower and 180 g/kg (standard protein) to 150 g/kg (30 g/kg reduction) for finisher, respectively.

Diversity indices of cecal microbial composition

On day 35, alpha diversity, expressed as Faith’s PD showed an interaction (P = 0.034) between RPS level and the quadratic effect of dietary CP concentration (Table 16). Birds fed 20 g/kg RPS diets exhibited a quadratic response, with the lowest diversity at the 165 g/kg CP diet, whereas diversity decreased linearly in birds fed 40 g/kg RPS diets as dietary CP concentration declined.Table 16. Alpha diversity of cecal microbial compositions of broilers fed experimental diets at the age of 351^,^2^,^3.Table 16 dummy alt textItemsEvennessFaith’s PDObserved featuresShannon indexRaw potato starch inclusionCP level, g/kg20 g/kg1800.71015.242505.6431650.72313.672155.5681500.79515.132516.28840 g/kg1800.73114.912385.7621650.73614.612305.7731500.74612.261925.640Pooled SE (n = 8)0.03740.70921.80.3366Raw potato starch inclusion20 g/kg0.74314.682395.83340 g/kg0.73713.932205.725Pooled SE (n = 24)0.02580.45913.50.2152CP level, g/kg1800.72115.072445.7021650.72914.142225.6701500.77013.702225.964Pooled SE (n = 16)0.02990.54916.50.2592P-values3Resistant starch0.8470.1780.2940.682CP linear0.1500.0490.3120.425CP quadratic0.5710.6770.5730.560Resistant starch × CP linear0.3040.0690.2810.244Resistant starch × CP quadratic0.6350.0340.1680.4021CP, crude protein; SE, standard error; Evenness, Pielou’s evenness index. Faith’s PD, Faith’s phylogenetic diversity index.2Experimental diets were separated into two stages: grower (day 7 to 21) and finisher (day 21 to 35). Raw potato starch was supplemented at 20 or 40 g/kg in experimental diets, crude protein concentrations of experimental diets were reduced from 200 g/kg (standard protein) to 170 g/kg (30 g/kg reduction) for grower and 180 g/kg (standard protein) to 150 g/kg (30 g/kg reduction) for finisher, respectively.3CP linear, linear effect of CP; CP quadratic, quadratic effect of CP.

Beta diversity indices using four distance methods revealed consistent differences (P < 0.05) among treatment groups, particularly between diets containing 40 g/kg RPS with 150 g/kg CP versus 165 or 180 g/kg CP, as well as between diets containing 40 g/kg RPS and 150 g/kg CP and those containing 20 g/kg RPS and 180 g/kg CP (Table 17).Table 17. Beta diversity of cecal microbial compositions of broilers fed experimental diets at the age of 351^,^2^,^3.Table 17 dummy alt textMethodsGroup 1Group 2q-valueMethodsGroup 1Group 2q-valueBray–Curtis distanceTreatment 1Treatment 20.008Unweighted UniFrac distanceTreatment 1Treatment 20.243Treatment 30.041Treatment 30.145Treatment 40.590Treatment 40.694Treatment 50.714Treatment 50.694Treatment 60.008Treatment 60.008Treatment 2Treatment 30.608Treatment 2Treatment 30.694Treatment 40.569Treatment 40.236Treatment 50.590Treatment 50.694Treatment 60.009Treatment 60.064Treatment 3Treatment 40.432Treatment 3Treatment 40.236Treatment 50.432Treatment 50.355Treatment 60.041Treatment 60.078Treatment 4Treatment 50.930Treatment 4Treatment 50.355Treatment 60.008Treatment 60.008Treatment 5Treatment 60.008Treatment 5Treatment 60.010Jaccard distanceTreatment 1Treatment 20.220Weighted UniFrac distanceTreatment 1Treatment 20.057Treatment 30.048Treatment 30.094Treatment 40.964Treatment 40.648Treatment 50.825Treatment 50.877Treatment 60.008Treatment 60.008Treatment 2Treatment 30.484Treatment 2Treatment 30.391Treatment 40.366Treatment 40.349Treatment 50.825Treatment 50.391Treatment 60.048Treatment 60.015Treatment 3Treatment 40.124Treatment 3Treatment 40.438Treatment 50.124Treatment 50.391Treatment 60.060Treatment 60.094Treatment 4Treatment 50.625Treatment 4Treatment 50.877Treatment 60.008Treatment 60.015Treatment 5Treatment 60.015Treatment 5Treatment 60.0081Experimental diets were separated into two stages: grower (day 7 to 21) and finisher (day 21 to 35). Raw potato starch was supplemented at 20 or 40 g/kg in experimental diets, crude protein concentrations of experimental diets were reduced from 200 g/kg (standard protein) to 170 g/kg (30 g/kg reduction) for grower and 180 g/kg (standard protein) to 150 g/kg (30 g/kg reduction) for finisher, respectively.2Each treatment represented an experimental diet for finisher phase containing a specific combination of raw potato starch (RPS) and crude protein (CP) levels as follows: Treatment 1, 20 g/kg RPS and 180 g/kg CP; Treatment 2, 20 g/kg RPS and 165 g/kg CP; Treatment 3, 20 g/kg RPS and 150 g/kg CP; Treatment 4, 40 g/kg RPS and 180 g/kg CP; Treatment 5, 40 g/kg RPS and 165 g/kg CP; and Treatment 6, 40 g/kg RPS and 150 g/kg CP.3Group dissimilarity was assessed using the PERMANOVA with 999 permutations.

Relative abundance of genus cecal microbiota

Birds fed a diet containing 20 g/kg RPS and 180 g/kg CP showed greater abundance of CAG-314, Porcipelethomonas, and SFLA01. At 165 g/kg CP, increased abundance of Faecalibacterium, Lachnoclostridium A 130679, Escherichia, and PeH17 were observed, whereas diets containing 150 g/kg CP were characterized by greater abundance of Romboutsia B, Acutalibacter, Caproiciproducens, and Enterocloster (Fig. 1).Fig. 1. Significant cecal microbiome markers of 35-day-old broiler chickens fed experimental diets at the genus level, identified by Linear Discriminant Analysis (LDA) Effect Size (LEfSe). Different colors indicate different dietary treatments (n = 8). Experimental diets were separated into two stages: grower (day 7 to 21) and finisher (day 21 to 35). Raw potato starch was supplemented at 20 or 40 g/kg in experimental diets, crude protein concentrations of experimental diets were reduced from 200 g/kg to 170 g/kg for grower and 180 g/kg to 150 g/kg for finisher, respectively. Abbreviations: RS, resistant starch; CP, crude protein.Fig 1 dummy alt text

In birds fed 40 g/kg RPS diets, diets containing 180 g/kg CP were associated with greater abundance of CAG-273, Alistipes A 871400, CAG-313, and SFMI01. Diet with 165 g/kg CP showed enrichment of Enterenecus, Ventrisoma, Butyricicoccus A 77030, Agathobaculum, and Fournierella, whereas those containing 150 g/kg CP were characterized by greater abundance of Borkfalkia, Mediterraneibacter A 155507, Laedolimicola, Negativibacillus, Suilimivivens, and Massilitercora.

Discussion

Oluseyifunmi and Olukosi (2025b) have shown that dietary inclusion of RPS or high-amylose cornstarch improved WG during the early post-hatch up to day 13. In another study (Oluseyifunmi and Olukosi, 2025a), birds fed RS-containing diets exhibited greater acetate, butyrate, and total SCFA concentrations than those fed control diets, indicating that RS effectively supplies indigestible carbohydrates for microbial fermentation in the ceca. Furthermore, Oluseyifunmi et al. (2024) compared RS sources and reported that birds fed RPS achieved growth performance comparable to the control diet, while inclusion of RPS up to 50 g/kg did not negatively affect apparent metabolizable energy or FCR. Based on these findings, RPS was selected as the RS source in the present study, and its inclusion level was limited to 40 g/kg to minimize potential adverse effects on energy utilization and growth. The higher inclusion of RS at levels above 50 g/kg in broiler diets is not economically feasible. The absence of a control diet without an RS source was intentional; several previous studies from our research group have already extensively evaluated the effects of RS using the control diet without RS (Oluseyifunmi et al., 2024; Oluseyifunmi and Olukosi, 2025a, b). Therefore, the primary objective of the current study was to specifically investigate the interaction between inclusion levels of RS and dietary CP reduction. Building on the reported benefits of RPS, the present study aimed to elucidate the interactive effects of reduced dietary CP and supplemental RPS mainly on growth performance, cecal microbiota patterns, and cecal microbial metabolites in broiler chickens.

Growth performance

The analyzed CP concentrations of the experimental diets containing 20 g/kg RPS deviated from the formulated target values by 7 to 19 g/kg as dietary CP levels decreased. This discrepancy has been extensively discussed previously (Olukosi et al., 2025) and is largely due to not accounting for N in the supplemental amino acids during feed formulation. In spite of this deviation in expected and analyzed CP, the growth performance in birds fed diets containing 20 g/kg RPS was linearly impaired with decreasing CP levels, indicating that a dose-dependent effect of CP reduction, as expected.

In the present study, BW, WG, and FCR impaired as dietary CP concentration decreased. Previous literature suggests that reduced growth performance associated with low-protein diets may result from imbalances in dietary electrolyte balance or insufficient total dietary Gly + Ser concentrations due to reduced SBM inclusion (Dean et al., 2006; Chrystal et al., 2020). However, in the current study, standardized ileal digestible AA concentrations and dietary electrolyte balance were maintained across CP levels. Analyzed total Gly + Ser concentrations in reduced-protein diets exceeded those of the standard-protein diet, minimizing potential confounding effects. Despite these adjustments, reductions in growth performance persisted, consistent with previous reports particularly when CP reduction exceeded 30 g/kg (Lemme et al., 2019; Lee et al., 2022a; Yoon et al., 2025). These findings suggest that growth retardation was likely driven by limitations in other dispensable amino acids rather than electrolyte imbalance or total Gly + Ser deficiency. Olukosi and Lin (2024) demonstrated that supplementation with a broad range of individual dispensable AA (Gly, Gln, Ser, Ala, Gly + Ser, or Ala + Ser) in 30 g/kg CP reduction in corn–SBM diets did not produce differences in growth performance, indicating other dispensable AA, such as Tyr + Phe, Pro, and Asp, may become limiting. Another potential contributor is the increased dietary starch:protein ratio associated with CP reduction. Greenhalgh et al. (2022) reported that increased dietary starch:protein ratios have been positively correlated to ileal starch:AA disappearance rates. Moss et al. (2018) showed that increased dietary starch:protein ratio eventually depressed ileal AA digestibility and showed negative correlations by competing with the sodium-dependent transporter systems. However, this mechanism was not fully explained by the present findings on AA digestibility, as no differences were observed, potentially due to the dietary starch:protein ratio. Instead, post-absorptive metabolism of glucose and amino acids, including first-pass splanchnic extraction, may influence their competitive availability at the site of protein utilization (Trommelen et al., 2021; Greenhalgh et al., 2022).

During the grower phase, improved FCR in birds fed RPS-containing diets aligns with previous observations that RS inclusion up to 50 g/kg improved WG of birds (Oluseyifunmi and Olukosi, 2025b). Resistant starch escapes enzymatic digestion in the small intestine and functions similarly to insoluble fiber, stimulating gizzard development and digestive enzyme secretion, thereby ensuring a more complete grinding of feed (Jha and Mishra, 2021; Tan et al., 2021). However, in the present study, higher inclusion of RPS negatively affected growth performance during the finisher phase, potentially due to prolonged accumulation of indigestible starch, which may entrap nutrients and limit absorption. Similarly, Oluseyifunmi et al. (2024) reported a stepwise reduction in WG and impairment in FCR in birds from day 8 to 21 as dietary RS inclusion increased up to 100 g/kg.

Notably, interactive effects of CP and RPS were observed for WG, FI, and BW at day 35, with birds fed a greater supplemental level of RPS (40 g/kg) exhibiting quadratic responses across CP levels. Although the underlying mechanism remains unclear, these results suggest a dynamic interaction between indigestible carbohydrates and protein in the intestine, particularly under conditions of greater dietary fermentable starch.

Digesta pH

Carbohydrate fermentation generates SCFAs, which typically decrease luminal pH, whereas excess dietary protein or poorly digested protein can increase cecal and colonic pH toward neutrality due to the production of alkaline metabolites such as ammonia and amines (Wang et al., 2018). This literature aligns with the present results, where jejunal pH was lower in birds fed 40 g/kg RPS than in those fed 20 g/kg RPS, and jejunal pH decreased as dietary CP concentration declined. Nyachoti et al. (2006) reported a dose-dependent reduction in ileal digesta pH with decreasing dietary CP. In contrast, cecal pH increased linearly as CP decreased in the current study. This discrepancy may be explained by the concurrent reduction in total cecal SCFA concentrations with CP reduction. Because SCFAs represent major acidic end products of fermentation, a lower SCFA pool would be expected to reduce acidification in the ceca, thereby increasing cecal pH.

Apparent ileal digestibility of nitrogen and amino acids

The increased apparent jejunal and ileal digestibility of DM and starch observed with decreasing CP concentration is likely attributable to the greater inclusion of corn at the expense of SBM in reduced-protein diets. Corn exhibits slightly higher DM and starch digestibility than SBM (McGhee and Stein, 2020), and similar increases in DM and starch digestibility have been reported previously in reduced-protein diets compared with standard-protein diets (Moss et al., 2018; Ajao and Olukosi, 2024).

A previous study from our research group reported linear reductions in starch digestibility across the jejunum, ileum, and total tract as dietary RS increased up to 100 g/kg (Oluseyifunmi et al., 2024). In the present study, however, RPS inclusion was capped at 40 g/kg to avoid potential adverse effects of higher inclusion levels; therefore, the difference between 20 and 40 g/kg RPS was likely insufficient to reduce jejunal and ileal starch digestibility.

The increased apparent ileal digestibility of Met, Gly, and Ser with decreasing CP concentration in the current study likely reflects the progressively higher inclusion of the corresponding supplemental AA in the experimental diets. Supplemental AA are considered completely digestible (Yoon and Kong, 2023), and were supplemented in the current experiment to meet digestible AA requirement in reduced-protein diets. By contrast, apparent ileal digestibility of Trp did not increase, which may relate to the absence of supplemental crystalline Trp because dietary Trp levels already met the requirement.

Relative expressions of nutrient transporter, tight junction protein, and hepatic protein metabolism genes

Lin et al. (2023b) reported increased expression of GLUT1 in birds challenged with Eimeria species, which was interpreted as a host response to enhance glucose transport under parasitic infection. This response may facilitate increased nutrient efflux from infected cells, potentially leading to intracellular nutrient depletion and apoptosis. In contrast, the present study demonstrated that GLUT1 expression decreased linearly with decreasing dietary CP levels. This reduction may be associated with elevated blood glucose concentrations, which are attributed to the greater dietary starch content in the reduced-protein diets. Increased systemic glucose concentrations may reduce the physiological demand for glucose transport into circulation, thereby contributing to the decreased GLUT1 expression.

During the grower phase, birds fed 40 g/kg RPS exhibited improved FCR, suggesting enhanced efficiency of nutrient utilization. Interestingly, mTOR expression on day 21 was lower in birds fed 40 g/kg RPS, which may indicate reduced protein synthesis metabolism under conditions of improved feed efficiency. Tight junction gene expression (OCLDN and CLDN1) decreased with decreasing CP concentration, whereas expression of hepatic protein synthesis-related genes (mTOR, RPS6KB1, and EIF4EBP1) increased. Collectively, these patterns may suggest a resource allocation response in which birds preferentially support hepatic protein synthesis rather than intestinal barrier protein synthesis under low-protein conditions during the grower phase. As dietary CP levels decrease, the reduced concentrations of dispensable AA in the body may induce birds to increase protein synthesis as a compensatory response.

By day 35, birds fed 40 g/kg RPS showed greater PEPT1 expression, which may represent a compensatory increase in peptide transport capacity in response to reduced growth performance during prolonged RPS feeding. Moreover, changes in the expression of neutral and cationic amino acid transporters (B^0,+^AT, CAT2, and y^+^LAT1) with decreasing CP concentration indicate dynamic modulation of jejunal transport activity for AA, such as Lys, Arg, Met, Leu, Ile, Val, and Trp, which were supplemented as dietary CP decreased. In addition, the lack of consistent effects in hepatic protein metabolism gene expression and the increased expression of tight junction protein genes in reduced-protein diets at day 35 may indicate adaptation to the reduced-protein diets. Birds may initially prioritize growth-related protein accretion, followed by a later shift toward maintaining intestinal barrier integrity during prolonged exposure to reduced CP; however, this remains speculative, and further research is warranted to elucidate these adaptive responses to reduced-protein diets.

Cecal short-chain fatty acid compositions and the correlation with cumulative growth performance

Total SCFA, particularly the UBCFA, can reflect the extent of carbohydrate fermentation and may contribute to host energy supply and regulation of cellular processes (Liu et al., 2021). In the present study, supplemental RPS level did not significantly affect cecal SCFA concentrations, which may indicate that the difference between 20 and 40 g/kg RPS was insufficient to elicit distinct fermentation outcomes.

A notable finding was that UBCFA and total SCFA decreased with decreasing dietary CP on days 21 and 35. This pattern agrees with previous observations in broilers and weaned pigs (Htoo et al., 2007; Yoon et al., 2025) and may be explained by reduced substrate availability associated with lower SBM inclusion. Soybean meal contributes protein and fiber substrates to the hindgut; although corn and SBM have comparable neutral detergent fiber contents, SBM contains substantially greater acid detergent fiber than corn (Rostagno and Albino, 2024). Therefore, reducing dietary SBM inclusion may have reduced the quantity of fermentable protein and fiber entering the ceca, thereby lowering overall metabolite production.

Branched-chain fatty acids are commonly used as indicators of protein fermentation, which also produces sulfur-containing metabolites, aromatic compounds, polyamines, and ammonia (Wang et al., 2018). In the present study, BCFA decreased as dietary CP decreased. Conversely, cecal protein concentrations increased with CP reduction in birds fed 40 g/kg RPS, suggesting that a greater proportion of protein may have remained unfermented by day 35, consistent with reduced production of protein fermentation metabolites. This interpretation aligns with previous studies showing that low-protein diets markedly reduced cecal ammonia and BCFA concentrations (Htoo et al., 2007; Wang et al., 2018). The observation that this reduction in protein fermentation was most evident at the higher RPS inclusion level suggests that greater RPS supplementation may attenuate protein fermentation.

On day 35, acetate, butyrate, UBCFA, and total SCFA exhibited quadratic responses across CP levels in birds fed 40 g/kg RPS, with the highest values at intermediate CP concentrations. This pattern mirrored cumulative growth performance responses, suggesting potential interactions between indigestible carbohydrate and protein availability in shaping hindgut fermentation and performance of birds. Pearson correlation analysis further supported these relationships; UBCFA was positively correlated with WG and negatively correlated with FCR, consistent with the role of SCFAs as energy substrates that can support growth. In contrast, BCFA showed negative correlations with WG and FI, indicating that elevated protein fermentation is associated with poorer performance. It is noteworthy that this finding has been consistently observed across several studies conducted by our research group. Lin et al. (2023b) reported that Eimeria-challenged birds exhibited reduced growth performance accompanied by increased BCFA concentrations, which is consistent with the findings in Lin et al. (2023a). In addition, the negative association between BCFA and cecal protein suggests that greater residual cecal protein coincided with reduced conversion of protein into BCFA.

Diversity indices of cecal microbial composition

Alpha diversity reflects within-treatment microbial richness and phylogenetic diversity (Li et al., 2022). Faith’s PD captures the breadth of evolutionary lineages represented, and a lower Faith’s PD indicates a simpler and more phylogenetically uniform microbial community (Faith, 2018). In the present study, Faith’s PD decreased with dietary CP reduction and followed a pattern similar to cecal SCFA profiles on day 35. This likely reflects reduced substrate availability associated with lower SBM inclusion, which may have constrained microbial niches and reduced metabolite production and community complexity.

Beta diversity describes between-treatment community dissimilarity and enables comparison of microbial communities via clustering and ordination (Lozupone and Knight, 2005). In the present study, beta diversity metrics indicated that birds fed 40 g/kg RPS combined with the lowest CP level (150 g/kg) harbored microbial communities distinct from those in birds fed 40 g/kg RPS with 165 or 180 g/kg CP, as well as birds fed 20 g/kg RPS with 180 g/kg CP. Together with the alpha diversity results, these findings suggest that the combination of higher RPS inclusion and lower CP generates a uniquely structured cecal microbiota.

Relative abundance of genus cecal microbiota

Diet-induced shifts in microbial composition are well established (De Cesare et al., 2019; Tan et al., 2021), however, genus-level responses remain less comprehensively characterized. In the current study, within the 20 g/kg RPS diets, birds fed the intermediate CP level (165 g/kg) exhibited enrichment of genera associated with both carbohydrate and protein metabolism. Faecalibacterium is a recognized butyrate-producing genus in nonruminant animals’ intestines, whereas Escherichia includes taxa associated with pathogenicity and indole production from Trp metabolism (Meiners et al., 2025). With further CP reduction, genera associated with carbohydrate utilization increased, including Romboutsia, a metabolically versatile genus capable of utilizing diverse substrates, and Caproiciproducens, which can produce lactate- and caproate-related metabolites and has been implicated in chain elongation pathways (Esquivel-Elizondo et al., 2020).

Within the 40 g/kg RPS diets, birds fed the highest CP level (180 g/kg) showed a higher abundance of Alistipes A 871400, which has been associated with protein-rich diets and putrefactive pathways via fermentation of undigested protein in the gastrointestinal tract (Mohr et al., 2024). At the intermediate CP level (165 g/kg), several butyrate-associated genera were enriched, including Butyricicoccus A 77030, Agathobaculum, and Fournierella (Chang et al., 2020; Lee et al., 2022b; Akinsuyi and Roesch, 2023; Stege et al., 2023). At the lowest CP level (150 g/kg), Mediterraneibacter A 155507 was implicated in polysaccharide degradation and propionate production, was enriched (Lin et al., 2025; Meiners et al., 2025), along with Negativibacillus, which has been reported in high-fiber diets and may contribute to complex carbohydrate digestion (Bernard et al., 2024).

Supporting the present findings, Sung et al. (2023) demonstrated that increased dietary indigestible protein in pigs reduced fecal fiber-fermenting microbes by shifting microbial metabolic pathways toward teichoic acid biosynthesis and heterolactic fermentation rather than fatty acids biosynthesis. The 16S rRNA gene sequencing in the current study indicated enrichment of fiber-fermenting taxa and a reduction in protein-fermenting and potentially pathogenic taxa as dietary CP decreased, regardless of RPS level. These results are consistent with the observed reduction in cecal BCFAs and support the conclusion that the 40 g/kg RPS diet combined with the lowest CP concentration most effectively attenuated protein fermentation, likely through suppression of taxa capable of protein degradation.

Conclusion

In conclusion, reducing dietary CP concentration, in combination with increased RPS inclusion, decreased cecal BCFA concentrations and promoted the enrichment of fiber-fermenting microbial populations in broiler chickens. Growth performance responses to CP reduction differed depending on RPS inclusion level and closely mirrored changes in cecal total SCFA profiles, indicating a strong link between fermentation metabolites and birds’ performance. Correlation analysis revealed that cecal UBCFA concentrations were positively associated with cumulative growth performance of birds, whereas elevated cecal BCFA levels were linked to impaired growth and increased cecal protein fermentation. Nutrient digestibility and relative gene expressions were primarily influenced by changes in ingredient composition, such as corn, SBM, and supplemental AA, rather than synergistic effects of reduced CP and increased RPS levels.

Collectively, these findings suggest that strategic incorporation of fermentable carbohydrates alongside reduced dietary protein may effectively mitigate protein fermentation and improve gut health, with fermentation metabolites serving as key indicators of growth performance outcomes. However, as our current design did not utilize a RPS-free control diet, the conclusions may be limited to comparisons between the 20 and 40 g/kg RPS inclusion levels and should not be interpreted as evidence of effects relative to diets without RPS.

CRediT authorship contribution statement

June Hyeok Yoon: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Jeferson Lourenco: Writing – review & editing, Formal analysis. Oluyinka A. Olukosi: Writing – review & editing, Supervision, Resources, Methodology, Funding acquisition, Conceptualization.

Disclosures

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.