Effects of Dietary Wolfberry Sprout Stems and Leaves on the Rumen Bacterial Community of Fattening Tan Sheep and Its Association with Fermentation and Growth Traits

Wolfberry (Lycium barbarum) sprouts are harvested on a large scale in Ningxia, and the residual tender stems and leaves constitute an abundant but largely unused by-product. This experiment was conducted to investigate the effects of dietary supplementation with different proportions of wolfberry (Lycium barbarum) sprout stems and leaves (WSSLs) on rumen fermentation and the rumen bacterial community of fattening Tan sheep, and to reveal the relationships between rumen bacteria and fermentation and growth traits. Forty-eight healthy 3-month-old weaned Tan rams with similar body weight ((22.75 ± 0.52) kg) were randomly divided into four groups of 12 sheep; the control group (G0) was fed a basal diet, while the G4, G8 and G12 groups were fed basal diets in which 4%, 8% and 12% of the corn straw was replaced by WSSLs. The adaptation period lasted 10 days and the experimental period 60 days. At the end of the experiment, six sheep per group (24 in total) were slaughtered and rumen contents were collected for fermentation measurements, and the V3–V4 region of the bacterial 16S rRNA gene was sequenced on the Illumina NovaSeq platform. Ruminal pH, ammonia nitrogen (NH3-N) and microbial crude protein (MCP) did not differ among groups (p > 0.05), whereas total volatile fatty acid (VFA) concentration differed (30.15–56.27 mmol/L; p = 0.025), together with acetate (p = 0.032) and butyrate (p = 0.010); these VFA differences were driven mainly by the high and variable values in the control group and showed no dose–response. Average daily gain was higher in G8 than in G0 and G4 (p = 0.010), and live weight before slaughter and carcass weight increased with the dietary WSSL level (p = 0.011 and 0.013). Seventeen phyla and 210 genera were identified, with Bacteroidota and Bacillota dominant; the relative abundance of Bacteroidota decreased while that of Bacillota tended to increase with the WSSL level, and Xylanibacter was the most abundant genus (15.85–18.18%). Alpha diversity indices did not differ among groups (p > 0.05), and principal coordinate analysis (PCoA) based on Bray–Curtis distance showed limited separation (Adonis R2 = 0.157, p = 0.002). Nine genera showed nominal between-group variation (uncorrected p < 0.05); however, after Benjamini–Hochberg FDR correction across all 210 genera tested, none remained significant (all q = 1.00). Howardella was negatively correlated with acetate, propionate, butyrate and total VFA (p < 0.05), U29-B03 was positively correlated with live weight before slaughter and carcass weight (p < 0.05), and bacteria of Christensenellaceae were positively correlated with propionate, butyrate and total VFA (p < 0.05); none of these associations survived FDR correction. Dietary supplementation with 4–12% WSSLs did not depress growth performance and did not significantly alter the abundance of any individual taxon after correction for multiple testing; it moderately alters rumen bacterial community composition, and WSSLs can be used safely at inclusion levels of 4–12% as an unconventional feed resource for fattening Tan sheep.