**Background:** Avian necrotic enteritis (NE) is a multifactorial enteric disease of commercial chickens and turkeys caused by pathogenic type G Clostridium perfringens, with major economic consequences for the poultry industry worldwide. The transition away from antibiotic growth promoters has intensified the need for effective vaccines, but no fully protective vaccine is currently available. Previous efforts targeting alpha toxin (CPA), NetB, and other virulence factors have yielded only partial protection. Comparative subtractive reverse vaccinology (CSRV) offers a cost-effective approach to rapidly identify surface-exposed antigens unique to disease-causing strains. In a prior in silico study, the research group identified 14 proteins unique to six NE-causing C. perfringens strains and absent from ten commensal strains. The present study aimed to validate these in silico results, produce recombinant proteins from the candidate sequences, evaluate their immunogenicity in broiler chickens, and assess the ability of raised antibodies to recognize both the recombinant and native forms of the proteins.
**Methods:** PCR and sequencing confirmed the presence of candidate protein-encoding genes in virulent C. perfringens MLG_7820 and screened for their presence in commensal strains. Seven candidate proteins were kept after PCR and sequencing validation, and five—P153, P264, P509, P537, and P561—were successfully cloned and expressed in E. coli. Bioinformatics analyses included VaxiJen v2.0 antigenicity prediction, GRAVY score calculation for hydrophobicity, and prediction of signal peptides and transmembrane regions using SignalP-5.0 and TMHMM-2.0. P537 (VaxiJen score 0.21) was included as a predicted weak antigen to serve as a negative control. For in vivo evaluation, 140 commercial day-old male Ross broiler chickens were randomly divided into seven groups (n=20): Quil-A adjuvant-only control, five groups each receiving one recombinant protein (50 µg) with Quil-A adjuvant, and a hyperimmunized group receiving whole-cell lysate of C. perfringens MLG_7820 with Quil-A. Immunizations were given intramuscularly at days 7, 14, and 21 of age, with blood collected at days 7, 14, 21, and 35. Serum IgY antibody levels were measured by ELISA and analyzed using generalized linear models with Gamma family, likelihood ratio tests, and post-hoc Tukey tests with Benjamini–Hochberg correction. Antibody specificity was assessed by Western blot and by ELISA using whole-cell lysate from C. perfringens MLG_7820.
**Key Results:** PCR confirmed the presence of all 14 candidate protein-encoding genes in six virulent C. perfringens strains with 100% sequence similarity. Seven candidate genes were also found in commensal strains with 97%-99% homology and were excluded, leaving seven candidates for cloning. Five proteins were successfully produced: P153 (5.6 kDa, VaxiJen 0.69), P264 (9.2 kDa, VaxiJen 1.47), P509 (17.3 kDa, VaxiJen 0.86), P537 (21 kDa, VaxiJen 0.21), and P561 truncated form (17 kDa, VaxiJen 0.64). Purification yields varied: P537 at 2.5 µg/µL, P153 at 3 µg/µL, and P264 at 2 µg/µL. P561 required denaturing conditions for purification. ELISA showed a significant interaction between time and candidate protein (GLM: P<0.001). All proteins elicited significant increases in IgY titers from day 7 to day 35, with fold changes of 8.8 (P153), 16.8 (P561), 23.8 (P264), 53 (P509), and 9110 (P537). The highest mean IgY titers at day 35 were observed for P537 (9.11E+05) and P509 (2.41E+05). P264 and P537 elicited significant antibody responses after a single dose, while other proteins required two or three doses. When tested against whole-cell lysate of C. perfringens MLG_7820, IgY antibody titers at day 35 showed fold changes of 12.8 to 41.2 across groups, all significantly higher than the Quil-A control (Bonferroni-adjusted P-values ranging from <0.001 to 0.007).
**Clinical Implications:** This study provides the first demonstration of a CSRV approach to identify surface-exposed vaccine candidate proteins unique to NE-causing C. perfringens and subsequently evaluate their immunogenicity in broiler chickens. The finding that all five recombinant proteins elicited significant immune responses after just two doses is promising, as early immunity is critical before chickens become susceptible to NE under commercial conditions. The unexpected result that P537—predicted as a weak antigen with cytoplasmic localization—induced the highest antibody titers (9.11E+05, fold change 9110) suggests that in silico antigenicity prediction alone may not reliably identify the most immunogenic targets. The identification of P509 as a prepilin N-terminal cleavage methylation domain protein with the second-highest antibody response supports growing evidence that pilus-associated proteins play important roles in NE pathogenesis and immunity. However, the ability of these antibodies to confer protection against NE challenge was not assessed. Future studies using NE experimental challenge models are needed to correlate antibody titers with protection. Additionally, alternative delivery routes compatible with field application must be developed, as the intramuscular immunization scheme used here is not practical for commercial broiler production. Despite these limitations, the CSRV pipeline successfully narrowed hundreds of proteins to five immunogenic candidates, demonstrating the utility of this approach for accelerating vaccine development against NE.