Paratuberculosis has been suspected to have a genetic component and estimations of heritability of about 0.15 have been reported. Caspase recruitment domain 15 (CARD15/NOD2) is a gene codifying for a cytosolic protein implicated in bacterial recognition by cells involved in innate immunity. This protein modifies inflammatory responses to bacterial triggers through activation of the nuclear factor-kB. Crohn's disease (CD) is an idiopathic inflammatory bowel disease in humans similar in many features to bovine paratuberculosis and involves an aberrant mucosal immune response in genetically susceptible individuals. The association between mutations in the CARD15/NOD2 gene and increased risk of CD has been described. The objective of this candidate gene case-control study was to characterize the distribution of two polymorphisms in the bovine CARD15/NOD2 gene and test their association with paratuberculosis infection in Florida dairy and beef cattle. The study population consisted of 432 adult cows composed of Holstein, Jersey and Brahman-Angus crosses distributed in four herds. The infection status for cases and controls was determined using five diagnostic tests (serum ELISA, milk/blood/fecal nested PCR, and fecal culture). Parallel interpretation of the results was used to compensate for limitations in sensitivity of available diagnostic tools. Two previously reported single nucleotide polymorphisms (SNP1; C733R and SNP2; Q1007L) in the bovine CARD15/NOD2 gene, responsible for two amino acid substitutions, were established for the study population by the TaqMan^® genotyping assay. The statistical analysis was based on Chi-square and Fisher's Exact Test and different models were proposed for the logistic regression analysis. It was our central hypothesis that a combination of particular alleles in our candidate gene would be present in higher frequency in controls compared to cases, suggesting a role in resistance to infection. The resulting ratio of cases to controls was 1:2.5. Frequencies for the major allele in SNP1 and SNP2 two were 0.957 and 0.543, respectively. The population was in Hardy-Weinberg equilibrium for SNP2 but not for SNP1. Values for coefficient of linkage disequilibrium (LD), the normalized LD, and the correlation between the two SNPs were -0.027, 1.0, and -0.23. Chi square test indicated that SNP1 and SNP2 are in linkage disequilibrium. The statistical analysis resulted in significant differences in allelic frequencies between cases and controls for SNP1 (p<0.001) indicating a significant association between infection and mutant allele. In the analysis of genotypes a significant association was found between SNP1 and infection status (P<0.0001). A significant association between allele combinations and infection status was found (P<0.0001) when both SNPs were considered in the genotype. The low representation of the variant allele for SNP1 in Holstein and Jersey breeds raises the prospect of a potential confounding role of breed for its connection with infection. However, a significant association between SNP1 and infection was confirmed when tested within the Brahman-Angus sub-population (P=0.02). Preliminary results suggest a role for CARD15/NOD2 gene in the susceptibility of cattle to paratuberculosis infection. Amino acid substitution C733R (SNP1) appears to be associated to paratuberculosis infection in Florida cattle. These results could be the basis for further research to create a rapid method to select for more resistant individuals, genetically contributing to the control of Johne's disease.