Arms race co-evolution of plant-pathogen interactions evolved sophisticated recognition mechanisms between host immune receptors and pathogen effectors. Different allelic haplotypes of an immune receptor in host mount distinct recognition against sequence or non-sequence related effectors in pathogens. We report the molecular characterization of the Piks allele of the rice immune receptor Pik against rice blast pathogen, which requires two head-to-head arrayed nucleotide binding site and leucine-rich repeat proteins. Like other Pik genes, both Piks-1 and Piks-2 are necessary and sufficient for Piks-mediated resistance. However, unlike other Pik alleles, Piks does not recognize any known AvrPik variants of M. oryzae. Sequence analysis of the genome of an avirulent isolate V86010 further revealed that its cognate avirulence (Avr) gene most likely has no significant sequence similarity to known AvrPik variants. We conclude that Piks breaks the canonical Pik/AvrPik recognition pattern. Piks-1 and Pikm-1 have only two amino acid differences within the integrated heavy metal-associated (HMA) domain. Pikm-1-HMA interact with AvrPik-A, -D and -E in vitro and in vivo, whereas Piks-1-HMA does not bind any AvrPik variants. Reciprocal exchanges of single amino acid residues between Piks-1 and Pikm-1 further reveal a dynamic recognition mechanism between Piks/Pikm alleles and their respective effectors. Piks-1E229Q/Pikm-1V261A can only activate immunity to AvrPik-D but not to other effectors, indicating that the amino acid change of E to Q at position 229 leads to its gain of a partial recognition spectrum of Pikm. By contrast, Piks-1A261V/Pikm-1Q229E confers immunity to the Piks cognate effector, indicating that the amino acid change of Q to E at position 229 leads to its shifts of the recognition from Pikm to Piks. Intriguingly, binding activities in both Y2H and analytical gel filtration assays are illustrated between Piks-1A261V/Pikm-1Q229E and AvrPik-D. However, it is unable to mount immunity against AvrPik-D, suggesting that biochemical activities based on in vitro and in vivo assays could be insufficient for sustaining biological function of receptor and effector pairs.