AVR-Mgk1 is predicted to be a MAX fold protein that belongs to a distinct family from AVR-Pik effectors.

(A) Domain architecture and amino acid sequence of AVR-Mgk1. We used SignalP v6.0 [106] to predict SP sequences in AVR-Mgk1. AVR-Mgk1 has the 2 cysteine residues (Cys27 and Cys67, indicated by black arrowheads) conserved in the MAX effector superfamily. (B) Clustering of putative M. oryzae AVR protein sequences using TRIBE-MCL [107]. Tribe-MCL assigned AVR-Mgk1 and AVR-PikD into different tribes. If a tribe includes an experimentally characterized protein, it is shown to represent the tribe. If a tribe includes an experimentally validated MAX effector protein or AVR-Mgk1, the tribe is shown in orange. Tribes having only 1 protein are not shown. (C) AVR-Mgk1 protein structure predicted by AlphaFold2 [108]. AVR-Mgk1 has antiparallel β sheets, characteristic of the MAX effector superfamily. (D) Protein structure of AVR-PikD (PDB ID: 6FU9 chain B) [71]. (E) Structure-based protein alignment between AVR-Mgk1 and AVR-PikD. TM-align [109] revealed significant structural similarity between AVR-Mgk1 and AVR-PikD, while the regions highlighted in pink structurally differ (C, D). This structural difference involves the highly polymorphic residues (His46-Pro47-Gly48) of AVR-Pik effectors that determine Pik-1 HMA domain binding and are probably modulated by arms race coevolution [70,96]. The data underlying Fig 6B and 6E can be found in S1 Data. AVR, avirulence; HMA, heavy metal-associated; SP, signal peptide.