Animal-vegetal (AV) polarity of most vertebrate eggs is established during early oogenesis through the formation and disassembly of the Balbiani Body (Bb). The Bb is a structure conserved from insects to humans that appears as a large granule, similar to a mRNP granule composed of mRNA and proteins, that in addition contains mitochondria, ER and Golgi. The components of the Bb, which have amyloid-like properties, include germ cell and axis determinants of the embryo that are anchored to the vegetal cortex upon Bb disassembly. Our lab discovered in zebrafish the only gene known to function in Bb disassembly, microtubule-actin crosslinking factor 1a (macf1a). Macf1 is a conserved, giant multi-domain cytoskeletal linker protein that can interact with microtubules (MTs), actin filaments (AF), and intermediate filaments (IF). In macf1a mutant oocytes the Bb fails to dissociate, the nucleus is acentric, and AV polarity of the oocyte and egg fails to form. The cytoskeleton-dependent mechanism by which Macf1a regulates Bb mRNP granule dissociation was unknown. We found that disruption of AFs phenocopies the macf1a mutant phenotype, while MT disruption does not. We determined that cytokeratins (CK), a type of IF, are enriched in the Bb. We found that Macf1a localizes to the Bb, indicating a direct function in regulating its dissociation. We thus tested if Macf1a functions via its actin binding domain (ABD) and plectin repeat domain (PRD) to integrate cortical actin and Bb CK, respectively, to mediate Bb dissociation at the oocyte cortex. We developed a CRISPR/Cas9 approach to delete the exons encoding these domains from the macf1a endogenous locus, while maintaining the open reading frame. Our analysis shows that Macf1a functions via its ABD to mediate Bb granule dissociation and nuclear positioning, while the PRD is dispensable. We propose that Macf1a does not function via its canonical mechanism of linking two cytoskeletal systems together in dissociating the Bb. Instead our results suggest that Macf1a functions by linking one cytoskeletal system, cortical actin, to another structure, the Bb, where Macf1a is localized. Through this novel linking process, it dissociates the Bb at the oocyte cortex, thus specifying the AV axis of the oocyte and future egg. To our knowledge, this is also the first study to use genome editing to unravel the module-dependent function of a cytoskeletal linker.