Complex neural circuits controlling body musculature that underlie animal movement comprise of diverse motor neuron (MN) classes. To investigate how such class identities are established, we turned to the ventral nerve cord cholinergic motor nervous system of Caenorhabditis elegans that consists of 6 distinct classes (DA/DB/VA/VB/AS/VC). In addition to morphological criteria, each MN class is molecularly defined by a unique combination of class-specific effector genes while sharing a multitude of commonly expressed ones. UNC-3/EBF is the terminal selector that confers cholinergic identity on the DA/DB/VA/VB/AS classes. Interestingly, despite its expression in all 5 MN classes, UNC-3 also directly controls the expression of the effector genes that are only selectively expressed in subsets of MN classes. We hypothesized that repressor proteins are the factors restricting the ability of UNC-3 to activate these genes in a MN class-specific manner. Forward genetic screens were performed using fluorescent reporters of MN class-specific effector genes [e.g. the DA/DB-specific
unc-129 (DA/DB)] to examine if such repressors do exist. Indeed, we discovered
bnc-1, a novel gene previously uncharacterized in C. elegans or any nervous systems which is the sole C. elegans ortholog of the vertebrate paired C2H2 zinc finger transcription factor, BNC. Loss of
bnc-1 function causes
unc-129 to be derepressed in VA/VB in an
unc-3-dependent manner. Further experiments showed that BNC-1 is expressed specifically in VA/VB and works as a repressor via a putative binding site near that of UNC-3 (COE motif) within the cis-regulatory region of
unc-129. Additionally, reporters of other
unc-3-dependent effector genes that are not expressed in VA/VB [
unc-8 (DA/DB/AS),
unc-53 (DA/AS) &
acr-16 (DB)] are also repressed in VA/VB by BNC-1. This concept of selective repression extends to the expression control of other
unc-3-dependent MN class-specific effector genes, as determined by additional mutant analyses. For example,
del-1 (VA/VB) is repressed in DA/DB by
mab-9/TBX20 (DA/DB/AS) and in AS redundantly by both
mab-9 and
unc-55/NR2F (AS) whereas
glr-4 (SAB/DA9) is repressed in DA/DB by
cfi-1/ARID (DA/DB). Similarly, the aforementioned
unc-53 is repressed in DB by
vab-7/EVX (DB) whereas
acr-16 is repressed in DA by
unc-4/UNCX (DA/VA). We hence propose a general principle of MN identity diversification whereby
unc-3-dependent MN class-specific effector genes are selectively repressed by evolutionarily conserved MN class-specific repressors via direct binding to cognate sites adjacent to the COE motif, giving rise to unique combinations of MN class-specific effector genes that consequentially define individual MN classes.