A Long Lost Key Opens an Ancient Lock: Drosophila Myb Causes a Synthetic Multivulval Phenotype in Nematodes

The five-protein MuvB core complex (LIN9/Mip130, LIN37/Mip40, LIN52, LIN54/Mip120, and LIN53/p55CAF1/RBBP4) has been highly conserved during the evolution of animals. This nuclear complex interacts with proteins encoded by the RB tumor suppressor gene family and its associated E2F-DP transcription factors to form DREAM complexes that repress the expression of genes that regulate cell cycle progression and cell fate. The MuvB core complex also interacts with proteins encoded by the Myb oncogene family to form the Myb-MuvB complexes that activate many of the same target genes. We show that animal-type Myb genes and proteins are present in Bilateria, Cnidaria, and Placozoa, the latter including some of the simplest known animal species. However, bilaterian nematode worms appear to have lost their animal-type Myb genes hundreds of millions of years ago. Nevertheless, the amino acids in the LIN9 and LIN52 proteins that directly interact with the MuvB-binding domains of human B-Myb and Drosophila Myb are conserved in C. elegans. Here we show that, despite greater than 500 million years since their last common ancestor, the Drosophila melanogaster Myb protein can bind to the nematode LIN9 and LIN52 family proteins in vitro and can cause a synthetic multivulval (synMuv) phenotype in vivo. This phenotype is similar to that caused by loss-of-function mutations in C. elegans synMuvB class genes including those that encode homologs of the MuvB core, RB, E2F, and DP. Furthermore, amino acid substitutions in the MuvB-binding domain of Drosophila Myb that disrupt its functions in vitro and in vivo also disrupt its activity in C. elegans. We speculate that nematodes and other animals may contain another protein that can bind to LIN9 and LIN52 in order to activate transcription of genes repressed by DREAM complexes.


INTRODUCTION
The Myb gene family was discovered due to the retroviral transduction of the c-Myb proto-oncogene that created the v-Myb oncogene of the avian myeloblastosis virus 1  in Drosophila Myb null mutant animals 6,9 .
Animal Myb-type proteins all contain a broadly conserved pan-eukaryotic aminoterminal DNA-binding domain and animal-specific domains 3

. The vertebrate A-Myb and c-
Myb proteins also share a central transcriptional activation domain that is not well-conserved in vertebrate B-Myb or invertebrate Myb proteins. Surprisingly, the animal-specific carboxyterminus of Drosophila Myb is both necessary and sufficient for rescue of the adult lethality of a Myb null mutant, for proper association with chromatin, for transcriptional activation of essential G2/M phase genes, and for mitotic cell cycle progression 10,11 . Alanine substitutions of evolutionarily conserved motifs identified a short peptide sequence that is required for all these functions 11 .
Biochemical purification of an activity that bound to DNA near a developmentally regulated origin of replication in a Drosophila chorion locus led to the discovery of a multiprotein complex that contains Myb and several Myb-interacting proteins (Mip) 12  identified in human cells 13 . In addition to Drosophila Myb or vertebrate B-Myb, these complexes contain Mip130/LIN9, Mip120/LIN54, Mip40/LIN37, p55CAF1/RbAp48, and LIN52. These five additional proteins, known as the MuvB core, can also associate with Drosophila E2F2, DP, and RBF1 or RBF2 or their vertebrate homologs (E2F4 or E2F5, DP1 or DP2, p107 or p130) to form complexes now called DREAM 13 . A large holocomplex containing Myb, E2F, DP, and RB family proteins together with the MuvB core and called dREAM was identified in Drosophila embryos, but has not been observed in human cell lines [13][14][15] .
These complexes have also been implicated in human cancer initiation and progression. For example, a high level of B-Myb/MYBL2 expression in breast cancer is a clinically useful predictor of tumor recurrence and decreased patient survival [27][28][29] . Furthermore, extensive DNA sequencing has revealed that approximately one-half of human breast cancer specimens contain a genetic alteration in at least one of the genes encoding subunits of these two complexes ( Figure   S1).
Remarkably, all of the proteins in the Myb-MuvB and DREAM complexes, with the exception of Myb itself, are encoded by homologs of synMuvB group genes in the nematode Caenorhabditis elegans 30 . In brief, dominant gain-of-function mutations in the EGF=>RAS=>RAF=>MEK=>MAPK=>ETS pathway caused a multivulval (Muv) phenotype in C. elegans 31 . Recessive loss-of-function mutations in lin-8 and lin-9 together caused a synthetic multivulval phenotype (synMuv) 32 . Additional genetic screens identified two groups of genes Vorster, et al, page 5 (synMuvA and synMuvB), in which any group A mutation could cooperate with any group B mutation to cause this synMuv phenotype 33 . The proteins encoded by synMuvA and synMuvB genes redundantly repress ectopic expression of the secreted LIN-3/EGF protein that normally controls vulval development via the RAS pathway 31,34,35 . The synMuvB genes also regulate transgene silencing, cell cycle progression, repression of germline-specific genes in somatic cells, RNA interference (RNAi), and X chromosome gene expression [36][37][38][39][40] .
In Drosophila, the DREAM complex encoded by homologs of nematode synMuvB genes represses the expression of G2/M phase genes and also represses ectopic expression of the carbon dioxide receptor in olfactory neurons 10,19,41 . The Drosophila Myb protein is required to relieve this DREAM-mediated repression for mitotic cell cycle progression and for carbon dioxide receptor expression in the appropriate neurons. Drosophila Myb also acts in opposition to the DREAM complex to regulate chorion gene amplification in ovarian follicle cells and programmed neuronal cell death 12,42,43 . Interestingly, recent studies in C. elegans have shown that the MuvB complex can effectively repress gene expression in the absence of the LIN-35 RB-family protein that was previously thought to be required for repression by DREAM complexes 44 . Although C. elegans and other nematode species contain two Myb-related genes that encode homologs of the CDC5/CEF1 splicing factor and the SNAPc small nuclear RNA transcription factor, they do not contain an animal-type Myb gene that might relieve repression by DREAM complexes 3 .
The animal-specific carboxy-terminus of Drosophila Myb is both necessary and sufficient for binding to the MuvB core complex in cell lysates 11

Evolutionary Conservation of Animal-Type Myb Proteins
Searches of public sequence repositories revealed that animal-type Myb proteins, characterized by an amino-terminal DNA-binding domain composed of three tandem Myb repeats, a central proline-rich "hinge", and a carboxy-terminal MuvB-binding domain, are present in species of all phyla of the superphylum Deuterostomia, including Chordata (human, lancelet, sea squirt), Hemichordata (acorn worm), and Echinodermata (sea urchin) (Figures 1 and   2). Animal-type Myb proteins, defined as having these three domains, are also present in species

Evolutionary Conservation of the Myb-Binding Domains of LIN9 and LIN52
Searches of public sequence repositories revealed that the Myb-binding domain of LIN9 proteins is conserved in a wide range of species within the Metazoa, including those in the Nematoda that lack an animal-type Myb protein (Figures 1 and 3). Sequences homologous to the Myb-binding domain of LIN9 were also identified in species within the Planta, Amoebozoa, and presence of both LIN9 and LIN52 Myb-binding domains, again suggests that these domains have an additional evolutionarily conserved function.

The MuvB-Binding Domain of Drosophila Myb Can Bind to a Nematode LIN9-LIN52
Although no sequenced species of Nematoda contains an animal-type Myb protein, they nevertheless do contain conserved Myb-binding domains in their LIN9 and LIN52 family proteins (Figures 1 and 3). Furthermore, the amino acids in the human LIN9 and LIN52 proteins that make direct contacts with human B-Myb in a structure determined by X-ray crystallography are well-conserved in the LIN9 and LIN52 proteins of the intensively studied nematode

Expression of Drosophila Myb in C. elegans Causes a Synthetic Multivulval Phenotype
The lin-9 gene was discovered in C. elegans because a loss-of-function mutation cooperated with a second loss-of-function mutation in lin-8 to cause a synthetic multivulval (synMuv) phenotype 32 . Additional genetic screens identified a group of genes (class A synMuv) for which a loss-of-function mutant could cooperate with a lin-9 mutant to cause a synMuv phenotype. A second group of genes (class B synMuv) including both lin-9 and lin-52 were identified for which a loss-of-function mutant could cooperate with a class A synMuv loss-offunction mutant to cause a synMuv phenotype 33  To test this hypothesis, GFP::Myb fusion proteins or a GFP-only (GFP) control protein were expressed in C. elegans under control of a heat-shock promoter using stably integrated single-copy transgenes. This was accomplished using the CRISPR-Cas9 system to promote transgene integration at a specific site on chromosome II 50 A mutant of Drosophila Myb protein lacking its DNA-binding domain (C-term) was previously shown to be capable of physically interacting with the Drosophila MuvB core complex 11 . Furthermore, in Myb null mutant flies, the C-term Myb mutant protein localizes to the cell nucleus, localizes to chromatin, activates the expression of G2/M phase genes, rescues G2/M phase cell cycle progression, and rescues adult viability at low temperatures 10,11 . The Cterm Drosophila Myb mutant protein also caused an elevated incidence of the synMuv phenotype when in combination with a lin-15A mutation, albeit less efficiently (approximately 10% incidence) than the full-length Myb protein (approximately 20% incidence) ( Figure 6).
Neither Drosophila Myb protein caused a multivulval phenotype in worms that were wild-type for lin-15A. These results show that the Drosophila Myb protein in combination with loss of lin-15A causes a synMuv phenotype in C. elegans, which itself has no animal-type Myb gene or protein of its own Furthermore, as was previously observed in Drosophila, the highly conserved Nematodes appear to have lost their animal-type Myb genes and proteins after their divergence from other clades of modern Metazoa (Figures 1 and 2). Nevertheless, the Myb-binding domains of nematode LIN9 and LIN52 have been highly conserved over more than 500 million years in the absence of Myb (Figures 1 and 3). Remarkably, the LIN9-LIN52 Myb-binding domain of C.  Figure   S4). However, this promoter may not function sufficiently early during development to turn on germline genes in somatic cells 52 . It is also possible that sustained rather than transient expression of Myb in a somatic tissue would be required to mimic this phenotype of endogenous synMuvB loss-of-function mutants 36,37 .
The conservation in C. elegans of the amino acids in LIN9 and LIN52 that contact Myb in the crystal structure of their human homologs suggests that there may be another as yet unknown protein in nematodes (and perhaps other species) that can bind to the same LIN9-LIN52 structure in order to activate genes that are repressed by MuvB and DREAM complexes ( Figure 7). Furthermore, the ability of C. elegans LIN9 and LIN52 to bind Drosophila Myb with a similar affinity and to discriminate among Myb mutants in a fashion similar to Drosophila LIN9 and LIN52 suggest a strong selective pressure during the evolution of the Myb-less Nematoda to retain the amino acids that directly contact Myb in humans and Drosophila.
It is possible that the amino acids in LIN9 and LIN52 that contact Myb are also essential for the structural integrity of these proteins, thus providing another explanation for their evolutionary conservation. In addition, the LIN9-LIN52 heterodimerization interface may be Vorster, et al, page 15 highly conserved because it is essential for incorporation of these proteins into the MuvB complex 45 . On the other hand, the presence of a conserved LIN9 Myb-binding domain in species that have neither an animal-type Myb protein nor a LIN52 protein suggests that this domain of LIN9 may also interact directly with other proteins (Figures 1 and 3).
Genes encoding components of the Myb-MuvB and DREAM complexes are frequently altered in human cancer. For example, 47% of a series of 2051 primary breast cancers were found to contain mutations in one of more of these genes ( Figure S1). Although the MYBL2 gene encoding the B-Myb protein is altered in only 4% of breast cancers, increased levels of expression of this gene occur more frequently, particularly in basal-like and triple-negative (ER-, PR-, HER2-) breast cancers that generally have a poor prognosis 27,29 . Indeed, MYBL2 is one of a small number of genes included in the Oncotype DX gene expression test that is widely used to predict clinical outcomes and plan treatment for patients with breast cancer 28,53 .
The remarkable conservation of the Myb-binding domains of LIN9 and LIN52 described above suggests that it might be difficult to develop resistance to therapeutic drugs that target this  Figures 1-3). These data in turn led to inferences about selective pressures for retention of these domains during long periods of evolution. These data also led to direct tests of whether a long lost "key" (the animal-type Myb protein of Drosophila) is capable of opening an ancient "lock" in vivo (the DREAM complex of C. elegans) despite over 500 million years of evolution in the absence of this "key" (Figure 7).

Database Searching and Sequence Alignment
Homologous protein sequences were identified by BLASTp or tBLASTn searches of the non-redundant protein or nucleotide sequence databases at NCBI as of August 2019

Recombinant Protein Production and Binding Assays
LIN9 and LIN52 Myb-binding domains were co-expressed in E. coli 45 . The open reading frames were synthesized as gBLOCK cassettes (IDT, Coralville, IA), cloned into the following plasmids, then verified by DNA sequencing. LIN9 (C. elegans residues 442-559 or D. melanogaster residues 571-699) was expressed from a pRSF plasmid without an affinity tag.

Transgenic Nematode Production
Plasmids containing transgenes for expression in C. elegans were constructed using the Gateway system to join four different elements 64

Phenotypic Analysis of Nematodes
Young adult hermaphrodites were incubated at 22 0 C on 10cm NGM plates with an OP50 bacterial lawn. After depositing embryos for 24 hours, adults were removed with an aspirator.               Vorster, et al, Figure S4 Part C