The adhesion GPCR Gpr56 regulates oligodendrocyte development via interactions with Gα12/13 and RhoA

In the vertebrate central nervous system, myelinating oligodendrocytes are postmitotic and derive from proliferative oligodendrocyte precursor cells (OPCs). The molecular mechanisms that govern oligodendrocyte development are incompletely understood, but recent studies implicate the adhesion class of G protein-coupled receptors (aGPCRs) as important regulators of myelination. Here, we use zebrafish and mouse models to dissect the function of the aGPCR Gpr56 in oligodendrocyte development. We show that gpr56 is expressed during early stages of oligodendrocyte development. Additionally, we observe a significant reduction of mature oligodendrocyte number and of myelinated axons in gpr56 zebrafish mutants. This reduction results from decreased OPC proliferation, rather than increased cell death or altered neural precursor differentiation potential. Finally, we show that these functions are mediated by Gα12/13 proteins and Rho activation. Together, our data establish Gpr56 as a regulator of oligodendrocyte development.

Oligodendrocytes form the myelin sheath by extending and iteratively wrapping their plasma membranes around axons, while glial cell cytoplasm is extruded to ultimately form compact myelin 4 . Impaired myelination causes debilitating symptoms in many severe neurological disorders such as multiple sclerosis, and loss of mature oligodendrocytes and myelin can lead to neuronal death and eventual paralysis. Therefore, continued research to uncover new pathways that promote oligodendrocyte development and remyelination is essential 5 .
Although myelin is required for human life, the molecular mechanisms that underlie glial cell development and myelination are incompletely understood. Previous studies indicate that extracellular matrix proteins and their receptors, in addition to intracellular signal transduction cascades, are essential for proper development of myelinating glial cells [6][7] . For this reason, adhesion G protein-coupled receptors (aGPCRs) represent good candidate regulators of glial cell development and myelination because they are purported to regulate both cell-cell and cell-matrix interactions as well as signal transduction by heterotrimeric G proteins 8 , and recent studies implicate aGPCRs as important regulators of myelination [9][10][11][12][13] . Indeed, the aGPCR Gpr126 is essential for SC, but not oligodendrocyte, myelination 10-13 . aGPCRs are the second largest class of GPCRs, but remain relatively understudied due to their structural complexity. In addition to the canonical seven-transmembrane (7TM) domain common to all GPCRs that usually facilitates intracellular signaling, aGPCRs are characterized by the presence of a very large N-terminal domain rich in functional motifs associated with cell-cell and cell-matrix adhesion, which is separated from the 7TM by a GPCR proteolytic site (GPS) 14 . The GPS motif is encompassed by a larger GPCR autoproteolysis-inducing (GAIN) domain 15 . Together, these domains are required for autoproteolytic cleavage of the N-terminus from the C-terminus, a process that may be required for proper intracellular aGPCR trafficking and signaling [15][16] . The dual roles of aGPCRs in promoting cell-cell/cell-matrix interactions and facilitating intracellular signaling, in addition to the importance of Gpr126 for SC myelination [10][11][12][13] , led us to hypothesize that additional aGPCRs regulate glial cell development and myelination in the CNS.
Previously published datasets indicate that Gpr56, an aGPCR related to Gpr126, is highly expressed during mouse CNS development in OPCs [17][18] . In addition, mutations in GPR56 in humans cause bilateral frontoparietal polymicrogyria (BFPP) 19 , a cortical brain malformation associated with a variety of devastating neurological impairments such as epilepsy and mental retardation. Interestingly, brains of BFPP patients also exhibit reduced white matter volume by MRI [20][21][22] , indicative of myelin defects. Taken together, these data suggest that Gpr56 may function in oligodendrocyte development. To test this hypothesis, we utilize zebrafish and mouse model systems to explore the necessity of Gpr56 in the CNS. We show that impaired Gpr56 function results in reduced numbers of myelinating oligodendrocytes and hypomyelination due to decreased OPC proliferation rather than failed specification of the oligodendrocyte lineage. Additionally, transient overexpression of gpr56 in wildtype (WT) zebrafish increases numbers of pro-OLs but inhibits terminal differentiation of mature oligodendrocytes. Finally, we show that these functions are mediated by interactions with Gα 12/13 G-proteins and RhoA signaling. Together, our studies implicate Gpr56 as an essential regulator of oligodendrocyte development and myelination.

gpr56 is expressed during oligodendrocyte development
Myelin is an evolutionary innovation of the jawed vertebrate lineage 23 , and zebrafish represent one of the simplest genetically tractable organisms in which to study myelinating glial cell development. We therefore took advantage of the zebrafish model system to interrogate the necessity of Gpr56 function in oligodendrocyte development and myelination. First, we analyzed the spatiotemporal expression of gpr56 by RT-PCR in developing zebrafish. We detected gpr56 expression during embryogenesis and in early larval stages (2 through 5 days post-fertilization (dpf)) when OPCs are first specified and differentiate 24 (Fig. 1a). In addition, whole-mount in situ hybridization (WISH) revealed low levels of gpr56 expression throughout the CNS at 1 dpf, prior to OPC specification 24 (Fig. 1b). These findings are consistent with previously published data in mammals showing that Gpr56 is also expressed in astrocytes, and to a lesser extent, in neurons 18 . At 2 dpf, when OPCs have been specified and are actively migrating and dividing 24 , there was a dramatic increase in gpr56 expression in the CNS (Fig. 1c). Cross-sections of 2 dpf WISH embryos revealed that gpr56 is expressed along the spinal cord midline and in a single distinct band running dorsal to ventral on either side of the midline, consistent with expression in neural precursors (Fig. 1d). In contrast, gpr56 expression was not detected by WISH in the CNS at larval stages >3 dpf (Supplementary Fig. 1a-b), during which time oligodendrocytes undergo terminal differentiation [24][25] . We also never detected signal in sense control WISH experiments ( Supplementary Fig. 1c). These findings demonstrate that gpr56 expression during CNS development is conserved from zebrafish to mammals and suggest that levels are highest at earlier stages.

Targeted disruption of gpr56 in zebrafish
Given the expression pattern of gpr56 in the developing CNS, we sought to test the hypothesis that Gpr56 regulates oligodendrocyte development. To this end, we generated Transcription Activator-Like Effector Nucleases (TALENs) 26 designed to cleave zebrafish gpr56 genomic sequences between exons 8 and 9 (Fig. 1e), which encode the GPS motif. We recovered two mutant alleles of gpr56: stl13 representing a 6 bp deletion that is predicted to remove a completely conserved tryptophan residue, and stl14 representing a 26 bp deletion ( Fig. 1f-g). The GPS is a cysteine-rich domain containing four cysteine residues and two tryptophan residues required for autoproteolytic cleavage 27 . Previous reports show that inhibiting autoproteolytic cleavage of GPR56 prevents receptor trafficking and therefore impairs signaling 20 . In addition, two missense mutations that disrupt the highly conserved cysteine and tryptophan residues in humans cause BFPP 27 . Together, these findings suggest that gpr56 stl13/stl13 mutants may generate a non-functional Gpr56 due to failed trafficking of the receptor to the plasma membrane. Therefore, we chose to primarily characterize and utilize stl13 as a tool for understanding gpr56 function.

Gpr56 is required for proper oligodendrocyte development
To determine whether Gpr56 is required for normal oligodendrogliogenesis, we performed marker analyses to observe oligodendrocyte development in gpr56 stl13/stl13 mutants. In the CNS, oligodendrocyte precursor cells (OPCs) differentiate from neural progenitors that reside within subventricular zones of the embryonic brain and spinal cord through coordinated expression of the transcription factors Olig2 and Sox10 28 . Some OPCs remain as non-myelinating OPC progenitors throughout life, while others are further specified by expression of the transcription factor Nkx2.2, which allows OPCs to enter the terminal differentiation pathway. Expression of Nkx2.2 is characteristic of axon-associated prooligodendrocytes 25 , and mature myelinating oligodendrocytes express Myelin basic protein (Mbp), which encodes a structural component of the myelin sheath 28 . Using transgenic sox10(−7.2):memgfp zebrafish (a kind gift from M. Voigt, manuscript in preparation), we did not detect any difference in the number of sox10+ OPCs in gpr56 stl13/13 larvae (N=19) compared to controls (N=20) at 28 hpf, the earliest time-point at which sox10 expression was detected ( Fig. 2a-c, p<.09, Student's t-test). These data suggest that specification of the oligodendrocyte lineage is not impaired in gpr56 stl13/stl13 mutants. Consistent with this observation, general dorsal-ventral patterning (visualized by Sonic hedgehog pathway markers) was also unaffected in gpr56 stl13/13 mutants ( Supplementary Fig. 2). Furthermore, we did not observe overt defects in the development of neurons, radial glia, or in microglia number at any stage examined in gpr56 stl13/13 mutants compared to controls ( Supplementary  Fig. 3). However, by 2.5 dpf, we detected a significant decrease (p<5.223×10 −6 , Student's t-test) in nkx2.2a+ pro-oligodendrocyte (pro-OL) number in gpr56 stl13/13 mutants (N=19) by WISH relative to controls (gpr56 +/+ and gpr56 stl13/+ : N=22), and we observed a subsequent reduction in CNS mbp expression in gpr56 mutants (gpr56 stl13/stl13 : N=20/27, gpr56 stl14/stl14 : N=14/19) by 4 dpf (Fig. 2d-i, WT control: N=13/13). Although gpr56 is maternally expressed (Fig. 1a), we did not observe enhanced defects in oligodendrocyte development in gpr56 stl13/stl13 maternal-zygotic mutants (no WT maternal contribution) compared to gpr56 stl13/stl13 zygotic mutants (Fig. 2j, N=5/5). Importantly, injection of synthetic mRNA encoding WT zebrafish (50 pg) or mouse (200 pg) Gpr56 was sufficient to rescue the CNS mbp deficits in gpr56 stl13/stl13 mutants, confirming the specificity of the TALEN-induced mutations ( Supplementary Fig. 4). Together, these results demonstrate that Gpr56 is required for the proper development of mature, myelinating oligodendrocytes.

Mutations in gpr56 cause CNS hypomyelination
To further delineate how impaired Gpr56 function affects oligodendrocyte myelination, we performed transmission electron microscopy (TEM) to measure the extent and quality of myelin present in the ventral spinal cord of gpr56 mutants (gpr56 stl13/stl13 : N=5, gpr56 stl14/stl14 : N=4, WT: N=6) during development ( Fig. 3a-k). We found a significant reduction (gpr56 stl13/stl13 , p<.019; gpr56 stl14/stl14 , p<.039, Student's t-test) in the percent of myelinated axons in mutant larvae compared to WT controls at 5 dpf (Fig. 3l), but total axon number was unchanged (gpr56 stl13/stl13 , p<.5; gpr56 stl14/stl14 , p<.78, Student's t-test Fig.  3m). We also did not observe a change in the number of myelin wraps surrounding those axons that were myelinated in gpr56 stl13/stl13 mutants compared to controls (Fig. 3n- . In contrast to the hypomyelination phenotype shown in the ventral spinal cord of gpr56 mutants at 5 dpf ( Fig. 3f-k), we did not observe any effect on oligodendrocyte myelination in the dorsal spinal cord at this developmental stage ( Supplementary Fig. 5). By 21 dpf, however, we observed a statistically significant decrease in the percent of myelinated axons in the ventral (p<.007, Student's t-test) and dorsal (p<.002) spinal cord of gpr56 stl13/stl13 (N=4) mutants relative to WT siblings (N=2), with no change in axon number ( Supplementary Fig. 6, dorsal spinal cord: p<.89, ventral spinal spinal: p<.96, Student's t-test). We suspect that we were unable to detect a difference in the dorsal spinal cord at larval stages because so few axons are myelinated in this region at 5 dpf. Interestingly, we also observed many oligodendrocytes with distended endoplasmic reticula in gpr56 stl13/stl13 mutants at 21 dpf, which was rarely seen in controls, supporting the hypothesis that the protein encoded by gpr56 stl13/stl13 does not traffic properly from the ER to the plasma membrane ( Supplementary Fig. 6). To determine if myelination is simply delayed in gpr56 stl13/stl13 mutants, we analyzed myelin ultrastructure in the spinal cord of gpr56 stl13/stl13 animals at 6 months of age ( Supplementary Fig. 7). At 6 months, we likewise found a statistically significant decrease in the percent of myelinated axons in gpr56 stl13/stl13 mutants (N=4) compared to WT controls (N=3) in both the ventral ( Supplementary Fig. 7g, axon diameter of .2-.5 µm (p<.021), .5-1 µm (p<.005), and 1-2 µm (p<.029), Student's ttest) and dorsal spinal cord ( Supplementary Fig. 7n, axon diameter of .2−.5 µm (p<.003) and .5-1 µm (p<.008), Student's t-test), without any change in myelin thickness for axons that were myelinated ( Supplementary Fig. 7h,o, linear regression analysis). Axon number was also unchanged in mutants compared to controls ( Supplementary Fig. 7f,m, dorsal spinal cord: p<.11, ventral spinal cord: p<.34, Student's t-test). Collectively, our analysis demonstrates that altered Gpr56 function causes developmental CNS hypomyelination that persists in the adult spinal cord.

Gpr56 is not required for neural stem cell differentiation
One possible explanation for the observed decrease in myelinating oligodendrocyte numbers in gpr56 stl13/stl13 mutants could be reduced differentiation of the oligodendrocyte lineage from neural stem cells. The fact that we did not detect defects in early OPC specification ( Fig. 2a-c) in gpr56 stl13/stl13 zebrafish mutants argued against this, but to definitively test this model, we harvested neurospheres from the subventricular zone of Gpr56−/− knockout mice 29 at postnatal day 3 and assessed their ability to differentiate into Tuj1+ neurons, GFAP+ astrocytes, and O4+ oligodendrocytes. Importantly, we found robust levels of Gpr56 protein in neurospheres harvested from WT and Gpr56 +/− animals, which was absent from mutant neurospheres ( Supplementary Fig. 8). We observed no differences in the ability of Gpr56−/− neural stem cells to differentiate into any cell type tested (Fig. 4), consistent with our analysis of gpr56 stl13/stl13 zebrafish mutants (Fig. 2a-c; Supplementary Fig. 3). Of note, neurospheres harvested from the spinal cord of Gpr56−/− animals showed a slightly increased potential to differentiate into neurons ( Supplementary Fig. 9), highlighting the variation that exists between neural progenitor populations from different tissues 30 . Importantly, neurospheres harvested from both tissues showed no difference in their ability to differentiate into oligodendrocytes compared to controls, supporting the hypothesis that Gpr56 is not required for specification and differentiation of the oligodendrocyte lineage.

Gpr56 regulates OPC proliferation
Because OPC specification is unaffected in gpr56 stl13/stl13 zebrafish mutants, and loss of Gpr56 does not affect the ability of neural precursors to differentiate into oligodendrocytes, reduced pro-OL numbers in gpr56 stl13/stl13 mutants could result from increased death and/or decreased proliferation of OPCs. To distinguish between these possibilities, we performed time-lapse imaging of gpr56 stl13/stl13 mutants and controls using transgenic sox10(−7.2):mRFP 24 zebrafish to visualize OPC behavior and development in vivo (Supplementary Movies 1-2). Embryos were imaged from 30 hpf to 46 hpf, allowing us to monitor the transition of many OPCs into pro-OLs. At 30 hpf, we found a statistically significant increase in the number of axons associated with pro-OLs in gpr56 stl13/stl13 mutants (N=7) compared to controls (N=7, p<.04, Student's t-test). Over time, however, the number of pro-OL associated axons in control embryos increased at a higher rate than in gpr56 stl13/stl13 mutants. This increase resulted in a statistically significant reduction in pro-OL associated axons in gpr56 stl13/stl13 mutants compared to controls by 46 hpf (Fig. 5a-c, p<.004, Student's t-test), consistent with our marker analysis ( Fig. 2d-f). These findings suggest that in gpr56 stl13/stl13 mutants, OPCs are born and prematurely form associations with axons, but that they do not proliferate sufficiently to account for the total number of axons that need to be myelinated. In accordance with this hypothesis, we also observed additional myelinated axons in the ventral spinal cord of gpr56 stl13/stl13 mutants (N=3) compared to controls (WT: N=3, gpr56 stl13/+− : N=3) by TEM at 3 dpf ( Fig. 5d-g, p<.048, Student's t-test), indicative of precocious myelination. In addition, we observed a statistically significant reduction in the number of proliferating OPCs (sox10+; pH3+) in gpr56 stl13/stl13 mutants compared to controls at 32 hpf, when most OPCs have not yet entered the differentiation program [24][25] (Fig. 5h-j, p<.0004, Student's t-test). Importantly, we did not observe any signs of cell death via time-lapse imaging, and there was no increase in the number of cell corpses assessed by acridine orange staining at 32 hpf, 2 dpf, or 3 dpf in gpr56 stl13/stl13 mutants compared to controls ( Fig. 5k-m). Thus, changes in pro-OL number can only be attributed to changes in OPC proliferation, and not cell death. Together, these data support a model in which Gpr56 promotes OPC proliferation at the expense of differentiation such that when Gpr56 function is impaired, insufficient OPC proliferation results in fewer pro-OLs and mature oligodendrocytes, which leads to hypomyelination.

gpr56 OE increases OPC number but inhibits myelination
Impaired Gpr56 function results in reduced numbers of myelinating oligodendrocytes and myelin due to decreased OPC proliferation ( Fig. 2 Fig. 4) by suppressing the OPC proliferation defect observed in gpr56 stl13/stl13 mutants ( Supplementary Fig. 10). Interestingly, we also observed enhanced mbp expression in a large fraction of gpr56 injected embryos ( Supplementary Fig. 4d, e). We therefore hypothesized that transient OE of gpr56 in WT embryos would cause hypermyelination by increasing oligodendrocyte number. Analysis of nkx2.2a expression revealed a statistically significant increase in the number of nkx2.2a+ pro-OLs in the anterior (p<8.7×10 −7 , Student's t-test) and posterior (p<.01, Student's t-test) spinal cord ( Fig. 6a-f) of embryos injected with 50 pg of WT gpr56 synthetic mRNA (anterior: N=25, posterior: N=29) relative to control injected embryos (anterior: N=26, posterior: N=31). In addition, we noted an increase in CNS mbp expression in gpr56-OE embryos relative to phenol-red injected controls by WISH ( Fig. 6g-j, N=19/34) at 2.5 dpf. Consistent with these findings, we observed an increase in the numbers of oligodendrocyte lineage nuclei in the spinal cord of gpr56-OE embryos relative to phenol-red injected controls by TEM at 2.5 dpf and 5 dpf (Fig. 6l-q; p<.037 at 2.5 dpf(OE: N=4, control-injected: N=4); p<.15 at 5 dpf(OE: N=5, control-injected: N=4), Student's t-test). We distinguished oligodendrocyte and OPC nuclei from neuronal nuclei by their distinctive heterochromatin conformations. OPC and oligodendrocyte heterochromatin is heavily condensed and localized along the perimeter of the nucleus for OPCs or in patches throughout for post-mitotic, terminally differentiated oligodendrocytes. In contrast, neuronal heterochromatin is organized uniformly throughout the nucleus, often with a single dense nucleolus [31][32][33] . At 2.5 dpf, generally only the large Mauthner axons are myelinated in WT zebrafish 34 , but in some OE embryos, additional myelinated axons were observed compared to controls at this time point by TEM (N=2/4 in the ventral spinal cord and N=1/4 in the dorsal spinal cord, Supplementary Fig. 11), although the difference was not significant. Taken alone, these data suggest that OE of gpr56 may induce hypermyelination; however, we observed hypomyelination in gpr56-OE embryos relative to controls at 5 dpf by TEM ( Fig. 6r-x). These findings indicate that OE of gpr56 increases mbp expression by increasing pro-OL number but that this mbp expression is not equivalent to true myelination, and OE of gpr56 inhibits terminal differentiation and myelination. Collectively, these data support the hypothesis that Gpr56 is both a positive regulator of OPC proliferation and a negative regulator of oligodendrocyte differentiation.

Gα 12/13 and RhoA manipulations alter gpr56 mutant phenotypes
Previous in vitro studies show that Gpr56 couples to Gα 12/13 and activates RhoA to control migration of mammalian neural precursor cells [35][36] . We likewise found that RhoA activity levels are high in control neurospheres when Gpr56 is abundant (Supplementary Figure 12). In OPCs, RhoA is highly expressed, but is later downregulated to facilitate terminal differentiation of oligodendrocytes in mammals 37 , which is consistent with the spatiotemporal expression profile we observe for gpr56 in zebrafish ( Fig. 1a-d). Taken together, these data suggest that Gpr56 may couple to Gα 12/13 and activate RhoA during oligodendrocyte development. To test these hypotheses, we first used established antisense morpholino oligonucleotides that transiently impair Gα 12/13 function in gpr56 stl13/stl13 mutants through translational inhibition of three previously identified Gα 12/13 transcripts ("3MO"; gna12, gna13a, and gna13b) 38 . We then analyzed mbp expression by WISH to assess oligodendrocyte development, and embryos were scored from 0-3 based on the strength of their expression: 0 = nearly absent, 1 = strongly reduced, 2 = reduced, 3 = wildtype ( Fig. 7a-d). We treated embryos with low doses of 3MO (1 ng each) such that WT, 3MO-injected embryos showed no phenotype (Fig. 7e-f). In contrast, sub-threshold doses of 3MO caused significant hypomyelination in gpr56 stl13/+ heterozygotes and severe hypomyelination in homozygous mutants, resulting in a dramatic shift in the distribution of their phenotypic classes and a significant reduction in their CNS mbp score ( Fig. 7e- . OE of rhoa also caused a dramatic shift in the distribution of phenotypic classes in gpr56 stl13/stl13 mutants such that the majority (76%) of gpr56 stl13/stl13 larvae expressed WT levels of mbp (Fig. 7i, N=13/17). Importantly, all embryos analyzed for these studies were morphologically normal, and thus changes in mbp expression could not be attributed to changes in overall heath in treated animals compared to controls ( Fig. 7g-h, k-l). Taken together, these experiments demonstrate that Gpr56 interacts with Gα 12/13 and signals through RhoA to promote oligodendrocyte development and subsequent myelination.

DISCUSSION
The timing of the transition from OPC proliferation to terminal differentiation is tightly regulated [40][41][42] . In the CNS, interactions between receptors on the OPC cell surface and axon-bound or axon-secreted molecules ensure that sufficient numbers of OPCs are generated to accommodate the numbers of axons that need to be myelinated. Availability of the axon-derived mitogen PDGF-A, which interacts with PDGFRα on the OPC plasma membrane, is a major determinant of the extent of OPC proliferation. Moreover, numerous mechanisms converge to ensure the timing and execution of terminal differentiation, including downregulation of Pdgfrα by Nkx2.2 in OPCs, cell-cycle exit, and vast morphological changes 37,[42][43][44][45] . Failed differentiation of resident OPC pools in response to demyelination is a major contributor to the degenerative nature of demyelinating diseases such as multiple sclerosis 46 ; thus, identification of signaling molecules that regulate OPC behavior and differentiation could provide new avenues for therapeutics.
In this study, using zebrafish and mouse models, we identified Gpr56 as a new regulator of OPC proliferation and oligodendrocyte myelination. Like PDGFRα,Gpr56 expression is restricted to early stages of oligodendrocyte development and is downregulated prior to terminal differentiation 18,42 , suggesting that Gpr56 may negatively regulate terminal differentiation and/or positively regulate OPC behavior/proliferation. Gpr17, another GPCR that is essential for oligodendrocyte development, is also expressed in OPCs and is downregulated in mature oligodendrocytes, and Gpr17 has been shown to directly inhibit terminal differentiation. In the absence of Gpr17, OPCs prematurely differentiate and form myelin, though myelination eventually proceeds to WT levels in Gpr17−/− mice 47 . We found that gpr56 stl13/stl13 mutant OPCs likewise formed associations with axons and myelinated prematurely, indicative of precocious differentiation due to relieved inhibition on terminal differentiation We found that OPCs are specified normally in gpr56 stl13/stl13 mutants but are less proliferative, resulting in reduced numbers of myelinating oligodendrocytes and consequently, hypomyelination. In addition, transient OE of gpr56 in WT embryos caused increased numbers of pro-OLs and increased mbp expression in gpr56 OE embryos at 2.5 dpf,. However, the apparent increase in mbp expression is likely the result of greater numbers of pro-OLs, as myelination was inhibited in the OE embryos at 5 dpf as assessed by TEM. We also observed many immature oligodendrocyte cell bodies with no axonal contacts in the center of the spinal cord of the overexpression embryos at this time-point. These findings indicate that that prolonged OE of gpr56 inhibits OPC differentiation and myelination. Together, these findings support the hypothesis that Gpr56 is both a positive regulator of OPC proliferation and an inhibitor of oligodendrocyte terminal differentiation.
Though we observed striking defects in oligodendrocyte development and myelination when Gpr56 function is perturbed, we did not observe any defects in the development of neurons, radial glia, or in microglia number in gpr56 stl13/stl13 mutants at any time point analyzed. These data suggest that the observed defects in oligodendrocyte development are a cellautonomous effect, although conditional mouse mutant studies could be leveraged to definitively address the question of cellular autonomy. Together, these data demonstrate that Gpr56 is a positive regulator of OPC proliferation and a negative regulator of OPC differentiation.
RhoA is the major downstream effector of Gα 12/13 and its expression is likewise restricted to early stages of oligodendrocyte development 37,39 . Inactivation of RhoA via Fyn is required for OPC morphological differentiation into pro-OLs, and sustained overexpression of RhoA inhibits process extension and differentiation of oligodendrocytes 37 . Reduction of Gα 12/13 function in gpr56 stl13/stl13 zebrafish mutants and heterozygotes causes enhanced loss of CNS mbp expression. In addition, transient OE of constitutively active RhoA prior to terminal differentiation suppressed gpr56 stl13/stl13 defects. In sum, these data support a model in which Gpr56 function in OPCs is mediated via interactions with Gα 12/13 and RhoA signaling.
RhoA has not been shown to directly promote OPC proliferation, though our findings do not preclude this possibility. We therefore propose a model in which Gpr56 interacts with Gα 12/13 and activates RhoA to prevent OPC differentiation, which facilitates continued OPC proliferation promoted by axon-glia interactions such as those mediated by PDGFRα and PDGF-A [42][43][44] . When Gpr56 function is perturbed, RhoA activation is therefore reduced, relieving RhoA inhibition of terminal differentiation and causing premature OPC differentiation that manifests in a reduced proliferation phenotype (Fig. 8). In this model, the OPCs that prematurely undergo terminal differentiation should myelinate normally. Accordingly, we detected an increase in axons associated with pro-OLs at 30 hpf and increased myelination at 3 dpf, consistent with precocious differentiation. Furthermore, we observed no myelin abnormalities or reduced myelin thickness in those axons that are myelinated in gpr56 stl13/stl13 mutants. In sum, these data support the model that Gpr56 positively regulates OPC proliferation via RhoA activation and inhibition of terminal differentiation.
At present, multiple ligands of Gpr56 have been identified in other cell types. In cancer studies, Gpr56 has been shown to inhibit melanoma cell growth via interactions with tissue transglutaminase (TG2) 48 . In the developing brain, collagen type III, alpha-1secreted by meningeal fibroblasts interacts with neuron-bound Gpr56 to facilitate proper neuronal migration 36 . Impaired differentiation of OPC progenitors is a major cause of failed remyelination in the CNS of patients suffering from demyelinating diseases such as multiple sclerosis 46 . Therefore, a better understanding of Gpr56 activation and signaling is crucial as targeted inhibition of this receptor could be used to promote OPC differentiation and remyelination after injury and disease-induced demyelination. Future studies are required to identify the endogenous ligand(s) and small molecules that modulate Gpr56 signaling in OPCs.

Zebrafish stocks and rearing conditions
Zebrafish (Danio rerio) were maintained in the Washington University Zebrafish Consortium facility (http://zebrafish.wustl.edu/husbandry.htm). All experiments were performed in compliance with Washington University's institutional animal protocols. Embryos were collected from pair-wise or harem matings and reared at 28.5° C in egg water (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl 2 , 0.33 mM MgSO 4 ). Embryos were staged in hours post-fertilization (hpf) or days post-fertilization (dpf) using morpholical markers 49 .

Genotyping
To genotype individual larvae for phenotypic analyses, we used the following primers to amplify a 182 bp fragment of gpr56 from genomic DNA: 5'-CATGGGATACAAAACAAGGT −3' and 5'-AGTGGCATTCTGTCTGTTCT -3'. For stl14, PCR products were run out on a 3% agarose gel and the smaller fragment (158 bp) is the mutant allele. For stl13, we performed restriction digest analysis using Hpy166II, which cleaves WT into a 61 bp and a 121 bp fragment, but is unable to cut mutant (176 bp).

TALEN generation and microinjection
TALEN constructs were assembled by the Hope Center for Neurological Disorders at Washington University using the Golden Gate TALEN assembly protocol 53 . The repeat variable diresidue sequences for the left and right TALENs were as follows: Left: NN NG NG NG HD NG HD HD HD NG NI NN NI HD NI NI NG and Right: NI NI HD NG NN NG NI NG HD NN HD NI NN HD HD NI NG HD NI NG. Synthetic TALEN mRNA was then transcribed using the mMESSAGE mMACHINE® SP6 ULTRA kit (Ambion), and RNA from each arm was combined and injected into 1-cell stage WT embryos (AB) at 100 pg quantities. To recover germ-line transmitted mutations, injected founders (F0s) were raised to adulthood, outcrossed to WT AB partners, and genomic DNA was extracted from individual F1 embryos for PCR amplification and restriction digest analysis of the targeted region. Genomic DNA from F1 embryos that showed disruption of the target site was then cloned using the TOPO® TA Cloning® kit (Invitrogen) and sequenced using Sanger sequencing.

Mice
All mouse (Mus musculus) experiments were performed in compliance with Washington University's institutional animal protocols. The Gpr56 knock-out mice were obtained from the Mutant Mouse Regional Resource Centers (MMRRC), were maintained on a C57BL/6 background, and genotyped as described previously 29 . P3 animals of presumed mixed gender were used for all neurosphere experiments, although pups were not genotyped for gender.

RT-PCR
Standard RT-PCR was performed on cDNAs derived from embryos at the 2-cell stage (Maternal), 8 hpf, 24 hpf, 48 hpf, 72 hpf, and 5 dpf. Analysis was performed 3 separate times with similar results. For all stages, cDNA was reverse-transcribed from RNA extracted from pools of 30 embryos with the SuperScript First-Strand Synthesis system for RT-PCR (Invitrogen) using Superscript II reverse transcriptase (RT) and oligo(dT) primers, and genomic DNA was controlled for by omitting RT from reactions using the same RNA.

Cloning
Full length, gpr56 cDNA was amplified from cDNA derived from 5 dpf embryos and cloned into pCS2+ using the following primer set: 5 -GATCTACTCGAGTTGGTAGAAAGTGGACTTCG -3 (XhoI site underlined) and 5 -AGGATATCTAGATTCCCTCACAACTTCACACT -3 (XbaI site underlined). Both vector and insert were digested with XhoI and XbaI and ligated together using standard T4 ligase (NEB), and the final vector (pCS2+_ zf_gpr56) was sequenced for verification.

In situ hybridization
Whole-mount in situ hybridizations were performed using standard protocols 54 . Briefly, embryos were fixed at the desired age with 4% paraformaldehyde for 2 hours and room temperature of 4° C overnight, and then dehydrated in 100% Methanol overnight. Following dehydration, embryos were washed in .2% PBS-Tween (PBSTw), permeabilized in proteinase K (20mg/ul diluted 1:1000 in .2% PBSTw), and incubated with the appropriate Digoxygenin-labeled riboprobe overnight at 65° C in hybridization buffer (50% formamide). Following probe-treatment, embryos were washed to remove formamide, blocked in 2% blocking medium supplemented with 10% normal sheep serum and .2% Triton, and incubated in primary antibody (Anti-Dig, Fab fragments (1:2000), Roche) overnight in block. Following primary antibody treatment, embryos were washed in Maleic Acid Buffer with .2% triton, and developed by alkaline phosphatase treatment. After complete, embryos were post-fixed in 4% paraformaldehyde and stored long-term in 70% glycerol.To synthesize gpr56 riboprobes, a 509 bp region spanning exons 11-13 was PCR amplified from a partially sequenced cDNA clone (Clone ID 7070653, Open Biosystems) using the following primer pair: 5' -ATTTCTGCTGTGCTTGTTCT -3' and 5' -CATGTGGTTCCATACAGACA -3'. The PCR product was cloned into pCR II using the pCR II-TOPO® TA Cloning® kit (Invitrogen) and sequenced. This construct was linearized by NotI and transcribed with SP6 for sense and linearized with BamHI and transcribed with T7 for antisense. All other riboprobes: mbp 55 , nkx2.2a 56 , apoeb 57 , shh 58 , ptc1 59 and ptc2 60 were described previously. Embryos stained with nkx2.2a were mounted in glycerol and nkx2.2a positive cells were counted manually in all focal planes of the spinal cord from segments 5-7 or 15-30. apoeb positive microglia were counted as previously described 57 . All quantifications were performed blinded, and genotypes were obtained following quantification.

Transmission electron microscopy
For zebrafish larvae and 21 dpf juveniles, transmission electron microscopy (TEM) was performed using a PELCO BioWave® Pro with SteadyTemp™ Digital Plus waterrecirculating system to ensure sufficient infiltration of the embryos 61 . Embryos were microwaved for all steps prior to EPON treatment. Briefly, embryos were fixed in Karnovsky's fix (4% paraformaldehyde, 2% glutaraldehyde, .1M sodium cacodylate) and left overnight at 4° C. After fixation, embryos were post-fixed in 2% osmium tetroxide in . 1M sodium cacodylate and .1M imidazole, washed, and treated with saturated uranyl acetate. Embryos were then dehydrated using increasing concentrations of ethanol followed by 100% acetone, and then infiltrated overnight in an acetone-EPON mixture. Embryos were then transferred to 100% EPON and embedded in molds for baking overnight at 65° C. We always examined fish of a given stage at approximately the same body segment level to control for variability along the anterior/posterior axis. For preparation of spinal cords from 6 months old adults, we anesthetized animals in an ice-water bath and then severed the hindbrain before dissecting a portion of the anterior spinal cord, and fixation of adult-tissue occurred as described previously 10 . In brief, adult tissues were fixed in Karnovsky's fix overnight, post-fixed in 2% osmium tetroxide in .1M sodium cacodylate, washed, and dehydrated in increasing concentrations of ethanol followed by 100% propylene oxide. The tissues were then submerged in a 2:1 mixture of propolyene oxide:EPON for 1 hour, followed by a 1:1 mixture overnight. The following day, the tissues were transferred to 100% EPON prior to embedding and baking at 65° C. Thin sections (70 nm) were mounted on mesh (5 dpf and 6 month samples) or slot (21 dpf samples) grids and stained with 8% uranyl acetate and Sato's lead stain. Sections were then viewed on a Jeol (JEM-1400) electron microscope and images were collected with an AMT V601 digital camera. All images were processed and analyzed using Image J software. At 3 dpf and 5 dpf, quantifications were performed on a stereotyped, 14 µm 2 region of the dorsal or ventral spinal cord containing approximately 400 and 250 axons, respectively (see Fig. 3b). At 21 dpf, quantifications were performed on three, 20 µm 2 sections of the ventral spinal cord and two, 20 µm 2 sections of the dorsal spinal cord containing of a minimum of 100 and 300 axons, respectively. Quantifications were performed on stereotyped positions of the spinal cord (see Fig. 3b). At 6 months, quantifications were performed on a stereotyped 17 µm 2 region of the dorsal and ventral spinal cord (Supplementary Fig. 7a). G-ratios were calculated using Adobe Photoshop and Image J software for approximately 50 axons per animal by taking the ratio of the axon diameter to the diameter of the whole myelinated fiber. At 3 dpf, N=3 for WT and gpr56 stl13/+ , N=4 for gpr56 stl13/stl13 . For quantification of gpr56 mutants and controls at 5 dpf, N= 6 for WT controls, N=5 for gpr56 stl13/stl13 mutants, and N=4 for gpr56 stl14/stl14 mutants. At 21 dpf, N=2 for WT controls and N=4 for gpr56 stl13/stl13 mutants. At 6 months, N=3 for WT controls, and N=4 for gpr56 stl13/stl13 mutants.
For quantification of gpr56 injected versus control injected embryos at 2.5 dpf and 5 dpf, we manually counted the number of wraps around Mauthner axons imaged at 10,000X. At 2.5 dpf, we quantified all neuronal/oligodendrocyte-lineage cell bodies in the dorsal half of the spinal cord (located above half the height of the whole spinal cord at 1200X). N=4 for each treatment. At 5 dpf (N=4 for control-injected embryos, N=5 for gpr56 injected embryos), we quantified oligodendrocyte-lineage cell bodies within a stereotyped 30 µm 2 region in the center of the spinal cord. Additional calculations were performed as described above. All quantifications were performed blinded to genotype and treatment.

Neurosphere immunocytochemistry
Immunocytochemistry was performed on trypsinized NPCs grown in 50 µg/mL poly-Dlysine-coated and 10 µg/mL fibronectin-coated 24-well plates containing defined medium [(5:3 mixture of DMEM low glucose (Life Technologies): neurobasal medium (Life Technologies), 0.5 mM 2-mercaptoethanol, 2mM L-glutamine, 5 IU of penicillin, and 5 µg/ml streptomycin (life technologies) supplemented with 1% N2 supplement (Life Technologies) and 2% B27 supplement (Life Technologies) without growth factors. After 5 days in culture, cells were fixed and stained with primary antibodies (Nestin: ab6142 (Abcam), GFAP: G9269 (Sigma-Aldrich), Tuj1: MMS-435P (Covance), and O4: MAB345 (Millipore)) at 4°C overnight. For fluorescence detection, Alexa Fluor-tagged secondary antibodies (Molecular Probes) were used and cells were counterstained with DAPI. Images were acquired using Metamorph software on a Nikon Eclipse TE300 microscope equipped with an optical camera (Optronics). Investigators were blinded to mouse genotypes and percentages of positive cells were calculated using the total number of cells in each image (DAPI nuclear stain). Data were analyzed with Prism 5 software (GraphPad Software) using a two-way ANOVA test as all data met the common assumptions (normally distributed, equal variances, independent samples) required for this test.

Neurosphere western blotting
Western blots were performed as previously described 62 . Images of full western blots can be seen in Supplementary Fig. 8

Time-lapse imaging
Embryos were mounted in 1% low melting agarose with tricaine to prevent movement and . 003% phenylthiourea to prevent pigmentation. Embryos were oriented laterally and all images were taken of spinal cord regions spanning segments 5-7 using a 40x, dry objective. Time-lapse images were taken from gpr56 stl13/stl13 mutants (N=7) and gpr56 stl13/+ controls (N=6) expressing tg(sox10:mRFP) 23 from 30 hpf to 46 hpf using a Quorum spinning disc confocal microscope on a heated stage, recorded using Metamorph software, and analyzed using Image J. All analyses were performed blinded to genotype.

Zebrafish immunohistochemistry and fluorescence imaging
For imaging of transgenes in fixed embryos, embryos were fixed in fresh, 4% paraformaldehyde supplemented with 0.1 M pipes, 1.0 mM MgSO 4 , and 2 mM EGTA in 1x PBS to preserve fluorescence 64 . For phospho-H3 antibody stains (06570MI, Millipore), embryos were fixed in 4% paraformaldehyde in 1X PBS overnight (4°C), washed in .3% PBS/Triton and then blocked for 2 hours in .3% PBS/Triton plus 10% normal goat serum and 4% BSA. Embryos were then incubated overnight at 4° C in primary diluted 1:3000 in blocking solution. For secondary staining, embryos were incubated in Alexa Flour® goat anti-rabbit 488 IgG (Life Technologies) diluted 1:2000 in .3% PBS/Triton plus 2% normal goat serum for 3 hours at room-temperature. Finally, embryos were washed and mounted in vectashield® (Vector Labs) for imaging. All quantifications were performed blinded, and genotypes were obtained following quantification.

Morpholino and synthetic mRNA injections
The morpholinos targeting gna12, gna13a, and gna13b are previously published 38 and were a kind gift from Fang Lin (University of Iowa). Morpholinos were combined and diluted in sterile water supplemented 1:10 with phenol-red dye, and then injected at a final concentration of 1 ng each in a total volume of 1 nl. To control for adverse side-effects from the injection process, we also injected control siblings with an equal volume of phenol-red diluted 1:10 in water. Contstitutively active Rho (rhov14) was kindly provided by Lila Solnica-Krezel and was transcribed and injected as described previously 38 . The full-length mouse Gpr56 cDNA clone was obtained from Open Biosystems (Clone ID: 3709247), linearized with NotI, transcribed using the mMESSAGE mMACHINE® SP6 ULTRA kit (Ambion), and injected at a final concentration of 200 pg in 2 nl. For transcription of pCS2+_ zf_gpr56, the plasmid was linearized with NotI and also transcribed using the mMESSAGE mMACHINE® SP6 ULTRA kit (Ambion) before injection of a total concentration of 50 pg in 3 nl.

Statistical Analyses
All data is presented as the mean + or ± SD. To represent significance: *, p < .05; **, p< . 01; ***, p<.001. A minimum of two technical replicates were performed for each described experiment. Although we did not perform statistical tests to predetermine sample size, our samples sizes are similar to previously published studies. For cell-count, molecular marker, and time-lapse imaging analyses in zebrafish, a minimum of six animals per genotype were used for each described marker, which is in line with previously published works 10,55,57 . Similarly, for TEM analyses, a minimum of two WT and four mutant animals were analyzed at each developmental stage as previously described 10,34 . For neurosphere studies, we used at least three animals per genotype, and for IHC experiments, cell counts were taken from three nonadjacent sections of each well in triplicate 62 . Finally, for injection studies, a minimum of ten animals were analyzed per genotype and treatment group, similar to previously published studies 10,55 . Unless otherwise noted, statistical analyses were performed using an unpaired Student's t-test with 2 tails, unequal variance. With few exceptions, all data sets analyzed by a Student's t-test met the appropriate assumptions, including normality. All datasets with a skewness and kurtosis with an absolute value of ≤ 1 were considered approximately normal. The datasets that did not follow a normal distribution were represented by the following graphs: Fig. 3o, Fig. 7j, and Supplementary  Fig. 3i. These datasets showed low variability, causing a kurtosis value of >1. Because the mean and standard deviation values were nearly identical between both control and experimental groups for these experiments, and we found no significant difference between these groups using the student's t-test, we did not feel that the test was inappropriate for these analyses.      Model of Gpr56 function in OPCs. In WT OPCs, Gpr56 couples to Gα 12/13 and activates RhoA, which prevents terminal differentiation and indirectly promotes OPC proliferation (dashed arrow) by maintaining the OPC in an immature, proliferative state. OPC proliferation is directly regulated via interactions between neuron-secreted PDGF-A and OPC bound PDGFRα. In Gpr56 (stl13/stl13) mutant OPCs, impaired autoproteolytic cleavage reduces Gpr56 signaling capacity, leading to loss of RhoA activation. Reduced RhoA signaling alleviates RhoA inhibition on terminal differentiation (green arrow), which indirectly causes a decrease in OPC proliferation (red dashed lines) as OPCs differentiate at the expense of proliferation.