GABA-A Signaling Maintains Melanocyte Stem Cell Quiescence in Larval Zebrafish

Home , Melanocyte

bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 1

1 GABA-A signaling maintains melanocyte stem cell quiescence in larval zebrafish

3 James R. Allen1*, James B. Skeath1, Stephen L. Johnson1†

5 1Department of Genetics, Washington University School of Medicine, St. Louis, Missouri,

6 63110, USA

8 *Corresponding Author

9 † Deceased

11 Abstract:

12 Adult stem cells (ASCs) contribute to long-term homeostasis and regeneration of many

13 adult tissues. Some ASCs proliferate continuously, others remain quiescent awaiting

14 activation. To identify pathways that regulate ASC quiescence and tissue homeostasis,

15 we study melanocyte stem cells (MSCs) that drive vertebrate pigmentation. In larval

16 zebrafish, MSCs are quiescent, but can be recruited to regenerate the larval pigment

17 pattern following melanocyte ablation. Through pharmacological experiments, we found

18 that inhibition of GABA-A receptor function, specifically the GABA-A rho subtype, induces

19 excessive melanocyte production in larval zebrafish. Conversely, pharmacological

20 activation of GABA-A inhibited melanocyte regeneration. We used CRISPR to generate

21 two mutant alleles of gabrr1, a subtype of GABA-A. Both alleles exhibited robust

22 melanocyte overproduction, while conditional overexpression of gabrr1 inhibited larval

23 melanocyte regeneration. Our data suggest that gabrr1 signaling is necessary and bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 2

1 sufficient to maintain MSC quiescence and prevent excessive pigmentation of the larval

2 zebrafish.

4 Introduction:

5 Vertebrate animals often rely on undifferentiated precursors to regulate growth and

6 homeostasis of specific tissues. These precursors, adult stem cells (ASCs), undergo long-

7 term self-renewal throughout the lifetime of the organism to maintain the growth and

8 regenerative potential of their target tissue. ASCs are found in many tissues including

9 blood (Bertrand, Kim et al. 2007), muscle (Cheung and Rando 2013), skin (Nishimura,

10 Jordan et al. 2002), and nervous system (Ma, Bonaguidi et al. 2009). While some ASCs

11 continuously proliferate to maintain their target tissue, others remain quiescent or dormant

12 and must be recruited in order to enter a proliferative state, often induced by depletion of

13 differentiated cells in their respective tissues (Li and Bhatia 2011). Understanding the

14 pathways that maintain ASC quiescence and that recruit quiescent ASCs to proliferate is

15 critical to elucidate vertebrate tissue growth and homeostasis.

17 Zebrafish pigmentation, specifically melanocyte development, is an excellent model

18 system to dissect the genetic and molecular basis of ASC quiescence and recruitment.

19 Both adult melanocytes (Johnson, Africa et al. 1995) and melanocytes that regenerate

20 (Rawls and Johnson 2000) appear to derive from recruitable melanocyte stem cells

21 (MSCs). For example, genetic studies indicate that the embryonic melanocyte pattern

22 develops from direct-developing melanocytes (Hultman, Budi et al. 2009) and is complete

23 by 3 days post fertilization (dpf). Under normal conditions, few new melanocytes develop bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 3

1 from 3 dpf until the onset of metamorphosis at approximately 15 dpf (Hultman and

2 Johnson 2010). However, when embryonic melanocytes are removed via chemical (Yang

3 and Johnson 2006) or laser (Yang, Sengelmann et al. 2004) treatment during this time, a

4 rapid and complete regeneration of the melanocyte pattern occurs through activation of

5 cell division in melanocyte precursors, MSCs (Yang and Johnson 2006). MSCs normally

6 lie dormant during larval zebrafish pigmentation, but can be recruited upon loss of

7 differentiated melanocytes. The pathways that regulate MSC quiescence and recruitment

8 are poorly understood.

10 Forward genetic studies have helped clarify the genetic regulatory hierarchy that controls

11 melanocyte production and MSC proliferation in zebrafish. These studies highlight the

12 importance of three genes in zebrafish pigmentation – the receptor tyrosine kinase

13 erbb3b, the transcription factor mitfa, and the receptor tyrosine kinase kita. Adult zebrafish

14 mutant for erbb3b, named picasso, exhibit defective melanocyte stripe formation, even

15 though larval erbb3b mutants exhibit a wild-type pigment pattern (Budi, Patterson et al.

16 2008). Critically, when picasso mutant zebrafish are challenged for melanocyte

17 regeneration during larval stages, melanocyte regeneration is completely abrogated

18 (Hultman, Budi et al. 2009), suggesting that regenerating melanocytes require erbb3b

19 function, while early embryonic melanocytes do not. This finding led to a model wherein

20 a subset migratory neural crest cells directly differentiate into embryonic melanocytes

21 (direct-developing melanocytes), and other erbb3b-dependent neural crest cells establish

22 undifferentiated melanocyte precursors, MSCs, that persist throughout zebrafish adult life bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 4

1 and can be recruited to form new (stem-cell derived) melanocytes throughout larval and

2 adult stages.

4 Additional insight into MSCs arose from experiments using a temp-sensitive mutation of

5 the microphthalmia-associated transcription factor (mitf; also referred to as melanocyte

6 inducing transcription factor), which is required for all melanocyte development and

7 survival across vertebrate biology (Lister, Robertson et al. 1999, Levy, Khaled et al.

8 2006). These studies revealed that mitfa function was required for the embryonic pigment

9 pattern, but was not required for the survival of MSCs (Johnson, Nguyen et al. 2011).

10 Therefore, while required for melanocyte survival, mitfa function is not required for the

11 survival of MSCs that can regenerate larval melanocytes and produces the adult pigment

12 pattern.

14 The receptor tyrosine kinase kit plays key roles during zebrafish pigment patterning.

15 Removal of kita function, as seen in the sparse mutant, results in a roughly 50% loss of

16 larval melanocytes, but the adult melanocyte pattern is largely normal (Parichy, Rawls et

17 al. 1999). Thus, kita function is required for the development of direct developing

18 melanocytes. kita does, however, regulate MSC function. For example, kita function is

19 required for melanocyte regeneration during larval stages (O'Reilly-Pol and Johnson

20 2013) and for melanocyte regeneration in the caudal fin at all stages (Rawls and Johnson

21 2001, Rawls and Johnson 2003).

22 bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 5

1 GABA is a major inhibitory neurotransmitter that transduces its signal by binding to and

2 activating GABA receptors, such as the GABA-A receptor class (Bormann 2000). GABA-

3 A receptors are voltage-gated chloride channels. When activated, they allow Cl- ions to

4 move down their electrochemical gradient into the cell, which hyperpolarizes the cell and

5 inhibits action potential propagation along axons (Sigel and Steinmann 2012). Although

6 GABA is best known to function as a neurotransmitter, prior studies indicated that GABA

7 can inhibit the proliferation of murine embryonic stem cells (Teng, Tang et al. 2013) and

8 peripheral neural crest stem cells (Young and Bordey 2009). A role for GABA in regulating

9 vertebrate pigment patterning, however, has not been shown.

11 Here, we show that pharmacological and genetic inhibition of GABA-A receptor function

12 leads to excessive melanocyte production during larval zebrafish development, with the

13 newly produced melanocytes likely arising from MSCs. Conversely, we show that

14 pharmacological or genetic activation of GABA-A signaling inhibits melanocyte

15 regeneration. Our work shows that GABA-mediated signaling promotes MSC quiescence

16 during zebrafish development and highlights the importance of membrane potential and

17 bioelectric sensing in regulating pigment patterning in vertebrates.

19 Methods:

20 Zebrafish Stocks and husbandry

21 Adult fish were raised and maintained at 14L:10D light-to-dark cycle according to

22 previously standardized protocols (Westerfield 2000). To facilitate melanocyte bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 6

1 quantification, the mlpha genotype (Sheets, Ransom et al. 2007) was used as wild-type

2 and all experiments were performed in a mlpha genetic background unless otherwise

3 indicated. To perform our melanocyte differentiation assay, we used mlpha fish carrying

4 Tg(fTyrp1:GFP)j900 (Tryon and Johnson 2012). To genetically ablate melanocytes, we

5 used mlpha fish homozygous for the temperature-sensitive mitfavc7 mutation (Johnson,

6 Nguyen et al. 2011). The kitb5 allele (Parichy, Rawls et al. 1999) in a mlpha background

7 was used in the study, and we used the mlpha background to generate our two CRISPR-

8 based mutations in gabrr1. Embryos of each genotype used in the present study were

9 generated from in vitro fertilization.

10 Pharmacological Reagents and drug screening

11 Drugs used in the study were purchased from commercial vendors (See Table S1). Each

12 drug was handled according to manufacture guidelines, but in general each compound

13 was dissolved in a solvent (dimethyl sulfoxide: DMSO or water) to a stock concentration

14 of 20 mM (if possible). For drug experiments, 10-12 embryos were placed into 24-well

15 plates with approximately 2 ml Egg water (60 mg / L Instant Ocean: 0.06 ppt). The stock

16 solution of each drug was then diluted and added to 2 ml Egg water to the indicated

17 concentration of the experiment to a final vehicle concentration of 0.5% DMSO or water

18 (See Table S1). For our melanocyte differentiation assay, we used phenylthiocarbamide

19 (PTU) at a final concentration of 200 µM in egg water. Experiments were performed as

20 parallel duplicates of each drug treatment.

21 Melanocyte Counting bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 7

1 We focused our analysis of melanocyte development on the larval dorsal stripe. We

2 quantified melanocytes along the dorsum, beginning with the melanocytes located

3 immediately caudal to the otic vesicle and along the stripe to the posterior edge of the

4 pectoral fin. Each larvae was analyzed once as a biological replicant for the indicated

5 experiment, and no further analysis on individual fish as technical replicants were used in

6 this study. Larvae with severe morphological defects, such as cardiac edema, were

7 excluded from analysis in the present study. P-values were calculated with student’s T-

8 test in Microsoft Excel to compare statistical significance between each experimental and

9 control group of larvae.

10 Generation of CRISPR/Cas9-mediated gabrr1 mutations

11 We used a previously described software tool (E-CRISP: http://www.e-crisp.org/E-

12 CRISP/(Heigwer, Kerr et al. 2014)) to design guide RNAs that target the gabrr1 locus.

13 We chose the target sequence (ggatgaaggagcgcttggag), since it targeted a highly

14 conserved stretch of residues within the gabrr1 coding sequence (Wang, Hackam et al.

15 1995). Briefly, we cloned our gabrr1 target sequence into pT7-gRNA (primer:

16 aattaatacgactcactataggatgaaggagcgcttggaggttttagagctagaaatagc). We used the

17 mMessage mMachine T7 RNA synthesis kit to synthesize non-capped guide RNA

18 targeting gabrr1. We used the mMessage mMachine SP6 RNA synthesis kit to generate

19 capped Cas9 mRNA from pCS2-nCas9n (Jao, Wente et al. 2013). We injected one-to-

20 two cell stage embryos with gabrr1 guide RNA, Cas9 mRNA, and phenol red. Injected

21 embryos were reared to adulthood, and sperm samples were screened for Cas9-induced

22 germline mutations. By amplifying a 500 bp amplicon flanking the gabrr1 target locus, bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 8

1 (tggacgggattaaactgagc; aaaatgcaagacccggagat), digesting with T7 endonuclease I and

2 Hpy166II, then analyzing with gel electrophoresis. T7-digested F0 founders with putative

3 gabrr1 lesions were outcrossed to mlpha and reared to adulthood. F1 individuals were

4 genotyped to identify mutation, outcrossed to Tg(fTyrp1:GFP)j900, and analyzed for PTU

5 melanocyte differentiation.

6 Generation of hspl:gabrr1 transgenic line

7 We generated a stable transgenic line to conditionally overexpress gabrr1 under the

8 heat shock promoter Tg(hspl:gabrr1). We used PCR-based methods to clone the full-

9 length gabrr1 cDNA obtained from (zv9 assembly; primers:

10 ggcgatcgcttaattaatgttgagggaaagacagctcca; cctgcaggttaattaatcactgtgagtagatggaccagt)

11 into the Pac1 site in pT2-hsp70l. We used the mMessage mMachine SP6 RNA

12 synthesis kit to synthesize capped transposase mRNA from pCS-TP (Kawakami,

13 Takeda et al. 2004). To create a germline integrated hsp:gabrr1, we injected embryos at

14 the one or two cell stage with plasmid, tranposase, and phenol red. Injected F0 animals

15 were screened at 2-3 dpf for the clonal marker Ef1:GFP, indicating genomic integration

16 of the construct. GFP+ embryos were reared to adulthood, outcrossed to mlpha, and the

17 resulting progeny were screened for germline transmission of the xenopus EF1a:GFP

18 clonal marker (Johnson and Krieg 1995). We established one stable hspl:gabrr1

19 transgenic line.

20 Heat Shock Induction bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 9

1 Adult zebrafish containing hspl:gabrr1 were outcrossed to mlpha strains to generate

2 clutches of hspl:gabrr1; mlpha. From 1-3 dpf, embryos were then treated with the

3 melanocyte pro-drug 4-hydroxyanisole (4-HA: 10 mg/ml in DMSO) to ablate melanocytes.

4 At 3 dpf, 4-HA was washed away, embryos were placed in 50 ml conical tubes, and heat

5 shocked at 37 °C in a water bath for 30 minutes. The heat shock treatment was repeated

6 every 24 hours at 3, 4, and 5 dpf. At 6 dpf, the experiment was terminated and larvae

7 were fixed in 3.7% formaldehyde for melanocyte quantification.

8 Results:

10 GABA-A antagonists increase melanocyte production in larval zebrafish

12 We sought to explore the molecular regulation of melanocyte stem cell quiescence by

13 searching for drugs that result in excess melanocyte development in the larval zebrafish.

14 Previously, our lab found that the larval pigment pattern develops from direct-developing

15 melanocytes and is largely complete by 3 dpf, but that melanocytes that develop after 3

16 dpf (Hultman and Johnson 2010) or those that regenerate the pigment pattern following

17 melanocyte ablation (Hultman, Budi et al. 2009) develop from MSCs. We took advantage

18 of this finding and designed a small molecule screen to identify compounds that increase

19 melanocyte output after 3 dpf. Our screen used larvae expressing the melanocyte marker

20 fTyrp1:GFPj900, incubated them in solution containing the screened compound and the

21 melanin inhibiting drug PTU, and quantified newly generated melanocytes, which can be

22 uniquely identified based on the lack of melanin (mel-) and expression of GFP (GFP+),

23 the mel-, GFP+ melanocytes. We focused on the dorsal larval stripe because we bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 10

1 previously found that less than two new melanocytes develop within this region between

2 3 dpf and 6 dpf (Hultman and Johnson 2010). This infrequent development of new

3 melanocytes provided a low background that allowed us to screen for compounds that

4 induced even a small increase in melanocyte production.

6 We screened over 500 compounds from the Pfizer repurposing panel and identified a

7 GABA-A receptor antagonist (CP-615003-27) that increased melanocyte production

8 between 3 and 6 dpf. Consistent with previous findings from our lab, zebrafish treated

9 with a vehicle control developed on average 1.05 mel-, GFP+ newly-formed melanocytes

10 in the dorsal larval stripe (Fig. 1B, 1F). Larvae treated with the GABA-A antagonist CP-

11 615003-27 developed on average 4.0 newly formed mel-, GFP+ melanocytes in the same

12 region (Fig. 1C, 1F), a significant increase over vehicle control treated fish (Fig. 1F). To

13 confirm the effect of GABA-A inhibition on melanocyte production and development, we

14 tested two other GABA-A antagonists. Zebrafish treated with the GABA-A antagonist

15 picrotoxin developed on average 3.7 mel-, GFP+ melanocytes (Fig. 1D, 1F), and zebrafish

16 treated with the GABA-A rho antagonist TPMPA developed on average 4.1 mel-, GFP+

17 melanocytes (Fig. 1E, 1F). Our finding that inhibition of GABA-A receptors through distinct

18 GABA-A antagonists increases melanocyte production suggested that GABA-A signaling

19 regulates MSC quiescence in larval zebrafish.

21 GABA-A antagonist induced melanocytes derive from MSCs

1 We next asked whether the newly formed melanocytes that develop following GABA-A

2 antagonist treatment arise from a melanocyte stem cell or precursor lineage. Our previous

3 work supported a model that melanocytes within the dorsal stripe primarily develop from

4 undifferentiated MSCs or melanocyte precursors after 3 dpf, suggesting that GABA-A

5 antagonists induce MSCs to produce new melanocytes. However, it remained formally

6 possible that pharmacological inhibition of GABA-A signaling could activate aberrant

7 melanocyte development from a non-stem cell source through an unknown mechanism.

8 To distinguish between these models, we treated zebrafish embryos with either DMSO or

9 the erbb3 inhibitor AG1478 from 8-48 h.p.f., washed out the drug, and then treated the

10 larvae with solution containing PTU and a GABA-A antagonist from 3-6 dpf (Figure 2A).

11 AG1478-mediated inhibition of erbb3 activity has been previously shown to inhibit

12 melanocyte regeneration and metamorphic melanocyte development in zebrafish. Early

13 treatment with this small molecule is thought to block establishment of MSCs, removing

14 the developmental source of new melanocytes. Therefore, if new melanocytes arise from

15 MSCs, we predicted that prior AG1478 treatment would inhibit the ability of GABA-A

16 antagonists to induce melanocyte production.

18 For this analysis, we focused on two representative GABA-A antagonists: TPMPA and

19 Picrotoxin. After each drug treatment, individual larvae were scored for mel-, GFP+

20 melanocytes, which we interpret as newly developed melanocytes in the presence of

21 PTU. Zebrafish larvae treated with DMSO and vehicle control developed 1.76 mel-, GFP+

22 melanocytes, while larvae treated with AG1478 and vehicle control developed 0.32 mel-,

23 GFP+ melanocytes (Figure 2B). This result suggested that the AG1478 treatment bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 12

1 effectively blocked late (3-6 dpf) melanocyte production. Thus, our PTU assay could

2 detect relatively small changes in melanocyte production, which allowed us to confidently

3 test the combinatorial effects of AG1478 and GABA-A antagonists on melanocyte

4 production. Larvae treated with DMSO and the GABA-A antagonist TPMPA developed

5 4.1 mel-, GFP+ melanocytes, but larvae treated with AG1478 and TPMPA developed only

6 0.28 mel-, GFP+ melanocytes. Similarly, larvae treated with DMSO and Picrotoxin

7 developed 5.2 mel-, GFP+ melanocytes, but larvae treated with AG1478 and Picrotoxin

8 developed only 0.17 mel-, GFP+ melanocytes (Figure 2B). We conclude that GABA-A

9 antagonists induce melanocyte production from erbb3-dependent undifferentiated

10 melanocyte precursors.

12 Pharmacological activation of GABA-A signaling inhibits melanocyte regeneration

14 Our data support the model that inhibition of GABA-A receptor signaling increases

15 melanocyte production from undifferentiated precursors. This provided a clear prediction

16 that activation of GABA-A signaling would inhibit melanocyte production. To test this

17 model, we chose to treat larvae homozygous for the temperature sensitive mitfavc7 allele

18 with drugs that activate GABA-A receptor signaling. When raised at restrictive

19 temperature (32°C), the temperature sensitive nature of the mitfavc7 allele prevents

20 melanoblast survival, and mitfavc7 larvae develop no melanocytes. When shifted to

21 permissive temperature (25°C), mitfa function is restored and mitfavc7 larvae exhibit a near

22 complete regeneration of the larval pigment pattern (Johnson, Nguyen et al. 2011). The bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 13

1 homozygous mitfavc7 allele then provided us with temporal control of melanocyte

2 development and regeneration to the test the effects of GABA-A receptor activation.

4 To determine if GABA-A agonists inhibit melanocyte production, we reared mitfavc7 larvae

5 to 4 dpf at 32°C, and then down-shifted to 25°C in the presence of a GABA-A receptor

6 drug. Each larvae was scored for the number of dorsal stripe melanocytes present at 7

7 dpf as a measure of melanocyte regeneration following downshift. Mitfavc7 larvae treated

8 with vehicle control regenerated 42.2 dorsal melanocytes. However, mitfavc7 larvae

9 treated with the endogenous ligand GABA or the GABA-A rho agonist GABOB

10 regenerated only 21.2 and 26.8 dorsal melanocytes, respectively (Figure 3B-C; Figure

11 3G). The reduction of melanocyte regeneration following treatment of GABA-A agonists

12 suggested that direct activation of GABA-A signaling partially inhibited melanocyte

13 regeneration. To further challenge this model, we treated mitfavc7 larvae with drugs that

14 indirectly activated GABA-A receptor signaling and challenged for melanocyte

15 regeneration. Larvae treated with the GABA-A partial agonists L-838,417 (Figure 3D;

16 Figure 3G) and MK 0343 (Figure 3F; Figure 3G) regenerated 16 dorsal melanocytes and

17 18.1 dorsal melanocytes respectively. Similarly, larvae treated with the GABA reuptake

18 inhibitor CI-966, which increases synaptic concentrations of GABA (Ebert and Krnjevic

19 1990), regenerated only 20.4 dorsal melanocytes (Figure 3E; Figure 3G). The effects of

20 these GABA-A activating drugs were not restricted to the dorsal stripe, and appeared to

21 reduce pigmentation across the ventral and lateral regions of the larvae as well (S. Figure

22 1). Our data suggest that pharmacological activation of GABA-A receptor signaling

23 inhibits melanocyte production from MSCs. bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 14

2 GABA rho 1 is necessary for restriction of melanocyte production in larval

3 zebrafish

5 To validate our pharmacology results, we sought to genetically remove GABA-A signaling

6 and assess the impact on melanocyte production. Here, we focused on the GABA-A rho

7 receptor subtype, as the GABA-A rho subtype-specific drug TPMPA yielded a robust

8 increase in melanocyte production. Fortunately, GABA-A rho receptors are homo-

9 pentameric (Martinez-Delgado, Estrada-Mondragon et al. 2010), allowing us to target a

10 single gene to disrupt receptor function. To target GABA-A rho receptor function, we used

11 a CRISPR-based strategy to target the gaba rho 1 (gabrr1) gene (see methods). We

12 specifically targeted a region in the ligand-binding domain that is critical for zinc inhibition

13 to increase the likelihood of disrupting endogenous protein function (Wang, Hackam et

14 al. 1995). Using a PCR- and restriction enzyme-based method, we identified two putative

15 gabrr1 alleles with altered DNA sequence at the targeted site. Sequence analysis of both

16 alleles revealed two gabrr1 in-frame mutations, both of which delete the conserved

17 residues VHS from position 146-148 of the polypeptide sequence (Figure 4A), with one

18 allele gabrr1j247, also substituting the lysine at position 149 to glutamic acid.

20 To determine if genetic reduction of gabrr1 function altered melanocyte development,

j900 21 we outcrossed carriers of each gabrr1 mutation to fTyrp1:GFP , treated the F1 progeny

22 with PTU from 3-6 dpf, and quantified newly generated dorsal melanocytes. Both alleles

23 demonstrated a robust dominant excess melanocyte phenotype. bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 15

1 Zebrafish heterozygous for the gabrr1j247 allele developed 9.56 mel-, GFP+ dorsal

2 melanocytes (Figure 4B; 4D), a five-fold increase over wild-type siblings (Figure 4B; 4C),

3 and zebrafish heterozygous for the gabrr1j248 allele developed 7.81 mel-, GFP+

4 melanocytes (Figure 4B; 4E). In the PTU assay, the dominant gabrr1 phenotypes were

5 mostly restricted to the dorsal stripe, as we observed little differences in ventral

6 pigmentation (S. Figure 2). In support of both mutations being dominant negative alleles

7 of gabrr1, the excess melanocyte phenotype of larval zebrafish homozygous or trans-

8 heterozygous for the alleles was essentially identical to the heterozygous phenotype of

9 each allele (Figure 4B). We infer that gabrr1 is necessary to inhibit excessive melanocyte

10 production in the larval zebrafish, suggesting that GABA signaling through gabrr1 is a key

11 regulatory pathway that normally maintains melanocyte stem cell quiescence in larval

12 zebrafish.

15 GABA rho 1 is sufficient to inhibit melanocyte regeneration in larval zebrafish

17 Our observation that gabrr1 function is necessary to inhibit melanocyte production led us

18 to test whether over-expression of gabrr1 was sufficient to inhibit melanocyte

19 regeneration. To address this question, we cloned the gabrr1 cDNA under control of the

20 heat shock promoter Hsp70l within the Tol2 germ-line transformation vector (Suster,

21 Kikuta et al. 2009) and obtained a stable transgenic line: Tg(Hsp70l:gabrr1) (see

22 methods). We then treated Tg(Hsp70l:gabrr1) and control larvae with the drug 4-HA from

23 1-3 dpf to ablate melanocytes, washed the drug out, induced heat shock at 37°C, and bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 16

1 then quantified melanocyte regeneration at 6 dpf (Figure 5A). Heat shocked wild-type and

2 Non-heat shocked larvae regenerated 49.9 and 51.1 melanocytes respectively, whereas

3 heat shocked Tg(Hsp70l:gabrr1) larvae regenerated 32.3 melanocytes, a roughly 40%

4 reduction in melanocyte production (Figure 5B; 5C). Thus, overexpression of gabrr1 can

5 repress production of melanocytes during periods of regeneration. The over-expression

6 of gabrr1 also appeared to inhibit melanocyte production both ventrally and laterally, but

7 this effect was not as obvious as the effect of the dorsal stripe (S. Fig 3). Expression of

8 gabrr1 then appears sufficient to inhibit melanocyte regeneration in larval zebrafish.

10 Kit signaling and gabrr1

12 We next determined whether kita function was required for GABAergic maintenance of

13 melanocyte stem cell quiescence. Zebrafish heterozygous for the kitab5 null allele

14 regenerate only about 50% of the larval pigment pattern (O'Reilly-Pol and Johnson 2013),

15 suggesting a reduction of the melanocyte stem cell pool consistent with the effects of kit

16 haploinsufficiency observed in mammals (Geissler, Ryan et al. 1988). To test for possible

17 interactions between kit and gabrr1, we asked whether kita haploinsufficiency inhibited

18 the melanocyte over-production phenotype observed in gabrr1 mutants. We generated

19 control and kitab5/+; gabrr1j247/+ double-heterozygous larvae, reared them to 3 dpf, treated

20 them with PTU, and then scored for excess melanocyte production at 6 dpf. As previously

21 observed, wild-type larvae developed 2 excess melanocytes between 3-6 dpf, kitab5/+

22 larvae developed 1.5 excess melanocytes, and gabrr1j247/+ developed 9 excess

23 melanocytes (Figure 6). Of note, kitab5/+; gabrr1j247/+ larvae developed 1.5 excess bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 17

1 melanocytes, indicating that the gabrr1 melanocyte over-production phenotype depends

2 entirely on full kita function, suggesting that GABA-A mediated MSC quiescence is

3 restricted within kita-dependent melanocyte lineages.

5 Discussion:

7 Our work provides evidence that GABAergic signaling promotes MSC quiescence in

8 larval zebrafish. Both pharmacological and genetic studies indicate that reduction of

9 GABA-A rho signaling increases melanocyte production, whereas over-expression of

10 GABA-A rho signaling inhibits melanocyte production. Although both classical and recent

11 studies implicate membrane potential in pigmentation and stem cell proliferation, to our

12 knowledge, our study is the first to uncover such a role for GABA-A receptor signaling in

13 vertebrate pigment biology. Below, we propose a model for how GABA-A signaling

14 regulates melanocyte development, discuss the nature of the GABA-A rho mutants, and

15 place our work in the context of old and new studies that highlight the importance of

16 membrane potential and cell excitability in regulating stem cell proliferation and

17 pigmentation.

18 Our work indicates that GABAergic signaling, directly or indirectly, maintains the MSC in

19 a quiescent state. Prior studies suggest that differentiated larval melanocytes inhibit

20 melanocyte differentiation by suppressing MSC proliferation. For example, chemical or

21 laser ablation of differentiated larval melanocytes induces melanocyte regeneration by

22 promoting the cell division of stem-cell like melanocyte progenitors. In this context, our

23 work provides a conceptual, albeit speculative, model for how the presence of bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 18

1 differentiated melanocytes promotes MSC quiescence and how their absence triggers

2 MSC proliferation and melanocyte production. Our pharmacological and genetic data

3 support a model wherein melanocytes release the neurotransmitter GABA, which

4 activates gabrr1 receptors on the melanocyte stem cell, maintaining the MSC in a

5 quiescent state. Conversely, loss of melanocytes would trigger a reduction in GABA

6 concentration and relieve gabrr1-mediated quiescence, triggering MSC proliferation and

7 melanocyte production. Alternatively, GABAergic signaling could act indirectly to regulate

8 MSC quiescence. For example, the melanocyte could release a non-GABA signal, which

9 triggers a GABA to GABA receptor relay in adjacent cells and tissues that ultimately

10 promotes MSC quiescence. Clearly, additional work is required to address whether

11 GABAergic signaling directly or indirectly controls MSC quiescence, but the presence of

12 GABA synthesis enzymes, such as GAD67 mRNA (Ito, Tanaka et al. 2007), in human

13 melanocytes hints that GABA signaling may be an evolutionarily conserved mechanism

14 that regulates vertebrate pigmentation.

16 Dominant-negative nature of gabrr1 alleles

18 Both of our gabrr1 alleles exhibit essentially identical melanocyte over-production

19 phenotypes when in the heterozygous or homozygous state. The apparent dominant-

20 negative nature of these alleles was initially surprising. But, both alleles remove a highly

21 conserved triplet of amino acids in the ligand-binding domain of the receptor (Wang,

22 Hackam et al. 1995). Thus, each allele likely produces a non-functional subunit. As

23 GABA-rho 1 receptors function as homo-pentamers, assuming the mutant form of the bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 19

1 protein is expressed at roughly wild-type levels, 97% of all gabrr1 channels would be

2 expected to contain at least one mutant subunit, providing a rational explanation for the

3 dominant nature of the gabrr1 mutant alleles.

5 Do multiple extrinsic pathways regulate MSC quiescence?

7 When challenged for regeneration, zebrafish larvae produces hundreds of new

8 melanocytes to repopulate the pigment pattern. These new melanocytes derive from a

9 pool of established precursors, MSCs. Though usually quiescent, MSCs are capable of

10 producing hundreds of new melanocytes throughout development. Using clonal analysis,

11 we previously estimated that the developing zebrafish establishes between 150-200

12 MSCs before 2 dpf (Tryon, Higdon et al. 2011). The MSC pool, though quiescent, then

13 maintains abundance in number and regenerative capability. Complete abrogation of the

14 mechanisms that maintain MSC quiescence would then be expected to generate excess

15 melanocytes proportional to the regenerative capabilities of all MSCs, i.e. generate

16 hundreds of excess melanocytes. Pharmacological or genetic inhibition of GABA-A

17 signaling, however, yields only 8-10 excess melanocytes (Fig. 1; Fig. 4). Thus, the full

18 regenerative capability of larval zebrafish likely involves the concerted actions of multiple

19 pathways that converge on activation of MSC proliferation.

21 Our genetic studies on kita and gabrr1 suggest GABA-A signaling may regulate a kita-

22 dependent pool of MSCs (Tu and Johnson 2010). For example, our prior work indicated

23 that haploinsufficiency for kita reduces the available MSC pool by about 50% (O'Reilly- bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 20

1 Pol and Johnson 2013). But, haploinsufficiency for kita completely suppressed the over-

2 production of melanocytes caused by reduced gabrr1 function. Our findings suggest that

3 kit signaling is required for all GABA sensitive MSCs, but not all MSCs within the larvae

4 are sensitive to either kita or GABA-A signaling. Interestingly, our finding that gabrr1-

5 specific melanocyte production is inhibited by kita haploinsufficiency also supports a

6 phenomena of kit signaling restriction observed in clinical melanomas. Acral (primarily in

7 palms of hands and soles of feet) and mucosal (primarily in oral or nasal cavity and other

8 mucosal surfaces) melanomas are unique melanoma subtypes often associated with

9 oncogenic mutations in c-kit (Curtin, Busam et al. 2006), suggesting that ectopic

10 activation of kit signaling promotes or supports spatially restricted presentations of

11 melanoma. Perhaps the requirement of kit signaling within the gabrr1-driven melanocyte

12 lineage of zebrafish is indicative of regulatory pathways that suppress melanocyte

13 production in a region-specific manner. While the functional relevance of this observation

14 remains unknown, it’s clear that complex and distinct aberrations in the melanocyte

15 lineage may play unique roles in melanocyte development and pathology.

17 Which pathways function in parallel with gabrr1 to suppress MSC proliferation remain

18 unclear. Recent work completed during the course of our study suggest the endothelin

19 receptor Aa (ednraa) acts in parallel to gabrr1 to maintain MSC quiescence. For example,

20 loss of ednraa function leads to ectopic melanocyte production (via a MSC intermediate)

21 specifically within the ventral truck of larval zebrafish (Camargo-Sosa, Colanesi et al.

22 2019), whereas we find that genetic reduction of gabrr1 function increased MSC-derived

23 melanocyte production in the dorsal stripe. Intriguingly, pharmacological (S. Fig 1) or bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 21

1 genetic (S. Fig 3) activation of gabrr1 signaling appeared to reduce pigmentation in a

2 larval-wide manner, rather than a region specific effect one. Regardless of the underlying

3 reasons behind the region-specific and larval-wide phenotypes between reduction and

4 activation of gabrr1 signaling, these studies support the idea that distinct genetic

5 pathways maintain MSC quiescence in a region-specific manner throughout zebrafish

6 development. Clearly, additional work is needed to determine whether other pathways

7 act with gabbr1 and ednraa to promote MSC quiescence, but our work hints that GABA-

8 A mediated quiescence may be a hallmark of vertebrate pigment biology. For example,

9 human melanocytes express enzymes such as aldehyde dehydrogenases (aldh1a1 and

10 aldh9a1) that can synthesize GABA (Ganesan, Ho et al. 2008). In addition, gabrr1

11 expression is reduced in zebrafish melanoma compared to melanocytes (Venkatesan,

12 Vyas et al. 2018), suggesting that down-regulation of GABA-A signaling could support

13 adult melanoma growth, consistent with our findings in larval melanocyte biology.

14 Whether GABA-A mediated MSC quiescence is a hallmark of vertebrate pigment biology,

15 was abandoned at some point in vertebrate evolution, or is a zebrafish restricted

16 phenomenon is of interest moving forward.

18 Bioelectric regulation of MSC quiescence and proliferation

20 Recent and more classical studies reveal that altering the bioelectric properties, or

21 membrane potential, of progenitor or stem cells may be a fundamental, but poorly

22 understood, developmental mechanism that regulates stem cell activity. For example,

23 pharmacological depolarization of glycine gated chloride channels induces severe bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 22

1 hyperpigmentation in xenopus larvae via melanocyte over-proliferation and over-

2 production (Blackiston, Adams et al. 2011). In addition, application of GABA and GABA-

3 A agonists, which hyperpolarizes cells by promoting Cl- influx, has been shown to inhibit

4 proliferation of embryonic stem cells (Teng, Tang et al. 2013) and peripheral neural crest

5 stem cells (Young and Bordey 2009) in mice. Our work complements these studies, as

6 we found that inhibition of gabrr1 promotes melanocyte production, likely through an MSC

7 intermediate. Thus, altering the membrane potential of Cl- appears to play a key role in

8 regulating vertebrate pigmentation and stem cell proliferation.

10 Recent work in mice and humans underlines the importance of membrane potential and

11 the bioelectric properties of cells in regulating stem cell activity and development. For

12 example, in the mouse neocortex, neural progenitors become increasingly hyperpolarized

13 as they produce their characteristics cell lineages (Vitali, Fievre et al. 2018). Of note,

14 artificially hyperpolarizing neural progenitor cells induced the premature production of

15 late-stage cell types, revealing a functional link between changes in membrane potential

16 and temporal birth order of cells in the neocortex. In humans, developmental defects in

17 cerebral cortex folding and neuronal migration strongly associate with mutations in

18 Nav1.3, a voltage-gated sodium channel primarily expressed at developmental stages

19 composed of early non-transmitting neurons (Smith, Kenny et al. 2018). This suggests

20 that the cellular basis for these neurological defects arises not from altered neuronal

21 transmission, but rather from alterations in the development and migration of neural

22 progenitors and neurons. Together, these studies and our work provide strong evidence

23 that changes in the membrane potential of stem and progenitor cells alters cell division bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 23

1 patterns and cellular behavior, highlighting the emerging theme of bioelectric regulation

2 of stem and progenitor cells during development. Future work that can systematically

3 assess the effect of membrane potential on the development of stem cells and their

4 progeny in a broad range of tissues is required to reveal the extent to which this

5 phenomenon occurs in vertebrate biology.

7 Acknowledgements:

8 We thank Brian Stephens and Sinan Li for fish husbandry during the majority of the study.

9 We are especially grateful to the Washington University School of Medicine Genetics

10 Department and Washington University Zebrafish Facility for providing critical support in

11 completing this research. We thank Rob Tryon and Ryan McAdow for assistance and

12 guidance generating mutant and transgenic lines. We thank Michael Nonet, Cristina

13 Strong, Charles Kaufman, and Douglas Chalker for critical comments on the manuscript.

14 James R. Allen was a Howard Hughes Medical Institute Gilliam Fellow during the course

15 of this study. This work was funded by NIH RO1-GM056988 to S.L.J. J.B.S was supported

16 by NIH RO1-NS036570.

18 Competing Interests:

19 The authors declare no competing financial interests.

21 Figure Legends:

1 Figure 1: GABA-A antagonists increase melanocyte production in larval zebrafish.

2 (A) Cartoon of experimental timeline for PTU melanocyte differentiation assay. Drugs and

3 PTU are added to zebrafish embryos between 3 – 6 dpf. (B-D) Images of representative

4 6 dpf larvae treated with vehicle control (B) or GABA-A antagonists CP-615003 (40 µM;

5 C), Picrotoxin (100 µM; D), and TPMPA (100 µM; E). (F) Quantification of the average

6 number of melanin-, GFP+ dorsal melanocytes for each treatment group. (Vehicle control:

7 0.92±1.15, S.E.M: 0.08, N=84; CP-615003: 4.27±2.12, S.E.M: 0.23, N=81; Picrotoxin:

8 3.76±1.38, S.E.M: 0.18, N=55; TPMPA: 4.10±2.14, S.E.M: 0.29, N=52). Error bars

9 represent standard error of mean, each experimental group compared to vehicle control

10 had P-values <0.001 (Students t-test).

12 Figure 2: GABA-A antagonist induced melanocytes derive from MSCs. (A) Cartoon

13 of experimental timeline for drug treatment. (B) Quantification of average melanin-, GFP+

14 dorsal melanocytes in each group. (Vehicle control: 1.76±1.26, S.E.M: 0.2, N=39; Vehicle

15 control + AG1478: 0.32±0.48, S.E.M: 0.10, N=28; TPMPA: 4.14±1.43, S.E.M: 0.27, N=29;

16 TPMPA + AG1478: 0.28±0.54, S.E.M: 0.11, N=25; Picrotoxin: 5.2±0.99, S.E.M: 0.18,

17 N=30; Picrotoxin + AG1478: 0.17±0.38, S.E.M: 0.06, N=30). Error bars represent

18 standard error of mean, each experimental group compared to vehicle control had P-

19 values <0.001 (Students t-test).

21 Figure 3: Pharmacological activation of GABA-A signaling inhibits melanocyte

22 regeneration. Images of representative mitfavc7 7 dpf larvae treated with vehicle control

23 (A), GABA (50 mM; B), GABOB (100 µM; C), L, 838-417 (100 µM; D), CI-966 HCL (20 bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 25

1 µM; E), and MK 0343 (100 µM; F). (G) Quantification of the average number of dorsal

2 melanocytes in each drug treatment group. (Vehicle control: 42.2±9.38, S.E.M: 1.06,

3 N=78; GABA: 21.2±10.4, S.E.M: 1.61, N=42; GABOB: 26.8±7.71, S.E.M: 1.20, N=41;

4 L,838-417: 16±11.2, S.E.M: 1.73, N=42, CI-966 HCL: 20.4±7.81, S.E.M: 1.19, N=43; MK

5 0343: 18.1±9.55, S.E.M: 1.61, N=35). Error bars represent standard error of mean, each

6 experimental group compared to vehicle control had P-values <0.001 (Students t-test).

8 Figure 4: gabrr1 mutations exhibit a dominant excess melanocyte phenotype during

9 larval stages. (A) Partial alignment of vertebrate gabrr1 protein sequence homology

10 within the ligand binding domain, with the amino-acid sequence of the two gabrr1 alleles

11 generated in the study. (B) Quantification of the average number of melanin-, GFP+ dorsal

12 melanocytes in each treatment group. (Wild Type: 2.04±0.94, S.E.M: 0.13, N=51;

13 gabrr1j247/+: 9.56±1.48, S.E.M: 0.23, N=39; gabrr1j247/j247: 9.13±1.41, S.E.M: 0.36, N=15

14 gabrr1j248/+: 7.81±1.17, S.E.M: 0.17, N=48; gabrr1j248/j248: 10.2±2.54, S.E.M: 0.71, N=13;

15 gabrr1j247/j248: 9.72±1.99, S.E.M: 0.37, N=29). Representative images of 6 dpf wild-type

16 (C), gabrr1j247/+ (D), gabrr1j248/+ (E) and gabrr1j247/j248 (F) larvae. Error bars represent

17 standard error of mean, each experimental group compared to wild type had P-values

18 <0.001 (Students t-test).

20 Figure 5: Over-expression of gabrr1 inhibits melanocyte regeneration in larval

21 zebrafish. (A) Cartoon of experimental timeline. Arrows indicate timing of three 30 minute

22 37 °C heat shock treatments. (B) Quantification of average dorsal melanocytes in each

23 treatment group. (mlpha + heat shock: 49.9±6.01, S.E.M: 1.06, N=32; Tg(Hsp70l:gabrr1): bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 26

1 51.1±6.86, S.E.M: 1.11, N=38; Tg(Hsp70l:gabrr1)+ heatshock: 32.3±5.54, S.E.M: 0.81,

2 N=46). Images of representative mlpha + heat shock (C), Tg(Hsp70l:gabrr1) (D), and

3 Tg(Hsp70l:gabrr1) + heatshock (E) larvae. Error bars represent standard error of mean,

4 Tg(Hsp70l:gabrr1) + heat shock compared to mlpha + heat shock had P-values <0.001

5 (Students t-test).

7 Figure 6: gabrr1 mediated maintenance of MSC quiescence is sensitive to kit

8 dosage. Quantification of the average number of melanin-, GFP+ dorsal melanocytes in

9 gabrr1j247/+ and kitb5 heterozygotes. (Wild-type: 2.35±1.03, S.E.M: 0.16, N=42; kitb5/+:

10 0.88±0.81, S.E.M: 0.11, N=56; Gabrr1j247/+: 9.55±1.43, S.E.M: 0.23, N=40; kitb5/+;

11 gabrr1j247/+: 0.87±0.87, S.E.M: 0.11, N=60). Error bars represent standard error of mean,

12 Each group compared to vehicle control had P-values <0.001 (Students t-test).

14 Table S1: List of GABA-A receptor pharmacology used in study. Each drug was

15 obtained from the indicated manufacturer and handled according to vendor guidelines.

17 Supplemental Figure 1: Pharmacological activation of GABA-A reduces larval

18 pigmentation across the body. (A-D) Images of representative mitfavc7 7 dpf larvae

19 treated with vehicle control (A), L,838-417 (B), CI-966 HCL (C), and MK 0343 (D).

21 Supplemental Figure 2: gabrr1 mutant phenotypes have no visible effect on ventral

22 pigmentation. (A-D) Images of representative 6 dpf wild-type (A), gabrr1j247/+ (B),

23 gabrr1j248/+ (C), and gabrr1j247/j248(D) larvae. bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 27

2 Supplemental Figure 3: Over-expression of gabrr1 partially reduces ventral

3 pigmentation. (A-C) Images of representative 6 dpf heat shocked mlpha (A),

4 Tg(Hsp70l:gabrr1) (B), and Tg(Hsp70l:gabrr1) + heat shock (C).

8 Bertrand, J. Y., A. D. Kim, E. P. Violette, D. L. Stachura, J. L. Cisson and D. Traver (2007).

9 "Definitive hematopoiesis initiates through a committed erythromyeloid progenitor in the

10 zebrafish embryo." Development 134(23): 4147-4156.

11 Blackiston, D., D. S. Adams, J. M. Lemire, M. Lobikin and M. Levin (2011).

12 "Transmembrane potential of GlyCl-expressing instructor cells induces a neoplastic-like

13 conversion of melanocytes via a serotonergic pathway." Dis Model Mech 4(1): 67-85.

14 Bormann, J. (2000). "The 'ABC' of GABA receptors." Trends Pharmacol Sci 21(1): 16-19.

15 Budi, E. H., L. B. Patterson and D. M. Parichy (2008). "Embryonic requirements for ErbB

16 signaling in neural crest development and adult pigment pattern formation." Development

17 135(15): 2603-2614.

18 Camargo-Sosa, K., S. Colanesi, J. Muller, S. Schulte-Merker, D. Stemple, E. E. Patton

19 and R. N. Kelsh (2019). "Endothelin receptor Aa regulates proliferation and differentiation

20 of Erb-dependent pigment progenitors in zebrafish." PLoS Genet 15(2): e1007941.

21 Cheung, T. H. and T. A. Rando (2013). "Molecular regulation of stem cell quiescence."

22 Nat Rev Mol Cell Biol 14(6): 329-340. bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 28

1 Curtin, J. A., K. Busam, D. Pinkel and B. C. Bastian (2006). "Somatic activation of KIT in

2 distinct subtypes of melanoma." J Clin Oncol 24(26): 4340-4346.

3 Ebert, U. and K. Krnjevic (1990). "Systemic CI-966, a new gamma-aminobutyric acid

4 uptake blocker, enhances gamma-aminobutyric acid action in CA1 pyramidal layer in

5 situ." Can J Physiol Pharmacol 68(9): 1194-1199.

6 Ganesan, A. K., H. Ho, B. Bodemann, S. Petersen, J. Aruri, S. Koshy, Z. Richardson, L.

7 Q. Le, T. Krasieva, M. G. Roth, P. Farmer and M. A. White (2008). "Genome-wide siRNA-

8 based functional genomics of pigmentation identifies novel genes and pathways that

9 impact melanogenesis in human cells." PLoS Genet 4(12): e1000298.

10 Geissler, E. N., M. A. Ryan and D. E. Housman (1988). "The dominant-white spotting (W)

11 locus of the mouse encodes the c-kit proto-oncogene." Cell 55(1): 185-192.

12 Heigwer, F., G. Kerr and M. Boutros (2014). "E-CRISP: fast CRISPR target site

13 identification." Nat Methods 11(2): 122-123.

14 Hultman, K. A., E. H. Budi, D. C. Teasley, A. Y. Gottlieb, D. M. Parichy and S. L. Johnson

15 (2009). "Defects in ErbB-dependent establishment of adult melanocyte stem cells reveal

16 independent origins for embryonic and regeneration melanocytes." PLoS Genet 5(7):

17 e1000544.

18 Hultman, K. A. and S. L. Johnson (2010). "Differential contribution of direct-developing

19 and stem cell-derived melanocytes to the zebrafish larval pigment pattern." Dev Biol

20 337(2): 425-431.

21 Ito, K., K. Tanaka, Y. Nishibe, J. Hasegawa and H. Ueno (2007). "GABA-synthesizing

22 enzyme, GAD67, from dermal fibroblasts: evidence for a new skin function." Biochim

23 Biophys Acta 1770(2): 291-296. bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 29

1 Jao, L. E., S. R. Wente and W. Chen (2013). "Efficient multiplex biallelic zebrafish genome

2 editing using a CRISPR nuclease system." Proc Natl Acad Sci U S A 110(34): 13904-

3 13909.

4 Johnson, A. D. and P. A. Krieg (1995). "A Xenopus laevis gene encoding EF-1 alpha S,

5 the somatic form of elongation factor 1 alpha: sequence, structure, and identification of

6 regulatory elements required for embryonic transcription." Dev Genet 17(3): 280-290.

7 Johnson, S. L., D. Africa, C. Walker and J. A. Weston (1995). "Genetic control of adult

8 pigment stripe development in zebrafish." Dev Biol 167(1): 27-33.

9 Johnson, S. L., A. N. Nguyen and J. A. Lister (2011). "mitfa is required at multiple stages

10 of melanocyte differentiation but not to establish the melanocyte stem cell." Dev Biol

11 350(2): 405-413.

12 Kawakami, K., H. Takeda, N. Kawakami, M. Kobayashi, N. Matsuda and M. Mishina

13 (2004). "A transposon-mediated gene trap approach identifies developmentally regulated

14 genes in zebrafish." Dev Cell 7(1): 133-144.

15 Levy, C., M. Khaled and D. E. Fisher (2006). "MITF: master regulator of melanocyte

16 development and melanoma oncogene." Trends Mol Med 12(9): 406-414.

17 Li, L. and R. Bhatia (2011). "Stem cell quiescence." Clin Cancer Res 17(15): 4936-4941.

18 Lister, J. A., C. P. Robertson, T. Lepage, S. L. Johnson and D. W. Raible (1999). "nacre

19 encodes a zebrafish microphthalmia-related protein that regulates neural-crest-derived

20 pigment cell fate." Development 126(17): 3757-3767.

21 Ma, D. K., M. A. Bonaguidi, G. L. Ming and H. Song (2009). "Adult neural stem cells in

22 the mammalian central nervous system." Cell Res 19(6): 672-682. bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 30

1 Martinez-Delgado, G., A. Estrada-Mondragon, R. Miledi and A. Martinez-Torres (2010).

2 "An Update on GABArho Receptors." Curr Neuropharmacol 8(4): 422-433.

3 Nishimura, E. K., S. A. Jordan, H. Oshima, H. Yoshida, M. Osawa, M. Moriyama, I. J.

4 Jackson, Y. Barrandon, Y. Miyachi and S. Nishikawa (2002). "Dominant role of the niche

5 in melanocyte stem-cell fate determination." Nature 416(6883): 854-860.

6 O'Reilly-Pol, T. and S. L. Johnson (2013). "Kit signaling is involved in melanocyte stem

7 cell fate decisions in zebrafish embryos." Development 140(5): 996-1002.

8 Parichy, D. M., J. F. Rawls, S. J. Pratt, T. T. Whitfield and S. L. Johnson (1999). "Zebrafish

9 sparse corresponds to an orthologue of c-kit and is required for the morphogenesis of a

10 subpopulation of melanocytes, but is not essential for hematopoiesis or primordial germ

11 cell development." Development 126(15): 3425-3436.

12 Rawls, J. F. and S. L. Johnson (2000). "Zebrafish kit mutation reveals primary and

13 secondary regulation of melanocyte development during fin stripe regeneration."

14 Development 127(17): 3715-3724.

15 Rawls, J. F. and S. L. Johnson (2001). "Requirements for the kit receptor tyrosine kinase

16 during regeneration of zebrafish fin melanocytes." Development 128(11): 1943-1949.

17 Rawls, J. F. and S. L. Johnson (2003). "Temporal and molecular separation of the kit

18 receptor tyrosine kinase's roles in zebrafish melanocyte migration and survival." Dev Biol

19 262(1): 152-161.

20 Sheets, L., D. G. Ransom, E. M. Mellgren, S. L. Johnson and B. J. Schnapp (2007).

21 "Zebrafish melanophilin facilitates melanosome dispersion by regulating dynein." Curr

22 Biol 17(20): 1721-1734. bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 31

1 Sigel, E. and M. E. Steinmann (2012). "Structure, function, and modulation of GABA(A)

2 receptors." J Biol Chem 287(48): 40224-40231.

3 Smith, R. S., C. J. Kenny, V. Ganesh, A. Jang, R. Borges-Monroy, J. N. Partlow, R. S.

4 Hill, T. Shin, A. Y. Chen, R. N. Doan, A. K. Anttonen, J. Ignatius, L. Medne, C. G.

5 Bonnemann, J. L. Hecht, O. Salonen, A. J. Barkovich, A. Poduri, M. Wilke, M. C. Y. de

6 Wit, G. M. S. Mancini, L. Sztriha, K. Im, D. Amrom, E. Andermann, R. Paetau, A. E.

7 Lehesjoki, C. A. Walsh and M. K. Lehtinen (2018). "Sodium Channel SCN3A (NaV1.3)

8 Regulation of Human Cerebral Cortical Folding and Oral Motor Development." Neuron

9 99(5): 905-913.e907.

10 Suster, M. L., H. Kikuta, A. Urasaki, K. Asakawa and K. Kawakami (2009). "Transgenesis

11 in zebrafish with the tol2 transposon system." Methods Mol Biol 561: 41-63.

12 Teng, L., Y. B. Tang, F. Sun, S. M. An, C. Zhang, X. J. Yang, H. Y. Lv, Q. Lu, Y. Y. Cui,

13 J. J. Hu, L. Zhu and H. Z. Chen (2013). "Non-neuronal release of gamma-aminobutyric

14 Acid by embryonic pluripotent stem cells." Stem Cells Dev 22(22): 2944-2953.

15 Tryon, R. C., C. W. Higdon and S. L. Johnson (2011). "Lineage relationship of direct-

16 developing melanocytes and melanocyte stem cells in the zebrafish." PLoS One 6(6):

17 e21010.

18 Tryon, R. C. and S. L. Johnson (2012). "Clonal and lineage analysis of melanocyte stem

19 cells and their progeny in the zebrafish." Methods Mol Biol 916: 181-195.

20 Tu, S. and S. L. Johnson (2010). "Clonal analyses reveal roles of organ founding stem

21 cells, melanocyte stem cells and melanoblasts in establishment, growth and regeneration

22 of the adult zebrafish fin." Development 137(23): 3931-3939. bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. 32

1 Venkatesan, A. M., R. Vyas, A. K. Gramann, K. Dresser, S. Gujja, S. Bhatnagar, S.

2 Chhangawala, C. B. F. Gomes, H. S. Xi, C. G. Lian, Y. Houvras, Y. J. K. Edwards, A.

3 Deng, M. Green and C. J. Ceol (2018). "Ligand-activated BMP signaling inhibits cell

4 differentiation and death to promote melanoma." The Journal of Clinical Investigation

5 128(1): 294-308.

6 Vitali, I., S. Fievre, L. Telley, P. Oberst, S. Bariselli, L. Frangeul, N. Baumann, J. J.

7 McMahon, E. Klingler, R. Bocchi, J. Z. Kiss, C. Bellone, D. L. Silver and D. Jabaudon

8 (2018). "Progenitor Hyperpolarization Regulates the Sequential Generation of Neuronal

9 Subtypes in the Developing Neocortex." Cell 174(5): 1264-1276.e1215.

10 Wang, T. L., A. Hackam, W. B. Guggino and G. R. Cutting (1995). "A single histidine

11 residue is essential for zinc inhibition of GABA rho 1 receptors." J Neurosci 15(11): 7684-

12 7691.

13 Westerfield, M. (2000). The Zebrafish Book. A Guide for The Laboratory Use of Zebrafish

14 (Danio rerio).

15 Yang, C. T. and S. L. Johnson (2006). "Small molecule-induced ablation and subsequent

16 regeneration of larval zebrafish melanocytes." Development 133(18): 3563-3573.

17 Yang, C. T., R. D. Sengelmann and S. L. Johnson (2004). "Larval melanocyte

18 regeneration following laser ablation in zebrafish." J Invest Dermatol 123(5): 924-929.

19 Young, S. Z. and A. Bordey (2009). "GABA's control of stem and cancer cell proliferation

20 in adult neural and peripheral niches." Physiology (Bethesda) 24: 171-185.

21 bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Figure 1: GABA-A antagonists increase melanocyte production in larval zebrafish

Drug panel A. PTU

0 dpf 3 dpf 6 dpf B. C.

Vehicle control CP-615003 D. E.

Picrotoxin TPMPA F. e

a *** v r a l

/ GABA-A antagonists

s e t y c o n a l e m

l a s r o d

+ P F G

- n i n a l e M

Vehicle control CP-615003 Picrotoxin TPMPA B. A.

- + certified bypeerreview)istheauthor/funder,whohasgrantedbioRxivalicensetodisplaypreprintinperpetuity.Itmadeavailableunder

Melanin / GFP dorsal melanocytes / larvae bioRxiv preprint AG1478 Figure2: doi: 0 https://doi.org/10.1101/619056 GABA-A antagonistinduced melanocytes derive from MSCs dpf Vehiclecontrol 3 - µ M AG1478 + ; this versionpostedApril25,2019. a CC-BY 4.0Internationallicense 2 dpf - TPMPA 3 dpf . antagonist The copyrightholderforthispreprint(whichwasnot + GABA-A PTU 6 - Picrotoxin dpf + bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. A. B. B.

Vehicle Control GABA C. D.

GABOB L,838-417 E. F.

CI-966 HCL MK 0343 G. e a v r a l

s e t y c o n a l e m

l a s r o D

Vehicle control GABA GABOB L-838,417 CI-966 HCL MK 0343

Figure 3: Pharmacological activation of GABA-A signaling inhibits melanocyte regeneration bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. A. Human (155-173): PDMFFVHSKRSFIHDTTT Mouse (156-174): PDMFFVHSKRSFIHDTTT Zebrafish (140-158): PDIFFVHSKRSFIHDTTT

Gabrr1j247 mutation: PDIFF---ERSFIHDTTT (K149E) Gabrr1j248 mutation: PDIFF---KRSFIHDTTT B. e a v r a l

s e t y c o n a l e m

l a s r o d

+ P F G

- n i n a l e M

Wild Type Gabrr1 j247/+ Gabrr1 j247/j247 Gabrr1 j248/+ Gabrr1 j248/j248 Gabrr1 j247/j248 C. D.

Wild-type Gabrr1j247/+ E. F.

Gabrr1j248/+ Gabrr1j247/j248

Figure 4: gabrr1 mutants have excess melanocyte phenotype during larval stage Heat shock (37 shock Heat B. A. certified bypeerreview)istheauthor/funder,whohasgrantedbioRxivalicensetodisplaypreprintinperpetuity.Itmadeavailableunder

Dorsal melanocytes / larvae bioRxiv preprint Figure 5: 0 dpf ° C) mlpha doi: 4-HA + Over-expression gabrr1of inhibitsmelanocyte regeneration https://doi.org/10.1101/619056 3 dpf Tg(hspl:

4 dpf - gabrr1 5 dpf ) ; this versionpostedApril25,2019. a CC-BY 4.0Internationallicense Tg(hspl: + gabrr1 ) E. D. C. . The copyrightholderforthispreprint(whichwasnot Tg(hspl Tg(hspl :gabrr1) mlpha: :gabrr1): :shock heat (37 shock heat (37 no shock heat ° ° C) C) certified bypeerreview)istheauthor/funder,whohasgrantedbioRxivalicensetodisplaypreprintinperpetuity.Itmadeavailableunder sensitive kit to dosage Figure6: bioRxiv preprint

Melanin- / GFP+ dorsal melanocytes / larvae gabrr1 mediated maintenance of MSC quiescence is doi: Wild Type Wild https://doi.org/10.1101/619056 ; this versionpostedApril25,2019. Kita a CC-BY 4.0Internationallicense b5/+ gabrr1 . The copyrightholderforthispreprint(whichwasnot j247/+ Kita b5/+ b5/+ ;gabrr1 j247/+ bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

A. B.

Vehicle Control L,838-417 C. D.

CI-966 HCL MK 0343

Supplemental Figure 1: GABA drugs - Lateral views bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

A. B.

Wild Type Gabrr1j247/+ C. D.

Gabrr1j248/+ Gabrr1j247/j248

Supplemental Figure 2: gabrr1 mutants - Ventral views bioRxiv preprint doi: https://doi.org/10.1101/619056; this version posted April 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

MK 0343 A.

mlpha: heat shock (37°C) B.

Tg(hspl:gabrr1): no heat shock C.

Tg(hspl:gabrr1): heat shock (37°C)

Supplemental Figure 3: hsp>gabrr1 - Lateral views