Microphthalmia and Disrupted Retinal Development Due to a LacZ Knock-in/Knock-Out Allele at the Vsx2 Locus

Introduction

Visual System Homeobox (VSX) genes encode paired-like homeodomain transcription factors required for visual system development in both vertebrates and invertebrates. Two paralogous genes, Vsx1 and Vsx2 (formerly Chx10), are present in vertebrates and have distinct roles in eye growth and retinal formation during development. In humans, clinical features of Vsx1 disruption include keratoconus and visual deficits whereas Vsx2 disruption include non-syndromic congenital bilateral microphthalmia and blindness. Vsx2 mutations also cause non-syndromic microphthalmia in other vertebrates. Understanding how Vsx2 functions and is regulated has provided important insights into ocular development.

Vsx2 expression is activated in the nascent retinal domain during the morphological transition from the optic vesicle to optic cup, remains expressed in retinal progenitor cells (RPCs) through development, and becomes restricted to bipolar cells and Müller glia in the adult retina.^1,2 Consistent with onset of expression, reductions in eye size and defects in retinal development commence soon after optic cup formation in murine Vsx2 mutants^1,3 and at an equivalent stage in Vsx2 mutant human organoids.⁴ In mice, retinal defects include ectopic expression of non-retinal genes and differentiation programs more akin to neighboring tissues (lineage infidelity), slowed proliferation, delayed onset of neurogenesis, disrupted laminar morphogenesis, and a lack of bipolar cells.^1,3,5–12

Multiple Vsx2 alleles encoding structural or regulatory variants exist in mice (informatics.jax.org; Quick Search term: Vsx2; Category: Alleles), but most phenotypic characterizations of Vsx2 disruption in the eye have been done with three protein coding variants. The most extensively studied is the ocular retardation J allele (orJ; MGI:1856112), a spontaneous mutation that replaces tyrosine at position 176 with a premature stop codon (Figure 1A).^1,5 Although Vsx2 mRNA remains expressed in the orJ mutant retina, VSX2 protein is not detected, and the orJ allele is considered a functional null.^1,3,10 The other two alleles are gene targeted missense mutations that are homologous to human variants associated with congenital microphthalmia and blindness: a glutamine substitution for arginine at amino acid position 200 (R200Q; MGI:5449355), and a tryptophan substitution of arginine at position 227 (R227W; MGI:5449356; (Figure 1A)).³ Like the orJ allele, both are recessive, and their phenotypes are consistent with disrupted Vsx2 function although they differ in severity. The R200Q mutant closely phenocopies orJ, while the R227W mutant is more severe as revealed by even smaller eyes, increased pigmentation of the retina, and a failure to initiate neurogenesis.³ The R200Q and R227W alleles also differ from orJ in that the mutant proteins are expressed.³ The R200Q mutation is in the homeodomain and abrogates DNA binding to canonical sites.^3,13 The R227W mutation is in the CVC domain, a conserved stretch of approximately 50 amino acids that is a defining feature of the VSX gene family.³ Although its mechanism of action remains elusive, the CVC domain also appears to promote DNA binding and sumoylation.^14,15 Consistent with its role in DNA binding, the R227W protein exhibits weakened binding to canonical VSX2 binding sites.³

Figure 1 Vsx2LacZ recombination confirmation and gross phenotype. (A) Schematic of Vsx2LacZ allele and resultant mRNA following Cre recombination. (B) Genomic DNA PCR shows Mendelian inheritance of Vsx2 wildtype and LacZ sequences confirming germline transmission. (C) Western blot analysis of E15.5 retinal lysates. The N-terminal specific VSX2 antibody was used in the top 2 rows to detect full length and truncated VSX2 proteins. bGal antibody was used in 4th row to detect bGal expression from LacZ allele. VSX2 and bGal antibodies were used on separate blot. The probing for ATP5A1 serves as a loading control for each blot. (D and E) Immunohistology at E15.5 reveals continued yet differentially localized expression of VSX2 in the LacZ mutant (retinas are contained within the dashed lines). (F–I) Recombination reporter, bGal, is co-expressed with VSX2 in the LacZ heterozygote INL at P28. (J–L) Microphthalmia phenotype of the LacZ mutant is observable and significant at E12.5 (****p<0.0001, T-test, sample sizes: 9 (Z/+), 10 (Z/Z)). Abbreviations: ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer. Scale bars: (D–H) 100 µm; (J and K) 2000 µm.

Here, we expand the genetic toolkit for Vsx2 with a new targeted allele that replaces exon 3 with an IRES-LacZ expression cassette (Vsx2^LacZ). We show that bGal protein expression from the LacZ cassette is a reliable reporter of endogenous VSX2 expression, and that the ocular phenotype of the Vsx2^LacZ homozygous mutant (LacZ mutant; Z/Z) exhibits the hallmarks of Vsx2 disruption in a manner similar to the orJ mutant. However, unlike orJ, the predicted truncated protein is expressed and there are subtle differences in the LacZ mutant phenotype – leaving open the possibility that the allele is not a complete knockout. This allele adds to the repertoire of Vsx2 mutants in which protein variants are expressed and provides an additional resource for probing Vsx2 function in vivo.

Results

The VSX2-encoding portion of the Vsx2^LacZ mRNA is solely composed of exons 1 and 2, therefore lacking the homeodomain and CVC domain. This is followed by an internal ribosome entry sequence (IRES) and LacZ cDNA, which encodes the Escherichia coli bGal protein (Figure 1A). The Vsx2^LacZ (LacZ) allele was generated by Cre recombination of the Vsx2^{tm1a(EUCOMM)Wtsi} allele (MGI ID:4453691),¹⁶ and genomic DNA PCR confirmed germline transmission (Figure 1B). Western blot analysis of E15.5 retinal lysates probed with an antibody specific to the N-terminal portion of the human VSX2 (aa1-131; see Table 1 for antibodies and dilutions) revealed that like the orJ allele, full length VSX2 is not expressed from the LacZ allele (Figure 1C; top row Supplemental Figure 1). Unexpectedly, the predicted, truncated VSX2 protein from the LacZ allele was detected, indicating stable expression (Figure 1C; second row). Detection of bGal confirmed that the gene targeting strategy was successful and that the IRES was sufficient to produce two distinct proteins from the same transcript (Figure 1C; fourth row). Immunohistology revealed that the mutant protein lacked nuclear localization, consistent with the retention of the nuclear export signal (NES) in exon 2 and the loss of the nuclear localization signal (NLS) in exon 3 (Figure 1D and E).¹⁷ We also confirmed that bGal accurately reports VSX2 expression, indicated by their extensive colocalization throughout development and easily visualized in the inner nuclear layer (INL) of LacZ heterozygous retinas (LacZ het; +/Z) at P28. (Figure 1F–I and Supplemental Figure 2A–F). Finally, the stereotypical microphthalmic phenotype was evident and significant at E12.5 (Figure 1J–L; see figure legends and Supplemental Table 1 for statistics; see Methods for quantification methods and statistical analyses).

Table 1 Primary Antibodies

We next performed immunohistology to characterize the LacZ mutant phenotype in the retina. Establishment of optic cup regionalization appeared intact at E10.5 as evidenced by VSX2 expression in the retina, LHX2 and PAX6 in the RPE and retina, OTX1/2 and MITF in the RPE (Figure 2A–E and G–K). The RPC and proliferation markers PCNA and CCND1 also appeared to be normal in the E10.5 LacZ mutant retina (Supplemental Figure 3A–D). However, MITF was ectopically expressed in the retina (Figure 2E and K) and PAX2 expression extended into the dorsal retina in the LacZ mutant (Figure 2F and L; arrows). At E11.5, VSX2, LHX2, and PAX6 expression remained unaltered (Figure 2M–O and S–U), but the ectopic expression of MITF (Figure 2Q and W and dorsal expression of Pax2 (Figure 2R and X; arrows) were still evident, and OTX1/2 expression in the central retina was elevated in the LacZ mutant (Figure 2P and V). These altered expression patterns are consistent with compromised retinal identity. Further evidence of altered retinal identity was observed at E12.5 with continued ectopic expression of OTX1/2 in the peripheral retina, most notably on the ventral side (Figure 4E and K; arrows). From E11.5 to E15.5, the RPC population, as indicated by PCNA, is maintained (Figure 3A–F; Supplemental Figure 3E and F), but proliferation progressively decreased, most notably in the peripheral retina, as assessed by reduced CCND1 expression (Figure 3G–L; Supplemental Figure 3G and H; arrows) and EdU incorporation (Figure 3M–R; Supplemental Figure 3I and J; arrows). This early embryonic phenotype emulates those of the orJ, R200Q, and R227W mutants.^1,3

Figure 2 Regionalization of the early embryonic LacZ mutant eye. (A–L) Markers of eye field specification and regionalization at E10.5. All markers exhibit similar expression patterns in the LacZ mutant compared to the LacZ het except for MITF, which is ectopically expressed in the retina (E and K) and PAX2, which is expanded into the dorsal retina (arrows; F, L). (M–X) Markers of eye field specification and regionalization at E11.5. MITF continues to be ectopically expressed (Q and W) and PAX2 extends into the dorsal retina of the LacZ mutants (arrows; R,X). All other markers were similar between the LacZ het and mutant. Abbreviations: R, retina; RPE, retinal pigment epithelium; CM, ciliary margin; L, lens; C, cornea. Scale bar: 100 µm.

Figure 3 Proliferation is reduced in the embryonic LacZ mutant retina. (A–F) The general proliferation marker PCNA remains expressed throughout the LacZ mutant retina, similar to the LacZ het at E11.5, E12.5, and E15.5. In contrast, the patterns of CCND1 expression (G–L) and EdU incorporation (M–R) became progressively restricted to the central retina in the LacZ mutants compared to LacZ hets, most notably at E12.5 and E15.5. Arrows in I-L and O-R point to the regions where staining is absent in the LacZ mutant retina compared to the LacZ het. Scale bar: 100 µm.

Neurogenesis and establishment of the inner and outer neuroblastic layers initiate at E11.5 in the central retina and progress out to the retinal peripheral margin by E15.5. This is readily observed in the LacZ het retina at E12.5, E13.5, and E15.5 using the expression of neurogenesis and cell-type specific markers including TUBB3, a marker of cytoskeletal maturation in early post-mitotic neuronal precursors, ATOH7, a marker expressed by neurogenic progenitors preceding cell cycle exit, BHLHB5 and POU4F2, markers of amacrine and RGCs respectively, and OTX1/2, a marker of neurogenic RPCs biased towards photoreceptor fates as well as photoreceptor precursors (Figure 4A–E and M–Q; Supplemental Figure 3K, M, O, Q and S). In contrast, none of these neurogenesis markers were detected in the central retina of the LacZ mutant at E12.5 (Figure 4G–K). OTX1/2 was expressed in the peripheral ventral retina (Figure 4E and K; arrows), but as mentioned above, this expression is associated with compromised retinal identity.⁷ By E15.5 in the LacZ mutant, all markers were expressed in patterns consistent with neurogenesis, although stained cells were restricted to the central retina (Figure 4S–4W; arrows). Correlated with the delayed neurogenesis is reduced expression of the neurogenic competence marker SOX2 at E12.5 (Figure 4F and 4L), and restricted expression in a small central region of the LacZ mutant retina at E13.5 (Supplemental Figure 3S and T; arrows) and E15.5 (Figure 4R and X; arrows).¹⁷ These observations are consistent with a delay in the onset of retinal neurogenesis, to a similar extent as the orJ and R200Q mutants.

Figure 4 Retinal neurogenesis is delayed and centrally restricted in the LacZ mutant. Immunohistology assessing cytoskeletal maturation with TUBB3 (A, G, M and S), the presence of neurogenic progenitors with ATOH7 (B, H, N and T), retinal ganglion cells with POU4F2 (C, I, O and U), amacrine cells with BHLHB5 (D,J,P and V) and photoreceptor precursors with OTX1/2 (E, K, Q and W), reveal a delay in neurogenesis in the LacZ mutant retina, with markers lacking at E12.5 and centrally restricted at E15.5 (arrows in S-W). OTX1/2 staining also reveals ectopic staining in the ventral peripheral retina at E12.5 (compare red arrows in (E) and (K)), which is associated with compromised retinal identity. (F, L, R and X) Expression of neurogenic competence regulator SOX2 in RPCs is reduced at E12.5 (F and L) and centrally restricted at E15.5 in the LacZ mutant retina (R, X; arrow denotes region of persistent SOX2 expression). Scale bar: 100 µm.

We observed variation in the severity of the microphthalmic phenotype of LacZ mutant embryos at E15.5, ranging from similar to more severe than in orJ embryos (Figure 5A). Although the mean reduction in eye size of the LacZ and orJ mutants were similar in magnitude when compared to their respective controls (Figure 5B and Supplemental Table 1), variability was greater in the LacZ mutants as determined by the coefficients of variation (CV) for each genotype (Figure 5C; see Methods for formula). We also observed variation in the extent of neurogenesis, as assessed by TUBB3 expression, between LacZ mutants (Figure 5D–G). While neurogenic length and density (defined in methods) were similar in the LacZ and orJ mutants (Figure 5H, I and Supplemental Table 1), the extent of variation in these measurements were similar except for neurogenic length in the LacZ mutants, which had a larger CV compared to the other genotypes (Figure 5J and K). Linear regression analysis revealed that neurogenic density and eye size were highly correlated (Figure 5L), but this was not the case for neurogenic length and eye size (Figure 5M). These observations suggest a link between eye size and neuron production, but not between eye size and progression of the neurogenic wave from the central to peripheral retina.

Figure 5 Timing of neurogenesis initiation and E15.5 variation in LacZ mutant deviate from the orJ mutant phenotype. (A) Microphthalmia was similar in the LacZ and orJ mutants at E15.5, but severity was variable in the LacZ mutants (compare iii and iv). (B and C) Quantifications of eye size revealed similar degrees of severity (B; ****p<0.0001; ANOVA; sample sizes: 11 (Z/+), 7 (Z/Z), 11 (orJ/+), 6 (orJ/orJ)), but the coefficient of variation (CV) was highest in the LacZ mutants compared to the other genotypes (C). (D–G) TUBB3 expression, indicating neurogenesis, is centrally restricted and seemingly covaries with microphthalmia severity in the LacZ mutant at E15.5. (H and I) Neurogenic density (H) and length (I) are similarly decreased in the LacZ and orJ mutants (****p<0.0001; ***p<0.001; nsp>0.05; ANOVA; sample sizes: 5 (Z/+ and orj/+ combined (Het)), 7 (Z/Z), 4 (orJ/orJ)). (J and K) CVs for neurogenic density were similar across genotypes (J), but the CV for neurogenic length was specifically increased in the LacZ mutant (K). (L) Linear regression analysis of eye size and neurogenic density in LacZ mutant retinas at E15.5 (R2=0.944; F=83.44; p=0.0003; sample size: 7). (M) Linear regression analysis of eye size and neurogenic length in LacZ mutant retinas at E15.5 (R2=0.364; F=2.856; p=0.1518; sample size: 7). (N–Q) Neurogenesis initiates at E13.5 in LacZ mutant (compare (N) to (P)). In contrast, neurogenesis initiates in the orJ retina by E13.25 (Q). Scale bars: (A) 500 µm; (D-G; N-Q) 100 µm.

Given the variation in neurogenesis at E15.5, we hypothesized that differences between the LacZ and orJ mutant phenotypes began earlier. We therefore focused on the onset of neurogenesis, which exhibits a near stereotypical delay of two days in the orJ retina, initiating at approximately E13.5.⁸ Expression of TUBB3 (Figure 5N; box and inset) and OTX1/2 (Supplemental Figure 3R; arrow) was observed at E13.5 in the central retina of the LacZ mutant, but cells were sparse compared to the LacZ het (Supplemental Figure 3Q). Similarly, cells expressing the neurogenic progenitor marker ATOH7 were present but sparse in the LacZ mutant retina (Supplemental Figure 3K and L; arrow), whereas POU4F2, and BHLHB5 were not detected (Supplemental Figure 3M–P). We therefore examined TUBB3 staining in retinas from the LacZ and orJ mutants at E13.25. Indeed, neurogenesis was not observed at E13.25 in the LacZ mutant retina (n=7) but was in the orJ mutant retina (n=4; Figure 5O–Q). The sum of these observations suggests that the LacZ mutant exhibits subtle increases in phenotypic severity compared to the orJ mutant (see discussion).

Elevated MITF expression in the retina is a central driver of the ocular phenotypes in orJ, R200Q, and R227W mutant mice, and a semi-dominant negative allele of Mitf (Mitf^mi) is sufficient to significantly reduce phenotypic severity.^3,8,9,18 We predicted a similar outcome with the LacZ allele, and to test this, we crossed LacZ mutant mice with our Vsx2°^rJ/orJ; MITF^mi/+ line to generate Vsx2°^rJ/Z; MITF^mi/+ embryos. At both E12.5 and E15.5, Vsx2°^rJ/Z; MITF^mi/+ eyes were significantly larger than those of Vsx2°^rJ/Z; MITF^+/+ (Figure 6). This suggests that, like the other Vsx2 mutants, Mitf has a prominent role in causing microphthalmia in the LacZ mutant retina.

Figure 6 Functional rescue of embryonic LacZ mutant microphthalmic phenotype with Mitfmi/+ allele. The semi-dominant negative allele MITFmi/+ reduced the severity of microphthalmia in Vsx2°rJ//Z mutants at E12.5 (A–C) and E15.5 (D–F). ****p<0.0001; T-test; sample sizes for E12.5: 10 (orJ/Z;+/+),20 (orJ/Z;mi/+); sample sizes for E15.5: 14 (orJ/Z;+/+), 22 (orJ/Z;mi/+)). Scale bar: 1000 µm.

The effects of early Vsx2 dysfunction are compounded in the postnatal retina as seen by severe hypocellularity and tissue disorganization. Additionally, due to the specific postnatal requirement of Vsx2 for the specification and differentiation of the last neuronal cell class, bipolar cells, they are not generated.^1,6,11,19 These features are also evident in the LacZ mutant retina. At P0 (Figure 7), the LacZ mutant retina is reduced to a 1–3 cell-thick, epithelium consisting of dispersed RPCs or newly forming Müller glia (EdU, MCM6, SOX9) and differentiating retinal neurons (TUBB3) including RGC’s (POU4F2), amacrine cells (BHLHB5), and photoreceptors (RECOVERIN, OTX1/2). At P10 (Figure 8), the thin apical epithelium, composed largely of photoreceptors (RECOVERIN, OTX1/2) remains intact while an additional population of disorganized aberrantly proliferating RPCs or Müller glia (EdU, MCM6, Sox9), and neurons (TUBB3) including RGCs (POU4F2), and amacrine cells (BHLHB5) arises basal to the organized epithelium. Signs of degeneration emerge at P14 (Supplemental Figure 4), as the cellular organization at the apical surface becomes disrupted and hypocellular regions, possibly plexiform outgrowths, arise in the disorganized basal cell layer. By P28 (Figure 9), the retina has severely degenerated - consisting of disorganized retinal progenitor cells,specified Müller glia, RGCs, amacrine cells, and photoreceptors. Consistent with the other Vsx2 mutants, bipolar cells are not observed in the LacZ mutant retina at P14 (Supplemental Figure 4Q–T) or P28 (Figure 10) as revealed by PKCα (PKC) and VSX1. Interestingly, sparse PKC-positive cells were observed at P14, which could suggest limited bipolar cell formation, but this antibody also detects amacrine cells and likely accounts for the pattern observed at P28.⁶

Figure 7 Immunohistology at P0. The LacZ mutant retina at P0 is reduced to a thin epithelium, containing interspersed (A–D, G and H) proliferating RPCs (EdU, MCM6, SOX9). (E–P) Similarly, markers of differentiating cells (TUBB3, POU4F2, BHLHB5, OTX1/2, RECOVERIN) are expressed in a disorganized manner. In addition to marking RPCs, SOX9 could indicate precocious Müller glia differentiation. Scale bar: 100 µm.

Figure 8 Immunohistology at P10. (A–L) A disorganized population of proliferating cells (EdU, MCM6), new neurons (TUBB3, POU4F2, BHLHB5), and Müller glia (SOX9), persist in the LacZ mutant retina. (M–P) In contrast, the photoreceptor markers OTX1/2 and RECOVERIN suggest the presence of an organized apical epithelium. Scale bar: 100 µm.

Figure 9 Immunohistology at P28. The LacZ mutant retina at P28 is severely degenerated and disorganized. Lacking cellular organization, spherical (A–D) aberrantly proliferating cells are interspersed with (E–P) progenitors (MCM6, SOX9), differentiating neuronal cells (TUBB3), Müller glia (SOX9), retinal ganglion cells (POU4F2), amacrine cells (BHLHB5), and photoreceptors (OTX1/2, RECOVERIN). Scale bar: 100 µm.

Figure 10 Bipolar cells are not detected in the LacZ mutant retina. (A–D) The rod-bipolar marker PKC is not detected in the LacZ mutant retina at P10, but basally displaced cells are observed at P28. This antibody also detects amacrine cells at this age. (E–H) The cone-bipolar marker VSX1 is not detected in the LacZ mutant retina at P10 or P28. Scale bar: 100 µm.

Discussion

Based on the observations reported here, we propose that the LacZ allele exhibits loss-of-function properties with strong phenotypic similarities to the orJ null allele, which lacks protein expression, and the R200Q allele, which produces a DNA binding-deficient protein.³ All three mutants exhibit bilateral microphthalmia, reduced retinal proliferation, delayed neurogenesis, an absence of bipolar cells, and are affected by ectopic Mitf activity.^3,8,9,12,18 Because the mutant VSX2 protein is expressed, we cannot conclude that the LacZ allele is a complete knockout, and consistent with this, we observed subtle differences in the timing of neurogenesis onset in the LacZ mutant retinas compared to orJ mutant, and greater variability amongst LacZ mutants at E15.5 when compared to that of orJ mutants in both eye size and parameters of neurogenesis.

There are two reasons that could account for these differences. One is that the mutant protein is not inert (see below). The other is that the LacZ mutant phenotype is sensitive to its genetic background, which has been documented for the orJ mutant phenotype. The orJ mutation was discovered in the 129svimj background, but when outcrossed, phenotypic variations occur.^5,11,20 For example, increased pigmentation and more severe microphthalmia resulted when orJ mice were bred with mice with mixed genetic backgrounds including C57Bl6 mice.²¹ We also observed similar increases in phenotypic severity in an outcross with Black Swiss mice (unpublished). In contrast, less severe phenotypes were observed when orJ mice were crossed to p27Kip1 (Cdkn1b) knockout mice on a C57BL/6 genetic background⁶ or outcrossed with wild type Mus castaneus.¹¹ Relevant here is that our characterization of the LacZ allele was done in a mixed C57BL/6;129sv mixed genetic background which could contribute to variations in phenotypic severity.

It is also possible that the mutant VSX2 protein has activity, but what this activity might be is not clear. The mutant coding sequence extends from amino acid 1 through 151 (out of 361 or 380; https://www.ncbi.nlm.nih.gov/protein/NP_031727) and lacks the NLS, homeodomain and CVC domain, making direct DNA binding unlikely. The NES is retained, however, and could contribute to non-nuclear localization.¹⁷ Polyubiquitination and subsequent degradation of zebrafish Vsx1 protein is enhanced by the CVC domain and a conserved, putative PEST sequence located close to the CVC domain, and their absence could explain the apparent stability of the mutant Vsx2 protein.²² Interactions of the mutant protein with other proteins is possible, but domains or motifs have not been identified in the retained N-terminal portion of VSX2 other than putative disordered regions. It is also possible that the mutant protein has novel or neomorphic activity because it is no longer localized to the nucleus, but how this might alter the phenotype is hard to reconcile because the allele is not dominant. Interestingly, the R227W allele exhibits properties of a recessive neomorph.³ We attributed this to the ability of the R227W mutant protein, retaining weak DNA binding, to activate a cryptic transcriptional circuit that increased non-retinal gene expression, including Mitf, beyond that observed in the orJ mutant retina.³ Further work is needed to determine if the truncated VSX2 protein from the LacZ allele has activity.

Limitations of the Study

It is unclear whether the subtle differences in the phenotypes between the LacZ and orJ mutants are due to the expression of the mutant protein in the LacZ mutant, differences in genetic background, or both. One approach to address this issue is to minimize differences in genetic background by generating a congenic 129sv/imj strain carrying the LacZ allele. Its also unclear if the mutant protein has activity or whether a similar truncated protein with pathogenicity could exist in humans. This limitation could be addressed by generating CRISPR mutations in human iPSCs that are differentiated into retinal organoids. Finally, although eye size is improved in the combined LacZ/orJ mutant with the mi mutation, the retinal phenotype was not characterized as part of this study. Mi mutant mice with the LacZ allele in a homozygous state would be needed to compare the effects of reduced Mitf activity on the LacZ mutant phenotype compared to the orJ mutant phenotype.

In summary, the LacZ allele is a new addition to the allelic series of Vsx2 mutations, and its functional properties make it useful for further probing Vsx2 function in the in vivo context.

Experimental Procedures

Animals

The Vsx2^LacZ line was generated from ES cells containing the Vsx2^{tm1a(EUCOMM)Wtsi} allele. Germline founders were crossed with CMV-Cre mice (RRID:IMSR_JAX:006054) to excise the LoxP flanked region containing the Neo cassette and Vsx2 exon 3 to create the Vsx2^LacZ strain. Vsx2°^rJ mice (RRID:IMSR_JAX:000395) and Mitf^mi mice (RRID:IMSR_JAX001573) are maintained in the laboratory.

Tissue Collection

Embryos and pups were generated using single night matings, with the day that the vaginal plug was observed considered E0.5. Vsx2^LacZ embryos were collected at E10.5, E11.5, E12.5, E13.5, E15.5 and Vsx2^LacZ pups were collected at P0, P10, P14, and P28. All embryos and pups were staged using whole-body attributes.^23–25

To assess proliferation, EdU was delivered via intra-peritoneal injection at a dose of 30 µg/g body weight at two hours prior to euthanasia of dam for embryonic dissection and 45 minutes prior to euthanasia for postnatal pups. Dams and pups from P10 and older were euthanized using CO₂ and cervical dislocation; P0 pups were anesthetized using hypothermia and thereafter euthanized via decapitation. At E15.5, embryo eyelids were removed for imaging to allow for accurate eye size quantification.

Tissue Preparation

Embryo heads and postnatal pup eyes were fixed at room temperature in 4% PFA-PBS solution for 45 minutes or 1 hour, respectively. After several PBS washes, a sucrose cryopreservation series was performed before the samples were frozen in TissueTek OCT (Sakura Finetech) and stored at −80°C. Tissue was sectioned in twelve micrometer sections at −20°C with a Leica CM1950 cryostat and adhered to Superfrost Plus slides (Fisher Scientific). The sections were then dried for 1–2 hours before storage at −20°C and −80°C.

PCR

DNA was extracted from embryonic tissue samples by adding 75 µL alkaline lysis buffer to sample tube, heating to 95°C for 10 minutes, cooling for 10 minutes, followed by adding 75 µL of neutralization buffer. DNA extractions were stored at −20°C. In a 10:1 ratio, 9 µL of master mix was added to 1 µL extracted DNA and run in a ProFlex thermocycler (Life Technologies). A 1.5% agarose gel was used to run the PCR product alongside a Quick-Load 1kb Plus DNA Ladder at 70 V for 95 minutes. A UVP Gel Doc-It^TS3 Imager was used to image the gel. The PCR and gel electrophoresis confirming germline transmission of the LacZ allele is a representative image from three replicates.

Western Blot

Freshly dissected E15.5 retinal tissue was flash frozen in liquid nitrogen and stored at −80°C. Lysates were prepared with RIPA buffer and EDTA-free protease inhibitor (Millipore Sigma) at a volume of 100 µL per 50 mg of tissue. Lysates for heterozygotes contained 2 retinas per sample (1 animal) and 4 retinas per sample (2 animals) for homozygotes. BCA assay (Quick Start Bradford, Bio rad) was used to determine protein concentration using a wavelength of 595 nm. 20 ug of protein per lane was loaded in a 20 µL volume. Gels were run for 40 min at 90 V and then 40 min at 110 V. Proteins were transferred using an iBlot2 (Fisher Scientific) system with a setting of 15 V in 7 mins. Blots were incubated for 1 hour at room temperature in 1X PBS, pH 7.4, 0.1% Tween 20 containing 5% bovine serum albumin (PBST/B), followed by incubation with constant rocking overnight at 4°C with primary antibody diluted in PBST/B (Table 1). Excess antibody was removed, and blots were washed 3 times for 10 minutes each with PBST/B, then incubated for 1 hour at room temperature with secondary antibody diluted in PBST (Table 2). Excess antibody was removed and blots were washed 6 times for 10 minutes each with PBST. Chemiluminescence HRP substrate (Immobilon Western, Millipore) was used for signal detection and images were captured on an Amersham Imager 600 (General Electric). The Western blot presented is a representative image from 3 replicates.

Table 2 Secondary Antibodies

Immunohistochemistry and EdU Detection

To remove OCT, slides were washed in PBS and incubated in blocking solution (PBST (0.05% Tween 20) containing 2% normal donkey serum (PBST/DS)) for 1 hour at room temperature. Slides were then incubated overnight at 4°C in primary antibody diluted in PBST/DS (Table 1). Excess antibody was removed, slides were washed 3 times for 5 minutes each with PBST, followed by incubation for 1 hour at room temperature in a dark humidified chamber with secondary antibodies diluted in PBST/DS (Table 2). Excess antibody was removed, and slides were washed 3 times for 10 minutes each with PBST. EdU incorporation was detected using AlexaFluor 568 or 647 Click-iT Cell Reaction Kit (Invitrogen-Molecular Probes, Eugene, Oregon). DAPI (1:40000 in PBS) was used to label nuclei. After final washes in dH₂O, slides were air dried and coverslipped with Fluoromount and stored at 4°C until imaging. All immunohistology expression patterns are representative images from a sample size of more than three, collected from at least two different litters for each stage and genotype.

Quantifications

Eye size was quantified using the pentagon-ROI tool of Fiji on embryo pictures taken on a Nikon C-FIDH SMZ1270. Eye size was defined by the area of the two-dimensional circumference of the pigmented eye (sphere). Area was measured in pixels² and converted to µm². Neurogenic length, as defined by the linear length of the retinal basal surface that TUBB3-positive cells occupy, was quantified using the freehand line-ROI tool of Fiji and the percentage calculated by dividing the neurogenic length by the total length of the retinal basal surface and multiplying by 100. Neurogenic density, as defined by the number of TUBB3-positive pixels divided by the total number of pixels in the neurogenic region (the apicobasal region of the retina containing TUBB3-positive cells), is used as a proxy for cells within the neurogenic region. This was quantified using Fiji by creating an ROI outlining the TUBB3-positive region of the retina and clearing the non-ROI regions using Edit>Clear Outside. The neurogenic region was then made binary using Threshold tool, then the number of positive (white) and negative (black) pixels were quantified using Analyze>Histogram. The neurogenic density was then calculated using the ratio of positive pixels to total pixels in the neurogenic region.

Statistical Analysis

Prism was used to run all statistical analyses and to create graphical representations. Unpaired two-tailed T-tests were used for comparisons limited to 2 conditions (Figure 1L; Figure 5B, H, I; Figure 6C and F). The coefficient of variation within a sample group was calculated using the formula CV=(Standard Deviation / Sample Mean)*100 (Figure 5C, J and K). One way ANOVAs accompanied by post-hoc analyses (Brown-Forsythe, Bartlett’s, Tukey’s multiple comparison tests) were run using Prism on the E15.5 eye size, neurogenic length, and neurogenic density data to thoroughly compare the heterozygous and homozygous phenotypes of the LacZ and orJ alleles (Figure 5B, H and I). Linear regression was used to describe the relationship between eye size and measures of neurogenesis (neurogenic density (Figure 5L) and neurogenic length (Figure 5M)). Supplemental Table 1 lists the results of all statistical analyses.

Imaging/Microscopy

Whole embryos were imaged using Nikon C-FIDH SMZ1270. Immunofluorescent staining at E10.5-E11.5 and postnatal stages were imaged using a Nikon Eclipse E600 while those at E12.5-E15.5 were imaged using a Zeiss LSM710 confocal microscope with 20X lens to capture a high-resolution image of the relatively larger ROI. Image histograms were adjusted using Fiji. Images were resized and trimmed using Adobe Photoshop. Figures were made using Adobe Illustrator.