A major effort in our lab is directed toward understanding the transcriptional regulatory networks that orchestrate the development and function of photoreceptors. We are employing a wide range of experimental and computational techniques to decipher these networks. Recently, we generated comprehensive maps of rod- and cone-specific open chromatin using ATAC-seq and have leveraged these maps to elucidate the differences in cis-regulatory grammar between these two cell types. We are now using a massively parallel reporter assay called CRE-seq to further interrogate the architecture of photoreceptor cis-regulatory elements. Our ultimate goal is to create a complete, quantitative model of photoreceptor transcriptional regulation including a detailed cis-regulatory grammar. This model will serve as a template for translating between both coding and non-coding variants and the complex cellular phenotypes of photoreceptors that result in blindness.

A prime goal of regenerative medicine is to direct cell fates in a therapeutically useful manner. Retinitis pigmentosa is one of the most common degenerative diseases of the eye and is associated with early rod photoreceptor death followed by secondary cone degeneration. We hypothesized that converting adult rods into cones, via knockdown of the rod photoreceptor determinant Nrl, could make the cells resistant to the effects of mutations in rod-specific genes, thereby preventing secondary cone loss. To test this idea, we engineered a tamoxifen-inducible allele of Nrl to acutely inactivate the gene in adult rods. This manipulation converted the rods into cone-like cells, thereby preventing retinal degeneration in a mouse model of blindness. This rescue is the first therapeutic effect that has ever been achieved via direct cellular reprogramming in the mammalian central nervous system. Current efforts are directed toward achieving complete reprogramming of rods into cones. If successful, reprogrammed rods could serve as a localized, in situ source of novel cones in diseases that preferentially afflict this cell type, such as age-related macular degeneration.

Recent studies have shown that inner retinal neurons undergo dramatic changes in their morphologic and physiological properties in response to photoreceptor degeneration. Surprisingly, almost nothing is known about the transcriptional and epigenetic changes that underlie these responses. In order to investigate this phenomenon, we are using expression profiling and epigenomic mapping to pinpoint the molecular responses of bipolar cells to photoreceptor disease. These studies will inform strategies for treating blindness that rely on optogenetic and chemogenetic transduction of bipolar cells.

Optogenetics holds tremendous potential for restoring vision to individuals with late-stage retinal degeneration. One promising therapeutic strategy is to express a light-sensitive protein (i.e., optogenetic actuator) in non-photosensitive retinal interneurons to restore vision. A major limitation of current approaches is that they require the application of very high-intensity blue light to stimulate the actuator, posing a risk of retinal photodamage. We are developing a novel biomimetic strategy for red-shifting the actuator such that it can be efficiently activated by far red light (> 650 nm), thus minimizing retinal injury. Our approach is based on a strategy used by migrating fish to enable better vision in turbid water. When salmon migrate from the open ocean (where incident light is in the 450-500 nm range) into inland streams (where incident light is significantly red-shifted), they switch from using retinal as their visual chromophore to 3,4-didehydroretinal which has red-shifted spectral properties. We recently identified Cyp27c1 as the enzyme that mediates this conversion, and we plan to co-express it with optogenetic actuators in mammalian neurons in vivo, thereby red-shifting their action spectra. In this way, we hope to create improved optogenetic strategies for treating blindness.

Some vertebrate species have evolved means of extending their visual sensitivity beyond the range of human vision. One mechanism of enhancing sensitivity to long-wavelength light is to replace the vitamin A₁-based chromophore in photopigments with one derived from vitamin A₂. We recently showed that a cytochrome P450 family member, Cyp27c1, mediates this switch by converting vitamin A₁ into vitamin A₂, thereby permitting fish and amphibians to dynamically enhance their ability to see infrared light. We are now exploring the molecular and genetic regulation of this switch in response to changing environmental variables. In addition, we are investigating the evolutionary origins of this remarkable sensory adaptation by investigating similar A₁/A₂ switches that occur in lampreys and crayfish. These studies promise to expand our understanding of the molecular and evolutionary origins of long-wavelength light sensitivity in vertebrates.

Color vision in birds is mediated by four types of cone photoreceptors whose maximal sensitivities (λ_max) are evenly spaced across the light spectrum. In the course of avian evolution, the λ_max of the most shortwave-sensitive cone, SWS1, has switched between violet (λ_max > 400 nm) and ultraviolet (λ_max < 380 nm) multiple times, thereby permitting birds to see very short-wavelength light that is invisible to humans. We are investigating the chemical and molecular mechanisms whereby birds have achieved ultraviolet vision. We have found that optimal short-wavelength vision depends on the presence of carotenoid filters that are deposited inside unique optical organelles called oil droplets that are present inside avian cone photoreceptors. We are investigating the enzymatic and cellular mechanisms whereby birds construct oil droplets and fine-tune their spectral properties for optimal color vision.

Unlike the human retina in which red and green cones are essentially randomly distributed, the five cone types of the avian retina are arrayed in highly regular, self-organizing mosaics that are spatially independent of one another. Optimal spatial sampling of light requires that photoreceptors be arranged in perfectly ordered arrays, yet avian photoreceptor mosaics are not perfectly regular. This finding suggests the existence of unknown constraints that limit their regularity. We have found evidence that physical packing of cells within the photoreceptor epithelium sets an upper limit to the spatial orderliness of avian photoreceptor mosaics. In other words, it appears that the bird's mosaics are as regular as they can be, given the constraints of packing within the epithelium. These findings show how fundamental physical principles can limit optimization within a biological system and have implications for the design of artificial sampling arrays such as the pixel boards of color digital cameras. Future studies will be directed toward understanding the developmental and cellular mechanisms whereby these remarkable avian photoreceptor mosaics are formed.

Birds are among the most colorful of vertebrate species. Many of the bright yellow, orange, and red colors of birds are produced by carotenoids deposited in the feathers. Birds obtain yellow carotenoids from their diet and either deposit them directly or modify them via the addition of ketone groups at the end of the molecules to produce red ketocarotenoids. In collaboration with Geoff Hill and Miguel Carneiro, we recently used a canary model to identify CYP2J19 as the enzyme that mediates the formation of ketocarotenoids in birds. In addition to answering the longstanding question, “What makes the cardinal red?”, this discovery promises to have a major impact on the field of avian ecology and evolution since red feather coloration is considered to be one of the most important sexually selected indicators of individual fitness in birds. In the future, we plan to elucidate the cis-regulatory mechanisms whereby certain bird species have evolved the ability to express CYP2J19 in the developing feather follicle, thereby endowing them with red feathers. In addition, we plan to explore the unique role of this enzyme in enhancing long-wavelength vision in birds.

A link between human cognition and neuropsychiatric disease has long been suspected, but individual genetic loci have not yet been identified. In analyzing the results of two recent genome-wide association studies (GWAS), we found a non-coding region upstream of POU3F2 that is associated with increased cognitive performance and increased susceptibility to bipolar disease. This region contains multiple human fetal brain-specific open chromatin regions, which may represent developmental enhancers. We hypothesize that a single causal variant resides in this region, and that this variant affects the activity of a brain-specific cis-regulatory element that regulates POU3F2 expression in the developing brain, thereby enhancing cognitive performance and increasing risk of bipolar disorder. Using a wide variety of techniques, including the creation of humanized mouse models carrying individual human SNPs and reporter assays in human iPSC-derived cerebral organoids, we have begun to evaluate candidate causal variants underlying this signal. These studies promise to demonstrate the first genetic linkage between cognitive performance and bipolar disorder, and open the door to a deeper understanding of the genetic basis of human cognition.