**3. The Emergence of a Specific and Stable Vitamin A Transport Mechanism that Coincided with Major Changes in Vitamin A Functions**

The tremendous expansion in the biological functions of retinoids, the dependence on vitamin A for survival, and toxicity associated with their random diffusion demand a specific and stable mechanism of vitamin A transport. The concomitant emergence of monostable pigments for vision also demands a specific and stable mechanism of vitamin A transport because the constant release of free retinal by monostable pigments (after every photoisomerization event) and the constant recycling of retinoid between two cell types in daylight inevitably causes loss of vitamin A (absorption of one photon initiates one cycle). Indeed, the diversification of vitamin A functions and the switching of visual pigments from bistable pigments to monostable pigments in evolution coincided with the emergence of a specific and dedicated vitamin A transport mechanism (Figure 3). This mechanism of vitamin A transport is mediated by the plasma retinol binding protein (RBP), a specific and sole carrier of vitamin A in the blood [131–136], and its specific membrane receptor STRA6, which mediates cellular vitamin A uptake [137].

**Figure 3.** Summary diagram of the key events in the evolution of vitamin A functions that coincide with the emergence of RBP/STRA6-mediated specific vitamin A transport.

Surprisingly, evolution seems to have produced the RBP receptor STRA6 from scratch because it is not homologous to any membrane receptors or transporters of known function and represents a new type of cell-surface receptor [138]. In contrast, ABCA4, a transporter for vitamin A derivatives, belongs to an ancient family of ATP-dependent transporters. STRA6 employs a membrane transport mechanism distinct from known celluar mechanisms including active transport, channels, and facilitated transport [139,140]. STRA6's vitamin A uptake is coupled to intracellular proteins involved in retinoid storage such as LRAT [137,141,142] or CRBP-I [139], but no single intracellular protein is absolutely required for its vitamin A uptake activity [139,140]. At the biochemical level, STRA6 has diverse catalytic activities such as catalyzing retinol release from holo-RBP [139,140], retinol loading into apo-RBP [139,142], retinol exchange between RBP molecules [140], and retinol transport from holo-RBP to apo-CRBP-I [139]. Depending on extracellular RBP species (the ratio of holo-RBP to apo-RBP) and intracellular proteins (the presence of CRBP-I or LRAT), STRA6 can promote retinol influx, retinol efflux or retinol exchange [140]. How STRA6 achieves its biological activities is not well understood. STRA6 has 9 transmembrane domains, 5 extracellular domains and 5 intracellular domains [143]. Between transmembrane 6 and 7 is an essential RBP binding domain [144].

Studies in human genetics and in animal models have revealed the critical functions of RBP and STRA6. Partial loss of RBP function leads to RPE dystrophy at a young age in humans [145,146]. Complete loss of RBP is embryonic lethal under vitamin A deficient conditions that mimic the natural environment [147]. RBP is required to mobilize liver-stored vitamin A [148]. Complete loss of STRA6 in human causes wide-spread pathogenic phenotypes in many organs [149,150]. Loss of STRA6 causes highly suppressed tissue vitamin A uptake in both zebrafish [142] and mouse [151]. Loss of STRA6 leads to the loss of most stored vitamin A in the eye and subsequent cone photoreceptor degeneration, consistent with previous findings that loss of visual chromophore causes cone photoreceptor degeneration [152–155].

STRA6 knockout causes the loss of 95% of the retinyl ester store in the RPE cells, the key cell type responsible for vitamin A uptake and storage for vision [151]. What is responsible for the STRA6-independent 5%? RBP/STRA6-mediated specific vitamin A transport is not the only mechanism of vitamin A delivery. Vitamin A, like many hydrophobic drugs, has a theoretically much simpler mechanism of transport by random diffusion. However, virtually all vitamin A in vertebrate blood is bound to RBP. The other most dominant mechanism is mediated by retinyl esters in the blood, as revealed by studies of RBP knockout mice [147,156]. Consistently, RPEspecific LRAT knockout also revealed that the RPE can take up retinyl esters without LRAT [157]. The LRAT-independent uptake of retinyl esters by the RPE is more than sufficient to account for the residual retinyl ester in STRA6 knockout mice [151]. This suggests that STRA6 is responsible for virtually all retinol accessible to LRAT in the RPE.

Retinyl ester bound to chylomicron is the primary vehicle that transports dietary vitamin A absorbed by the small intestine to the liver, the primary organ for vitamin A storage [158,159]. There is also strong experimental evidence that a fraction of the retinyl esters can be absorbed by peripheral organs as well [133,159]. This vitamin A transport mechanism is independent of RBP/STRA6. If retinyl ester in the blood can deliver vitamin A, why do we need RBP/STRA6? The many differences between the two mechanisms can answer this question (Table 5). The RBP/STRA6-mediated transport is a sustained and specific mechanism. The high affinity and specificity in RBP's binding to STRA6 can target the vitamin A/RBP complex to specific cells that specialize in vitamin A uptake and storage (e.g., the RPE cell). Although retinyl ester in the blood is capable of partially compensating for the loss of RBP or STRA6 under vitamin A sufficient or excessive conditions, it "borrows" lipid transport pathways, which target a much wider variety of cell types (beyond those specialized in vitamin A uptake and storage) and cannot be relied on during vitamin A deficiency, which is common in natural environments. Studies in both animals [160] and humans [111] revealed that more toxicity is associated with vitamin A delivery independent of RBP. An increase above 10% in retinyl ester in the blood is regarded as a sign of vitamin A overload [111,131].



There exists a STRA6 homolog. The function of this homolog is an intriguing question [161,162]. A recent study found that it is mostly expressed in the liver and the small intestine in mice and can take up vitamin A from holo-RBP similarly to STRA6 [163]. Since transfer of retinol within the liver does not depend on RBP, and liver largely obtains its stored vitamin A from chylomicron remnants [159], this receptor may help certain liver cells to obtain vitamin A from holo-RBP in the circulation. The small intestine absorbs vitamin A or its precursors from food and secretes retinyl esters bound to chylomicrons to be delivered to the liver for storage [158,159]. Because there is no retinol/RBP complex in the intestinal lumen, this receptor likely helps small intestine cells not directly accessible to vitamin A from food to obtain vitamin A from the circulation.

#### **4. The Eye and Vitamin A**

The earliest structure remotely related to an eye is the eyespot, a light sensing structure in the green alga *Chlamydomonas*. Although the human eye is vastly more complex than the eyespot, and the structures are separated by billions of years of evolutionary time, both serve a similar biological function in perceiving light, and both depend on vitamin A (Figure 4). Despite the growing dependence of other organs on vitamin A in evolution, the eye is still the organ most dependent on vitamin A. For human, the eye is the organ most sensitive to vitamin A deficiency, the loss of RBP, or the loss of STRA6 (Table 6). Given both the essential functions and toxicity of retinoids, how the eye regulates its vitamin A uptake to obtain a sufficient but not excessive amount is still poorly understood.

**Figure 4.** Comparison of two retinal-based light sensing structures: the eyespot in Chlamydomonas reinhardtii and the human eye. The human eye depends on vitamin A not only for light sensing for vision and the biological clock, but also for embryonic development and for the maintenance of the cornea. Cells or structures that depend on vitamin A are labeled in red.


**Table 6.** In both mice and humans, the eye is the organ most sensitive to vitamin A deficiency, loss of RBP, or loss of STRA6.

Nutritional blindness due to vitamin A deficiency is still a leading cause of blindness in the world. Vitamin A deficiency can deprive the photoreceptor cells of the visual chromophore [164]. In addition, vitamin A deficiency causes the disorganization of rod photoreceptor outer segments, degeneration of cone photoreceptor cells, and the loss of LRAT expression in the RPE [165]. If rod and cone photoreceptor cells that sense light for vision depend on vitamin A, what about sensing light for the biological clock, which needs to be frequently readjusted by light? An early study using a mammalian model showed that the spectral sensitivity of the photoreceptors that mediate light's entrainment of the biological clock is indicative of a vitamin A-based light sensor that peaks at 500 nm [166]. Although there was a debate on whether it might be flavin-based, recent studies confirmed that it is vitamin A based and revealed that visual pigments in rod and cone and melanopsin in light-sensitive ganglion cells all contribute to this light sensing function.

Vitamin A, a chemical originally used only for light sensing, is now also an essential molecule for eye development. Retinoic acid, the acid form of vitamin A, plays critical roles in retina and eye development [167–172]. The human eye does not develop without STRA6, the RBP receptor that mediates vitamin A uptake [149,150,173]. STRA6's influence on eye development may not be limited to its expression within the eye itself. One of the organs that expresses the highest level of STRA6 is the placenta, the maternal-fetal barrier which supplies essential nutrients for fetal development. STRA6 can also influence eye development by supplying retinoid to developing embryos in general.

In addition to sensing light for vision and circadian rhythm and eye development, vitamin A also plays crucial roles in maintaining a healthy cornea [174,175]. Without vitamin A, the cornea develops ulceration. Corneal dryness due to vitamin A deficiency is another common cause of human blindness. This role of vitamin A is likely related to one of vitamin A's general functions in epithelial maintenance and stem cell differentiation. How the cornea absorbs vitamin A physiologically is still poorly understood.

Although human vision in a sense perfectly serves our daily needs, we are living with the consequences of the choice of monostable pigments in evolution. If this choice helped our ancestors survive at night, it came at surprisingly high costs. It is astonishing to realize that "every" photon we see depends on a complex enzymatic cycle that consumes cellular energy and releases free toxic retinoid. As we see using our cones in natural daylight or artificial light, a staggering amount of energy is consumed, and a constant flux of toxic free retinoid is cycling between cells to regenerate rhodopsin, which plays no role in daylight vision. In a sense, a whole range of human diseases, from our vision's high sensitivity to vitamin A deficiency to Stargardt macular dystrophy, are the price we pay for this evolutionary choice.

#### **5. Conclusion**

For most of evolutionary history starting about 3 billion years ago, vitamin A has functioned as a light sensor. Vitamin A-based light sensors span a wide range of absorption maxima from UV to near infrared. This range matches the peak irradiance of sunlight on earth's surface, the dominant light source in evolution that determines the "visible" light for fish in deep sea or human beings. The major changes during the evolution of vitamin A-based light sensors are the switch from light-gated ion channels to light-activated G-protein coupled receptors and the switch from bistable pigments to monostable pigments for vision. Vitamin A's biological functions have also been tremendously expanded to include its crucial roles in regulating cell growth and differentiation from embryogenesis to adulthood. The likely driving forces for the evolution of a sustained, efficient and precise system of vitamin A transport are the high demand for vitamin A by vision (due to monostable pigments that constantly release the chromophore in daylight), the high toxicity associated with excess vitamin A, and the need to survive vitamin A deficiency, which is common in the natural environments. Because an imbalance in vitamin A homeostasis is associated with diverse human diseases including blindness and birth defects, a better understanding of how vitamin A is transported to the right cell type in the appropriate amount will help to devise new strategies to treat many human diseases caused by insufficient or excessive tissue retinoid levels or to use retinoids as therapeutic agents.

#### **Acknowledgment**

Supported by National Institutes of Health grant R01EY018144. H.S. is an Early Career Scientist of the Howard Hughes Medical Institute.

### **References**

