Quantitative Analysis of Dietary Vitamin A Metabolites in Murine Ocular and Non-Ocular Tissues Using High-Performance Liquid Chromatography

Summary

Here, a normal-phase, high-performance liquid chromatography method is described to detect and quantify critical retinoids involved in the facilitation of visual function in both ocular and systemic tissue, in the context of the systemic vitamin A supply to generate the essential photosensitive rhodopsin chromophore 11-cis-retinal.

Abstract

G protein-coupled receptors (GPCRs) are a superfamily of transmembrane proteins that initiate signaling cascades through activation of its G protein upon association with its ligand. In all mammalian vision, rhodopsin is the GPCR responsible for the initiation of the phototransduction cascade. Within photoreceptors, rhodopsin is bound to its chromophore 11-cis-retinal and is activated through the light-sensitive isomerization of 11-cis-retinal to all-trans-retinal, which activates the transducin G protein, resulting in the phototransduction cascade.

While phototransduction is well understood, the processes that are involved in the supply of dietary vitamin A precursors for 11-cis-retinal generation in the eye, as well as diseases resulting in disruption of this supply, are not yet fully understood. Once vitamin A precursors are absorbed into the intestine, they are stored in the liver as retinyl esters and released into the bloodstream as all-trans-retinol bound to retinol-binding protein 4 (RBP4). This circulatory RBP4-retinol will be absorbed by systemic organs, such as the liver, lungs, kidney, and eye. Hence, a method for the quantification of the various metabolites of dietary vitamin A in the eye and systemic organs is critical to the study of proper rhodopsin GPCR function.

In this method, we present a comprehensive extraction and analytical method for vitamin A analysis in murine tissue. Through normal-phase, high-performance liquid chromatography analysis, all relevant isomers of retinaldehydes, retinols, and retinyl esters can be detected simultaneously through a single run, which allows for the efficient use of experimental samples and increases internal reliability across different vitamin A metabolites within the same sample. With this comprehensive method, investigators will be able to better assess systemic vitamin A supply in rhodopsin GPCR function.

Introduction

G protein-coupled receptors (GPCRs) are one of the most studied and characterized superfamily of proteins known. In its most well-known function, GPCRs serve as a cell surface receptor in signal transduction, initializing intracellular responses upon binding with a specific ligand. GPCRs are characterized by seven transmembrane (TM) helical domains and six total loop domains. Of the six loops, three loops are oriented extracellularly to facilitate ligand binding, while the other three intracellular loops are coupled to a heterotrimeric G protein consisting of the Gα, Gβ, and Gγ subunits^1,2.

GPCRs are classified into several classes, including Class A Rhodopsin-Like, Class B Secretin Receptor family, Class C Glutamate, Class D Fungal Mating Pheromone Receptors, Class E Cyclic AMP receptors, and Class F Frizzled/Smoothened^3,4. As its name suggests, the GPCR rhodopsin-like Class A subclass includes rhodopsin, the critical GPCR responsible for phototransduction and visual function. Rhodopsin contains all the pertinent key characteristics and structural elements that are found in the canonical model of GPCRs, including the previously mentioned seven TM helical domains, the six extracellular and intracellular loops, and association with a heterotrimeric G protein, also known as transducin (G_t) in photoreceptors^1,5,6,7. Within the binding pocket of rhodopsin, 11-cis-retinal, the light-sensitive chromophore ligand, binds to rhodopsin on lysine 296 through a covalent Schiff base linkage, thus forming 11-cis-retinylidene^1,8. Upon absorption of a photon, 11-cis-retinylidene photoisomerizes into all-trans-retinylidene, inducing a conformational change within rhodopsin. Therefore, the 11-cis-retinal ligand is critical to the function of the rhodopsin GPCR, and a robust and efficient supply of 11-cis-retinal must be continuously maintained to overcome the high turnover rate within photoreceptors.

Retinaldehydes such as 11-cis-retinal belong to a group of molecules collectively called retinoids, and biologically relevant retinoids are more widely referred to as vitamin A. Retinoids are characterized by a cyclic end group connected to a conjugated polyene chain, with a polar end group at the other end. Retinaldehydes and associated vitamers of vitamin A are no exception to this characterization, which contain the β-ionone ring as the cyclic end group, a diterpene polyene chain, and a differing polar end group depending on the vitamer, that is, aldehyde group for retinaldehydes, hydroxyl group for retinols, carboxyl group for retinoic acids, ester bond for retinyl esters, etc (Figure 1)^9,10.

Mammals cannot synthesize vitamin A de novo, but plants can; therefore, all retinoids within mammalian systems must originate from the diet of plant-based producers to the consumers in the food chain. In the canonical model of vitamin A metabolism, β-carotene, the archetypal plant provitamin A, is absorbed into the intestinal enterocyte through the scavenger receptor class B, member 1 (SCARB1), cleaved into two molecules of all-trans-retinal by β-carotene oxygenase 1 (BCO1/BCMO1), which binds to retinaldehyde binding protein 2 (RBP2) and is reduced to all-trans-retinol by retinol dehydrogenases (RDH), converted into retinyl esters by lecithin retinol acyltransferase (LRAT), and then sent to the bloodstream in chylomicrons^11,12,13,14. Retinyl esters, such as retinyl palmitate, on the other hand, serve as the predominant provitamin A from animal sources. Retinyl palmitate from the intestinal lumen is hydrolyzed into all-trans-retinol by carboxylesterase 1 (CES1) and diffuses into the intestinal enterocyte¹⁵. The liver is the primary storage and homeostatic organ for vitamin A homeostasis, which absorbs the retinyl esters within these chylomicrons, which are hydrolyzed into all-trans-retinol bound to cellular retinol-binding protein 1 (CRBP1) by retinyl ester hydrolases, enters hepatic stellate cells and is converted back into retinyl esters by LRAT for storage^13,16,17. To maintain a homeostatic level of vitamin A in the organism, the liver releases vitamin A in the form of all-trans-retinol bound to a serum transport complex, consisting of retinol-binding protein 4 (RBP4) and transthyretin (TTR)^15,18,19. This complex will be referred to as holo-RBP4 in this manuscript.

To use this systemic vitamin A supply in the blood, systemic tissues, including ocular tissue where a robust source of vitamin A is maintained, must have a method to absorb holo-RBP4 into tissue. Within the photoreceptor-rich retina in ocular tissue, the membrane receptor stimulated by retinoic acid 6 (STRA6) is the transporter implicated in this function. In mechanistic studies, STRA6 has been shown to be capable of facilitating the intake of extracellular all-trans-retinol from holo-RBP4 into the RPE²⁰. This imported all-trans-retinol will then enter the visual cycle, which is the process by which all-trans-retinol is converted into 11-cis-retinal within the RPE and the photoreceptor outer segment, thereby facilitating visual function when bound to rhodopsin^9,21.

Once all-trans-retinol from circulatory holo-RBP4 crosses the blood-retina barrier into the RPE within ocular tissue through STRA6, all-trans-retinol in the RPE is first esterified into retinyl esters by LRAT, then hydrolyzed into 11-cis-retinol by retinal pigment epithelium-specific 65 kDa protein (RPE65). 11-cis-retinol is then converted into 11-cis-retinal by the retinol dehydrogenase 5. This 11-cis-retinal is then carried into the photoreceptor's outer segment (OS) by the interphotoreceptor retinoid-binding protein (IRBP)^9,21. Within the endoplasmic reticulum that surrounds the photoreceptor nucleus within the outer nuclear layer (ONL), the opsin GPCRs are synthesized and transported across the connecting cilium (CC). The motor proteins that are involved in this transport across the CC are contentious, but current hypotheses implicate kinesin and dynein-based intraflagellar transport (IFT) or myosin-based transport as being likely facilitators of this process^{14,22,23,24,25,26}. Once these two components meet within the membranous disks within the OS, 11-cis-retinal and opsin form 11-cis-retinylidene through a Schiff base covalent linkage at lysine 196 on rhodopsin, ready for phototransduction⁸.

While the expression of STRA6 within the RPE of the retina helps facilitate the intake of all-trans-retinol from holo-RBP4, STRA6 was not found to be expressed in the liver, despite its role as the main homeostatic organ for vitamin A and exhibiting capabilities in intaking all-trans-retinol from holo-RBP4^{15,19,27,28,29,30,31}. Eventually, an analogous receptor called retinol-binding protein 4 receptor 2 (RBPR2) was discovered, exhibiting the capability to intake all-trans-retinol from holo-RBP4, much like STRA6, but is expressed in hepatic tissue³².

Therefore, a complete understanding of the role of rhodopsin in visual function necessitates an understanding of the biological processes that culminate in the regeneration of the visual pigment. This is, in turn, intimately related to the previously described processes, including the metabolism of provitamin A precursors, storage within the liver, release of holo-RBP4 by the liver, and eventual uptake of holo-RBP4 through STRA6 and RBPR2 membrane receptors. As mentioned above, animal models such as mice remain one of the premier models in the study of such processes. Hence, we would like to present an extraction method for retinoids in murine tissue, as well as a normal-phase high-performance liquid chromatography (HPLC) method that can detect and quantify these retinoids. Using these methods, the important retinoids described above, such as the 11-cis-retinal rhodopsin ligand or the main transport retinoid all-trans-retinol, can be analyzed in ocular, hepatic, and systemic organs. By assessing retinoid supply in murine tissue, our understanding of the disease states and pathologies related to the logistical supply of retinoids can be further advanced.

Besides functioning as a chromophore in visual function through association with opsin GPCRs, retinoids also play a major role in mammalian cell signaling through retinoic acid signaling, facilitated by two families of nuclear receptors, retinoic acid receptors (RARs) and retinoid X receptors (RXRs), that bind directly to DNA and regulated gene transcription³³. These two families or receptors both utilize retinoids in the form of retinoic acids as the ligand. RARs have been shown to have affinity for both all-trans-retinoic acid and 9-cis-retinoic acid, whereas RXRs express affinity for only 9-cis-retinoic acid^34,35. Retinoic acids in uncontrolled quantities are teratogenic, and retinoic acid signaling must be extremely tightly controlled³⁶. Production of retinoic acids for signaling must occur locally and at very specific time points for the proper development of tissues, such as in hindbrain and limb development, but countless other examples utilize retinoic acid signaling^37,38. Within cells participating in retinoic acid signaling, retinoic acids are synthesized by two groups of enzymes, alcohol/retinol dehydrogenases (ADHs/RDHs) that facilitate the oxidation of retinols taken in by STRA6 or RBPR2 to retinaldehydes, and retinaldehyde dehydrogenases (RALDHs) that facilitate oxidation of retinaldehydes to retinoic acids³⁹. While not participating in GPCR signaling per se, retinoic acids nonetheless present as a crucial retinoid that also functions as a ligand for signaling receptors.

While not described in detail here, we would like to acknowledge the previously established methods for retinoid detection using HPLC across various contexts, such as in food research and the study of microbial rhodopsin. These methods employ different goals and approaches to retinoid detection, including the use of reverse-phase techniques that require less volatile and hazardous mobile phases^40,41,42, the detection of retinoic acids and their associated isomers^40,41, and purification and extraction from different biological sources⁴³. Our method focuses specifically on the detection of retinyl palmitate, retinaldehyde isomers, and retinol isomers from mammalian tissue. Different protocols should be considered if the intended use case differs from this specific application.

Protocol

NOTE: All animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Minnesota (protocol # 2312-41637A) and performed in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Perform all extractions in the dark, under a dim red light for illumination. Be aware of residual light emitted by instrument displays and accessory LEDs.

1. Spectrophotometric retinoid standard generation and external standard curve generation

NOTE: Prepare a dry ice container for temporary storage of retinoids before analysis with the HPLC.

1. Weigh an arbitrary but appropriate quantity of retinoids and dissolve them in an appropriate solvent for quantification through spectrophotometry.
   NOTE: The solvent of choice used in this study is ethanol, and the molar absorptivity values of retinoids dissolved in ethanol are detailed in Table 1. Anhydrous, HPLC-grade ethanol must be used to dissolve standards and samples. Ethanol purified solely through distillation forms an azeotropic mixture with water, containing approximately 4% water by volume. Water is immiscible with the hexanoic mobile phase and will result in the hydration of the silica stationary phase, leading to eventual degradation of the column.
2. Utilizing the Beer-Lambert Law and the molar absorptivity of the pertinent retinoid, quantify the concentration of the generated standard (Table 1).
   Absorbance = Molar Absorbtivitty (ε) × Molar Concentration × Path Length
3. Perform serial dilutions to create concentrations that can generate a standard curve within the range of quantitation for the desired tissue.
   NOTE: Be aware of the fundamental limitations of the Beer-Lambert Law, such as its lack of validity with high concentrations of analyte. To avoid this issue, the diluted retinoid standards should avoid generating absorbance values greater than 1. For our applications in retinol and retinaldehyde quantification in murine organs, we found that a calibration curve with a range of 1-10 ng was able to cover the typical quantities found. For retinyl palmitate quantification from murine liver, we found that a calibration curve with a range of 20-80 µg was able to cover the typical quantities found.
4. By incrementally altering the injection volume from the previously created stock solution, thereby incrementally altering the injected amount of retinoid, integrate the peaks to generate an external standard curve suitable for retinoid quantification where peak integration is directly proportional to the amount of retinoid injected.

2. Tissue harvest and sample collection

NOTE: Prepare a dry ice container for temporary storage of tissue before tissue homogenization and retinoid extraction. Recommended tissue harvest amounts are detailed in Table 2. To account for retinoid variations due to variations in blood content for each tissue, tissue extraction should be done on fully perfused mice, and blood extraction should be completed on separate mice.

1. Euthanize the mice following the guidelines set forth by the Institutional Animal Care and Use Committee (IACUC) dictated protocol (CO₂ asphyxiation here).
2. Blood: Immediately following euthanasia, decapitate the mice with a pair of scissors, and drain the blood from the main trunk of the mice into a 1.5 mL tube.
3. Eye: Using a pair of forceps, remove the eyes from the decapitated head.
4. Brain: Using a pair of small scissors, cut into the decapitated head and remove the brain using a pair of forceps.
5. Kidney, liver, spleen, heart, and lung: Using a pair of small scissors, make an incision into the abdomen, cut in the superior direction along the midline, and cut through the sternum and ribcage. Remove the exposed tissue using a pair of forceps.

3. Tissue homogenization

NOTE: If analysis of smaller partitions of organs is desired, such as in larger organs (e.g., liver or lung tissue), the whole organ should be homogenized to avoid differences in retinoid content in different parts of the tissue. Instead, partition the homogenate if smaller quantities of tissue are desired. A schematic for the protocol is detailed in Figure 2. This modified protocol was adapted from Kane and Napoli⁴⁴.

1. Place the tissue in the tissue grinder tube, along with 50% ice-cold saline (0.9%) and 50% methanol. See Table 2 for the volume used for each tissue type.
2. Place the pestle into the grinder tube, slowly and gently perform five full rotations with the pestle to obtain a homogenate.
3. Transfer samples into 15 mL tubes immediately after homogenization.
4. Add 2 mL of methanol and let sit for 15 min at room temperature.
   NOTE: If analysis of retinaldehyde oxime derivatives is desired, add 1 mL of 0.1 M hydroxylamine hydrochloride in 0.1 M HEPES (pH 6.5) (Figure 1).

4. Retinoid extraction

CAUTION: Hexane is highly flammable, highly volatile, and highly toxic. National Institute for Occupational Safety and Health (NIOSH)-approved respirators, eye protection, butyl gloves, and a fume hood must be used when handling hexane. When evaporating hexane from samples, some form of enhanced air circulation apparatus is recommended to prevent solvent fume buildup, for example, a snorkel suction apparatus.

1. Add 10 mL of hexane to the homogenate and vortex mix the tube horizontally for at least 10 s.
   NOTE: It is critical that the phases mix fully.
2. Centrifuge the homogenate/hexane mixture for 3 min at 1,000 × g to facilitate phase separation.
3. Conduct the extraction 2x to ensure total extraction of retinoids from the homogenate. Repeat steps 4.1 and 4.2.
4. Draw off the hexane layer using a pipette and place the hexane layer into a separate set of glass 15 mL tubes for vacuum evaporation.
   NOTE: For evaporation, use GLASS 15 mL tubes to avoid adhesion of retinoids to the walls of the tube.
5. Using a vacuum centrifuge, completely evaporate the hexane.

5. Resuspension and HPLC analysis

NOTE: Since the HPLC system used in this manuscript was a binary pump system, the four-component mobile phase was premixed into a singular bottle prior to operation.

CAUTION: All four organic solvents used in this method are highly flammable, highly volatile, and highly toxic. 1,4-dioxane is susceptible to explosive peroxide formation upon exposure to oxygen. Keep all vessels containing 1,4-dioxane closed when not in use. National Institute for Occupational Safety and Health (NIOSH)-approved respirators, eye protection, butyl gloves, and a fume hood must be used when handling these solvents. While running these solvents in an HPLC, some form of enhanced air circulation apparatus is recommended to prevent solvent fume buildup, for example, a snorkel suction apparatus.

1. Resuspend the dried 15 mL tube with 100 µL of hexane; vortex well to make sure all retinoids are dissolved.
2. Pipette all 100 µL of hexane into a single glass insert for HPLC analysis.
3. Set up the HPLC run (adapted from Landers and Olson⁴⁵): mobile phase: 85.4% hexane (v/v), 11.2% ethyl acetate (v/v), 2% dioxane (v/v), 1.4% 1-octanol (v/v); column: two 4.6 mm ID x 250 nm, 5 µm columns, connected in series; a multicolumn Thermostat Temperature: 25 °C; injection volume: 100 µL; flow rate: 1 mL/min; run time: 40 min. Use UV spectrum absorbance detection; keep the option checked in to acquire UV spectrum from 200 nm to 400 nm.

6. Peak Identification and Integration

1. Identify peaks using retention time and UV spectra of each retinoid of interest as observed from analysis of retinoid standards (Figure 3, Table 3, and Table 4).
2. Using the chromatographic data system of the chosen HPLC system, integrate the identified peaks. The integration, or area under the curve is directly proportional to the amount of analyte. Reference the external standard curve generated in step 1 to quantify the analyte.
   1. For analysis of chromatograms generated from biological tissue, use manual integration over automatic integration offered by typical chromatographic data systems, since variabilities in parameters such as retention time are often observed in such samples.
   2. Ensure that chromatograms do not exhibit irregularities that might indicate issues during the run, such as noisy baselines or non-Gaussian peaks. These issues indicate contaminants in the HPLC or wear on columns and should be rectified for valid analysis.

Results

Here, we utilized the method described above to detect and quantify retinoids in murine ocular and systemic tissue and generated representative chromatograms. We will additionally give a summary of the typical retinoids that can be detected in these tissues.

At 6 months of age, mice were euthanized through CO₂ asphyxiation. To maintain ocular retinoid content, mice were dark-adapted for 2 days prior to euthanization and extraction. Two eyes, 0.2 g of liver, and 75 μL of blood were harvested for retinoid extraction and subsequent HPLC analysis. For ocular tissue, one eye was extracted without the addition of hydroxylamine, while the other was subjected to hydroxylamine treatment. For liver and blood, hydroxylamine treatment was not needed since retinaldehydes are not typically detected in these tissues.

In the chromatogram of the mouse eye not treated with hydroxylamine, 13-cis-retinal, one of the two non-canonical retinaldehyde isomers, was identified. Additionally, both 11-cis-retinal and all-trans-retinal, the two canonical retinaldehyde isomers, were identified. In both chromatograms, the visual cycle intermediate 11-cis-retinol, as well as the main vitamin A transport form all-trans-retinol, were both identified (Figure 4A).

In the chromatogram of the mouse eye treated with hydroxylamine, all previously mentioned retinaldehyde isomers were still present. However, the retention times for these isomers were significantly increased. Moreover, these retinaldehyde isomers now present as both syn and anti isomers. During quantification through peak integration, both the syn and anti peaks must be summed to gain an accurate integration value. Both 11-cis-retinol and all-trans-retinol were still present (Figure 4B).

In the chromatogram of mouse hepatic tissue, retinyl palmitate and all-trans-retinol were identified within these tissues. Retinyl palmitate, a retinyl ester, serves as the main storage form of vitamin A in mammals and can be found in significant quantities within mammalian hepatic tissue. The liver releases vitamin A into systemic tissues as holo-RBP4, containing all-trans-retinol within the RBP4-transthryetin complex. Subsequently, a large all-trans-retinol peak can be identified within this chromatogram (Figure 4C).

In the chromatogram of mice blood, a large all-trans-retinol peak was identified. Given that holo-RBP4 from the liver is released to systemic tissues through the circulatory system, this is as expected (Figure 4D). The characteristic UV spectrum absorbance (~250-400 nm) of the retinoid isoforms on the time-dependent resolved HPLC peaks can be helpful in initial cross-confirming the quality of the isoform presence in a particular respective peak (Figure 5).

Figure 1: Chemical structures of retinoids. (A) These retinoids are typically found in extracted murine tissue. Note the differing polar end groups, as well as the differing locations of cis double bonds in retinoid isomers. (B) Conversion of retinaldehydes to retinal oximes increases extraction efficiency, prevents co-elution of peaks with retinyl palmitate, as well as co-elution within retinaldehyde isomers. Please click here to view a larger version of this figure.

Figure 2: Workflow for tissue harvest, homogenization, retinoid extraction and analysis. Retinas and tissue samples from dark-adapted mice were homogenized and extracted. Step 1: After euthanizing mice according to guidelines, blood and various tissues are collected. Steps 2-4: Tissue homogenization is then done with saline and methanol, with hydroxylamine hydrochloride for quantitative analyses. Steps 5-7: Retinoids are treated using hexane, followed by evaporation. Steps 8,9: Resuspended with 100 µL of hexane to make sure all retinoids are dissolved. Step 10: Finally, samples are analyzed using HPLC. Abbreviation: HPLC = High-performance Liquid Chromatography. Please click here to view a larger version of this figure.

Figure 3: Chromatogram of retinoid standards and hydroxylamine-treated retinoid standards. (A) This chromatogram contains retinyl palmitate, 13-cis-retinal, 11-cis-retinal, all-trans-retinal, 11-cis-retinol, 13-cis-retinol, and all-trans-retinol standards. (B) This chromatogram contains retinyl palmitate, 13-cis-retinal, 11-cis-retinal, all-trans-retinal, 11-cis-retinol, 13-cis-retinol, and all-trans-retinol standards. This chromatogram contains the syn and anti oxime retinaldehyde isomers, in addition to retinol isomers. Please click here to view a larger version of this figure.

Figure 4: Representative Chromatograms. (A) Representative chromatogram of mouse ocular tissue.This chromatogram contains retinoids typically found in murine ocular tissue, including retinaldehyde isomers, in addition to retinol isomers. (B) Representative chromatogram of hydroxylamine-treated mouse ocular tissue. Note that in lieu of the non-chemically modified retinaldehydes found in the previous chromatograms (Figure 4A), this chromatogram contains the syn and anti oxime retinaldeyde isomers, in addition to retinol isomers. (C) Representative chromatogram of mouse liver tissue.This chromatogram contains retinoids typically found in murine hepatic tissue, including retinyl palmitate and all-trans-retinol. (D) Representative chromatogram of mouse blood. This chromatogram contains retinoids typically found in murine blood, including all-trans-retinol. Please click here to view a larger version of this figure.

Figure 5: UV absorbance spectrum of retinaldehyde isomers from murine ocular tissue. Absorbance spectra of (A) 13-cis-retinal, (B) 11-cis-retinal, (C) 9-cis-retinal, (D) all-trans-retinal, (E) syn 11-cis-retinal oxime, (F) anti 11-cis-retinal oxime. Please click here to view a larger version of this figure.

Figure 6: Overview of HPLC modules, mobile phase flow, and other instrumentation. Mobile phase flow begins from the prepared solvent bottles and is pumped at high pressure by the binary pump into the autosampler, where the analyte is incorporated into the solvent flow. The mobile phase and analyte reach the temperature-regulated column compartment, and finally, into the diode array detector for measurement. (A) Red tinted monitor screen minimizing photoisomerization of retinoids. (B) Binary pump. (C) Vial sampler. (D) Multicolumn thermostat. (E) Diode array detector. (F) Analytical fraction collector. (G) High-density poly waste container. (H) From left to right, aluminum crimp cap, 250 μL insert with polymer feet, amber tinted crimp top vial, final assembly of injection vial. (I) Left, refrigerated centrifuge. Right, centrifuge concentrator. Please click here to view a larger version of this figure.

Compound	Molar Absorptivity (ε)	Solvent	Reference
Retinyl Palmitate	49260	Ethanol	Hubbard et al.49
13-cis-Retinal	35500	Ethanol	Hubbard et al.49
11-cis-Retinal	24935	Ethanol	Hubbard et al.49
9-cis-Retinal	36100	Ethanol	Robeson et al.51
All-trans-Retinal	42880	Ethanol	Hubbard et al.49
11-cis-Retinol	34890	Ethanol	Hubbard et al.49
13-cis-Retinol	48305	Ethanol	Robeson et al.50 and Robeson et al.51
All-trans-Retinol	52770	Ethanol	Hubbard et al.49

Table 1: Molar absorptivity of retinoids in ethanol. Molar absorptivity values of retinoids were gathered from Kane and Napoli⁴⁴, Hubbard et al.⁴⁶,Robeson et al.⁴⁷, and Robeson et al.⁴⁸. The molar absorptivity depends on the solvent. The values listed here are specific to retinoids dissolved in ethanol.

Tissue	Tissue Harvest Amount	Volume for Homogenization
Blood	75 µL	N/A
Liver	0.2 g	2 mL
Kidney	2 Whole Kidneys	2 mL
Skin	0.2 g	1 mL – 1.5 mL
Spleen	1 Whole Spleen	1 mL
Eye	Both Retina, pool 4 Retina	1 mL
Brain	1 Whole Brain	2 mL
Heart	1 Whole Heart	1 mL
Lung	0.2 g	2 mL

Table 2: Recommended tissue quantity per analysis. Tissue quantity per analysis and volume of 50% ice-cold saline (0.9%) and 50% methanol for tissue homogenization, as recommended by Kane and Napoli⁴⁴.

Compound	Retention Time (min)	UV Absorbance Maxima (nm)
Retinyl Palmitate	5.177	326
13-cis-Retinal	7.268	366
11-cis-Retinal	7.691	366
9-cis-Retinal	7.931	364
All-trans-Retinal	8.993	370
11-cis-Retinol	17.003	318
13-cis-Retinol	17.83	328
All-trans-Retinol	23.933	326

Table 3: Retention times and maximum absorbance wavelengths of retinoids. These values were observed when used with the described method and mobile phase, as detected with the referenced diode array detector.

Compound	Retention Time (min)	UV Absorbance Maxima (nm)
Syn 11-cis-Retinal Oxime	9.59	346
Syn All-trans-Retinal Oxime	10.6	356
Syn 13-cis-Retinal Oxime	11.3	352
Anti 13-cis-Retinal Oxime	13	356
Anti 11-cis-Retinal Oxime	15.1	350
Anti All-trans-Retinal Oxime	19.7	360

Table 4: Retention times and maximum absorbance wavelengths of hydroxylamine-treated retinaldehydes, resulting in syn and anti oxime isomers of each retinaldehyde. These values were observed when used with the described method and mobile phase, as detected with the referenced diode array detector.

Discussion

In this method, normal-phase HPLC is used to detect and quantify relevant retinoids, including retinyl esters, retinaldehydes, and retinols. Given the importance of 11-cis-retinal as the critical chromophore in the activation of the rhodopsin GPCR, a method that can detect the metabolites that is related to the production of 11-cis-retinal is critical to the study of overall visual function. The main advantage of this method is that all relevant isomers of both retinaldehydes and retinols can be simultaneously detected and quantified with a single run. Given how scarce viable experimental tissues are, this method allows for efficient use of tissue in retinoid detection. While the use of hexane and ethyl acetate-based mobile phases has been reported in the separation of retinoids⁴⁹, the addition of 1,4-dioxane and 1-octanol improves the resolution of retinoid isomers. 1,4-dioxane is necessary for the resolution of 13-cis-retinol and 11-cis-retinol⁵⁰, while long-chain alcohols such as 1-octanol further improve the separation of retinol isomers^45,51,52.

For quantification of retinaldehyde isomers, prior conversion to the corresponding retinaldehyde oximes with hydroxylamine is required, since the reactive aldehyde group is prone to react with other biomolecules such as proteins and lipids^45,53,54. Retinaldehyde oximes are more readily extracted and cause retinaldehyde peaks to elute later in the run, thus preventing peak co-elution with the unknown group of peaks and the retinyl palmitate peak that elutes at ~5-8 min for biological tissues. Treatment of retinaldehydes with hydroxylamine also prevents co-elution of retinaldehydes with similar elution times⁴⁵. The conversion to retinaldehyde oximes generates both syn and anti isomers, which elute at two different times with two different peaks for each retinaldehyde isomer. The integration for both peaks must be summed to properly quantify each retinaldehyde isomer. Besides retention time, the UV spectra provide another metric that can be utilized in peak identification. Each retinoid exhibits a slightly different UV absorbance maximum. When taken into consideration along with retention time, these two metrics allow for the accurate identification of peaks (Table 3 and Table 4).

While the presence of canonical retinaldehyde isomers such as 11-cis-retinal and all-trans-retinal are expected in ocular tissue, with 11-cis-retinal being the critical rhodopsin ligand chromophore and all-trans-retinal being the photoisomerized ligand, other non-canonical retinaldehyde isomers are detected with this method. In particular, 13-cis-retinal was readily detected in HPLC runs of murine ocular tissue during our analysis (Figure 4A-D). The appearance of these non-canonical retinaldehydes is not completely unexpected however, both 13-cis-retinal and 9-cis-retinal are known byproducts resulting from light-induced stress upon the retina^55,56. These retinaldehyde isomers do not participate in the known canonical phototransduction cascade pathway and their metabolic fates remain relatively unknown. However, 9-cis-retinal has been shown to also stably bind to the binding pocket of rhodopsin, forming isorhodopsin. The role of isorhodopsin has yet to be fully characterized, but studies have shown that isorhodopsin does exhibit photosensitive capabilities much like the 11-cis-retinal based rhodopsin^57,58. At the very least, the capability of 9-cis-retinal to stably form isorhodopsin and photoisomerize into all-trans-retinal provides a reasonable explanation for its absence in our HPLC analysis⁵⁶. 13-cis-retinal, on the other hand, has not been shown to participate in any known ocular biochemical pathway, and its accumulation within ocular tissue might result in oxidative stress within the retina^55,59,60.

Although 9-cis-retinal is typically not found in murine tissue, it should be mentioned that the hydroxylamine-treated method is not a viable method for the separation and quantification of 9-cis-retinal. The peaks for syn 9-cis-retinal and syn 13-cis-retinal will coelute at 11.2 min. In cases where significant amounts of 9-cis-retinal are expected, such as when it is applied exogenously, only quantification of combined 9-cis-retinal and 13-cis-retinal will be possible. For qualitative identification, performing the method without treating the tissue with hydroxylamine with allow for the complete separation of 13-cis-retinal and 9-cis-retinal peaks.

Besides ocular tissue, this method can also be readily applied to systemic tissue. While retinaldehydes in their various isomeric forms are typically not found in systemic tissue, all-trans-retinol bound to retinol-binding protein 4 (holo-RBP4) is the main transport form of vitamin A in mammalian organisms and all-trans-retinol can be readily detected in the blood using this method. Besides the blood, all-trans-retinol can also typically be detected in all systemic tissues. Hence, using this method, a complete systemic tissue-wide retinoid profile can be created using this method.

Since retinaldehydes, especially the critical rhodopsin ligand 11-cis-retinal, are extremely susceptible to photoisomerization, it is critical that preventative measures are taken to prevent accidental exposure to light during any of the described steps above. Besides the obvious sources from room lighting and instrument panels, we have found that errant light from accessory LEDs in various instruments, such as those found behind instruments, including the HPLC itself, to be major sources of accidental light exposure (Figure 6). Additionally, we have found that a dedicated dark room was vital to the proper handling of retinaldehyde isomers and experimental tissues. Besides light minimization, we recommend that tissue is swiftly and promptly processed following euthanization and harvest since degradation of retinaldehyde isomers occurs even in the absence of light. Ideally, tissue harvest, extraction, and HPLC analysis should all be completed on the same day.

Tissue that is not processed immediately must be stored in the dark at -80 °C and should not be stored for more than a week before processing. As was mentioned above, all solvents used in the mobile phase and in the extraction steps in the above method are highly toxic, volatile, and flammable. Proper procedures and equipment must be used when handling these hazardous solvents. Proper personal protective equipment (PPE) such as butyl gloves, NIOSH-approved respirators, and eye protection must be donned prior to solvent handling, and all solvent transfers must be done inside a fume hood. Be aware that 1,4-dioxane is susceptible to explosive peroxide formation upon exposure to oxygen and keep all solvent containers with 1,4-dioxane tightly sealed from air.

If a normal-phase HPLC method is unviable due to inadequate materials or equipment, more typical reverse-phase methods with less volatile solvents such as acetonitrile and water do exist to detect and quantify retinaldehydes and retinols. However, these methods are not able to separate isomers and are only able to detect total retinaldehydes or total retinols^44,54.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Reagent
1-Octanol, suitable for HPLC, ≥99.5%	Sigma-Aldrich, Millipore Sigma	203-917-6
1,4-Dioxane, suitable for HPLC, ≥99.5%	Sigma-Aldrich, Millipore Sigma	204-661-8
11-cis-retinal	National Eye Institute	N/A
11-cis-Retinol	Toronto Research Chemicals	TRC-R252105
13-cis-retinal	Toronto Research Chemicals	TRC-R239900
13-cis-retinol	Toronto Research Chemicals	TRC-R252110
All-trans-Retinal	Toronto Research Chemicals	TRC-R240000
All-trans-Retinol	Toronto Research Chemicals	TRC-R252002
Ethyl Acetate, suitable for HPLC, ≥99.7%	Sigma-Aldrich, Millipore Sigma	205-500-4
Hexane, HPLC Grade	Fisher Scientific, Spectrum Chemical	18-610-808
Methanol (HPLC)	Fisher Scienctific	A452SK-4
Retinyl Palmitate	Toronto Research Chemicals	TRC-R275450
Sodium Chloride (Crystalline/Certified ACS)	Fisher Scientific	S271-500
Instruments
1260 Infinity II Analytical Fraction Collector	Agilent	G1364F
1260 Infinity II Binary Pump	Agilent	G7112B
1260 Infinity II Diode Array Detector	Agilent	G7115A
1260 Infinity II Multicolumn Thermostat	Agilent	G7116A
1260 Infinity II Vialsampler	Agilent	G7129A
ST40R Refrigerated Centrifuge	Thermo Scientific	TSST40R
Vacufuge Plus Centrifuge Concentrator	Eppendorf	22820168
Consumables
2 mL Amber Screw Top Vials	Agilent	5188-6535
Crimp Cap with PTFE/red rubber septa, 11 mm	Agilent	5183-4498
Disposable Glass Conical Centrifuge Tubes	Millipore Sigma	CLS9950215
Screw cap tube, 15 mL	Sarstedt	62.554.502
Vial insert, 150 µL, glass with polymer feet	Agilent	5183-2088