Hesperetin prevents selenite-induced cataract in rats.

PURPOSE: This study investigated the ability of hesperetin, a natural flavonoid, to prevent selenite-induced cataracts in a rat model. METHODS: Animals were divided into four treatment groups: G1 (control group), G2 (hesperetin-treated group), G3 (selenite-induced cataract group), and G4 (hesperetin-treated selenite cataract group). Animals in the G1 and G3 groups were injected with vehicle alone, while those in the G2 and G4 groups received a subcutaneous injection of hesperetin (0.4 µg/g bodyweight on days 0, 1, and 2, corresponding to P13, P14, and P15). Sodium selenite (20 µmol/g bodyweight given 4 h after the hesperetin injection on day 0) was administered to rats in the G3 and G4 groups to induce cataract formation. Lenses were observed with slit-lamp microscopy, and filensin degradation and the decreased glutathione (GSH) and ascorbic acid levels in the lens were measured on day 6. RESULTS: Lenses in the G3 group showed mature central opacity, while some lenses in the G4 group lacked central opacity and had lower-grade cataracts. All lenses in the G1 and G2 groups were transparent. Expression of the 94 kDa and 50 kDa forms of filensin was significantly decreased in the lenses in the G3 group compared with those in the G1 and G2 groups. Interestingly, these forms of filensin rescued the rat lenses in the G4 group. In the G3 group lenses, the GSH and ascorbic acid levels were lower than in the control group but were normalized in the G4 group lenses. CONCLUSIONS: The results suggest that hesperetin can prevent selenite-induced cataract formation.

Quantitative analysis of ascorbic acid permeability of aquaporin 0 in the lens.

Aquaporin 0 (AQP0) is a lens-specific protein comprising more than 30% of lens membrane protein content and is a member of the aquaporin family. Water permeates through AQP0 much more slowly than other aquaporin family members, and other compounds, such as glycerol, also permeate AQP0. In the lens, ascorbic acid (AA) is found at high concentrations, protecting the lens from photochemical events such as photo-oxidation. The aim of the present study was to clarify the function of AQP0. Mouse fibroblast L-cells stably expressing AQP0 were established and incubated in medium containing AA, and intracellular AA levels were measured by high-performance liquid chromatography (HPLC) and 2,6-dichlorophenol-indophenol (DCPIP) analysis. Intracellular AA levels in AQP0-expressing cells quickly rose and reached saturation 10 min after incubation in medium containing 1000 µM AA. In contrast, AA levels in cells slowly decreased when AA was washed out from the medium. Cells overexpressing AQP0 increased the cellular uptake of AA in a time- and concentration-dependent manner. These data suggest that AA as well as water permeates AQP0. AQP0 expression on Xenopus oocyte membranes was achieved by the injection of AQP0 cRNA into oocytes that were incubated in medium containing AA. Intracellular AA levels were then measured by HPLC. AA uptake was demonstrated in the AQP0-expressing oocytes and was shown to quickly reach saturation. Intracellular AA concentration in oocytes increased in a time- and concentration-dependent manner. The data in the present study show that AA permeates AQP0, reveal the role of AQP0 in AA permeability ex vivo, and also indicate that there is a difference between the import and export of AA via AQP0. These findings suggest that AQP0 plays an important role in controlling lens AA content.

The effect of the interaction between aquaporin 0 (AQP0) and the filensin tail region on AQP0 water permeability.

PURPOSE: To study the interaction between the lens-specific water channel protein, aquaporin 0 (AQP0) and the lens-specific intermediate filament protein, filensin, and the effect of this interaction on the water permeability of AQP0. The effect of other factors on the interaction was also investigated. METHODS: Expression plasmids were constructed in which glutathione-S-transferase (GST) was fused to the AQP0 COOH-terminal region (GST-AQP0-C), which contains the major phosphorylation sites of the protein. Plasmids for AQP0 COOH-terminal mutants were also constructed in which one, three or five sites were pseudophosphorylated, and the proteins expressed from these GST-fusion plasmids were assayed for their interaction with lens proteins. Expressed recombinant GST-fusion proteins were purified using glutathione beads and incubated with rat lens extract. Western blotting was used to identify the lens proteins that interacted with the GST-fusion proteins. Filensin tail and rod domains were also expressed as GST-fusion proteins and their interactions with AQPO were analyzed. Additionally, the water permeability of AQP0 was calculated by expressing AQP0 with or without the filensin peptide on the cell membrane of Xenopus oocytes by injecting cRNAs for AQP0 and filensin. RESULTS: The GST-AQP0-C construct interacted with the tail region of lens filensin and the GST-filensin-tail construct interacted with lens AQP0, but the GST-filensin-rod construct did not interact with AQP0. GST-AQP0-C also interacted with a purified recombinant filensin-tail peptide after cleavage from GST. The AQP0/filensin-tail interaction was not affected by pseudophosphorylation of the AQP0 COOH-terminal tail, nor was it affected by changes in pH. Xenopus oocytes expressing AQP0 on the plasma membrane showed increased water permeability, which was lowered when the filensin COOH-terminal peptide cRNA was coinjected with the cRNA for AQP0. CONCLUSIONS: The filensin COOH-terminal tail region interacted with the AQP0 COOH-terminal region and the results strongly suggested that the interaction was direct. It appears that interactions between AQP0 and filensin helps to regulate the water permeability of AQP0 and to organize the structure of lens fiber cells, and may also help to maintain the transparency of the lens.

Characterization and localization of side population cells in the lens.

PURPOSE: Side population (SP) cells were isolated and the possibility whether lens epithelial cells contain stem cells was investigated. METHODS: Mouse lens epithelial cells were stained by Hoechst 33342 and then sorted by fluorescence-activated cell sorting (FACS). The expression of stem cell markers in sorted SP cells and the main population of epithelial cells were analyzed by quantitative real-time PCR. Localization of SP cells in the mouse lens was studied by fluorescence microscopy. RESULTS: The sorted cells contained SP cells, which were no longer separable by FACS following treatment with verapamil. The number of SP cells decreased with aging, but the adult mouse lens still contained SP cells. Phase contrast microscopy revealed that SP cells were smaller in size than non-SP cells. SP cells were localized near the equator region in lens epithelial cell layers. SP cells expressed higher levels of the stem cell markers ATP-binding cassette transporter G2 (ABCG2), p75 neurotrophin receptor (p75NTR), nestin (nes), B-cell lymphoma 2 (Bcl2), and cell surface antigen Sca-1 mRNA than the main population cells. These results suggest that SP cells contain a high proportion of stem cells. CONCLUSIONS: The SP cells in the lens contain stem cells, and these stem cells are localized around the germinative zone.

The function of filensin and phakinin in lens transparency.

PURPOSE: Beaded filaments are lens cell-specific intermediate filaments composed of two proteins: filensin and phakinin (CP49). Filensin and phakinin are believed to function in the maintenance of lens transparency. To elucidate the function of filensin and phakinin at the molecular level, we examined the degradation of these two proteins in normal and cataractous rat lenses. METHODS: A hereditary cataract model, the Shumiya cataract rat (SCR), was used for these studies. Anti-filensin antibodies were raised against three different regions of the protein, the rod domain, the inner region of the tail domain, and the outer region of the tail domain. Anti-filensin and anti-phakinin antibodies were used to examine the conformation of degradation of filensin and phakinin by western blot analysis and fluorescent immunocytochemistry of cryosectioned lenses. RESULTS: In the normal lens, filensin was processed from a 94 kDa protein to proteins of 50 kDa and 38 kDa. Similarly, phakinin was processed from a 49 kDa protein to one of 40 kDa. The concentrations of filensin and phakinin in the rat lens cortex fluctuated with age and decreased during cataractogenesis. The 50 kDa form of filensin decreased significantly before opacification. In the normal lens, phakinin, the filensin rod domain, and the filensin inner tail domain localized to membrane lining regions in the shallow cortex and to the central region of the cytoplasm in the deep cortex. The COOH-terminal domain of filensin localized to the membrane lining region in the deep cortex. In pre-cataractous lenses, phakinin and the filensin rod domain localized primarily to the membranes lining the shallow cortex region and were distributed throughout the cytoplasm of lens fiber cells in the deep cortex. CONCLUSIONS: The 50 kDa form of filensin is important for the localization of beaded filaments in lens fiber cells and for lens transparency.

NADH photo-oxidation is enhanced by a partially purified lambda-crystallin fraction from rabbit lens.

PURPOSE: In the rabbit lens, high levels of reduced nicotinamide adenine dinucleotide (NADH) can function as a near-ultraviolet light (near-UV) filter, an effect apparently achieved by specific nucleotide binding to lambda-crystallin. The present investigation asks whether lambda-crystallin enhances NADH photo-oxidation by superoxide radicals produced via a photosensitization reaction of near-UV with NADH. METHODS: Lambda-crystallin was partially purified from rabbit lens soluble fraction by a two-step gel filtration and affinity column chromatography procedure. NADH solutions with or without partially purified lambda-crystallin were subjected to near-UV irradiation or exposed to superoxide generated enzymatically by the xanthine/xanthine oxidase system. NADH oxidation was determined by assaying the decrease of absorbance at 340 nm. RESULTS: When irradiated with near-UV, free NADH was oxidized very little in the absence of lambda-crystallin. In contrast, NADH photo-oxidation was rapidly initiated in the presence of partially purified lambda-crystallin. This lambda-crystallin-enhanced NADH photo-oxidation was totally inhibited by adding superoxide dismutase. We also found that lambda-crystallin largely increased NADH oxidation by a superoxide that is generated enzymatically. These results indicate that NADH bound to lambda-crystallin rapidly reacts with superoxides. The reactivity of bound NADH with superoxide was almost equivalent to that of ascorbic acid. However, lambda-crystallin-enhanced NADH oxidation exceeded the superoxide levels generated by NADH photosensitization and xanthine/xanthine oxidase. CONCLUSIONS: We conclude that NADH binding to lambda-crystallin enhances NADH photo-oxidation through a radical chain reaction mechanism that is initiated by superoxides generated by NADH photosensitization and propagated by another superoxide from the molecule oxygen. High concentrations of NADH bound to lambda-crystallin may be beneficial to the rabbit lens in scavenging the low amounts of superoxide produced by near-UV absorption, since oxygen tension in the lens is very low. We also discuss the function of near-UV-filtering and the anti-photo-oxidation systems in other vertebrate lenses.

Model for Studying Anti- Allergic Drugs for Allergic Conjunctivitis in Animals.

Allergic conjunctivitis (AC), which is characterized by ocular itching, hyperemia, and edema, deteriorates quality of life. In this study, effects of anti-allergic drugs were evaluated by assessing eye-scratching behavior, the number of eosinophils in conjunctiva epithelial tissues, and concentrations of chemical mediators in the tears of the guinea pig model of ovalbumin (OA)-induced AC. METHODOLOGY: On day 0, 3-week-old guinea pigs were sensitized by OA subconjunctival injections. On days 15, 17, and 19, OA solution was administered. Anti-allergic eye drops were administered 5 and 15 min before the final OA challenge on day 19. Scratching behavior within 1 h after OA exposure was studied. Eosinophils in the conjunctiva were stained with Giemsa reagent. Histamine and substance P (SP) concentrations in tears were measured using ELISA. RESULTS: Subconjunctivally injected guinea pigs were observed for clinical symptoms. Scratching responses significantly reduced with ketotifen or olopatadine treatment. Eosinophil numbers reduced in animals treated with ketotifen, levocabastine, or tranilast. Histamine and/or SP concentrations in tears were inhibited by some of these anti-allergic drugs. CONCLUSIONS: It is important to assess the anti-allergic AC drugs objectively because there are several of these drugs currently available. This model allows for an objective evaluation of anti-allergic drugs for AC.

The Extracellular C-loop Domain Plays an Important Role in the Cell Adhesion Function of Aquaporin 0.

PURPOSE: Although aquaporin 0 (AQP0) is a member of the AQP family, it has limited water permeability compared with other members. AQP0 may also have cell adhesion-related functions, but the evidence is still limited. Here, we studied the relationship of AQP0 to cell adhesion and determined the region required for cell adhesion. METHODS: L-cell fibroblasts stably expressing AQP0 or AQP1 (L-AQP0 or L-AQP1) were established. One group of cells was stained with CellTracker Red and cultured into a confluent monolayer, whereas the other group was loaded with CellTracker Blue and seeded over the monolayer. To study cell adhesion, the percentages of lower and upper layer cells were measured using flow cytometry. To determine the region of AQP0 required for adhesion, activity was done by pull-down assay using glutathione S-transferase fusion proteins. To study the water permeability, Xenopus laevis oocyte expressing AQP0 wild-type or AQP0 mutated in C-loop was transferred to a hypotonic solution and photographed, and the diameter was measured to calculate the volume. RESULTS: More cells adhered to the lower cells in the L-AQP0 homotypic pair than other pairs such as L-AQP1 homotypic or L-AQP0/L-AQP1 heterotypic pairs. Pull-down assays revealed that AQP0 could bind to itself via the C-loop extracellular domain. Furthermore, we determined that 109Pro and 110Pro in the C-loop were important for cell adhesion. However, mutation of the C-loop in AQP0 did not affect its water permeability. CONCLUSIONS: AQP0 is known to bind lipids in the opposing membrane. Our data suggest that this cell-to-cell adhesion occurs not only in the AQP0/liquids but also via AQP0/AQP0 interaction through the C-loop domain. Mutations in the C-loop amino acids did not affect the water permeability of AQP0 but did affect its cell adhesion function. These independent dual functions of AQP0 are important for lens transparency.

NADH binding properties of rabbit lens lambda-crystallin.

PURPOSE: The present investigation aims to evaluate the NADH binding ability of lambda-crystallin, a taxon-specific enzyme-crystallin, in the rabbit lens. METHODS: A lambda/betaL1-crystallin fraction was separated from the rabbit lens soluble fraction by gel filtration and the enzyme-crystallin was partially purified by subsequent affinity column chromatography. Analysis of NADH bound to the lambda-crystallin preparation was performed using spectrophotometric and enzymological methods. Binding of added NADH to the enzyme-crystallin preparation was also analyzed using a simple ultrafiltration method, which was theoretically equivalent to equilibrium dialysis, to study additional NADH binding to the protein. RESULTS: The prepared lambda-crystallin samples clearly exhibited an absorption maximum at 340 nm, even though they were thoroughly dialyzed. This was due to the presence of nondialyzable NADH bound tightly to the protein. The bound NADH was removed by charcoal treatment, and extracted by 0.1% SDS or 70 degrees C heat treatment. A dissociation constant (Kd) of less than 5 nM indicated tight binding of NADH. The quantity of bound NADH in the 88% purified 33 kDa enzyme-crystallin was estimated to be 20.5 nmol/mg protein, suggesting a stoichiometry of 0.7 mol of the nucleotide/mol of the 33 kDa protein. Additional looser binding of added NADH to lambda-crystallin was observed in both the lambda/betaL1-crystallin fraction (including the full-length 33 kDa protein: 34%; 25-30 kDa proteins, most of which might be generated by cleavage of the 33 kDa protein: 64%) and the partially purified enzyme-crystallin. It was assumed from the analysis of binding titration that some (about 30%) of the 33 kDa protein and most of the lower molecular weight proteins still possessed the ability to loosely bind NADH. Kd values of their lower affinity binding were determined to be 2 or 6 microM. CONCLUSIONS: From the present study, we conclude that lambda-crystallin plays a sufficiently important role as a NADH binding protein to maintain high levels of this nucleotide in the rabbit lens.

Effect of hesperetin on chaperone activity in selenite-induced cataract.

BACKGROUND: Chaperone activity of α-crystallin in the lens works to prevent protein aggregation and is important to maintain the lens transparency. This study evaluated the effect of hesperetin on lens chaperone activity in selenite-induced cataracts. METHODOLOGY: Thirteen-day-old rats were divided into four groups. Animals were given hesperetin (groups G2 and G4) or vehicle (G1 and G3) on Days 0, 1, and 2. Rats in G3 and G4 were administered selenite subcutaneously 4 hours after the first hesperetin injection. On Days 2, 4, and 6, cataract grades were evaluated using slit-lamp biomicroscopy. The amount of a-crystallin and chaperone activity in water-soluble fraction were measured after animals sacrificed. RESULTS: G3 on day 4 had developed significant cataract, as an average cataract grading of 4.6 ± 0.2. In contrast, G4 had less severe central opacities and lower stage cataracts than G3, as an average cataract grading of 2.4 ± 0.4. The a-crystallin levels in G3 lenses were lower than in G1, but the same as G4. Additionally, chaperone activity was weaker in G3 lenses than G1, but the same as in G4. CONCLUSIONS: Our results suggest that hesperetin can prevent the decreasing lens chaperone activity and a-crystallin water solubility by administered of selenite.

Type XVI collagen is expressed in factor XIIIa+ monocyte-derived dermal dendrocytes and constitutes a potential substrate for factor XIIIa.

We have previously reported that connective tissue cells in the superficial dermis preferentially express alpha1(XVI) collagen rather than those in the lower dermis. Double immunofluorescence labeling using the antibodies for alpha1(XVI) collagen and factor XIIIa (plasma transglutaminase), which is a marker of dermal dendrocytes, demonstrated that both antibodies reacted with the same cells in the superficial dermis of normal skin as well as the lesional skins of dermal dendrocyte-related disorders, dermatofibroma, and psoriasis. Dermal dendrocytes are considered to be established by a culture of peripheral blood monocytes in the presence of granulocyte macrophage-colony stimulating factor and interleukin-4. Reverse transcription--polymerase chain reaction, metabolic labeling, and immunofluorescence studies demonstrated that treatment of CD14+ peripheral blood monocytes with granulocyte macrophage-colony stimulating factor/interleukin-4 over a period of 8 d resulted in the induction of alpha1(XVI) collagen as well as factor XIIIa. The physiologic significance of colocalization of alpha1(XVI) collagen and factor XIIIa in the tissue and their coordinate induction in CD14+ monocyte-derived dendritic cells in vitro was studied. Considerable incorporation of [3H]putrescine by factor XIIIa into recombinant noncollagenous domain (NC) 11 but not into collagenous domain (COL) 1.NC1 domain of the alpha1(XVI) polypeptide was found. Incubation of recombinant NC11 of alpha1(XVI) polypeptide with factor XIIIa in vitro produced a covalent cross-linking complex on sodium dodecylsulfate-polyacrylamide gel electrophoresis. The results indicate that alpha1(XVI) collagen is constitutively expressed by most dermal dendrocytes in the skin and dendritic cells differentiated from peripheral blood monocytes in vitro. Type XVI collagen is expressed in factor XIIIa+ dermal dendrocytes and may form an intermolecular cross-linking through NC11 domain by the reaction catalyzed by factor XIIIa contributing to the structural integrity of factor XIIIa+ dendritic cell-rich tissues.

Heterogeneity of ascorbate free radical reductase in the human lens.

The soluble ascorbate free radical (AFR) reductases in the human lens were separated into many isoforms in the range of pI 5-7 by native isoelectric focusing. In the two-dimensional gel electrophoresis, however, two main proteins with molecular weights of 20-25 kD were commonly identified to each isoform. The observed heterogeneity of the human lens AFR reductase is very similar to those reported for beta- and gamma-crystallins in aged and cataractous human lenses. From these results, it is suggested that some of the isoforms of the lens AFR reductase, especially the more acidic isoforms, may be formed by posttranslational modifications.

Comparison of Lp82- and m-calpain-mediated proteolysis during cataractogenesis in Shumiya cataract rat (SCR).

PURPOSE: It is well known that m-calpain, a ubiquitous calpain, is involved in cataract formation in rodent lens. Involvement of Lp82, a lens-specific calpain, in the cataract formation is also suggested. However, the exact relationship between Lp82-mediated proteolysis and lens opacification has not yet been established. We therefore compared Lp82- and m-calpain-mediated proteolyses of alphaA-crystallin during cataractogenesis to clarify whether Lp82 is involved in cataract formation. METHODS: In order to analyze the Lp82- and m-calpain-mediated proteolyses, we developed antibodies exclusively specific to the proteolytic products of alphaA-crystallin produced by Lp82 and m-calpain actions, respectively. The proteolytic profiles of alphaA-crystallin by Lp82 and m-calpain during cataractogenesis in SCR lenses were analyzed by Western blotting and immunohistochemical staining. RESULTS: While m-calpain-mediated proteolysis was detected predominantly in cataractous lenses, Lp82-mediated proteolysis was detected not only in cataractous but in normal lenses. The m-calpain-mediated proteolysis was observed in restricted areas developing and destined to develop opacification, i.e., the nuclear and perinuclear regions of lens. On the other hand, Lp82-mediated proteolysis was observed not only in the same regions but also in the cortical region where opacity does not develop. Unlike m-calpain-mediated proteolysis, Lp82-mediated proteolysis was not inhibited by the oral administration of aminoguanidine (AG), which acts to prevent lens opacification. CONCLUSIONS: From these results, it is shown that there is no direct contribution of Lp82-mediated proteolysis to cataract formation in SCR. Rather, Lp82 may function in fiber cell development and/or fiber cell remodeling during lens maturation under physiological conditions, since Lp82-mediated proteolysis occurs in the cortical region of normal lens.

Effects of dorzolamide hydrochloride on ocular tissues.

We studied the intraocular pharmacokinetics of dorzolamide hydrochloride eye drops and the effect of dorzolamide on carbonic anhydrase activity and localization in ocular tissues. Carbonic anhydrase activity was detected in normal ocular tissues. The activity was inhibited in corneal endothelial cells, the ciliary body, lens epithelial cells, or the retina 1 to 8 hours after instillation of dorzolamide eye drops. In lens epithelial cells and the retina, the enzyme activity had not recovered even 10 hours after instillation of the drug. Immunostaining did not reveal any differences between the group administered dorzolamide eye drops and the control group administered a physiologically balanced solution. Time-related changes in dorzolamide concentrations in ocular tissues were measured by high-performance liquid chromatography (HPLC). In the cornea, anterior aqueous, iris, ciliary body and retina, drug concentrations increased 15 minutes after the instillation and peaked within 1 hour. These results suggest that dorzolamide immediately suppresses carbonic anhydrase activity in ocular tissues, and is rapidly distributed among the tissues of the eye when administered as eye drops.

lambda-crystallin related to dehydroascorbate reductase in the rabbit lens.

PURPOSE: To evaluate the relationship of lambda-crystallin to reduced nicotinamide adenine dinucleotide (NADH)-dependent dehydroascorbate (DHA) reductase found specifically in the rabbit lens. METHODS: DHA reductase Fractions I-IV were separated from the lambda/betaL1-crystallin fraction of rabbit lens soluble protein by diethylaminoethyl (DEAE)-cellulose ion-exchange column chromatography, and then the enzyme was partially purified from Fraction II by rechromatography on the same ion-exchange column. The isolated DHA reductase fractions were characterized by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), Western blotting, native isoelectric focusing and two-dimensional gel electrophoresis. RESULTS: Using Western blot and a probe of antiserum to recombinant lambda-crystallin, the main 33-kDa protein band was strongly stained in all the rabbit lens DHA reductase fractions, and most of the additional protein bands of approximately 25-30 kDa were also detectable. In the partially purified enzyme, the 33-kDa subunit alone was identified as a distinct protein band by SDS-PAGE, and a main basic protein was found at pI 7.6 by native isoelectric focusing. In addition, many bands of more acidic proteins were separated from other enzyme fractions, and protein spots corresponding to the 33 and/or approximately 25-30-kDa subunits were detected in each of the more acidic proteins by two-dimensional gel electrophoresis. CONCLUSION: These results suggest that lambda-crystallin is closely related to the DHA reductase in the rabbit lens. The above heterogeneity of the enzyme-crystallin may arise from posttranslational modifications.

WITHDRAWN: The role of ascorbic acid transporter in the lens of streptozotocin-induced diabetic rat.

The Publisher regrets that this article is an accidental duplication of an article that has already been published, http://dx.doi.org/10.1016/j.bionut.2010.09.008. The duplicate article has therefore been withdrawn.

Isolation of ascorbate free radical reductase from rabbit lens soluble fraction.

Ascorbate free radical (AFR) reductase with diaphorase activity was isolated from the rabbit lens soluble fraction to characterise some molecular properties of the enzyme. The isolation was accomplished using gel filtration (Sephadex G-75 superfine or Sephacryl S-200 HR), affinity chromatography (Affi-Gel Blue), native isoelectric focusing and two-dimensional gel electrophoresis. A major soluble AFR reductase was found at an isoelectric point of 8.4 and a molecular weight of 31 kDa, and a few minor enzymes were also detected in the range of pI 7.0-8.6. An unknown N-terminal partial amino acid sequence was determined in one peptide fragment prepared from the major enzyme fraction. From the sequence analysis, it is discussed that the lens soluble AFR reductase may differ from NADH-cytochrome b5 reductase reported to be involved in the membrane-bound AFR reductase activity of mitochondria, microsomes and plasma membrane.