Diurnal, localized exposure of phosphatidylserine by rod outer segment tips in wild-type but not Itgb5-/- or Mfge8-/- mouse retina

Abstract

In the mammalian retina, life-long renewal of light-sensitive photoreceptor outer segments (POS) involves circadian shedding of distal rod POS tips and their subsequent phagocytosis by the adjacent retinal pigment epithelium (RPE) every morning after light onset. Molecular mechanisms that promote or synchronize POS tip shedding have thus far remained unknown. Here we examined plasma membrane asymmetry of living POS by quantifying surface exposure of the membrane phospholipid phosphatidylserine (PS) using antibodies, annexin V, and pSIVA (polarity-sensitive indicator of viability and apoptosis), an annexin-based biosensor with switchable states of fluorescence. We found that isolated POS particles possess externalized PS, whose blockade or removal reduces their binding and engulfment by RPE in culture. Imaging of live photoreceptors in freshly dissected mouse retina detected PS externalization restricted to POS tips with discrete boundaries. In wild-type mice, frequency of rod tips exposing PS and length of tips with exposed PS peak shortly after light onset. In contrast, PS-marked POS tips do not vary in mice lacking the diurnal phagocytic rhythm of the RPE due to loss of either the phagocytosis receptor αvβ5 integrin, expressed by the RPE but not by photoreceptors, or its extracellular ligand milk fat globule-EGF factor 8 (MFG-E8). These data identify a molecular distinction, localized PS exposure, that is specific to the surface of rod POS tips. Enhanced PS exposure preceding rod shedding and phagocytosis suggests that surface PS promotes these processes. Moreover, our results demonstrate that the diurnal rhythm of PS demarcation of POS tips is not intrinsic to rod photoreceptors but requires activities of the RPE as well.

Fig. 1.

Competitive binding and inhibition of phagocytosis of purified POS fragments by αPS and A5. (A) Detection of opsin POS load (Upper, POS) and bound mouse IgG light chains (Upper, IgG, arrowhead) and bound A5 (Lower, A5) in lysates of POS particles after incubation with nonimmune IgG (lanes 1), αPS (lanes 2), or rhodopsin antibody (lanes 3) in the presence of recombinant A5 (+ A5) or β-galactosidase (+ β-gal), which served as negative control. One representative blot is shown of four independent experiments performed with similar results. (B and C) POS particle binding (B) and internalization (C) by RPE cells of particles preincubated with αPS or A5 (black bars, αPS, A5), rhodopsin antibody (gray bar, αPS), or β-galactosidase (gray bar, A5). Bars show relative values compared with levels of bound or internalized POS particles preincubated with nonimmune IgG (set as 100%), displayed as mean ± SD of three independent experiments, each with duplicate samples. *P < 0.05 of αPS and A5 samples compared with the respective control by Student t test.

Fig. 2.

Increased binding of A5 to WT but not Itgb5^−/− retina at light onset. (A) Representative immunoblot membrane sequentially probed for proteins as indicated comparing FITC signal detecting of FITC-A5 in working solution used for labeling (load) and in individual neural retinas excised from eyes from two different WT or Itgb5^−/− mice at light onset (0 h) or 7 h later (+ 7 h) after incubation with FITC-A5 (lanes +). Lanes - show neural retina incubated with buffer without FITC-A5. (B) Quantification of blots as in A. Bars show relative binding of FITC-A5 to WT and Itgb5^−/− retina at time points as indicated, with binding at 0 h to WT retina set as 1 (mean ± SD of six retinas of six mice tested in three independent experiments. *P < 0.05 relative to value for WT at 0 h by Student t test. For each sample, FITC-A5 values were normalized to rod opsin to account for differences in tissue yield.

Fig. 3.

Diurnal variation in POS binding of 488-A5 in WT but not in Itgb5^−/− retina. (A–C, E–G) Representative whole mounts of retina from WT and Itgb5^−/− mice harvested at times of day as indicated, labeled live with 488-A5 followed by washing and fixation before imaging. Maximal projections are shown. (Scale bar, 25 μm.) (D and H) Quantification of labeling intensity per area for WT retina (D) and Itgb5^−/− retina (H) relative to labeling of WT retina at 0 h, which was set as 1. Bars show mean ± SEM, n = 6 retinas of six mice. *P < 0.05 relative to labeling in WT at 0 h (Student t test).

Fig. 4.

Imaging of PS exposed at POS tips in live, dissected mouse retina. Images show maximal projections of WT mouse retina harvested 15 min after light onset and imaged immediately while incubating in pSIVA. (B–D) Field shows costain of pSIVA (B), CellMask membrane stain (C), and overlay of both (D). (Scale bar in A, 10 μm; in B for B–D, 5 μm.)

Fig. 5.

Elongation of PS-exposing tips in WT but not Itgb5^−/− or Mfge8^−/− retina after light onset. Retinas were harvested at time points and from mice of different genotypes as indicated and imaged immediately during incubation with pSIVA. (A) Fields show maximal projections of two representative tips from the same retina for each sample type as indicated. (Scale bar, 1 μm.) (B) Quantification of PS-positive POS tip length. Bars show mean ± SD, n = 4–5 independent experiments each analyzing both eyes of the same mouse. *P < 0.05 relative to labeling in WT at 0 h (Student t test).