Fig 1.
Loss-of function of Tmem98 results in an elongated eye shape.
(A) LacZ stained Tmem98tm1b/+ E16.5 eye cryosections. The expression pattern of Tmem98, indicated by the blue staining for the reporter knockout allele Tmem98tm1b, is found predominantly in the RPE and is also observed in the ganglion cell layer. (B) H&E stained eye sections of wild-type (left), Tmem98tm1b/+ (centre) and Tmem98tm1b/tm1b (right) E16.5 littermate embryos. The eye shape is elongated in the Tmem98 tm1b/tm1b embryo compared to the wild-type and heterozygous embryos. Abbreviations: GCL, ganglion cell layer and RPE, retinal pigment epithelium. Scale bars represent 100 μm (A), 250 μm (B).
Fig 2.
Loss-of-function of Tmem98 in the eye leads to an enlarged eye and retinal defects.
(A) Tmem98tm1c/+ (left) and Tmem98tm1c/tm1d; Tyr-Cre (right) 9 week old female littermates are shown. When Tmem98 is knocked-out by Tyr-Cre the eyes are noticeably enlarged. (B) Slit-lamp pictures of 9 week old littermates. The eyes of Tmem98tm1c/+ (female), Tmem98tm1c/+; Tyr-Cre (male) and Tmem98tm1c/tm1d (female) mice are normal indicating that haploinsufficiency for Tmem98 and expression of Tyr-Cre does not affect eye size. In contrast the Tmem98tm1c/tm1d; Tyr-Cre (female) eye, where Tmem98 expression is lost, is enlarged and bulges out of the head. (C) Comparison of Tmem98tm1c/+ (top) and Tmem98tm1c/+; Tyr-Cre (bottom) enucleated eyes. The eyes were collected from female littermates at three months of age. The posterior segment of the eye is enlarged in the Tmem98tm1c/+; Tyr-Cre eye. Scale bar represents 1mm. (D) Fundus images of Tmem98tm1c/+ (female), Tmem98tm1c/+; Tyr-Cre (male), Tmem98tm1c/tm1d (female) and Tmem98tm1c/tm1d; Tyr-Cre (female) mice. The pictures were taken at 7 weeks of age for the first three and at 10 weeks of age for the fourth. There is extensive retinal degeneration in the Tmem98tm1c/tm1d; Tyr-Cre whilst the retinas of the other genotypes are normal. (E) Scanning laser ophthalmoscope images of control (Tmem98tm1c/+) and mutant (Tmem98tm1c/tm1d; Tyr-Cre) eyes from female littermates at 15 weeks of age. Abbreviations: IR, Infrared; BAF, blue autofluorescence and IRAF, near infrared.
Fig 3.
Characterisation of the phenotype caused by loss-of-function of Tmem98 in the eye.
(A) Staining of RPE flat mounts with an anti-ZO1 antibody (green) and DAPI (blue) reveals the regular hexagonal shape of the cells which contain one or two nuclei in the control Tmem98tm1c/tm1d RPE. In contrast, the cells in the Tmem98tm1c/tm1d; Tyr-Cre RPE vary greatly in size, the number of sides they have and many are multinuclear. The RPEs were collected from male littermates at 7 weeks of age. (B) H&E stained adult eye sections. The Tmem98tm1c/tm1d; Tyr-Cre retina is hugely expanded and very thin compared to the control Tmem98tm1c/tm1d retina. In addition, it appears to have lost structural integrity as indicated by the folding of the retina that occurred during processing which is not seen in the control. All the layers of the retina appear to be present (bottom) but are all extremely thin. In addition the choroid is compressed and the sclera is very thin. The eyes were collected from female littermates at 16 weeks of age. Scale bars represent 50 μm (A), 500 μm (B, top) and 100 μm (B, bottom).
Fig 4.
Loss-of-function of Tmem98 in the eye results in an attenuated ERG response.
Three Tmem98tm1c/tm1d; Tyr-Cre female mice and three control female mice (two Tmem98tm1c/tm1d and one Tmem98tm1c/+; Tyr-Cre) were tested at six months of age (S2 Dataset). (A) ERG traces of Tmem98tm1c/tm1d; Tyr-Cre mice (red lines) and control mice (black lines). Shown are the responses at 3 cd.s/m2 (average of four flashes) for all eyes. (B) Comparison of a-wave amplitudes between the mutant Tmem98tm1c/tm1d; Tyr-Cre mice and control mice. There is a significant difference between mutant and control mice (unpaired t-test, P < 0.0001). (C) Comparison of b-wave amplitudes between the mutant Tmem98tm1c/tm1d; Tyr-Cre mice and control mice. There is a significant difference between mutant and control mice (unpaired t-test, P < 0.0001). **** indicates P < 0.0001.
Fig 5.
Characterisation of the retinal phenotype caused by loss-of-function of Tmem98.
Immunostaining of adult retinal sections from control mice (Tmem98tm1c/tm1d (left) and Tmem98tm1c/+; Tyr-Cre (centre)) and mutant mice (Tmem98tm1c/tm1d; Tyr-Cre (right)). (A) Staining with an anti-GFAP antibody (green) shows that GFAP localisation is normal and seen only in the ganglion cell layer in the control retinas but extends down towards the inner nuclear layer in the mutant retina indicating retinal stress. (B) Staining with anti-TMEM98 (green) and anti-EZRIN (red) antibodies. In the control retinas TMEM98 staining is seen in the apical and basal layers of the RPE and EZRIN is localised to the apical microvilli, co-localisation of the two proteins is indicated by yellow staining. In the mutant retina TMEM98 staining is absent and EZRIN appears to be mislocalised. (C) Staining with anti-RHODOPSIN (green) and anti-OPSIN (red) antibodies marking rods and cones respectively. Normal staining is observed in the control retinas but in the mutant although RHODOPSIN and OPSIN are present the outer segment layer is very thin. In the mutant the retina has folded back completely on itself. This is a processing artefact and reflects the fragility of the mutant retina. DAPI staining of DNA is in blue. Abbreviations: inner nuclear layer (INL), outer nuclear layer (ONL), outer segments (OS) and retinal pigment epithelium (RPE). The retinas shown in (A) and (B) were collected from P21 littermate male mice. The retinas shown in (C) were collected from 9 week old littermate male mice. Scale bars represent 50 μm.
Fig 6.
TMEM98 is a type II transmembrane protein and interacts with MYRF.
(A) Western blot analysis of ARPE-19 subcellular fractions probed with the indicated antibodies. The control antibodies against MEK1/2, AIF and VIMENTIN are found in the expected fractions. TMEM98 is present in the membrane fraction. Uncropped Western blot images are shown in S8 Fig. WCL = whole cell lysate (B) Topology of TMEM98. NIH/3T3 cells were transiently transfected with GFP-TMEM98-V5 (green) and immunostained with anti-GFP or anti-V5 antibodies (red). Non-permeabilised cells are shown on the left and permeabilised cells are shown on the right. On the top row transfected cells stained with an anti-V5 antibody are shown and on the bottom row transfected cells stained with an anti-GFP antibody are shown. Accessibility to the antibodies indicates that the GFP epitope is intracellular and that the V5 epitope is extracellular. A schematic of the expression GFP-TMEM98-V5 construct is shown below. (C) Volcano plot of the mass spectrometry data comparing proteins biotinylated by TMEM98-BirA* with those by BirA* alone. The x axis is log2 fold change of TMEM98-BirA* peptides versus BirA* peptides and the y axis is–log10 p-value of significance. The horizontal dashed line shows where p = 0.05 (−log10(0.05) = 1.3) and the vertical dashed lines show where fold change is four. Proteins with significant difference and change greater than four-fold are shown in red, other proteins are shown in blue. The two most highly enriched proteins, TMEM98 and MYRF, are labelled.
Fig 7.
TMEM98 prevents MYRF self-cleavage and binds to the C-terminal part of MYRF.
(A) Co-immunoprecipitation experiment where HEK293T cells were transiently transfected with the indicated epitope-tagged expression constructs and immunoprecipitated with anti-V5.The two MYRF constructs were either full-length (MYRF) or lacked exon 19 (MYRF Δ19). Western blot analysis of the inputs (left) and immunoprecipitated fractions (right) probed with anti-MYC (Cell Signaling Technology, 2276), anti-FLAG (Cell Signaling Technology, 2368) and anti-V5 antibodies are shown. Anti-tubulin antibody was used to probe the input Western as a loading control. The Western of the input samples shows that MYRF cleaves when transfected alone but remains largely intact when co-transfected with TMEM98-V5. The Western of the immunoprecipitated fractions shows that intact MYC-MYRF-FLAG and the C-terminal part tagged with FLAG are co-immunoprecipitated with TMEM98-V5 indicating that TMEM98 interacts with the C-terminal part of MYRF. Uncropped Western blot images are shown in S9 Fig. (B-D) ARPE-19 cells were transiently transfected with TMEM98-V5 and/or MYC-MYRF-FLAG and immunostained with anti-V5 (magenta), anti-MYC (Cell Signaling Technology, 2278) (red) and anti-FLAG (Biolegend, 637302) (green) antibodies as indicated. DAPI staining is in blue. (C) When transfected alone MYC-MYRF-FLAG cleaves and the N-terminal part tagged with MYC translocates to the nucleus whilst the C-terminal part tagged with FLAG is membrane-bound. (D) When MYC-MYRF-FLAG is co-transfected with TMEM98-V5 it remains intact and colocalises with TMEM98-V5 in the membrane. Scale bars represent 20 μm.
Fig 8.
MYRF is mislocalised in the RPE when Tmem98 is knocked-out in the eye.
Immunostaining of control Tmem98tm1c/tm1d (top) and mutant Tmem98tm1c/tm1d; Tyr-Cre RPE flat mounts with DAPI (blue), anti-MYRF N-terminal (red) and anti-MYRF C-terminal (green) antibodies. MYRF staining is aberrant and when compared to the control more of N-terminal part of MYRF appear to be present in the nucleus in the mutant when compared to the control. The RPEs were collected from P26 male littermate mice. Scale bars represent 50 μm.
Table 1.
Primary antibodies.