3.2.4. Prph2K153Δ/+ and Prph2K153Δ/K153<sup>Δ</sup>

Another mutation in *PRPH2* that leads to variable phenotypes among patients is the deletion of codon 153 (K153 Δ) that results in the elimination of the lysine at position 153 in the D2 loop of Prph2. This mutation is found to associate with RP, pattern dystrophy and fundus flavimaculatus [80]. K153 Δ-Prph2 knockin mouse model was generated and provided evidence that the mutant protein cannot form the complexes required for OS formation [79]. The heterozygous knockin mice (*Prph2K153*Δ/+) displayed a shortened OS with minor structural defects at P30 [79] (Figure 1). At P180, these animals also exhibited a reduction in ONL thickness without further structural deterioration of the OS [79]. The overall Prph2 protein level in these mice was around 80% compared to the level in WT mice, thus demonstrating that the observed structural defects are most likely caused by the dominant e ffect of the mutant protein rather than due to haploinsu fficiency.

The heterozygous animals displayed a significant progressive reduction in scotopic responses that started as early as P30 (Figure 2) and worsened with age. The photopic response in the *Prph2K153*Δ/+ was also significantly reduced at P30, albeit this reduction did not worsen as the animals aged (Figure 2). In the homozygous animals, the scotopic and photopic responses were minimal at P30. Biochemical analysis showed that the formation of covalently linked Prph2 dimers in *Prph2K153*Δ/*K153*<sup>Δ</sup> mice was abolished while Rom1 homodimers were present [79].

Due to the observed e ffects on cone function, *Prph2K153*Δ/+ animals were crossed into the *Nrl*−/− background in order to assess the e ffects of the K153 Δ mutation on cones. Photoreceptor cells in the resulting *Prph2K153*<sup>Δ</sup>/+/*Nrl*−/− retina displayed a highly disrupted structure with many photoreceptors having no lamellae what so ever. Interestingly, the mutant Prph2 was able to interact with Rom1 in the rod-dominant *Prph2K153*Δ/*K153*<sup>Δ</sup> retinas, but this interaction was abolished in *Prph2K153*Δ/*K153*<sup>Δ</sup>/*Nrl*−/− retina, indicating a defect in the Prph2/Rom1 interaction specific to cones [79]. Sucrose gradient velocity sedimentation showed no significant alteration in complex formation in *Prph2K153*Δ/+ retina when compared to WT while in the *Prph2K153*Δ/*K153*<sup>Δ</sup> retina, the formation of intermediate and higher order complexes is abolished [79]. Here, both Prph2 and Rom1 are restricted to the tetramer fractions. The same was observed in the *Prph2K153*Δ/*K153*<sup>Δ</sup>/*Nrl*−/− retinas. Sedimentation profile showed that the amount of higher order complexes in the *Prph2K153*<sup>Δ</sup>/+/*Nrl*−/− retina was reduced, while una ffected in *Prph2K153*Δ/+ retina [79]. This provides further evidence for a di fferential role of the lysine at position 153 in rods and cones, and hence emphasizes the notion for potential varied roles for Prph2 in rods versus cones. Localization studies performed in *Prph2K153*Δ/+ mice demonstrated that most of Prph2 and Rom1 were successfully transported to the OS with some amount of Prph2 mislocalized to the IS (arrows in Figure 3). Small amount of rhodopsin (Rho) and M-opsin were also found to be mislocalized to the ONL and outer plexiform layer (OPL) in this model [79].

In the *Prph2K153*Δ/*K153*<sup>Δ</sup> retinas, the majority of Rho and Prph2, but not Rom1, were found to be mislocalized to the ONL. As stated above, patients with the K153 Δ mutation exhibit a highly variable phenotype, ranging from rod dominant RP to more cone related defects in the macula [80]. The K153 Δ knockin mouse model displayed functional and structural defects in both rods and cones, and thus mimics the phenotype seen in patients carrying this mutation. While the lack of a macula in the murine retina made it impossible to observe the macular pattern dystrophy often found in patients [80], the knockin mouse still showed funduscopic anomalies [79] which are characteristic of the pattern dystrophy in patients. *Prph2K153*Δ/+ mice show a flecking in the fundus at P180 which is more severe in the *Prph2K153*Δ/*K153*<sup>Δ</sup> mice. This phenotype does not deteriorate further in P365 heterozygous animals while the flecking was replaced by large splotches in the homozygous animals [79].

Gene supplementation using the NMP mouse that over-express Prph2 (*NMP*/*Prph2K153*Δ/+ and *NMP*/*Prph2K153*Δ/*K153*<sup>Δ</sup>) rescued the structural defects but failed to rescue the functional decline seen in scotopic and photopic ERGs [79]. This indicates that the presence of the mutant protein alone is su fficient to deteriorate the photoreceptor function and sugges<sup>t</sup> that gene silencing along with gene augmentation is the best strategy for this model.

The knockin models for *PRPH2* related patient mutations proved to be very useful in highlighting dominant-e ffects of the *Prph2* mutations. In general, the models were more successful in mimicking patient phenotypes related to a decline in rod function. Patient phenotypes related to functional defects in the cones or pattern dystrophies in the macula were more di fficult to reproduce in mice due to the lack of a macula and a lower overall percentage of cones in the murine retina. Crossing the knockin mouse models with *Nrl*−/− mice, as done in the studies with the K153 Δ model [79], has proven to be a successful approach in studying the e ffects of the mutation on cones in detail. While the knockin models could not reproduce the macular pattern dystrophy, they successfully reproduced the funduscopic aberrations, which connect with the pattern dystrophy.
