Loss-of-function and overexpression of Ror

4.2.1 Ror loss of function does not lead to lethality but results in a mild fasciculation defect while overexpression does not affect development

The lack of Ror protein was not lethal, indicating that the proteins function is not essential. The viability however was decreased, but only when compared to the wild type control. When compared to otk, otk2 double mutant embryos, the difference was not statistically relevant (Figure 22). This was also the case when compared to other fly lines (e.g da>Gal4 or UAS-Ror-Myc). So this increased embryonic lethality is probably artificial. In order to clarify this matter, one would have to increase the number of analyzed embryos (n=300) and also compare to the viability of the P-element line, which was used for the generation of the Ror⁴ allele. The ubiquitous overexpression of Ror-Myc clearly did not influence the viability (Figure 31).

Because of the specific expression of Ror in the nervous system, I was interested to see if I could uncover any defects in neural development. When looking at all axon tracts of the CNS or at the pattern of glial cells, the mutant embryos were unremarkable (Figure

23 B’/J’). But in the FasII staining several differences to the wt control became obvious.

The first one, disruptions in the lateral fascicles, was not statistically significant (Figure 24). More noticeable was that all three fascicles appeared to be wavy and many axons appeared not tightly associated with the fascicles (Figure 23 F’). This defect was more pronounced in the lateral and intermediate fascicles than in the medial one. The fact that these defects are not visible in the BP102 staining could indicate that only a subset of the longitudinal axons was affected. The pathway choices of all axons seemed not to be disturbed, all major axon tracts were formed and no defects in midline crossing were observed. So although Ror expression in the CNS begins at a point in development when differentiating axons begin to extend axons, the main requirement for the protein might be in the late stages of CNS development. To confirm this, it is necessary to examine the establishment of the longitudinal pathways in earlier embryos. Without a thorough analysis of Ror mutant CNS development covering all stages between stage 12 and stage 17 one cannot be sure if the observed phenotype is due to a defect in inter-axonal adhesion, axon guidance or axon fasciculation. Also important would be to see whether the observed defect is still visible in the larval and adult CNS.

To verify that the observed phenotype is specific to the loss of Ror function, one would have to examine whether it can be rescued by supplying wild type Ror either by expressing a UAS-Ror construct using a neuronal driver line (e.g. elav>Gal4) or by simply analyzing the CNS of Ror-eGFP embryos in a Ror⁴ mutant background. If this was indeed the case, one could identify the required protein domain(s) by performing rescue experiment with different deletion constructs.

Interestingly, a phenotype with similarities has been described for the double mutant of two receptor protein tyrosine phosphatases (RPTPs), Ptp4E and PtP10D (Jeon et al., 2008). In Drosophila, all six existing RPTPs are involved in CNS and motor axon guidance.

Among them, there is extensive redundancy and the observed phenotypes are of varying severity (Jeon et al., 2008; Sun et al., 2001; Schindelholz et al., 2001). In Ptp4E¹, PtP10D¹ double mutant embryos, the longitudinal fascicles appear wavy and exhibit some fraying. The longitudinally projecting SemaIIB-positive axons also do not form a tight bundle and appear frayed. However, sometimes there are also discontinuities in the fascicles and additionally they exhibit defects in motor axon guidance (Jeon et al., 2008).

It would be important to look for motor axon guidance defects in Ror mutants and quite interesting to analyze the CNS of embryos mutant for Ror and one or several RPTPs.

Genetic redundancy between Wnt signaling components has been shown several times already (see 1.2 and 1.4). Therefore is likely that this is also the case between the two Drosophila Ror family members. The second member Nrk has been implicated in the maintenance of adult muscles and in axon guidance and rhabdomere elongation during eye development (Kucherenko et al., 2011; Marrone et al., 2011). We have analyzed larval brains in which we downregulated Nrk via RNAi, but did not observe any obvious defects (Loth, 2014) and also downregulated Ror and Nrk together via RNAi. The resulting flies were viable, displayed no PCP defects but we did not examine their CNS (data not shown). A first indication as to whether Ror and Nrk act together could come from co-overexpressing the two proteins, but to fully understand the function of Drosophila Ror proteins, the generation of a Ror, Nrk double mutant is crucial.

In some of the analyzed otk, otk2 mutant embryos, the fascicles appeared wavy as well, but breaks or fraying could be seen (Figure 23 G’). For otk it has previously been suggested that its loss leads to guidance defects of motor axons and aberrant photoreceptor axon projections in the brain (Winberg et al., 2001; Cafferty et al., 2004).

However, the allele used in these reports (otk³) is lethal, which is in contrast to our recently published otk single and double mutant alleles (Linnemannstöns et al., 2014).

Thus, the lethality of the published phenotype might be due to a second site lethal mutation and all previous findings concerning otk function might be misleading (Linnemannstöns, 2012). The analysis of the morphology of the otk, otk2 mutant nervous system was not within the scope of this thesis. The CNS of otk, otk2 double mutant embryos was used as a control for the Ror, otk, otk2 triple mutant line. This was done in order to ensure that possible defects observed in the triple mutant cannot be attributed to the lack of otk and otk2 and the number of otk, otk2 double mutant nervous systems analyzed was quite low. Until the morphology of the otk, otk2 mutant nervous system has been thoroughly re-examined, the involvement of the two proteins in nervous system development remains unclear.

4.2.2 Neither Ror loss of function nor Ror overexpression affect PCP

In vertebrates, the absence of either Ror2 or PTK7 leads to characteristic PCP defects such as disturbed orientation of the hairs in the mouse inner ear and characteristic gastrulation and neurulation defects (Lu et al., 2004; Paudyal et al., 2010; Hikasa 2002;

Yamamoto 2008; Ho et al., 2012). However, in Drosophila, the absence of Ror, otk and otk2 by themselves or in combination has no influence upon the establishment of planar cell polarity (Figure 26; Figure 27; Linnemannstöns et al., 2014). The same is true for the overexpression of Myc (Figure 32 and Figure 33). This is not surprising, since Ror-eGFP expression in the imaginal discs is only expressed in small very distinct clusters. For a protein with an essential function in PCP signaling, one would expect a much broader expression domain.

4.2.3 The combined loss of Ror, Otk and Otk2 increases the lethality rate and some embryos display CNS defects

The combined loss of Otk and Otk2 did not increase the embryonic lethality (Figure 22;

Linnemannstöns et al., 2014) and the loss of Ror alone did not seem to influence the viability as well. However, when all three genes were lacking, the embryonic lethality appeared to be significantly increased (Figure 22). But this increase is not due to embryos dying during embryogenesis but rather to an increased number of unfertilized eggs. Noticeable was that of all embryos that hatched, only 27 % developed to adulthood. So the lack of Ror, Otk and Otk2 together increases the overall lethality during post-embryonic stages of development.

Some triple mutant embryos displayed a phenotype within the CNS (Figure 23 M). These embryos explain the high standard deviation in Figure 24 and most likely represent embryos, which would not survive to adulthood. The outermost fascicle in these nervous systems is discontinuous and displays numerous fascicle breaks (Figure 23 M). This phenotype is somewhat unexpected since the CNS of Ror⁴ mutant embryos displayed different defects. The FasII staining of the CNS of other triple mutant embryos resembled the wild type (Figure 23 H’) and in the BP102 staining showing all CNS axons, all analyzed embryos were unremarkable (Figure 23 D’). To quantify the fraction of nervous systems

carrying the severe phenotype, it is necessary to examine a much higher number of embryos. To ensure that the analyzed embryos are indeed late staged and their CNS is fully developed, it would be wise to co-stain for a second protein, which makes it possible to follow CNS condensation (e.g. Even-skipped).

While the intermediate fascicle was mostly intact in the triple mutant CNS and they displayed no commissural phenotype, their CNS phenotype was still somewhat similar to that of wnt5 mutant embryos. Due to the failure of the inner and outer fascicle to defasciculate from each other during stage 14, their intermediate fascicle shows breaks, and the outer fascicle is discontinuous as well (Fradkin et al., 2004).

The formation of the longitudinal axon pathways requires interactions between neurons and glial cells. While earlier in CNS development the pioneer axons act as support for glia cells, later the longitudinal glia cells provide axon guidance cues and direct fasciculation and defasciculation (Hidalgo et al., 2000). However, since the number and positions of longitudinal glia were neither affected in the Ror⁴ single mutant embryos nor in the triple mutant embryos (Figure 23 J’/L’), this can be ruled out as the cause of the observed phenotypes.

To ensure that the observed phenotype is indeed due to the lack of Ror, otk and otk2, rescue experiments with Ror, otk or otk2 alone have to be performed.