Intriguingly, treatment with PI-PLC strongly reduced the growth cone collapse rate of T→N interactions to a level similar to those observed for the other three interactions (Figure 2). While PI-PLC treatment affects all GPI-anchored proteins, the most parsimonious explanation seems to be that the removal of ephrinAs from nasal axons in these cultures essentially reduces growth cone collapse to baseline levels. Our in vitro data therefore confirm F. Bonhoeffer’s early in vitro findings and offer a good molecular candidate, i.e., ephrinAs, for the growth

cone collapse-inducing activity of nasal axons. To analyze the role of ephrinAs on RGC axons in vivo, we used ephrinA5 conditional “KO-first” mice (Skarnes et al., 2011) (Figure S2A available online), which we obtained from the International Knockout Dabrafenib mouse Mouse Consortium (IKMC) and the European Conditional Mouse Mutagenesis (EUCOMM)

project. In these mice, loxP sites flank the second exon of the ephrinA5 gene, which contains most of the coding region of ephrinA5, while the first exon contains only the first 20 amino acids (aa) of the mature protein, which comprises in total 228 aa. Thus, a conditional, Cre-mediated excision of exon 2 abolishes the synthesis of a functional ephrinA5 transcript ( Figure S2C). For the widely published full KO of ephrinA5, a comparable approach was taken, that is, deletion of exon BI 2536 mouse 2 by homologous recombination ( Feldheim et al., 1998 and Frisén et al., 1998). In addition, these KO-first mice harbor a splice acceptor-IRES-lacZ cassette, flanked by frt sites, in the intron preceding exon 2 (Figure S2A) (Skarnes et al., 2011), which abolishes the normal splicing from exon 1 to exon 2. Since we were interested here in analyzing retinocollicular mapping after a conditional

next inactivation of ephrinA5 in either the retina, the SC, or both (Figure S2C), we first removed this SA-IRES-lacZ cassette by breeding them with mice expressing flp recombinase ubiquitously (http://www.jax.org) thereby restoring the normal splicing/expression of the ephrinA5 gene while retaining the loxP sites flanking exon 2 ( Figure S2B). In order to abolish expression of ephrinA5 in the retina, a mouse line was chosen in which Cre is expressed under the control of the rx promoter (rx:cre) (Swindell et al., 2006), and for inactivation of ephrinA5 in the SC, the En1cre/+ line, in which Cre is expressed from the endogenous engrailed-1 promoter (Basson et al., 2008). Both lines have been extensively characterized (Basson et al., 2008 and Swindell et al., 2006). Expression of rx starts at E8.5 in the entire prospective optic vesicle; accordingly Cre will be expressed in all RGCs. While there is additional expression in the entire forebrain (see also Ackman et al., 2012 and Pinter and Hindges, 2010), expression of ephrinA5 in the SC is unaffected.