Liprin-alpha has LAR-independent functions in R7 photoreceptor axon targeting

Abstract

In the Drosophila visual system, the color-sensing photoreceptors R7 and R8 project their axons to two distinct layers in the medulla. Loss of the receptor tyrosine phosphatase LAR from R7 photoreceptors causes their axons to terminate prematurely in the R8 layer. Here we identify a null mutation in the Liprin-alpha gene based on a similar R7 projection defect. Liprin-alpha physically interacts with the inactive D2 phosphatase domain of LAR, and this domain is also essential for R7 targeting. However, another LAR-dependent function, egg elongation, requires neither Liprin-alpha nor the LAR D2 domain. Although human and Caenorhabditis elegans Liprin-alpha proteins have been reported to control the localization of LAR, we find that LAR localizes to focal adhesions in Drosophila S2R+ cells and to photoreceptor growth cones in vivo independently of Liprin-alpha. In addition, Liprin-alpha overexpression or loss of function can affect R7 targeting in the complete absence of LAR. We conclude that Liprin-alpha does not simply act by regulating LAR localization but also has LAR-independent functions.

Fig. 1.

Liprin-α is required for R7 photoreceptor targeting. (A–C) Horizontal sections of adult heads carrying glass-lacZ to label all photoreceptors, stained with anti-β-gal. (A) WT. Photoreceptors project to the lamina (la) and medulla (me). R7 and R8 terminate in two distinct layers in the medulla (arrowheads). (B and C) In Liprin-α^oos homozygous mutants (B) and in LAR null mutants (C), only one layer of termini is present in the medulla (arrowhead). (D) oos flies have a nonsense mutation predicted to truncate the Drosophila Liprin-α gene within the N-terminal coiled-coil domain. The C-terminal liprin homology domain (LHD) of Liprin-α binds to the PTP-D2 domain of LAR. The extracellular domain of LAR, containing three Ig domains and nine fibronectin type III domains, is not shown. Three tyrosine residues (Y) within the LHD of Liprin-α are highly conserved throughout evolution. (E) Tangential section of a retina containing a small Liprin-α^oos mutant clone. The central R7 rhabdomere is present in both pigmented WT regions (yellow arrows) and unpigmented, mutant regions (red arrows). (F) Horizontal section of an adult head with Liprin-α^oos mutant clones carrying the PanR7-lacZ reporter to label all R7 photoreceptors, stained with X-Gal. Liprin-α^oos mutant R7 axons (brackets) terminate more superficially than surrounding WT R7 photoreceptors. (G and H) R1–R6 project from the retina (re) to the lamina (la) in both WT (G) and Liprin-α^oos (H) horizontal sections of adult heads carrying the Rh1-lacZ reporter and stained with anti-β-gal. (I) shows a Liprin-α^oos mutant whole-mount optic lobe at 24 h after puparium formation stained with anti-β-gal to reveal glass-lacZ expression. At this stage, most R7 termini are clearly separated from the R8 layer. A maximum intensity projection of six optical sections spanning ≈3 μm along the z axis is shown. (J) Horizontal head sections of the indicated genotypes carrying glass-lacZ were stained with anti-β-gal and scored for R7-targeting defects. In Liprin-α^oos homozygous mutants, only 37% of R7 axons project beyond the R8 layer. Expression of HA-tagged Liprin-α from two independent transgene insertions (#21 and #43) in all neurons with elav-GAL4 or in R7 with sev-GAL4 rescued R7 targeting to almost the WT level. HA-Liprin-α expression driven in R8 by 109.68-GAL4 did not improve R7 targeting. Number of cartridges counted are given in parentheses in this and subsequent figures.

Fig. 2.

Interactions between LAR and Liprin-α. (A) Myc-tagged Liprin-α and HA-tagged LAR expressed ubiquitously under da-GAL4 control can be coimmunoprecipitated from embryonic extracts with anti-Myc. Input lanes contain 1% of the total protein used in the immunoprecipitation. (B–D) Liprin-α can be coimmunoprecipitated with full-length LAR or the LAR PTP-D2 domain from S2R+ cells. S2R+ cells were transfected with the indicated constructs. Immunoprecipitations were carried out with anti-Myc (B) or anti-HA (C and D). Input lanes contain 0.5% of the total protein used in the immunoprecipiation. (E) Liprin-α is tyrosine phosphorylated. S2R+ cells were transfected with HA- or Myc-tagged Liprin-α. Liprin-α was immunoprecipitated from lysates with anti-HA or anti-Myc. Western blots with anti-phosphotyrosine, anti-HA and anti-Myc are shown. (F–K) LAR and Liprin-α localize to the growth cones of R1–R6 (arrowheads) in third instar larval optic lobes. (F–H) WT. (I–K) GMR-GAL4; UAS-HA-Liprin-α. Photoreceptors are stained with anti-β-gal to reveal glass-lacZ expression (red in F, H, I, and K); anti-LAR staining (cyan in G and H) is detected on R1–R6 growth cones and in the medulla neuropil (asterisk); Liprin-α was localized to R1–R6 growth cones as monitored by anti-HA staining (cyan in J and K). (L) Horizontal head sections of the indicated genotypes carrying glass-lacZ were stained with anti-β-gal and scored for R7 targeting defects. LAR²¹²⁷/LAR^C12 null mutants were rescued by expression of full-length LAR but not by LAR lacking the D2 domain.

Fig. 3.

Liprin-α is not essential for LAR localization in vivo. (A–F) LAR can be detected on the growth cones of R1–R6 photoreceptors (arrowheads in A) in third instar larval optic lobes (A–C) and on R7 termini in the adult medulla (D–F) in Liprin-α^oos mutants. All photoreceptors were labeled with anti-β-gal to reveal glass-lacZ expression (red in A, C, D, F, G, and I). LAR localization was monitored with anti-LAR antibody in the larval brain (cyan in B and C) where LAR is also strongly expressed in the medulla (asterisk). (D–I) Horizontal sections from adults carrying a GMR-HA-LAR insertion expressed in all photoreceptors were stained with anti-HA (cyan in E, F, H, and I). There is no obvious difference between the adult medulla of Liprin-α^oos heterozygotes (G–I) and Liprin-α^oos homozygous mutants (D–F) in HA-LAR distribution on R7 (open arrowheads) and R8 (filled arrowheads). (F and I Insets) Enlargements of R7 termini that have grown beyond the R8 layer to illustrate their aberrant shape in Liprin-α mutants.

Fig. 4.

Liprin-α is not essential for LAR localization in S2R+ cells. (A–F) Liprin-α and LAR colocalize with the focal adhesion marker Talin in S2R+ cells. S2R+ cells were transfected with HA-Liprin-α (A–C) or HA-LAR (D–F) and actin5c-GFP (blue in A and D) as a transfection marker. Cells were fixed and stained with anti-HA (B and E; red in A and D) and anti-Talin (C and F; green in A and D). (G–I) Liprin-α RNA interference did not alter LAR colocalization with Talin. S2R+ cells were transfected with HA-LAR and a double-stranded RNA hairpin construct directed against Liprin-α, and stained with anti-HA (H, red in G) and anti-Talin (I, green in G).

Fig. 5.

Liprin-α acts partially in parallel to LAR in R7 targeting. (A–C) Horizontal head sections of the indicated genotypes carrying glass-lacZ were stained with anti-β-gal and scored for R7 targeting defects. (A) Liprin-α overexpression in all neurons by elav-GAL4 or in photoreceptors by GMR-GAL4 partially rescued the R7 targeting defect in LAR mutants. Asterisks indicate significant differences from LAR null with no rescue construct (P < 0.0001 by Student’s t test). (B) Whole-eye mosaics of the indicated mutations were generated with ey-FLP and FRT40, M (2)24F. In the absence of both Liprin-α and LAR, fewer R7 axons were correctly targeted than in the absence of either Liprin-α or LAR alone. Asterisks indicate a significant difference from LAR^c12 mosaics (P < 0.0001 by Student’s t test). (C) In a Liprin-α^oos mutant background, we used elav-GAL4 to drive expression of the indicated constructs. ΔN-Liprin-α, which lacks the N-terminal coiled-coil domain, failed to rescue R7 targeting. Overexpression of LAR had no effect on R7 targeting.