In this study we reveal Lpd as a novel substrate of Abl kinases that is transiently phosphorylated upon PDGF and netrin-1 stimulation. We found that Lpd is phosphorylated by c-Abl at its PH domain and at the C terminus. Phosphorylation by Abl kinases at the PH domain does not regulate the interaction of Lpd with the plasma membrane because Lpd localizes at lamellipodia in Abl−/−Arg−/− fibroblasts ().
We observed that the interaction between Lpd and Ena/VASP proteins is positively regulated by c-Abl. However, Ena/VASP-binding sites in Lpd are not directly phosphorylated by Abl kinases. The phosphorylation of Lpd might alter its tertiary structure, thereby unmasking the Ena/VASP-binding sites. Furthermore, we noticed differences in the biochemical behavior of individual Ena/VASP proteins. We found that wild-type, dominant-active, and kinase-inactive c-Abl can coprecipitate with Lpd (Figure S2
A). Wild-type and dominant-active, but not kinase-inactive, c-Abl may bind directly to phosphorylated Lpd via the c-Abl SH2 domain. In contrast, kinase-inactive c-Abl may bind indirectly to Lpd via Ena/VASP, because Ena/VASP proteins can directly interact with the c-Abl SH3 domain [5
] and Ena/VASP can interact with FP4 motifs in Lpd [9
]. Furthermore, VASP localization to the leading edge is independent of Abl kinases (Figure S4
B), suggesting that its recruitment to this site can also occur without Lpd phosphorylation by Abl kinases. In addition, only Lpd-Mena, and not Lpd-EVL or Lpd-VASP interaction, is inhibited by expression of kinase-inactive c-Abl, suggesting that c-Abl may also regulate Mena directly. This might be achieved by phosphorylation of Ena or Mena, because they are substrates of Abl [5
]. We believe that this is unlikely because we did not observe any tyrosine phosphorylation of Mena when we co-overexpressed GST-Lpd, Mena, and c-Abl (data not shown; ). Furthermore, direct phosphorylation of Ena by D-Abl is not required for regulation of Ena function because expression of a nonphosphorylatable Ena mutant rescued axon guidance defects in the Drosophila ena
mutant [5, 23
]. However, Abl kinase activity is required for its function in axon guidance [24, 25
], suggesting that D-Abl must phosphorylate some other component. It was postulated that Abl regulates Ena's location in cells because Ena is mislocalized in d-abl
mutant flies [14
]. Taken together, this suggests that Abl regulates Ena localization indirectly by phosphorylating an unknown protein. In this study we have discovered that the interaction between Lpd and Ena/VASP proteins is positively regulated by c-Abl. Therefore, we suggest that Lpd is this hitherto unknown intermediary between Abl and Ena/VASP and that the differential formation of trimolecular complexes between Lpd, c-Abl, and individual Ena/VASP proteins allows for fine tuning of signaling responses.
This positive regulation of Lpd and Ena/VASP by Abl is surprising because it was postulated that Drosophila d-abl
negatively regulates ena
function [1, 2
]. How can both be reconciled? Unexpectedly, comparing known cellular functions of Abl kinases and Ena/VASP proteins revealed that many functions are very similar and not in opposition to each other. First, overexpression of both Abl kinases and Ena/VASP proteins increases filopodia formation [20, 22, 26
]. Second, analysis of knockout fibroblasts lacking Ena/VASP proteins or Abl kinases revealed that they migrated faster compared to cells re-expressing physiological levels of the respective proteins [8, 19, 27, 28
]. This indicates that both Abl kinases and Ena/VASP proteins negatively regulate whole-cell migration [8, 19, 27, 29, 30
]. Third, both Abl/Arg or Ena/VASP knockout mice have neurulation defects during development and die of hemorrhage [18, 31–34
]. Finally, both Drosophila ena
mutants have defects in longitudinal and commissural axon tracts [2, 4, 6
How can mutations in ena
phenotypes when Abl is a positive regulator of Ena? We suggest that Abl positively regulates the correct subcellular localization of Ena/VASP proteins, which is consistent with our observations and with those in d-abl
mutant flies [14
]. When D-Abl is absent, Ena accumulates ectopically where excess F-actin is observed [14
]. Consistently, reduction of ena
mutant flies restores viability [1, 2
]. Therefore, rescue of viability can be equally well explained by a positive regulation of Ena by Abl.
Lpd is transiently phosphorylated upon stimulation of primary cortical neurons with netrin-1 and is accompanied by an increased Lpd-Mena interaction. Treatment of neurons with netrin-1 causes increased formation of lamellipodia within 5 to 10 min [20
], which correlates well with the time course of Lpd phosphorylation and increased interaction with Mena. In C. elegans
function together to mediate unc-6
/netrin-dependent axon guidance decisions [4
]. Our data indicate that cooperation between Lpd and Ena/VASP proteins downstream of the netrin receptor also occurs in vertebrates.
We observed a transient phosphorylation of Lpd upon PDGF receptor stimulation and found that Lpd and Ena/VASP proteins contribute to the PDGF-induced dorsal ruffle response of fibroblasts. Our data indicate that Lpd function in the dorsal ruffle response is regulated by Abl kinases and mediated by Ena/VASP proteins. In lamellipodia, Ena/VASP proteins increase elongation of actin filaments by antagonizing actin filament capping [8
]. N-WASP and the Arp2/3 complex facilitate actin nucleation off of the side of existing actin filaments, and cortactin stabilizes these actin branches [35
]. N-WASP [6
], cortactin [36
], and Lpd (this study) are positively regulated by Abl kinases and function to regulate dorsal ruffling. A fine balance of Arp2/3-mediated branching and Ena/VASP-mediated elongation of actin filaments regulates the actin ultrastructure during lamellipodia protrusion [8
] and might also regulate dorsal ruffle formation. Therefore, positive regulation by Abl kinases and cooperation between Lpd, Ena/VASP, N-WASP, Arp2/3, and cortactin might be a more general mechanism to modulate the actin ultrastructure for different types of membrane protrusions.
In conclusion, we have identified Lpd as a component of PDGF and netrin-1 signaling pathways. Lpd is a substrate of Abl kinases, and c-Abl cooperates with Lpd in an Ena/VASP-dependent manner during the PDGF-induced dorsal ruffle response and axonal morphogenesis. Lpd's interaction with Ena/VASP proteins is positively regulated by Abl kinases. We propose that Lpd is the hitherto unknown intermediary between Abl and Ena/VASP. Our data do not support the suggested negative regulatory role of Abl for Ena, and we propose an alternative hypothesis that Abl kinases, via Lpd, positively regulate Ena/VASP proteins.