Property impaired in active Src-Y527FGFP mutant, presumably due to interactions with other membrane proteins in a Src SH2-dependent manner. Interestingly, we found that BiFC puncta were significantly reduced when the Y529F mutant of Src was analyzed, suggesting that PTP1B targets plasma membrane Src on this residue. Our 3,4,5-Trimethoxyphenylacetic acid time-lapse experiments showed that BiFC puncta move apparently at random and locate at the tip of ER tubules, and under TIRFM they were frequently visualized as bright spots at one tip of comet-like Ginsenoside-Ro fluorescent structures, suggesting a “dipping down” of ER tubules towards the plasma membrane in contact with the substrate. Parallel analysis of GFPPTP1BDA showed similar results suggesting that PTP1B localization at the tip of ER tubules imposes a restriction to the spatial propagation of the interaction with Src molecules associated to the cytosolic side of the plasma membrane. A schematic view derived from our results is shown in Figure 6. ERbound PTP1B positioning in the cell cortex requires microtubules. The cometlike fluorescent figures observed under TIRFM suggest that ER tubules approach to the membrane at different angles, as previously shown for microtubules. In the context of the substrate trapping mutant PTP1BDA, interactions with plasma membrane-associated Src are stabilized, and therefore BiFC puncta enhanced. Mechanisms underlying the fusion and split of BiFC puncta, which were also seen for GFP-PTP1B, are currently unknown, but likely depend on the dynamics of ER tubules. Our analysis of BiFC was extended to YC-PTP1BWT/Fyn-YN and YC-PTP1BDA/Fyn-YN pairs. In both cases we found a positive BiFC signal that distributes in puncta, as it was described for Src. Fyn has a more tight association to the plasma membrane than Src due to additional palmitoylation, and quickly associates with the plasma membrane after being synthesized. This strengthens the view that BiFC puncta between PTP1B and Src kinases most frequently occur in association with the plasma membrane. However, we cannot completely rule out transient interactions between PTP1B and Src within the endosomal compartment. To elucidate this, a more extensive co-localization analysis with markers for different endosomes would be required, as well as high resolution double time-lapse analyses. In our study we showed that a truncation of the Src Nterminus, which removes the myristoylation target site and the polybasic motif involved in membrane association, eliminates the production of BiFC. Thus, membrane-bound Src is a requisite for BiFC to occur. Myristate is added to Src cotranslationally by the N-myristoyl-CoA-protein transferase enzyme. Beyond that, little is known about the regulation of Src myristoylation. We identified the active site of PTP1B and the Src tyrosine 529 at the C-tail as major determinants underlying BiFC. Tyrosine 529 is phosphorylated by the C-terminus Src kinase, Csk, and dephosphorylated by PTP1B and other tyrosine phosphatases. Replacement of tyrosine 529 by phenylalanine significantly reduced but did not completely eliminate BiFC puncta. In addition, a single mutation converting the PTP1B active site in a substrate trap significantly enhanced BiFC puncta throughout the cell, provided that the tyrosine 529 of Src remained unchanged. A second determinant contributing to BiFC is a proline-rich motif of PTP1B which fits the consensus sequence for class 
II SH3 domain-binding motifs. Using PTP1BPA, a proline mutant in which the SH3binding motif was disrupted, moderately reduced the BiFC signal. Remarkably, PA mutation had no effect when combined with the substrate trap mutation DA.