Recombinant antibodies against subcellular fractions used to track endogenous Golgi protein dynamics in vivo.

Generation of specific antibodies against enriched subcellular fractions is a powerful strategy to identify and characterize cellular components. We show that recombinant antibodies can be selected in vitro by phage display against complex subcellular fractions, namely microtubule-binding proteins and Golgi stacks. This technique has allowed us to overcome many limitations of the classical animal-based approach and generate cell biology-compliant antibodies. In addition, we show that intracellular expression of GFP-tagged recombinant antibodies can reveal the dynamics of endogenous proteins in vivo. Endogenous Giantin is very static and outlines the Golgi in living cells. It accumulates neither onto Golgi-derived tubules upon Brefeldin A treatment before Golgi disappearance, nor onto de novo formed Golgi mini-stacks upon microtubule depolymerization, and remains instead on the 'old' pericentriolar Golgi. This suggests that, in contrast to other Golgi matrix proteins, endogenous Giantin is very stably associated with the Golgi and does not efficiently recycle to the ER. Altogether, we show that the antibody phage display technique represents an efficient alternative to rapidly generate versatile antibodies that represent new tools to study protein function.

CLIP-170 interacts with dynactin complex and the APC-binding protein EB1 by different mechanisms.

CLIP-170 is a "cytoplasmic linker protein" implicated in endosome-microtubule interactions and in control of microtubule dynamics. CLIP-170 localizes dynamically to growing microtubule plus ends, colocalizing with the dynein activator dynactin and the APC-binding protein EB1. This shared "plus-end tracking" behavior suggests that CLIP-170 might interact with dynactin and/or EB1. We have used site-specific mutagenesis of CLIP-170 and a transfection/colocalization assay to address this question in mammalian tissue culture cells. Our results indicate that CLIP-170 interacts, directly or indirectly, with both dynactin and EB1. We find that the CLIP-170/dynactin interaction is mediated by the second metal binding motif of the CLIP-170 tail. In contrast, the CLIP-170/EB1 interaction requires neither metal binding motif. In addition, our experiments suggest that the CLIP-170/dynactin interaction occurs via the shoulder/sidearm subcomplex of dynactin and can occur in the cytosol (i.e., it does not require microtubule binding). These results have implications for the targeting of both dynactin and EB1 to microtubule plus ends. Our data suggest that the CLIP-170/dynactin interaction can target dynactin complex to microtubule plus ends, although dynactin likely also targets MT plus ends directly via the microtubule binding motif of the p150(Glued) subunit. We find that CLIP-170 mutants alter p150(Glued) localization without affecting EB1, indicating that EB1 can target microtubule plus ends independently of dynactin.

CLIPR-59, a new trans-Golgi/TGN cytoplasmic linker protein belonging to the CLIP-170 family.

The microtubule cytoskeleton plays a fundamental role in cell organization and membrane traffic in higher eukaryotes. It is well established that molecular motors are involved in membrane-microtubule interactions, but it has also been proposed that nonmotor microtubule-binding (MTB) proteins known as CLIPs (cytoplasmic linker proteins) have basic roles in these processes. We report here the characterization of CLIPR-59, a CLIP-170-related protein localized to the trans-most part of the Golgi apparatus. CLIPR-59 contains an acidic region followed by three ankyrin-like repeats and two CLIP-170-related MTB motifs. We show that the 60-amino acid-long carboxy-terminal domain of CLIPR-59 is necessary and sufficient to achieve Golgi targeting, which represents the first identification of a membrane targeting domain in a CLIP-170-related protein. The MTB domain of CLIPR-59 is functional because it localizes to microtubules when expressed as a fragment in HeLa cells. However, our results suggest that this domain is normally inhibited by the presence of adjacent domains, because neither full-length CLIPR-59 nor a CLIPR-59 mutant missing its membrane-targeting region localize to microtubules. Consistent with this observation, overexpression of CLIPR-59 does not affect the microtubule network. However, CLIPR-59 overexpression strongly perturbs early/recycling endosome-TGN dynamics, implicating CLIPR-59 in the regulation of this pathway.