BioID Identification of Lamin-Associated Proteins

Abstract

A- and B-type lamins support the nuclear envelope, contribute to heterochromatin organization, and regulate a myriad of nuclear processes. The mechanisms by which lamins function in different cell types and the mechanisms by which lamin mutations cause over a dozen human diseases (laminopathies) remain unclear. The identification of proteins associated with lamins is likely to provide fundamental insight into these mechanisms. BioID (proximity-dependent biotin identification) is a unique and powerful method for identifying protein–protein and proximity-based interactions in living cells. BioID utilizes a mutant biotin ligase from bacteria that is fused to a protein of interest (bait). When expressed in living cells and stimulated with excess biotin, this BioID-fusion protein promiscuously biotinylates directly interacting and vicinal endogenous proteins. Following biotin-affinity capture, the biotinylated proteins can be identified using mass spectrometry. BioID thus enables screening for physiologically relevant protein associations that occur over time in living cells. BioID is applicable to insoluble proteins such as lamins that are often refractory to study by other methods and can identify weak and/or transient interactions. We discuss the use of BioID to elucidate novel lamin-interacting proteins and its applications in a broad range of biological systems, and provide detailed protocols to guide new applications.

Introduction

Eukaryotes have a nuclear envelope (NE) that delineates the nucleus and partitions the DNA from other cellular structures, and includes nuclear pore complexes (NPCs), which mediate transport between the nucleoplasm and cytoplasm (Ptak, Aitchison, & Wozniak, 2014). The NE is continuous with the endoplasmic reticulum (ER), but specialized: the outer membrane (outer nuclear membrane, ONM) is compositionally similar to the ER, whereas the inner nuclear membrane (INM) contains a myriad of integral membrane proteins (NE transmembrane proteins, “NETs”), likely retained by associations with nuclear constituents (Wong, Luperchio, & Reddy, 2014). One special class of NETs, named SUN-domain proteins, associate translumenally with ONM KASH-domain proteins to physically link the nucleoskeleton and cytoskeleton (LINC-complex) (Kim et al., 2015, Sosa et al., 2013).

The metazoan NE is structurally supported by networks of nuclear intermediate filaments, called the nuclear “lamina,” near the inner face of the INM. The nuclear lamina is composed of A- and B-type lamins. The LMNA gene encodes lamin-A and lamin-C (LaA and LaC), whereas LMNB1 and LMNB2 encode lamin-B1 (LaB1) and lamin-B2 (LaB2). Lamins have a central coiled-coil α-helical “rod” domain, a nuclear localization sequence, and a C-terminal immunoglobulin-like fold (Simon & Wilson, 2013). With the exception of LaC, lamins have a C-terminal CaaX motif that is posttranslationally prenylated and carboxymethylated. LaA is further processed by the removal of the C-terminal 15 residues from pre-LaA to generate mature LaA. Lamins form stable homodimers that can further assemble via head–tail associations into filaments (Simon & Wilson, 2013). Subsets of lamins localize and function in the nucleoplasm. Early in mitosis, the nuclear lamins are disassembled via specific phosphorylation that occurs in conjunction with NE disassembly (Kochin et al., 2014, Ottaviano and Gerace, 1985). Lamins impact diverse pathways including cell proliferation and senescence, differentiation, transcription, DNA repair, peripheral heterochromatin organization, epigenetic chromatin modifications, NE assembly and disassembly, and the dimensions of the nucleus (Dechat, Adam, Taimen, Shimi, & Goldman, 2010). Even during mitosis, when the NE and nuclear lamina are disassembled, lamins are implicated in the function of the mitotic spindle matrix (Tsai et al., 2006).

Mutations in LMNA cause at least 11 rare clinically distinct degenerative diseases including muscular dystrophy, cardiomyopathy, lipodystrophy, and progeria (Chi, Chen, & Jeang, 2009). Duplication of LMNB1 is associated with adult-onset leukodystrophy and mutations in LMNB2 lead to partial acquired lipodystrophy (Gao et al., 2012, Hegele et al., 2006, Padiath et al., 2006). These diverse tissue-specific pathologies are consistent with the tissue-specific phenotypes seen in mice that lack one, two, or all three lamin genes (Chen, Zheng, & Zheng, 2015). Mutations in NE membrane proteins, most of which bind lamins, can also cause diseases (“envelopathies”; Dauer and Worman, 2009, Worman et al., 2010) that phenotypically mimic or overlap with laminopathies, suggesting shared mechanisms. A better understanding of lamin-associated proteins may uncover the fundamental mechanisms by which lamins contribute to normal cellular function and human disease.

Protein–protein interactions provide critical insights into biological mechanisms, since proteins function collaboratively to drive cellular signaling cascades, regulation, differentiation, and proliferation. Proteins can also become “rogue” if they are mutated or modified. Rogue proteins have the potential to interact aberrantly with other proteins, organelles, or structures and thereby cause disease states.

One hurdle to overcome in studying lamins and associated proteins is the highly insoluble nature of the nuclear lamina itself. Ultrasonification with high concentrations of ionic- or chaotropic detergents to solubilize the lamina can disrupt lamin–protein interactions (Kubben et al., 2010). Various strategies have been used to identify lamin-associated proteins, including immunoprecipitation, cross-linking, fractionation of isolated nuclear membranes coupled with mass spectrometry (MS) analysis, and yeast two-hybrid (Y2H) (Simon & Wilson, 2013). These strategies can be successful but have specific drawbacks. Immunoprecipitation is done using nonphysiological conditions that can be insensitive, lead to loss of weak or transient interactions, or recover nonspecific proteins if wash conditions are insufficiently . Cross-linking prior to harsh lysis is attractive, but interpretation can be complicated by protein aggregation. Fractionation techniques are useful but require large amounts of sample that can be difficult to obtain, and do not reveal specific protein associations. The Y2H method can yield many false-positives. Results can also be affected by posttranslational modifications (PTMs) that may influence protein recovery, or be required for specific partners to interact. A novel technique to screen for potential protein–protein interactions in living cells, named BioID for proximity-dependent biotin identification, avoids most of these issues and offers a powerful new tool for studying lamin-associated proteins (Roux, Kim, Raida, & Burke, 2012).

BioID is a recently developed approach that uses a “bait” protein fused to a mutant biotin ligase which promiscuously biotinylates directly interacting or nearby primary amines of “prey” proteins over a period of time. BioID takes advantage of a 35-kDa enzyme, named biotin ligase (BirA), found in prokaryotic cells (Escherichia coli). Biotin is a water-soluble molecule found in living cells in small amounts as an enzyme cofactor. Under normal conditions, BirA uses ATP to activate biotin, forming biotinyl-AMP (bioAMP), which then reacts with lysine residues of very specific protein sequences (Chapman-Smith and Cronan, 1999, Lane et al., 1964). The BioID approach utilizes an R118G-mutated biotin ligase (hereafter called BioID) that prematurely releases bioAMP, allowing it to react with any free primary amines nearby (Kwon and Beckett, 2000, Roux et al., 2012). When this promiscuous ligase is fused to a specific protein bait, the resulting BioID-fusion protein can biotinylate directly interacting and/or vicinal proteins, leaving a “trail” of modified proteins that reveal its history of interactions over time.

The BioID method is shown schematically in Fig. 1. BioID can be used to screen potential partners of the bait in virtually any genetically tractable cell type. The bait ORF is recombined in-frame with the BioIDORF in an expression vector plasmid. A stable cell line expressing the BioID protein is then established using common transfection methods or virus-mediated DNA transfer. These stable cells are then incubated with excess biotin (typically 50 μM). Most conventional cell culture media do not contain biotin and the low level of biotin needed by cells is supplied by serum. Under these conditions, there is minimal biotinylation by the BioID-fusion protein. The level of biotinylation can be regulated by varying the time of incubation with excess biotin: 15–18 h is typically sufficient to saturate the biotinylation signal (Kim et al., 2014, Roux et al., 2012). The cells are then stringently lysed using SDS/Triton X-100, sonicated, and affinity-captured in a single step using magnetic beads coated with streptavidin, which has a well-known high affinity for biotin. These biotinylated proteins are then analyzed by MS or immunoblotting (IB). Fluorescently labeled streptavidin can be used to visualize biotinylated proteins in fixed and stained cells by immunofluorescence (IF). If antibodies against the bait are unavailable, the BioID-fusion protein can include a MYC or HA tag (Evan et al., 1985, Field et al., 1988). Antibodies specific for BirA are also now available for detection of BioID-fusion proteins by IF and IB (see Section 2.4.1).

BioID is especially powerful for screening protein–protein interactions in living cells. Partners and vicinal proteins are covalently biotinylated in their normal cellular environments and can therefore be recovered using stringent lysis conditions that completely solubilize and denature the proteins. Other major strengths of BioID are that candidate interactors are rapidly and selectively enriched by affinity capture and can be directly identified by MS.

BioID has been successfully applied to abroad range of proteins and cell structures including the nuclear lamina (Roux et al., 2012), NPCs (Kim et al., 2014), chromatin-associated protein complexes (Lambert, Tucholska, Go, Knight, & Gingras, 2014), the trypanosome bilobe (Morriswood et al., 2013), cell junction complexes (Fredriksson et al., 2015, Guo et al., 2014, Steed et al., 2014, Ueda et al., 2015, Van Itallie et al., 2013, Van Itallie et al., 2014), and centrosomes (Comartin et al., 2013, Firat-Karalar et al., 2014). This method has been used to screen for proteins involved in the Hippo signaling pathway (Couzens et al., 2013). BioID was used to study a novel protein secreted by a bacterial pathogen, Chlamydia psittaci, that targets the NE (Mojica et al., 2015), and to identify novel components of the inner membrane complex of Toxoplasma gondii (Chen, Kim, et al., 2015), analyze HIV-1 Gag protein interactions (Ritchie, Cylinder, Platt, & Barklis, 2015), and detect c-MYC interacting partners in cultured cells and mice xenograft tumors (Dingar et al., 2014).