Myeloid translocation gene 16b is a dual A-kinase anchoring protein that interacts selectively with plexins in a phospho-regulated manner

2.1. Vector constructs 

Preparation of MTG16b constructs is described previously [4], [23]. Constructs expressing full-length human PlexinA3-pcDNA, PlexinB1-pcDNA, and human/mouse chimera PlexinA1-pcDNA (all VSV-G tagged) were kindly provided by Dr. Luce Tamagnone. For bacterial expression, restriction enzyme digestion was used to obtain just the cytoplasmic domain of each plexin: XhoI for PlexinA1, EcoRI and NcoI for PlexinA3 and EcoRI for PlexinB1. Fragments were ligated into pET30 (S-protein-tagged) or pGEX-5X [glutathione S-transferase (GST)-tagged]. For PlexinA3 cloning into pGEX-5X, primers were designed to add BamHI and NotI restriction sites, respectively: forward 5′-CGCGGATCCATGCCATGGTGGCCCTGCAGAGC-3′ and reverse 5′-ATAAGAATGCGGCCGCCCTCCTCACCGATTCCACCAC-3′. Due to poor protein expression using PlexinB1-pcDNA to transfect COS7 cells, PlexinB1 was also expressed in pEGFP (cut from pcDNA construct as above).

2.2. Transformation, expression and pulldown assays 

Methods for transformation, bacterial protein induction and expression, and pulldown assays are described previously [24].

2.3. Transfection and immunoprecipitation (IP) and immunoblot (IB) 

Transfection was performed using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) as described previously [24], using 2.5μg of MTG-pcDNA (myc-his tagged) and 5μg of plexin-pcDNA (VSV-G tagged). Immunoprecipitation and immunoblot methods are described in [24], with modifications. Briefly, a glycerol lysis buffer was used (20mM Tris–HCl pH 7, 4, 5mM EDTA, 150mM NaCl, 10% glycerol, 1% Triton-X 100, and protease inhibitors as in [24]. IP antibodies were: rabbit polyclonal anti-VSV-G (8μg; Sigma–Aldrich, St. Louis, MO), mouse monoclonal anti-myc (10μg; Santa Cruz Biotechnology, Santa Cruz, CA; used to IP MTG rather than PlexinB1-GFP in co-IP experiments due to the myc antibody′s superior performance in IP assays compared to the green fluorescent protein (GFP) antibody), or mouse monoclonal anti-RI (4μg, BD Biosciences, San Jose, CA). Isotype appropriate IgG (all BD Biosciences) in amounts equal to IP antibody was used as a negative control in each IP. IB antibodies were VSV-G (1:2000), myc (1:500), RI (1:300), and rabbit polyclonal anti-GFP (1:250; Clontech, Mountain View, CA). Genscript (Piscataway, NJ) One Hour IP Western Kit was used according to manufacturer′s instructions for RI western blots to minimize interference from IP/IgG cross reaction with the IB antibody.

2.4. Phosphorylation 

Phosphorylation for pulldown experiments is described in [3].

3.1. MTG interacts with plexins 

To determine whether MTG interacts with plexins, two types of binding assays were performed. Co-immunoprecipitation experiments using co-transfected COS7 cells revealed that full-length (FL) MTG interacts with plexins A1 and A3, but not PlexinB1 (Fig. 1A, compare lane 3 in upper panels). To confirm the results of IP experiments using a second method, to delineate the binding domains involved, and to examine the effects of phosphorylation on binding, MTG16b (FL and fragments: amino acids 200–700, 700–1510, and 1510–2000) was bacterially expressed. Due to the fact that MTG is not a membrane or extracellularly expressed protein, we hypothesized that it would interact with the highly conserved cytoplasmic domain of plexins [4]. As such, we subcloned and expressed just the cytoplasmic domains (C1 and C2) of plexins A1, A3 and B1. Consistent with co-immunoprecipitation assays, the cytoplasmic domains plexins A1 and A3 bound to FL MTG16b, while PlexinB1 did not (Fig. 1B, compare lane 3 in upper panels). Interestingly, in fragment pulldowns, plexins A1 and A3 bound well to FL MTG, but they both bound only weakly and inconsistently to the MTG fragments (primarily binding to the 200–700 and 700–1510 fragments, data not shown). This may indicate that proper folding of the intact FL MTG protein is required for optimal binding of plexins.

View Large Image Download to PowerPoint

Fig. 1. MTG interacts with PlexinA1 and PlexinA3, but not PlexinB1. (A) COS7 cells were co-transfected with full-length MTG-pcDNA(myc-His) and PlexinA1-pcDNA, PlexinA3-pcDNA (both VSV-G tagged), or PlexinB1-pEGFP. Lane 1 in all panels is a lysate control for transfection. In the PlexinA1 and PlexinA3 panels, immunoprecipitations (IP) were performed with negative control IgG (lane 2) or VSV-G polyclonal antibody (lane 3). After SDS-PAGE and transfer, membranes were cut in half and immunoblotted (IB) with anti-VSV-G antibody to confirm IP (top panel), or anti-myc antibody (bottom panel), demonstrating that MTG co-IP with plexins A1 and A3. In the PlexinB1 panel, IPs were performed with negative control IgG (lane 2) or anti-myc monoclonal antibody (lane 3). Anti-myc antibody IB was performed to confirm IP, and polyclonal anti-GFP antibody IB was performed, demonstrating that MTG does not co-IP with PlexinB1. (B) Pulldown assays using bacterially expressed proteins and glutathione sepharose (GST) beads. Lane 1 in each panel is the plexin-S tagged protein. Plexin-S tagged protein lysates were incubated with GST beads alone (data not shown), GST beads bound to GST lysates (lane 2), or MTG-GST lysates (lane 3). Western analysis was used to detect binding of plexins (upper blot in each panel) to GST (lower blot in each panel). In (A) and (B) the results are representative blots of three independent experiments.

3.2. MTG is a dual AKAP 

We used co-IP and pulldown assasy to determine whether MTG interacts with RI. Co-IP experiments using MTG transfected COS7 cells revealed that MTG interacts with endogenously expressed RI (Fig. 2A). To confirm a direct interaction, purified RIα and MTG were used in pulldown assays. Consistent with co-immunoprecipitation assays, RIα interacts with FL MTG (Fig. 2B). Together with our previous studies that demonstrate RII-MTG interactions [4], these data indicate that MTG is a dual AKAP. This result is supported by recently published sequence analyses suggesting that in addition to the well-characterized RII-binding amphipathic helix domain, dual AKAPs contain a PKA binding region called the RI Specifier Region (RISR) [25]. Sequence alignment with the dual AKAPs presented in the Jarnaess study indicate that MTG16b contains the RISR motif (Supplementary data).

View Large Image Download to PowerPoint

Fig. 2. MTG interacts with RI. (A) COS7 cells were transfected with full-length MTG-pcDNA(myc-His). Lane 1 in both panels is a lysate control to demonstrate size and confirm transfection. IPs were performed with negative control IgG (lane 2) or RI monoclonal antibody (lane 3). After SDS-PAGE and transfer, membranes were cut in half and immunoblotted (IB) with anti-RI antibody to confirm IP (top panel), or anti-myc antibody (bottom panel) to demonstrate that MTG co-IPs with RI. Both IBs used Genscript’s (Piscataway, NJ) one-step IP Western blot kit. Since different antibodies are used, resulting bands should not be considered quantitative measures of protein. (B) Pulldown assays using S-protein beads. Lane 1 is RIα protein bacterial lysate compared to pulldown lanes 2 and 3. Bacterially expressed, affinity-purified RIα was incubated with S-protein beads alone (data not shown), S-protein beads bound to S-protein bacterial lysates (lane 2), or MTG-S protein bacterial lysates (lane 3). Western analyses with anti-RIα monoclonal antibody were used to detect RI binding. In (A) and (B) the results are representative blots of three independent experiments.

3.3. MTG complexes with plexin and RI 

In functioning as scaffolding proteins that target PKA-phosphorylation, AKAPs will often simultaneously bind the R subunits of PKA and PKA substrate(s) [23], [24]. To illustrate that MTG binds RIα and plexins concurrently, forming a signaling complex, we performed pulldown assays. Results indicate that MTG is capable of and necessary for forming a complex with plexins and RIα (Fig. 3, lane 2) as RIα and PlexinA3 do not interact directly (Fig. 3, lane 1).

View Large Image Download to PowerPoint

Fig. 3. MTG can form a complex with plexin and RIα. In modified pulldown assays, PlexinA3-GST was immobilized on GST beads and incubated with either negative control S-protein (lane 1) or MTG-S protein (lane 2). Both sets of beads were then washed and incubated with bacterially expressed, cAMP affinity-purified RIα (lanes 1 and 2). The samples were separated by SDS–PAGE, transferred to Immobilon PVDF membranes, and western blots were performed with anti-RIα monoclonal antibody and goat anti-mouse-HRP-conjugated secondary antibody in order to detect binding of RI (lower blot). Anti-GST-HRP and anti-S protein-HRP western blots were also performed in order to confirm loading of PlexinA3 (upper blot) and MTG (middle blot), respectively. These results are representative blots of three independent experiments.

3.4. Phosphorylation affects MTG interactions with RI and plexins 

Phosphorylation is a key event in many signaling pathways, often manifesting its effects by altering binding affinities. Recent studies have demonstrated regulation of AKAP protein interactions via PKA-phosphorylation of the AKAPs. Furthermore PlexinA, the drosophila plexin that interacts with MTG family member Nervy, may be a target for PKA-phosphorylation [3], [12]. In addition, tyrosine phosphorylation of plexins is important in semaphorin signaling [26]. As such, we performed studies to test the hypothesis that PKA-phosphorylation affects plexin/MTG/RI interactions. To begin, we performed experiments to determine the sites of PKA-phosphorylation on MTG, which are S536 and S411 (see Supplementary data).

Using FL MTG, purified RIα, and the cytosolic domains of PlexinA1 or PlexinA3 (see above), we performed pulldown assays to determine whether PKA-phosphorylation of either MTG or plexins alters interactions (RIα is constitutively phosphorylated in vivo). First, MTG was immobilized on beads and PKA-phosphorylated, plexins or purified RIα were added. Binding was compared to unphosphorylated MTG. Results indicate that PKA-phosphorylation of MTG significantly increases its interaction with PlexinA1, PlexinA3, and RIα (Fig. 4A). In reverse pulldowns, plexins were immobilized and PKA phosphorylated as above, then MTG was added. Results indicate that PKA-phosphorylation of PlexinA3 increases its interaction with MTG (Fig. 4B, right), while PKA-phosphorylation of PlexinA1 has no effect on its interaction with MTG (Fig. 4B, left).

View Large Image Download to PowerPoint

Fig. 4. PKA phosphorylation affects MTG interactions with plexins and RI. (A) MTG-S protein was immobilized on S-protein beads and either left unphosphorylated (bar 1 and lane 1 in each blot) or phosphorylated by the catalytic subunit of PKA (bar 2 and lane 2 in each blot). GST-PlexinA1 (left), GST-PlexinA3 (middle), or purified RIα (right) lysates were then incubated with the bead-bound MTG. Samples were electrophoresed, transferred and anti-GST-HRP western blots were performed to detect binding of the plexins or RIα to MTG. Phosphorylation of MTG by PKA resulted in a significant increase in binding to all proteins tested (PlexinA1 P=0.04, PlexinA3 P=0.02, and RIα P=0.02 by t-test; * denote significance). (B) PlexinA1-GST (left) or PlexinA3-GST (right) is immobilized on GST beads and either left unphosphorylated (bar 1 and lane 1 in each blot) or phosphorylated by the catalytic subunit of PKA (bar 2 and lane 2 in each blot). MTG-S protein lysate was then incubated with the bead-bound plexins. Western analysis detects binding of the MTG to plexins. Phosphorylation of PlexinA3 by PKA resulted in a significant increase in binding to MTG (P=0.01 by t-test, * denotes significance), while phosphorylation of PlexinA1 did not affect its interaction with MTG. The blots in A and B are each one representative of three separate experiments in which equal loading was determined by coomassie staining and western blot (anti-S protein-HRP, A, anti-GST-HRP, B). PKA-phosphorylation of proteins was confirmed using the anti-phospho-serine PKA substrate antibody. Densitometric analyses were performed using NIH Image software. Error bars are±S.E.M.