Trio, a Rho Family GEF, Interacts with the Presynaptic Active Zone Proteins Piccolo and Bassoon

Abstract

Synaptic vesicles (SVs) fuse with the plasma membrane at a precise location called the presynaptic active zone (AZ). This fusion is coordinated by proteins embedded within a cytoskeletal matrix assembled at the AZ (CAZ). In the present study, we have identified a novel binding partner for the CAZ proteins Piccolo and Bassoon. This interacting protein, Trio, is a member of the Dbl family of guanine nucleotide exchange factors (GEFs) known to regulate the dynamic assembly of actin and growth factor dependent axon guidance and synaptic growth. Trio was found to interact with the C-terminal PBH 9/10 domains of Piccolo and Bassoon via its own N-terminal Spectrin repeats, a domain that is also critical for its localization to the CAZ. Moreover, our data suggest that regions within the C-terminus of Trio negatively regulate its interactions with Piccolo/Bassoon. These findings provide a mechanism for the presynaptic targeting of Trio and support a model in which Piccolo and Bassoon play a role in regulating neurotransmission through interactions with proteins, including Trio, that modulate the dynamic assembly of F-actin during cycles of synaptic vesicle exo- and endocytosis.

Fig 1. Developmental expression of Trio isoforms and their association with synaptic junctions.

(A) Schematic diagram depicting rat forebrain isoforms of Trio as previously defined [62]. Inverted “Y”s depict the epitopes recognized by the two antibodies (αGEF-2 and αC-term) used throughout this study. (B) Western blots of postnuclear supernatants (PNS) generated from total rat brains at embryonic (E) or postnatal (P) ages. Note that while SV (Synaptophysin) and CAZ (Bassoon and Piccolo) proteins were upregulated during development, Trio isoforms (denoted by arrows to the right of blots) were partially downregulated. (C) Western blots depicting the amounts of PSD-95, Synaptophysin, Bassoon, Piccolo, ELKS2, and Trio in the following fractions successively prepared from adult rat forebrain: PNS, supernatant/cytosolic (S2), pellet (P2), synaptosomal (SYN), and synaptic junction/postsynaptic density (SJ/PSD). Although the smallest Trio isoform (Duet) was enriched in the S2 fraction, the three larger isoforms were enriched in the SJ fraction, similar to PSD-95, Bassoon, Piccolo, and ELKS2. Note, while the SJ is traditionally referred to as the PSD fraction, this is a misnomer as it contains scaffold proteins of both the PSD and AZ.

Fig 2. Trio localizes to synapses in cultured hippocampal neurons.

(A, B, and C) Images of 16–19 DIV neurons fixed in methanol and immunostained for the proteins indicated. Arrows depict puncta of either Piccolo or Synaptophysin that are positive for PSD-95 and, thus, likely excitatory synapses that were also immune-positive for Trio. Arrowheads depict puncta of Piccolo or Synaptophysin that were negative for PSD-95 (and thus either inhibitory synapse or transport vesicles) that were also Trio positive. Insets in A-C show higher mag regions of boxed areas. (D) Quantification of Trio puncta that colocalize with PSD-95 positive (+) or negative (-) puncta that were also positive for Piccolo or Synaptophysin puncta. Piccolo or Synaptophysin puncta were separated into PSD-95 positive (+) or negative (-) groups, which were then each scored as positive or negative for Trio. * = p < 0.0005, paired t test comparing PSD-95 positive (+) or PSD-95 negative (-) groups in individual images. n = 10 images for each group.

Fig 3. GFP-Trio-FL targets to dendritic spines and presynaptic terminals.

(A) Montage of several images of a neuron transfected with GFP-Trio-FL (green) growing near untransfected neurons immunostained with antibodies against the dendritic microtubule associated protein MAP2 (red). (B and C) Higher magnification images from panel A with GFP-Trio-FL (green), Piccolo (red), and MAP2 (blue), illustrating that GFP-Trio-FL targeted to dendritic shafts and spines (arrowheads panel B) as well as to MAP2-negative axons (arrows in panel B). Furthermore, GFP-Trio-FL had a distinctly punctate presynaptic localization in axons as demonstrated by the colocalization of these GFP-Trio-FL puncta (arrows in panel C) with Piccolo positive puncta (red) situated along dendrites of untransfected neurons (C).

Fig 4. Spectrin repeats 3 and 4 of Trio mediate binding to Bassoon.

(A) Examples of COS7 cells co-transfected with full-length (FL) Myc-tagged Trio-FL (red) and GFP tagged Bassoon (GFP-Bsn) or GFP-ELKS2 (green), used to evaluate binding interactions between these proteins. Note that GFP-Bsn effectively clustered Myc-Trio-FL. (B) COS7 cells co-transfected with Myc-tagged Trio polypeptides (red) and GFP-Bsn (green), used to evaluate which domain of Trio mediate binding to Bassoon. Note that Myc-Trio (382–624) clustered with GFP-Bsn, while Myc-Trio (607–859) did not.

Fig 5. Spectrin repeats 3 and 4 of Trio are required for co-immunoprecipitation (co-IP) with Bassoon (Bsn).

(A) Schematic diagram depicting FL-Trio and the polypeptides used in the clustering and co-IP assays. Column labeled “Clustering” summarizes the ability of different Trio mutants to bind GFP-Bsn (95–3938). Column labeled “Co-IP” contains a similar analysis for Trio constructs co-immunoprecipitated with GFP-Bsn (95–3938)(as shown in panel B). ND = not determined. (B) Western blot showing co-IP of GFP-Bsn95-3938 with Myc-Trio polypeptides (using Myc antibody) from COS7 cell lysates. Blots in left column (Myc) show levels of Myc-Trio constructs in total lysates (input), after control IP with rabbit IgG (IgG IP), and after IP with Myc antibody (α-Myc ip). Middle column (GFP-Bsn) shows GFP-Bsn 95–3938 levels for the same conditions, as detected with anti-GFP antibody. Right column shows Rabbit IgG, demonstrating that similar levels of antibody were used in the α-Myc and control conditions.

Fig 6. Trio interacts with the PBH9/10 regions in Piccolo and Bassoon.

(A) Western blots evaluating co-IP of GFP-Bassoon (Bsn) and Piccolo (Pclo) polypeptides with Myc-Trio1-1203 (using Myc antibody) from COS7 cell lysates. IP and Westerns were performed as in Fig 5B. (B) Schematic of Bassoon (top) and Piccolo (bottom) illustrating the following domains: Zinc finger (Zn), coiled-coil (CC), Piccolo/Bassoon homology (PHB; only 2 of the 10 PHB domains depicted), poly-Q (Q) PDZ, and C2 domains. Below the domain diagram are single lines depicting the regions of Bassoon and Piccolo that bound to Trio in panel A. The PBH 9 and PBH 10 domains are the only conserved domains in this region of these proteins.

Fig 7. Trio requires Spectrin repeats 3 and 4 to be targeted to presynaptic boutons.

Images of axons from neurons transfected with GFP-Trio-FL, GFP-Trio-FL(Δ395–606) or GFP-Trio(1–1203) growing along the dendrites of untransfected cells that have been fixed and immunostained with Piccolo (red) and MAP2 (blue). Neurons were fixed with methanol where indicated or formaldehyde (when fixative is not specified). Note that when GFP-Trio-FL is lacking amino acids 395–606 (Spectrin repeats 3 and 4), it did not localize to the presynapse and was readily extracted by methanol fixation. GFP-Trio (1–1203), which contains all the spectrin repeats, was targeted effectively to presynaptic sites.

Fig 8. The C-terminus of Trio inhibits binding with Piccolo/Bassoon on PTVs.

(A) Images of axons from immature neurons (5DIV) immunostained with Piccolo, Bassoon, and Trio. In these neurons, Piccolo and Bassoon co-localized (arrows in top panel), while Piccolo and Trio did not (arrowheads in bottom panel). (B) Axons from neurons (5DIV) expressing GFP-Trio-FL or GFP-Trio (1–1203), stained for Piccolo and MAP2. In these immature axons, GFP-Trio-FL remained diffuse, while GFP-Trio (1–1203) was punctate and colocalized with Piccolo. (C) Western blot of co-immunoprecipitation (IP) experiments comparing the ability of Myc-Trio-FL and Myc-Trio (1–1203) to co-IP with GFP-tagged Bassoon [GFP-Bsn(95–3938)]. Tubulin was included as an additional control demonstrating a lack of non-specific co-IP. (D) Quantification of panel C, obtained by dividing the pixel intensity of the GFP-Bassoon band in the anti-Myc lane by the GFP-Bsn band in the input lane and normalizing the results to the total amount of GFP-Bsn present in the lysate. Trio1-1203 immunoprecipitated Bassoon more effectively than full-length (FL) Trio. * = p < 0.05. n = 5.