Differential regulation of MAP2 by phosphorylation events in proline-rich versus C-terminal domains

Abstract

MAP2 is a critical cytoskeletal regulator in neurons. The phosphorylation of MAP2 (MAP2-P) is well known to regulate core functions of MAP2, including microtubule (MT)/actin binding and facilitation of tubulin polymerization. However, site-specific studies of MAP2-P function in regions outside of the MT-binding domain (MTBD) are lacking. We previously identified a set of MAP2 phosphopeptides which are differentially expressed and predominantly increased in the cortex of individuals with schizophrenia relative to nonpsychiatric comparison subjects. The phosphopeptides originated not from the MTBD, but from the flanking proline-rich and C-terminal domains of MAP2. We sought to understand the contribution of MAP2-P at these sites on MAP2 function. To this end, we isolated a series of phosphomimetic MAP2C constructs and subjected them to cell-free tubulin polymerization, MT-binding, actin-binding, and actin polymerization assays. A subset of MAP2-P events significantly impaired these functions, with the two domains displaying different patterns of MAP2 regulation: proline-rich domain mutants T293E and T300E impaired MT assembly and actin-binding affinity but did not affect MT-binding, while C-terminal domain mutants S426E and S439D impaired all three functions. S443D also impaired MT assembly with minimal effects on MT- or actin-binding. Using heterologous cells, we also found that S426E but not T293E had a lower capability for process formation than the wild-type protein. These findings demonstrate the functional utility of MAP2-P in the proline-rich and C-terminal domains and point to distinct, domain-dependent regulations of MAP2 function, which can go on to affect cellular morphology.

Keywords: actin; cytoskeleton; microtubule-associated proteins; microtubules; mutagenesis; neurobiology.

Figure 1.. Isolated phosphomimetic MAP2C constructs.

Coomassie blue-stained SDS-PAGE gel (top) or western blot (bottom) of all MAP2C constructs used in the present study. Ladder band (far left) masses are in kDa. The predicted molecular weight of MAP2C is approximately 50 kDa but the protein typically migrates at ~70 kDa. Faint bands are observed at ~140 kDa and may reflect minor dimerization products.

Figure 2.. MAP2 phosphomimetic mutations in the C-terminal and proline-rich domain impair tubulin polymerization.

(A) Sequence map of the MAP2C protein. Positions of grouped phosphorylated sites are delineated with colored vertical lines. (B-D) Averaged polymerization curves for each group. Average is taken for each timepoint of the curve. All genotypes are compared to a reaction lacking MAP2 protein (Control). Vertical lines indicate SEM. (E-G) The presence of MAP2 significantly reduces tenth time (t_1/10), while several phosphomimetic mutants increase t_1/10 relative to WT. (H-J) The presence of MAP2 does not affect k_obs, the pseudo-first order rate constant of tubulin polymerization, yet several phosphomimetic constructs significantly reduce k_obs relative to WT. * p < 0.05; ** p < 0.01; **** p < 0.0001.

Figure 3.. C-terminal MAP2 phosphomimetic mutations impair MT-binding.

(A) ~77% of WT MAP2 was pulled down in the ultracentrifugation-based assay. (B,D,F) Representative stained SDS polyacryamide gels are shown for each group. (C,E,G) Several C-terminal phosphomimetic constructs show reduced levels of pull-down with preassembled MTs relative to WT MAP2. S = supernatant, P = pellet, Tub = Tubulin band in P. * p < 0.05; *** p < 0.001; **** p < 0.0001.

Figure 4.. C-terminal and proline-rich domain MAP2 phosphomimetic mutations impair actin-binding.

(A) ~36% of WT MAP2 was pulled down in the ultracentrifugation-based assay. (B,D,F) Representative stained SDS polyacrylamide gels are shown for each group. (C,E,G) Several proline-rich and C-terminal phosphomimetic constructs show reduced pull-down with preassembled actin filaments relative to WT MAP2. S = supernatant, P = pellet, F-actin = actin band in P. * p < 0.05; ** p < 0.01.

Figure 5.. T293E and T300E have differential effects on MT- and actin-binding.

(A) An XY plot showing the relative MT-binding (X-axis) and actin-binding (Y-axis) (as defined by P/Tub and P/F-actin in Figures 3–4) of tested genotypes. The unity line (black line) represents equivalent changes in MT- and actin-binding capabilities relative to WT. Data are shown as mean ± standard deviation (SD). SD for WT was averaged across groups. (B) P/Tub (MT) and P/F-actin (Actin) values are compared within-genotype. * p < 0.05; ** p < 0.01.

Figure 6.. MAP2 does not regulate actin polymerization.

(A) Average polymerization curve for the indicated levels of MAP2. Data were also obtained for a positive control reaction containing jasplakinolide (JSP), a negative control reaction lacking MAP2 protein (Control), and a negative control reaction lacking actin polymerization buffer (ABP-) (see Materials & Methods). Average is taken for each timepoint of the curve. Vertical lines indicate SEM. (B-C) MAP2 fails to change the half-time of the reaction (t_1/2) (B) or the polymerization rate at t_1/2 (AP_1/2) (C). (D) Average polymerization curve for all MAP2 genotypes (1 μg each) and JSP, Control and ABP- reactions. (E-F) All MAP2 genotypes fail to affect t_1/2 (E) or AP_1/2 (F). * p < 0.05; *** p < 0.001.

Figure 7.. The S426E, but not the T293E, mutation impairs process formation in HEK293T cells.

(A) Exemplar images of transfected cells. Scale bars = 10 um. (B-C) Mean intensities of MAP2 (B) and β-tubulin (C). (D) Exemplar images of cells with 0 (top left), 1 (top right), 2 (bottom left) or 3 (bottom right) process(es). Scale bars = 10 um. (E) Number of cells with or without at least 1 process per genotype. (F) Frequencies for number of processes per genotype. Numbers within the pie charts represent cells counted per category.

Figure 8.. Model of MAP2 regulation by phosphorylation in proline-rich and C-terminal domains.

Diagram of functional consequences resulting from MAP2-P. When MAP2 is relatively dephosphorylated (middle), MAP2 (purple) binds and crosslinks MTs (green) and actin (red) while promoting MT assembly. We here represent this crosslinking function as a multivalent interaction with both filaments by a single MAP2 molecule, as has been suggested for Tau. Following phosphorylation of MAP2 at sites in the proline-rich domain (left), the ability of MAP2 to assemble MTs is reduced (either due to increased incidence of catastrophe [pictured], or reduced rates of polymerization and/or bundling). MT/actin crosslinking is also presumed to be disrupted, as actin-binding affinity is reduced. However, MT-binding is preserved. After MAP2-P in C-terminal domain sites (right), MAP2 partially dissociates from both MTs and actin, also concurrent with reduced MT assembly and a presumed reduction in MT/actin crosslinking. Created with BioRender.com.