Phosphorylation-Induced Mechanical Regulation of Intrinsically Disordered Neurofilament Proteins

Abstract

The biological function of protein assemblies has been conventionally equated with a unique three-dimensional protein structure and protein-specific interactions. However, in the past 20 years it has been found that some assemblies contain long flexible regions that adopt multiple structural conformations. These include neurofilament proteins that constitute the stress-responsive supportive network of neurons. Herein, we show that the macroscopic properties of neurofilament networks are tuned by enzymatic regulation of the charge found on the flexible protein regions. The results reveal an enzymatic (phosphorylation) regulation of macroscopic properties such as orientation, stress response, and expansion in flexible protein assemblies. Using a model that explains the attractive electrostatic interactions induced by enzymatically added charges, we demonstrate that phosphorylation regulation is far richer and versatile than previously considered.

Figure 1

Structure of bottlebrush NF filaments. (A) A schematic of two neighboring interacting filaments. Each filament consists of three different subunit proteins whose protruding tails are shown in color (red, blue and green). (B) Transmission electron microscopy images of de-phosphorylated filaments (from left to right) NF-LH, NF-LM, and NF-LMH. Scale bar, 100 nm. (C–F) Tail charge distributions of (C) phosphorylated NF-M, (D) phosphorylated NF-H, (E) de-phosphorylated NF-M, and (F) de-phosphorylated NF-H. The charge distributions were calculated at pH 6.8 and averaged over a five-amino-acid window (see Materials and Methods). The gray-shaded areas highlight the protein segments that are most affected by the phosphate charge removal. To see this figure in color, go online.

Figure 2

NF network phase behavior, as determined by cross-polarized microscopy. (A) Bright field and cross-polarized microscopy images of phosphorylated and de-phosphorylated networks at low (10²–10³ Pa) and high (10⁵ to 2 × 10⁵ Pa) osmotic pressure in quartz capillaries. White dashed lines demarcate the capillary boundaries, as observed with bright field (see Fig. S4). The filaments in all networks are aligned (nematic), except in the case of deNF-LM and deNF-LMH, where they are isotropic (i.e., un-oriented) at low $\Pi$. Each capillary is ∼1.5 mm wide. (B) Phase diagram showing the network phase behavior at different osmotic pressures $(\Pi )$. Each star denotes a measurement point. To see this figure in color, go online.

Figure 3

Comparison of native and de-phosphorylated networks using SAXS and osmotic pressure. (A) Intensity curves of NF-LH, NF-LM, and NF-LMH native and de-phosphorylated networks at 1% (w/w) PEG ($\Pi =2.2\times {10}^3$ Pa). (B–D) Semi-log plot of osmotic pressure, $\Pi$, versus inter-filament distance, d, for different network compositions: (B) NF-LMH, (C) NF-LH, and (D) NF-LM. For calculation of E in Eq. 2, we integrate over the textured area in (D). The osmotic bulk modulus, $B_{\text{T}}$, is shown in the insets. The data are fitted with smoothing splines, which are then used to calculate $B_{\text{T}}$. Isotropic and nematic samples are denoted by solid and open symbols, respectively. Subunit molar ratios for native and de-phosphorylated NF-LM, NF-LH, and NF-LMH filaments are 7:3 (NF-L/NF-M), 4:1 (NF-L/NF-H) and 10:3:2 (NF-L/NF-M/NF-H). To see this figure in color, go online.

Figure 4

Schematic of phosphorylation regulation of NF network expansion, alignment, and osmotic stress response, together with the conjectured tail configuration. (A and B) NF-H phosphorlyation aligns the isotropic deNF-LH network and increases the inter-filament distance (A), whereas NF-M phosphorylation collapses the nematic deNF-LM network (B). (C and D) Except for NF-LM networks, all protruding tails organize in two corona layers at low osmotic pressure. The outer layer is formed by the long tails (either NF-H or NF-M) and is denoted by yellow shading. Upon phosphorylation, deNF-LM tails transition from a flower (C) to a truffle conformation (D). (E and F) Under significant osmotic compression, filaments align and compress (E), whereas opposite tails increasingly inter-penetrate (F). To see this figure in color, go online.

Figure 5

Handshake analysis of short-range electrostatic interactions between NF tail regions. (A) Two opposite nine-amino-acid-long segments interacting. Details of the electrostatic interaction energy calculations are found in (32, 37). (B–E) Energy matrices for all possible segment pairs for two opposing NF-H (B), deNF-H (C), NF-M (D), or deNF-M (E) tails. Interactions between two oppositely charged segments, which are more electrostatically viable, are denoted in blue in the interaction matrices. The solid box in (B) marks the last 200 amino acids that are known to engage in attractive interactions. Dashed boxes in (E) denote phosphosite-rich segments (see also Fig. 1). Comparison of (D) with (E) reveals that de-phosphorylation forms new negative energy pairs between segments found farther away from the filament backbone. To see this figure in color, go online.