Diverse Supramolecular Nanofiber Networks Assembled by Functional Low-Complexity Domains Bolin An,†,‡,§,△ Xinyu Wang,†,‡,¶,△ Mengkui Cui,†,‡,§ Xinrui Gui,⊥ Xiuhai Mao,† Yan Liu,∥,# Ke Li,†,‡,¶ Cenfeng Chu,∥,# Jiahua Pu,†,‡ Susu Ren,†,‡,§ Yanyi Wang,†,‡ Guisheng Zhong,∥,# Timothy K. Lu,& Cong Liu,⊥ and Chao Zhong*,† †
School of Physical Science and Technology, ShanghaiTech University, Shanghai 201210, China University of Chinese Academy of Sciences, Beijing 100049, China ¶ Shanghai Institute of Ceramics, Chinese Academy of Sciences, Shanghai 200050, China § Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences, Shanghai 200032,China ⊥ Interdisciplinary Research Center on Biology and Chemistry, Shanghai Institute of Organic Chemistry, Chinese Academy of Science, Shanghai 200032,China ∥ iHuman Institute, ShanghaiTech University, Shanghai 201210, China # School of Life Science and Technology, ShanghaiTech University, Shanghai 201210, China & Department of Electrical Engineering and Computer Science and Department of Biological Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139-4307, United States ‡
S Supporting Information *
ABSTRACT: Self-assembling supramolecular nanofibers, common in the natural world, are of fundamental interest and technical importance to both nanotechnology and materials science. Despite important advances, synthetic nanofibers still lack the structural and functional diversity of biological molecules, and the controlled assembly of one type of molecule into a variety of fibrous structures with wide-ranging functional attributes remains challenging. Here, we harness the low-complexity (LC) sequence domain of fused in sarcoma (FUS) protein, an essential cellular nuclear protein with slow kinetics of amyloid fiber assembly, to construct random copolymer-like, multiblock, and self-sorted supramolecular fibrous networks with distinct structural features and fluorescent functionalities. We demonstrate the utilities of these networks in the templated, spatially controlled assembly of ligand-decorated gold nanoparticles, quantum dots, nanorods, DNA origami, and hybrid structures. Owing to the distinguishable nanoarchitectures of these nanofibers, this assembly is structure-dependent. By coupling a modular genetic strategy with kinetically controlled complex supramolecular self-assembly, we demonstrate that a single type of protein molecule can be used to engineer diverse onedimensional supramolecular nanostructures with distinct functionalities. KEYWORDS: supramolecular nanofibers, self-assembly, low-complexity sequence domain, modular genetic design, nanoparticles, functional amyloid he controlled assembly of discrete molecules into fibrous structures is a widespread phenomenon in nature. The ability to form one-dimensional (1D) ordered nanostructures by supramolecular assembly is key to the extraordinary functionalities of many biomacromolecules found in living organisms1,2 and has also been of fundamental interest and technical significance in polymer chemistry, nanotechnology, and materials science,3−6 with a wide range of applications in
biomimetic mineralization,7,8 nanoelectronics,9,10 regenerative medicine,11,12 and energy-relevant areas.13,14 Natural 1D supramolecular structures are often assemblies of multiple, identical copies of a few building blocks. These building
T
© 2017 American Chemical Society
Received: April 2, 2017 Accepted: June 13, 2017 Published: June 13, 2017 6985
DOI: 10.1021/acsnano.7b02298 ACS Nano 2017, 11, 6985−6995
Article
www.acsnano.org
Article
ACS Nano
Figure 1. Functional single-component supramolecular nanofibers based upon EGFP-FUS LC and mCherry-FUS LC proteins. (a) Schematic of self-assembly and reversible solution-to-hydrogel phase transition of the genetically fused FUS LC proteins. The genetically fused FUS LC monomers first form liquid-like droplets (Figure S11), which locally increases protein concentration and in turn leads to the self-assembly of nanofibers. When the fiber concentration reaches a threshold, a hydrogel forms by increased physical entanglement among fibers. The resultant nanofibers typically comprise several pristine fibrils that are laterally stacked to each other. (b) AFM and fluorescent images of EGFP-FUS LC (left) and mCherry-FUS LC nanofibers (right). Scale bars, 10 μm. The green and red fluorescent images were acquired in the EGFP (488 nm) and the dsRed (545 nm) channel, respectively. (c) Far-UV circular dichroism (CD) spectra of EGFP-FUS LC and mCherry-FUS LC nanofibers. The appearance of minimum and maximum peaks around 220 and 200 nm, respectively, in the curves indicated that all samples were rich in β-sheet structures. (d) X-ray fiber diffraction patterns of EGFP-FUS LC and mCherry-FUS LC nanofibers. The data showed a typical diffraction pattern of cross-β spine of amyloid fibers in which the meridional reflection was at 4.71 and 4.54 Å, and the equatorial reflection was 10.07 and 10.08 Å for EGFP-FUS LC and mCherry-FUS LC, respectively.
assembly of one single type of molecule. Such construction faces major challenges mainly because (1) many existing selfassembling fiber systems have very fast kinetics of fiber formation, so it is difficult to achieve kinetically driven selfassembly, and (2) chemical approaches for constructing artificial self-assembling molecules often lack the flexibility and diversity of molecular design enabled by the modular genetic design strategy typically utilized in biology. Low-complexity (LC) sequence domains, frequently found in many RNA-binding proteins, can form “reversible amyloid” fibers under certain circumstances.28 Because of their controlled self-assembly at high local concentrations, they can further form hydrogels in vitro and possibly in vivo, as well. As the hydrogels form, the LC sequence domains cluster, sometimes with other molecules, into punctate structures called granules or intracellular bodies.29,30 In particular, the LC sequence domain of fused in sarcoma (FUS) protein, an RNA-binding protein with
blocks usually have well-defined repetitive sequences, so the assembly pathway is deterministic, arising from intramolecular folding and intermolecular noncovalent associations, such as hydrogen bonding and van der Waals interaction.3 Various artificial self-assembling systems have been developed with either well-defined protein/peptide sequences or tailor-designed molecules. These include functional amyloid-containing proteins,15,16 silk-like fusion proteins,17,18 β-sheet fibrillizing peptides,19,20 alkylated oligopeptides,21,22 and many precisely designed polymeric or small-molecule systems.23,24 Although a variety of supramolecular fibrous structures have been reported based on various types of molecules with distinct features,18,19,25,26 synthetic 1D supramolecular nanoarchitectures so far have less structural and functional diversity and less complexity than their biological counterparts.6,27 Furthermore, a broad spectrum of finely tuned structures with variable functionalities has not yet been constructed by the controlled 6986
DOI: 10.1021/acsnano.7b02298 ACS Nano 2017, 11, 6985−6995
Article
ACS Nano
Figure 2. Random copolymer-like supramolecular nanofibers based upon simultaneous complex assembly of EGFP-FUS LC and mCherry-FUS LC monomers. (a) Schematic of the formation of random copolymer-like supramolecular nanofibers by mixing the monomer solutions. (b) Digital camera images of protein solutions (40 μM) prepared by mixing EGFP-FUS LC and mCherry-FUS LC monomer solutions at different molar ratios. The actual color of the mixed solutions matched well with the predicted results. Scale bars, 5 mm. (c) Digital camera images of hydrogels prepared by mixing EGFP-FUS LC and mCherry-FUS LC monomer solutions (800 μM) at different molar ratios. Scale bars, 5 mm. (d) Fluorescent images of resultant random supramolecular copolymer-like nanofibers comprising different molar ratios of EGFP-FUS LC and mCherry-FUS LC monomers. Scale bars, 10 μm. The green and red images for FUS LC nanofibers were acquired in the EGFP (488 nm) and dsRed (545 nm) channels, respectively. The merged images clearly indicated that the nanofibers were composed of two FUS LC proteins with different proportions. The dense fiber networks shown in each case were obtained at a concentration of 40 μM, whereas the imaging of single fibers was acquired with a 1000-fold dilution.
RESULTS AND DISCUSSION Single-Component Supramolecular Fibrous Structures. The LC domain can self-assemble into fibers after incubation in solution for several hours under appropriate fiberforming conditions.31,33 However, these fibers tend to precipitate from solution over time. To increase the solubility of fibers in aqueous solution, we fused a wide range of soluble proteins or tags at the N-terminus of the FUS LC domain: EGFP, mCherry, mMaple3, and PAtagRFP (Figure S1). Two of these recombinant proteins, EGFP-FUS LC and mCherry-FUS LC, assembled into large numbers of nanofibers with enhanced stability in solution and functional fluorescence because of the flanking fusion domains (Figure 1b and Figure S12), consistent with previous findings.33 In contrast, when FUS LC was fused with less soluble tags, such as mMaple3 or PAtagRFP, the fusion proteins formed significantly fewer nanofibers in solution (Figure S12) and precipitated at higher concentrations (40 μM), similar to what had been observed with FUS LC alone (Figure S13a−c). These observations thus suggested that fusion partners contributed to fiber solution stability. All the nanofibers exhibited typical cross-β-structures, revealed by CD spectrum and X-ray fiber diffraction (Figure 1c,d and Figure S14). The
essential roles in RNA transcription, processing, and transport, is enriched in four amino acids (glycine, serine, glutamine, tyrosine) at the N-terminus region of 200 amino acids of FUS.31,32 Recent studies have revealed that this LC domain, with its intrinsically disordered nature, can form liquid-like droplets, which then assemble into highly dynamic amyloid fibers. These fibers undergo reversible solution-to-hydrogel phase transition under certain conditions.33 In addition, the FUS LC domain can be endowed with additional functionalities by the genetic fusion of various functional domains without affecting its selfassembling capability (Figure 1a). These attractive attributes of FUS LC make it a potentially useful building block with which to construct complex supramolecular nanoarchitectures. In this study, we leveraged the intrinsic self-assembling propensity of the FUS LC domain, along with a modular genetic strategy, to construct tailor-designed supramolecular structures with variable functionalities. We demonstrate that this integrative strategy is suitable for creating random copolymer-like, multiblock, and self-sorted supramolecular fibers that displayed distinct fluorescent functionalities and allowed the spatially controlled assembly of nano-objects in structure-dependent fashions. 6987
DOI: 10.1021/acsnano.7b02298 ACS Nano 2017, 11, 6985−6995
Article
ACS Nano
Figure 3. Multiblock supramolecular nanofibers based upon sequential assembly of EGFP-FUS LC and mCherry-FUS LC monomer proteins on preformed fiber seeds. (a) Schematic of the formation of multiblock nanofibers with alternating fluorescent features through the three-stage selfassembling processes. EGFP-FUS LC fiber seeds were first added to mCherry-FUS LC monomer solution (Stage 1), and the resultant fibers were sequentially incubated in EGFP-FUS LC (Stage 2) and mCherry-FUS LC monomer solution (Stage 3), eventually leading to multiblock nanofibers with alternating fluorescent features (Stage 2 and Stage 3). (b) Fluorescent images of the resultant block nanofibers after Stage 1 selfassembly. The 3D confocal fluorescent image of the resultant fibrous networks is shown on the right (corresponding movie shown in Supplementary Movie 1). (c) Fluorescent images of resultant multiblock nanofibers after Stage 2 self-assembly. (d) Fluorescent images of resultant multiblock nanofibers after Stage 3 self-assembly. (e) Percentages of single-, di-, and triblock suparmolecular structures obtained after the first-stage self-assembly at molar ratios of seeds to monomers of 1:3,1:6, and 1:9, respectively. (f) Percentages of the multiblock supramolecular structures comprising a specific number of segments obtained at various stages of self-assembly. The molar ratio of seeds to monomers was 1:3. The statistical analysis was based on 10 replicates acquired from three independent experiments. One replicate refers to counts based on at least 100 isolated pieces of supramolecular nanofibers randomly picked up from fluorescent images. Scale bars, 10 μm for dense fibers and 5 μm for single fiber.
nanfibers assembled in 40 μM solutions at 4 °C typically were several microns in length. The size of EGFP-FUS LC and mCherry-FUS LC nanofibers ranged from several to tens of nanometers, with an average fiber diameter of approximately 8.5 ± 2.4 and 8.8 ± 2.3 nm for EGFP-FUS LC and mCherry-FUS LC nanofibers, respectively (Figure S15). The size of nanofibers also seemed to be affected by solution temperature and concentration, with higher temperature leading to nanofibers with slightly reduced diameter and higher concentration, resulting in nanofiber bundles due to aggregation (Figures S16 and S17). These observations thus imply that the average diameters of nanofibers can be carefully regulated when temperature and concentration are both under control.
Colored hydrogels derived from EGFP-FUS LC and mCherryFUS LC formed in 3 h after increasing the concentration of the fiber solution to 800 μM because of increased physical entanglement among the fibrous networks (Figure S13d,e). Upon heating at or above 37 °C in a short time (
