Self-Assembled 2D Free-Standing Janus Nanosheets with Single-Layer Thickness.

We report the thermodynamically controlled growth of solution-processable and free-standing nanosheets via peptide assembly in two dimensions. By taking advantage of self-sorting between peptide β-strands and hydrocarbon chains, we have demonstrated the formation of Janus 2D structures with single-layer thickness, which enable a predetermined surface heterofunctionalization. A controlled 2D-to-1D morphological transition was achieved by subtly adjusting the intermolecular forces. These nanosheets provide an ideal substrate for the engineering of guest components (e.g., proteins and nanoparticles), where enhanced enzyme activity was observed. We anticipate that sequence-specific programmed peptides will offer promise as design elements for 2D assemblies with face-selective functionalization.

Two-dimensional (2D) nanostructures are an important material class with diverse potential applications.[1−4] It remains a challenge to produce free-standing 2D materials without a templating surface or confined space. Recently, covalent (e.g., click chemistry)[5−7] and noncovalent[8−12] strategies have been developed to construct 2D structures. Of particular interest are the biological units (e.g., protein, DNA, peptide, and peptoid) that can be engineered in a well-defined manner to favor self-assembly in 2D.[13−20] For example, homoligomeric 2D protein assemblies with spatial arrangements and patterns have previously been achieved by exploiting the directionality of metal coordination interactions.[13] However, although amphiphilic peptides represent an important class of supramolecular units with a proclivity toward aggregating into 1D nanomaterials (e.g., nanotubes and fibrils),[21−28] the design of 2D peptide structures is infrequently reported. This is because the H-bonding arising from amide groups dominates 1D peptide self-assembly preferentially along the direction parallel to H-bonding.

Herein, we construct 2D materials via the manipulation of peptide self-assembly in two dimensions (Figures a and S1). A peptide amphiphile F6C11 containing hexa-phenylalanine, a hydrophobic tail, and glutamic acids was designed. The hexa-phenylalanine segment was incorporated to provide H-bonding to form β-sheets in the x-axial direction, with concomitant β-sheet stacking in the y-axial direction via aromatic interactions (Figure b).[24−26,29,30] The resulting self-assembly in 2D is consistent with previous studies where tyrosine and phenylalanine-rich peptides have been reported to form 2D nanobelts and nanosheets.[20,29] An alkyl tail was incorporated to strengthen the hydrophobic effect in both directions (x and y), where enthalpy (H)-governed self-sorting between hexa-phenylalanine and the alkyl tail enables us to create asymmetric peptide layers with the capability of heterofunctionalization on the opposing faces (Figure b).[31]

Figure 1

(a, b) Schematic of F6C11 self-assembly into Janus nanosheets following self-sorting between phenylalanine (blue) and hydrocarbon tails (red). (c) Fluorescence spectra of Nile Red in the presence or absence of F6C11. (d) SIM, (e) TEM, (f) AFM images of nanosheets. The inset in (f) shows the height profile across the marked section indicated by A and B in (f). (g) SAXS profile plot for solution structures of nanosheets (black circles) and fit to lamellar sheets of thickness 4.8 nm (red trace). (h) CD spectrum of nanosheets.

The aggregation of F6C11 was confirmed by Nile Red fluorescence, where an increase of emission intensity and a spectral blue shift was observed, indicating the solubilization of Nile Red in a hydrophobic domain (Figure c). The solution of F6C11 was found to be cloudy with low viscosity, suggesting the formation of large but multidimensional aggregates that scattered incident light strongly. Structured illumination microscopy (SIM), transmission electron microscopy (TEM), and tapping mode atomic force microscopy (AFM) revealed homogeneous nanosheets hundreds of nanometers wide and several microns long (Figures d–1f, S2). The observation of a fraction of overlapping plates suggested that they were free-standing in solution and stacked during sample preparation. In two overlapped sheets, the shape of the second layer can still be identified, indicating that the nanosheets were ultrathin. Indeed, AFM showed a homogeneous thickness of ∼5 nm, in agreement with the theoretical monolayer thickness. Overlapped nanosheets were also visualized by AFM, giving a surface height of ∼10 nm. The single-layer nanosheets could be easily isolated from solution by centrifugation and recovered by redispersion without affecting the nanosheet morphology and structural integrity (Figure S3). The recovery yield of nanosheets after a centrifugation/redispersion process was determined to be as high as 92%. Such a property is important for nanosheet separation, purification, and functionalization.

Synchrotron radiation small-angle X-ray scattering (SAXS) confirmed the plate-like configuration with a Q–2 decay in the Guinier region (Figure g, Table S1). A quantitative assessment of the thickness of the structures was accomplished by fitting the data as lamellar sheets and indicated a thickness of 4.8 nm. The mechanism of 2D self-assembly was characterized by circular dichroism (CD), Thioflavin T (ThT) assay, and Fourier transform infrared spectroscopy (FTIR). Strong CD signals were observed for F6C11 nanosheets (Figure h), where a minimum at 225 nm and a maximum at 202 nm are representative of β-sheets. The minimum at 225 nm was red-shifted compared to ∼216 nm of a typical β-sheet, due to the aromatic interactions between phenylalanine.[29,32,33] The FTIR spectrum of F6C11 nanosheets exhibited the vibration peaks of amide I (1625 cm–1) and amide II (1543 cm–1), suggesting the parallel β-sheets (Figure S4). The ThT assay confirmed the existence of rigid β-sheets with an increase of fluorescence emission in nanosheets (Figure S5).

The surface of the nanosheets can selectively display specific reactive groups via coassembly of F6C11 and functionalized peptides that share the same backbone (Figure S1). Since the main driving forces for 2D self-assembly originate from the configuration of the hexa-phenylalanine and hydrocarbon chains, the incorporation of nonionic groups is not supposed to alter the hydrophobicity/hydrophilicity balance or the self-assembly pathway. Indeed, the insertion of a biotinylated peptide (Bio-F6C11) did not cause notable morphological changes to the nanosheets (Figure a, b). The biotin-displaying nanosheets can be quantitatively decorated with streptavidin, a 52.8 kDa protein, via biotin–streptavidin affinity due to the high affinity (Figure S6). AFM imaging demonstrated small features that included a height ∼5 nm, close to the size of a protein (Figure c, d); the thickness of nanosheets remained unchanged (∼5 nm, Figure e).

Figure 2

(a) Adsorption of streptavidin on a nanosheet. (b) TEM image of biotin-displaying nanosheets. (c) AFM image of streptavidin-anchored nanosheets. (d, e) AFM height profiles across nanosheet indicated as A and B in (c), respectively, suggesting the size of protein to be ∼5 nm and the single-layer thickness of nanosheet. (f) Surface assembly of HRP on the nanosheet via biotin–avidin affinity, and HRP-catalyzed TMB oxidization. (g) Kinetics of HRP-catalyzed TMB oxidation by H2O2..

The biotin-displaying nanosheets can be used as a substrate to immobilize enymes, e.g., avidin D-horseradish peroxidase (AvD-HRP) (Figure f) while retaining the enzyme activity. In the presence of H2O2, HRP catalyzed the oxidation of 3,3′,5,5′-tetramethylbenzidine (TMB) to produce a reaction product with an absorption at 650 nm. Interestingly, the immobilization of HRP on the nanosheets was found to increase the rate of catalytic reaction, which is not due to the intrinsic enzymatic activity of peptides (Figure g). A similar phenomenon was found when using negatively charged ABTS as a substrate (Figure S7). The enhanced enzymatic activity likely results from the highly negatively charged nanosheets acting to lower the local pH, which has previously been shown to increase the activity of HRP.[34]

Beyond biotin–avidin interactions, the peptide nanosheets can serve as a substrate to host inorganic nanomaterials (e.g., Au nanoparticles, quantum dots, Au nanorods) and fluorophores via thiol–Au bonding, electrostatic force, and copper-free click chemistry (Figures S8, S9).

Due to the self-sorting between hexa-phenylalanine and the alkyl chain, the single-layered nanosheets possess two different surfaces, with N-termini on Face A and C-termini on Face B (Figure b). To confirm the Janus nature of the nanosheets and their surface heterofunctionalization, we incoporated biotin and cysteine on opposing faces, via co-assembly of F6C11, Bio-F6C11, and F6C11-Cys, and immobilized the nanosheets on a silicon wafer via the thiol–maleimide reaction. The biotin-displaying face was therefore exposed to bulk solution and could be modified with streptavidin-coated AuNPs (Figure a, b). In contrast, the nanosheets displaying biotin and cysteine on the same surface (Surface A) could not be used to assemble streptavidin-AuNPs once they were anchored to the silicon wafer (Figure c, d) since the biotin moieties were embedded toward silicon and not exposed to solution. All together, we believe the peptide nanosheets are highly ordered with C-termini on one face and N-termini on the other, enabling the presentation of two different groups on opposing surfaces.

Figure 3

(a) Schematic and (b) scanning electron microscopy image showing nanosheets functionalized with cysteine and biotin (denoted as “B”) on opposing faces can be immobilized on silicon surfaces and modified with streptavidin-AuNPs. (c) Schematic and (d) SEM image showing nanosheets with cysteine and biotin on the same surface did not undergo specific AuNP binding. Scale bar: (b) 200 nm; (d) 1 μm.

Peptide nanostructures of varied morphology could be formed by varying the peptide sequences. For example, twisted fibrils were formed by F5C11 and F4C11 (Figure a, b), when the number of phenylalanines was reduced to 5 and 4, respectively. This is because a reduction in the number of phenylalanines weakens the β-sheet stacking in the y-axial direction, resulting in the formation of 1D structures along the x-axial direction. Meanwhile, the electrostatic repulsions between terminal charges in F4C11 and F5C11 became overwhelming, which can favor the intrinsic twist of β-sheet.[29,35] Similarly β-sheet fibrils were formed when valine was substituted for phenylalanine (V6C11) (Figures c, S10). More subtly, the size of nanosheets in both dimensions could be reduced by shortening the alkyl chain from C11 to C6, resulting in 2D plates with a width of ∼178 nm and a length of ∼1.05 μm (Figures d, S11). This can be rationalized since a shortening of the alkyl chain would reduce the hydrophobic interactions and therefore weaken the peptide association in both the x and y directions.

Figure 4

TEM images of nanostructures: (a) F4C11, (b) F5C11, (c) V6C11, and (d) F6C6. Scale bar: (a–c) 100 nm, (d) 500 nm. The arrows in (a) and (b) denote helical features of F4C11 and F5C11 fibrils, respectively.

These morphological transitions were accompanied by changes in internal ordering (e.g., β-sheet strength, micropolarity, and fluidity) of the peptide structures. As shown by ThT and Nile Red assays (Figure S12), the β-sheets were weakened and the micropolarity of the peptide structures was increased as the number of phenylalanines was reduced or substituted by valine (F6C11 > V6C11 > F5C11 > F4C11). Similarly, shortening of the alkyl chain from C11 to C6 was also found to weaken β-sheet strength and increase micropolarity. The fluidity of the peptide layers was probed by fluorescence polarization (FP) using 1,6-diphenyl-1,3,5-hexatriene (DPH). The FP value of DPH was found to increase (F6C11 ≈ F6C6 ≈ V6C11 > F5C11 > F4C11) when more amide H-bonds were present, indicating the decisive role of amide bonds in determining the fluidity (Figure S13). We therefore propose that F6C11 nanosheets display a high structural ordering cooperatively stabilized by H-bonding, aromatic stacking, and the hydrophobic effect as defined by the alkyl tail length, which compete with the electrostatic repulsion between glutamic acids.

In conclusion, we have designed Janus nanosheets with single-layer thickness via peptide self-assembly, which offers the possibility of dual functionalization on opposing surfaces. Benefiting from the versatility of surfaces displaying different reactive groups, the nanosheets can serve as substrates for various guest components. In particular, the surface-confined enzyme assembly on nanosheets was observed to enhance the catalytic activity, which can mimic the enzymatic reaction in living systems that take place at biological membranes. The important role of amino acid side groups (especially aromatic groups) was highlighted in promoting β-sheet stacking to favor peptide self-assembly in 2D. We anticipate these 2D Janus nanomaterials can be engineered with surface ligands (e.g., cell-targeting and analyte-binding properties) to inspire broad applications such as drug delivery and bioelectronic sensing.