Dual Self-Assembled Nanostructures from Intrinsically Disordered Protein Polymers with LCST Behavior and Antimicrobial Peptides

Abstract

Antimicrobial peptides (AMPs) have attracted great interest as they constitute one of the most promising alternatives against drug-resistant infections. Their amphipathic nature not only provides them antimicrobial and immunomodulatory properties but also the ability to self-assemble into supramolecular nanostructures. Here, we propose their use as self-assembling domains to drive hierarchical organization of intrinsically disordered protein polymers (IDPPs). Using a modular approach, hybrid protein-engineered polymers were recombinantly produced, thus combining designer AMPs and a thermoresponsive IDPP, an elastin-like recombinamer (ELR). We exploited the ability of these AMPs and ELRs to self-assemble to develop supramolecular nanomaterials by way of a dual-assembly process. First, the AMPs trigger the formation of nanofibers; then, the thermoresponsiveness of the ELRs enables assembly into fibrillar aggregates. The interplay between the assembly of AMPs and ELRs provides an innovative molecular tool in the development of self-assembling nanosystems with potential use for biotechnological and biomedical applications.

Graphical Abstract

1. INTRODUCTION

Self-assembly is ubiquitous in nature and is a powerful strategy for the fabrication of materials.¹ Understanding the self-assembling processes of biological systems facilitates the fabrication of novel supramolecular materials and vice versa. Proteins, one of the most abundant macromolecules in living systems, constitute an important source of inspiration.² Given their multiple and unique functions, as well as their simple composition, protein-inspired materials enable the development of advanced self-assembling nanosystems with virtually limitless variations of their biochemical and bioactive properties for biotechnological and material engineering applications.^3–5 In this sense, proteins that undergo phase transition and organize into hierarchical assemblies have been especially relevant. A common feature of these proteins is the presence of structural disorder. Despite the lack of a defined tertiary structure, intrinsically disordered proteins (IDPs) and protein regions (IDRs) are involved in vital cell functions. Indeed, structural disorder plays a fundamental role in the mechanical properties of elastomeric proteins and in protein phase transition.^6,7 A number of IDPs and IDRs phase-separate under physiological conditions driving the formation of subcellular membraneless compartments (so-called biomolecular condensates).^7,8

Intrinsically disordered protein polymers (IDPPs) are artificial polypeptides composed of the repetition of conserved motifs found in IDRs, typically in structural proteins.⁹ This confers them stimuli responsiveness and valuable mechanical properties that make them interesting candidates for the biofabrication of hierarchical materials.¹⁰ This is the case of elastin-like recombinamers (ELRs), a class of protein-engineered polymers based on low-complexity sequences found in the hydrophobic domains of tropoelastin.¹¹ ELRs are biocompatible polymers that undergo a reversible lower critical solution temperature (LCST) phase transition in aqueous solution.¹² Rational modular approaches lead to the synthesis of elastin-like multiblock copolymers that self-assemble at physiological temperature into different nanostructures and hydrogels.¹⁰ Moreover, the thermoresponsive self-assembly can be extended further to create more complex architectures. Other self-assembling protein domains can be introduced in the modular design with exquisite control, thanks to their recombinant production,¹³ thus controlling the self-assembly process and giving access to new synthetic designs. As such, IDPPs offer a tailored platform for the fabrication of self-assembling nanosystems for fundamental or applied sciences, including molecular models for IDPs and biomolecular condensates,^14,15 advanced nanovehicles for drug delivery, or multifunctional scaffolds for tissue engineering.^16–20 Nevertheless, there is a need for innovative self-assembling domains (SADs) that expand the range of molecular designs and functionalities of hybrid IDPPs.

Antimicrobial peptides (AMPs) are short (10–50 amino acids) and generally cationic peptides found in a wide variety of multicellular organisms.²¹ Natural and designer AMPs have gained increasing attention in recent years because of their broad-spectrum activity and immunomodulatory properties as well as their ability to form supramolecular assemblies.^22,23 Their tendency to aggregate and self-assemble results from their amphipathicity, which allows them to interact with several molecular targets.^23–25 Recent studies have demonstrated the possibility to obtain different supramolecular architectures from different AMPs, including fibers,²⁶ spherical nano-particles,²⁷ twisted nanoribbons, or hydrogels.^28–32 Additionally, AMPs can be chemically conjugated with synthetic and natural self-assembling polymers or functionalized with peptide amphiphiles in order to enhance their antimicrobial performance.^33–35 However, despite the increasing interest on AMPs and AMP conjugates, no one to the best of our knowledge has studied their potential as SADs within protein-engineered polymers.

Herein, this study aims to investigate the interplay of potential self-assembling AMPs and IDPPs, in this case ELRs. We propose an alternative approach for the design of hierarchically self-assembled nanomaterials that exploits the self-organizing capability of AMPs and the thermoresponsiveness of ELRs. We hypothesized that the recombinant synthesis of hybrid polymers combining AMPs within IDPPs could lead to hierarchical supramolecular assembly by way of a dual process. Thus, aggregating and self-assembling tendencies of AMPs could be synergistically combined with the stimuli responsiveness of the IDPPs in order to fabricate functional supramolecular materials for biotechnological and biomedical applications. De novo designed AMP-ELRs were recombinantly produced, and their thermal behavior and self-assembly dynamics were characterized via turbidimetry, circular dichroism (CD) spectroscopy, dynamic light scattering (DLS), and electron microscopy. This approach seeks to shed light on the complex mechanisms that govern the supramolecular assembly of hybrid AMP-IDPP systems, thus setting the basis for the synthesis of advanced bioinspired materials that synergistically combine their complex self-assembly dynamics.

2. EXPERIMENTAL SECTION

2.1. Gene Construction.

Gene construction of AMP-IDPPs was performed using previously described procedures.³⁶ Encoding genes for the AMPs were purchased from NZYTech, Lda. (Portugal) and cloned into a modified pDrive plasmid flanked by EarI restriction sites, using Escherichia coli XL-1 blue (Agilent, USA) as the cloning strain. The final genetic constructs were then completed using the iterative recursive method.³⁶

2.2. Bioproduction and Purification.

All the protein polymers used in this work were recombinantly bioproduced. Briefly, encoding genes were cloned into pET-25b (+) expression vectors and transformed into E. coli BLR (DE3) for heterologous expression. After overnight fermentation in a 15 L bioreactor (Applikon Biotechnology, the Netherlands), the biopolymers were purified taking advantage of the LCST phase transition of the ELRs by inverse transition cycling (ITC), adding 1.5 M NaCl for warm precipitation.³⁶ After three cycles, the biopolymers were found to be pure and monodisperse by sodium dodecyl sulfate-polyacrylamide gel electro phoresis (SDS-PAGE) and were then dialyzed against ultrapure water, lyophilized, and stored at −20 °C. The yields observed ranged from 380 to 600 mg L⁻¹ of purified IDPPs (ELR/AMP-ELR) per liter of bacterial culture.

2.2.1. Protective Block Cleavage and Purification.

AMP-ELRs were designed and bioproduced in E. coli as propolypeptides (Table S1). After recombinant expression, the purity of the pro-AMP-ELRs was verified by SDS-PAGE (Figure S1). The sacrificial block was then removed. To that end, propolypeptides were incubated with CNBr solution (70% formic acid, FA) at a Met/CNBr molar ratio of 1:200. The reaction was performed for 20 h at room temperature in the darkness and under anaerobic conditions. CNBr was then eliminated on a rotary evaporator. The ELRs were resuspended in ultrapure water and dialyzed. After four dialysis steps against cold ultrapure water and lyophilization, the cleaved AMP-ELR was purified using HisPur Ni-NTA resin (ThermoFisher Scientific, USA) following a batch methodology. Briefly, the lyophilized products were dissolved in denaturing buffer (4 M urea, 20 mM sodium phosphate, and 500 mM NaCl) in order to prevent physical interactions between the AMPs, mixing 30 mL of the dialyzed solution with 15 mL of the resin in 50 mL tubes and incubating at 200 rpm and 4 °C for 3 h. The resin was then centrifuged. Because of the presence of the His tag, the sacrificial block and the uncleaved copolymers bonded to the resin, whereas the AMP-ELR remained in the supernatant. After two purification steps, AMP-ELRs were completely purified (Figure S2). Finally, the protein polymer solutions were dialyzed, filtered (0.22 μm Nalgene, ThermoFisher Scientific, USA), lyophilized, and stored at −20 °C until further use.

The monodispersity and purity of the hybrid AMP-ELRs were assessed by SDS-PAGE (Figures 1 and S2), matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) (Table S2 and Figure S3), and high-performance liquid chromatography (HPLC) (Table S3). MALDI-TOF and HPLC analysis were performed in the “Laboratorio de Técnicas Instrumentales” (LTI) at the University of Valladolid (Spain).

Figure 1.

(a) Molecular scheme of the modular design of the hybrid protein polymers (AMP-ELRs). Individual blocks (AMP, spacer and ELR) are not to scale. Additional information regarding the molecular weights can be found in the Supporting Information. (b) Copperstained SDS-PAGE of the pure recombinant products: SI ELR and the hybrid AMP-ELRs (GL13K-SI and 1018-SI). (c) Scheme of the thermally driven self-assembly of the ELR control, SI, into spherical micelles.

2.3. Phase-Transition Characterization.

The thermal behavior was evaluated by turbidimetry, measuring the optical density (OD) at 350 nm in the range 5–40 °C with a scan step of 0.5 °C using a Cary 100 UV-vis spectrophotometer (Agilent). OD or absorbance is defined as the logarithmic ratio of the intensity of light falling upon a material and the intensity transmitted. Heating and cooling ramps were performed at 0.25 °C min⁻¹ while stirring. T_t was determined as the temperature corresponding to the maximum of the first derivative of the OD versus temperature, and the thermal hysteresis was the difference between T_{t (heating)} and T_{t (cooling)}. All samples were prepared in ultrapure water at a concentration of 25 μM and measured in triplicates.

2.4. Physical Characterization of the Nanostructuration.

Because ELRs exhibit a LCST phase transition in solution, the self-assembly dynamics of the hybrid protein polymers (AMP-ELRs) were evaluated below and above the transition temperature (T_t) of the ELR. To that end, 25 μM solutions in ultrapure water were prepared under sterile conditions and incubated at 5 or 37 °C for 10 min, 1 h, 4 h, 1 d, 3 d, and 7 d. Nanostructuration of the ELR/AMP-ELRs was then analyzed by DLS and TEM.

Nanoparticle size distribution was evaluated by DLS using a Zetasizer Nano (Malvern Instruments, UK), with a 173° scattering angle equipped with a HeNe laser (633 nm) with an output power of 10 mW. Each sample was measured in triplicates.

TEM samples were prepared on 300-mesh carbon-coated copper grids with negative staining. First, grids were rendered hydrophilic by plasma treatment in a PDC-002 plasma cleaner (Harrick Plasma, USA) at a low power setting (7.2 W applied to the RF coil) for 20 s. Then, 15 μL of the preincubated polymers, ultrapure water, and uranyl acetate (1% w/v) solution were dropped onto a Parafilm strip over prechilled (5 °C) or preheated (37 °C) glass surfaces. Plasma-treated grids were placed onto the polymer drop for 90 s, on ultrapure water for 60 s, and, finally, on the negative staining solution for another 60 s. Blotting filter paper was used to remove excess solution after every step.

Images were taken using a Tecnai Thermionic T20 microscope operating at 200 kV (SAI, University of Zaragoza, Spain).

2.5. Circular Dichroism.

ELR or AMP-ELR solutions were prepared at 5 μM in prechilled ultrapure water and incubated at 5 or 37 °C for 10 min, 1 h, 4 h, 1 d, 3 d, and 7 d. The CD signal was measured for a 200 μL solution in a quartz cuvette (1 mm path-length) using a CD spectrometer (Jasco J-815, Easton, MD, USA), scanning over a range of 260–190 nm with a data pitch of 1 nm, a scanning rate of 50 nm min⁻¹, and a response time of 2 s. All measurements were subtracted from the background signal from ultrapure water in the quartz cuvette and repeated in triplicates.

2.6. Cryogenic TEM.

To test the thermoresponsiveness of the fibrillar structures formed by the AMP-ELRs, nanofibers were preformed, and the behavior of the nanostructures was evaluated by cryo-TEM. To that end, 25 μM solutions of the three protein-engineered polymers in ultrapure water were incubated at 5 °C for 24 h in order to drive the AMP fibrillar assembly. Samples were then heated at 37 °C for 30 min. Cryo-TEM samples were prepared before and after heating to evaluate changes in nanostructuration, as well as DLS measurements were performed.

Preparation and visualization of the cryo-TEM samples were carried out at the Electron Microscopy Platform (CICbioGUNE, University of the Basque Country, Spain). To that end, 4 μL of the sample was placed onto a glow-discharged 300-mesh lacey carbon-coated grid (Lacey Carbon film on 300 mesh copper; LC300-Cu; Electron Microscopy Sciences) and incubated inside the chamber of a Vitrobot Mark III (FEI Inc., The Netherlands) at 4 °C and at a relative humidity close to saturation (95% RH) for 30 s. Most of the liquid in the grid was removed by blotting (3 s at an offset of −3 mm) and vitrified by plunging into liquid ethane, previously cooled with liquid nitrogen at approximately −180 °C.

Images were collected at liquid nitrogen temperature using a JEM-2200FS/CR (JEOL Europe, Croissy-sur-Seine, France) field emission gun transmission electron microscope operating at 200 kV. An in-column energy filter (Omega filter) produced images with improved contrast and signal-to-noise ratio by zero-loss filtering. The energy slit width was set to 15 eV. Digital images were recorded using a 4 k × 4 k Ultrascan4000 charge-coupled device camera (Gatan, Inc.) running DigitalMicrograph (Gatan, Inc.) software.

3. RESULTS AND DISCUSSION

3.1. Molecular Design and Bioproduction of AMP-IDPP Protein Polymers.

Hybrid protein polymers were engineered and biosynthesized by recombinant DNA technology using a modular design in which two different domains can be differentiated (AMP and ELR, Figure 1a). We chose ELRs as the model IDPP because they mimic the physicochemical and biological properties of tropoelastin.³⁷ ELRs exhibit intrinsic molecular disorder, tunable LCST phase behavior, and biocompatibility. Moreover, ELRs have shown to be useful as purification tags for the expression of different proteins and peptides,³⁸ including AMPs,^39,40 in high yield, thus resulting in a variety of possibilities for supramolecular assembly.^41–43 In this approach, we employed an ELR, referred to as SI, with amphiphilic diblock design. The SI protein polymer (ELR control) contains the hydrophilic S-block [(Val-Pro-Gly-Ser-Gly)₅₀] and the hydrophobic I-block [(Ile-Pro-Gly-Val-Gly)₆₀], which have LCSTs below and above physiological temperature, respectively.⁴⁴ At 37 °C, the hydrophobic block (I) collapses into hydrophobic cores surrounded by a hydrophilic corona (S), thus driving the formation of micellar nanostructures (schematically represented in Figure 1c).

The AMP was located at the N-terminus, connected to the hydrophilic block (S) via a flexible poly-Gly spacer. This meant that both potential SADs, namely, the AMP and the hydrophobic block of the ELR, were located at opposite ends of the molecule, thus facilitating the identification of the self-assembly properties of the individual blocks. For AMPs, we chose the well-characterized designer peptides GL13K and 1018.^45,46 Although these AMPs have similar molecular properties (i.e. number of amino acids, charge, hydrophobicity, and hydrophobic moment),⁴⁷ they have shown different self-assembly properties in solution. While GL13K peptides self-assemble into defined twisted nanoribbons under basic conditions,²⁸ 1018 peptides tend to aggregate into nondefined and polydisperse nanostructures.⁴⁸

AMP-ELRs were produced as propolypeptides by introducing a sacrificial ELR block, referred to as HE, at the N-terminus of the AMP-ELR construct (Table S1). This block plays a key role during bioproduction because (a) it protects the host bacterial strain from the toxic side-effects of the AMP during fermentation; (b) it increases the expression levels; and (c) it enables site-specific cleavage. A Met was incorporated at the C-terminus to allow us to release the AMP-ELRs with no extra amino acid that may affect their bioactivities; and (d) it facilitates purification of the AMP-ELR with a designed histidine tag intended for selective removal of the HE block and the uncleaved products from the AMP-ELRs.

After recombinant production, ITC purification and chemical cleavage of the sacrificial block, AMP-ELRs were purified on a nickel-charged agarose resin (Figures S1 and S2). Monodisperse and highly pure products were obtained, as revealed by SDS-PAGE, MALDI-TOF mass spectrometry, and HPLC analysis (Figures 1b and S3 and Tables S2 and S3).

3.2. Thermal Behavior Characterization.

Given the thermosensitivity of the ELRs, the thermal behavior of the hybrid AMP-ELRs in aqueous solution was studied by monitoring the evolution of the OD at 350 nm (OD₃₅₀) during consecutive heating and cooling cycles in the range 5–40 °C of semidilute polymer solutions at 25 μM, above the critical aggregation concentration (Figures 2 and S5). The critical aggregation concentration of the ELR/AMP-ELRs was determined using surface tension measurements at physiological temperatures and was found to be around 2–4 μM (Figure S4), which is in accordance with previous reports in the literature.⁴⁹

Figure 2.

Thermal behavior of the hybrid AMP-ELRs (a,b) and the ELR (c) monitored by turbidimetry. (d) Images of the protein polymer solutions at 37 °C (25 μM in ultrapure water). Evolution of OD as a function of temperature demonstrated that all protein polymers exhibit a reversible phase transition. The presence of AMPs contributed to the phase transition and thermal hysteresis. Solid red lines represent heating cycles, and dashed blue lines represent cooling ones.

Turbidimetry analysis revealed that all three protein-engineered polymers (SI, GL13K-SI, and 1018-SI) showed a reversible LCST phase transition with thermal hysteresis and a transition temperature (T_t) below the physiological temperature (Table 1). Additionally, a slight increase in the absorbance (OD₃₅₀) after cooling the samples was observed. This may indicate that the process is not completely reversible. SI polypeptides were soluble below T_t. Upon increasing the temperature above T_t, collapse of the I-block triggered the formation of hydrophobic cores, which were stabilized in solution by surrounding hydrophilic coronas (S-blocks), thus meaning that SI self-assembled into micellar nanostructures.⁴⁴ When the samples were cooled, these micelles disassembled, and OD₃₅₀ decreased. However, a minimal fraction of the hydrophobic interactions between isoleucine side-chains seemed to remain, thus meaning that disassembly was not completely reversible. Consequently, OD₃₅₀ was slightly greater than prior to heating, progressively increasing after each consecutive heating-cooling cycle (Figure S5). Furthermore, and consistent with previous studies with the IDPP analogue poly(Val-Pro-Gly-Leu-Gly),⁵⁰ thermal hysteresis was observed during cooling cycles, possibly as a result of these hydrophobic interactions.

Table 1.

Transition Temperatures (T_t) and Hysteresis (ΔT_t) of the ELR (SI) and the Hybrid AMP-ELRs (GL13K-SI and 1018-SI)

	SI	GL13K-SI	1018-SI
Tt heating	23.6 ± 0.4	19.6 ± 0.5	17.5 ± 0.6
Tt cooling	21.7 ± 0.3	13.0 ± 0.6	13.1 ± 0.4
hysteresis (ΔTt)	1.9 ± 0.3	6.6 ± 0.3	4.5 ± 0.4

Similarly, hybrid AMP-ELRs also underwent a quasi-reversible phase transition, although with substantial differences (Figure 2). SI self-assembles into stable nanostructures in solution, and as a consequence, the OD₃₅₀ is less than 0.05. In contrast, for the hybrid AMP-ELRs, OD₃₅₀ was much higher at temperatures above T_t. As observed previously for AMPs in solution,^28,48 AMP domains can self-interact, and AMP-ELRs may assemble into larger aggregates, thus increasing OD₃₅₀ (Figure 2). Moreover, divergences were observed between the turbidity profiles of the hybrid protein polymers. Whereas 1018-SI showed a thermal behavior similar to the ELR control (SI), GL13K-SI underwent a sharper transition with a peak, which suggested that GL13K tendency to fold may alter the assembly process of the ELR.

The presence of AMP affected the phase transition when compared with the SI polymer (Table 1). The T_t of the hybrid polymers decreased in comparison with the ELR control, which indicates that the noncovalent interactions between the AMPs favor the phase transition of the ELR. For the GL13K-SI and the 1018-SI polymers, T_{t (heating)} was decreased by an average of about 4 and 6°, respectively, as compared to the ELR control. As such, these results suggest differential aggregation mechanism of the AMPs. 1018 peptides are likely to display a greater tendency to aggregate than GL13K peptides. The differences between the assembly mechanisms of the AMPs are discussed below (Section 3.3).

The AMP domains also contributed to decrease the T_{t (cooling)}, thereby enhancing thermal hysteresis (Table 1). Aggregation of the GL13K and 1018 peptides is associated with the formation of secondary β-structures,⁴⁷ which would increase the intermolecular order in the coacervate state and, therefore, the hysteresis. This behavior is consistent with previous studies in which order-promoting domains (e.g., poly-Ala) were introduced into the IDPP backbone.²⁰ Besides that, coacervation of the ELR-domain seemed to promote the assembly of the AMPs. After heating-cooling cycles, the OD₃₅₀ of the AMP-ELR samples gradually increased (Figure S5a,b), and this increase was substantially greater compared with the control (SI), which also showed the synergy between the AMP and the ELR domains in the assembly of the hybrid protein polymers.

3.3. Self-Assembly Dynamics.

Phase-transition characterization suggested that the AMP domains may play an important role in the supramolecular assembly of AMP-ELRs by contributing to the formation of aggregates of higher order complexity than micellar assemblies and stabilization of the assembled structure, thus increasing thermal hysteresis. As such, we proceeded to characterize the self-assembly dynamics below and above T_t of the I-block in detail.

At 5 °C, a temperature well below T_t, ELR molecules were completely soluble, and the ELR control (SI) did not form any nanostructures (Figures 3 and 4a). DLS distribution of around 10 nm corresponds to the value for soluble macromolecules.⁵¹ In contrast, the presence of AMPs within the recombinant polymers triggers self-assembly into fibrillar nanostructures. During DLS characterization of the nanostructures in situ, it was observed larger size distributions in the hybrid AMP-ELR samples than in the SI polymer samples (Figures 4a and S6). Besides that, despite the intensity size distributions of both hybrid polymers overlapped, and the nanostructures formed by the 1018-SI polymer were bigger and more polydisperse than those formed by GL13K-SI (Table S4).

Figure 3.

Negatively stained TEM micrographs of the ELR/AMP-ELRs after incubation at 5 °C. Presence of the AMPs drove the formation of nanofibers after short incubation periods (10 min, 1 h) which evolved over time, thus indicating a dynamic behavior.

Figure 4.

DLS intensity distributions of the three protein-engineered polymers below [(a) 5 °C] and above [(b) 37 °C] the T_t.

TEM imaging confirmed these differences (Figure 3). After 10 min of incubation, nanofibers with lengths of 27.0 ± 16.5 and 118.2 ± 68.6 nm appeared in the GL13K-SI and 1018-SI polymer samples, respectively (Table S4). The fibrillar nanostructures formed by the AMP-ELRs exhibited a dynamic behavior in solution with substantial differences in size and shape of the nanostructures as a function of the AMP sequence (Figure 3, middle and right column and Table S4). Although the nanofibers evolved over time in both cases, different aggregation patterns could be observed for each hybrid protein polymer. 1018-SI formed longer fibrils (506.6 ± 425.9 nm) than GL13K-SI (314.7 ± 220.8 nm) after one-week incubation. The growth of GL13K-SI nanofibers seems to be spatially constrained, thus limiting fiber elongation in favor of nanofibers with repeated patterns. TEM also revealed differences in contrast of the protein-polymer fibrils. The capability of the GL13K peptide to self-assemble into nanofibril and nanoribbon structures in solution²⁸ seems to be associated with a higher packing density of the protein polymers in GL13K-SI nanofibers than in 1018-SI nanostructures.

Supramolecular assembly of the peptides 1018 and GL13K has been previously observed in solution,^28,48 surfaces, and model membranes,^47,52,53 although they require alkaline conditions or the presence of salts to self-assemble. Electro-static repulsion between the positively charged side-groups has to be overcome to favor interaction between the peptides and their organization into supramolecular assemblies through the formation of β-structures. Interestingly, in our case, self-assembly of the hybrid protein polymers occurred in salt-free solution. Both AMPs seemed to behave similarly to other peptide domains that induce the formation of β-structures in protein polymers.^54,55 Their tendency to acquire configurations with high proportions of β-sheets⁴⁷ promotes the formation of fibrils in block polypeptides that grow over time.

In addition to DLS and TEM characterization, structural studies of the hybrid protein polymers were performed using CD spectroscopy (Figure S7). Nevertheless, the CD spectra for all polymers were very similar at the different time points below T_t, with a negative peak at ≈197 nm and a weak positive shoulder near 220 nm that are characteristic for unordered polypeptides.⁵⁶ Although the overwhelmingly higher molecular weight of the ELRs compared to the AMP in the ELR-AMP molecules (46 kDa vs ≈1.5 kDa) dominated the CD signal, a decrease in ellipticity of the 197 nm peak (random-coil peak) was detected, probably due to the self-assembly of the AMPs.

In parallel, we also studied the nanostructuration at physiological temperature and found that, above T_t, the I-block underwent a phase transition, thus leading to the formation of defined nanostructures. These micellar assemblies were observed in the three protein polymers after short incubation periods (10 min, 1 h, and 4 h). However, in the hybrid AMP-ELRs, micellar populations coexisted with small nanofibers similar to those observed below T_t, which may be a result of the early assembly of the AMPs (Figure 5).

Figure 5.

Negatively stained TEM micrographs of the ELR/AMP-ELRs after incubation at 37 °C. The presence of the AMP drives a second self-assembly, which triggers the formation of hierarchical structures. Fibrillar aggregates with globular or amorphous shapes are found when the GL13K or the 1018 peptide, respectively, are found within the hybrid polypeptide.

The assembly of the elastin-like hydrophobic block correlated with shift in the secondary structure (Figure S7). In the three CD spectra above T_t appeared a lower proportion of random coil (minimum at 197 nm) and a higher proportion of type II β-turns and distorted β-sheets conformations (maximum at ≈210 nm and pronounced minimum at 220 nm) comparing with the spectra below T_t, which is characteristic in ELRs.^57,58 In addition, the presence of the AMP in the hybrid polymers resulted in a substantial decrease in the intensity of the random coil peak (197 nm), thus confirming that the early aggregation of the AMPs induced an increase of the structural order in the hybrid polymers comparing with the SI polymer. It must also be noted that after one day of incubation at 37 °C, CD signals of 1018-SI decreased significantly, probably due to the phase separation and the consequent precipitation of the aggregates.

The SI polymer self-assembled into stable and monodisperse micellar nanostructures (D_h = 41.0 nm, PdI = 0.125, after 1 week at 37 °C). In contrast, the presence of the AMP in the hydrophilic corona induced the aggregation of the micelles, thus driving a secondary self-assembly in the hybrid protein polymers (Figure 4b). Given that the ELRs are IDPPs, they retain a high degree of disorder in the coacervate state and water, allowing chains to assemble into transient structure conformations above T_t.^6,11 As such, the interaction of the AMP domains and subsequent formation of stable ordered structures (β-sheets) impaired the micellar assemblies, leading to the formation of hierarchical assemblies based on interconnected fibers (Figure 5). After one incubation day above T_t, AMP-ELRs formed fibrillar aggregates with a mean length of 327.0 ± 117.1 nm and 293.9 ± 170.9 nm for GL13K-SI and 1018-SI, respectively (Table S5). Consistent with the characterization below T_t, the self-organization of the AMP-ELRs varied depending on the AMP, with different shapes being observed. GL13K-SI preferably formed spherical aggregates while 1018-SI aggregates were more elongated with undefined shapes (Figures 4b, 5, and Table S5). Moreover, fibrillar aggregates also exhibited dynamic behaviors. Because both SADs were located at opposite ends of the polymer chains, the aggregates showed a continuous growth until precipitation (Figure S6 and Table S5).

Both AMPs have similar molecular weights and molecular properties (Figure 6), which may suggest similar self-assembly mechanisms, even though their self-assembly dynamics are substantially different. In the case of 1018, a stretch of the hydrophobic residues between Leu-3 and Val-7 seems to be crucial to trigger the aggregation of the peptide.⁴⁸ In contrast, the spatial disposition of the hydrophobic and hydrophilic amino acids in the GL13K is decisive for the supramolecular self-assembly into twisted nanoribbons. Recent studies demonstrated that the randomization of the sequence or even a single modification of charged or hydrophobic residues impaired the ability of the peptides to self-assemble into supramolecular structures.^28,59 Consequently, the different spatial distribution of the charged and hydrophobic residues in the AMPs also modulates the driving forces for peptide aggregation. In this regard, although both mechanisms of aggregation are pH-dependent, the tendency to acquire β-sheet conformations in solution is greater in 1018 peptides than GL13K at the same pH.⁴⁷ GL13K self-assembly requires pH values of 9.6 or greater,^28,59 whereas the tendency of 1018 to aggregate is noticeable from pH values greater than 2 and is highly influenced by the presence of anions.⁴⁸

Figure 6.

Molecular structures and molecular weights (M_W) of the GL13K and 1018 peptides. Ionic residues are marked in red and large hydrophobic residues are marked in green.

Therefore, the diverse self-assembly properties of both AMPs have a direct impact in the nanostructuration of the hybrid protein polymers (GL13K-SI and 1018-SI), as we observed by DLS and TEM imaging. The C-terminal conjugation of the 1018 peptide with the IDPP chain seemed to contribute to enhance the spatial self-organization of the peptide, which enabled the formation of large fibrillar nanostructures below T_t (Figure 3 and Table S4) and to guide the supramolecular assembly of the AMP-ELR toward fibrillar networks above T_t. In contrast, GL13K peptide tendency to self-assemble into well-defined twisted nanoribbons in solution²⁸ seemed to be hampered in the hybrid polymer because of steric effects, thus resulting in the formation of nanostructured patterns below T_t (Figure 3). Likewise, this effect seemed to contribute to the aggregation of GL13K-SI into spherical fibrillar aggregates above T_t (Figure 5).

3.4. Thermoresponsive Nanostructures.

Additionally, we evaluated whether the fibrillar nanostructures formed upon self-assembly of the AMPs retained the LCST behavior of the ELRs. To that end, we incubated the ELR/AMP-ELR solutions at 5 °C for 24 h to preform the nanofibers and then carried out a second incubation at physiological temperatures for 30 min to assess the influence of the temperature on nanostructuration. Samples were analyzed via DLS and cryogenic transmission electron microscopy (cryo-TEM) visualization after each incubation (Figures 7 and S8).

Figure 7.

(a) Intensity size distributions of the nanostructures formed by the hybrid protein polymers after the incubation below (5 °C) and above (37 °C) T_t. (b) Cryo-TEM micrographs of the GL13K-SI and 1018-SI samples after initial incubation at 5 °C for 24 h, where the AMP triggered fibrillar assembly, and (c) after subsequent incubation at 37 °C, where thermally triggered coacervation of the ELR drove aggregate formation. (d) Schematic representation of the hierarchical self-assembly of the hybrid polymers (AMP-ELR) and magnification of the nanostructures formed at physiological temperatures after the incubation at 5 °C.

As expected, incubation below T_t allowed AMP domains to drive the supramolecular assembly of the hybrid polypeptides into nanofibers (Figure 7a,b). A subsequent second incubation at physiological temperatures induced coacervation of the hydrophobic block of the ELR (I-block), thus exposing the hydrophilic ELR block (S-block). Nanofibers aggregated into fibrillar networks with spherical or undefined shapes, depending on the AMP (Figure 7c). In both cases, the AMPs seemed to be hidden inside the fibrillar network. To assess this hypothesis, we performed minimal inhibitory concentration (MIC) assays against Gram-positive Streptococcus gordonii and Gram-negative Pseudomonas aeruginosa to verify if the AMPs were exposed on the surface of the nanostructures. Both AMPs have demonstrated that they retain their bactericidal potential when they are folded and immobilized on surfaces.⁴⁷ As such, the lack of antimicrobial activity of the hybrid polypeptides may indicate that the AMPs were buried inside the nanostructures (Table S6).

We have demonstrated both the ability of AMPs to self-assemble and induce the supramolecular structuration of larger protein polymers and that the thermosensitivity of the ELRs is maintained in the fibrillar assembly. Furthermore, these results suggest that AMPs act as order-promoting domains in the hybrid protein polymers, thus inducing thermal hysteresis and controlling the supramolecular assembly of the ELR domain below and above T_t. Recent studies have shown that ordered domains highly modulate the hierarchical assembly of the IDPPs into complex architectures.^20,60 In this regard, the synergistic combination of AMPs and IDPPs opens up a wide range of possibilities for the fabrication of hierarchical nanomaterials with advanced functionalities for biotechnology and biomedical engineering.

Lastly, it is important to note that we used this modular approach in order to recognize the identity and properties of the individual building blocks in the hybrid construction. However, further re-engineering of the design would be needed for their future application. As such, our system can be easily modified to enable optimization for the intended application, thanks to its recombinant nature. In light of these results, the AMP domains were likely to be hidden inside the fibrillar nanostructures. Although this effect hinders the antimicrobial activity of the assembly, it might be useful for designing nanovehicles or nanoreservoirs that protect the AMP from the environment or physically crosslinked three-dimensional scaffolds for tissue engineering applications. In fact, the stimuli responsiveness of the ELR could be used to switch the condensation of the nanofibers, thereby increasing the local AMP concentration. This could allow the manufacture of switchable nanomaterials with potential applications as nanostructured reservoirs for AMP delivery. In addition, it is noteworthy that the presence of the spacer between the AMP domain and the ELR diblock does not compromise hierarchical assembly. Thus, the introduction of functional spacers, including sequences sensitive to biological or physical stimuli (e.g. protease degradation or pH), would facilitate the controlled release of the AMPs, thereby enabling the development of smart AMP nanocarriers. Moreover, hydrophobic collapse of the elastin-like domain could be used for the encapsulation of hydrophobic drugs,⁶¹ including conventional antibiotics, antioxidants, fluorophores, or anticancer drugs.

4. CONCLUSIONS

The combined use of AMPs with stimuli-responsive protein polymers is a promising strategy for the design of self-assembled nanomaterials for biotechnological and biomedical applications. We have shown that AMPs can be used as SADs to trigger the assembly of larger IDPPs, with different nanostructures being achieved depending on the AMP. Moreover, their combination with thermoresponsive protein polymers in modular designs enables the manufacture of hierarchical architectures formed by a dual assembly process.

Consequently, our nanosystem represents a sound strategy for the fabrication of smart biomaterials incorporating AMPs. Their recombinant nature facilitates edition of the modular design and the incorporation of other bioactive motifs with extreme control, in addition to a scalable method for their sustainable production and potential widespread use. This investigation provides a new insight into the protein engineering of self-assembling materials that combine the properties of IDPPs and AMPs in a synergistic manner.