Protein- and peptide-based proton-conducting biomaterials have been touted as particularly promising for bioelectronics applications because of their advantageous chemical and physical characteristics, typically excellent biocompatibilities, and readily understood electrical properties. Within this context, our laboratory has previously discovered and systematically investigated bulk proton conduction for a unique family of cephalopod structural proteins called reflectins. Herein, we leverage a combination of experimental and computational methodologies to investigate the bulk electrical properties of hierarchically nanostructured films self-assembled from a previously reported truncated reflectin variant. Our findings indicate that the truncated reflectin variant exhibits protonic conductivities and associated figures of merit on par with those reported for both full-length reflectins and other proteinaceous proton-conducting materials. The combined studies enhance current understanding of reflectins’ functional properties within the framework of bioengineering and bioelectronics applications and may ultimately facilitate the development of other protein- and peptide-based conductive biomaterials.
INTRODUCTION
Protein- and peptide-based proton-conducting biomaterials have been touted as particularly promising for bioelectronics applications because of their numerous advantageous characteristics, including well-defined amino acid sequences, amenabilities to chemical modification, controllable self-assembly properties, straightforward processabilities, and excellent biocompatibilities.1–6 Within this context, nanostructured films from such biomaterials have generally exhibited excellent electrical properties because of their formation of internal conduits that support effective long-range proton transport.1–20 For example, bovine serum albumin (BSA) mats consisting of interconnected fibers have shown proton conductivities of ∼0.04 to ∼0.3 mS/cm because of proton translocation mediated by the fibers’ oxidized amino acids and trapped water molecules.7,8 In addition, cyclic octapeptide films consisting of self-assembled overlapping nanotubes have featured proton conductivities of ∼0.2 to ∼0.3 mS/cm because of proton transport mediated by the nanotube octapeptides’ side chains and intervening water molecules.9,10 Moreover, tandem-repeat squid ring teeth protein membranes containing amorphous and crystalline nanoscale domains have possessed proton conductivities of ∼0.5 to ∼2 mS/cm because of proton hopping through the domains’ hydrogen bonded networks.11,12 Last, hybrid tyrosine-rich peptide films/manganese oxide composites containing inorganic nanoparticles in peptidic matrices have demonstrated mixed electron–proton conductivities of ~0.01 to ~18.6 mS/cm in part because of electron transfer-coupled proton migration between oxidized amino acids.13,14 These and other seminal studies have reinforced the promise and potential of proteinaceous biomaterials for bioelectronics applications while helping establish broadly relevant design guidelines for other biomolecular proton conductors.7–20
Within the area of nanostructured protein- and peptide-based biomaterials, our laboratory initially discovered and then systematically investigated bulk proton conduction in unique cephalopod structural proteins called reflectins.2,21–27 These proteins, which are well known for their critical roles in light-reflecting cephalopod skin cells, possess unusual amino acid sequences that contain multiple conserved repeating domains and feature high aromatic, charged, and polar amino acid contents.28–35 Due to their unique primary sequences, both reflectins and reflectin-like polypeptides spontaneously self-assemble into spheroidal nanoparticles and can be readily processed into nanostructured thin films under various conditions.27,29–31,33–36 Such nanostructured films formed from either reflectins or reflectin-like polypeptides exhibit proton conductivities of ∼0.07 to ∼0.4 mS/cm at room temperature because of Grotthuss-type proton transport among charged amino acids within hydrophilic nanoscale domains.21,23,26,27 Accordingly, the reported favorable electrical characteristics have made reflectin variants suitable as active materials for protonic transistors, photochemically dopable platforms, and protochromic devices.21,22,24,25 Moreover, reflectin-based thin films and nanostructures have exhibited excellent biocompatibilities, as verified by their ability to support human and murine neural stem cell differentiation and their utility as high contrast markers for visualizing or tracking mammalian cells.37–41 When considered together, the aforementioned studies have suggested that reflectins constitute promising biomaterials for further research and development from the perspective of bioelectronics and bioengineering applications.
Here, we describe proton conduction in nanostructured films self-assembled from a previously reported Doryteuthis pealeii reflectin A1 truncated variant (RfA1TV), as illustrated in Fig. 1.35 Initially, we comparatively investigate the concentration-dependent structural characteristics of monomeric and self-assembled RfA1TV ensembles in solution, as well as study the morphologies and structural characteristics of nanostructured films fabricated from the self-assembled RfA1TV nanoparticles. Subsequently, we interrogate the electrical properties of our nanostructured RfA1TV films with a combination of alternating current (AC) and direct current (DC) electrical measurements. Finally, we probe the molecular-level arrangements of our monomeric and self-assembled RfA1TV ensembles with both computational and experimental methods. Our findings enhance current understanding of reflectins’ reported functional properties and may also help guide the development of other protein- and peptide-based conductive biomaterials.
RESULTS AND DISCUSSION
We initially comparatively investigated monomeric and self-assembled RfA1TV ensembles in solution. For this purpose, we prepared aqueous solutions containing RfA1TV at low and high concentrations and then characterized them with dynamic light scattering (DLS) and circular dichroism (CD) spectroscopy [see, for example, Fig. 2(a) and the supplementary material for details].27,34,35,40 For monomeric RfA1TV, the DLS volume distributions revealed a primary peak at typical RH values of ∼1 to ∼2 nm, which corresponded to a single dominant nanoparticle population [Fig. 2(b)]. The protein’s corresponding CD spectra exhibited a main negative peak at ∼200 nm followed by smaller negative shoulders, indicating a partially ordered secondary structure containing a ∼0.10 helical fraction and a ∼0.56 beta structure fraction [Fig. 2(c)]. For self-assembled RfA1TV, the DLS volume distributions revealed peaks at typical RH values of ∼10 to ∼900 nm, which corresponded to multiple distinct nanoparticle populations [Fig. 2(b)]. The protein’s corresponding CD spectra exhibited two negative peaks between ∼203 and ∼215 nm, indicating a more ordered secondary structure containing a ∼0.37 helical fraction and a ∼0.41 beta structure fraction [Fig. 2(c)]. Notably, the concentration-dependent size distribution and secondary structure trends observed for RfA1TV were similar in water and ammonium formate buffer [compare Figs. 2(b), 2(c), and S1] but were quite different from the previously reported analogous mechanical agitation-dependent trends at low concentrations.35 These findings provided insight into the evolution of RfA1TV’s structural characteristics upon concentration increase-induced self-assembly into nanoparticles in solution.
We next investigated the morphologies of films processed from RfA1TV nanoparticles. For this purpose, we prepared continuous films from high-concentration RfA1TV solutions and then characterized the films with grazing incidence small angle x-ray scattering (GISAXS), atomic force microscopy (AFM), and solid state circular dichroism (ssCD) spectroscopy (note that films prepared from low-concentration RfA1TV solutions were typically discontinuous and, therefore, were challenging to characterize) [see, for example, Fig. 2(d) and the supplementary material for details].27,34,35,40 First, the GISAXS scattering intensity data collected and computationally analyzed for the RfA1TV films indicated a radius of gyration (Rg) value of ∼170 nm and a Porod exponent (P) of ∼2, which were consistent with the presence of large ellipsoidal nanoparticles [Fig. 2(e)]. Second, the AFM imaging performed for continuous RfA1TV films confirmed that their surfaces consisted of coalesced nanoparticles with variable sizes of hundreds of nanometers (Fig. S2). Third, the ssCD spectra obtained for the films revealed a positive peak at ∼195 nm and two negative peaks at ∼208 and ∼222 nm, and therefore, indicated an ordered secondary structure containing a ∼0.43 helical fraction and a ∼0.35 beta structure fraction [Fig. 2(f)]. Here, the sizes of the RfA1TV nanoparticles comprising the films were consistent with those of nanoparticles dispersed in solution [Fig. 2(b)] or cast onto solid substrates (Fig. S3). Moreover, the secondary structure of the RfA1TV biomolecules comprising the nanoparticles remained comparable in solution [Fig. 2(c)] and within films [Fig. 2(f)]. These findings showed that self-assembled RfA1TV nanoparticles readily formed films with a distinct nanoscale order and that the nanoparticles' constituent biomolecules generally maintained their secondary structure in the solid state.
We next interrogated the electrical properties of the RfA1TV films formed from our nanoparticles via alternating current (AC) electrical measurements. Toward this end, we fabricated two-terminal devices consisting of continuous nanostructured RfA1TV films contacted with gold electrodes and characterized such devices with electrochemical impedance spectroscopy (EIS) in the presence of water (H2O) or deuterium oxide (D2O) vapor at a relative humidity (RH) of ∼90% [see, for example, Figs. 3(a), S4, and the supplementary material for details].21,23,26,27 The representative Nyquist plots recorded for the RfA1TV-based devices revealed semi-circles and inclined spurs corresponding to bulk film and interfacial film-electrode impedances, respectively, for both H2O and D2O vapor, as expected for proton-conducting materials [Fig. 3(b)].2,7,8,21,23,26,27,42,43 The representative Nyquist plots recorded for the RfA1TV-based devices revealed that their average conductivities decreased by ∼34% from 0.29 ± 0.08 to 0.19 ± 0.04 mS/cm upon moving from H2O to D2O, with this kinetic isotope effect suggesting that proton transport occurred via the Grotthuss mechanism [Fig. 3(b)].2,5,10,21,23,26,27,44 Here, the Nyquist plots were reasonably modeled with a standard equivalent circuit validated for proton-conducting materials, wherein an RC element represented the bulk film and a constant phase element represented the film-electrode interface (Fig. S5).21,27,42,43 Moreover, the electrical characteristics and conductivities measured for RfA1TV were similar to those previously reported for full-length reflectins and reflectin-like polypeptides in analogous device configurations under comparable conditions (Table S1).21,23,26,27 The measurements together indicated that nanostructured RfA1TV films supported proton transport in the solid state.
We in turn interrogated the electrical properties of RfA1TV films formed from our nanoparticles via direct current (DC) electrical measurements. Toward this end, we fabricated two-terminal devices consisting of continuous nanostructured RfA1TV films contacted with palladium (Pd) electrodes and characterized the devices via current–voltage (I–V) measurements when the electrodes were converted in situ from Pd to palladium hydride (PdHx) at a RH of ∼90% [see, for example, Figs. 3(c), S6, and the supplementary material for details].21–24,26 The representative I–V curves recorded for the RfA1TV-based devices contacted with proton-blocking Pd electrodes revealed slight hysteresis and current densities of 3.0 ± 0.3 mA/cm2 at 1.5 V [Figs. 3(d) and S7]. The representative I–V curves recorded for the RfA1TV-based devices contacted with proton-injecting PdHx electrodes also revealed slight hysteresis but much higher current densities of 65 ± 22 mA/cm2 at 1.5 V [Figs. 3(d) and S7]. Here, the substantial ∼20-fold increase in the current densities upon transitioning from proton-blocking to proton-injecting electrical contacts closely matched previous observations for various biomolecular proton-conducting materials [Fig. 3(d)].13,15–19,21–24,26,45 Moreover, the electrical characteristics and current densities measured for RfA1TV were similar to those previously reported for full-length reflectins in analogous device configurations under comparable conditions (Table S1).21,23,26 The measurements together confirmed that nanostructured RfA1TV films supported proton transport in the solid state.
We subsequently experimentally probed the molecular-level structure of monomeric and self-assembled RfA1TV in vitro. To achieve this goal, we prepared solutions containing 13C- and 15N-labeled RfA1TV at low and high concentrations and then characterized the labeled RfA1TV with two-dimensional (2D) heteronuclear single quantum coherence (1H–15N HSQC) nuclear magnetic resonance (NMR) spectroscopy [see, for example, Fig. 4(a) and the supplementary material for details].35 The 1H–15N HSQC NMR spectra recorded for monomeric and self-assembled RfA1TV revealed different cross-peaks between 15N chemical shifts of ∼105 to ∼125 ppm and 1H chemical shifts of ∼7 to ∼9 ppm, which could be readily assigned to the individual amino acids [Fig. 4(b)].35 Here, the chemical shifts recorded for specific amino acids upon RfA1TV’s self-assembly into nanoparticles displayed average perturbations Δδ(H,N) = ((ΔδH)2 + (0.154 * ΔδN)2)1/2 with magnitudes as large as ∼0.12 ppm, indicating a concentration increase-induced rearrangement of the protein’s secondary structure [Fig. 4(c)]. Interestingly, the larger chemical shift perturbations found for some charged amino acid residues, e.g., aspartic acid, suggested their participation in local hydrogen bonding or electrostatic interactions within nanoparticles [Figs. 4(a)–4(c)]. Moreover, the pronounced chemical shift perturbations found for RfA1TV’s amino acid tracts with a predisposition toward forming alpha helices suggested substantial structural changes and a transition from beta to alpha helical character during self-assembly [Figs. 4(a)–4(c)], in good agreement with our CD observations [Figs. 2(c), 2(f), and S1]. These experiments afforded molecular-level insight into the global ordering that occurred for RfA1TV upon concentration increase-induced self-assembly.
We last computationally probed the molecular-level arrangements of monomeric and self-assembled RfA1TV ensembles in silico. To achieve this goal, we performed all-atom molecular dynamics (MD) simulations for 27 RfA1TV biomolecules starting from their experimentally determined monomeric NMR structures and analyzed the evolution of these proteins’ aggregation states and secondary structural characteristics [see, for example, Fig. 4(d) and the supplementary material for details].35,40 The structural snapshots and particle size distributions calculated for a representative RfA1TV ensemble indicated that the monomeric proteins (Rg of ∼1 nm at 0 ns) formed multiple dimers and oligomers (Rg’s of ∼2 and ∼3 nm at 10 ns) before merging into a single large aggregate (Rg of ∼7 nm at 40 ns), which subsequently underwent some compaction (Rg of ∼5 nm at 200 ns) [Figs. 4(d) and 4(e)]. This multi-stage assembly process appeared driven by the reduction in the total system potential and protein–protein interaction energies for the system (Fig. S8). The structural snapshots and changes in the total fractions of secondary structure calculated for a representative RfA1TV ensemble indicated that the individual proteins generally transitioned from less ordered to more ordered conformations [Figs. 4(d), 4(f), and S9]. This structural transition was evidenced by a notable increase in the fraction of helices (as large as +0.33) and a concomitant decrease in the fraction of beta structures (as large as −0.30) for the ensemble [Figs. 4(d), 4(f), and S9], in close overall agreement with our CD and NMR observations [Figs. 2(c), 2(f), Fig. S1, and Figs. 4(b), 4(c)]. These simulations confirmed and reinforced our molecular-level understanding of the structural changes that occurred for RfA1TV ensembles during concentration increase-induced self-assembly.
CONCLUSION AND SIGNIFICANCE
In summary, we have demonstrated the concentration-dependent self-assembly of RfA1TV into nanoparticles, characterized the morphological and structural characteristics of films fabricated from the RfA1TV nanoparticles, and investigated the electrical properties of such nanostructured RfA1TV films, with our combined observations holding significance for several reasons. First, our synergistic experimental and computational methods consistently indicate that the self-assembled RfA1TV comprising our nanoparticles features altered secondary structures, i.e., shows an increase in the fraction of helices and a corresponding decrease in the fraction of beta structures (relative to monomeric RfA1TV). These findings afford additional insight into RfA1TV’s previously predicted ability to sample a broad conformational landscape and thus could ultimately facilitate the improved design of nanostructured reflectin-based optical systems.35,40,41 Second, the described RfA1TV films processed from the self-assembled nanoparticles exhibit not only global nanoscale-level morphological ordering, i.e., packing of their constituent nanoparticles, but also local molecular-level ordering, i.e., prominent secondary structure changes for some charged amino acids. These findings are consistent with the broader widely reported paradigm that hierarchical multi-length-scale organization promotes effective proton transport in materials derived from either reflectins or other biomolecules,7–27 thus providing added insight into reflectins’ solid state structure-electrical function relationships.21,34 Third, our nanostructured RfA1TV films surprisingly feature electrical figures of merit, i.e., bulk proton conductivities and associated current densities, which are comparable not only to those reported for full-length reflectins and reflectin-like polypeptides but also to many other proteins or peptides.7–14,21–27 These findings suggest that our minimal reflectin variant constitutes a promising and tractable template for the design of multifunctional reflectin-inspired conductive biomaterials that could be capable of interfacing with living systems.2,37,38 Altogether, our findings improve current understanding of reflectins’ functional properties within the framework of bioengineering and bioelectronics applications and may more generally help advance ongoing efforts in the development of protein- and peptide-based conductive biomaterials.
SUPPLEMENTARY MATERIAL
See the supplementary material for the detailed experimental procedures, atomic force microscopy images of nanoparticles and films from RfA1TV, optical microscopy images of devices from RfA1TV films contacted with gold and palladium electrodes, computational fits of the electrochemical impedance spectroscopy data obtained for RfA1TV films, expanded views of the current–voltage characteristics obtained for RfA1TV films, the evolution of the total system potential and protein–protein interaction energies for the RfA1TV ensembles, and the changes in the fractions of total secondary structure, helices, and beta structures for the RfA1TV ensembles.
ACKNOWLEDGMENTS
The authors are grateful to the Air Force Office of Scientific Research (Grant No. FA9550-17-1-0024 to A.A.G.), the Office of Naval Research (Grant No. N00014-20-1-2265 to A.A.G.), and the Army Research Office (Grant No. W911NF-23-1-0052 to Y.G.Y.) for financial support. The authors are grateful to the Slovenian Research and Innovation Agency (Grant No. P1-0242 to J.P.) for financial support. The authors acknowledge the Laser Spectroscopy Labs at the University of California, Irvine, for providing access to experimental facilities. The authors acknowledge the CERIC-ERIC Consortium for access to experimental facilities and financial support. The authors acknowledge North Carolina State University’s Office of Information Technology for access to its Henry2 high performance computing cluster.
AUTHOR DECLARATIONS
Conflict of Interest
The authors have no conflicts to disclose.
Author Contributions
A.A.G., J.P., and Y.G.Y. conceived and supervised the project; P.P. and A.C. expressed, purified, and characterized unlabeled RfA1TV; P.P. expressed, purified, and characterized 13C- and 15N-labeled RfA1TV; P.P., N.K., and A.P. performed and analyzed the DLS and CD measurements; M.K. performed and analyzed the NMR spectroscopy experiments; A.L.K. performed and analyzed the MD simulations with assistance from A.P.; K.L.N., B.S., B.M., and S.B. performed and analyzed the GISAXS experiments; P.P. fabricated the RfA1TV-based devices contacted with gold electrodes, characterized the devices with optical microscopy and AFM, and performed the EIS experiments and modeling with assistance from N.E.T.; R.K., M.K.D., and P.P. fabricated the RfA1TV-based devices contacted with palladium electrodes, characterized the devices with optical microscopy and AFM, and performed the I–V measurements; A.A.G. and P.P. wrote the paper with input from all of the other co-authors.
Preeta Pratakshya: Data curation (lead); Formal analysis (lead); Investigation (lead); Methodology (lead); Visualization (lead); Writing – original draft (lead); Writing – review & editing (equal). Albert L. Kwansa: Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Matic Kovačič: Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Writing – original draft (equal); Writing – review & editing (supporting). Nikhil Kaimal: Formal analysis (supporting); Investigation (supporting); Methodology (supporting); Writing – original draft (supporting); Writing – review & editing (equal). Arsenii Panteleev: Formal analysis (supporting); Investigation (supporting); Visualization (supporting); Writing – original draft (supporting); Writing – review & editing (equal). Atrouli Chatterjee: Investigation (supporting). Nadia E. Tolouei: Investigation (supporting). Rylan Kautz: Investigation (supporting). Kyle L. Naughton: Investigation (supporting). Barbara Sartori: Investigation (supporting). Benedetta Marmiroli: Investigation (supporting). MyAnh K. Dao: Investigation (supporting). Sigrid Bernstorff: Formal analysis (supporting); Investigation (supporting); Writing – review & editing (supporting). Janez Plavec: Conceptualization (equal); Formal analysis (supporting); Funding acquisition (lead); Methodology (supporting); Writing – review & editing (supporting). Yaroslava G. Yingling: Conceptualization (equal); Formal analysis (supporting); Funding acquisition (lead); Methodology (supporting); Writing – review & editing (supporting). Alon A. Gorodetsky: Conceptualization (lead); Formal analysis (supporting); Funding acquisition (lead); Methodology (supporting); Supervision (lead), Writing – review & editing (equal).
DATA AVAILABILITY
The data that support the findings of this study are available within the article and its supplementary material and are available from the corresponding author upon reasonable request.