Nature utilizes pH changes to regulate protein assembly, as seen in spider silk, CTP synthase filaments, and R bodies. This inspired the development of pH-responsive protein materials for applications in tissue engineering, drug delivery, and self-healing biomaterials. While some progress has been made in designing pH-responsive materials using methods like mimicking silk proteins or introducing histidine residues into coiled coils, the de novo design of filaments with precisely defined structures and tunable pH transition points remains a significant challenge. This study aimed to address this gap by computationally designing self-assembling protein filaments from subunits containing multiple buried residues that alter their protonation state over a small pH range, triggering unfolding/folding transitions. The high density of such residues in the filament would amplify the response to even minor pH variations, creating a highly sensitive and cooperative system. Previously designed pH-dependent trimeric helical bundles, with their sharp unfolding transitions at tunable pH values and nine buried histidines, served as promising starting points.

Several studies have explored pH-sensitive protein assembly. Askarieh et al. (2010) showed that spider silk protein assembly is controlled by a pH-sensitive relay. Hansen et al. (2021) revealed the cryo-EM structures of CTP synthase filaments and their pH-sensitive assembly during yeast starvation. Polka & Silver (2016) demonstrated a tunable protein piston responding to pH changes. Beun et al. (2014) made progress in creating pH-responsive assembling nanomaterials by mimicking silk protein domains. Zimenkov et al. (2006) and Boyken et al. (2019) reported on the rational design and de novo design of reversible pH-responsive switches for peptide self-assembly and tunable pH-driven conformational changes, respectively. These existing approaches provide a foundation for this research, but lacked the precise control over structure and tunability achieved in this study.

The researchers designed short loops to connect three protomers of a pH-dependent trimer (pRO-2.3) containing six buried histidines into a single chain. This addressed the trimer's internal symmetry, preventing unwanted bundling and allowing the design of asymmetric interfaces for fibre assembly. They employed a filament design method to dock these monomers into various helical filament arrangements, designing interface amino acid residues to drive assembly. 18 designs were selected for experimental testing based on predicted energy and other metrics. These designs (DpHFs) were expressed in *E. coli* and purified. Cryo-EM was used to determine the high-resolution structures of selected DpHFs. A redesign (DpHF19) was created to correct for deviations from the design model observed in DpHF18. Negative stain EM, fluorescence microscopy (TIRFM), and liquid-phase atomic force microscopy (AFM) were employed to characterize pH responsiveness, including assembly dynamics and disassembly kinetics. A version with nine buried histidines (DpHF19_9his) was created to explore tunability. Photoacid activation with a UV source was used to induce local or global disassembly, controlling the location and pattern of disassembly.

Two designs, DpHF7 and DpHF18, initially formed filaments. Cryo-EM revealed that DpHF18's structure deviated slightly from the model, exhibiting antiparallel dihedral symmetry. Subsequent redesign (DpHF19) yielded a structure closely matching the model (1.4 Å RMSD). DpHF19 showed a one-start helical symmetry with five subunits forming a ring in cross-section. Negative stain EM showed that DpHF18, DpHF19, and DpHF19_9his disassembled at pH 3.5, 3, and 4.2, respectively, with reversible assembly upon pH increase. TIRFM revealed end-to-end fibre growth and disassembly within less than 1 second after a pH drop. Liquid-phase AFM confirmed a sharp pH transition over 0.3 pH units for DpHF18 and 0.1 pH units for DpHF19_9his. Disassembly at pH 4.4 showed that fibres disassemble from both ends at approximately the same rate, however the rate of disassembly was much higher at a pH of 4.1. Photoacid activation by UV light allowed precise control of fibre disassembly, creating spot or line patterns. The pH-dependent response was highly cooperative, likely due to the large number of buried histidines per subunit and their close packing.

This study successfully demonstrates the de novo design of micrometre-scale pH-responsive protein filaments with remarkable tunability. The sharpness of the pH response highlights the cooperative nature of the disassembly process driven by the numerous buried histidine residues. The close structural agreement between the designed model and the cryo-EM structure of DpHF19 validates the design approach. The ability to control disassembly with both global and local pH changes using photoacids opens new avenues for applications such as controlled drug release. This approach overcomes challenges inherent in traditional protein engineering and offers a powerful tool for creating advanced, environmentally responsive nanomaterials.

The de novo design of pH-responsive protein filaments represents a significant advance in the creation of environmentally responsive protein nanomaterials. The tunable and sharp pH dependence of these filaments, along with the ability to control their disassembly using photoacids, opens up exciting possibilities for applications like controlled drug delivery. Future research could explore the incorporation of these filaments into hydrogels or other higher-order structures to further exploit their pH-responsiveness. The precise design of reconfigurable protein systems with atomic-scale resolution also provides opportunities for developing more sophisticated environmentally sensitive nanomaterials.

The study notes that even small deviations in interface geometry from the design model can lead to aggregation or altered fibre geometry. This highlights the need for improved design accuracy; deep learning protein design methods could improve the success rate. Further investigation into the structural intermediates during filament assembly and disassembly remains challenging, although future advances in liquid-phase transmission electron microscopy may help to address this.