Design of multifunctional protein crystals by DNA-mediated symmetry breaking

Abstract

The present disclosure provides methods for programmable biomaterial synthesis, providing the ability to influence symmetry of a protein crystal via DNA conjugation to wild-type or mutated amino acid residues in or near PPIs.

Description

Docket No. 30938 / 2024-145DESIGN OF MULTIFUNCTIONAL PROTEIN CRYSTALS BY DNA-MEDIATED SYMMETRY BREAKINGSTATEMENT OF GOVERNMENT SUPPORT

[0001] This invention was made with government support under grant numbers FA9550- 22-1 -0300 and FA9550-17-1 -0348 awarded by the Air Force Office of Scientific Research and grant number DMR2104353 awarded by the National Science Foundation. The government has certain rights in the invention.INCORPORATION BY REFERENCE OF MATERIAL SUBMITTED ELECTRONICALLY

[0002] The Sequence Listing, which is a part of the present disclosure, is submitted concurrently with the specification as a computer-readable file. The name of the computer- readable file containing the Sequence Listing is “2024-145_SeqListing. xml”, which was created on August 28, 2025 and is 9,312 bytes in size. The subject matter of the Sequence Listing is incorporated herein in its entirety by reference.FIELD

[0003] The disclosure relates to methods of reducing protein crystal structural symmetry via conjugating DNA to the protein surface.BACKGROUND

[0004] Low symmetry inorganic materials, synthetic polymers, amino acids and peptide crystals exhibit symmetry-dependent optical and electrical properties, offering promising applications in biomedicine, optics, and electronics. The periodic arrangement of atoms in three dimensions influences the mechanical, electrical, and optical properties of crystalline materials. While such structure-property relationships are found in atomic, metallic and ionic crystals, they remain poorly understood in complex biomolecular systems, and deliberately reducing symmetry in complex biomolecular systems remains a formidable challenge. This difficulty arises from the stochastic nature of protein crystallization, which involves multiple non-covalent interactions and the inherent tendency to maximize protein contacts in crystal packing. Current strategies in protein interface design, such as de novo protein design, metal coordination, peptide fusion, and disulfide linkages, primarily promote protein oligomerization, resulting in highly symmetric assemblies that limit their potential for emergent functionalities.SUMMARY

[0005] The present disclosure addresses the foregoing challenges by introducing a number of DNA ligands conjugated to precise locations on protein surfaces to control macromolecular structure during crystallization, where DNA conjugated to amino acids in orDocket No. 30938 / 2024-145 near crystal protein-protein interaction (PPI) regions can obstruct PPIs to a certain extent, contributing to the overall structure observed. This method enables the structure of protein crystals to be programmed and controlled. Using the methods of the disclosure, protein crystal structure is controlled through programming DNA sequence, length, and placement.

[0006] The present disclosure provides methods for conjugating DNA strands to proteins, reducing symmetry of the resultant DNA-protein conjugate crystals that concomitantly increases biocompatibility and broadens the functionality of the crystals. The methods disclosed herein for producing a protein crystal comprise: (i) conjugating one or more DNA strands to a surface of a wild-type (non-mutated) or mutated protein using a cross-linker to form a protein-DNA conjugate, wherein the mutated protein is a protein wherein one or more amino acids in or near the crystal PPI region are mutated; and (ii) crystallizing the protein- DNA conjugate to form DNA-modified protein crystals, wherein the mutated protein crystals have reduced symmetry when compared to protein crystals crystallized without the DNA conjugation, generated under similar conditions. In some embodiments, the symmetry of a DNA-modified protein crystal is reduced by at least 1 when compared to protein crystals derived from the wild-type protein generated under similar conditions, wherein symmetry is defined by the number of symmetry operators of a given space group. In some embodiments, the non-mutated or mutated protein is a globular protein, fibrous protein, or membrane protein that can be crystallized.

[0007] In some embodiments, the one or more amino acids that participate in PPI are mutated to cysteine. In various embodiments, the one or more amino acids that participate in PPI are mutated into cysteine, lysine, aspartic acid, glutamic acid, or another amino acid comprising an azide, carboxylic acid, sulfhydryl, carbonyl, or amine-containing natural or unnatural amino acid, or a combination thereof. In various embodiments, the one or more amino acids that participate in PPI meet one or more of the following conditions: (i) the one or more amino acids in the PPI region are involved in disulfide bonding, salt bridging, electrostatic interactions, hydrogen bonding, covalent bonding, hydrophobic interactions, and / or van der Waals interactions; (ii) the one or more amino acids in the PPI region are surface-exposed; (iii) the one or more amino acids in the PPI region are at unstructured locations; and (iv) the one or more amino acids in the PPI region are located along the symmetry axis of the protein molecule.

[0008] In some embodiments, the DNA strands are non-complementary. In some embodiments, the DNA strands are self-complementary or complementary. In some embodiments, the DNA strands are between 1 and 50 nucleotides in length. In various aspects, the DNA strands are 10 nucleotides or less in length. In various aspects, the DNA strands are between 4 and 6 nucleotides in length.Docket No. 30938 / 2024-145

[0009] In some embodiments, the DNA is attached to the one or more amino acids in or near the PPI region via a cross-linker. In various aspects the cross-linker is a PEG-based cross-linker, optionally wherein the PEG-based cross-linker has no more than 10 PEG units. In various aspects, the cross linker comprises one or more functional groups or chemical moieties including DBCO, azide, alkyne, phosphine, maleimides, thiol, pyridyl disulfies, haloacetyls, NHS-esters, imidoesters, carbodiimides, hydrazides, alkoxyamines, aryl azides, and / or diazirines.

[0010] In some embodiments, 1 or more DNA strands are attached to the one or more mutated amino acids on the mutated protein, or to one or more amino acids on the wild-type protein. In some embodiments, 1 to 500 strands are attached to the one or more mutated amino acids on the mutant protein, or to one or more amino acids on the wild-type protein.

[0011] In some embodiments, the DNA loading density for DNA functionalization ranges from 1-100 pmol / cm2. In some embodiments, the concentration of the protein-DNA conjugate to be crystallized ranges from 1 mg / mL to 50 mg / mL.BRIEF DESCRIPTION OF THE DRAWINGS

[0012] FIGs. 1 A-1 E show the design of DNA-modified ferritin crystals, crystal packing of Synechococcus sp. CC9311 ferritin (SynFtn) with F432 space group. FIG. 1 A provides optical microscopy images of SynFtn crystals exhibiting halved octahedral morphology sectioned along the <111> direction. Scale bars: 200 pm, (inset)100 pm. FIGs. 1 B-1 C provide the unit cell (FIG. 1 B) and 110 plane (FIG. 1 C) of SynFtn crystals in the F432 space group. FIG. 1 D shows the protein-protein interface in SynFtn crystals, showing intermolecular hydrogen bonds between Asp90 and Thr88. FIG. 1 E is a schematic illustration of DNA-directed symmetry breaking. Conjugation of non-hybridizing DNA at the protein-protein interface disrupts native crystal contacts along the C2 interface, leading to crystal formation with reduced symmetry. In contrast, DNA conjugation at residues along the four-fold axis preserves native protein-protein interactions and packing arrangement.

[0013] FIGs. 2A-2G demonstrate symmetry breaking in DNA-modified SynFtn crystals. FIG. 2A provides size exclusion chromatograms of SynFtn-DNA conjugates. FIG. 2B provides optical microscopy image of SynFtnN161c-DNA crystals exhibiting truncated octahedral morphology. Scale bar: 200 pm. Fig. 2C provides 1.39 A-resolution structure of SynFtnN161c-TTTT conjugates in the F432 space group, showing electron density at the crystal contact formed via Thr88-Asp90 interactions and two alternative conformations of Cys161-maleimide adducts. The 2Fo-Fc electron density map is contoured at 2.0 o. FIG. 2D shows an optical microscopy image of SynFtnT88C-DNA crystals exhibiting cubic morphology. Scale bar: 200 pm. FIG. 2E shows a unit cell of the 1 .77 A-resolution structureDocket No. 30938 / 2024-145 of SynFtnT88C-AGCT conjugates in the 123 space group. FIG. 2F shows the 100-crystal plane of SynFtnT88C-AGCT conjugates, with monomers forming four-fold symmetric pores shown in dark gray. FIG. 2G shows the protein-protein interface in the symmetry-reduced SynFtnT88C-AGCT crystals, showing electron density surrounding Cys88, hydrogen bondforming residues, and the Mn2+(larger sphere) coordination environment. Water molecules are shown as the smaller spheres. The 2Fo-Fc electron density map is contoured at 2.0o for amino acids and 1 .Oo for the Mn-coordination complex.

[0014] FIGs. 3A-3B demonstrate the biocompatibility of SynFtn-DNA crystals. FIG. 3A provides viability of NIH3T3 fibroblasts after incubation with crystals of SynFtnN161c-DNA (top) and SynFtnT88C-DNA conjugates (bottom) over 24, 48, or 96 h. NT, no treatment. FIG. 3B shows the morphology of NIH3T3 cells following treatment with SynFtnT88C-DNA crystals in the 123 space group, with Actin filaments (F-actin and nuclei being shown). Scale bar, 50 pm. NT, no treatment.

[0015] FIGs. 4A-4E demonstrate the mechanical properties of SynFtn-DNA crystals. FIGs. 4A-4C are representative load-depth curves from nanoindentation experiments on crosslinked (dark colors) and unprocessed (light colors) crystals of wild-type SynFtn (FIG. 4A), SynFtnN161c-DNA (FIG. 4B), and SynFtnT88C-DNA (FIG. 4C) under displacement control. Insets show representative optical images of protein crystals on substrates. Scale bars: 80 pm (FIGs. 4A-4B); 40 pm (FIG. 4C). FIG. 4D provides a comparison of reduced Young’s modulus (Er) across different space groups of unprocessed and crosslinked SynFtn or SynFtn-DNA crystals. Moduli were determined from the load-depth curves. Data represent n = 5 crystals per group with at least three technical replicates per crystal. FIG. 4E provides a comparison of hardness values across the same crystal types, derived from the load-depth curves. Data represent N = 5 crystals per group in at least three technical replicates. Box plots in FIGs. 4D-4E represent the 25-75th percentiles, with whiskers indicating the range within 1 .5 IQR. Median values are shown as horizontal lines, means as hollow rectangles, and outliers as filled diamonds.

[0016] FIGs. 5A-5F demonstrate the optical and electrical properties of symmetry-reduced protein crystals. FIG. 5A provides second harmonic generation (SHG) of SynFtnT88C-DNA crystals, showing the SHG image of an 123 crystal (left) and the corresponding emission spectrum (right) under the excitation wavelength of 1200 nm. FIG. 5B provides SHG of wildtype SynFtn crystals, showing the SHG image of an F432 crystal (left) and the corresponding emission spectrum (right). FIG. 5C shows the emission spectra of SynFtnT88C-DNA crystals under excitation wavelengths of 1 150, 1200, 1250, and 1300 nm, showing frequency doubling. FIG. 5D provides amplitude of piezoresponse force microscopy (PFM) signals as a function of applied frequency from a SynFtnT88C-DNA crystal. Inset:Docket No. 30938 / 2024-145 optical image of the corresponding 123 crystal; scale bar, 50 pm. FIG. 5E is a topography image of the 123 crystal characterized by PFM. Scale bar, 1 pm. FIG. 5F provides vertical and lateral piezoelectric responses of the same crystal, with phase and amplitude maps in vertical mode (top) and lateral mode (bottom).DETAILED DESCRIPTION

[0017] Disclosed herein are methods that utilize DNA conjugation to control crystallization of proteins and the degree of symmetry exhibited by the resulting DNA-protein conjugate crystal.

[0018] Proteins are the central building blocks of biological systems, and are powerful synthons for supramolecular materials because of their well-defined structures and sophisticated chemical functions. Their assembly into well-defined 1 , 2 and 3D functional structures in nature has inspired efforts to engineer the assembly of proteins into designed architectures. The assembly of proteins, however, is difficult to control synthetically owning to the chemical heterogeneity of their surfaces, representing a major challenge. For example, the organization of proteins into single crystals with high positional, orientational, and translational order results in materials where the location of every atom can be known, and yet controlling the organization of proteins is challenging due to the myriad interactions that define protein interfaces within a single crystal. Efforts to control protein crystallization have included modifications that affect charge, hydrophobicity, protein structure, ligand binding, and metal binding characteristics.

[0019] Provided herein are methods to modulate the organization of protein crystals with DNA such that the resulting DNA-modified protein crystals exhibit reduced symmetry when compared to the protein crystals alone. In some aspects, the disclosed methods utilize site- directed mutagenesis to precisely control DNA conjugation to the protein, enabling the user to tightly control the resulting DNA-modified protein crystal. The methods for producing a protein crystal comprise: (i) conjugating one or more DNA strands to the surface of a wildtype (non-mutated) or mutated protein using a cross-linker to form a protein-DNA conjugate, wherein the mutated protein is a protein wherein one or more amino acids in or near the crystal PPI region are mutated; and (ii) crystallizing the protein-DNA conjugate to form DNA- modified protein crystals, wherein the mutated protein crystals have reduced symmetry when compared to protein crystals crystallized without the DNA conjugation, generated under similar conditions. Due to DNA interactions that span the nanometer length scale, DNA can eliminate entire protein-protein interfaces along selected symmetry axes, engendering designed low-symmetry protein crystals with emergent optical, mechanical, and electrical properties.Docket No. 30938 / 2024-145

[0020] The structural arrangement of crystalline materials underlies their mechanical, electronic, and optical properties. While these structure-property relationships are well established in inorganic, metallic, or ionic systems, they remain poorly understood in more complex, macromolecular crystals. Methods of the disclosure provide a bottom-up strategy to program the symmetry of protein crystals using DNA as a nanoscopic chemical modifier. Conjugating non-hybridizing DNA strands to a protein, such as for example, ferritin, was observed to selectively disrupts native protein-protein interactions, steering crystallization towards low-symmetry lattices. The disclosed symmetry-reduced crystals regained angstrom-level order but exhibited increased solvent content, enhanced mechanical compliance, and emergent second-harmonic generation and piezoelectricity — properties absent in high-symmetry counterparts. Such programmable symmetry breaking reveals a previously untapped structure-function relationship in biomolecular materials and demonstrates the versatility of the platform for designing multifunctional materials, with applications in bioelectronics, photonics, medicine, and soft robotics.

[0021] Crystal symmetry describes the packing arrangement of atoms and molecules within a unit cell and plays a crucial role in determining the material properties of a crystal. Centrosymmetric crystals, which contain an inversion center, possess more symmetry elements than non-centrosymmetric ones. Consequently, high-symmetry crystals do not exhibit non-linear optical properties, such as second harmonic generation (SHG), because their second-order susceptibility is zero. Additionally, the bonding configuration within crystals dictates their mechanical and electrical properties, as seen in planar sp2-bonded graphite, which is soft and conductive, versus tetrahedral sp3-bonded diamond, which is hard and insulating. Designing atomic, metallic, and molecular crystals with reduced symmetries has proven to be an effective strategy for inducing emergent optical, mechanical, and electrical properties. However, while some discrete properties, such as piezoelectricity1-4and SHG5, have been observed in natural protein crystals, engineering protein crystals with programmable symmetries to systematically alter material properties remains largely unexplored. Disclosed herein are methods of engineering protein crystals via conjugating DNA to one or more amino acids in or near the crystal PPI region, yielding programmable symmetries in protein crystals.

[0022] Desirable material properties such as piezoelectricity, the reversible coupling between mechanical stress and electric charge, and second harmonic generation (SHG), a nonlinear optical process, require the absence of inversion symmetry. Non-centrosymmetric inorganic crystals, synthetic polymers, and ceramics, such as lithium niobate, potassium dihydrogen phosphate, barium titanate, and zinc oxide, have been widely used in nonlinear optics, photonics, sensing, actuation, and energy harvesting. Biomaterials including aminoDocket No. 30938 / 2024-145 acids, peptides3, and biopolymers1have also been shown to exhibit piezoelectricity and SHG, making them attractive for biomedical and bioelectronic applications.

[0023] Proteins, composed of chiral L-amino acids, are inherently asymmetric and crystalize in one of 65 non-enantiogenic space groups. This intrinsic non-centrosymmetry makes protein crystals a promising, yet largely unexplored class of biomaterials with the potential to exhibit multifunctional properties.5Indeed, the direct piezoelectric effect has been observed in natural crystals of lysozyme2and bas cytochrome c oxidase4. However, the mechanistic origins of nonlinear optical and electromechanical responses in macromolecular crystals remain elusive, and it is unknown whether these properties can be engineered into designed protein assemblies. The structural complexity of proteins, arising from their folded, higher-order architectures composed of thousands of atoms, complicates mechanistic insight and limits both the discovery and rational design of protein-based functional materials. It has surprisingly and beneficially been found that, contrary to conventional materials, the nonlinear optical and electromechanical responses of macromolecular crystals are not dictated by local dipoles inherent to individual molecular structures, but by global molecular packing and macroscopic polarization.

[0024] As used herein, “symmetry” is a description of protein subunits within a crystal, wherein all protein subunits have identical interactions. Protein assemblies are typically symmetrical — they are built from many identical subunits that each interact with their neighbors in identical ways. The symmetry of a DNA-modified protein crystal is defined by the number of symmetry operators of a given space group. As used herein, a “symmetry operator” is an operation that interchanges the positions of atoms within a crystal, and results in the molecule or crystal appearing exactly the same as before the operation. As used herein, a “space group” is a symmetry group of a repeating pattern in space, usually in three dimensions. A space group is a collective set of geometrical symmetry operations that can change the orientation of a crystal without changing the position of its atoms. These changes (via symmetry operators) include, but are not limited to, translations, rotations, reflections, inversions, and sequential rotary inversions.

[0025] As used herein, the term “protein” refers to a polymer comprised of amino acid residues. Proteins are understood in the art and include without limitation an antibody, an enzyme, a structural protein and a hormone. Thus, proteins contemplated by the disclosure include without limitation those having structural, catalytic, signaling, therapeutic, or transport activity.

[0026] Proteins of the present disclosure may be either naturally occurring or non- naturally occurring. Naturally occurring proteins, referenced herein as “wild type” or “nonDocket No. 30938 / 2024-145 mutated” proteins, include without limitation biologically active proteins (including antibodies) that exist in nature or can be produced in a form that is found in nature by, for example, chemical synthesis or recombinant expression techniques. Naturally occurring proteins also include lipoproteins and post-translationally modified proteins, such as, for example and without limitation, glycosylated proteins. Antibodies contemplated for use in the methods and compositions of the present disclosure include without limitation antibodies that recognize and associate with a target molecule either in vivo or in vitro. Structural proteins contemplated by the disclosure include without limitation actin, tubulin, collagen, elastin, myosin, kinesin and dynein.

[0027] Non-naturally occurring proteins contemplated by the present disclosure, also referred to as “mutated proteins”, are proteins wherein one or more amino acids have been mutated when compared to the wild-type protein. Mutated proteins include but are not limited to synthetic proteins, as well as fragments, analogs and variants of naturally occurring or non-naturally occurring proteins as defined herein. Non-naturally occurring proteins include proteins or protein substances that have D-amino acids, modified, derivatized, or non- naturally occurring amino acids in the D- or L- configuration and / or peptidomimetic units as part of their structure. Non-naturally occurring proteins also include proteins designed by de novo computational methods, such as artificial intelligence. The sequence and / or folding patterns of these proteins can be distinct (i.e., de novo) from natural analogs. These de novo proteins, either with or without mutations, can also be classified as non-naturally occurring proteins. The term "peptide" typically refers to short polypeptides / proteins.

[0028] Non-naturally occurring proteins are prepared, for example, using an automated protein synthesizer or, alternatively, using recombinant expression techniques using a modified polynucleotide which encodes the desired protein.

[0029] As used herein, “functionalization” is the process of conjugating a DNA strand to an amino acid residue. As used herein, a “protein-DNA conjugate” is the complex that results from the functionalization of one or more DNA strands on the surface of a protein. The protein-DNA conjugate can be present in solution. As used herein, a “DNA-modified protein crystal” is obtained from the crystallization of the protein-DNA conjugate.

[0030] In various embodiments of the disclosure, the DNA is single-stranded. In some embodiments, the DNA is double stranded. Single stranded DNA also includes DNA with secondary structure, such as, for example and without limitation, G-quadruplexes and i- motifs.

[0031] In some embodiments of the disclosure, wild-type protein candidates of interest are selected and synthesized. In various aspects, protein targets can be selected from theDocket No. 30938 / 2024-145Protein Data Bank (PDB), or unpublished protein structures. In various aspects, the protein sequence can be naturally occurring or computationally designed. The protein can be synthesized or expressed using natural organisms or via solid-phase peptide synthesis. The chirality of proteins can be left-handed (L-polypeptides) or right-handed (R-polypeptides)(i.e., not racemic). In some embodiments, the protein candidate of interest is a globular, fibrous, or membrane protein.

[0032] In some embodiments, the protein crystal interfaces are analyzed to identify one or more amino acid residues that participate in crystal PPIs within the crystal contacts. The one or more amino acids selected will serve as conjugation sites for DNA. In various aspects, the crystals are obtained via crystallization of the wild-type proteins. In various other embodiments, the protein interface is examined via PDB, if the published structures are available.

[0033] Controlling the position and orientation of proteins within single crystals is a challenging endeavor due to the myriad weak noncovalent interactions on protein surfaces. These native PPIs dictate the packing of proteins within single crystals, and are difficult to predict. Conjugating a single DNA strand at a residue that exists between protein surfaces leads to changes in the packing of proteins within single crystals, and the PPIs that arise.

[0034] In some embodiments, one or more amino acid residues that participate in PPIs within the crystal contacts are selected for mutation. In some embodiments, the one or more amino acids selected for mutation is mutated to cysteine. In some aspects, the one or more amino acids selected for mutation are mutated to lysine, aspartic acid, or glutamic acid. In various aspects, the one or more amino acids selected for mutation are mutated into an azide-, carboxylic acid-, sulfhydryl-, carbonyl-, or amine-containing natural or unnatural amino acid. In various embodiments, the one or more amino acids are involved in at least one of the following types of bonding: disulfide bonding, salt bridges, electrostatic interactions, hydrogen bonding, covalent bonding, hydrophobic and / or van der Waals interactions. In some aspects, the one or more amino acids are surface-exposed. In some aspects, the one or more amino acids are located at unstructured locations or structured but flexible locations, wherein an “unstructured location” is an intrinsically disordered region (i.e., a protein region without a proper folding structure). Amino acid residues within structured but flexible locations include, but are not limited to, residues in flexible loops, or residues with weak or absent electron densities within the crystal structures. In some embodiments the one or more amino acids are located along the symmetry axis of the protein molecule.

[0035] In other embodiments, the amino acid residues that participate in PPIs within the crystal contacts are not mutated.Docket No. 30938 / 2024-145

[0036] In some embodiments, a protein with one or more amino acid mutations in PPIs is expressed and synthesized, and DNA strands are conjugated to one or more mutated protein surface. In other embodiments, a wild-type protein is expressed and synthesized, and DNA strands are conjugated to one or more native amino acid residues on the protein surface. In some embodiments, a single DNA strand is conjugated to a single amino acid residue.

[0037] DNA ligands can be chemically tethered to the surfaces of proteins, at specific locations, to drive the assembly of proteins into one- and three-dimensional structures and crystals. Protein mutagenesis has been used to site-specif ically encode multiple, orthogonal DNA interactions onto protein surfaces to program directional assembly. Chemical functionalization of DNA strands on protein surfaces alters protein packing within single crystals. Without intending to be bound by theory, it is believed that proteins modified with DNA strands as formed in the methods of the disclosure possess reduced point group symmetry and can be crystallized in lower symmetry space groups compared to their unmodified counterparts. Due to DNA interactions that span the nanometer length scale, DNA can eliminate entire protein-protein interfaces along selected symmetry axes, engendering designed low-symmetry protein crystals with emergent optical, mechanical, and electrical properties.

[0038] To append DNA to the surface of proteins, surface amines (lysines) or thiols (cysteines), can be selectively reacted with oligonucleotides through conjugation strategies that include, but are not limited to, reaction with an NHS-ester-azide crosslinker and cyclooctyne-terminated DNA, and reaction with pyridyl disulfide terminated DNA.

[0039] As used herein, the term “conjugated” includes both covalent and non-covalent interactions between the protein and the DNA e.g., covalent conjugation or ligand binding, such as sugar binding (e.g., functionalizing a DNA with a sugar moiety (such as a monosaccharide) such that the DNA-sugar conjugate is attached to a protein via binding of the sugar moiety). Appropriate chemistries for conjugating DNA to a protein disclosed herein are known to those skilled in the art. Conjugation of the DNA to the protein may be accomplished using, for example, a bio-orthogonal copper catalyzed or copper-free click chemistry reaction (e.g., DBCO-azide) an inverse-electron demand Diels-Alder (IEDDA) reaction, azide-alkyne cycloaddition, azide-phosphine reaction (i.e. Staudinger ligation), NHS-ester amine reaction, imidoester-amine reaction, carbodiimides chemistry (dicyclohexyl carbodiimide, 1 -Ethyl-3-(3-dimethylaminopropyl)carbodiimide), maleimide-sulfhydryl reaction, pyridyl disulfide exchange, haloacetyl-sulfhydryl reaction, hydrazides-aldehyde reaction, alkoxyamines-carbonyl reaction, reductive aminoation, protein-ligand interactions,Docket No. 30938 / 2024-145 or photoreactive conjugation chemistries involving aryl azides, hydroxyphenyl azides, tetrafluorophenyl azide, nitrophenyl assize, diazirine, azido-methylcoumarin, and psoralen.

[0040] The protein to be conjugated to DNA may comprise an azide, a tetrazine, or a combination thereof. In some cases, the protein to be conjugated to DNA comprises one azide. In some cases, the protein to be conjugated to DNA comprises a plurality of azides. In some cases, the protein to be conjugated to DNA comprises one tetrazine. In some cases, the protein to be conjugated to DNA comprises a plurality of tetrazines. In some cases, the protein to be conjugated to DNA comprises a combination of azides and tetrazines. The azide may be located at the C-terminus or N-terminus of the protein, or it may be an internal azide (e.g., an azide located on the side chain of an amino acid residue in the protein). The azide may be introduced into the protein via an azide-containing linker, or via a non-naturally occurring amino acid. DNA to be conjugated may comprise an alkene or alkyne, which acts as the complimentary click reagent. The alkyne may be introduced into the polynucleotide via a linker containing the alkene or alkyne, e.g., trans-cyclooctene (TCO), dibenzocyclooctyne (DBCO), or bicyclononyne (BCN). Alternatively, the polynucleotide may comprise the azide or tetrazine and the protein may comprise the alkene or alkyne. It is to be understood that other conjugation methods may be used to effect conjugation of the polynucleotide to the protein, e.g., NHS ester conjugation, isocyanate conjugation, isothiocyanate conjugation, maleimide conjugation, iodoacetamide conjugation, and other conjugation methods known to those skilled in the art.

[0041] The programmability of DNA allows for selective tuning of protein interactions during crystallization (i.e., facilitation or inhibition of PPIs) via complementary, self- complementary, and non-complementary sequences without compromising protein synthesis and crystallization. In some embodiments, the one or more DNA strands functionalized onto the protein surface are self-complementary or complementary. In other embodiments, the one or more DNA strands functionalized onto the protein surface are non-complementary. In some embodiments, the one or more DNA strands functionalized onto the protein surface are between 1 -50 nucleotides in length, between 1-10 nucleotides in length, or between 4-5 nucleotides in length.

[0042] In some embodiments, the DNA strands are conjugated to an amino acid via a cross-linker. In various aspects, the cross-linker is a cleavable cross-linker, a non-cleavable cross-linker (e.g., p-maleidopropyl-oxysuccinimide ester (BMPS)), a traceless cross-linker, or a combination thereof. In various aspects, the cross-linker comprises one or more functional groups or chemical moieties including DBCO, azide, alkyne, phosphine, maleimides, thiol, pyridyl disulfies, haloacetyls, NHS-esters, imidoesters, carbodiimides, hydrazides, alkoxyamines, aryl azides, and / or diazirines.Docket No. 30938 / 2024-145

[0043] In some embodiments, the number of DNA strands functionalized onto the protein surface ranges from 1 strand (selective functionalization) to 500 strands (dense DNA coverage), with the number of strands referring to the number of strands functionalized on the surface of a single protein. In some embodiments, the DNA loading density during the functionalization process ranges from 1 to 100 pmol / cm2. In some embodiments, the degree of DNA functionalized on the protein surface versus the total available conjugation sites ranges from 5% to 100%, with 100% being full functionalization.

[0044] In some embodiments, DNA functionalization is followed by crystallization of the protein-DNA conjugate. In some embodiments, the concentration of the protein-DNA conjugate used for crystallization ranges from 1 mg / mL to 50 mg / mL. In other embodiments, the concentration of the protein-DNA conjugate used for crystallization ranges from 5mg / mL to 20mg / mL.

[0045] Crystallization methods compatible with the present methods include, but are not limited to, sitting drop vapor diffusion, handing drop vapor diffusion, microbatch crystallization, microdialysis crystallization, slow-cooling or annealing within a temperature gradient, or annealing at a constant temperature. In some embodiments, crystallization of the protein-DNA conjugate is followed by determination of the structures of DNA-modified protein crystals, which can occur through one or a combination of the following methods: single-crystal X-ray diffraction, small- or wide-angle X-ray scattering, cryogenic or negativestain transmission electron microscopy, microcrystal electron diffraction (MicroED), atomic force microscopy, and nuclear magnetic resonance (NMR) spectroscopy.

[0046] In some embodiments, the methods disclosed herein yield DNA-protein conjugate crystals with emergent optical properties. In various aspects, the optical properties of the DNA-protein conjugate crystals are determined through one or a combination of the following methods: Z-scan technique, SHG and third harmonic generation (THG) (multiphoton confocal microscopy), pump-probe spectroscopy, photoluminescence spectroscopy, time- resolved photoluminescence, circular dichroism (CD) spectroscopy, and optical roatory dispersion (ORD). Optical properties of the DNA-protein conjugate crystals include nonlinear optical properties, photoluminescence, and chiral optical properties. Optical properties, particularly non-linear optical properties, are influenced by crystal symmetry. For example, all crystals without an inversion center are non-centrosymmetric. In the same crystal system, centrosymmetric structures have more symmetry elements than non-centrosymmetric ones. Centrosymmetric crystals do not have SHG, one typical non-linear optical property, due to the second-order susceptibility being zero. However, breaking the inversion center and converting centrosymmetric structures into non-centrosymmetric ones can make the crystals exhibit SHG properties due to the nonzero susceptibility. Reducing crystal symmetryDocket No. 30938 / 2024-145 changes the crystal structure in such a way that some non-linear properties, eliminated by the formation of a highly symmetrical crystal, will re-appear when asymmetry is introduced into the crystal structure.

[0047] In some embodiments, the methods disclosed herein yield DNA-protein conjugate crystals with emergent mechanical properties. In various aspects, the mechanical properties of the DNA-protein conjugate crystals are determined through one or a combination of the following methods: nanoindentation, tensile testing, electromechanical test system, Vickers hardness testing, Rockwell hardness test, three-point bending test, dynamic mechanical analysis (DMA), microscopy techniques (scanning or transmission electron microscopy), and mechanical testing (fatigue and fracture tests). Mechanical properties of the DNA-protein conjugate crystals include Young’s modulus, stiffness, hardness, shape memory, and self- healing. Altering the symmetry of protein crystals will influence the crystal structure, and thus the packing of the protein molecules in the crystals. Mechanical properties are determined by the packing of atoms and molecules in the crystals. A reduction in symmetry will alter the molecular packing of proteins within a crystal, leading to changes in solvent content and porosity of the crystal. Densely packed crystals will be mechanically tough because of the dense connectivity between molecules. In the case of exemplary SynFtn (further described below), 123 crystal has less dense packing than F432 crystal due to symmetry reduction, therefore leading to a lower Young’s modulus.

[0048] In some embodiments, the methods disclosed herein yield DNA-protein conjugate crystals with emergent electrical properties. In various aspects, the electrical properties of the DNA-protein conjugate crystals are determined through one or a combination of the following methods: piezoelectric or piezoresponse force microscopy (PFM), dynamic mechanical analysis (DMA) with piezoelectric measurement, hysteresis loop measurements, pyroelectric coefficient measurement, thermally stimulated depolarization current (TSDC), bending tests with electrical measurement, and scanning probe microscopy techniques. Electrical properties of the DNA-protein conjugate crystals include piezoelectricity, ferroelectricity, pyroelectricity, and flexoelectricity. Reducing the symmetry of protein crystals will influence the electrical properties the resultant crystals. Piezoelectricity and ferroelectricity are highly dependent on crystal structures. The exemplary SynFtn (further described below) without DNA conjugation forms a 432 structure, which has more symmetry elements that prevent piezoelectricity. The combination of its symmetry operations prevents the material from developing a net dipole moment under mechanical stress. The symmetry cancels out any dipole moments that could arise, rendering the material non-piezoelectric. In contrast, DNA-modified SynFtn forms 123 crystals, which is a noncentrosymmetric structure with fewer symmetry elements (without four fold rotation axis), potentially allowing forDocket No. 30938 / 2024-145 piezoelectricity due to the symmetry breaking. If further modifying the symmetry with DNA enables the formation of polar structures in SynFtn (among 20 noncentrosymmetric structures, only 10 are polar structures), the DNA-modified SynFtn can potentially exhibit ferroelectricity and pyroelectricity.

[0049] The effect of packing structures in crystals on mechanical and / or electrical properties is well known in the art for a variety of materials, both protein and non-protein based crystals. For example, graphite and diamond both consist of carbon atoms, but exhibit different mechanical properties (i.e., the softness of graphite and the hardness of diamond) due to the different crystals structures and packing of carbon atoms. As a result of the different crystal structures, they also exhibit vastly different electrical properties, with diamond being insulating and graphite being conducting. The crystal structure, such as a tetrahedral lattice through strong sp3covalent bonds verses hexagonal lattices in which an atom is bonded to three other atoms via strong sp2covalent bonds, forming a planar structure provide for different interactions with free electrons. In the former, no free electrons are left to conduct electricity, while in the hexagonal lattice, an electron not involved in bonding within the plane can allow for the crystal structure to be conductive.

[0050] Methods of the disclosure, which can controllably reduce symmetry and affect the packing and bonding between protein molecules, can be used to tailor mechanical and / or electrical properties of the crystals.

[0051] Any properties affected by crystal structure and packing are candidates for being tuned by the methods disclosed herein and the reduction of symmetry inside the crystals. In addition to optical, mechanical, and electrical properties, chemical properties are also influenced by symmetry reduction. For example, a reduction in symmetry can lead to a lower protein backing density and higher solvent content within a protein crystal, as protein crystals are porous and a large volume is filled by water or solvent. Therefore, the porosity of protein crystals will be affected by DNA functionalization on the protein surface, and the protein crystals can be utilized for encapsulation of guest molecules, molecular separation, or as a crystalline template. In addition, different arrangements of proteins within crystals may affect their catalytic properties due to proximity / access to various substrates and products.

[0052] The following examples are given merely to illustrate the present disclosure, and do not in any way limit its scope.EXAMPLES

[0053] Crystalline lysozyme, rich in a-helices, and its amyloid fibrils, dominated by - sheets28, were found to exhibit comparable piezoelectric coefficients despite vastly different secondary and quaternary structures. Contrary to conventional materials, it is believed that,Docket No. 30938 / 2024-145 the nonlinear optical and electromechanical responses of macromolecular crystals are not dictated by local dipoles inherent to individual molecular structures, but by global molecular packing and macroscopic polarization. In this example, ferritin protein crystals were designed in accordance with the disclosure with reduced crystallographic symmetry using DNA, and it was investigated how crystal packing and symmetry breaking influence the multifunctionality of these crystals.

[0054] Methods of the disclosure were used to prepare a crystal of reduced symmetry from isotropic Synechococcus sp. CC9311 ferritin (SynFtn), which naturally crystalizes in the F432 space group with 432 point groups. The symmetry of ferritin crystals can be programmed through chemical modifications of DNA on the protein surface, resulting in single crystals with symmetry-dependent material properties. This principal is exemplified by the foregoing SynFtn crystal, wherein DNA-functionalized protein crystalized in the I 2 3 space group resulted in new material properties including second-harmonic generation, reduced Young’s modulus and stiffness, and piezoelectricity compared to unmodified high- symmetry crystals. These low-symmetry crystals have increased solvent content, are mechanically soft, and exhibit second-harmonic generation and piezoelectricity — properties absent in their wild-type counterparts.

[0055] Strategies to design and restructure protein interfaces — such as residue-specific interactions via metal coordination6-9and disulfide bonds910, the installation or fusion of molecular glues via macrocycles11, ligands12-14, and peptides15-19, and computational design20-22— have generated well-defined protein assemblies and crystalline materials. However, these designed interactions, combined with the inherent symmetries of protein oligomers, often result in highly symmetric structures. Chemical functionalization of DNA on the protein surface has been previously explored as a strategy to modulate protein crystallization, with distinct changes observed in protein packing in DNA-modified protein crystals.2324Here, it was hypothesized that modifying the surface topology of a highly symmetric protein with non-hybridizing DNA can reduce the symmetry of the protein-DNA conjugate and eliminate selected protein-protein interactions during crystallization. Through this “interface-deletion” approach, protein crystals with reduced symmetry can be designed to exhibit emergent material properties.

[0056] Design of DNA-Modified Ferritin Crystals. Among the 21 non-centrosymmetric crystallographic point groups, the 432-point group is uniquely neither piezoelectric nor SHG- active due to its high rotational symmetry, which cancels both net polarization and second- harmonic generation. It was hypothesized that breaking the symmetry of protein crystals in the 432-point group could reveal latent functionalities arising from intrinsic molecular packing.Docket No. 30938 / 2024-145

[0057] An ideal starting point for protein topology engineering requires quaternary structures and crystal packing that are highly symmetric. For this purpose, ferritin from Synechococcus sp. CC9311 (SynFtn) was selected, as a targeted search of the Protein Data Bank identified cyanobacterial ferritin protein SynFtn, which crystallizes exclusively in the F432 space group. Ferritin’s cage-like structure, assembled from 24 monomers, possesses 432-point group symmetry. Analysis of existing structures from the Protein Data Bank (PDB) revealed the cubic F432 as the only observed space group, suggesting a lack of crystal polymorphism. Indeed, vapor diffusion experiments of SynFtn yielded octahedral and truncated octahedral crystals, consistent with the morphology of face-centered cubic symmetry (FIG. 1A). This morphology is characteristic of substrate-supported nucleation and growth of face-centered cubic crystals.25Single-crystal X-ray diffraction confirmed the expected packing arrangement (FIGs. 1 B-1 C), with crystal contacts stabilized by intramolecular hydrogen bonds (FIG. 1 D). Interface analysis of single crystals revealed hydrogen bonding between Asp90 and Thr88 as the key crystal contact between neighboring ferritin cages (FIG. 1 D). Within a ferritin cage, these residues are located on a looped structure in contact with two ferritin monomers and near the four-fold and three-fold channels (FIG. 1 E). These locations are therefore desired for DNA attachment to disrupt native protein-protein contacts and lower the crystal symmetry.

[0058] Each SynFtn molecule included 24 identical monomers arranged with overall 432- point symmetry (FIG. 1 E). Residues involved in crystal contacts are located on looped regions between adjacent monomers at the two-fold (C2) interface. To design SynFtn crystals with reduced symmetry, an ideal strategy would eliminate this native interface without disrupting crystallinity. Protein assembly strategies, such as de novo interface design,21 27metal coordination,29’8 9supramolecular linkages,11 13or peptide fusion,10 30’17’19 16typically introduce new protein-protein interactions that reinforce or enhance symmetry. In contrast, it was previously found that proteins conjugated with DNA can crystallize with lower symmetry than their unmodified counterparts.2324In these cases, the DNA strands remain non-hybridized and sterically disrupt native protein-protein interfaces at the nanoscopic level, making this approach well suited for symmetry reduction.

[0059] To apply this “interface deletion” strategy, Thr88, a residue at the native interface, was mutated to cysteine (SEQ ID NO: 9) and conjugated the mutant with maleimide- terminated, non-hybridizing DNA. The attached DNA sterically interfered with the Thr88- Asp90 interaction, favoring crystallization through alternative interfaces away from the C2 axis. As a control, a cysteine mutation was introduced at Asn161 (SEQ ID NO: 10), a site near the fourfold channel that does not participate in crystal contacts, allowing the SynFtnN161c-DNA conjugate to retain the native Thr88-Asp90 interactions.Docket No. 30938 / 2024-145

[0060] To functionalize DNA on the SynFtn surface, a T88C SynFtn mutant was recombinantly expressed (SEQ ID NOs: 6 and 9) and subsequently reacted with 5’- maleimide-modified poly-thymine DNA (SEQ ID NO: 4) via thiol-maleimide addition. DNA- SynFtn conjugates were characterized using denaturing gel electrophoresis, where a slight decrease in the electrophoretic mobility of the conjugates compared to the unmodified protein was observed, suggesting successful DNA attachment. Due to the non-hybridizing nature of the DNA sequence, the formation of a double helix was not possible during crystallization. Consequently, the DNA strands dangle from the protein surface, leading to a reduction in the point group symmetry of the SynFtn-DNA conjugate and the disruption of Asp90 and Thr88 interactions during crystallization (FIG 1 E).

[0061] DNA Induces Symmetry Breaking. Under size-exclusion chromatography (SEC), the protein-DNA conjugates exhibited a higher 260 / 280 nm absorbance ratio and eluted earlier than the unconjugated protein mutants, indicating an increased hydrodynamic radius (FIG. 2A). The reduced mobility of the conjugates in gel electrophoresis further supports DNA functionalization. Crystallization of SynFtn-N161 C-TTTT yielded crystals with morphologies and packing identical to those of native F432 crystals, indicating that DNA conjugation at non-interfacial sites has minimal impact on crystallization (FIG. 2B-2C). The Cys161-maleimide adduct adopts multiple conformations, and electron density for the remainder of the DNA strand was absent, suggesting the unhybridized DNA occupies a range of conformational states. Furthermore, conjugation with an AGCT sequence yielded the same crystal structure, indicating that clustered DNA functionalization does not promote hybridization but rather eliminate protein-protein interfaces.

[0062] In contrast, the SynFtn-T88C-DNA conjugate produced crystals with cubic-like morphology within two weeks (FIG. 2D). X-ray diffraction experiments revealed that these crystals adopt the 123 space group (FIG. 2E), which has 72 fewer symmetry operators (24 in total), fewer than the 96 found in F432. This change in packing eliminated twofold symmetries along the face diagonals and reduced the fourfold symmetry along the unit cell axes to twofold (FIG. 2F). Protein interface analysis confirmed that the native Thr88 — Asp90 interaction was eliminated; the corresponding residues were exposed to solvent channels within the crystal pores. Interestingly, proteins functionalized with either TTTT or AGCT also yielded crystals in the same space group with nearly identical unit cell parameters. Notably, protein interface analysis suggests the emergence of two new sets of protein-protein interfaces in the 123 crystals: hydrogen bonding between Pro9 and Gln7, and between Arg63 and Asp97 (FIG. 2G). This indicates a reorganization of packing interactions. These structural changes demonstrate that DNA conjugation at Cys88 effectively disrupts the native interface and induces a predictable reduction in crystallographic symmetry.Docket No. 30938 / 2024-145

[0063] Biocompatibility of Protein-DNA Crystals. To evaluate the biocompatibility of the protein-DNA crystals, their cytotoxicity was assessed in murine fibroblast (NIH3T3) cells. Crystals were tested both in their native form and after crosslinking with glutaraldehyde, a treatment known to enhance their chemical and thermal stability. Incubation with either noncrosslinked or crosslinked F432 and 123 crystals showed no observable difference in cell viability compared to untreated controls (FIG. 3A). This finding is supported by actin filament staining, which revealed healthy cell morphology and active proliferation from 24 to 96 hours across all conditions (FIG. 3B).

[0064] Mechanical and Optical Properties. For nanoindentation measurement of each type of protein with different crystal structures, approximately 20 indents were done on 3 different crystals to ensure the accurate evaluation. A typical curve consists of loading, holding period and unloading segments. A tangent drawn at the upper part of the unloading segment provides the material’s elastic Contact Stiffness (S), which is the change in load with respect to per unit change in the indentation depth S = dP / dh. Reduced modulus Er can be calculated from elastic contact stiffness and contact area.Er=(TTA0.5 / 2p)*(SA0.5 / A), where A is the contact area, p is a constant related to the indenter geometry, Er is the indentation modulus (reduced modulus). All indentation curves measured from these protein crystals (FIGs. 4A-4C) are smooth without significant mechanical noises, indicating a robust test with protein crystals firmly sticking to the substrates.

[0065] The average reduced elastic modulus measured for the DNA-Protein crystal ( 123) was 1 .2 ± 0.6 GPa, whereas a significantly higher average modulus of 3.3 ± 1 .5 GPa was observed in the pristine protein crystal (F432). The distribution of these measurements is shown in FIG. 4D. The difference between the 123 DNA-protein crystal and the F432 pristine protein can be attributed to the varying packing densities within the crystals. Specifically, in 123 crystals, each SynFtn molecule is bonded to 8 nearest neighbors (FIG. 2E), while in F432 crystals, each molecule connects with 12 nearest neighbors (FIG. 1 B) within a unit cell. As a result, 123 crystals are more porous, with a solvent content of 64.6%, compared to 57.6% in F432 crystals, indicating a lower packing density. This reduced packing density leads to a lower Young’s modulus and hardness, which aligns well with the results shown in FIGs. 4D-4E.

[0066] The nonlinear optical properties of protein crystals in solution were studied using a multiphoton imaging system (FIG. 5). For DNA-protein crystal with 123 structure, when excited at a wavelength of 1200 nm, a strong and relatively uniform optical response based on imaging was detected inside the crystal (FIG. 5A). To further validate the response dueDocket No. 30938 / 2024-145SHG instead of fluorescence, this crystal was excited with various wavelengths spanning from 1150 to 1300 nm. An expected peak shift of output emission from 575 nm to 650 was observed in all collected spectra (FIG. 5C). Notably, the center of emission peak under each excitation aligns well with the expected emission wavelengths, which are precisely half the excitation wavelengths. On the contrary, no significant imaging and spectra (FIG. 5B) response were detected in the when excited under the irradiations with the same wavelengths and intensities. These results indicate the robust SHG response of the 123 crystal due to the symmetry-breaking structure.

[0067] For nanoindentation measurement of each type of protein with different crystal structures, approximately 20 indents were done on 3 different crystals to ensure the accurate evaluation. A typical curve consists of loading, holding period and unloading segments. A tangent drawn at the upper part of the unloading segment provides the material’s elastic Contact Stiffness (S), which is the change in load with respect to per unit change in the indentation depth S = dP / dh. Reduced modulus Er can be calculated from elastic contact stiffness and contact area.

[0068] Discussion. Inorganic materials with optical, mechanical, and electronic properties have been widely discovered and used. For biomedical and bioelectronic applications, however, materials must also exhibit mechanical flexibility, softness, and biocompatibility. Proteins are an attractive class of biomolecules, but their molecular complexity has limited the understanding of their emergent properties in both natural and designed materials. Here, it was investigated whether the intrinsic chirality and non-centrosymmetry of macromolecular crystals could be harnessed to control symmetry-dependent multifunctional properties. However, symmetry control in protein crystals remains challenging, as protein crystallization is largely empirical, and existing assembly strategies typically aim to maximize symmetry to facilitate crystallization. To overcome this, DNA was used as a nanoscopic ligand to deliberately reduce symmetry without compromising structural order. Molecular dynamics simulations revealed a direct relationship between macromolecular packing and optical / electronic properties, and, unexpectedly, showed that piezoelectric behavior originates from the ordered water molecules within the crystal lattice, rather than from the protein itself. This finding shows that when the system gets increasingly complex, the traditional rules of piezoelectricity do not apply. Given the high porosity of macromolecular crystals, this mechanism may extend to other biomolecular systems and opens new avenues for the discovery of protein-based crystalline materials with emergent functionalities.

[0069] It has been demonstrated herein that modifying the protein surface topology with DNA results in a symmetry reduction of protein-DNA conjugates in both discrete and crystalline forms. DNA represents a unique class of synthons, enabling nanoscaleDocket No. 30938 / 2024-145 engineering of protein surfaces and effectively eliminating protein interfaces. While other macromolecules, such as synthetic polymers and glycans, could theoretically achieve similar functions, DNA offers distinct advantages in ease of design, chemical addressability, and programmability. With the support of computational tools, mechanistic details can be elucidated at the molecular or atomic level. Given the vast array of natural and designed proteins with diverse shapes, chemical structures, and functions,2627protein crystal engineering holds the potential to create biocompatible materials with advanced properties and functions that may extend far beyond their traditional role in structural biology.Materials and Methods

[0070] DNA synthesis. Oligonucleotides were coupled with 5'-maleimide-modifier phosphoramidite (Glen Research) on UnyLinker CPG solid supports (ChemGenes) using a MerMade 12 synthesizer (BioAutomation). 4,5-Dicyanoimidazole and iodine were used as the activator and oxidizer, respectively. Oligonucleotides were cleaved from the solid support using standard AMA deprotection [1 :1 mixture of 30% ammonium hydroxide (aq) and 40% methylamine (aq) for 25 min at 55°C] and purified using reversed-phase high-performance liquid chromatography (HPLC) (Agilent 1260 Infinity) equipped with an Agilent Dynamax Microsorb C18 column, with a gradient of 0 to 75% acetonitrile in triethylammonium acetate buffer over 45 min. With collected fractions, dimethoxytrityl (DMT) protecting groups were cleaved under 20% (v / v) acetic acid for 1 hour and extracted using ethyl acetate three times. The deprotected DNA strands were lyophilized overnight. The identities of the purified oligonucleotides were confirmed by matrix-assisted laser desorption ionization mass spectrometry (MALDI MS) using a Bruker MALDI Rapiflex Tissue Typer in linear negative mode, with 2',6'-dihydroxyacetophenone as the matrix and diammonium hydrogen citrate as the co-matrix. A complete list of the synthesized DNA sequences along with their experimentally determined masses is shown in Table 1.

[0071] Table 1 : DNA sequences and corresponding masses determined by MALDI-MSDocket No. 30938 / 2024-145

[0072] A retro-Diels-Alder reaction was performed to deprotect the maleimide moiety. The dried DNA strands were dissolved in anhydrous methanol and dried by rotary evaporation. The oligonucleotides were further dried through two co-evaporations with anhydrous acetonitrile, followed by three co-evaporations with anhydrous toluene. The DNA strands (10 pmol) was then suspended in 50 mL of anhydrous toluene and heated in an oil bath at 90 °C for six hours. After evaporating the toluene, the deprotected oligonucleotides were obtained as a white powder. The masses after retro-Diels-Alder reaction was determined using MALDI MS as shown in Table 1 .

[0073] Protein synthesis. Gene fragments encoding Synechococcus sp. CC9311 ferritin (SynFtn, PDB ID: 50UW) were obtained from Integrated DNA Technologies and inserted into a pET28a vector using Gibson assembly. The assembled plasmids were transformed into BL21 (DE3) electrocompetent cells (Sigma-Aldrich) via electroporation, and the cells were grown overnight on agar plates containing kanamycin (50 pg / mL). Single colonies were selected to inoculate 7-mL cultures in LB broth, and the plasmids were extracted for sequence confirmation via Sanger sequencing (T7 and T7 Reverse primers). Gene fragments used and protein sequences are shown in Table 2 and Table 3, respectively.

[0074] Table 2: Gene fragments used for Gibson assembly.Docket No. 30938 / 2024-145Docket No. 30938 / 2024-145Overlapping ends with pET28a vector are underlined.

[0075] Table 3: Protein sequences for the wild-type (WT) SynFtn and mutants.Mutations are underlined.

[0076] Subsequently, starter cultures (7 mL) were grown in 2x Yeast Extract / Tryptone / Potassium Phosphate (YTP) broth containing kanamycin (50 pg / mL) at 37°C until an optical density of 0.4 to 0.8 was reached. Protein overexpression was induced by the addition of 0.5 mM isopropyl-|3-d-thiogalactopyranoside (IPTG; Invitrogen) for the wildtype SynFtn (SEQ ID NOs: 5 and 8) and N161 C mutant (SEQ ID NOs: 7 and 10), or 0.2 mM IPTG for the T88C mutant (SEQ ID NOs: 6 and 9). After cultures were grown with shaking at 200 rpm overnight at 25 °C, the cells were harvested by centrifugation, resuspended in 20 mM Hepes buffer (pH 8.0), and lysed using a high-pressure homogenizer. Lysates were clarified by centrifugation at 16,000g for 25 min, and the supernatant was heated to 65eC in a water bath for 60 min. The heat-treated lysate was clarified by centrifugation at 8,000g for 25 min, and further digested by DNase I (New England Biolabs, 300 units per 45 mL lysate) at 37eC for 5 h. Digested DNA fragments were removed, and proteins were concentrated via five rounds of spin filtration using a 100 kDa cutoff ultra centrifugal filter (Amicon). Proteins were further purified using a HiLoad Superdex 200 pg preparative size exclusion chromatography column (Cytiva) equilibrated with 20 mM Hepes buffer (pH 8.0).Docket No. 30938 / 2024-145

[0077] DNA conjugation. Protein was reduced with 2 mM tris(2-carboxyethyl)phosphine (TCEP) for 20 min, and then desalted using ENrich SEC 650 columns (Bio-Rad) with 20 mM Hepes buffer (pH 7.0) as the mobile phase. The eluent was pooled, and the concentration of SynFtn monomer was adjusted to 150-200 pM. Maleimide-modified DNA (SEQ ID NOs: 2 or 4) was added in a 3-fold molar excess relative to the concentration of free cysteines, with sodium chloride added to a final concentration of 100 mM. The reaction was incubated overnight at room temperature and then analyzed using 4-15% sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) or 7.5% native polyacrylamide gel electrophoresis (PAGE). Excess DNA was removed through two rounds of spin filtration using a 100 kDa molecular weight cut-off ultracentrifugal filter (Amicon), and the reaction was further purified using ENrich SEC 650 columns (Bio-Rad) equilibrated with 20 mM Hepes buffer (pH 8.0).

[0078] Crystallization of proteins and protein-DNA conjugates. Proteins or protein- DNA conjugates were concentrated to 10 or 20 mg / mL and crystalized using the sitting-drop vapor-diffusion method. Crystallization screening was conducted in a 96-well plate (Intelli- Plate 96-3 well, Art Robbins Instruments) using a Crystal Gryphon liquid-handling robot (Art Robbins Instruments). The Helix and JCSG+ screens (Molecular Dimensions) were used to determine crystallization conditions. Each reservoir in the 96-well plate was filled with 70 pL of the crystallization solution, while each sitting drop consisted of 1 pL of the crystallization solution mixed with 1 pL of the protein or protein-DNA solution. The plates were incubated at 22 °C, undisturbed, for 2 to 4 weeks, and crystal formation was monitored under an optical microscope. Crystals were transferred to nylon loops and immediately flash-frozen in liquid nitrogen. Diffraction data were collected at National Synchrotron Light Source II (Brookhaven National Laboratory, USA) or European Synchrotron Radiation Facility (France).

[0079] Within the same crystallization trials, multiple crystals formed under different conditions. Many of these crystals were looped for diffraction experiments, and all datasets were indexed, scaled, and merged. The best-quality data, based on overall completeness, resolution, R factors, and l / sig(l) for structures with nearly identical unit cell parameters, were chosen for building and refining the protein models.

[0080] Structure determination and model refinement. Diffraction data were indexed, integrated, and scaled with autoPROC (Global Phasing Limited) and merged using Aimless (CCP4 Program Suite) with a resolution cut-off criterion of CC1 / 2 > 0.30 for most datasets. Structures were determined by molecular replacement with Phaser MR, using the structure of SynFtn (PDB: 50UW) as the template. The initial molecular replacement structures were refined with Refmac, and the protein models were further refined via successive rounds of manual model building in Coot and Refmac refinements until Rwork / Rfree were unchanged.Docket No. 30938 / 2024-145Metal coordination sites, water molecules, and ligands (if applicable) were built, and structures were validated against the wwPDB Validation System to refine model geometry, Ramachandran outliners, and rotamers in Coot.

[0081] Second harmonic generation. SHG imaging and PL spectra were taken using a Leica Dive Sp8 multiphoton confocal laser scanning microscope. The SHG signals of protein crystals in the solution were measured under 690-1040 nm radiation from built-in Physics Mai Tai tunable laser and under 700-1300 nm radiation from built-in Insight laser.

[0082] Nanoindentation. The protein crystals were transferred from the growth solution to a highly doped silicon substrate (University Wafer). Triboindenter (Hysitron, Minneapolis, MN) fitted with a Berkovitch tip (tip radius of 0.6 pm, angle of 142.3°) was used. During this test, the indenter came into contact with the sample surface with a trapezoidal load function, as defined by a loading time of 5 s, a holding time of 2 s at a maximum depth of 150 to 200 nm, and an unloading time of 5 s. This loading program was chosen to minimize the shortterm creep and size effect in cementitious materials.

[0083] Piezoelectricity. The protein crystals were transferred from the growth solution to a highly doped silicon substrate (University Wafer). The measurements were performed on a Bruker Dimension Icon Atomic Force Microscope. The fully metal-coated conductive tip (SCM-PIT V2) was used to perform electromechanical imaging in air. The Pt / lr coating probe (SCM-PIT V2) holds a spring constant (k) of ~3 N m-1 , and a resonant frequency of ~75 kHz. The piezoelectric response of protein crystals was measured by PFM Vertical and Horizontal Domains Operation workspaces within the NanoScope 9.40 R2 software. A contact mode PFM was employed with an AC voltage applied between the conductive probe and grounded sample.

[0084] Table 4: Sequence TableDocket No. 30938 / 2024-145Docket No. 30938 / 2024-145References: Lee, B. Y. et al. Virus-based piezoelectric energy generation. Nat Nanotechnol 7 , 351-356 (2012). Stapleton, A. et al. The direct piezoelectric effect in the globular protein lysozyme. Appl Phys Lett VI , (2017). Bera, S. et al. Molecular engineering of piezoelectricity in collagen-mimicking peptide assemblies. Nat Commun 12, (2021). O’Donnell, J. etal. Piezoelectricity of the Transmembrane Protein ba3 Cytochrome c Oxidase. Adv Fund Mater 31 , (2021). Haupert, L. M., Dewait, E. L. & Simpson, G. J. Modeling the SHG activities of diverse protein crystals. Ada Crystallogr D Biol Crystallogr 68, 1513-1521 (2012). Choi, T. S. & Tezcan, F. A. Overcoming universal restrictions on metal selectivity by protein design. Nature 603, 522-527 (2022). Zhang, L., Bailey, J. B., Subramanian, R. H. & Tezcan, F. A. Hyperexpandable, self-healing macromolecular crystals with integrated polymer networks. Nature 557, 86-91 (2018). Malay, A. D. et al. An ultra-stable gold-coordinated protein cage displaying reversible assembly. Nature 569, 438-442 (2019). Suzuki, Y. etal. Self-assembly of coherently dynamic, auxetic, two-dimensional protein crystals. Nature 533, 369-373 (2016). Ringler, P. & Schulz, G. E. Self-assembly of proteins into designed networks. Science 302, 106-109 (2003). Ramberg, K. O., Engilberge, S., Skorek, T. & Crowley, P. B. Facile Fabrication of Protein- Macrocycle Frameworks. J Am Chem Soc 143, 1896-1907 (2021).Docket No. 30938 / 2024-145 Dotan, N., Arad, D., Frolow, F. & Freeman, A. Self-assembly of a tetrahedral lectin into predesigned diamondlike protein crystals. Angewandte Chemie - International Edition 38, 2363-2366 (1999). Sakai, F. et al. Protein crystalline frameworks with controllable interpenetration directed by dual supramolecular interactions. Nat Common 5, 1-8 (2014). Yang, G. et al. Highly Ordered Self-Assembly of Native Proteins into 1 D, 2D, and 3D Structures Modulated by the Tether Length of Assembly-Inducing Ligands. Angewandte Chemie - International Edition 56, 10691 -10695 (2017). Padilla, J. E., Colovos, C. & Yeates, T. O. Nanohedra: Using symmetry to design self assembling protein cages, layers, crystals, and filaments. Proc Natl Acad Sci U SA 98, 2217-2221 (2001). Fletcher, J. M. etal. Self-assembling cages from coiled-coil peptide modules. Science 340, 595-599 (2013). Sciore, A. et al. Flexible, symmetry-directed approach to assembling protein cages. Proc Natl Acad Sci U S A AA 3, 8681-8686 (2016). Merg, A. D. et al. 2D Crystal Engineering of Nanosheets Assembled from Helical Peptide Building Blocks. Angewandte Chemie - International Edition 58, 13507-13512 (2019). Banatao, D. R. et al. An approach to crystallizing proteins by synthetic symmetrization. Proc Natl Acad Sci U S A 103, 16230-16235 (2006). Huddy, T. F. et al. Blueprinting extendable nanomaterials with standardized protein blocks. Nature (2024) doi : 10.1038 / s41586-024-07188-4. Li, Z. et al. Accurate computational design of three-dimensional protein crystals. Nat Mater (2023) doi:10.1038 / S41563-023-01683-1 . Ben-Sasson, A. J. et al. Design of biologically active binary protein 2D materials. Nature 589, 468-473 (2021). Partridge, B. E., Winegar, P. H., Han, Z. & Mirkin, C. A. Redefining Protein Interfaces within Protein Single Crystals with DNA. J Am Chem Soc 143, 8925-8934 (2021). Winegar, P. H. et al. DNA-Directed Protein Packing within Single Crystals. Chem 6, 1007- 1017 (2020). Han, K., Zhang, Z. & Tezcan, F. A. Spatially Patterned, Porous Protein Crystals as Multifunctional Materials. J Am Chem Soc 145, 19932-19944 (2023). Howarth, M. Say it with proteins: an alphabet of crystal structures. Nat Struct Mol Biol 22, 349-349 (2015). Kortemme, T. De novo protein design — From new structures to programmable functions. Ce / / vol. 187 526-544 Preprint at https: / / doi.Org / 10.1016 / j.cell.2023.12.028 (2024). Gito, D. A. D. et al. Self-Assembled Piezoelectric Films from Aligned Lysozyme Protein Fibrils. Biomacromolecules 26, 514-527 (2025). Churchfield, L. A. & Tezcan, F. A. Design and Construction of Functional Supramolecular Metalloprotein Assemblies. Acc Chem Res 52, 345-355 (2019). Laniado, J. & Yeates, T. O. A complete rule set for designing symmetry combination materials from protein molecules. Proceedings of the National Academy of Sciences 117, 31817-31823 (2020).

Claims

Docket No. 30938 / 2024-145What is claimed is:1 . A method of producing a protein crystal comprising:(i) conjugating one or more DNA strands to a surface of a wild-type (non-mutated) or mutated protein using a cross-linker to form a protein-DNA conjugate, wherein the mutated protein is a protein wherein one or more amino acids in or near the crystal protein-protein interaction (PPI) region are mutated; and(ii) crystallizing the protein-DNA conjugate to form DNA-modified protein crystals, wherein the mutated protein crystals have reduced symmetry when compared to protein crystals crystallized without the DNA conjugation, generated under similar conditions.

2. The method of claim 1 , wherein the symmetry of DNA-modified protein crystals, defined by the number of symmetry operators of a given space group, is reduced by at least 1 when compared to protein crystals derived from the wild-type protein generated under similar conditions.

3. The method of claim 1 or 2, wherein the non-mutated or mutated protein is a globular protein, fibrous protein, or membrane protein that can be crystalized.

4. The method of any one of claims 1-3, wherein the one or more amino acids that participate in the crystal PPI are mutated into cysteine.

5. The method of any one of claims 1-3, wherein the one or more amino acids that participate in the crystal PPI are mutated into cysteine, lysine, aspartic acid, glutamic acid, or another amino acid comprising an azide, carboxylic acid, sulfhydryl, carbonyl, or amine- containing natural or unnatural amino acid, or a combination thereof.

6. The method of any one of claims 1-5, wherein one or more of the following conditions are met:(i) the one or more amino acids mutated in or near the crystal PPI region are involved in disulfide bonding, salt bridging, electrostatic interactions, hydrogen bonding, covalent bonding, hydrophobic and / or van der Waals interactions;(ii) the one or more amino acids mutated in or near the crystal PPI region are surface-exposed;(iii) the one or more amino acids mutated in or near the crystal PPI region are at unstructured locations; and(iv) the one or more amino acids in or near the crystal PPI region are located along the symmetry axis of the protein molecule.Docket No. 30938 / 2024-1457. The method of any one of claims 1-6, wherein the DNA strands are self- complementary or complementary.

8. The method of any one of claims 1-7, wherein the DNA strands are between 1 and 50 nucleotides in length.

9. The method of claim 8, wherein the DNA strands are 10 nucleotides or less in length.

10. The method of claim 9, wherein the DNA strands are between 4 and 6 nucleotides in length.11 . The method of any one of claims 1-10, wherein DNA is attached to a mutated amino acid via a cross-linker.

12. The method of claim 11 , wherein the cross-linker is a PEG-based cross linker, and optionally wherein the PEG-based cross linker has no more than 10 PEG units.

13. The method of claim 11 , wherein the cross linker comprises one or more functional groups or chemical moieties including DBCO, azide, alkyne, phosphine, maleimides, thiol, pyridyl disulfies, haloacetyls, NHS-esters, imidoesters, carbodiimides, hydrazides, alkoxyamines, aryl azides, and / or diazirines.

14. The method of any one of claims 1-13, wherein 1 or more DNA strands are attached to the one or more mutated amino acids on the mutated protein, or to one or more amino acids on the wild-type protein.

15. The method of claim 14, wherein 1-500 DNA strands are attached to the one or more mutated amino acids on the mutant protein, or to one or more amino acids on the wild-type protein.

16. The method of any one of claims 1-15, wherein the DNA loading density ranges from 1-100 pmol / cm2.

17. The method of any one of claims 1-16, wherein the concentration of the protein-DNA conjugate to be crystalized ranges from 1 mg / mL to 50 mg / mL.

Images