METHOD FOR CONJUGATION OF BIOMOLECULES AND NEW USE OF GOLD DONOR FOR BIOMOLECULAR COMPLEX FORMATION

Abstract

The subject matter of the invention is a method for conjugation of free thiol group(s) containing biomolecules, leading to the biomolecular complex formation, comprising a reaction to connect biomolecules using a gold-donor agent in which a —S—Au—S— bond is formed, characterised in that a gold-donor agent is halogen(triarylphosphine)gold (I). The subject matter of the invention is also the use of halogen(triarylphosphine)gold (I) molecules as the gold-donor agent in the method of biomolecular complex formation.

Claims

1-22. (canceled)

23. A method for conjugating a free thiol group of a moiety of a biomolecule, comprising contacting the biomolecule with a gold-donor agent to form a —S—Au—S— bond, characterised in that the gold-donor agent is halogen(triarylphosphine)gold (I).

24. The method of claim 23, wherein the biomolecule is selected from the group consisting of peptides, polypeptides, and proteins.

25. The method of claim 23, wherein a conjugated complex composed of multiple units of the same biomolecule is formed.

26. The method of claim 25, wherein the complex is symmetric.

27. The method of claim 23, wherein the moiety is cysteine.

28. The method of claim 27, wherein cysteine moiety is a naturally occurring moiety in the biomolecule.

29. The method of claim 27, wherein cysteine moiety is artificially introduced into the biomolecule.

30. The method of claim 23, wherein the halogen is selected from the group consisting of chloro, bromo, iodo, and fluoro; and wherein the aryl is selected from the group consisting of unsubstituted phenyl- or ortho-, meta- or para- mono or polysubstituted phenyl.

31. The method of claim 23, wherein the gold-donor agent is chloro[diphenyl(3-sulfonatophenyl)phosphine]gold (I).

32. The method of claim 23, wherein gold-donor agent is chloro(triphenylphosphine)gold (I).

33. The method of claim 23, further comprising purifying the conjugation product.

34. The method of claim 33, wherein the biomolecule is prepared by expression in a suitable expression system and purification of the expression product prior to conjugation.

35. The method of claim 34, wherein at least one cysteine is introduced into the biomolecule.

36. The method of claim 33, wherein the conjugation is performed in aqueous solution, at room temperature, for up to 3 days and the molar ratio of biomolecule:gold-donor is from 3:1 to 1:4

37. The method of claim 33, wherein the purification of the conjugation product is performed by a method comprising at least one of filtration, crystallization, centrifugation, and column chromatography.

38. The method of claim 23, wherein the biomolecule complex is a protein cage.

39. The method of claim 38, wherein the biomolecule is a TRAP protein.

40. The method of claim 38, wherein the protein complex consists of 24 biomolecule units.

41. A modified protein cage obtainable by the method of claim 1.

42. A modified protein cage comprising 24 TRAP rings.

43. The modified protein cage of claim 42 wherein the TRAP protein contains a K35C mutation.

44. The modified protein cage of claim 43 wherein the TRAP protein additionally contains a R64S mutation.

45. The modified protein cage of claim 42 wherein the cage is held together by linear, —S—Au—S— coordinate bonds between sulphurs of 240 of the 264 available cysteines in the cage.

Description

BRIEF DESCRIPTION OF THE FIGURES

[0042] FIG. 1: illustrates examples of gold(I)-containing compounds for gold-stitching reaction.

[0043] FIG. 1A. Monosulfonated triphenylyphosphine gold(I) chloride.

[0044] FIG. 1B. Triphenylyphosphine gold(I) chloride.

[0045] FIG. 2. illustrates the gold-stitching reaction and an example of the resulting complex formed.

[0046] FIG. 2A. Structure of a single TRAP ring (pdb 4v4f) shown in two mutually orthogonal views.

[0047] FIG. 2B. Pseudo-atomic model of the structure of the hollow cage structure.

[0048] FIG. 2C. The obtained cryo EM densities for the TRAP cage.

[0049] FIG. 2D. Close up views of two neighbouring TRAP rings in the produced protein cage showing the presence of the bridging gold atom.

[0050] FIG. 2E. Chemical bond formed between opposing cysteine side chains by the action of monosulfonated triphenylyphosphine gold(I) chloride, “R” refers to the remainder of the TRAP protein.

[0051] FIG. 3A: illustrates LC-MS data for three forms of TRAP monomer: (from left) unliganded protein (dark grey), monomer bound to a single gold atom (grey), and monomer bound to a gold atom and TPPMS ligand (light grey).

[0052] FIG. 3B: illustrates Native MS of intact TRAP cages performed at high collisional activation.

[0053] FIG. 3C: Expansion of low-m/z region in FIG. 3B.

[0054] FIG. 4 shows stability of protein complexes held together by the results of gold-stitching reaction.

[0055] FIG. 4A illustrates native PAGE gel showing the high thermal stability of the formed TRAP-cage.

[0056] FIG. 4B illustrates TEM images of the formed TRAP cage without heat treatment (scale bar 200 nm).

[0057] FIG. 4C: shows TEM images of the formed TRAP cage after 3 hour incubation at 95° C. showing no significant degradation of the cage structure (scale bar 100 nm).

EXAMPLES

[0058] Techniques Employed in the Realisation of the Invention

[0059] Transmission Electron Microscopy (TEM)

[0060] Samples were typically diluted to a final protein concentration of 0.025 mg/ml, centrifuged briefly in a desktop centrifuge and the supernatant applied onto hydrophilized carbon-coated copper grids (STEM Co.), negatively stained with 4% phosphotungstic acid, pH 8, and visualized using a JEOL JEM-1230 80 kV instrument.

[0061] Native PAGE

[0062] Samples were run on 3-12% native Bis-Tris gels following the manufacturer's recommendations (Life Technologies). Samples were mixed with 4× native PAGE sample buffer (200 mM BisTris, pH 7.2, 40% w/v Glycerol, 0.015% w/v Bromophenol Blue). As a qualitative guide to molecular weights of migrated bands, NativeMark unstained protein standard (Life Technologies) was used. Where blue native PAGE was performed, protein bands were visualized according to the manufacturer's protocol (Life Technologies), otherwise InstantBlue™ protein stain (Expedeon) was used.

[0063] Electrothermal Atomic Absorption Spectrometry (ETAAS)

[0064] A sample mass of approx. 2 mg was dissolved in 25 ml with 0.2% HCl. The solution was then diluted 25× before determination of total Au performed by an ETAAS spectrometer (PinAAcle 900Z, Perkin Elmer, Waltham, Mass.), with Zeeman background correction, at a wavelength of 242.80 nm (slit width of 0.7 nm). The measured volume of the sample solution was 10 μl and to each sample a mixture of matrix modifiers: 5 μg of Pd(NO.sub.3).sub.2 and 3 μg of Mg(NO.sub.3).sub.2 was added. 5 sets of measurements were carried out with each set consisting of 3 repeats.

[0065] Protein Expression and Purification

[0066] In a typical purification, E. coli BL21(DE3) cells (Novagen) transformed with pET21b plasmid harboring the TRAP-CS gene were grown at 37° C. with shaking in 3 L of LB medium with 100 μg/ml ampicillin until OD.sub.600=0.6, induced with 0.5 mM IPTG then further shaken for 4 h. Cells were harvested by centrifugation and the pellet kept at −80° C. until use. Cells were lysed by sonication at 4° C. in 50 ml of 50 mM Tris-HCl, pH 7.9, 50 mM NaCl in presence of proteinase inhibitors (Thermo Scientific) and presence or absence of 2 mM DTT, and lysates were centrifuged at 66,063 g for 0.5 h at 4° C. The supernatant fraction was heated at 70° C. for 10 min, cooled to 4° C., and centrifuged again at 66,063 g for 0.5 h at 4° C. The supernatant fraction was purified by ion exchange chromatography on an ÄKTA purifier (GE Healthcare Life Sciences) using 4×5 ml HiTrap QFF columns with binding in 50 mM Tris-HCl, pH 7.9, 0.05 M NaCl, +/−2 mM DTT buffer and eluting with a 0.05-1 M NaCl gradient. Fractions containing TRAP protein were pooled and concentrated using Amicon Ultra 10 kDa MWCO centrifugal filter units (Millipore) and the sample subjected to size exclusion chromatography on a HiLoad 16/60 Superdex 200 column in 50 mM Tris-HCl, pH 7.9, 0.15 M NaCl at room temperature. Protein concentrations were calculated using the BCA protein assay kit (Pierce Biotechnology).

[0067] Liquid-Chromatography Mass Spectrometry

[0068] TRAP cage sample was denatured in 50 mM Tris⋅HCl buffer (pH 8.0) with 8 M urea at 56° C. for 30 min, then buffer-exchanged to 50 mM Tris⋅HCl buffer (pH 8.0) using a centrifugal filtration device (Amicon 3 kDa MWCO, Millipore). For denaturing LC-MS analysis, the TRAP protein was desalted on a C18 pre-column (Acclaim PepMap100, C18, 300 μm×1 cm; Thermo Scientific), then separated on a C18 column (Acclaim PepMap100, C18, 75 μm×15 cm; Thermo Scientific) by Dionex UltiMate 3000 RSLCnano System connected to a hybrid LTQ Orbitrap XL mass spectrometer (Thermo Scientific) via a dynamic nanospray source. A binary buffer system was used, with buffer A 0.1% formic acid in H.sub.2O, and buffer B 0.1% formic acid in acetonitrile. The proteins were separated at 25° C. with a gradient of 1% to 90% buffer B at a flow rate of 300 nL min.sup.−1 over 60 min. The LTQ-Orbitrap XL was operated in positive ion mode with a nanoelectrospray voltage of 1.6 kV and capillary temperature of 275° C. Survey full-scan MS spectra were acquired in the orbitrap (m/z 300-4000) with a resolution of 60000. The data were processed using Xcalibur 2.2 (Thermo Scientific).

[0069] Native Mass Spectrometry

[0070] TRAP cage samples at 0.8 mg ml.sup.−1 were prepared for native MS by buffer-exchanging into ammonium acetate (pH 6.9) using miniature spin columns (Micro Bio-Spin P-6, BioRad). This was performed in two steps: the first exchanged into 2.5 M ammonium acetate, the second into 200 mM ammonium acetate. Native MS experiments were performed using methods described previously.sup.14, employing a Q-ToF2 instruments (Waters Corp.), modified for the analysis of large protein ions.sup.15. Relevant instrument parameters were: nanoelectrospray capillary voltage: 1.9 kV; sample cone: 200 V; extractor cone: 10 V, acceleration into collision cell: 200 V. The collision cell was pressurized with argon at ≈35 μbar. Data was calibrated externally using MassLynx software (Waters Corp.), and are shown without background subtraction and minimal smoothing.

Example 1

[0071] TRAP Complex Preparation—Reaction with Au-TPPMS

[0072] (see FIG. 1)

[0073] Gold Compounds:

[0074] Chloro[diphenyl(3-sulfonatophenyl)phosphine]gold (I), sodium salt hydrate (Au-TPPMS, MDL number MFCD19443491) was purchased from STREM chemicals UK, limited and was made up to the desired concentration (typically 5 mM) by dissolving in water. The gold nanoparticle (GNP) used was a diphenyl(m-sulfonatophenyl)phosphine-gold nanocluster with a 1-3 nm core diameter (MDL number MFCD17018839) from STREM Chemicals UK.

[0075] TRAP Preparation:

[0076] The protein used that exemplifies the successful use of Au-TPPMS was TRAP protein with an introduced cysteine. Expression and purification of TRAP containing the mutation of residue lysine (K) number 35 to cysteine and an additional mutation of residue arginine (R) 64 to serine (S) (called “TRAP-CS”) was similar to as described previously for TRAP-CS.sup.11 (and as detailed above) with the notable change that TCEP (tris(2-carboxyethyl)phosphine) was not included in the lysis step. The final buffer was typically 20 mM Tris-HCl, pH 8.0, 0.15 M NaCl

[0077] Reaction of Modified TRAP Protein with Au(I)-TPPMS.

[0078] Purified TRAP protein was reacted with Au-TPPMS (FIG. 1A) in aqueous buffer at room temperature. S—Au—S bonds were formed in the reaction resulting in assembly of TRAP rings into the TRAP-cage which was then purified and further characterised.

[0079] Formation of TRAP-cage was carried out by mixing purified TRAP-CS and Au-TPPMS in aqueous solution. The exact concentrations of reactants were tailored for each reaction but were typically as follows: 1 mM TRAP-CS (8.3 mg ml.sup.−1) and 1 mM Au-TPPMS in 50 mM Tris-HCl, pH 7.9, 0.15 M NaCl. Reactions were incubated for at least 3 days at room temperature. Formation of TRAP-cage was confirmed using TEM and native PAGE. Any precipitated material was removed by centrifugation at 12 045×g for 5 min, and TRAP-cage was purified by size exclusion chromatography on either Superose 6 Increase 10/300 GL or HiPrep 16/60 Sephacryl S-500 HR column (GE Healthcare) or a HiLoad 16/600 Superdex 200 pg. Fractions containing the cage protein were pooled, concentrated using Amicon Ultra 0.5 100 kDa MWCO, and protein concentrations were measured using the BCA protein assay (Pierce Biotechnology).

Example 2

[0080] TRAP Complex Preparation—Reaction with Au(I)-Triphenylphosphine

[0081] A similar reaction can be carried out with triphenylyphosphine gold(I) chloride (Au-TPP, FIG. 1B) instead of Au-TPPMS. Au-TPP is dissolved in DMSO and the reaction is carried out in 50 mM Tris, 150 mM NaCl, pH 7.9 and Au-TPP, with DMSO not exceeding 10% of the final volume and the TRAP monomer is present at a concentration of approximately 1 mM. Ratios of Au-TPP to TRAP range from 4:2 to 4:3. This also results in TRAP-cage formation with the same structure as obtained in the reaction with Au-TPPMS. The amounts of the reagents were adjusted respectively. The ratio of the TRAP monomer/gold donor and conditions of the reaction were the same as in the reaction where Au-TPPMS was the gold-donor.

[0082] Two examples with different halogen(triarylphosphine)gold (I) gold-donor agents were performed above. It shows that halogen(triarylphosphine)gold (I) with different aryl moiety are suitable for the complex formation according to the invention.

Example 3

[0083] Confirmation of TRAP-Complex Structure Using Cryo-EM

[0084] The initial (low resolution) cryo-EM structure of TRAP-cage was obtained using cryo-EM single particle reconstruction techniques for TRAP-cage formed using GNPs..sup.10,11 and this structural data was used as an initial model for solving the high-resolution cryo-EM structure of TRAP-cage formed in the reaction with halogen(triarylphosphine)gold (I) obtained according to the invention.

[0085] Cryo-EM was used to solve the structure of the TRAP-cage to 3.9 Angstrom resolution. This was sufficient to show the arrangement of the 24 TRAP rings and to demonstrate the presence of a linking density (assigned to Au) between opposing cysteine side chains of the rings (See FIG. 2).

[0086] FIG. 2. illustrates the gold-stitching reaction and an example of the resulting complex formed. FIG. 2A shows structure of a single TRAP ring (pdb 4v4f) shown in two mutually orthogonal views. Mutated residues 35 and 64 are shown as spheres on each TRAP monomer with residue 35 being on the outer perimeter of the ring. Reaction with gold(I)-containing compound (arrow) results in structure shown in FIG. 2B—pseudo-atomic model of the structure of the hollow cage structure. Here each of the 24 rings is shown in cartoon format with cysteines bridging the rings shown as sticks with spheres representing gold atoms between them. This model is built using structure illustrate in FIG. 2C that shows the obtained cryo EM densities. FIG. 2D is a close up views of two neighbouring TRAP rings in the produced protein cage showing the presence of the bridging gold atom. Cryo-EM map is shown as a grey net and protein is shown in cartoon format with cysteine residues shown as sticks. Four cysteine residues are highlighted by arrows and their bridging gold atoms are shown as spheres. Refined distances between Au—S linkages are indicated in FIG. 2E where chemical bond formed between opposing cysteine side chains by the action of monosulfonated triphenylyphosphine gold(I) chloride, “R” refers to the remainder of the TRAP protein, is shown.

[0087] Cryo-EM Single Particle Reconstruction of TRAP-Cage Formed Using Au-TPPMS at Higher Resolution

[0088] Purified sample (3 μl of 0.89 mg ml.sup.−1) formed using Au-TPPMS was applied to glow-discharged holey carbon grids (Quantifoil R 1.2/1.3, Mo 200 mesh) with a thin amorphous carbon film of .sup.˜10 nm thickness over the holes and incubated for 30 s at 4° C. and 100% humidity. Grids were then blotted for 3.0 s and plunged into liquid ethane using a Vitrobot Mark IV (FEI). Data were recorded semi-automatically using the EPU software on a transmission electron cryo-microscope (FEI Titan Krios) operated at an accelerating voltage of 300 kV and at a nominal magnification of 75,000×. Images (0.91 Å/pixel) were recorded at applied underfocus values ranging from approximately −0.9 to −3.4 μm on a Falcon II direct electron detector (FEI) as 32 frames in 2.0 s exposure with a total electron dose of 40 electrons/Å.sup.2. Data were subsequently aligned and summed using MotionCor2.sup.21 to obtain a final dose weighted image and then 2× binning was performed using the Bsoft program package,.sup.22 resulting in a pixel size of 1.82 Å for further image processing. Estimation of the contrast transfer function was performed using CTFFIND4..sup.23 Micrographs exhibiting poor power spectra based on the extent and regularity of the Thon rings were rejected (96 micrographs). Initially, approximately 2,000 particles were manually picked and subjected to reference-free two-dimensional (2D) classification using EMAN 2.1..sup.18 Ten representative 2D class averages were selected as templates for automated particle picking using Gautomatch (http://www.mrc-lmb.cam.ac.uk/kzhang/). All subsequent processing steps were performed in RELION 2.0..sup.20 A total of 1,085,623 auto-picked particles from 10,290 micrographs were subjected to reference-free 2D classification to remove aberrant particles. Particles in 5 representative classes showing spherical shapes were selected (578,865 particles) for the following processes. The selected particles were subjected to three-dimensional (3D) classification into three classes using an angular sampling of 3.7° for 25 iterations without any symmetry (C1 symmetry), where the initial low-resolution structure as described above was used for the reference in the 3D classification after low-pass filtered to 60 Å. The particles (176,463 particles) in a class showing the most symmetrical cage structure with regular density distribution were selected for the following processes. However, although the density map clearly showed the overall TRAP-cage structure as a sphere with 24 11-membered rings, the structure at the level of the individual rings was curiously devoid of protein chiral features and showed mixed features of two mirrored protein structures, contrary to expectations from the protein structure previously determined by x-ray crystallography,.sup.24 which is suggestive of the existence of chiral cage structures. Therefore, to separate the two chiral cage particles, we performed a second round of 3D classification into two classes using a finer angular sampling of 1.8° for 25 iterations without any symmetry (C1 symmetry). The resultant two maps clearly showed left-handed and right-handed structures at the level of the individual protein rings, respectively. Each structure (class I: 94,338 particles and class II: 82,125 particles) was refined individually with the C1 (asymmetric reconstruction), C4 and D4 symmetries. The resolutions of the class I were estimated to 3.9 (D4 sym.), 4.1 (C4 sym.), and 4.4 Å (C1 sym.) and the resolutions of the class II were estimated to 3.9 (D4 sym.), 4.2 (C4 sym.), and 4.5 Å (C1 sym.) by the gold-standard Fourier shell correlation (FSC=0.143 criterion), after applying a soft spherical mask on the two reconstructions refined from the half of the data sets independently. According to the individual protein structures, the handedness of the class I map was corrected to the opposite one (resulting in class I: right-handed cage structures and class II: left-handed cage structures). The maps of the class I and II were sharpened with B-factors of −229 and −231 Å.sup.2, respectively. Local resolution was estimated using ResMap..sup.25 Figures were prepared using UCSF Chimera..sup.26

[0089] Structural Refinement

[0090] The initial atomic coordinate model was based on the TRAP crystal structure (PDB accession 4V4F.sup.9), with the Cys.sup.35 and Ser.sup.64 substitutions modelled in Coot.sup.27 to generate TRAP-CS ring structures. Note that residue positions have been renumbered from the initial deposited PDB to reflect the actual positions in the coding sequence of TRAP from G. stearothermophilus (e.g. the mutated Lys->Cys residue was assigned to residue number 37 in the original PDB file 4V4F but corresponds to residue number 35 in our analyses). Initial inspection of the density maps revealed areas of weak or missing density, and thus the structure of each TRAP subunit was truncated to residues 6-72; in addition residues 22-32 (corresponding to a loop that exhibits high flexibility in the apo-form of TRAPS) were omitted from the model to reflect this. Refinement of the LH and RH structures followed a similar regime. Twenty-four copies of TRAP-CS rings were initially fit into the cage density by rigid body refinement using Phenix real-space refinement..sup.28 Optimization of the original cryo-EM map voxel size using the high-resolution TRAP crystal structure.sup.24 as a reference was performed as follows, in a manner analogous to previous reports..sup.29,30 Comparison of cross-correlation scores of the fits between a simulated map of the TRAP-CS ring atomic model and the cryo-EM map at varying voxel scales (starting from the original 1.82 Å voxel.sup.−1 and varying by 0.01 increments) was performed using Chimera, with the optimal results corresponding to a map scale of 1.74 Å voxel.sup.−1. Similar results were obtained by performing rigid body refinement of individual subunits of 24 TRAP-CS rings onto the cryo-EM density at varying scales using Phenix..sup.28 Au.sup.I atoms (120 in total) were docked manually into the prominent blobs of density between the Cys.sup.35 side chains from neighbouring rings of the rigid-body fitted model, and subsequently 15 macro cycles of Phenix real-space refinement were run using the 1.74 Å voxel.sup.−1 map, including rigid-body refinement, global minimization, a single round of simulated annealing, and adp refinement; restraints on the Au—S bond lengths and S—Au—S bond angles were applied during the later stages of refinement. Validation of the refined models was carried out using MolProbity..sup.31 Analysis of interfacial contacts in the TRAP-cage models was performed using PDBePISA (http://www.ebi.ac.uk/pdbe/pisa/)..sup.32

[0091] Mass Spectometry:

[0092] Mass spectrometry was further used to support the presence of a gold atom linking TRAP monomers within the TRAP cage structure.

[0093] The results of mass spectrometry experiments are presented in the FIG. 3A that illustrates LC-MS data for three forms of TRAP monomer: (from left) unliganded protein (dark grey), monomer bound to a single gold atom (grey), and monomer bound to a gold atom and TPPMS ligand (light grey). Only 10+ charge states are show for clarity, and magnifications of the different peaks allows accurate mass determination for unambiguous assignment. The other, minor peaks correspond to salt adducts and/or other charge states. Inset table shows list of TRAP masses, and the mass additions expected due to the different modifications. These correspond very well to the masses measured, taking into account the 10 protons responsible for the 10+ charge state. FIG. 3B illustrates Native MS of intact TRAP cages performed at high collisional activation. Native MS of intact TRAP-cages performed at high collisional activation reveals a broad, unresolved region of signal at high m/z, and a series of peaks at low m/z. These features correspond to dissociation of intact cages, resulting in the release of cage fragments. illustrates expansion of low-m/z region. FIG. 3C shows expansion of low-m/z region expressed in FIG. 3B, showing assignment of the various charge state series. Monomeric TRAP, in both modified and unmodified forms (greys, same colouring as A), are the major fragments observed. Mass spectra described above shows notably, peaks that can be assigned unambiguously to a TRAP dimer containing a single gold atom are observed, validating the TRAP-Au(I)-TRAP linkage hypothesis.

[0094] Electrothermal atomic absorption spectrometry (ETAAS) showed 112±8 gold atoms per cage assembly, in close agreement with the predicted value of 120 (Table 1 below).

TABLE-US-00001 TABLE 1 Determination of Au content of TRAP cages using ETAAS Table 1: Determination of Au content of TRAP cages using ETAAS Gold SD of Equivalent Measure- concentration measurement number of SD in no. ment (mg/g) (mg/g) Au per cage Au per cage 1 9.34 0.6 103 7 2 9.23 0.3 102 3 3 9.58 1.0 106 11 4 10.99 0.2 121 2 5 11.12 0.3 123 3 Results of 5 ETAAS measurements of TRAP-cage, each performed in triplicate, showing the measured mass of gold and its translation into number of gold atoms per TRAP-cage. Measurement 3 was discarded in calculation of overall averages due to the large observed error.

[0095] The Experiments above confirmed the structure of the TRAP-complex. The structure of the TRAP-complex obtained in the reaction with halogen(triarylphosphine)gold (I) is the same as obtained in the reaction with GNPs that was described before in the inventors' paper.

Example 4

[0096] Stability Tests of TRAP Complex

[0097] Thermostability tests were performed as follows. Samples (7.5 μl) containing 1 μg TRAP cage protein in aqueous buffer were heated to 95° C. for different times (0-180 mins). After heating, samples were centrifuged at 10,000 rpm for 5 minutes in a bench-top centrifuge. Supernatants were taken and mixed with 2.5 μl of 4× NativePAGE sample buffer and the samples subjected to native PAGE analysis, the same sample were further analysed by TEM Typical results are shown in FIG. 4 A-C.

[0098] The stability of protein complexes held together by the results of gold-stitching reaction is presented in FIG. 4, in which. FIG. 4A illustrates Native PAGE gel showing the high thermal stability of the formed TRAP-cage. Samples were treated at the indicated temperatures for the indicated times before being applied to the gel. The band corresponding to the TRAP cage is indicated by the arrow. C=Control lane (no heat treatment). M=protein molecular weight marker (weights shown in kDa on left hand side of gel). FIG. 4B shows TEM images of the formed TRAP cage without heat treatment (scale bar 200 nm). FIG. 4C illustrates TEM images of the formed TRAP cage after 3 hour incubation at 95° C. showing no significant degradation of the cage structure (scale bar 100 nm).

REFERENCES

[0099] 1 Antson, A. A. et al. Structure of the trp RNA-binding attenuation protein, TRAP, bound to RNA. Nature 401, 235-242 (1999). [0100] 2 Antson, A. A. et al. The structure of trp RNA-binding attenuation protein. Nature 374, 693-700 (1995). [0101] 3 Chen, X. et al. Regulatory features of the trp operon and the crystal structure of the trp RNA-binding attenuation protein from Bacillus stearothermophilus. J. Mol. Biol. 289, 1003-1016, doi:10.1006/jmbi.1999.2834 S0022283699928346 [pii] (1999). [0102] 4 Watanabe, M. et al. The nature of the TRAP-Anti-TRAP complex. Proc. Natl. Acad. Sci. USA 106, 2176-2181 (2009). [0103] 5 Malay, A. D., Watanabe, M., Heddle, J. G. & Tame, J. R. H. Crystal structure of unliganded TRAP: implications for dynamic allostery. Biochem. J. 434, 429-434 (2011). [0104] 6 Ballister, E. R., Lai, A. H., Zuckermann, R. N., Cheng, Y. & Mougous, J. D. In vitro self-assembly of tailorable nanotubes from a simple protein building block. Proc. Natl Acad. Sci. USA 105, 3733-3738 (2008). [0105] 7 Wason, A., Pearce, F. G., Gerrard, J. A. & Mabbutt, B. C. Archaeal Lsm rings as stable self-assembling tectons for protein nanofabrication. Biochem. Biophys. Res. Commun. 489, 326-331, doi:10.1016/j.bbrc.2017.05.129 (2017). [0106] 8 Miranda, F. F. et al. A Self-Assembled Protein Nanotube with High Aspect Ratio. Small 5, 2077-2084 (2009). [0107] 9 Nagano, S. et al. Understanding the Assembly of an Artificial Protein Nanotube. Adv. Mater. Interfaces 3 (2016). [0108] 10 Malay, A. D. et al. Gold Nanoparticle-Induced Formation of Artificial Protein Capsids. Nano Lett. 12, 2056-2059 (2012). [0109] 11 Imamura, M. et al. Probing structural dynamics of an artificial protein cage using high-speed atomic force microscopy. Nano Lett. 15, 1331-1335, doi:10.1021/n15045617 (2015). [0110] 12 Pan, Y. et al. Gold Nanoparticles of Diameter 1.4 nm Trigger Necrosis by Oxidative Stress and Mitochondrial Damage. Small 5, 2067-2076, doi:10.1002/smll.200900466 (2009). [0111] 13 Pan, Y. et al. Size-Dependent Cytotoxicity of Gold Nanoparticles. Small 3, 1941-1949, doi: 10.1002/smll.200700378 (2007). [0112] 14 Kondrat, F. D., Struwe, W. B. & Benesch, J. L. Native mass spectrometry: towards high-throughput structural proteomics. Methods in molecular biology 1261, 349-371, doi:10.1007/978-1-4939-2230-7_18 (2015). [0113] 15 Sobott, F., Hernandez, H., McCammon, M. G., Tito, M. A. & Robinson, C. V. A tandem mass spectrometer for improved transmission and analysis of large macromolecular assemblies. Anal Chem 74, 1402-1407 (2002). [0114] 16 Li, X. et al. Electron counting and beam-induced motion correction enable near-atomic-resolution single-particle cryo-EM. Nat. Methods 10, 584-590, doi:10.1038/nmeth.2472 (2013). [0115] 17 Yasunaga, T. & Wakabayashi, T. Extensible and object-oriented system Eos supplies a new environment for image analysis of electron micrographs of macromolecules. J. Struct. Biol. 116, 155-160 (1996). [0116] 18 Tang, G. et al. EMAN2: an extensible image processing suite for electron microscopy. J. Struct. Biol. 157, 38-46, doi:10.1016/j.jsb.2006.05.009 (2007). [0117] 19 Ludtke, S. J., Baldwin, P. R. & Chiu, W. EMAN: semiautomated software for high-resolution single-particle reconstructions. J. Struct. Biol. 128, 82-97 (1999). [0118] 20 Scheres, S. H. RELION: implementation of a Bayesian approach to cryo-EM structure determination. J. Struct. Biol. 180, 519-530, doi:10.1016/j.jsb.2012.09.006 (2012). [0119] 21 Zheng, S. Q. et al. MotionCor2: anisotropic correction of beam-induced motion for improved cryo-electron microscopy. Nat. Methods 14, 331-332, doi:10.1038/nmeth.4193 (2017). [0120] 22 Heymann, J. B. Bsoft: image and molecular processing in electron microscopy. J. Struct. Biol. 133, 156-169, doi:10.1006/jsbi.2001.4339 (2001). [0121] 23 Rohou, A. & Grigorieff, N. CTFFIND4: Fast and accurate defocus estimation from electron micrographs. J. Struct. Biol. 192, 216-221, doi:10.1016/j.jsb.2015.08.008 (2015). [0122] 24 Hopcroft, N. H. et al. The interaction of RNA with TRAP: the role of triplet repeats and separating spacer nucleotides. J. Mol. Biol. 338, 43-53 (2004). [0123] 25 Kucukelbir, A., Sigworth, F. J. & Tagare, H. D. Quantifying the local resolution of cryo-EM density maps. Nat. Methods 11, 63-65, doi:10.1038/nmeth.2727 (2014). [0124] 26 Pettersen, E. F. et al. UCSF Chimera—a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605-1612, doi: 10.1002/jcc.20084 (2004). [0125] 27 Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta. Cryst. D 66, 486-501, doi:10.1107/S0907444910007493 (2010). [0126] 28 Adams, P. D. et al. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta. Cryst. D 66, 213-221, doi:10.1107/S0907444909052925 (2010). [0127] 29 Wang, Z. et al. An atomic model of brome mosaic virus using direct electron detection and real-space optimization. Nat. Commun. 5, 4808, doi:10.1038/ncomms5808 (2014). [0128] 30 Natchiar, S. K., Myasnikov, A. G., Kratzat, H., Hazemann, I. & Klaholz, B. P. Visualization of chemical modifications in the human 80S ribosome structure. Nature 551, 472-477, doi: 10.1038/nature24482 (2017). [0129] 31 Chen, V. B. et al. MolProbity: all-atom structure validation for macromolecular crystallography. Acta. Cryst. D 66, 12-21, doi:10.1107/S0907444909042073 (2010). [0130] 32 Krissinel, E. & Henrick, K. Inference of macromolecular assemblies from crystalline state. J. Mol. Biol. 372, 774-797, doi:10.1016/j.jmb.2007.05.022 (2007).