“Silent” Amino Acid Residues at Key Subunit Interfaces Regulate the

“Silent” Amino Acid Residues at Key Subunit Interfaces Regulate the Geometry of Protein Nanocages Shengli Zhang,†,§ Jiachen Zang,†,‡,§ Xiaorong Zhang,† Hai Chen,† Bunzo Mikami,*,‡ and Guanghua Zhao*,† †

Beijing Advanced Innovation Center for Food Nutrition and Human Health, College of Food Science and Nutritional Engineering, China Agricultural University, Key Laboratory of Functional Dairy, Ministry of Education, Beijing 100083, China ‡ Laboratory of Applied Structural Biology, Division of Applied Life Sciences, Graduate School of Agriculture, Kyoto University, Gokasho, Uji, Kyoto 611-0011, Japan S Supporting Information *

ABSTRACT: Rendering the geometry of protein-based assemblies controllable remains challenging. Protein shelllike nanocages represent particularly interesting targets for designed assembly. Here, we introduce an engineering strategykey subunit interface redesign (KSIR)that alters a natural subunit−subunit interface by selective deletion of a small number of “silent” amino acid residues (no participation in interfacial interactions) into one that triggers the generation of a non-native protein cage. We have applied KSIR to construct a non-native 48-mer nanocage from its native 24-mer recombinant human H-chain ferritin (rHuHF). This protein is a heteropolymer composed of equal numbers of two different subunits which are derived from one polypeptide. This strategy has allowed the study of conversion between protein nanocages with different geometries by re-engineering key subunit interfaces and the demonstration of the important role of the above-mentioned specific residues in providing geometric specificity for protein assembly. KEYWORDS: “silent” amino acid residues, subunit interface redesign, geometry regulation, nanocage, ferritin reassembly

P

conditions and provide a controlled microenvironment, which is chemically well-defined by virtue of their high symmetrical nature. However, nanotechnologists have subverted these nature functions and used self-assembled cage structures of nanometer dimensions for the preparation of nanomaterials and encapsulation of guest molecules with potential applications14−25 due to their high symmetry, solubility, stability, and monodispersity. Protein cages represent a standard structural component in ferritin.26 Ferritins are superfamily proteins that self-assemble into hollow cage-like structures which are ubiquitously found in both prokaryotes and eukaryotes. The ferritins can selfassemble into two types of nanocages: maxi-ferritins and mini-ferritins. The classical ferritins are considered maxiferritins, whereas DNA-binding proteins from starved cells (Dps) are mini-ferritins.27 All classical ferritins share the highly

rotein self-assembly is ubiquitous and vitally important in nature. Subunit−subunit interactions (SSIs) are required to drive assembly in nearly all self-assembling protein architectures. Reported approaches to designing selfassembling proteins have satisfied this requirement in different ways, such as the use of engineered disulfide bonds,1,2 electrostatic interactions,3 chemical cross-links,4 metal-mediated interactions,5 ligand-induced association,6 computation interface design,7,8 or genetic fusion of multiple protein domains or fragments.9,10 The ability to control SSIs can lead to the construction of the nanoscale protein materials customtailored to specific applications. However, the task of rendering SSIs controllable is complicated by the fact that SSIs are mediated by weak, noncovalent interactions over large surfaces. Recently, there has been growing evidence that established shell-like protein cages are a common architectural paradigm of the subcellular world. For example, clathrin cages are involved in endocytosis,11 carboxysomes in CO2 fixation,12 and viral capsids in nucleic acid storage and transport.13 In all of these cases, the cages shield their cargo from the influence of external © 2016 American Chemical Society

Received: September 15, 2016 Accepted: November 4, 2016 Published: November 9, 2016 10382

DOI: 10.1021/acsnano.6b06235 ACS Nano 2016, 10, 10382−10388

Article

www.acsnano.org

Article

ACS Nano

the key subunit interface, which triggers reassembly of the protein architecture into a non-native assembly. Specifically, we applied KSIR to control the self-assembly of recombinant human H-chain ferritin (rHuHF) because (1) it is a homopolymer which can simplify the design challenge; (2) its sequence and structural information are available to guide genetic manipulations; and (3) it has become a vehicle for tumor imaging and drug delivery.15−18 By using the KSIR strategy, introduction of small (hexapeptide) deletion into helix D of each subunit located at the C3−C4 interface results in conversion of a native 24-mer ferritin cage into its 48-mer nonnative analogue having a 17 nm diameter (Figure 1). Notably, the crystal structure reveals that equal numbers of two different subunits that originate from the same polypeptide coassemble into this 48-mer ferritin-like cage.

conserved architecture that comprises 24 subunits assembling into a spherical shell with 432 symmetry (Figure 1a,c). Each

RESULTS AND DISCUSSION Redesign of Key Subunit Interfaces of Ferritin. There are four interfaces responsible for ferritin assembly, which correspond to C2, C3, C4, and C3−C4. The genetic modification in the C2 interface facilitates ferritin shell disassociation into a single subunit.31 In contrast, the genetic modification with both C 3 and C 4 interfaces hardly alters protein shell-like structure.32−34 Similar results were obtained by amino acid residue deletion at the N-terminal and C-terminal of ferritin.35−37 However, so far, there has been no report focusing on the C3−C4 interface of ferritin. By comparison, we found that the C3−C4 interface (Figure S1b,c) has the largest total surface area among the above-mentioned four interfaces in the ferritin shell, suggesting that it plays an important role in controlling protein assembly. Therefore, we chose the C3−C4 interface for redesign by amino acid deletion (Figure 1a,b). By scanning the C3−C4 interface, we identified six amino acid residues 139NEQVKA144 located at the D-helix that do not participate in SSIs. To elucidate the contribution of these specific residues to ferritin assembly, we genetically modified the ferritin subunit by individually introducing four deletion mutations into a critical region of the D-helix containing amino acid residues 139NEQVKAIK146 (Figures S2 and S3a). Four mutants Δ139NE, Δ139NEQV, Δ139NEQVKA, and Δ139NEQVKAIK, correspond to deletion the dipeptide Asn-Glu, the tetrapeptide Asn-Glu-Gln-Val, the hexapeptide Asn-Glu-Gln-Val-Lys-Ala, and the octapeptide Asn-Glu-Gln-Val-Lys-Ala-Ile-Lys behind L138, respectively. Δ139NE and Δ139NEQV mutants exhibited electrophoretic behavior similar to that of native ferritin (Figure S3b). In contrast, the electrophoretic behavior of Δ139NEQVKA is markedly distinct from that of native ferritin, indicative of the formation of a non-native protein. As expected, deletion of octapeptide (Δ139NEQVKAIK) destroyed the protein shelllike structure into subunit dimers (Figure S3b) because Lys146 is involved in SSIs. Purification and Characterization of Δ139NEQVKA. To gain insight into the characteristics of Δ139NEQVKA, we expressed it in Escherichia coli BL21 cells and purified it by a combination of gel and ion-exchange chromatography. Native PAGE of this mutant exhibited a single band (Figure 2a), indicating that it was purified to homogeneity. As expected, SDS-PAGE analyses revealed that this protein is composed of one type of subunit, the molecular weight (MW) of which is markedly lower than that of the native ferritin subunit (Figure 2b). This result was confirmed by MALDI-TOF-MS analyses. The MW of the deletion mutant is determined as 20557.3 ±

Figure 1. Schematic representation of conversion of native human ferritin into its non-native analogue by deleting “silent” amino acid residues. (a) Ferritin shell consists of 12 subunit dimers with the approximate geometry of a rhombic dodecahedron symmetry, producing 24 subunit interfaces highlighted with the black line. (b) Helix D of each subunit is directly involved in subunit−subunit interactions at the interfaces, which is highlighted as a cylindrical helix model. (c) Crystal structure of a tetraeicosameric recombinant human H-chain ferritin (rHuHF) cage with external diameter of about 12 nm (PDB ID: 2FHA). (d) Structure of native ferritin subunit. The “silent” amino acid residues, N139−A144 (no participation in interfacial interactions) are highlighted in cyan, which are located at the D-helix. (e) Deletion of the six “silent” amino acid residues triggers the formation of a non-native 48-mer protein cage (about 17 nm) which consists of two non-native subunits at a ratio of 1:1.

subunit consists of a four-α-helix bundle containing two antiparallel helix pairs (A, B and C, D) and a fifth short Ehelix (Figure S1a). The ferritin shell has the approximate geometry of a rhombic dodecahedron, consisting of 12 subunit dimers,28 as shown in Figure 1a and Figure S1b. Dps effectively protects DNA against oxidative agents both in vitro and in vivo. The subunit of Dps is similar in structure to that of typical ferritin and is an elaboration of a four-helix bundle motif (Figure S1d). However, unlike classical ferritin, Dps proteins assemble in a quasi-spherical dodecamer with 23 symmetry (Figure S1e,f).27 Compared to other protein cages, the ferritin nanocage has one major advantage: high selectivity for cancer cells which overexpress two kinds of receptorsthe scavenger receptor class A member 5 (SCARA5) for L-ferritin29 and transferrin receptor 1 (TfR1) for H-ferritin.30 Therefore, it has emerged as a class of drug delivery vehicles and imaging agents.15−21 However, to date, the ferritin assembly has been limited in scope to a single size and shape. Here, we describe an engineering strategy termed key subunit interface redesign (KSIR) that could be used for fabrication of non-native multisubunit protein architectures. The first step of KSIR calls for determination of key subunit interfaces of a target symmetric protein architecture by using crystal structural information as a guide. The second step is to identify “silent” amino acid residues (SAAR, no participation in interfacial interactions) located at the key subunit interfaces. The last step corresponds to the deletion of SAAR to redesign 10383

DOI: 10.1021/acsnano.6b06235 ACS Nano 2016, 10, 10382−10388

Article

ACS Nano

Figure 2. Characterization of Δ139NEQVKA. (a) Native PAGE and (b) SDS-PAGE analyses of Δ139NEQVKA. Lane 1, rHuHF; lane 2, Δ139NEQVKA; and lane M, protein markers and their corresponding molecular masses. (c) Mass spectra of the rHuHF subunit (black line) and Δ139NEQVKA subunit (blue line). (d) Sedimentation coefficient distribution c(s) versus s20,w for rHuHF (black line) and Δ139NEQVKA (blue line). Conditions: [rHuHF] = [Δ139NEQVKA] = 0.6 mg/mL, 50 mM Tris, pH 8.0, 4 °C. (e) Typical equilibrium distribution of Δ139NEQVKA (0.3 mg/mL protein at 9600 rpm and 4 °C). Inset is random distribution of residuals as a function of radial distance. (f) Comparison of the measured molecular weight of Δ139NEQVKA based on sedimentation equilibrium experiments performed at different protein concentrations and centrifugation speeds. All experiments were repeated three times with different protein preparations.

with a hole at the center, and its exterior diameter is approximately 10.0 nm (Figure S5b). Similar results were obtained with protein samples from three different protein preparations/purifications. Thus, the above deletion at the C3− C4 interface of native ferritin resulted in a significant decrease in protein size. Crystal Structure of Δ139NEQVKA and Its Possible Formation Pathway. To obtain structural information on Δ139NEQVKA, we solved its X-ray crystal structure at 2.81 Å resolution (F432 space group) by molecular replacement (Figure 1e and Table S1). Upon introduction of a small (hexapeptide) deletion into the native ferritin subunit region defined by residues 139−144, we would expect that two different subunits are generated to fill the vacancy in the Dhelix. One part close to the C-terminal moves backward to meet another located at the N-terminal (Figure 3b), producing a subunit termed Hα (Figure 3a) in which the native E-helix structure was lost. In contrast, the part near the N-terminal of the native D-helix moves forward to the C-terminal to connect with another one (Figure 3b), forming the second non-native subunit termed Hβ (Figure 3c), in which the initial E-helix is kept at a cost of loss of the CD turn. Indeed, we noted that this non-native protein in the crystal is a heteropolymer composed of 48 subunits of two types at a ratio of 1:1 that assemble into a spherical shell. Interestingly, these two subunits are derived from an identical polypeptide chain (Figure S6a) folding into a four-helix bundle which closely resembles that of native rHuHF (Figure 3d,e) and other ferritin subunits.28 The 48 subunit composition in the crystals causes this protein to exhibit a size (∼17 nm) larger than that of this mutant in solution (∼10 nm) (Figures 1e and S5b). The striking difference in size is derived from the possibility that the 48-mer is unstable in solution relative to its 8-mer analogue, while it is much more stable in the crystal due to its much larger surface area. Consistent with

0.5 Da, a value in good agreement with the theoretical value of single subunit (20555.9 Da) (Figure 2c). The difference in MW between the native ferritin subunit (21223.3 ± 0.5 Da) and this mutant subunit is 666.0, confirming the deletion of residues 139−144. Analytical ultracentrifugation showed that Δ139NEQVKA sedimented as a single discrete species with s20,w = 7.4 ± 0.1 S (Figure 2d). Thus, the deletion of the 139 NEQVKA144 sequence in the ferritin subunit largely decreased the average oligomer mass as compared to 18.8 ± 0.1 S (24 subunits/oligomer) for the native protein. Interestingly, approximately a 3% decrease in subunit mass leads to a significantly larger decrease in oligomeric mass. This might be due to the reduction in the number of subunits that assemble to form an oligomer. To corroborate this interpretation, we used equilibrium sedimentation analyses to directly measure the MW of this non-native protein in solution. The equilibrium sedimentation data fit well to a single-ideal species model that yielded random residuals (Figure 2e). By using this method, we determined the MW of this non-native protein to be 167.3 ± 2.0 kDa, which was nearly constant over a wide concentration range (0.15−0.45 mg/mL) and different centrifugation speeds (6000−12000 rpm), as shown in Figure 2f. Based on a ratio of the MW of this non-native protein to that of its subunit (167.3 kDa/20557.3 Da ≈ 8), we can draw a conclusion that this non-native protein consists of eight subunits. This result was further approved by MS analyses showing that the MW of Δ139NEQVKA in a native form is 163.6 ± 1.0 kDa (Figure S4). Subsequently, we used transmission electron microscopy (TEM) to visualize the morphology of this non-native protein with native ferritin as the control. Native ferritin appeared as discrete protein cages with an exterior diameter of around 12 nm (Figure S5a), consistent with previous observation.14−17 In contrast, we noted that the deletion mutant likewise exhibits a roughly disc-like structure 10384

DOI: 10.1021/acsnano.6b06235 ACS Nano 2016, 10, 10382−10388

Article

ACS Nano

To confirm this idea, we observed the stability of the crystal in three different buffers (acetate pH 4.0, Tris-HCl pH 8.0, and glycine pH 10, 50 mM) by TEM as a function of time. We found that only 30−50% of the crystals can be dissolved in the buffers within 5 h at 37 °C, suggesting that the crystal is exceptionally stable. This observation is perhaps not surprising because of the large outer surface area (803.8 nm2) of the 48mer nanocage, which would considerably enhance intermolecular interactions in the crystal. After being dissolved in the buffers, the 48-mer supramolecules occurred as discrete protein nanocages with a diameter of ∼17 nm (Figure 4a), a size identical to that in the crystal. However, at 2 h, smaller protein species appeared (Figure 4b). After 10 h, all 17 nm species completely converted into these smaller ones (Figure 4e) having the same size (∼10 nm) and MW as those of the 8-mer protein originally produced in E. coli (Figures 4f and S6b). Taken together, these findings demonstrate that the 48-mer stems from its 8-mer analogue, and that the 8-mer and 48-mer can interconvert with each other depending on experimental conditions. Further support for this conclusion comes from comparison between protein crystal structure and 2D average results. According to the crystal structure of the 48-mer nanocage, each 8-mer moiety looks like a concave disc with a hole in the center upon viewing down the four-fold channel, while its shape resembles a bow from the side view (Figure S8a). If the above purified 8-mer was used as building blocks for the 48-mer assembly, one would expect that it has a shape like a concave disc or bow. Indeed, two-dimensional class averages by TEM analyses revealed that, from either the top view or the side view, the shape and size of the purified 8-mer is virtually identical to those of the 8-mer moiety in the crystal structure (Figure S8c). Effect of Different Experimental Conditions on the Stability of the 8-mer Protein Species in Solution. In Tris-HCl buffer (pH 8.0), we can only observe the 8-mer protein species but not its analogue 48-mer by different

Figure 3. Ribbon diagram and structure alignment of the subunit structure of native ferritin rHuHF and its non-native analogue Δ139NEQVKA. (a) Hα structure of Δ139NEQVKA. Some amino acid residues at the C-terminal cannot be observed due to a disorder of the partial D-helix. (b) Proposed conversion mechanism of the ferritin subunit into its two non-native analogues (Hα and Hβ) upon deletion of six “silent” amino acid residues (yellow) at the D-helix. Such deletion results in a shift of two ends of helix D in two different directions; namely, the helix part near the C-terminal (purple) moves backward to the N-terminal (cyan), producing a subunit termed Hα, while another part located at the N-terminal moves forward to the C-terminal, generating a subunit termed Hβ. (c) Hβ structure of Δ139NEQVKA. (d) Superposition of the rHuHF subunit (gray) and Hα (blue) yields a root-meansquare deviation (rmsd) of the Cα positions of 0.477 Å, indicating that the monomeric folds are very similar. (e) Value of rmsd between Cα positions in the cores of the aligned bundles of rHuHF subunit and Hβ (red) is 0.436 Å.

this view, the crystal structure of the 48-mer nanocage reveals that, besides the salt bridges and hydrophobic interactions at the C2 interfaces (Figure S7a,b), electrostatic repulsive forces also occur at the C2 interfaces, which involve E107 of one Hα subunit and E141 of another Hα (Figure S7c). On one hand, these electrostatic repulsive interactions are required for the formation of the 48-mer nanocages. On the other hand, such interactions destabilize the 48-mer proteins in solution, resulting in its disassembly along the C2 interfaces into the 8mer protein species.

Figure 4. Kinetics of the conversion of Δ139NEQVKA (48-mer) to Δ139NEQVKA (8-mer) revealed by TEM. (a−e) TEM images of crystal in liquid at 0.5, 2, 5, 8, and 10 h, respectively. (f) TEM image of the 8-mer deletion mutant originally produced in E. coli. Conditions: 50 mM Tris, pH 8.0, 25 °C. Scale bars represent 100 nm. The insets show the magnification, in which scale bars represent 20 nm. All experiments were repeated three times with different protein preparations. 10385

DOI: 10.1021/acsnano.6b06235 ACS Nano 2016, 10, 10382−10388

Article

ACS Nano

Figure 5. Native ferritin rHuHF and its non-native analogue Δ139NEQVKA subunit symmetry relations. View of (a) rHuHF and (b) Δ139NEQVKA molecules down the two-fold symmetry axis. View of (c) rHuHF and (d) Δ139NEQVKA molecules down the three-fold symmetry axis. View of (e) rHuHF and (f) Δ139NEQVKA molecules down the four-fold symmetry axis.

by helices C and D from three adjacent subunit monomers, in such a way that, with respect to the hollow tetraeicosameric ferritin shell, the C-terminal ends of helices C define the outer entrance to the channel and the N-terminal of helices D define the inner entrance with pore sizes between 3 and 5 Å (Figure 5c). Differently, the three-fold channel of the non-native protein involves three pairs of Hα (AB turn, BC loop, and Chelix) and Hβ (the N-terminal, B-helix, and BC loop) subunits contributed from three adjacent tetramers, producing large pores with a size of ∼25 Å (Figure 5d). This structural feature is reminiscent of the ferritin having tetrahedral symmetry from the hyperthermophilic Archaeon Archaeoglobus fulgidus, which opens four large (∼45 Å) pores in the protein shell; larger molecules such as FeS2 and iron complexes were thought to potentially diffuse through these massive pores.38 In contrast, the four-fold interactions shared in common the non-native and native ferritins, resulting in the formation of six channels where four E-helices lie roughly parallel (Figure 5e,f).

methods, as shown in Figure S5b, suggesting that the 8-mer non-native protein is more stable than the 48-mer in solution. To confirm this idea, we likewise visualized the morphology of the protein species in different buffers such as MOPS, HEPES, and others at different pH values in the presence or absence of NaCl (50−300 mM), and we found that only protein species with ∼10 nm can be observed (Table S2). Similar results were obtained with protein samples from three different protein preparations/purifications. Thus, the 8-mer protein species appear to be very stable under different experimental conditions, and its existence in aqueous solution is not sample/buffer/pH/salt concentration-dependent. Thus, it appears that the 8-mer is more stable in solution, while its 48-mer analogue is more stable in the crystal. Comparison between the Crystal Structures of Δ139NEQVKA and Native Ferritin. Despite the same octahedral symmetry displayed by the shell of rHuHF and its 48-mer analogue, their quaternary structures differ strikingly from each other. The geometry of the native protein shell is approximately that of a rhombic dodecahedron, where each face consists of a subunit homodimer in antiparallel position (Figure 5a). Such a homodimer, however, is replaced by a tetramer in the 48-mer, which comprises two identical heterodimers (HαHβ) related by a two-fold axis (Figure 2b). The interfaces related to the two-fold interactions in native ferritin involve the helices A and B, as well as the BC loops, the N-terminus, and the AB turn. As expected, these interfaces were conserved well in this non-native 48-mer protein cage because we did not genetically modify the helices A and B which are responsible for the two-fold interactions in native ferritin. However, different C2 interfaces in this 48-mer protein were formed, where the helices C and D and the CD turn of the Hα subunit from adjacent heterodimers are involved. This represents the first major difference in structure and assembly between these two proteins. The second structural difference between native ferritin and the 48-mer corresponds to the three-fold interactions. The eight hydrophilic channels along the three-fold axes are formed

CONCLUSION Our results established a protein engineering strategy, KSIR, by which a non-native protein nanocage with different geometries could be constructed. The construction strategy that focuses on “silent” amino acid residues at the key subunit interfaces is conceptually simple. Since the SAAR are widely distributed in other multisubunit protein architectures, such as Dps,39 heat shock protein,40 and the E2 protein,41 our general engineering approach, KSIR, should, in principle, be applicable to these proteins. This would lead to the generation of a variety of nonnative protein nanomaterials with different geometries which could impart unexplored properties and functions to these nanomaterials. METHODS Protein Preparation. cDNA encoding the full-length amino acid sequence of rHuHF was cloned into the pET-21d (Novagen) and verified by DNA sequencing.42 Mutagenesis of the rHuHF cDNA was performed with the fast site-directed mutagenesis kit (TIANGEN 10386

DOI: 10.1021/acsnano.6b06235 ACS Nano 2016, 10, 10382−10388

Article

ACS Nano Biotech Co., Ltd.). PCR amplification was carried out using the pET21d plasmid with the rHuHF gene as a template. PCR amplification was as follows: denaturation at 95 °C for 3 min, followed by 18 cycles of 94 °C for 30 s, 55 °C for 30 s, and 68 °C for 3 min, and a final extension cycle of 68 °C for 10 min. After the PCR reaction, the parental DNA template was digested with DMT enzyme. The PCRamplified plasmid was separated on agarose gel, extracted, and inserted into E. coli competent cells. Ampicillin-resistant colonies were selected from which the plasmids were extracted. The extracted plasmid was sequenced for confirmation of the mutation. Native ferritin rHuHF was purified as previously described.42 The non-native ferritin-like protein Δ139NEQVKA was purified as follows. The E. coli strain BL21 (DE3) which contained Δ139NEQVKA expression plasmid was grown at 37 °C. Protein expression was likewise induced with 200 μM of isopropyl β-D-1-thiogalactopyranoside. After the cell density reached an absorbance of 0.6 at 600 nm, the cells were harvested by centrifugation after 5 h of induction and resuspended in 50 mM Tris-HCl (pH 8.0) to a concentration of 40 g fresh weight bacteria per liter, followed by disruption by sonication. The supernatant of the resulting crude extract was collected by centrifugation and fractionated by 30−50% saturation of ammonium sulfate. The pellet was resuspended in 50 mM Tris-HCl (pH 8.0) and dialyzed against the same buffer. The protein solution was applied to an ion-exchange column (Q-Sepharose Fast Flow, GE Healthcare), followed by gradient elution with 0−0.5 M NaCl. Finally, the protein solution was concentrated and purified on a gel filtration column (Superdex 200 pg 16/60, GE Healthcare), equilibrated with 50 mM Tris-HCl and 150 mM NaCl (pH 8.0). Protein concentrations were determined according to the Lowry method with bovine serum albumin as standard. Crystallization, Data Collection, and Structure Determination. Purified Δ139NEQVKA was concentrated to 10 mg/mL in a buffer consisting of 10 mM Tris-HCl at pH 8.0 and 150 mM sodium chloride. Crystals of Δ139NEQVKA were obtained using the hanging drop vapor diffusion method by mixing equal volumes of the protein sample and mother liquid, which was composed of 0.1 M imidazoleHCl at pH 7.5, 10% reagent alcohol, and 0.2 M MgCl2. Snowflake-like crystals appeared within 4 days at 20 °C. Diffraction data of the crystal were collected to resolutions of 2.81 Å at SSRF (BL18U) after flash cooling with 30% ethylene glycol as a cryo-protectant. Data were processed, merged, and scaled with the HKL-2000 (HKL Research). Data processing statistics are shown in Table S1. The structure of Δ139NEQVKA was determined by molecular replacement using coordinates of human H ferritin (PDB code 2FHA) as an initial model using the MOLREP program in the CCP4 program package. Structure refinement was conducted using the Refmac5 program and PHENIX software. The structure was rebuilt using COOT, which made the model manually adjusted. Figures of protein structures were prepared using the PyMOL program. The final model (1/12 of the asymmetric unit) contains residues 4−146 of the Hα chain and 10−171 of the Hβ chain. The missing C-terminal residues of the Hα chain and N-terminal residues of the Hβ chain appear to be disordered. The model geometry is consistent with the high quality and resolution of the diffraction data.

AUTHOR INFORMATION Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. ORCID

Guanghua Zhao: 0000-0001-8587-9680 Author Contributions §

S.Z. and J.Z. contributed equally.

Author Contributions

G.Z. supervised research. G.Z. and S.Z. conceived and designed experiments. S.Z., J.Z., X.Z., and H.C. prepared the samples and performed the experiments. J.Z. and B.M. performed and analyzed the X-ray diffraction experiments. All authors wrote the paper, discussed the results, and commented on the manuscript. Notes

The authors declare no competing financial interest.

ACKNOWLEDGMENTS This work was supported by the National Natural Science Foundation of China (Nos. 31271826 and 31471693) and the Chinese Universities Scientific Fund for Graduate Students of the China Agricultural University (2015sp004). REFERENCES (1) 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. U. S. A. 2008, 105, 3733−3738. (2) Fletcher, J. M.; Harniman, R. L.; Barnes, F. R.; Boyle, A. L.; Collins, A.; Mantell, J.; Sharp, T. H.; Antognozzi, M.; Booth, P. J.; Linden, N.; Miles, M. J.; et al. Self-Assembling Cages from Coiled-Coil Peptide Modules. Science 2013, 340, 595−599. (3) Kostiainen, M. A.; Hiekkataipale, P.; Laiho, A.; Lemieux, V.; Seitsonen, J.; Ruokolainen, J.; Ceci, P. Electrostatic Assembly of Binary Nanoparticle Superlattices Using Protein Cages. Nat. Nanotechnol. 2013, 8, 52−56. (4) Ringler, P.; Schulz, G. E. Self-Assembly of Proteins into Designed Networks. Science 2003, 302, 106−109. (5) Brodin, J. D.; Ambroggio, X. I.; Tang, C.; Parent, K. N.; Baker, T. S.; Tezcan, F. A. Metal-directed, Chemically Tunable Assembly of One-, Two- and Three-Dimensional Crystalline Protein Arrays. Nat. Chem. 2012, 4, 375−382. (6) McAllister, K. A.; Zou, H.; Cochran, F. V.; Bender, G. M.; Senes, A.; Fry, H. C.; Nanda, V.; Keenan, P. A.; Lear, J. D.; Saven, J. G.; Therien, M. J.; et al. Using α-Helical Coiled-Coils to Design Nanostructured Metalloporphyrin Arrays. J. Am. Chem. Soc. 2008, 130, 11921−11927. (7) King, N. P.; Bale, J. B.; Sheffler, W.; McNamara, D. E.; Gonen, S.; Gonen, T.; Yeates, T. O.; Baker, D. Accurate Design of Co-Assembling Multi-Component Protein Nanomaterials. Nature 2014, 510, 103− 108. (8) Bale, J. B.; Gonen, S.; Liu, Y.; Sheffler, W.; Ellis, D.; Thomas, C.; Cascio, D.; Yeates, T. O.; Gonen, T.; King, N. P.; Baker, D. Accurate Design of Megadalton-Scale Two-Component Icosahedral Protein Complexes. Science 2016, 353, 389−394. (9) Sinclair, J. C.; Davies, K. M.; Vénien-Bryan, C.; Noble, M. E. Generation of Protein Lattices by Fusing Proteins with Matching Rotational Symmetry. Nat. Nanotechnol. 2011, 6, 558−562. (10) Lai, Y. T.; Cascio, D.; Yeates, T. O. Structure of a 16-nm Cage Designed by Using Protein Oligomers. Science 2012, 336, 1129−1129. (11) Royle, S. J. The Cellular Functions of Clathrin. Cell. Mol. Life Sci. 2006, 63, 1823−1832.

ASSOCIATED CONTENT S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsnano.6b06235. Supplementary methods, additional materials, and characterizations (PDF) Accession Codes

Atomic coordinates and structure factors have been deposited into the Protein Data Bank with the identification number: 5GN8. 10387

DOI: 10.1021/acsnano.6b06235 ACS Nano 2016, 10, 10382−10388

Article

ACS Nano (12) Tanaka, S.; Kerfeld, C. A.; Sawaya, M. R.; Cai, F.; Heinhorst, S.; Cannon, G. C.; Yeates, T. O. Atomic-Level Models of the Bacterial Carboxysome Shell. Science 2008, 319, 1083−1086. (13) Canady, M. A.; Larson, S. B.; Day, J.; McPherson, A. Crystal Structure of Turnip Yellow Mosaic Virus. Nat. Struct. Biol. 1996, 3, 771−781. (14) Uchida, M.; Kang, S.; Reichhardt, C.; Harlen, K.; Douglas, T. The Ferritin Superfamily: Supramolecular Templates for Materials Synthesis. Biochim. Biophys. Acta, Gen. Subj. 2010, 1800, 834−845. (15) Liang, M.; Fan, K.; Zhou, M.; Duan, D.; Zheng, J.; Yang, D.; Feng, J.; Yan, X. H-Ferritin−Nanocaged Doxorubicin Nanoparticles Specifically Target and Kill Tumors with a Single-Dose Injection. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, 14900−14905. (16) Zhen, Z.; Tang, W.; Guo, C.; Chen, H.; Lin, X.; Liu, G.; Fei, B.; Chen, X.; Xu, B.; Xie, J. Ferritin Nanocages to Encapsulate and Deliver Photosensitizers for Efficient Photodynamic Therapy Against Cancer. ACS Nano 2013, 7, 6988−6996. (17) Huang, P.; Rong, P.; Jin, A.; Yan, X.; Zhang, M. G.; Lin, J.; Hu, H.; Wang, Z.; Yue, X.; Li, W.; Niu, G.; et al. Dye-Loaded Ferritin Nanocages for Multimodal Imaging and Photothermal Therapy. Adv. Mater. 2014, 26, 6401−6408. (18) Fan, K.; Cao, C.; Pan, Y.; Lu, D.; Yang, D.; Feng, J.; Song, L.; Liang, M.; Yan, X. Magnetoferritin Nanoparticles for Targeting and Visualizing Tumour Tissues. Nat. Nanotechnol. 2012, 7, 459−464. (19) Conti, L.; Lanzardo, S.; Ruiu, R.; Cadenazzi, M.; Cavallo, F.; Aime, S.; Crich, S. G. L-Ferritin Targets Breast Cancer Stem Cells and Delivers Therapeutic and Imaging Agents. Oncotarget 2016, 7, 66713− 66727. (20) Cutrin, J. C.; Crich, S. G.; Burghelea, D.; Dastrù, W.; Aime, S. Curcumin/Gd Loaded Apoferritin: A Novel “Theranostic” Agent to Prevent Hepatocellular Damage in Toxic Induced Acute Hepatitis. Mol. Pharmaceutics 2013, 10, 2079−2085. (21) Crich, S. G.; Cadenazzi, M.; Lanzardo, S.; Conti, L.; Ruiu, R.; Alberti, D.; Cavallo, F.; Cutrin, J. C.; Aime, S. Targeting Ferritin Receptors for the Selective Delivery of Imaging and Therapeutic Agents to Breast Cancer Cells. Nanoscale 2015, 7, 6527−6533. (22) Douglas, T.; Young, M. Host−Guest Encapsulation of Materials by Assembled Virus Protein Cages. Nature 1998, 393, 152−155. (23) Wörsdörfer, B.; Woycechowsky, K. J.; Hilvert, D. Directed Evolution of a Protein Container. Science 2011, 331, 589−592. (24) Maity, B.; Fujita, K.; Ueno, T. Use of the Confined Spaces of Apo-Ferritin and Virus Capsids As Nanoreactors for Catalytic Reactions. Curr. Opin. Chem. Biol. 2015, 25, 88−97. (25) Dickerson, M. B.; Sandhage, K. H.; Naik, R. R. Protein- and Peptide-Directed Syntheses of Inorganic Materials. Chem. Rev. 2008, 108, 4935−4978. (26) Honarmand Ebrahimi, K.; Hagedoorn, P. L.; Hagen, W. R. Unity in the Biochemistry of the Iron-Storage Proteins Ferritin and Bacterioferritin. Chem. Rev. 2015, 115, 295−326. (27) Zhang, Y.; Orner, B. P. Self-Assembly in the Ferritin Nano-Cage Protein Superfamily. Int. J. Mol. Sci. 2011, 12, 5406−5421. (28) Crichton, R. R.; Declercq, J. P. X-Ray Structures of Ferritins and Related Proteins. Biochim. Biophys. Acta, Gen. Subj. 2010, 1800, 706− 718. (29) Li, J. Y.; Paragas, N.; Ned, R. M.; Qiu, A.; Viltard, M.; Leete, T.; Drexler, I. R.; Chen, X.; Sanna-Cherchi, S.; Mohammed, F.; Williams, D.; et al. Scara5 Is a Ferritin Receptor Mediating Non-Transferrin Iron Delivery. Dev. Cell 2009, 16, 35−46. (30) Li, L.; Fang, C. J.; Ryan, J. C.; Niemi, E. C.; Lebrón, J. A.; Björkman, P. J.; Arase, H.; Torti, F. M.; Torti, S. V.; Nakamura, M. C.; Seaman, W. E. Binding and Uptake of H-Ferritin Are Mediated by Human Transferrin Receptor-1. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 3505−3510. (31) Huard, D. J.; Kane, K. M.; Tezcan, F. A. Re-Engineering Protein Interfaces Yields Copper-Inducible Ferritin Cage Assembly. Nat. Chem. Biol. 2013, 9, 169−176. (32) Chen, H.; Zhang, S.; Xu, C.; Zhao, G. Engineering Protein Interfaces Yields Ferritin Disassembly and Reassembly under Benign Experimental Conditions. Chem. Commun. 2016, 52, 7402−7405.

(33) Ardejani, M. S.; Chok, X. L.; Foo, C. J.; Orner, B. P. Complete Shift of Ferritin Oligomerization Toward Nanocage Assembly via Engineered Protein−Protein Interactions. Chem. Commun. 2013, 49, 3528−3530. (34) Lv, C.; Zhang, S.; Zang, J.; Zhao, G.; Xu, C. Four-Fold Channels Are Involved in Iron Diffusion into the Inner Cavity of Plant Ferritin. Biochemistry 2014, 53, 2232−2241. (35) Luzzago, A.; Cesareni, G. Isolation of Point Mutations that Affect the Folding of the H Chain of Human Ferritin in E. coli. EMBO J. 1989, 8, 569−576. (36) Levi, S.; Luzzago, A.; Franceschinelli, F.; Santambrogio, P.; Cesareni, G.; Arosio, P. Mutational Analysis of the Channel and Loop Sequences of Human Ferritin H-Chain. Biochem. J. 1989, 264, 381− 388. (37) Levi, S.; Luzzago, A.; Cesareni, G.; Cozzi, A.; Franceschinelli, F.; Albertini, A.; Arosio, P. Mechanism of Ferritin Iron Uptake: Activity of the H-Chain and Deletion Mapping of the Ferro-Oxidase Site. A Study of Iron Uptake and Ferro-Oxidase Activity of Human Liver, Recombinant H-Chain Ferritins, and of Two H-Chain Deletion Mutants. J. Biol. Chem. 1988, 263, 18086−18092. (38) Johnson, E.; Cascio, D.; Sawaya, M. R.; Gingery, M.; Schröder, I. Crystal Structures of a Tetrahedral Open Pore Ferritin From the Hyperthermophilic Archaeon Archaeoglobus f ulgidus. Structure 2005, 13, 637−648. (39) Roy, S.; Gupta, S.; Das, S.; Sekar, K.; Chatterji, D.; Vijayan, M. X-Ray Analysis of Mycobacterium Smegmatis Dps and a Comparative Study Involving Other Dps and Dps-like Molecules. J. Mol. Biol. 2004, 339, 1103−1113. (40) Kim, K. K.; Kim, R.; Kim, S. H. Crystal Structure of a Small Heat-Shock Protein. Nature 1998, 394, 595−599. (41) Knapp, J. E.; Mitchell, D. T.; Yazdi, M. A.; Ernst, S. R.; Reed, L. J.; Hackert, M. L. Crystal Structure of the Truncated Cubic Core Component of the Escherichia Coli 2-oxoglutarate Dehydrogenase Multienzyme Complex. J. Mol. Biol. 1998, 280, 655−668. (42) Masuda, T.; Goto, F.; Yoshihara, T.; Mikami, B. The Universal Mechanism for Iron Translocation to the Ferroxidase Site in Ferritin, Which is Mediated by the Well Conserved Transit Site. Biochem. Biophys. Res. Commun. 2010, 400, 94−99.

10388

DOI: 10.1021/acsnano.6b06235 ACS Nano 2016, 10, 10382−10388