Encapsulation of Biomolecules in Bacteriophage MS2 Viral Capsids

Home , Bacteriophage MS2, Circovirus

Jeff Edward Glasgow

A dissertation submitted in partial satisfaction of the

requirements for the degree of

Doctor of Philosophy

Chemistry

in the

Graduate Division

of the

University of California, Berkeley

Committee in Charge:

Professor Matthew Francis, Co-Chair Professor Danielle Tullman-Ercek, Co-Chair Professor Michelle Chang Professor John Dueber

Spring 2014 Encapsulation of Biomolecules in Bacteriophage MS2 Viral Capsids

By: Jeff Edward Glasgow Abstract

Encapsulation of Biomolecules in Bacteriophage MS2 Viral Capsids

Jeff Edward Glasgow

Doctor of Philosophy in Chemistry

University of California, Berkeley

Professor Matthew Francis, Co-Chair

Professor Danielle Tullman-Ercek, Co-Chair

Nanometer-scale molecular assemblies have numerous applications in materials, catalysis, and medicine. Self-assembly has been used to create many structures, but approaches can match the extraordinary combination of stability, homogeneity, and chemical flexibility found in viral capsids. In particular, the bacteriophage MS2 capsid has provided a porous scaffold for several engineered nanomaterials for drug delivery, targeted cellular imaging, and photodynamic therapy by chemical modification of the inner and outer surfaces of the shell. This work describes the development of new methods for reassembly of the capsid with concomitant encapsulation of large biomolecules. These methods were then used to encapsulate a variety of interesting cargoes, including RNA, DNA, protein-nucleic acid and protein-polymer conjugates, metal nanoparticles, and enzymes. To develop a stable, scalable method for encapsulation of biomolecules, the assembly of the capsid from its constituent subunits was analyzed in detail. It was found that combinations of negatively charged biomolecules and protein stabilizing agents could enhance reassembly, while electrostatic interactions of the biomolecules with the positively charged inner surface led to encapsulation. To investigate the role protein shells play in encapsulated enzymatic processes, this method was then used to a model enzyme in a series of capsids with altered characteristics around the pores. The method was then used to develop a potential system for delivery of therapeutic proteins to the cytoplasm of cancer cells by encapsulating conjugates of these proteins to a negatively charged, membrane-lysing polymer.

1 Dedicated to my parents, Joe and Heidi Glasgow.

i Table of Contents

Abstract...... 1

Acknowledgments...... iv

1. Production and Applications of Engineered Viral Capsids...... 1

1.1 Abstract...... 1 1.2 Introduction...... 2 1.3 Production: Recombinant vs. Infection...... 2 1.4 Production: Capsid Size and Assembly...... 3 1.5 Production: Surface Chemistry...... 4 1.6 Applications: Drug Delivery...... 4 1.7 Applications: Imaging Agents...... 6 1.8 Applications: Templated Synthesis...... 7 1.9 Applications: Nanoreactors and Scaffolds...... 9 1.10 Nonviral Protein Scaffolds...... 10 1.11 Conclusions and Perspectives...... 11 1.12 References...... 11

2. Osmolyte-Mediated Encapsulation of Proteins inside MS2 Viral Capsids...... 21

2.1 Abstract...... 21 2.2 Introduction...... 22 2.3 Reassembly of Wild-type MS2...... 23 2.4 Encapsulation of GFP...... 25 2.5 Encapsulation and Activity of Alkaline Phosphatase...... 27 2.6 Summary...... 29 2.7 Materials and Methods...... 30 2.8 References...... 34

ii 3. The Influence of Electrostatics on Small Molecule Flux through a Protein Nano- reactor...... 38

3.1 Abstract...... 38 3.2 Introduction...... 39 3.3 PhoA Kinetics in MS2 Pore Mutants...... 41 3.4 Modeling of Encapsulated Reactions...... 43 3.5 Conclusion...... 45 3.6 Materials and Methods...... 46 3.7 References...... 52

4. Toward Therapeutic Protein Delivery with MS2-Encapsulated Protein-polymer Conjugates...... 55

4.1 Abstract...... 55 4.2 Targeted Delivery of Therapeutic Proteins ...... 56 4.3 Polymer Conjugates in Cytosolic Protein Delivery...... 57 4.4 Conjugation of Negatively Charged Polymer Amphiphiles...... 57 4.5 Encapsulation of Protein-polymer Conjugates...... 61 4.6 Therapeutic Protein Encapsulation...... 63 4.7 Materials and Methods...... 66 4.8 References...... 73

iii Acknowledgments

Berkeley has been such a great place for me to grow into a scientist, with so many people eager to lend help or support in developing this dissertation. As such there are innumerable people I should thank. Here are a few:

Matt and Danielle—thank you for agreeing to take me on and start a new collaboration. You’ve both been so helpful with your ideas, enthusiasm, discussion, and support.

The DTE lab—thank you for giving me such a fun, stimulating place to work. I have really en- joyed watching how the lab’s personality has grown and shifted since the first year. You have all made a huge impact on me as a scientist. Thanks to Chris for all the math help. Thanks to Anum, who has always been a great partner.

The Francis lab—thank you for also taking me in and teaching me a little chemistry. The diversity of the both the research ideas and the people that have come through the lab has helped me see things from many different perspectives. Thanks to Stacy for lots of help and puppy pictures and Adel, Allie, Ioana, and Farkas for all their contributions.

Other labs—there have been several other groups that have helped me over the years. Thanks to the M. Chang lab for basically being my third home in the department. Thanks to the Fréchet, Ber- tozzi, and Sarpong labs for being so generous with useful discussion and equipment.

My Berkeley friends—both in and out of lab, I was lucky to have a great group of friends. I couldn’t have done it without you all!

My Oklahoma friends—thanks for still being there, even from over a thousand miles away.

My family—mom, dad, Susan, Christina, and Jeremy, thank you for all the support. The apprecia- tion for knowledge you all instilled in me has been a major driving force for everything I’ve done.

iv Chapter 1

Production and Applications of Engineered Viral Capsids

1.1 Abstract As biological agents, viruses come in an astounding range of sizes, with varied shapes and surface morphologies. The structures of viral capsids are generally assemblies of hundreds of copies of one or a few proteins which can be harnessed for use in a wide variety of applications in biotechnology, nanotechnology, and medicine. Despite their complexity, many capsid types form as homogenous populations of precise geometrical assemblies. This is important in both medicine, where well-defined therapeutics are critical for drug performance and federal approval, and nanotechnology, where precise placement affects the properties of the desired material. Here we review the production of viruses and virus-like particles with methods for selecting and manipulating the size, surface chemistry, assembly state, and interior cargo of capsid. We then discuss many of the applications used in research today and the potential commercial products from engineered viral capsids. Portions of this chapter appear in a separate publication.

1 1.2 Introduction Viruses, initially described at the end of the 19th century, are infectious agents found throughout the biosphere. Though they are highly diverse, viruses all consist of a nucleic acid genome contained inside a protein coat. In some viruses, this capsid is further confined in a lipid en- velope generally derived from the host-cell membrane. Viruses have enormous impact on human health and the environment; as a result, every aspect of their life cycle, composition, and assembly has been studied for over a century. This vast amount of literature has provided the foundation for engineering applications of viruses in a much newer field—bionanotechnology. The shapes and sizes of viral capsids are often highly regular, with weak protein-protein and protein-nucleic acid interactions orchestrating hundreds of subunits into precise locations around the capsid.1 Most well-studied viral capsids, particularly those with applications in bionanotechnology, form spherical shells with icosahedral symmetry or rod shaped structures with helical symmetry. Klug and coworkers developed much of the theory of assembly of both types of viruses.2 Icosahedral viruses typically follow specific geometric quasi-equivalence rules, recently outlined elsewhere.3 The high level of structural control obtained in these systems enables many scientists to create new molecular devices, scaffolds for materials, and containers for diverse cargoes. Many capsids can form structures, called virus-like particles (VLPs), lacking components integral for full infection. Several applications of viral capsids in biotechnology are beyond the scope of this review, including native viral vaccines,4 gene delivery,5 and epitope display methods.6

1.3 Production: Recombinant vs. Infection Expression of viral proteins can be accomplished by either infecting a native host and collecting fully assembled virions or by expressing structural proteins recombinantly. The appropriate method and resulting structure are highly case specific; the downstream applications, safety precautions, and the nature of the capsid itself can determine the appropriate production method. Many viral capsids, particularly those with more simple structures, benefit from the convenience and genetic flexibility of recombinant expression. For example,Escherichia coli-expressed recombinant hepatitis B virus (HBV), cowpea chlorotic mottle virus (CCMV), and polyomavirus capsids have been studied and manipulated extensively despite being of eukaryotic origin.3,7,8 Other viruses were engineered to be simpler for recombinant VLP production, such as rotavirus VP6. This capsid normally consists of 6 different structural proteins, but recently one major capsid protein, VP6, was expressed in inclusion bodies in E. coli and refolded and assembled in vitro.9 Recombi- nant expression has also allowed exploration of even more diverse capsids with novel properties, such as artificial amino acids and altered architectures. Expression of capsid proteins in vitro is also a viable approach for some systems; however, many viruses require specific queues in their life cycle to form properly, and are at best difficult to express heterologously.10 Infection of host organisms or cultured host cells is an effective way of producing large amounts of natively assembled capsid. This technique has been successfully applied at large scale for several biotechnologically relevant capsids, such as tobacco mosaic virus (TMV)11 and cowpea mosaic virus (CPMV).12 This is often the best choice for the expression of uncharacterized viruses with new properties, such as thermal stability.13,14 While native expression assures production of uniform particles, there are several disadvantages. Often for VLP applications, the genome must be removed by either chemical degradation or disassembly and separation. Host organisms may

2 Figure 1. Applications of Viral Capsids. Rod-shaped or spherical viral capsids can be used in many applications in medicine, materials, and catalysis. Genetic engineering creates many new possible structures and means of modification. require specialized equipment, culture techniques, or genetic tools. Finally, propagation is often deleterious to engineered mutations as a result of genetic instability.15

1.4 Production: Capsid Size and Assembly Viruses come in a huge range of sizes from the 17 nm porcine circovirus 2 (PCV2) to the nearly 1 μm Megaviridae.16,17 Capsid size becomes important in several of the applications of VLPs including drug delivery18 and the templated synthesis of nanoparticles.19 Capsid assembly properties can also vary widely. Many capsids, including HBV,3 CCMV,20 polyomaviridae,8 TMV,21 and brome mosaic virus (BMV)22 undergo pH- and ionic strength-dependent assembly and disassembly. Still other capsids, such as the bacteriophages MS2 and Qβ, require harsher conditions to disassemble their highly stable structures in a reversible way.23 Some viruses, such as red clover necrotic mottle virus (RCNMV) and bacteriophage P22, can also undergo structural transitions affecting the porosity of the capsid when treated with metal chelators24 or high temperature,25 respectively. These structural features can be used to make new materials with properties incredibly difficult to obtain with standard synthetic molecules. Scientists are now taking the self-assembly of capsid components further, creating new, unnatural structures with many of the desirable properties of known capsids. Several studies have taken advantage of the tendency of capsids to assemble around negatively charged polymers of different sizes to alter supramolecular architecture. The BMV capsid assembles into 18 nm, 25 nm, or 28 nm capsids depending on the assembly conditions and encapsulated RNA.26,27 The related icosahedral CCMV coat protein will preferentially assemble into 18 nm, 22 nm or 28 nm particles upon incubation with different molecular weight poly(styrene sulfonate) chains28 or nucleic acids.29,30 Similarly, the normally spherical CCMV capsid can be induced to form tube-like structures when its coat protein is incubated with large DNA templates.31,32 Other spherical viruses such as human immunodeficiency virus (HIV)33 and bacteriophage Qβ34 can also form tube-like struc- 3 tures by introducing mutations to the coat protein. In one recent example, a single point-mutant of Qβ was shown to form three different non-natural assemblies—two spherical and one prolate- shaped.35 Simian virus 40 (SV40) also assembles into several diverse structures depending on the nature of the cargo.36 The assembly of rod-shaped viruses has also been studied extensively over several de- cades. TMV has long been a model system for different assembly states, many of which are used for nanotechnology applications.11,21 Mutants of TMV have resulted in stable rods of various lengths,37,38,39 stable disc-shaped assemblies,40 and even spheres.41 Another interesting engineered assembly comes from filamentous bacteriophage.42 The length of these viruses is controlled by the number of nucleotides in their genome; by manipulating phage DNA, capsids can be made from several micrometers long down to 50 nm. As the rules of self assembly of viral coat proteins become more well-defined, protein design,43,44 targeted mutations,35 calculated genetic recombina- tion,45 and mixed protein assemblies46,47 can lead to even more diverse protein nanostructures.

1.5 Production: Surface Chemistry Virus-like particles, as hollow protein shells, have a defined interior and exterior surface. The properties of the exterior surface can be altered to target capsids to specific tissues,48 change the overall in vivo tropism,49 or scaffold other molecules for precise spatial localization.50 Many approaches use lysine chemistry to append new functionality to the surfaces of capsids; this is often the most facile method but the lack of specificity can counteract the advantages of uniformity in capsid structure. Several groups have incorporated artificial amino acids for bioorthogonal labeling, either by using genetically-encoded Amber stop codon suppression51 or replacement of methionine residues with uniquely reactive analogs during translation.52,53 These approaches allow high-yield, site-selective modification of the capsid surface with a wide variety of useful moieties. As an alternative to the intricacies of these unnatural amino acid incorporation methods, Banerjee and coworkers took a unique approach for bioorthogonal labeling of capsid surfaces.54 Using the known glycosylation site of adenovirus and the ability of mammalian glycosylation machinery to incorporate N-azidoacetylglucosamine (GlcNAz) in place of N-acetylglucosamine (GlcNAc), they produced infective viruses with a “click” reactive handle by metabolic GlcNAz incorporation during virus production.

1.6 Applications: Drug Delivery The uniformity and biocompatibility of viral capsids, along with the myriad of chemical tools available for their modification, have made these structures attractive candidates for drug delivery devices.55,56,57 Research in this area focuses two main goals: 1) attachment of different groups on the surface to target specific cells or evade the immune system of a patient; and 2) methods to load drugs into the interior of the capsid. Covalent attachment of surface targeting groups by the methods described above has led to a wide variety of capsid/target combinations. The vita- min folate is a small molecule that is used frequently as a targeting group because of its receptor’s broad overexpression in many cancers and its relatively straightforward derivitization for attachment.58 Various forms of this molecule have been conjugated to Cucumber mosaic virus (CMV)59, CPMV,60 Hibiscus chlorotic ringspot virus (HCRSV),61 and adenovirus,54 and in each case was shown lead to uptake by target cells. While the targeting group was the same in each of these cases, the capsid size and surface, therapeutic cargo and loading, and capsid production method differ.

4 Folate-targeted CMV was loaded with doxorubicin by simple diffusion of the drug into the capsid interior and physical adsorption; the drug was released slowly over several days. Similarly, in the case of HCRSV, doxorubicin was adsorbed onto poly(styrene sulfonate), which was then loaded into the folate-labeled capsid by disassembly and reassembly. CPMV was labeled with either a simple folate derivative or a poly(ethylene glycol) (PEG) derivative of folate; the researchers did not deliver any drugs, but instead examined the effects of PEG labeling on capsid delivery properties. The polymer significantly reduced non-specific binding and uptake of the VLP. Finally adenovirus capsids modified with folate carried foreign genetic material, loaded during production, to cancer cells. Another small molecule targeting group, lactobionic acid, has been attached to rotavirus VP6 along with doxorubicin for delivery to hepatic cancer cells.9 Collectively, these studies demonstrate the feasibility of targeting capsids using small molecule groups; however, for more specific and efficient delivery to target cells, researchers have turned to the use of large, high- affinity biomolecular targeting groups. The use of large biomolecules as targeting groups opens up possibilities for an astronomi- cal number of combinations of highly specific capsid vehicles carrying drugs tailored for the disease of choice. Synthetic peptides, nucleic acid aptamers, carbohydrates, and entire proteins can be attached to capsid surfaces and combined with engineered drug-loading mechanisms to make incredibly complex yet well-ordered structures. Several viruses have proven amenable to the intri- cate engineering required to scaffold both a biomolecular targeting agent and drug cargo, several of which we will highlight here. RCNMV, for instance, was shown to be cytotoxic to HeLa cells when a 16 amino acid, CD46-binding synthetic peptide was attached to its surface and the capsid interior was loaded with doxorubicin.62 Similarly BMV, conjugated with ferritin and loaded with fluorophores or doxorubicin, was delivered to cancer63 cells. Polyomavirus-like particles, constructed from engineered capsid subunits, were also studied for therapeutic protein delivery64,65 or cellular targeting via single-chain antibody variable fragments (ScFv’s).66 The Finn lab has engineered bacteriophage Qβ for delivery to multiple cell lines. In one study, the Qβ coat protein was coexpressed with an epidermal growth factor (EGF) fusion coat protein, forming hybrid VLPs.67 The hybrid particles were labeled with fluorophores and delivered to A431 cancer cells; the inherent signaling properties of EGF caused a significant decrease in cancer cell viability. Alternatively, whole transferrin proteins along with organic fluorophores were attached to the Qβ surface using click chemistry.68 These assemblies were also efficiently delivered to cells, and it was shown that increasing transferrin labeling levels enhanced cell binding through a polyvalency effect. Recently Rhee and coworkers targeted Qβ containing a photodynamic therapy agent to CD22-containing cells for light-initiated toxicity.69 Transferrin was also used as a targeting group on engineered HK97 capsids to deliver drug-like fluorophores70; it remains to be seen which capsid/targeting group combination is most effective, or what rules can be established for vehicles and target cells to guide further developments. Several other bacteriophages have also been effective scaffolds for targeted drug delivery. Bacteriophage MS2 has been used by many labs to target a variety of cells and tissues. In a pioneering example, Wu and coworkers encapsulated the cytotoxic protein ricin A chain in MS2 and targeted it to cells by surface-attached transferrin.71 Several years later, Ashley and coworkers greatly expanded on this work, encapsulating proteins, drugs, nucleic acids, and nanoparticles and delivering them via surface targeting peptides.72 This work shows the versatility of capsid-based delivery vehicles with the ability to easily change targeting groups and therapeutic cargo for large spectrum of delivery devices. Mutants of this capsid have also been used to photodynamic thera-

5 Alternate 3D Cargo Mineral Drug Tissue PDT/Light Imaging Structures Loading Templating Delivery Targeting Harvesting Spherical viruses PET, MS2 - Several No MRI, Yes Yes Yes Optical Reduced Proteins, Qβ No Optical Yes Yes Yes size, prolate RNA P22 Expanded Proteins Yes MRI No No No Reduced CCMV Several Yes MRI Yes No Yes size, Tubes CPMV - - Yes MRI Yes Yes No Reduced Nano- BMV No Optical Yes Yes No size particles Reduced Polymers, HBV No No No No No Size Proteins Polyoma- Proteins, - No No Yes Yes No virus nucleic acids AAV - Nucleic acids No No No Yes No Rod-shaped and filamentous viruses TMV Several No Yes MRI Yes No Yes Altered Optical, M13/fd No Yes Yes Yes Yes lengths MRI PVX - No No Optical No Yes No Table 1. Applications of several model viral capsids. Abbreviations: AAV, adeno-associated virus; BMV, brome mosaic virus; CCMV, cowpea chlorotic mottle virus; CPMV, cowpea chlorotic mottle virus; HBV; hepatitis B virus; MS2, male- specific bacteriophage 2; PDT, photodynamic therapy; PVX, potato virus X; TMV, tobacco mosaic virus py agents to Jurkat cells.73,74 Using an artificial amino acid, the researchers attached a controlled number of Jurkat-specific aptamers to the MS2 outer surface, in addition to singlet-O2-generating moieties on the inner surface. These constructs effectively killed the cancer cells in a matter of minutes. The use of virus-like particles in drug delivery is still a rapidly growing field. As researchers continue to search for the most effective targeting motifs, labeling techniques, and drug loading, we will obtain a clearer picture of how to make effective therapeutic viral nanoparticles. Many concerns remain, however, about the immunogenicity, biodistribution, and fate of virus-like particles in vivo. A variety of studies have already begun to address these issues.75,76,77,78

1.7 Applications: Imaging Agents Like drug delivery vehicles, targeted imaging agents seek to replace invasive or harm- ful medical procedures by delivering cargo specifically to an affected tissue. Early detection of disease can be critical to successful treatment; a key challenge is to develop materials to increase the ability to visualize diseased tissue in vivo. Imaging can be performed with a wide variety of techniques, but most viral capsid agents focus on enhancing one of three major modes: positron emission tomography (PET), magnetic resonance imaging (MRI); or optical imaging (see refer- ence 79 for a general review). PET imaging involves detection of radiation emitted when certain 6 atomic isotopes decay. The most common isotope for this application is 18F, although 64Cu, 11C, 124I, 68 15 13 74 Ga, O2, N, and As agents are occasionally applied. MRI obtains images based on relaxation of proton spins in an applied magnetic field. Gd3+ chelates and iron oxide particles are commonly used to add contrast to images around targeted tissue. Finally, optical imaging can be performed in some cases, but limited tissue penetration of visible light requires near-infrared fluorescent molecules to be used. The key challenge in making PET imaging agents is to develope fast labeling chemistry 18 79 with little purification to limit the decay of the short-lived isotopes (e.g. F t1/2 = 109.8 min). Untargeted bacteriophage MS2 capsids were used for both 18F PET80 and 64Cu PET.78 Both of these studies analyzed the biodistribution of the untargeted particles, an important first step toward fully characterizing the agents. For targeted imaging, Li and coworkers attached 64Cu chelators and RGD peptides, which bind to integrins often found on tumor cell surfaces, to the bacteriophage T7 capsid.81 Despite high liver uptake, the particles were found to effectively accumulate in tumor tissue. Capsids may also be an effective platform for non-PET nuclear imaging agents, such as 111In-labeled filamentous phage in prostate cancer82 or 99mTc-labeled bacteriophages to visualize bacterial infection.83 Several different capsids have been engineered to serve as enhanced MRI contrast agents. Attachment of paramagnetic molecules to a capsid brings about enhancement in two primary ways: multiple paramagnetic molecules can increase the signal, and a capsid’s large size gives the structure a long tumbling time in solution, which is important for the agent’s relaxivity.79 Many researchers have developed capsids appended with Gd3+ ligands, starting with bacteriophage MS2.84 Later research investigated an often overlooked concept, the rigidity of the capsid-ligand linker.85 This study, using MS2-based agents, suggested more rigid linkers can have a significant effect on relaxivity. Bacteriophage P22 has also provided an effective scaffold; its large size and high loading capacity give it a very large relaxivity on a per particle basis.86,87,88 Notably, when P22 capsids containing polymeric ligand conjugates were used as Gd3+ scaffolds, the per particle relaxivity was among the highest reported. TMV conjugates were recently reported to have even higher per particle relaxivity, though the enhancement per Gd3+ ion was less than that of P22.89 The capsids of Qβ and CPMV have also been examined for Gd3+ ligand scaffolding, indicating this approach is broadly effective.90,91 Despite its wide applicability, Gadolinium (III) is not the only MRI contrast agent. Iron oxide particles are also effective, but less widely used. M13, featuring phage display-developed targeting groups, was coated with the particles and shown to enhance imaging of cancer cells in vivo.92 Spherical capsids of BMV93 and simian virus 40 (SV40)94 were used to encapsulate nanoparticles for high-contrast targetable imaging agents. Capsids have also begun to emerge as “theranostic” agents, combining therapeutic properties with diagnostic imaging. Iron oxide- and dox-loaded rotavirus VP4 VLPs were shown to allow simultaneous imaging and killing of cancer cells using the same capsid-based vehicle.95 Finally, unconventional MRI-based imaging using hyperpolarized 129Xe bound to fd phage can be targeted to cancer cells, clearing the way for capsid technology to enter a new era of imaging.96

1.8 Applications: Templated Synthesis The use of viral capsids to direct the synthesis of metal and mineral nanoparticles has exploded in popularity in the past decade.97 Consequently, we will focus on recent examples that utilize the versatility of capsids, and direct the reader to more specialized reviews for thorough 7 coverage.19,98,99,100,101,102 Nanoparticles and nanowires are inorganic materials with vast potential for use in chemical catalysis, diagnostic imaging, electronic circuits, sensors, batteries, and memory storage.102 Because these structures require nanoscale precision to function properly, researchers have turned to the uniformity, stability, and chemical functionality of capsids for use as templates in the directed assembly of the metal components. Nanoparticle synthesis can take place in the interior of a spherical capsid for uniform size solid particles, or on the exterior for metal-coated proteins or hollow particles. Because viruses are genetically encoded, specific functionality can be easily incorporated into the structure at the DNA level, allowing for placement of metal-active amino acids or peptide sequences with unprecedented spatial precision. P22 was used in a recent example for the synthesis of very uniform iron (III) oxide nanoparticles.103 By genetically fusing an anionic peptide sequence to the interior scaffolding protein, the researchers were able to direct mineralization of iron into the capsid, where particle growth continued until it was spatially constrained by the protein shell. Conversely, metal structures could be grown on the outer surface of P22 by peptide incorporation into the outer capsid protein followed by crystal growth.104 Growth of cad- mium sulfide or zinc sulfide on the surface was templated by the peptide and the protein was gently removed, leaving a hollow metal shell. The CPMV capsid has also been used as a template for many different types of particles. Aljabali and coworkers recently released a series of papers examining different approaches to assembling metals on the surface of the capsid. Certain metals—cobalt, nickel, iron, platinum, and some mixtures—were directed to assemble on the capsid by first depositing a layer of palladium.105 Cobalt–platinum, iron–platinum, or zinc sulfide mixtures could be assembled by first chemically attaching directing peptides to the capsid surface.106 Iron (III) oxide growth was similarly directed by appending negatively charged succinate groups,107 while gold was deposited by first adsorbing a positively charged polymer, poly(allyl amine) hydrochloride, to the capsid surface.108 Each of these assemblies resulted in uniform, capsid-sized nanoparticles with the potential for use in different devices. While none of these examples represents a drastic leap forward for the field, this series highlights the versatility one can achieve by using a stable, highly-expressed, uniform viral capsid for template synthesis. Like nanoparticles, nanowires can also be templated on the interior or exterior of protein surfaces using rod-shaped or filamentous capsids. The M13 capsid has been used for numerous nanowire-type applications because of its genetic tractability and ease of production. M13’s wide success in phage display methods allows functional capsids to be selected, enriched, and easily adapted for new applications.42 In a recent example, a gold-binding peptide was incorporated into the phage P8 proteins, which cover the length of the filament.109 The gold-binding phage was able to efficiently template the synthesis of gold and gold-platinum core-shell nanowires, which were useful for catalytic oxidation reactions. Recently Oh and coworkers improved the performance of lithium-oxygen battery electrodes by plating engineered M13 with a series of metal oxides, polymers, and nanoparticles.110 These highly complex structures enhanced both battery capacity and cycle life. Perhaps the most widely used capsid for templated synthesis is TMV.11 The wild-type capsid has been used extensively, but we will focus on engineered variants. While it has not been shown to be as tolerant to insertions as M13, the capsid can accommodate some mutations, such as recently reported changes that allowed assembly of considerably longer capsids with the same diameter.37,111 By removing a negative charge from the interior and fusing different peptides to the

8 N-terminus, these researchers altered the binding and assembly properties of extra-long capsids expressed in plants111 and E. coli.37 The capsid was also engineered to form head-to-tail assemblies by coating it with poly(aniline).112 These constructs could be further elaborated with silica or titanium, and even incorporated into a petroleum sensing device. A mutant of TMV featuring an added cysteine for metal binding was used to template the growth of palladium nanoparticles.113 The metal-protein protein particles were easily immobilized and recycled, and were found to be active in catalyzing Suzuki coupling reactions. The researchers went on to characterize growth of the particles by small-angle X-ray scattering (SAXS) and found that the capsid template was important for particle size, shape, and uniformity.114 For palladium nanowires, a similar cysteine- containing mutant was used as a template using a reducing agent-free metal deposition reaction.115 This method results in a smooth palladium coating with a palladium oxide surface, which can be reduced with hydrogen gas. Alternative assembly states of TMV also have potential for creating new nanomaterials, such as “nanorings” templated by the disc-shaped state at intermediate pH values.116 Other researchers examined TMV’s in vitro self assembly to create previously unheard of structures; Eber and co-workers coated gold nanoparticles with thiolated RNA molecules, which templated growth of TMV outward to create “nanostars.”117 A similar capsid assembly mechanism using silica- immobilized RNA was used to create hybrid assemblies of capsid mutants featuring alternating blocks of mutant proteins along the rod.118 An alternating capsid of cysteine containing mutants and cysteine-free mutants was modified with a thiol-selective biotin probe, which was then bound to a streptavidin-linked peroxidase. The resulting structures, resembling tubes with alternating surface characteristics, could prove useful for further hybrid nanowire. These unusual shapes suggest that the limits of virus-templated nanotechnology have yet to be reached.

1.9 Applications: Nanoreactors and Scaffolds Engineered capsid particles are also receiving increased attention for applications in nanoscale scaffolding for catalytic reactions, both on the capsid surface and in the interior. Often referred to as nanoreactors, these assemblies can impart significant advantages over free catalysts, be it enzymatic or small molecule catalysis.119 Enzymes localized to the interior of capsids can be useful for protecting and enhancing stability of enzyme cargoes, increasing local concentration and molecular crowding, or even modulating catalytic activity. Horseradish peroxidase was encapsulated by statistical incorporation during reassembly of the CCMV capsid in the presence of the enzyme.120 The researchers used a more controlled interaction between the capsid and another enzyme, Candida antarctica lipase B, to show that enzymatic activity could be enhanced by encapsulation within the shell.121 Encapsulation of proteins in vivo has been demonstrated using both the Qβ and P22 capsids.122,123,124 Fiedler and coworkers co-expressed the Qβ capsid protein, enzymes of interest, and a bivalent RNA aptamer to link the two during capsid assembly,122 while O’Neil and coworkers took advantage of the P22 scaffolding protein’s (SP) ability to accommodate fusions and still incorporate into the structure to encapsulate SP-protein fusions into the capsid.124 The P22 encapsulation system was carried on to encapsulate biotechnologically relevant glyco- hydrolase CelB125 and Brevundimonas diminuta phosphotriesterase.126 Encapsulation of CelB had little effect on enzyme activity and allowed efficient immobilization of the construct in a polymer matrix; encapsulation of the phosphotriesterase significantly increased the enzyme’s stability for bioremediation applications, but with a trade-off in catalytic activity. Our lab has explored the use of charge-based interactions to encapsulate enzymes in the MS2 capsid in vitro,127 akin to non-viral 9 capsid work using the lumazine synthase protein shell.128 We found that encapsulation of model enzyme alkaline phosphatase had little effect on enzyme activity; the varied effects seen among different enzymes in capsids raises questions about what specific parameters are important for efficient encapsulation while preserving catalytic rate. Inhibited substrate access, increased product inhibition, reduced conformational flexibility, enzyme-capsid interaction geometry, and number and size of enzymes within the nanoreactor could all play a role in modulation of activity, indicating further studies are needed to fully understand this potentially useful phenomenon. Nanoreactors are not limited to encapsulated enzymes—immobilization of enzymes or small molecule catalysts on the surface of capsids can also be useful for biotechnology applications. Immobilization of enzymes on the surfaces of capsids has been examined for a handful of enzyme types with the eventual goals of increased enzyme stability, catalytic arrays of multiple enzymes, and immobilized biosensors.119 Glucose oxidase and horseradish peroxidase were conjugated to the surface of CPMV in hopes of creating a sensitive enzymatic glucose sensor for dia- betes patients.129 Though no enhancement was reported, the paper outlined a general strategy for chemically attaching proteins to capsid in vitro. Filamentous potato virus X (PVX) was expressed as a heterologous assembly of wild-type coat proteins and coat proteins fused to Candida antarctica lipase B, resulting in a ratio of approximately 1:3 wild-type: fusion.130 The enzyme activity decreased upon incorporation into the capsid, but the potential kinetic advantage of assembling an enzymatic pathway together in a line on a capsid could outweigh the kinetic cost of immobilization. A noncovalent strategy using an antibody as a binding mediator between a tagged enzyme and filamentous zucchini yellow mosaic virus surface was also shown to bind enzymes to the capsid, though the expense of this approach limits its industrial utility.131 Scaffolding of small molecule catalysts on capsids has been utilized in a wide variety of applications where nanoscale arrangement is crucial. An engineered T4 capsid protein provided an interesting “nanocup” structure useful for embedding and protecting iron-heme catalysts.132 Qβ was also used for a similar scaffolding of iron-heme; engineering poly-histidine tags to express on the exterior of the capsid allowed simple non-covalent attachment of the catalysts, potentially useful for oxidation reactions.47 Nanoantennae, used for collecting light energy and converting it to useful chemical energy, have particularly benefited from this scaffolding onto helical capsids. The TMV capsid has proven to be a very interesting model system to study how structure relates to energy transfer between appended light-harvesting molecules; transfer was much more efficient in the rod structure than in the disc structure, a useful insight for both engineered and natural light-harvesting.133,134,135 Bacteriophage M13 has been used for photocatalytic water splitting for hydrogen production.136 The researchers attached both zinc porphyrin molecules and iridium oxide clusters to the outside of the capsid and found that the system was stable, recyclable, and effective for producing hydrogen. Similarly, M13 could scaffold organic dyes, titanium dioxide, and gold nanoparticles to create a highly efficient dye-sensitized solar cell.137 Finally, MS2 was also used to learn more about a poorly understood light-harvesting interaction between organic fluorophores and metal nanoparticles. By encapsulating gold nanoparticles inside and positioning external fluorophores specific distances away, the effect of the nanoparticle on the fluorescence lifetime was elucidated.138

1.10 Nonviral Protein Scaffolds In addition to the wealth of useful structures offered by viral capsids, there are several other natural and engineered protein assemblies with similar applications. Ferritin and ferritin-like 10 proteins, for example, are popular scaffolds for metal nanoparticle synthesis due to their natural metal-binding properties.139 The Hilvert lab has engineered the lumazine synthase assembly for protein encapsulation both in vitro128 and in vivo.140 This structure was evolved to shield the expressing cells from toxic enzymes directed to the interior. Vault protein particles, common in eukaryotic organisms, are also being developed for encapsulation, drug delivery, and materials.141 With protein design methods coming of age, designed protein assemblies with potential for new shapes are now possible.142 Finally, natural protein assemblies involved in metabolic systems such as bacterial microcompartments and cellulase scaffolds can be utilized for new reactions in vivo.143

1.11 Conclusions and Perspectives The astounding diversity and versatility of viral capsids have proven incredibly useful for engineering nanotechnological devices. From a self-assembly perspective, few other structures can match the combination of precise molecular placement from inter-subunit interactions and chemical flexibility afforded by genetic manipulation. Combined with chemical modifications, these properties make viral capsids ideal for scaffolding nanodevices of all types. Despite many recent developments with these structures, there is still much to learn to facilitate further engineering efforts. The rules of capsid subunit folding and assembly are still being defined at the molecular level—knowledge of such rules will enable better prediction of mutations that can be tolerated and new structures that can be formed. To reduce immunogenicity for therapeutic agents, enhance ad- juvant activity for vaccine carriers, and understand the factors affecting biodistribution, researchers will also need to develop insights into how capsid protein sequence and structure govern interactions with the human body. Finally, the production of capsid-based materials needs to become more cost-effective to permit commercial application in electronics, sensors, and catalysts. Future applications of viral capsids will likely use cues from both nanotechnology and nature to yield novel structures and devices. For example, many natural capsids have key structural changes in response to external stimuli, such as disassembly or enzyme activation, which could be exploited to make “smart” biological devices. Encapsulation methods that incorporate or control the addition of foreign materials—nucleic acids, proteins, polymers, or drugs—into these structures are also on the horizon. Finally, developing methods of interfacing capsids with electronics can enhance performance or create new functions for bionanotechnology.

1.12 References 1. Mateu, M.G. (2013). Assembly, Stability and Dynamics of Virus Capsids. Arch. of Biochem. Biophys. 531, 65–79. 2. Caspar, D. and Klug, A. (1962). Physical Principles in the Construction of Regular Viruses. Cold Spring Harb. Symp. Quant. Biol. 27, 1-24. 3. Katen, S. and Zlotnick, A. (2009). Chapter 14: The Thermodynamics of Virus Capsid As- sembly. In Methods in Enzymology, (Elsevier), pp 395–417. 4. Roldão, A., Mellado, M.C.M., Castilho, L.R., Carrondo, M.J., and Alves, P.M. (2010). Virus- like Particles in Vaccine Development. Exp. Rev. Vaccines 9, 1149–1176. 5. Kay, M.A. (2011). State-of-the-art Gene-based Therapies: The Road Ahead. Nat. Rev. Genet. 12, 316–328. 6. Bratkovič, T. (2010). Progress in Phage Display: Evolution of the Technique and its Applica- tions. Cell Mol. Life Sci. 67, 749–767.

11 7. Zhao, X., Fox J.M., Olson, N.H., Baker, T.S., and Young, M.J. (1995). In Vitro Assembly of Cowpea Chlorotic Mottle Virus from Coat Protein Expressed in Escherichia coli and In Vitro-Transcribed Viral cDNA. Virology 207, 486–494. 8. Teunissen, E.A., de Raad, M., and Mastrobattista, E. (2013). Production and Biomedical Applications of Virus-like Particles Derived from Polyomaviruses. J. Control. Release 172, 305–321. 9. Zhao, Q., Chen, W., Chen, Y., Zhang, L., Zhang, J., and Zhang, Z. (2011). Self-Assembled Virus-like Particles from Rotavirus Structural Protein VP6 for Targeted Drug Delivery. Bio- conjugate Chem. 22, 346–352. 10. Bundy, B.C., Franciszkowicz, M.J., and Swartz, J.R. (2008). Escherichia coli-based Cell- free Synthesis of Virus-like Particles. Biotechnol. Bioeng. 100, 28–37. 11. Alonso, J.M., Górzny, M.Ł., and Bittner, A.M. (2013). The Physics of Tobacco Mosaic Virus and Virus-based Devices in Biotechnology. Trends Biotech. 31, 530–538. 12. Sainsbury, F., Cañizares, M.C., and Lomonossoff, G.P. (2010). Cowpea Mosaic Virus: The Plant Virus–Based Biotechnology Workhorse. Ann. Rev. Phytopathol. 48, 437–455. 13. Steinmetz, N.F., Bize, A., Findlay, K.C., Lomonossoff, G.P., Manchester, M., Evans, D.J., and Prangishvili, D. (2008). Site-specific and Spatially Controlled Addressability of a New Viral Nanobuilding Block: Sulfolobus islandicus Rod-shaped Virus 2. Adv. Funct. Mater. 18, 3478–3486. 14. Rice, G., Stedman, K., Snyder, J., Wiedenheft, B., Willits, D., Brumfield, S., McDermott, T., and Young, M.J. (2001). Viruses from Extreme Thermal Environments. Proc. Nat. Acad. Sci. U.S.A. 98, 13341–13345. 15. Rabindran, S., and Dawson, W.O. (2001). Assessment of Recombinants That Arise from the Use of a TMV-Based Transient Expression Vector. Virology 284, 182–189. 16. Khayat, R., Brunn, N., Speir, J.A., Hardham, J.M., Ankenbauer, R.G., Schneemann, A., and Johnson, J.E. (2011). The 2.3-Angstrom Structure of Porcine Circovirus 2. J. Virol. 85, 7856–7862. 17. Philippe, N., Legendre, M., Doutre, G., Couté, Y., Poirot, O., Lescot, M., Arslan, D., Selt- zer, V., Bertaux, L., Bruley, C., Garin, J., Claverie, J., Abergel, C. (2013). Pandoraviruses: Amoeba Viruses with Genomes Up to 2.5 Mb Reaching That of Parasitic Eukaryotes. Sci- ence 341, 281–286. 18. Zhang, S., Li, J., Lykotrafitis, G., Bao, G., and Suresh, S. (2009). Size-Dependent Endocyto- sis of Nanoparticles. Adv. Mater. 21, 419–424. 19. Crookes-Goodson, W.J., Slocik, J.M., and Naik, R.R. (2008). Bio-directed synthesis and assembly of nanomaterials. Chem. Soc. Rev. 37, 2403. 20. Adolph, K.W., and Butler, P.J.G. (1974). Studies on the Assembly of a Spherical Plant Virus: I. States of Aggregation of the Isolated Protein. J. Mol. Biol. 88, 327–341. 21. Durham, A.C.H., Finch, J.T., and Klug, A. (1971). States of Aggregation of Tobacco Mosaic Virus Protein. Nature 229, 37–42. 22. Pfeiffer, P., and Hirth, L. (1974). Aggregation States of Brome Mosaic Virus Protein. Virol- ogy 61, 160–167. 23. Hung, P.P., Ling, C.M., Overby, L.R. (1969). Self Assembly of QB and MS2 Phage Particles Possible Function of Initiation Complexes. Science 166, 1638-1640. 24. Sherman, M.B., Guenther, R.H., Tama, F., Sit, T.L., Brooks, C.L., Mikhailov, A.M., Orlova, E.V., Baker, T.S., and Lommel, S.A. (2006). Removal of Divalent Cations Induces Structural

12 Transitions in Red Clover Necrotic Mosaic Virus, Revealing a Potential Mechanism for RNA Release. J. Virol. 80, 10395–10406. 25. Jiang, W., Li, Z., Zhang, Z., Baker, M.L., Prevelige, P.E., and Chiu, W. (2003). Coat Protein Fold and Maturation Transition of Bacteriophage P22 Seen at Subnanometer Resolutions. Nat. Struct. Mol. Biol. 10, 131–135. 26. Krol, M.A., Olson, N.H., Tate, J., Johnson, J.E., Baker, T.S., and Ahlquist, P. (1999). RNA- Controlled Polymorphism in the In Vivo Assembly of 180-subunit and 120-subunit Virions from a Single Capsid Protein. Proc. Nat. Acad. Sci. U.S.A. 96, 13650–13655. 27. Lucas, R.W., Kuznetsov, Y.G., Larson, S.B., and McPherson, A. (2001). Crystallization of Brome Mosaic Virus and T = 1 Brome Mosaic Virus Particles Following a Structural Transi- tion. Virology 286, 290–303. 28. Hu, Y., Zandi, R., Anavitarte, A., Knobler, C.M., and Gelbart, W.M. (2008). Packaging of a Polymer by a Viral Capsid: The Interplay between Polymer Length and Capsid Size. Bio- phys. J. 94, 1428–1436. 29. Minten, I.J., Ma, Y., Hempenius, M.A., Vancso, G.J., Nolte, R.J.M., and Cornelissen, J.J.L.M. (2009). CCMV Capsid Formation Induced by a Functional Negatively Charged Polymer. Org. Biomol. Chem. 7, 4685-4688. 30. Cadena-Nava, R.D., Comas-Garcia, M., Garmann, R.F., Rao, A.L.N., Knobler, C.M., and Gelbart, W.M. (2012). Self-Assembly of Viral Capsid Protein and RNA Molecules of Differ- ent Sizes: Requirement for a Specific High Protein/RNA Mass Ratio. J. Virol. 86, 3318–3326. 31. Mukherjee, S., Pfeifer, C.M., Johnson, J.M., Liu, J., and Zlotnick, A. (2006). Redirecting the Coat Protein of a Spherical Virus to Assemble into Tubular Nanostructures. J. Am. Chem. Soc. 128, 2538–2539. 32. De la Escosura, A., Janssen, P.G.A., Schenning, A.P.H.J., Nolte, R.J.M., and Cornelissen, J.J.L.M. (2010). Encapsulation of DNA-Templated Chromophore Assemblies within Virus Protein Nanotubes. Angew. Chem. Int. Edit. 49, 5335–5338. 33. Gross, I., Hohenberg, H., Huckhagel, C., and Kräusslich, H.G. (1998). N-Terminal Extension of Human Immunodeficiency Virus Capsid Protein Converts the In Vitro Assembly Pheno- type from Tubular to Spherical Particles. J. Virol. 72, 4798–4810. 34. Cielens, I., Ose, V., Petrovskis, I., Strelnikova, A., Renhofa, R., Kozlovska, T., and Pumpens, P. (2000). Mutilation of RNA Phage Qβ Virus-like Particles: From Icosahedrons to Rods. FEBS Letters 482, 261–264. 35. Fiedler, J.D., Higginson, C., Hovlid, M.L., Kislukhin, A.A., Castillejos, A., Manzenrieder, F., Campbell, M.G., Voss, N.R., Potter, C.S., Carragher, B., Finn, M.G. (2012). Engineered Mutations Change the Structure and Stability of a Virus-Like Particle. Biomacromolecules 13, 2339–2348. 36. Kler, S., Wang, J.C., Dhason, M., Oppenheim, A., and Zlotnick, A. (2013). Scaffold Proper- ties Are a Key Determinant of the Size and Shape of Self-Assembled Virus-Derived Par- ticles. ACS Chem. Biol. 8, 2753–2761. 37. Brown, A.D., Naves, L., Wang, X., Ghodssi, R., and Culver, J.N. (2013). Carboxylate-Di- rected In Vivo Assembly of Virus-like Nanorods and Tubes for the Display of Functional Peptides and Residues. Biomacromolecules 14, 3123–3129. 38. Bruckman, M.A., Soto, C.M., McDowell, H., Liu, J.L., Ratna, B.R., Korpany, K.V., Zahr, O.K., and Blum, A.S. (2011). Role of Hexahistidine in Directed Nanoassemblies of Tobacco Mosaic Virus Coat Protein. ACS Nano 5, 1606–1616.

13 39. Butler, P.J.G., and Finch, J.T. (1973). Structures and Roles of the Polymorphic Forms of Tobacco Mosaic Virus Protein: VII. Lengths of the Growing Rods During Assembly into Nucleoprotein with the Viral RNA. J. Mol. Biol. 78, 637–649. 40. Dedeo, M.T., Duderstadt, K.E., Berger, J.M., and Francis, M.B. (2010). Nanoscale Protein Assemblies from a Circular Permutant of the Tobacco Mosaic Virus. Nano Lett. 10, 181–186. 41. Atabekov, J., Nikitin, N., Arkhipenko, M., Chirkov, S., and Karpova, O. (2011). Thermal Transition of Native Tobacco Mosaic Virus and RNA-free Viral Proteins into Spherical Nanoparticles. J. Gen. Virol. 92, 453–456. 42. Rakonjac, J., Bennett, N.J., Spagnuolo, J., Gagic, D., and Russel, M. (2011). Filamentous Bacteriophage: Biology, Phage Display and Nanotechnology Applications. Curr. Issues Mol. Biol. 13, 51–76. 43. King, N.P., and Lai, Y. (2013). Practical Approaches to Designing Novel Protein Assemblies. Curr. Opin. Struct. Biol. 23, 632–638. 44. Van Eldijk, M.B., Wang, J.C.Y., Minten, I.J., Li, C., Zlotnick, A., Nolte, R.J.M., Cornelissen, J.J.L.M., and van Hest, J.C.M. (2012). Designing Two Self-Assembly Mechanisms into One Viral Capsid Protein. J. Am. Chem. Soc. 134, 18506–18509. 45. Ho, M.L., Adler, B.A., Torre, M.L., Silberg, J.J., and Suh, J. (2013). SCHEMA Computation- al Design of Virus Capsid Chimeras: Calibrating How Genome Packaging, Protection, and Transduction Correlate with Calculated Structural Disruption. ACS Synth. Biol. 2, 724–733. 46. Rumnieks, J., Ose, V., Tars, K., Dislers, A., Strods, A., Cielens, I., and Renhofa, R. (2009). Assembly of Mixed Rod-like and Spherical Particles from Group I and II RNA Bacterio- phage Coat Proteins. Virology 391, 187–194. 47. Udit, A.K., Hollingsworth, W., and Choi, K. (2010). Metal- and Metallocycle-Binding Sites Engineered into Polyvalent Virus-Like Scaffolds. Bioconjugate Chem. 21, 399–404. 48. Smith, M.T., Hawes, A.K., and Bundy, B.C. (2013). Reengineering viruses and virus-like particles through chemical functionalization strategies. Curr. Opin. Biotech. 24, 620–626. 49. Kwon, I., and Schaffer, D.V. (2008). Designer Gene Delivery Vectors: Molecular Engineer- ing and Evolution of Adeno-Associated Viral Vectors for Enhanced Gene Transfer. Pharm. Res. 25, 489–499. 50. Plummer, E.M., and Manchester, M. (2011). Viral Nanoparticles and Virus-like Particles: Platforms for Contemporary Vaccine Design. Wiley Interdisciplinary Reviews: Nanomedi- cine and Nanobiotechnology 3, 174–196. 51. Carrico, Z.M., Romanini, D.W., Mehl, R.A., and Francis, M.B. (2008). Oxidative Coupling of Peptides to a Virus Capsid Containing Unnatural Amino Acids. Chem. Commun. 10, 1205-1207. 52. Strable, E., Prasuhn, D.E., Udit, A.K., Brown, S., Link, A.J., Ngo, J.T., Lander, G., Quispe, J., Potter, C.S., Carragher, B., Tirrell, D., Finn, M.G. (2008). Unnatural Amino Acid Incor- poration into Virus-Like Particles. Bioconjugate Chem. 19, 866–875. 53. Patel, K.G., and Swartz, J.R. (2011). Surface Functionalization of Virus-Like Particles by Di- rect Conjugation Using Azide−Alkyne Click Chemistry. Bioconjugate Chem. 22, 376–387. 54. Banerjee, P.S., Ostapchuk, P., Hearing, P., and Carrico, I. (2010). Chemoselective Attach- ment of Small Molecule Effector Functionality to Human Adenoviruses Facilitates Gene Delivery to Cancer Cells. J. Am. Chem. Soc. 132, 13615–13617. 55. Yildiz, I., Shukla, S., and Steinmetz, N.F. (2011). Applications of Viral Nanoparticles in Medicine. Curr. Opin. Biotech. 22, 901–908.

14 56. Ma, Y., Nolte, R.J.M., and Cornelissen, J.J.L.M. (2012). Virus-based Nanocarriers for Drug Delivery. Advanced Drug Delivery Reviews 64, 811–825. 57. Ren, Y., Wong, S.M., and Lim, L.Y. (2010). Application of Plant Viruses as Nano Drug De- livery Systems. Pharm. Res. 27, 2509–2513. 58. Low, P.S., and Kularatne, S.A. (2009). Folate-targeted Therapeutic and Imaging Agents for Cancer. Curr. Opin. Chem. Biol. 13, 256–262. 59. Zeng, Q., Wen, H., Wen, Q., Chen, X., Wang, Y., Xuan, W., Liang, J., and Wan, S. (2013). Cucumber Mosaic Virus as Drug Delivery Vehicle for Doxorubicin. Biomaterials 34, 4632– 4642. 60. Destito, G., Yeh, R., Rae, C.S., Finn, M.G., and Manchester, M. (2007). Folic Acid-Mediated Targeting of Cowpea Mosaic Virus Particles to Tumor Cells. Chem. Biol. 14, 1152–1162. 61. Ren, Y., Wong, S.M., and Lim, L. (2007). Folic Acid-Conjugated Protein Cages of a Plant Virus: A Novel Delivery Platform for Doxorubicin. Bioconjugate Chem. 18, 836–843. 62. Lockney, D.M., Guenther, R.N., Loo, L., Overton, W., Antonelli, R., Clark, J., Hu, M., Luft, C., Lommel, S.A., and Franzen, S. (2011). The Red Clover Necrotic Mosaic Virus Capsid as a Multifunctional Cell Targeting Plant Viral Nanoparticle. Bioconjugate Chem. 22, 67–73. 63. Yildiz, I., Tsvetkova, I., Wen, A.M., Shukla, S., Masarapu, M.H., Dragnea, B., and Stein- metz, N.F. (2012). Engineering of Brome Mosaic Virus for Biomedical Applications. RSC Advances 2, 3670-3677. 64. Schmidt, U., Günther, C., Rudolph, R., and Böhm, G. (2001). Protein and Peptide Delivery via Engineered Polyomavirus-like Particles. FASEB J. 15, 1646–1648. 65. Abbing, A., Blaschke, U.K., Grein, S., Kretschmar, M., Stark, C.M.B., Thies, M.J.W., Wal- ter, J., Weigand, M., Woith, D.C., Hess, J., Reiser, C.O. (2004). Efficient Intracellular Deliv- ery of a Protein and a Low Molecular Weight Substance via Recombinant Polyomavirus-like Particles. J. Biol. Chem. 279, 27410–27421. 66. Stubenrauch, K., Gleiter, S., Brinkmann, U., Rudolph, R., and Lilie, H. (2001). Conjugation of an Antibody Fv Fragment to a Virus Coat Protein: Cell-specific Targeting of Recombinant Polyoma-virus-like Particles Biochem. J. 356, 867–873. 67. Pokorski, J.K., Hovlid, M.L., and Finn, M.G. (2011). Cell Targeting with Hybrid Qβ Virus- Like Particles Displaying Epidermal Growth Factor. ChemBioChem 12, 2441–2447. 68. Banerjee, D., Liu, A.P., Voss, N.R., Schmid, S.L., and Finn, M.G. (2010). Multivalent Dis- play and Receptor-Mediated Endocytosis of Transferrin on Virus-Like Particles. ChemBio- Chem 11, 1273–1279. 69. Rhee, J., Baksh, M., Nycholat, C., Paulson, J.C., Kitagishi, H., and Finn, M.G. (2012). Gly- can-Targeted Virus-like Nanoparticles for Photodynamic Therapy. Biomacromolecules 13, 2333–2338. 70. Huang, R.K., Steinmetz, N.F., Fu, C., Manchester, M., and Johnson, J.E. (2011). Transferrin- mediated Targeting of Bacteriophage HK97 Nanoparticles into Tumor Cells. Nanomedicine (Lond) 6, 55–68. 71. Wu, M., Brown, W.L., and Stockley, P.G. (1995). Cell-specific Delivery of Bacteriophage- encapsidated Ricin A Chain. Bioconjugate Chem. 6, 587–595. 72. Ashley, C.E., Carnes, E.C., Phillips, G.K., Durfee, P.N., Buley, M.D., Lino, C.A., Padilla, D.P., Phillips, B., Carter, M.B., Willman, C.L., Brinker. C.J., Caldeira, J., Chackerian, B., Wharton, W., Peabody, D.S. (2011). Cell-Specific Delivery of Diverse Cargos by Bacterio- phage MS2 Virus-like Particles. ACS Nano 5, 5729–5745.

15 73. Tong, G.J., Hsiao, S.C., Carrico, Z.M., and Francis, M.B. (2009). Viral Capsid DNA Aptamer Conjugates as Multivalent Cell-Targeting Vehicles. J. Am. Chem. Soc. 131, 11174–11178. 74. Stephanopoulos, N., Tong, G.J., Hsiao, S.C., and Francis, M.B. (2010). Dual-Surface Modi- fied Virus Capsids for Targeted Delivery of Photodynamic Agents to Cancer Cells. ACS Nano 4, 6014–6020. 75. Singh, P., Prasuhn, D., Yeh, R.M., Destito, G., Rae, C.S., Osborn, K., Finn, M.G., and Manchester, M. (2007). Bio-distribution, Toxicity and Pathology of Cowpea Mosaic Virus Nanoparticles In Vivo. J. Control. Release 120, 41–50. 76. Kaiser, C.R., Flenniken, M.L., Gillitzer, E., Harmsen, A.L., Harmsen, A.G., Jutila, M.A., Douglas, T., and Young, M.J. (2007). Biodistribution Studies of Protein Cage Nanoparticles Demonstrate Broad Tissue Distribution and Rapid Clearance In Vivo. Int. J. Nanomedicine 2, 715–733. 77. Plummer, E.M., and Manchester, M. (2013). Endocytic Uptake Pathways Utilized by CPMV Nanoparticles. Mol. Pharmaceutics 10, 26–32. 78. Farkas, M.E., Aanei, I.L., Behrens, C.R., Tong, G.J., Murphy, S.T., O’Neil, J.P., and Francis, M.B. (2013). PET Imaging and Biodistribution of Chemically Modified Bacteriophage MS2. Mol. Pharmaceutics 10, 69–76. 79. Fass, L. (2008). Imaging and Cancer: A Review. Mol. Oncol. 2, 115–152. 80. Hooker, J.M., O’Neil, J.P., Romanini, D.W., Taylor, S.E., and Francis, M.B. (2008). Ge- nome-free Viral Capsids as Carriers for Positron Emission Tomography Radiolabels. Mol. Imaging Biol. 10, 182–191. 81. Li, Z., Jin, Q., Huang, C., Dasa, S., Chen, L., Yap, L., Liu, S., Cai, H., Park, R., and Conti, P.S. (2011). Trackable and Targeted Phage as Positron Emission Tomography (PET) Agent for Cancer Imaging. Theranostics 1, 371–380. 82. Newton-Northup, J.R., Figueroa, S.D., Quinn, T.P., and Deutscher, S.L. (2009). Bifunctional Phage-based Pretargeted Imaging of Human Prostate Carcinoma. Nucl. Med. Biol. 36, 789– 800. 83. Rusckowski, M., Gupta, S., Liu, G., Dou, S., and Hnatowich, D.J. (2008). Investigation of Four 99mTc-labeled Bacteriophages for Infection-specific Imaging. Nucl. Med. Biol. 35, 433–440. 84. Anderson, E.A., Isaacman, S., Peabody, D.S., Wang, E.Y., Canary, J.W., and Kirshenbaum, K. (2006). Viral Nanoparticles Donning a Paramagnetic Coat: Conjugation of MRI Contrast Agents to the MS2 Capsid. Nano Lett. 6, 1160–1164. 85. Garimella, P.D., Datta, A., Romanini, D.W., Raymond, K.N., and Francis, M.B. (2011). Mul- tivalent, High-Relaxivity MRI Contrast Agents Using Rigid Cysteine-Reactive Gadolinium Complexes. J. Am. Chem. Soc. 133, 14704–14709. 86. Min, J., Jung, H., Shin, H., Cho, G., Cho, H., and Kang, S. (2013). Implementation of P22 Vi- ral Capsids As Intravascular Magnetic Resonance T1 Contrast Conjugates via Site-Selective Attachment of Gd(III)-Chelating Agents. Biomacromolecules 14, 2332–2339. 87. Lucon, J., Qazi, S., Uchida, M., Bedwell, G.J., LaFrance, B., Prevelige, P.E., and Doug- las, T. (2012). Use of the Interior Cavity of the P22 Capsid for Site-specific Initiation of Atom-transfer Radical Polymerization with High-density Cargo Loading. Nature Chemistry 4, 781–788. 88. Qazi, S., Liepold, L.O., Abedin, M.J., Johnson, B., Prevelige, P., Frank, J.A., and Douglas, T. (2013). P22 Viral Capsids as Nanocomposite High-Relaxivity MRI Contrast Agents. Mol.

16 Pharm. 10, 11–17. 89. Bruckman, M.A., Hern, S., Jiang, K., Flask, C.A., Yu, X., and Steinmetz, N.F. (2013). Tobac- co Mosaic Virus Rods and Spheres as Supramolecular High-relaxivity MRI Contrast Agents. J. Mater. Chem. B 1, 1482-1490. 90. Prasuhn, Jr. D.E., Yeh, R.M., Obenaus, A., Manchester, M., and Finn, M.G. (2007). Viral MRI Contrast Agents: Coordination of Gd by Native Virions and Attachment of Gd Com- plexes by Azide-alkyne Cycloaddition. Chem. Commun. 12, 1269-1271. 91. Shriver, L.P., Plummer, E.M., Thomas, D.M., Ho, S., and Manchester, M. (2013). Localiza- tion of Gadolinium-loaded CPMV to Sites of Inflammation During Central Nervous System Autoimmunity. J. Mater. Chem. B 1, 5256-5263. 92. Ghosh, D., Lee, Y., Thomas, S., Kohli, A.G., Yun, D.S., Belcher, A.M., and Kelly, K.A. (2012). M13-templated Magnetic Nanoparticles for Targeted In Vivo Imaging of Prostate Cancer. Nat. Nanotechnol 7, 677–682. 93. Huang, X., Bronstein, L.M., Retrum, J., Dufort, C., Tsvetkova, I., Aniagyei, S., Stein, B., Stucky, G., McKenna, B., Remmes, N., Baxter, D., Kao, C.C., Dragna, B. (2007). Self- Assembled Virus-like Particles with Magnetic Cores. Nano Lett. 7, 2407–2416. 94. Enomoto, T., Kawano, M., Fukuda, H., Sawada, W., Inoue, T., Haw, K.C., Kita, Y., Saka- moto, S., Yamaguchi, Y., Imai, T., Hatakeyama, M., Saito, S., Sandhu, A., Matsui, M., Aoki, I., Handa, H. (2013). Viral Protein-coating of Magnetic Nanoparticles Using Simian Virus 40 VP1. J. Biotechnol. 167, 8–15. 95. Chen, W., Cao, Y., Liu, M., Zhao, Q., Huang, J., Zhang, H., Deng, Z., Dai, J., Williams, D.F.,

and Zhang, Z. (2012). Rotavirus Capsid Surface Protein VP4-coated Fe3O4 Nanoparticles as a Theranostic Platform for Cellular Imaging and Drug Delivery. Biomaterials 33, 7895–7902. 96. Palaniappan, K.K., Ramirez, R.M., Bajaj, V.S., Wemmer, D.E., Pines, A., and Francis, M.B. (2013). Molecular Imaging of Cancer Cells Using a Bacteriophage-Based 129Xe NMR Bio- sensor. Angewandte Chemie 125, 4949–4953. 97. Bronstein, L.M. (2011). Virus-Based Nanoparticles with Inorganic Cargo: What Does the Future Hold? Small 7, 1609–1618. 98. Soto, C.M., and Ratna, B.R. (2010). Virus hybrids as nanomaterials for biotechnology. Curr. Opin. Biotech. 21, 426–438. 99. Lomonossoff, G.P., and Evans, D.J. (2011). Applications of Plant Viruses in Bionanotechnol- ogy (Springer Berlin Heidelberg), pp. 1–27. 100. Lee, S., Lim, J., and Harris, M.T. (2012). Synthesis and Application of Virus-based Hybrid Nanomaterials. Biotechnol. Bioeng. 109, 16–30. 101. Liu, Z., Qiao, J., Niu, Z., and Wang, Q. (2012). Natural Supramolecular Building Blocks: From Virus Coat Proteins to Viral Nanoparticles. Chem. Soc. Rev. 41, 6178-6194. 102. Li, F., and Wang, Q. (2013). Fabrication of Nanoarchitectures Templated by Virus-Based Nanoparticles: Strategies and Applications. Small 10, 230-245. 103. Reichhardt, C., Uchida, M., O’Neil, A., Li, R., Prevelige, P.E., and Douglas, T. (2011). Tem- plated Assembly of Organic–inorganic Materials Using the Core Shell Structure of the P22 Bacteriophage. Chem. Commun. 47, 6326-6328. 104. Shen, L., Bao, N., Prevelige, P.E., and Gupta, A. (2010). Fabrication of Ordered Nanostruc- tures of Sulfide Nanocrystal Assemblies over Self-Assembled Genetically Engineered P22 Coat Protein. J. Am. Chem. Soc. 132, 17354–17357. 105. Aljabali, A.A.A., Barclay, J.E., Lomonossoff, G.P., and Evans, D.J. (2010). Virus Templated

17 Metallic Nanoparticles. Nanoscale 2, 2596-2600. 106. Aljabali, A.A.A., Sainsbury, F., Lomonossoff, G.P., and Evans, D.J. (2010). Cowpea Mosaic Virus Unmodified Empty Viruslike Particles Loaded with Metal and Metal Oxide. Small 6, 818–821. 107. Aljabali, A.A.A., Barclay, J.E., Cespedes, O., Rashid, A., Staniland, S.S., Lomonossoff, G.P., and Evans, D.J. (2011). Charge Modified Cowpea Mosaic Virus Particles for Templated Min- eralization. Adv. Funct. Mater. 21, 4137–4142. 108. Aljabali, A.A.A., Shah, S.N., Evans-Gowing, R., Lomonossoff, G.P., Evans, D.J. (2011). Chemically-coupled-peptide-promoted Virus Nanoparticle Templated Mineralization. Inte- grative Biology 3, 119-125. 109. Lee, Y., Kim, J., Yun, D.S., Nam, Y.S., Shao-Horn, Y., and Belcher, A.M. (2012). Virus- templated Au and Au–Pt Core–shell Nanowires and their Electrocatalytic Activities for Fuel Cell Applications. Energ. Environ. Sci. 5, 8328. 110. Oh, D., Qi, J., Lu, Y., Zhang, Y., Shao-Horn, Y., and Belcher, A.M. (2013). Biologically En- hanced Cathode Design for Improved Capacity and Cycle Life for Lithium-oxygen Batteries. Nat Commun. 4, DOI: 10.1038/ncomms3756. 111. Kadri, A., Maiß, E., Amsharov, N., Bittner, A.M., Balci, S., Kern, K., Jeske, H., and Wege, C. (2011). Engineered Tobacco Mosaic Virus Mutants with Distinct Physical Characteristics In Planta and Enhanced Metallization Properties. Virus Research 157, 35–46. 112. Rong, J., Oberbeck, F., Wang, X., Li, X., Oxsher, J., Niu, Z., and Wang, Q. (2009). Tobacco Mosaic Virus Templated Synthesis of One Dimensional Inorganic–polymer Hybrid Fibres. J. Mater. Chem. 19, 2841-2845. 113. Manocchi, A.K., Seifert, S., Lee, B., and Yi, H. (2011). In Situ Small-Angle X-ray Scatter- ing Analysis of Palladium Nanoparticle Growth on Tobacco Mosaic Virus Nanotemplates. Langmuir 27, 7052–7058. 114. Yang, C., Manocchi, A.K., Lee, B., and Yi, H. (2011). Viral-templated Palladium Nanocata- lysts for Suzuki Coupling Reaction. J. Mater. Chem. 21, 187-194. 115. Lim, J., Kim, S., Lee, S., Stach, E.A., Culver, J.N., and Harris, M.T. (2010). Biotemplat- ed Aqueous-Phase Palladium Crystallization in the Absence of External Reducing Agents. Nano Lett. 10, 3863–3867. 116. Zahr, O.K., and Blum, A.S. (2012). Solution Phase Gold Nanorings on a Viral Protein Tem- plate. Nano Lett. 12, 629–633. 117. Eber, F.J., Eiben, S., Jeske, H., and Wege, C. (2013). Bottom-Up-Assembled Nanostar Col- loids of Gold Cores and Tubes Derived From Tobacco Mosaic Virus. Angewandte Chemie 125, 7344–7348. 118. Geiger, F.C., Eber, F.J., Eiben, S., Mueller, A., Jeske, H., Spatz, J.P., and Wege, C. (2013). TMV Nanorods with Programmed Longitudinal Domains of Differently Addressable Coat Proteins. Nanoscale 5, 3808–3816. 119. Cardinale, D., Carette, N., and Michon, T. (2012). Virus Scaffolds as Enzyme Nano-carriers. Trends Biotechnol. 30, 369–376. 120. Comellas-Aragonès, M., Engelkamp, H., Claessen, V.I., Sommerdijk, N.A.J.M., Rowan, A.E., Christianen, P.C.M., Maan, J.C., Verduin, B.J.M., Cornelissen, J.J.L.M., and Nolte, R.J.M. (2007). A Virus-based Single-enzyme Nanoreactor. Nat. Nanotechnol. 2, 635–639. 121. Minten, I.J., Claessen, V.I., Blank, K., Rowan, A.E., Nolte, R.J.M., and Cornelissen, J.J.L.M. (2011). Catalytic Capsids: The Art of Confinement. Chemical Science 2, 358-362.

18 122. Fiedler, J.D., Brown, S.D., Lau, J.L., and Finn, M.G. (2010). RNA-Directed Packaging of Enzymes within Virus-like Particles. Angew. Chem. Int. Edit. 49, 9648–9651. 123. O’Neil, A., Reichhardt, C., Johnson, B., Prevelige, P.E., and Douglas, T. (2011). Genetically Programmed In Vivo Packaging of Protein Cargo and Its Controlled Release from Bacterio- phage P22. Angew. Chem. Int. Edit. 50, 7425–7428. 124. Patterson, D.P., Prevelige, P.E., and Douglas, T. (2012). Nanoreactors by Programmed En- zyme Encapsulation Inside the Capsid of the Bacteriophage P22. ACS Nano 6, 5000–5009. 125. Patterson, D.P., Schwarz, B., El-Boubbou, K., van der Oost, J., Prevelige, P.E., and Douglas, T. (2012). Virus-like particle nanoreactors: programmed encapsulation of the thermostable CelB glycosidase inside the P22 capsid. Soft Matter 8, 10158–10166. 126. O’Neil, A., Prevelige, P.E., and Douglas, T. (2013). Stabilizing Viral Nano-reactors for Nerve-agent Degradation. Biomaterials Science 1, 881-886. 127. Glasgow, J.E., Capehart, S.L., Francis, M.B., and Tullman-Ercek, D. (2012). Osmolyte-Me- diated Encapsulation of Proteins inside MS2 Viral Capsids. ACS Nano 6, 8658–8664. 128. Wörsdörfer, B., Pianowski, Z., and Hilvert, D. (2012). Efficient In Vitro Encapsulation of Protein Cargo by an Engineered Protein Container. J. Am. Chem. Soc. 134, 909–911. 129. Aljabali, A.A.A., Barclay, J.E., Steinmetz, N.F., Lomonossoff, G.P., and Evans, D.J. (2012). Controlled Immobilisation of Active Enzymes on the Cowpea Mosaic Virus Capsid. Na- noscale 4, 5640-5645. 130. Carette, N., Engelkamp, H., Akpa, E., Pierre, S.J., Cameron, N.R., Christianen, P.C.M., Maan, J.C., Thies, J.C., Weberskirch, R., Rowan, A.E., Nolte, R.J.M., Michon, T., van Hest, J.C.M. (2007). A Virus-based Biocatalyst. Nat. Nanotechnol. 2, 226–229. 131. Pille, J., Cardinale, D., Carette, N., Di Primo, C., Besong-Ndika, J., Walter, J., Lecoq, H., van Eldijk, M.B., Smits, F.C.M., Schoffelen, S., van Hest, J.C.M., Mäkinen, K., Michon, T. (2013). General Strategy for Ordered Noncovalent Protein Assembly on Well-Defined Na- noscaffolds. Biomacromolecules 14, 4351–4359. 132. Koshiyama, T., Yokoi, N., Ueno, T., Kanamaru, S., Nagano, S., Shiro, Y., Arisaka, F., and Watanabe, Y. (2008). Molecular Design of Heteroprotein Assemblies Providing a Bionano- cup as a Chemical Reactor. Small 4, 50–54. 133. Miller, R.A., Presley, A.D., and Francis, M.B. (2007) Self-Assembling Light-Harvesting Systems from Synthetically Modified Tobacco Mosaic Virus Coat Proteins. J. Am. Chem. Soc. 129, 3104–3109. 134. Miller, R.A., Stephanopoulos, N., McFarland, J.M., Rosko, A.S., Geissler, P.L., and Francis, M.B. (2010). Impact of Assembly State on the Defect Tolerance of TMV-Based Light Har- vesting Arrays. J. Am. Chem. Soc. 132, 6068–6074. 135. Endo, M., Fujitsuka, M., and Majima, T. (2007). Porphyrin Light-Harvesting Arrays Con- structed in the Recombinant Tobacco Mosaic Virus Scaffold. Chem.-Eur. J. 13, 8660–8666. 136. Nam, Y.S., Magyar, A.P., Lee, D., Kim, J., Yun, D.S., Park, H., Pollom, T.S., Weitz, D.A., and Belcher, A.M. (2010). Biologically Templated Photocatalytic Nanostructures for Sustained Light-driven Water Oxidation. Nat. Nano. 5, 340–344. 137. Chen, P., Dang, X., Klug, M.T., Qi, J., Courchesne, N.M., Burpo, F.J., Fang, N., Hammond, P.T., and Belcher, A.M. (2013). Versatile Three-Dimensional Virus-Based Template for Dye- Sensitized Solar Cells with Improved Electron Transport and Light Harvesting. ACS Nano 7, 6563–6574. 138. Capehart, S.L., Coyle, M.P., Glasgow, J.E., and Francis, M.B. (2013). Controlled Integration

19 of Gold Nanoparticles and Organic Fluorophores Using Synthetically Modified MS2 Viral Capsids. J. Am. Chem. Soc. 135, 3011–3016. 139. Uchida, M., Kang, S., Reichhardt, C., Harlen, K., and Douglas, T. (2010). The Ferritin Super- family: Supramolecular Templates for Materials Synthesis. Biochim. Biophys. Acta 1800, 834–845. 140. Worsdorfer, B., Woycechowsky, K.J., and Hilvert, D. (2011). Directed Evolution of a Protein Container. Science 331, 589–592. 141. Rome, L.H., and Kickhoefer, V.A. (2013). Development of the Vault Particle as a Platform Technology. ACS Nano 7, 889–902. 142. King, N.P., Sheffler, W., Sawaya, M.R., Vollmar, B.S., Sumida, J.P., Andre, I., Gonen, T., Yeates, T.O., and Baker, D. (2012). Computational Design of Self-Assembling Protein Nano- materials with Atomic Level Accuracy. Science 336, 1171–1174. 143. Agapakis, C.M., Boyle, P.M., and Silver, P.A. (2012). Natural strategies for the spatial opti- mization of metabolism in synthetic biology. Nat. Chem. Biol. 8, 527–535.

20 Chapter 2

Osmolyte-Mediated Encapsulation of Proteins inside MS2 Viral Capsids

2.1 Abstract The encapsulation of enzymes in nanometer-sized compartments has the potential to enhance and control enzymatic activity, both in vivo and in vitro. Despite this potential, there are little quantitative data on the effect of encapsulation in a well-defined compartment under varying conditions. To gain more insight into these effects, we have characterized two improved methods for the encapsulation of heterologous molecules inside bacteriophage MS2 viral capsids. First, attaching DNA oligomers to a molecule of interest and incubating it with MS2 coat protein dimers yielded reassembled capsids that packaged the tagged molecules. The addition of a protein stabilizing osmolyte, trimethylamine N-oxide (TMAO), significantly increased the yields of reassembly. Second, we found that expressed proteins with genetically encoded negatively charged peptide tags could also induce capsid reassembly, resulting in high yields of reassembled capsids containing the protein. This second method was used to encapsulate alkaline phosphatase tagged with a 16 amino acid peptide. The purified encapsulated enzyme was found to have the same Km value and a slightly lower kcat value than the free enzyme, indicating that this method of encapsulation had a minimal effect on enzyme kinetics. This method provides a practical and potentially scalable way of studying the complex effects of encapsulating enzymes in protein-based compartments.

Reproduced with permission from Glasgow, Jeff E., Stacy L. Capehart, Matthew B. Francis, and Danielle Tullman-Ercek. Osmolyte-mediated encapsulation of proteins inside MS2 viral capsids. ACS Nano 6, 2012, 8658-8664. Copyright 2012 American Chemical Society.

21 2.2 Introduction Recent research has shown that the A) confinement of biochemical reactions within 1. 66% AcOH nanometer-sized compartments can have a 0°C, 30 min profound effect on enzymatic reaction rate and selectivity.1,2 Many types of cells from all 2. TMAO kingdoms of life are known to compartmen- talize enzymes to take advantage of local environment effects, channeling, and substrate Figure 1. Bacteriophage MS2 reassem- B) control.3,4 Notable examples thought to capi- bly process. A) MS2 capsids are disas- T T A A talize on these advantages include bacterial sembled into dimers with acetic acid and centrifuged to remove RNA (ref. 25). G C microcompartments, such as the carbon-fix- Reassembly is initiated with DNA, DNA AG C ing carboxysome.5 Due to their common oc- conjugated molecules, or highly nega- G C currence, there is a growing interest to mimic tively charged molecues in the presence T A of trimethylamine N-oxide (TMAO). A T such systems, both in vitro and in vivo, us- Shown is a model of E. coli alkaline 6,7,8 C G ing enzymes encapsulated in viral capsids. phosphatase (green) with a negatively A T Encapsulating enzymes in such protein charged peptide tag (red). B) DNA se- A quence TR in predicted hairpin (ref. 28). 3’ compartments and studying the effects on MS2 PDB 1ZDQ, alkaline phosphatase 5’ catalyzed reactions could advance our under- modeled from PDB 1ED8. standing of the advantages brought about by “TR-DNA” reaction space confinement, or could be used to alter substrate selectivity through selective diffusion through the shell.9,10 Recent approaches to the encapsulation of enzymes inside viral capsids have taken advantage of the reversible assembly of cowpea chlorotic mottle virus11,12,13 and hepatitis B capsids,14 triggered through changes in pH or salt concentration. This allows enzymes to be trapped inside. Other approaches have relied on specific interactions of fusion proteins with the bacteriophage P22 capsid15,16 and Simian Virus 40,17 or of nucleic acids with the interior surfaces of bacteriophage Qβ.18 The latter case provides a particularly elegant example, as the nucleic acid strands also pos- sessed aptamer domains that bound to co-expressed proteins and packaged them in vivo. Non-viral protein shells have also been used to encapsulate proteins through electrostatic interactions both in vivo19 and in vitro.20 Bacteriophage MS2 capsid has also been used extensively for virus-like particle applications, such as drug delivery,21,22,23 MRI contrast enhancement,24,25,26 and protein and nanoparticle encapsulation.27,28 Many of these applications have taken advantage of the presence of multiple ~2 nm pores, which allow small molecules to access the interior volume of the capsids, but are too small to allow folded proteins to pass through. This feature, in addition to unusually high stability of the MS2 capsid toward heat and denaturants, its tolerance of mutations, and its extensive structural characterization make it a particularly compelling protein shell for enzyme encapsulation.29 Previous efforts to encapsulate heterologous molecules in the MS2 capsid have similarly taken advantage of a specific interaction of a short RNA oligonucleotide30 with the interior surface of the capsid coat protein.27 In this approach, RNA is attached to the molecule of interest, which is then mixed with capsid coat protein dimers to initiate assembly around the molecule to be encapsulated. This method has been used to encapsulate several kinds of molecules, including the ricin A chain, fluorescent quantum dots, doxorubicin, and anti-cyclin siRNAs;28 however, the cost and instability 22 of the RNA used limits the practicality of encapsulation on larger synthetic scales. In this work, a convenient new osmolyte-based method is reported for cargo encapsulation using MS2 viral capsids. Osmolytes are a large class of small, neutral organic molecules often used by cells to regulate osmotic pressure,31 with common examples including methylamines, polyols, and amino acids. These molecules can have a profound effect on protein stability and solubility; for example, urea is well known to solubilize proteins while stabilizing the unfolded state.32 Tri- methylamine N-oxide (TMAO) is a common osmolyte often found in urea-rich organisms, which counteracts urea’s effects.33,34 As a result, the presence of TMAO can decrease unfolding while increasing the thermal stability of a wide variety of proteins.35 Herein we demonstrate that the attachment of DNA to a molecule of interest and incubation with MS2 coat protein (MS2-CP) dimers initiates reassembly around the molecule, similar to Ash- ley et al.,28 but with a significantly more convenient packaging signal. The replacement of RNA with DNA greatly reduces cost while increasing nucleic acid stability, making large scale encapsulation more feasible. Second, we show that addition of a genetically encoded poly(anionic) tag to a protein of interest also initiates reassembly and allows encapsulation. This method is advantageous as the target protein requires no further in vitro modification and can be fully characterized before encapsulation. This latter method was used to encapsulate derivatives of green fluorescent protein and E. coli alkaline phosphatase. The encapsulated alkaline phosphatase retained its activity, with kinetic parameters nearly equal to those of the free enzyme. All these processes are possible through the addition of a stabilizing osmolyte, trimethylamine N-oxide (TMAO).

2.3 Reassembly of Wild-type MS2 To develop a simple method of heterologous molecule encapsulation inside the MS2 capsid, we first attempted to initiate assembly of coat protein dimers with different oligonucleotides using similar conditions to those previously reported27 (Figure 1A). Extended incubation of MS2- CP dimers (15 µM, based on monomer) with yeast tRNA or the 20 nucleotide DNA sequence corresponding to the MS2 translational repressor sequence30 (“TR-DNA,” Figure 1B) at various concentrations resulted in visible aggregation of the coat protein and little to no capsid formation, as measured by size exclusion chromatography (SEC). To stabilize the coat proteins and suppress aggregation, we repeated the experiments in the presence of various osmolytes, such as glycine, arginine, proline, urea, guanidinium chloride, and TMAO. The resulting protein was analyzed by SEC (Figure 2). As shown in Figure 3A, increasing concentrations of TMAO in the presence of 50 μM TR-DNA led to increased yields of intact capsid. Similar results were obtained when TR-DNA was replaced with yeast tRNA (Figure 2A). Dynamic Light Scattering (DLS) of the reassembled capsid indicated a particle diameter of 27 nm, and transmission electron microscopy (TEM) images showed spherical particles matching untreated capsids (Figure 4). Interestingly, high concentrations of the osmolyte alone could also induce capsid reassembly, although the yield was greatly increased with the addition of TR-DNA. At even higher concentrations of TMAO, the yield of reassembled capsid decreased, probably due to a “salting out” effect (TMAO is known to decrease protein solubility at high concentrations36). Though the RNA translational repressor sequence has been shown to trigger capsid reassembly specifically,37 we found that in the presence of TMAO, a random DNA sequence could also induce reassembly (Figure 2A). At higher TR-DNA concentrations, the reassembly yields were reduced. A similar effect has been observed in TR-RNA-induced reassembly experiments, where binding kinetically traps a significant pool of the capsid dimers in a non-assembling conformation.38 For comparison, 23 A) DNA/0.25 M TMAO P(AA)/0.25 M TMAO MS2 Dimer MS2 Capsid e c

n 0.1 mg/ml tRNA/TMAO MS2 CP e c s Wild type MS2 e r o u l F

0 2 4 6 8 10 12 14 Time (min) B) DNA/0.25 M Arginine DNA/0.5 M Proline DNA/1 M Glycine e c n e c s e r o u l F

0 2 4 6 8 10 12 14 Time (min) Figure 2. Osmolyte-mediated reassembly. A) SEC traces of intact wild-type MS2 capsids, disassembled MS2 dimers, and capsids reassembled with TR-DNA, poly (acrylic acid) [P(AA)], and tRNA from Baker’s Yeast (tRNA) in the presence of TMAO. B) SEC-HPLC traces are shown for MS2 capsids assembled in the presence of other osmolytes and TR- DNA. Intact capsids eluted at 8.2 min, while MS2-CP dimers eluted around 10 min.

TR-RNA was also used to initiate reassembly. Also shown in Figure 3A, the TR-RNA induced maximum reassembly at low micromolar concentrations, followed by a decrease in yield at higher TR-RNA levels. The difference between optimal TR-RNA and TR-DNA concentrations can be attributed to specific interactions between the coat protein and 2’-OH groups on TR-RNA, giving the RNA higher affinity.39 Previous studies have shown that negatively charged polymers can initiate viral capsid assembly and even induce the formation of previously uncharacterized structures.40,41 There is evidence that other capsids from the Leviviridae family can assemble in this fashion, but little is known about the interaction of MS2-CP with anionic polymers.42 To test whether a negatively charged polymer could initiate MS2 reassembly, disassembled coat protein was incubated with varying amounts of 1.8k poly(acrylic acid) and 0.25 M TMAO at pH 7.2. This polymer has approximately 20 acidic monomers and, at this pH, should have a significant negative charge. Similar to the TR-DNA, the poly(acrylic acid) was also found to induce significant amounts of reassembly in the presence of TMAO (Figure 3B). Presumably, the electrostatic attraction between the nega-

24 tively charged molecules and the positively A) charged interior surface of the capsid initiated 80 reassembly with the negative charge inside. 50 μM TR-DNA At the concentrations used, the inhibitory ef- 70 No DNA fect seen with a large excess of TR-DNA was 60 y l not observed. b m

e 50 Based on these experiments, we s s a emerged with three options for the encapsula- e

R 40

t tion of proteins inside the MS2 capsid: 1) the n e c covalent attachment of a negatively charged r 30 e polymer, such as DNA, RNA, or poly(acrylic P 20 acid), to a cargo group and using the conjugate to initiate reassembly; 2) the genetic ad- 10 dition of a negatively charged amino acid sequence to the protein of interest and using the 0 0 0.2 0.4 0.6 0.8 1 1.2 purified protein to initiate reassembly; or 3) [TMAO] (M) the use of naturally negatively charged pro- B) teins to reassemble the capsid. 100 90

2.4 Encapsulation of GFP 80 y l Before encapsulating enzymes, we b 70 m e first attempted to encapsulate GFP as an eas- s

s 60 a ily detectable model protein. The monomeric e R 50

43,44 t

form of enhanced GFP (mEGFP) has a n e

c 40 charge of approximately -7 at pH 7.2. Incuba- r

e TR-RNA tion of free mEGFP with disassembled capsid P 30 TR-DNA resulted in reassembly of less than 10% of the 20 Poly(acrylic acid) capsid dimers. Based on a comparison of the No Polymer fluorescence of the MS2 tryptophans and the 10 GFP fluorophore, there were approximately 0 five GFP molecules per capsid. 0 20 40 60 80 100 120 To attach DNA to GFP we used an [Polymer] (μM) oxidative coupling strategy developed in our Figure 3. Encapsulation based on negative charge. a) Re- 44 assembly occurs with increasing concentrations of TMAO group (Figures 5 and 6). TR-DNA with a in the presence and absence of TR-DNA. b) Reassembly is 5’ amino group was incubated with N,N-di- enhanced with increasing concentrations of TR-DNA and ethyl-N’-acylphenylene-diamine NHS ester poly(acrylic acid) 1.8k in 250 mM TMAO. 1 to yield TR-DNA containing a phenylene- diamine moiety (2). GFP was incubated with isatoic anhydride to install an aniline derivative on surface lysine residues. The phenylene-diamine containing TR-DNA was then incubated with the aniline GFP in the presence of 5 mM NaIO4 for 1 h to yield GFP-DNA conjugate 3 at 35% yield by densitometry (Figure 5). After purification by anion exchange FPLC, the conjugate was analyzed by SDS-PAGE (Figure 6B), confirming a shift in mass of about 6 kDa, corresponding to the TR- DNA attachment. The GFP-DNA conjugate was next incubated with disassembled capsids in the presence of 0.25 M TMAO, as described above. After incubation the free GFP was purified away using 100 25 25 A B Wild Type E DNA t 20 n Poly(acrylate) e c

r PhoA-neg e 15 P

3xFLAG GFP e m u

l 10 o V 5

0 0.1 1 10 100 1000 10000 Diameter (nm)

C D Reassembly with: D (nm)

DNA 26.7

Poly(acrylate) 26.6

PhoA-neg 25.4

3xFLAG GFP 29.1

Wild Type 27.2

Figure 4. Structural characterization of reassembled MS2 capsids. TEM images of capsids reassembled with A) TR- DNA, B) poly(acrylic acid), C) PhoA-neg, and D) 3xFLAG GFP. E) DLS traces of purified capsids in 20 mM Tris-HCl,

pH 7.2, and table of average peak diameters. TEM images were negatively stained with UO2(OAc)2. kDa spin concentrators. The GFP-DNA conjugate was able to initiate, on average, 35% (±6%, n = 3) of the MS2-CP dimers to assemble. As shown in Figure 6C, a significant amount of GFP fluorescence co-eluted with the MS2 capsid. No such peak was seen when GFP and MS2 were co-injected, implying that the GFP-DNA was trapped inside the capsid when it was used to initiate reassembly. Based on the fluorescence of GFP eluting with the capsid, there was an average of 6.5±1.9 GFP molecules per capsid. This corresponds roughly to a concentration of 2 mM GFP within the capsid volume. To test if the addition of a negatively charged amino acid tag would enable the encapsulation of mEGFP, we next added a 3xFLAG tag to the N-terminus. As shown in Figure 7A, increasing amounts of 3xFLAG mEGFP corresponded to higher yields of capsid reassembly, whereas mEGFP gave low reassembly yield at all concentrations. Incorporation of mEGFP fluorescence into the capsid was confirmed by SEC-HPLC. To support the observation that negative charge was responsible for capsid reassembly, disassembled capsids were incubated with lysozyme (a well known basic protein) and poly(ethylene) glycol 8k (a neutral polymer). Neither of these experiments resulted in significant amounts of reassembled capsid (Figure 8). We then increased the negative charge by adding a C-terminal acidic peptide tag, inserting arbitrarily chosen codons for aspartate and glutamate residues (EEEEDDDEDDDDEEDD). An N-terminal 6xHis tag was also added for purification purposes. Reassembly of MS2-CP dimers with this construct, referred to as His-GFP-neg, is also shown in Figure 7A. Yield of intact capsid increased significantly using this construct, presumably due to the higher negative charge of the neg tag over the 3xFLAG tag (16 vs. 9 total charges added). With an effective encapsulation strategy developed, we then proceeded to enzyme encapsulation.

26 900.00 120.00 A B 800.00 A280 GFP-DNA 100.00 700.00 Fractions 600.00 Percent B TR-DNA 80.00 B

500.00 t 0

60.00 n 8 e

2 400.00 GFP c r A GFP+DNA 300.00 40.00 e P GFP 200.00 20.00 100.00 0.00 1 2 3 4 5 6 0.00 -100.00 -20.00 5.00 15.00 25.00 35.00 45.00 mL Figure 5. Analysis and purification of GFP-DNA conjugation.A) Gel of TR-DNA GFP conjugation. Lanes: 1) Protein Ladder 2) Untreated mEGFP 3) Aniline treated mEGFP, unpurified 4) Aniline mEGFP, purified 5) Aniline mEGFP treated

with NaIO4; 6) Aniline mEGFP+ phenylene-diamine TR-DNA + NaIO4. B) FPLC trace of the purification of TR-DNA conjugated mEGFP. A) - GFP DNA O O O NH O 2.5 Encapsulation and Activity of Alkaline NH2 P + O N 4 NH O H 2 Phosphatase O E. coli alkaline phosphatase (PhoA) exists as a homodimer of a 49 kDa protein. 1 2 N After in vivo secretory processing,45 each O monomer is predicted to have a charge of DNA N O 5 mM NaIO O O O approximately -8.6 at pH 7.2, giving the 4 P - PB pH 6.5 N N 4 O O holoenzyme an approximate charge of -17. H H 1M NaCl N Incubation of MS2 coat protein dimers H N with free wild type enzyme in presence of GFP O TMAO resulted in low (<10%) yield of re- B) C) 3 assembled capsid. To increase the yield of Trp Fl.

GFP Fl. MS2-encapsulated PhoA, a large acidic pep- e GFP-DNA n c e c

tide (EEEEDDDEDDDDEEDD) was added s e r o

to the C-terminus (PhoA-neg). As shown in u l Figure 7A, increasing amounts of PhoA-neg 3 F incubated with disassembled capsid resulted 1 in increasing reassembly in a concentration- 0 5 10 15 Time (min) dependent fashion. At higher PhoA-neg A B concentrations, PhoA-neg inhibited capsid Figure 6. DNA-Protein bioconjugation for encapsula- reassembly, possibly due to an undesired as- tion experiments. a) mEGFP (40 µM, 10 mM PB pH 8.0) sociation between the large, dimeric enzyme was incubated with isatoic anhydride (1mM) for one hour and the positively charged MS2-CP surface, at room temperature to yield aniline-GFP 1. Phenylene diamine DNA 2 (ref. 41) was incubated with 1 in 25 mM PB blocking assembly. The reassembled capsid pH 6.5, with 5mM NaIO4 for 1 hour to yield 3. b) The DNA was precipitated using a solution of 10% PEG bioconjugation reaction was followed using SDS-PAGE 8k and 500 mM NaCl, and the obtained mate- (Coomassie Stain). Lane A: Unreacted mEGFP; Lane B: GFP-DNA conjugate 3. c) HPLC trace of MS2 with encap- rial was resuspended in ST buffer and purified sulated (green) and free (red) GFP-DNA. by size exclusion HPLC. Fractions containing the intact capsids were collected and found 27 to possess significant alkaline phosphatase A) activity, whereas corresponding fractions 80 PhoA-neg collected when preincubated mixtures of 70 GFP-neg untreated, intact capsids and alkaline phos- 3xFLAG GFP mEGFP phatase were injected contained no activity y 60 l

b WT PhoA

(Figure 9). A sample of the isolated capsids m

e 50 s was analyzed using SDS-PAGE. Densitom- s a e

etry analysis after Coomassie staining indi- R

40 t cated an average of 1.6 PhoA-neg dimers per n e c

r 30 capsid, corresponding to an effective enzyme e P concentration of 0.5 mM. The lower incor- 20 poration of PhoA-neg molecules per capsid compared to GFP is probably due to the en- 10 zyme’s size and dimerization. 0 The kinetics of the encapsulated 0 0.2 0.4 0.6 0.8 1 PhoA-neg were assayed by monitoring the B) [Initiator Protein] (mg/ml) hydrolysis of 4-methylumbelliferyl phos- 1 phate to yield fluorescent 4-methylumbelliferone in 100 mM 3-(N-morpholino)propane- 0.9 sulfonic acid (MOPS) buffer with 500 mM 0.8 NaCl. As shown in Figure 7B, the encap- 0.7 sulated enzyme dimer followed Michaelis- 0.6 Menten kinetics with a Km equal to that of 0.5 the free enzyme dimer. The kcat was slightly reduced when the enzyme was encapsulated, 0.4 (μM/sec/μg enzyme)

possibly due to the constrained enzyme en- t

a 0.3 vironment. Alkaline phosphatase is known to R PhoA-neg 46 0.2 be a “nearly perfect” enzyme, and therefore MS2/PhoA-neg is more susceptible to inhibited diffusion, but 0.1 PhoA-neg fit this did not appear to be the case here. Previ- MS2/PhoA-neg fit 13 0 ous studies have shown both enhanced and 100 300 500 18 inhibited kinetics for viral capsid based- [4-MUP] (μM) enzyme nanoreactor systems, suggesting that complex influences are operational in these Kinetic Parameters -1 -1 -1 systems. Other encapsulation systems based Km (μM) kcat (s ) kcat/Km (M s ) on liposome- or polymersome-encapsulated 6 enzymes, have shown similar variations in Free PhoA-neg 14.3 17.9 1.25x10 47,48 kinetic effects. These systems are often 5 Encapsulated 14.5 14.0 9.6x10 complicated by permeability issues and un- desirable enzyme conditions during syn- Figure 7. Negatively charged protein encapsulation. a) thesis. Because of the synthetic ease of this Reassembly yield with increasing concentrations of PhoA- system, mild encapsulation conditions used, neg (n = 2), 3xFLAG GFP (n = 2), mEGFP (n = 1), and wild-type PhoA (n = 1) in 0.25 M TMAO. b) Rate profiles of and modest effect on enzyme activity, MS2 free (red) vs. MS2 encapsulated (blue) alkaline phosphatase encapsulated enzymes have the potential to dimer. The lines show the result of fitting the data to the elucidate the rich kinetic effects associated Michaelis-Menten equation.

28 1 uM 2 uM MS2 CP 5 uM Lysozyme 10 uM

e 20 uM c

n 50 uM

e MS2 Capsid c

s 0 uM e r o u l F

0 2 4 6 8 10 12 14 Time (min) Figure 8. Reassembly attempts with increasing concentrations of lysozyme. The concentration of intact capsid is near background reassembly or less for every experiment. with confined enzyme compartments.

2.6 Summary In conclusion, we have demonstrated two improved methods for encapsulating proteins inside MS2 capsids using a protein stabilizing osmolyte to increase yields. In the first method, the capsid is reassembled around a protein of interest by means of a conjugated negatively charged polymer. DNA and poly(acrylic) acid are cheaper and more stable than RNA and offer a much more practical trigger for reassembly. The use of these polymers is not limited to protein encapsulation; any molecule larger than the MS2 capsid pores can be attached to a polymer via a wide variety of different conjugation reactions and used to trigger assembly. Our second method uses genetically encoded, negatively charged amino acid tags to specifically incorporate purified proteins into the capsid. This method allows facile encapsulation of enzymes into an easily modified protein shell, which will aid in studies on the precise effect that encapsulation has on enzyme activity, stability, and specificity. Furthermore, by altering the pore characteristics of the capsids, it may be possible to confer additional substrate selectivity or restrict the diffusion of inhibitory molecules. In this way, we intend to use this versatile encapsulation system to study electrostatic and steric effects on substrate diffusion into the capsid nanoreactor.

29 A)

e 1200 0.001598 c ) n Fluorescence 0.001398 c e

1000 e c s / s Activity 0.001198 5 e

r 800 0.000998 0 o MS2 Capsid 4 u l 600 0.000798 A (

F PhoA-neg

y n 0.000598 t i 400 i v e i t

0.000398 t o c r 200 A

P 0.000198 0 -2E-06 0 2 4 6 8 10 12 14

B) e 1000 0.000248 c )

n Fluorescence c e e

c 800 0.000198 Activity s / s 5 e r 0

o 600 0.000148 4 u l A ( F

400 0.000098 y n t i i v e i t t o

200 0.000048 c r A P 0 -2E-06 0 2 4 6 8 10 12 14

C) 1200 0.00198 e ) c Fluorescence c n 1000 e e Activity s c

0.00148 / s 5

e 800 0 r 4 o u A

l 600 0.00098 (

y t i n

i 400 v i e t t 0.00048 c o

r 200 A P 0 -0.00002 0 2 4 6 8 10 12 14

D) e 300 0.01998 c )

n Fluorescence c

e 250 e c Activity s 0.01498 / s 5 e 200 r 0 o 4 u l 150 0.00998 A ( F

y n t i 100 i v e i t 0.00498 t o c r 50 A P 0 -0.00002 0 2 4 6 8 10 12 14 Time (min) Figure 9. Purification of MS2-PhoA-neg.A) Untreated MS2 and PhoA-neg were coinjected onto a Biosep SEC S4000 column. B) Untreated MS2 and PhoA-neg were precipitated with 10% PEG 8k, resuspended and then injected. C) Un- treated MS2 and PhoA-neg were precipitated, and supernatant was injected. D) MS2 was reassembled with PhoA-neg, precipitated, resuspended, and then injected. 2.7 Materials and Methods General Procedures

Unless otherwise noted, all chemical reagents were purchased from Aldrich and used without further purification. Water (ddH2O) used in biological procedures or as a reaction solvent was de-ionized using a MilliQ academic system (Millipore). High Pressure Liquid Chromatography

30 (HPLC) columns were purchased from Phenomenex, Inc. (www.phenomenex.com) and used as specified, unless otherwise noted.

Instrumentation and Sample Analysis Preparations

High Perfomance Liquid Chromatography. HPLC was performed on an Agilent 1100 Series HPLC System (Agilent Technologies, USA). Size exclusion chromatography was accomplished on a Phenomenex PolySep-GFC-P 5000 (PS5K) column (300 x 7.8 mm, flow rate 1.0 mL/min) or a BioSep SEC-S-4000 (300 x 7.8 mm, flow rate 1.0 mL/min) column using an aqueous mobile phase (50 mM Tris-HCl 100 mM NaCl, pH 7.2) . Sample analysis for all HPLC experiments was achieved with an inline diode array detector (DAD) and an inline fluorescence detector (FLD).

Gel Analyses. Sodium dodecyl sulfate-poly(acrylamide) gel electrophoresis (SDS-PAGE) was accomplished on a Mini-Protean apparatus from Bio-Rad (Hercules, CA) 10% or 12.5% poly(acrylamide), following the protocol of Laemmli.49 All electrophoresis protein samples were mixed with SDS loading buffer in the presence of dithiothreitol (DTT) and heated to 95 °C for 5 min to ensure reduction of disulfide bonds and complete denaturation unless otherwise noted. Commercially available molecular EZ Run mass markers from Fisher Scientific (Hampton, NH) were applied to at least one lane of each gel for calculation of the apparent molecular masses. Gels were imaged using a Bio-Rad ChemiDoc XRS+ system (Bio-Rad).

Centrifugations. Centrifugations were conducted with an Avanti J30I Tabletop Centrifuge (Beck- man Coulter, Inc., USA). General desalting and removal of other small molecules of biological samples were achieved using NAP-10 gel filtration columns (GE Healthcare). Protein samples were concentrated by way of centrifugal ultrafiltration using Amicon Ultra-4 or Ultra-15 100 kDa molecular weight cut off (MWCO) centrifugal filter units (Millipore), or Amicon Microcon 10 kDa, 30 kDa, and 100 kDa MWCO (Millipore) centrifugal filter units. Dialysis was performed using Slide-A-Lyzer Dialysis Cassettes (Pierce/Thermo Scientific, Rockford, lL). Protein concentrations were determined using the BCA Assay (Pierce/Thermo Scientific).

Fast Protein Liquid Chromatography. FPLC was performed on an Akta Purifier (GE Healthcare). Anion exchange was performed using a HiPrep DEAE column (GE Healthcare).

Dynamic Light Scattering. DLS measurements were performed on a Zetasizer Nano ZS (Malvern Instruments, UK). Samples were taken in 20 mM Tris buffer pH 7.2 at 24 ºC. Data points are calculated from an average of three measurements, each of which consists of 10 runs of 45 seconds each.

Transmission Electron Microscopy (TEM). TEM images were obtained at the University of Cali- fornia – Berkeley Electron Microscope Lab using a FEI Tecnai 12 transmission electron microscope with 120 kV accelerating voltage. Samples were prepared by pipetting 5 µL onto formvar- coated copper mesh grids (400 mesh, Ted Pella, USA) for 5 minutes, followed by rinsing with 8 µL dd-H2O. The grids were then exposed to 8 µL of a solution of uranyl acetate (15 mg/mL in dd-H2O) for 90 seconds as a negative stain. Excess stain was then removed and the grids were allowed to dry in air.

31 mEGFP and 3xFLAG mEGFP expression. The monomeric form of enhanced green fluorescent protein (mEGFP, eGFP A206K) was expressed and purified as previously described.43,44 To add a 3xFLAG tag to mEGFP, the mEGFP gene was amplified from mEGFP-V2G-pTXB1 with the concomitant addition of a 5’ BamHI site and a 3’ HindIII site (forward primer 5’- AAATGTA- AGCTTTTATTACTTGTACAGCTCGTCCATGCC-3’, reverse primer 5’- AAATGTGGATC- CGGCGAGGAGCTGTTCACC-3’) and inserted into a pTrc99a vector using standard cloning techniques. The 3xFLAG tag was then assembled by PCR with a 5’ NcoI site and 3’ BamHI site (forward primer 5’- ACACTGCCATGGGTGATTATAAAGATGACGATGACAAGGAT-TATA- AAGATGACGATGACA-3’, reverse primer 5’- ACACTGGGATCCCTTGTCATCGTCATCTT- TATAATCCTTGTCATCGTCATCTTTATAATC-3’). The 3xFLAG tag was inserted into mEGFP- pTrc99a and transformed into DH1 cells. Cells containing the plasmid were grown to OD600=0.5 and induced overnight with 1 mM isopropyl β-D-1-thiogalacto-pyranoside (IPTG). The cells were harvested by centrifugation and lysed by sonication. The bright green soluble fraction was then purified by the ethanol extraction method of Samarkina et al.50 followed by dialysis into bis-Tris pH 6.0. The protein was then applied to a DEAE column (HiPrep DEAE, GE Healthcare), and eluted with a 0-1 M NaCl gradient. Green fractions were pooled and stored at 4 oC.

His-GFP-neg expression and purification.The gene for mEGFP was amplified from mEGFP-V2G- pTXB1 with the addition of a 5’ NcoI restriction site and the sequence encoding a 6xHis tag, and a 3’ Neg tag followed by a 3’ XbaI site (forward primer 5’-ACATACCCATGGCGAAAACCCAT- CACCATCATCATCACGGCGAGGAGCTGTTCACC-3’, reverse primer 5’-ACATACTCTA- GATTATTAATCGTCTTCCTCGTCATCGTCATCTTCGTCATCATCCTCTTCTTCCTCCTT- GTACAGCTCGTCCATGCC-3’). The resulting gene was inserted into a pTrc99a vector and transformed into BL21 Codon+RIL cells. The cells were grown to an OD600 of 0.6 and induced overnight with 1 mM IPTG. The resulting green cells were harvested, resuspended in 50 mM phosphate/Tris buffer pH 7.5 with 300 mM NaCl and 10 mM imidazole, and lysed by sonication. The soluble fraction was applied to a His Gravitrap column (GE Healthcare), washed with the same buffer containing 75 mM imidazole, and eluted with the same buffer containing 300 mM imidazole. The bright green elution fractions were pooled and concentrated and stored at 4 oC for encapsulation assays.

PhoA and PhoA-neg Expression, Purification and Assay. The phoA gene was amplified from E. coli DH10B genomic DNA with a 5’ XbaI site and a 3’ HindIII site (forward primer 5’-ACACT- GTCTAGAGTGAAACAAAGCACTATTGCAC-3’, reverse primer 5’-CATTGTAAGCTTCGC- GGTTTTATTTCAGCC-3’) and inserted into pTrc99a using standard cloning techniques. The protein, which contains an N-terminal periplasmic localization tag, was expressed overnight in DH10B cells with 1 mM IPTG. The enzyme was purified by isolating the periplasm using lysozyme/EDTA in 50 mM Tris pH 8.0. The periplasmic fraction was precipitated with 50% saturated ammonium sulfate, leaving the PhoA in solution. The supernatant was extensively dialyzed against 20 mM Tris pH 8.0, applied to a DEAE column (HiPrep DEAE, GE Healthcare), and eluted with a 0-1 M NaCl gradient. Fractions containing PhoA were pooled and stored at 4 oC.

To add the negative charge tag to PhoA, the phoA gene was PCR amplified with primers, adding a C-terminal sequence of EEEEDDDEDDDDEEDD (forward primer 5’-ACACTGTC- TAGAGTGAAACAAAGCACTATTGCAC-3’, reverse primer 5’-ACACTGAAGCTTTTAATC-

32 GTCTTCCTCGTCATCGTCATCTTCGTCATCATCCTCTTCTTCCTCTTTCAGCCCCAGAG CG) and reinserted into pTrc99a. The enzyme was purified similarly to the wild type enzyme and was found to have similar activity.

Reassembly Experiments. MS2 coat protein dimers (15 μM) were mixed on ice with the desired concentration of reassembly initiator and osmolyte in 50 mM Tris 100 mM NaCl. The mixture was incubated at 4 oC for 36-48 h and assayed for assembly by HPLC using a Polysep GFC-P-5000 column (Phenomenex) or a Biosep SEC-S-4000 column (Phenomenex). Reassembly yield was quantified using the Trp fluorescence peak (ex. 280/em. 330) at 7.9-8.3 min (Polysep) or 5.9-6.3 min (Biosep), which corresponded to the assembled MS2 capsids. The concentration of reassembled capsids could be found by comparing the integrated peak area to that obtained using a standard curve prepared with known concentrations of authentic MS2 capsids.

Protein-DNA Oxidative Coupling. Isatoic anhydride (100 mM in DMSO) was added to a 1 mg/ mL solution of GFP in 20 mM phosphate buffer, pH 8.0, to a final concentration of 1 mM and incubated at room temperature for 1 h. The excess isatoic anhydride was then removed by passage over a NAP column (GE Healthcare) followed by multiple rounds of centrifugal concentration through a 10 kDa MWCO centrifugal filter (Millipore).

TR-DNA with a 5’ amino group (Integrated DNA Technologies) was incubated with phenylene-diamine NHS ester in 50% dimethyl formamide for 2 hours as previously described.44 The DMF and excess linker were extracted into 50:1 chloroform:acetic acid 3 times. The remaining phenylene diamine-TR-DNA was then further purified using multiple rounds of spin concentration against a 3 kDa MWCO centrifugal filter (Millipore).

Aniline GFP and phenylene diamine-TR-DNA were then mixed in 20 mM phosphate buffer, pH

6.5, and 1 M NaCl in a 1:15 molar ratio. Freshly prepared 50 mM NaIO4 in water was added to a final concentration of 5 mM and allowed to react for 1 h. The resulting DNA-GFP conjugate was then purified and buffer exchanged into 20 mM piperazine, pH 9.5, by multiple rounds of spin concentration against a 30 kDa MWCO centrifugal filter, followed by application to a 1 mL HiTrap Q XL strong anion exchange column. The conjugate was eluted using a linear gradient of 0-1 M NaCl. Fractions containing the GFP-DNA conjugate were pooled and concentrated, then stored at 4 oC.

GFP-DNA conjugate analysis and purification. TR-DNA and mEGFP were coupled as described in the main text and analyzed by SDS-PAGE. When the conjugation reaction was complete the mixture was quenched with 1 μl of ethylene glycol desalted with a 30kDa MWCO spin concentrator. The solution was then diluted into 20 mM 2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)-1,3- propanediol (bis-Tris) pH 6.0. The protein was applied to a HiPrep DEAE column and eluted with a 0-0.5M gradient of NaCl in 20 mM bis-Tris pH 6.0.

MS2 Capsid Expression, Purification & Disassembly.Wild type bacteriophage MS2 capsids were expressed and purified as previously described.25 Protein concentrations were determined using a BCA assay (Pierce). The capsids were disassembled using the method of Wu et al.27 Briefly, a 10 mg/mL solution of intact MS2 capsids in ST buffer (50 mM Tris, 100 mM NaCl) was mixed with

33 cold glacial acetic acid to give a final concentration of 66% acid. The mixture was incubated on ice for 30 min, then centrifuged at 16,000 x g for 20 min at 4 oC to remove any nucleic acid con- taminants. The supernatant was then desalted into 1 mM acetic acid using commercially available gel filtration columns (NAP-5, GE Healthcare). Fractions containing the disassembled capsid were maintained on ice and used on the same day.

Determination of Capsid Reassembly. The extent of capsid reassembly was determined by size exclusion chromatography (SEC) on either a Polysep 5k column (Capsids elute at 8.2 minutes @1ml/min flow rate) or Biosep 4k (Capsids elute at 6.2 minutes @1ml/min flow rate). Trypto- phan fluorescence (ex 280/em 330 nm) was tracked, and the area under the MS2 capsid peak was used to quantify concentration. Reassembled capsids were compared to intact capsid standards for absolute quantification. Reassembled capsids were also analyzed by TEM and DLS. To show the necessity of negative charge in reassembly, disassembled CP was incubated with lysozyme from chicken egg white type VI (MP Biomedicals, Solon, OH) at several concentrations. No reassembly above background was observed.

PhoA-neg encapsulation and purification. PhoA-neg (10 μM) was incubated for 36-48 h with disassembled MS2 (15 μM monomer concentration) in the presence of 0.25 M TMAO in ST buffer. The solution was then centrifuged at 10,000 x g for 10 min. To the supernatant was added PEG 8k to 10% w/v and NaCl to a concentration of 0.5 M. The solution was rotated on a benchtop rock- er at 4 oC for 2 h and centrifuged at 17,800 x g for 45 min. The resulting pellet was resuspended in a minimal amount of ST buffer and again spun at 10,000 x g for 10 min. The supernatant was then applied to a Biosep SEC-S-4000 using a flow rate of 1 mL/min ST buffer. Fractions corresponding to intact MS2 capsids were collected, concentrated, and stored at 4 oC. Fractions were collected (1 minute fractions at 1 mL/min) and assayed for hydrolysis of p-nitrophenyl phosphate in 1 M Tris-

HCl, 10 μM MgCl2, 1 μM ZnCl2. As controls, intact capsids were also incubated with PhoA-neg and purified and assayed in a similar manner. No detectable phosphatase activity was associated with the intact capsid fractions without using the disassembly/reassembly protocol, implying the PhoA-neg was not associating with the outside of the capsids and co-purifying on the column. A small amount of free PhoA-neg precipitated with the capsids, but was fully purified away by SEC.

Alkaline Phosphatase Activity Assays. Alkaline phosphatase activity was characterized in two ways. First, the liberation of p-nitrophenol from p-nitrophenylphosphate in 1 M Tris pH 8.0, 10 mM MgSO4, 10 mM ZnCl2 was followed using the absorbance at 405 nm. Second, the liberation of 4-methylumbelliferone from 4-methylumbelliferyl phosphate was followed in 0.1 M 3-(N-morpholino)propanesulfonic acid (MOPS), 500 mM NaCl (ex. 362/em 448 nm).

2.8 References

1. Kim, K. T., Meeuwissen, S. A., Nolte, R. J. M., van Hest, J. C. M. (2010). Smart Nanocontain- ers and Nanoreactors. Nanoscale 2, 844–858. 2. Vriezema, D., Comellas-Aragonès, M., Elemans, J., Cornelissen, J. J. L. M., Rowan, A. E., Nolte, R. J. M. (2005). Self-Assembled Nanoreactors. Chem. Rev. 105, 1445–1490. 3. Huang, X., Holden, H. M., Raushel, F. M. (2001). Channeling of Substrates and Intermediates in Enzyme-Catalyzed Reactions. Ann. Rev. Biochem. 70, 149-180. 34 4. Yeates, T., Kerfeld, C., Heinhorst, S., Cannon, G., Shively, J. (2008) Protein-Based Organelles in Bacteria: Carboxysomes and Related Microcompartments. Nature Reviews 6, 681-691. 5. Kinney, J. N., Axen, S. D., Kerfeld, C. A. (2011). Comparative Analysis of Carboxysome Shell Proteins. Photosynth. Res. 109, 21-32. 6. de la Escosura, A., Nolte, R. J. M., Cornelissen, J. J. L. M. (2009). Viruses and Protein Cages as Nanocontainers and Nanoreactors. J. Mat. Chem. 19, 2274–2278. 7. Conrado, R. J., Varner, J. D., DeLisa, M. P. (2008). Engineering the Spatial Organization of Metabolic Enzymes: Mimicking Nature’s Synergy. Curr. Opin. Biotech. 19, 492-499. 8. Aniagyei, S. E., DuFort, C., Kao, C. C., Dragnea, B. (2008). Self-Assembly Approaches to Nanomaterial Encapsulation in Viral Protein Cages. J. Mat. Chem. 18, 3763-3774. 9. Conrado, R. J., Mansell, T. J., Varner, J. D., DeLisa, M. P. (2007). Stochastic Reaction-Diffu- sion Simulation of Enzyme Compartmentalization Reveals Improved Catalytic Efficiency for a Synthetic Metabolic Pathway. Metab. Eng. 9, 355-363. 10. Renggli, K., Baumann, P., Langowska, K., Onaca, O., Bruns, N., Meier, W. (2011). Selective and Responsive Nanoreactors. Adv. Func. Mat. 21, 1241-1259. 11. Comellas-Aragonès, M., Engelkamp, H., Claessen, V. I., Sommerdijk, N. A. J. M., Rowan, A. E., Christianen, P. C. M., Maan, J. C., Verduin, B. J. M., Cornelissen, J. J. L. M., Nolte, R. J. M. (2007). A Virus-based Single-Enzyme Nanoreactor. Nat. Nanotechnol. 2, 635–639. 12. Minten, I. J., Wilke, K. D. M., Hendriks, L. J. A., van Hest, J. C. M., Nolte, R. J. M., Cornelis- sen, J. J. L. M. (2011). Metal-Ion-Induced Formation and Stabilization of Protein Cages Based on the Cowpea Chlorotic Mottle Virus. Small 7, 911-919. 13. Minten, I. J., Claessen, V. I., Blank, K., Rowan, A. E., Nolte, R. J. M., Cornelissen, J. J. L. M. (2011). Catalytic Capsids: The Art of Confinement. Chem. Sci. 2, 358-362. 14. Lee, K. W., Tan, W. S. (2008). Recombinant Hepatitis B Virus Core Particles: Association, Dissociation and Encapsidation of Green Fluorescent Protein. J. Virol. Meth. 151, 172-180. 15. O’Neil, A., Reichhardt, C., Johnson, B., Prevelige, P., Douglas, T. (2011). Genetically Pro- grammed In Vivo Packaging of Protein Cargo and Its Controlled Release from Bacteriophage P22. Angew. Chem. Int. Ed. 50, 7425-7428. 16. Patterson, D. P., Prevelige, P. E., Douglas, T. (2012). Nanoreactors by Programmed Enzyme Encapsulation Inside the Capsid of the Bacteriophage P22. ACS Nano 6, 5000–5009. 17. Inoue, T., Kawano, M., Takahashi, R., Tsukamoto, H., Enomoto, T., Imai, T., Kataoka, K., Handa, H. (2008). Engineering of SV40-based nano-capsules for delivery of heterologous proteins as fusions with the minor capsid proteins VP2/3. J. Biotechnol. 134, 181–192. 18. Fiedler, J. D., Brown, S. D., Lau, J. L., Finn, M.G. (2010). RNA-Directed Packaging of En- zymes within Virus-Like Particles. Angew. Chem. Int. Ed. 49, 9648–9651. 19. Wörsdörfer, B., Woycechowsky, K. J., Hilvert, D. (2011). Directed Evolution of a Protein Con- tainer. Science 331, 589-592. 20. Wörsdörfer, B., Pianowski, Z., Hilvert, D. (2012). Efficient In Vitro Encapsulation of Protein Cargo by an Engineered Protein Container. J. Am. Chem. Soc. 3, 9-11. 21. Wu, W., Hsiao, S. C., Carrico, Z. M., Francis, M. B. (2009). Genome-Free Viral Capsids as Multivalent Carriers for Taxol Delivery. Angew. Chem. Int. Ed. 48, 9493–9497. 22. Stephanopoulos, N., Carrico, Z., Francis, M. B. (2009). Nanoscale Integration of Sensitizing Chromophores and Porphyrins with Bacteriophage MS2. Angew. Chem. Int. Ed. 48, 9498 -9502. 23. Stephanopoulos, N., Tong, G. J., Hsiao, S. C., Francis, M. B. (2010). Dual-surface Modified

35 Virus Capsids for Targeted Delivery of Photodynamic Agents to Cancer Cells. ACS Nano 4, 6014-6020. 24. Anderson, E. A., Isaacman, S., Peabody, D. S., Wang, E. Y., Canary, J. W., Kirshenbaum, K. (2006). Viral Nanoparticles Donning a Paramagnetic Coat: Conjugation of MRI Contrast Agents to the MS2 Capsid. Nano Lett. 6, 1160-1164. 25. Meldrum, T., Seim, K. L., Bajaj, V. S., Palaniappan, K. K., Wu, W., Francis, M. B., Wemmer, D. E., Pines, A. A (2010). Xenon-based Molecular Sensor Assembled on an MS2 Viral Capsid Scaffold. J. Am. Chem. Soc. 132, 5936-5937. 26. Garimella, P. D., Datta, A., Romanini, D. W., Raymond, K. N., Francis, M. B. (2011). Mul- tivalent, High-Relaxivity MRI Contrast Agents Using Rigid Cysteine-Reactive Gadolinium Complexes. J. Am. Chem. Soc. 133, 14704-14709. 27. Wu, M., Brown, W., Stockley, P. (1995). Cell-specific Delivery of Bacteriophage-Encapsidated Ricin A Chain. Bioconjugate Chem. 6, 587-595. 28. Ashley, C. E., Carnes, E. C., Phillips, G. K., Durfee, P. N., Buley, M. D., Lino, C. A., Padilla, D. P., Phillips, B., Carter, M. B., Willman, C. L., et al. (2011). Cell-Specific Delivery of Di- verse Cargos by Bacteriophage MS2 Virus-Like Particles. ACS Nano 5, 5729-5745. 29. Stonehouse, N., Stockley, P. (1993). Effects of Amino Acid Substitution on the Thermal Stabil- ity of MS2 Capsids Lacking Genomic RNA. FEBS Letters 334, 355-359. 30. Beckett, D., Uhlenbeck, O. C. (1988). Ribonucleoprotein Complexes of R17 Coat Protein and a Translational Operator Analog. J. Mol. Biol. 204, 927–938. 31. Kumar, R. (2009). Role of Naturally Occurring Osmolytes in Protein Folding and Stability. Arch. Biochem. Biophys. 491, 1-6. 32. Bolen, D. W. (2004). Effects of Naturally Occurring Osmolytes on Protein Stability and Solu- bility: Issues Important in Protein Crystallization. Methods 34, 312-322. 33. Yancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., Somero, G. N. (1982). Living With Water Stress: Evolution of Osmolyte Systems. Science 217, 1214-1222. 34. Wang, A., Bolen, D. A (1997). Naturally Occurring Protective System in Urea-Rich Cells: Mechanism of Osmolyte Protection of Proteins Against Urea Denaturation. Biochemistry 36, 9101-9108. 35. Street, T. O., Bolen, D. W., Rose, G. D. (2006). A Molecular Mechanism for Osmolyte-Induced Protein Stability. Proc. Nat. Acad. Sci. U.S.A. 103, 13997-14002. 36. Zhang, Y., Cremer, P. S. (2010). Chemistry of Hofmeister Anions and Osmolytes. Ann. Rev. Phys. Chem. 61, 63-83. 37. ElSawy, K., Caves, L., Twarock, R. (2010). The Impact of Viral RNA on the Association Rates of Capsid Protein Assembly: Bacteriophage MS2 as a Case Study. J. Mol. Biol. 400, 935–947. 38. Rolfsson, O., Toropova, K., Ranson, N. A, Stockley, P. G. (2010). Mutually-Induced Confor- mational Switching of RNA and Coat Protein Underpins Efficient Assembly of a Viral Capsid. J. Mol. Biol. 401, 309-322. 39. Valegård, K., Murray, J. B., Stockley, P. G., Stonehouse, N. J., Liljas, L., (1994). Crystal Struc- ture of an RNA Phage Coat Protein-Operator Complex. Nature 371, 623-626. 40. Douglas, T., Young, M. (1998). Host–Guest Encapsulation of Materials by Assembled Virus Protein Cages. Nature 393, 1996-1999. 41. Hu, Y., Zandi, R., Anavitarte, A., Knobler, C. M., Gelbart, W. M. (2008). Packaging of a Poly- mer by a Viral Capsid: The Interplay Between Polymer Length and Capsid Size. Biophys. J. 94, 1428-1436.

36 42. Hohn, T. (1969). Role of RNA in the Assembly Process of Bacteriophage fr. J. Mol. Biol. 43, 191-200. 43. Zacharias, D. A., Violin, J. D., Newton, A. C., Tsien, R. Y. (2002). Partitioning of Lipid-Mod- ified Monomeric GFPs Into Membrane Microdomains of Live Cells. Science 296, 913-916. 44. Hooker, J. M., Esser-Kahn, A. P., Francis, M. B. (2006). Modification of Aniline Containing Proteins Using an Oxidative Coupling Strategy. J. Am. Chem. Soc. 128, 15558-15559. 45. Karamyshev, A. L., Karamysheva, Z. N., Kajava, A. V., Ksenzenko, V. N., Nesmeyanova, M. A. (1998). Processing of Escherichia coli Alkaline Phosphatase: Role of the Primary Structure of the Signal Peptide Cleavage Region. J. Mol. Biol. 277, 859-870. 46. Simopoulos, T. T., Jencks, W. P. (1994). Alkaline Phosphatase is an Almost Perfect Enzyme. Biochemistry 33, 10375-10380. 47. Peters, R., Louzao, I., van Hest, J. C. M. (2011). From Polymeric Nanoreactors to Artificial Organelles. Chem. Sci. 3, 335–342. 48. Jesorka, A., Orwar, O. Liposomes: Technologies and Analytical Applications. (2008). Ann. Rev. Anal. Chem. 1, 801–832. 49. Laemmli, U.K. (1970). Cleavage of Structural Proteins during the Assembly of the Head of Bacteriophage T4. Nature 227, 680–685. 50. Samarkina, O.N., Popova, A.G., Gvozdik, E.Y., Chkalina, A.V., Zvyagin, I.V., Rylova, Y.V., Rudenko, N.V., Lusta, K.A., Kelmanson, I.V., Gorokhovatsky, A.Y., et al. (2009). Universal and Rapid Method for Purification of GFP-like Proteins by the Ethanol Extraction. Prot. Expr. Purif. 65, 108–113.

37 Chapter 3

The Influence of Electrostatics on Small Molecule Flux through a Protein Nanoreactor

3.1 Abstract Compartmentalization of proteins is a promising technique to enhance activity and stability of these molecules. Encapsulation in nanometer-sized containers to create nanoreactors has the potential to elicit interesting, unexplored effects resulting from deviations from well-understood bulk processes. Self-assembled protein shells for encapsulation are especially desirable for their uniform structures and ease of perturbation through genetic mutation. Here, we use the MS2 capsid, a well-defined porous 27 nm protein shell, as an enzymatic nanoreactor to explore pore-structure effects on substrate and product flux during the catalysed reaction. We find that the shell can influence the enzymatic reaction based on charge repulsion between small molecules and point mutations around the pore structure. These findings lend support to the potential influence of protein compartments on small molecules in metabolic reactions and in vitro catalysis. Portions of this work were performed in collaboration with Chris Jakobson.

38 3.2 Introduction A) Enzymes in natural metabolic systems are co-localized by genetic fusion, direct binding interactions, scaffolding, or PhoA-neg 1,2,3,4 compartmentalization. This is thought TMAO o to facilitate pathway flux by increasing- lo pH 6.0, 4 C cal concentration of intermediate molecules around downstream enzymes while limiting B) n n = n n HO O O off-target reactions, which can lead to un- 4-MU 5,6,7 productive accumulation or even toxicity. - - O O- - - - = P - HO - Many researchers have attempted to replicate - n - O - and use this effect in vitro and in vivo by add- - - n = 8,9,10,11,12 - - - n - ing artificial scaffolding, localization - - - O O O to specific organelles,13,14 protein engineer- - -O P O- 4-MUP ing,15,16,17 or encapsulation in protein compart- O ments18,19,20,21,22 or polymersomes.23,24,25,26,27 = PhoA-neg Despite numerous technological advances, C) one aspect of compartmentalization that has yet to be fully explored is the ability to influence passive diffusion of substrates, products, and intermediates. Kinetically favoring passage of substrates and products over intermediates through a compartment containing MS2 Capsid WT Pore appropriate enzymes is postulated to be an effective way to enhance pathways without altering individual proteins.5 Permeability effects have been examined in a variety of inorganic and polymeric 28,29,30,31,32 nanoreactors, but there are only a T71E Pore T71K/V72R Pore handful of proteins complexes studied for such properties. Ferritin cages, for example, Figure 1. Testing the effects of pore charge on substrate direct Fe(II) ions for oxidation using nega- and product flux. A) Bacteriophage MS2-based enzymatic tively charged pores.33 Natural membrane nanoreactors are created through self-assembly of the capsid around negatively charged alkaline phosphatase (PhoA-neg) protein ion channels have both ion-specific variants. B) Capsid pore charge is varied by mutagenesis, and nonspecific electrostatic interactions and reactions of negatively charged phosphatase substrates with their substrates, affecting their transfer are monitored. C) Vacuum electrostatic representations of pore mutants modeled in PyMol. Negative potential is rep- 34 kinetics and selectivity. The carboxysome, resented by red, and positive by blue. (top left) The capsid a protein-based bacterial microcompartment, features ~1.8 nm pores at each of its 5- or quasi-6-fold axes. is also hypothesized to use positively charged (top right) Wild-type MS2 has negative charge around the pore periphery, but not inside. (bottom left) Mutant T71E pores to allow passive diffusion of bicarbon- has significant negative charge throughout the pore. (bottom right) Mutant T71K/V72R has positive charge in the pore. ate ions while keeping in hydrophobic CO2 and excluding molecular oxygen, which can lead to undesired side reactions.35,36 Inspired by these natural systems, we sought to develop a protein-based nanoreactor with mutatable pores to examine the effects of electrostatics on substrate and product flux (Figure 1). 39 A) E Reassembly vs. TMAO B) ED Reassembly vs. TMAO 35 35 With PhoA-neg With PhoA-neg 30 30 No PhoA-neg No PhoA-neg ) ) 25

% 25 % ( (

y y l l b 20 b 20 m m e e s 15 s 15 s s a a R e 10 R e 10

5 5 n.d n.d n.d n.d 0 0

[TMAO] (M) [TMAO] (M)

C) Size Exclusion Chromatography of Purified MS2/PhoA-neg Variants 150 Untreated WT WT/PhoA-neg E/PhoA-neg E/PhoA-neg-GN U . ) .

b 100 KR/PhoA-neg WT/GN r A

( ED/PhoA-neg . l F

. 50 p r T

0 0 5 10 15 Time (min) Figure 2. Reassembly of capsid mutants around PhoA-neg. A) T71E MS2 reassembly around PhoA-neg in Bis Tris pH 6.0 is greatly enhanced by high concentrations of trimethylamine N-oxide (TMAO). B) T71E/V72D MS2 reassembly is also enhanced by TMAO. No significant reassembly was observed below 1 M. Background reassembly, though minimal, may result in some empty capsids. C) Size exclusion chromatography of all purified enzyme/capsid combinations. All have the same retention time and show no detectable free enzyme, even after 2 months of storage at 4 oC.

We previously developed the bacteriophage MS2 viral capsid as a stable, versatile compartment for encapsulation of negatively charged macromolecules enhanced by the presence of a protein-stabilizing osmolyte, trimethylamine N-oxide (TMAO).37,38 The capsid is also very well characterized for tolerance to mutations, particularly in the assembly-directing pore residues of the FG-loop.39,40 This loop forms the inter-subunit contacts lining the 1.8 nm pores at the 5-fold and quasi-6-fold axes of the icosahedral capsid. Point mutations in this loop can add 5 or 6 charges directly around the capsids pores. Here, we show here that mutant capsids are able to encapsulate negatively charged alkaline phosphatase enzymes. When compartmentalized within capsids with the pores uncharged or oppositely charged as the substrate, the enzyme kinetics are relatively unaffected; however when pore and substrate charge are the same the apparent Km of the enzyme increases significantly. Kinetic modeling suggests this is caused by a mix of inhibition of substrate influx and product efflux, with the latter leading to inhibition of the enzyme. These experi-

40 10000 ED

) WT m n (

E r 1000 e t

e KR m D i a

e g

a 100 r e v A

50 nm 50 nm 10 30 40 50 60 70 Temperature (oC) Figure 3. Structure and stability of MS2-encapsulated PhoA-neg derivatives. A) DLS analysis of temperature stability of encapsulated enzymes. Aggregation occurs above 50 oC for all mutants. B) Negative stain TEM of WT MS2 with PhoA-neg inside. C) Negative stain TEM of T71E MS2 with PhoA-neg inside. Both samples were stained with

UO2(OAc)2. Scale bars represent 50 nm. ments represent a first step in creating selec- Capsid and Average tive nanoreactors with the potential to either Enzyme Type Enzyme/Capsid protect cargo by inhibiting entrance of inter- WT/PhoA-neg 9.6 fering molecules or enhance multi-step path- T71E/PhoA-neg 5.8 ways by concentrating intermediates. T71E/V72D/PhoA-neg 3 T71K/V72R/PhoA-neg 9.4 3.3 PhoA Kinetics in MS2 Pore Mutants WT/PhoA-neg-GN 9.5 We generated several mutants of T71E/PhoA-neg-GN 6 the wild-type (“WT”) MS2 capsid by Qui- Table 1. Enzyme loading in MS2 derivatives. Enzyme kchange mutagenesis and expressed them as concentration was measured by SDS-PAGE densitometry. Total protein was measured by A280 using ε = 0.71 cm∙ml/ described (Figure 1C). The T71E (“E”) and mg for PhoA-neg and ε = 1 cm∙ml/mg for MS2. T71E/V72D (“ED”) mutants add one or two negative charges to each subunit, respectively, while the T71K/V72R (“KR”) mutant adds two positive charges. This results in capsids with -5/-6 (E), -10/-12 (ED), or +10/+12 (KR) extra charges around the 5- and quasi-6-fold axis pores on each capsid. The mutations had little or no effect on expression and assembly in E. coli, and had only a small effect on the thermal stability of the capsids (Figures 2 and 3). We subjected each of these to our previously developed disassembly/ reassembly protocol, which encapsulates negatively charged molecules, such as enzymes, based on electrostatic interactions with the positive residues on the capsid interior surface. To initiate reassembly, we used a model enzyme, E. coli alkaline phosphatase, with a C-terminal negatively charged sequence (PhoA-neg).37 The KR mutant reassembled similarly to the WT capsid in these conditions (50 mM Tris HCl pH 7.2, 100 mM NaCl, 36-48 hours, 4 oC; Figure 2). The E and ED mutants, however, gave significantly lower yields of reassembled product. We varied the salt, pH, and TMAO concentrations of the reassembly reaction to determine which conditions favored reassembly of these mutants. While salt concentration did not significantly affect the yield of reassembled capsid, lowering the pH from 7.2 to 6.0 and increasing the TMAO concentration from 0.25 M to 1.8 M enhanced the yield of reassembled E and ED capsids by 8-fold (Figure 2A and 2B). Using these improved reassembly conditions, adequate amounts of encapsulated enzyme in each of the MS2 mutants were readily obtained. 41 20 20 A) B) PhoA-neg 18 Med. Salt . ) 18

z PhoA-neg FITC

16 Low Salt n 16 PhoAneg Fit e Medium salt Med. Salt Fit 14 -1 Low Salt Fit 14 PhoA-neg FITC Fit kcat = 18.2 ± 0.1 s / M 12 Km = 1.47 ± 0.1 µM 12

10 U / s 10 PhoA-neg k = 17.6±0.42 s 8 8 cat -1

4- M Km = 15.2±1.7 μM 6 Low salt 6 k = 10.1 ± 0.3 s-1 M PhoA-neg FITC cat ( k = 16.2±1.3 s-1 v (4-MU/s/M enz.) 4 K = 1.00 ± 0.1 µM 4 cat m v 2 2 Km = 16.5±0.77 μM 0 0 0 20 40 60 0 200 400 600 [4-MUP] (µM) [4-MUP] (µM) Figure 4. Analysis of free PhoA-neg kinetics. A) The reassembly conditions have little effect on the kinetics of FITC- labeled PhoA-neg in 4 mM or 100 mM NaCl. B) The kinetics of FITC-labeled and unlabeled PhoA-neg are within experimental error in high salt MOPS.

Free WT E ED KR k (s-1) 23.8 ± 3.8 13.21 ± 1.0 19.7 ± 1.6 16.9 ± 2.0 18.3 ± 2.46 Tris High Salt cat Km,app (µM) 2.13 ± 0.09 2.59 ± 0.30 13.4 ± 0.24 5.72 ± 0.22 2.76 ± 0.12 k (s-1) 11.6 ± 0.83 5.40 ± 0.34 4.75 ± 0.57 4.87 ± 1.0 8.79 ± 1.1 Tris Medium Salt cat Km,app (µM) 0.97 ± 0.02 1.0 ± 0.02 4.8 ± 0.63 2.18 ± 0.29 1.0 ± 0.07 k (s-1) 5.23 ± 0.25 2.51 ± 0.12 2.74 ± 0.30 2.59 ± 0.30 4.68 ± 0.57 Tris Low Salt cat Km,app (µM) 0.54 ±0.003 0.89 ± 0.03 2.13 ± 0.15 1.89 ± 0.16 0.56 ± 0.03 k (s-1) 7.58 ± 0.02 4.89 ± 0.33 4.40 ± 0.34 4.06 ± 0.57 n.d. MOPS Low Salt cat Km,app (µM) 2.31 ± 0.07 3.79 ± 0.29 6.87 ± 0.71 6.42 ± 0.50 n.d. -1 Tris Medium Salt kcat (s ) 43.3 ± 8.1 31.4 ± 5.3 19.8 ± 5.3 n.d. n.d. (GN mutant) Km,app (µM) 1.43 ± 0.16 2.45 ± 0.10 15.0 ± 1.8 n.d. n.d.

Table 2. Kinetic parameters for enzyme constructs used in this study. Kinetic constants for free and encapsulated

PhoA-neg and PhoA-neg-GN derivatives in several conditions all show substantial increases in Km,app. Error represents aggregation of standard deviations from at least three kinetic assays using nine substrate concentrations and at least three SDS-PAGE densitometry measurements to determine enzyme concentration.

To ensure complete encapsulation of PhoA-neg, the enzyme was labeled with fluorescein 5(6) isothiocyanate (FITC). The labeling had no effect on the enzymatic activity, nor did incubation of the free enzyme in reassembly conditions (Figure 4). The labeled enzyme was encapsulated in each of the capsid mutants. Upon encapsulation, a change in optical properties of the dye was observed; therefore the fluorescein was only used as a qualitative indicator of presence of the enzyme within the capsid; enzyme quantification was performed by running the purified constructs on SDS-PAGE gels followed by coomassie staining and optical densitometry. The capsids each contained a similar amount of enzyme, with the WT and KR capsids carrying slightly more, followed by E then ED (Table 1). It is not surprising that added negative charge decreases encapsulated enzyme, due to the electrostatic nature of the interaction between PhoA-neg and the capsid. It was also found that increasing salt lowered the number of enzymes per KR capsid, supporting the importance of electrostatics in the enzyme/capsid interaction. We tested the Michaelis-Menten kinetics of the hydrolysis of 4-methylumbelliferyl phosphate to 4-methylubelliferone catalyzed by PhoA-neg in each of the capsids as shown in Figure

42 A) 5. We chose to do the experiments in low

) 14 concentration Tris buffer, although similar e Free

m 12 trends were observed in MOPS (Table 2). y KR z

n 10 WT Although the presence of Tris can lead to e ED transphosphorylation onto the buffer, artifi- M / 8 E s / cially increasing kcat, this effect is negligible

U Free Fit 6 at the low concentrations used (10 mM).41 At

- M KR Fit 4 4 WT Fit the pH of the reaction, both the substrate and

( M ED Fit 2 product are negatively charged, which could v E Fit lead to electrostatic effects on flux both in and 0 0 10 20 out of the compartment. As we observed pre- [4-MUP] (μM) viously, the WT capsid had no effect on the B) 7 apparent Km (Km,app) but a modest inhibitory

effect on the kcat. Interestingly encapsulation

p 6

p Free a , in KR had virtually no effect on the kinetics m 5 WT K

at all. On the contrary, both the negative pore d E e 4 mutants significantly increased the K of z m,app i l ED a 3 the reaction while decreasing the overall kcat.

m KR r Curiously, E increased the Km,app significantly o 2

N more than ED; we speculate that because D72 1 in the ED mutant is located somewhat exter- 0 High Medium Low nal to the pore, the mutation could change the Salt Salt Salt structure of the pore or even alter the pKa of E71, effectively increasing pore size or de- Figure 5. Enzyme assays with free and encapsulated creasing repulsion on the substrate and prod- PhoA-neg. A) Michaelis-Menten curves of the enzyme encapsulated in each MS2 pore variant in 10 mM Tris uct molecules. The increase in Km,app persists

pH 8.0, 100 mM NaCl (medium salt). B) Apparent Km at a variety of salt concentrations. Theoretical of the enzyme in each pore variant in 10 mM Tris pH studies suggest electrostatic barriers in nano- 8.0 with 4 mM (low salt), 100 mM (medium salt), or pores can be overcome by high ionic strength 500 mM NaCl (high salt). The Km,app of both negative pore mutants, E and ED, increases significantly. Each even when the substrate size approaches the K was normalized to the free enzyme in the cor- m,app debye length of the pore; our data suggests responding buffer conditions. Error bars represent standard deviation of 3 assays. our system is not in this regime, even at 0.5 M NaCl.42 We next tested the encapsulation effects on a mutant of alkaline phosphatase that has different kinetic parameters. This mutant, D153G/D330N (PhoA-neg-GN), was developed to have a 43 significantly higherk cat and Km, while suffering less inhibition from free phosphate. We expressed this mutant and encapsulated it in WT and E capsids. The kinetics of the free PhoA-neg-GN enzyme, as expected, showed an increase in both kcat and Km relative to the free wild-type enzyme.

When encapsulated in WT capsid, this enzyme had moderate increase in Km,app; encapsulation in the E capsid led to an even more pronounced increase in the Km,app, suggesting that charge repulsion remains important with the mutant enzyme (Table 2).

3.4 Modeling of Encapsulated Reactions The MS2/PhoA-neg nanoreactor system is significantly more complex than the classical Michaelis-Menten model of enzyme kinetics. Therefore, using the data from the above experi- 43 ments, we developed a kinetic model of substrate transport and enzymatic reaction kinet- A) ics in the MS2 nanoreactor system. The MS2 Pi O O O E•P i HO - nanoreactor was modeled using the reaction -O P O O k2 P O- scheme shown in Figure 6A, incorporating Ki O - k1 4-MUP O substrate diffusion between the bulk and the 4-MUP + E + Pi capsid, Michaelis-Menten enzyme kinetics k1 E with competitive product inhibition by phos- + 4-MU k phate, and phosphate diffusion out of the k f r kcat 4-MU capsid. Kinetic studies of the PhoA-neg and E•4-MUP PhoA-neg-GN enzymes in T71E MS2 capsid were modeled by numerical integration -O O O of this reaction scheme for each substrate B) 35 concentration tested experimentally, fol- WT/GN lowed by linear regression to determine the 30 (Experiment) initial reaction rate (Figure 6). The four ki- 25 E/GN netic parameters (k1, k2, Ki(WT), and Ki(GN)) 20 were fitted to the experimental data using the (Experiment) 15 built-in lsqnonlin minimization function in E/GN 10 MATLAB. The intrinsic Michaelis-Menten k (wt) (Model) v (M 4-MU/s/M enz.) 2 parameters k and K fitted to the WT 5 ≈100 cat,app m,app k2 (T71E) capsid case were held constant in the T71E 0 case for each enzyme. The transport param- 0 20 40 60 [4-MUP] (µM) eters k1 and k2 capture the effects of the pore mutation T71E on substrate and inhibitor Figure 6. Kinetic model of encapsulated reactions. A) System of reactions used as parameters for the ki- transport, respectively, relative to a WT MS2 netic model. Variables highlighted in green defined the capsid, and the two competitive inhibition kinetics of the system. Experimental kinetic data from parameters K (WT) and K (GN) capture the WT and E encapsulated PhoA-neg were used as inputs i i to solve k , k , and K . B) Results of model for encap- differences in competitive phosphate inhibi- 1 2 i sulated PhoA-neg-GN. WT/GN was used to define kcat tion between the PhoA-neg enzyme and the and Km, and T71E/PhoA-neg values were used for k1 and k . PhoA-neg-GN mutant. The modeled enzyme 2 kinetics for both the PhoA-neg and PhoA- neg-GN enzymes in T71E MS2 are shown alongside the WT MS2 and T71E MS2 experimental data in Figure 6B. The competitive inhibition parameters Ki(WT) and Ki(GN) were fitted to be 1.74 μM and 2.67 μM, respectively, comparable to the experimental values of 1.4 μM and 6.4 μM found in the case of free enzyme inhibition by phosphate (Figure 7). Simulated concentration profiles for the product, substrate, and competitive inhibitor are shown for each experimental substrate concentration in Figure 8, indicating a steady- state buildup of phosphate concentration inside the capsid. The model does not account for effects on the intrinsic enzyme kinetics relative to the WT MS2 capsid case due to the T71E pore mutation, changes in the competitive inhibition constants

Ki(WT) and Ki(GN) due to encapsulation, nor for modes of inhibition other than product inhibition by phosphate, but nonetheless we find that a model in which substrate and inhibitor diffusion are limited due to the negatively-charged T71E pore can account for the difference in apparent kinetics between an enzyme encapsulated in a WT MS2 capsid and one encapsulated in a T71E mutant

44 Phosphate Inhibition of PhoA-neg Phosphate Inhibition of PhoA-neg-GN 5 6

y = 0.2458x + 1.4005 4 y = 0.6946x + 0.4298 5 R² = 0.9995 R² = 0.9922 4 p p

p 3 p a a , ,

m m 3 K K 2 2

1 1

0 0 0 2 4 6 0 5 10 15 20 [Phosphate] (µM) [Phosphate] (µM)

Figure 7. Experimental determination of Ki for A) Free PhoA-neg and B) PhoA-neg-GN in medium salt Tris buffer. capsid (Figures 9 and 10). This supports the T71E/ T71E/ Parameter hypothesis that changes in pore charge can PhoA-neg PhoA-neg-GN affect transport in and out of the nanoreactor k 5.08 5.08 and thus affect enzyme kinetics. A relaxation 1 k 3.04x106 3.04x106 of the transport limitation of the phosphate 2 K (WT) 1.74 μM - ion (modeled as a 100-fold increase in the i

Ki(GN) - 2.67 μM phosphate transport parameter k2) approximates the behavior observed in the WT MS2 Table 3. Parameter values of the kinetic model fitted capsid for both PhoA-neg (Figure 9D) and to the T71E/PhoA-neg and T71E/PhoA-neg-GN enzyme systems. Km,app and kcat,app values are held con- PhoA-neg-GN (Figure 10D), further support- stant at the values fitted to the WT MS2 cases. ing this hypothesis.

3.5 Conclusion Here we show that porous protein compartments can indeed have an effect on encapsulated enzymatic reactions. In our model system a large negatively charged substrate must overcome a coulombic barrier to reach the enzyme. Similarly, the product must overcome a barrier to escape, potentially leading to product inhibition of the enzyme. It is important to note that this system cannot enhance the enzyme’s rate; we instead demonstrate the potential of both controlled access to the interior of a protein and product buildup, which could be useful for compartmentalized me- tabolons. In the future, this study can be used to develop computational models for other designed compartments with selectivity based on size, charge, or hydrophobicity. Theories generated can then be applied to real systems, such as bacterial microcompartments for production of fuels, chemical feedstocks, and pharmaceuticals in bacterial hosts.

45 PhoA−neg PhoA−neg−GN 200 800

150 600

100 400 Product [uM] 50 Product [uM] 200

0 0 0 10 20 30 40 50 0 10 20 30 40 50 Time (sec) Time (sec) PhoA−neg PhoA−neg−GN 20 60

15 40 10 20 5 Substrate [uM] Substrate [uM]

0 0 0 10 20 30 40 50 0 10 20 30 40 50 Time (sec) Time (sec) PhoA−neg PhoA−neg−GN 15 60

10 40

5 20 Inhibitor [uM] Inhibitor [uM]

0 0 0 10 20 30 40 50 0 10 20 30 40 50 Time (sec) Time (sec) Figure 8. Simulated concentration profiles for the product, substrate, and competitive inhibitor are shown for each experimental substrate concentration. A linear regression was performed to determine the initial rate of product formation at each substrate concentration. Substrate and inhibitor concentrations were both observed to reach a pseudo-steady state over the timescale of the simulation.

3.6 Materials and Methods General Methods

Unless otherwise noted, reagents were used as received from commercial sources. In all experie- ments, water was deionized using a MilliQ system (Millipore). Chemicals were obtained from Fisher Chemical except Tris base (VWR), TMAO dihydrate (Alpha Aesar), triethanolamine (Sig- ma), 4-methylumbelliferyl phosphate (Invitrogen) and 4-methylumbelliferone (Sigma). Molecular biology reagents were obtained from New England Biolabs.

Instrumentation and Sample Analysis

High Performance Liquid Chromatography. HPLC was performed using an Agilent 1100 series HPLC using ST buffer (50 mM Tris pH 7.2, 100 mM NaCl) or sodium phosphate buffer (10 mM, pH 7.2) as mobile phase. All size exclusion chromatography was performed using a Biosep SEC-4000 column (Phenomenex). Liquid Chromatography/Mass Spectrometry (LC/MS) was performed using an Agilent 1200 series HPLC connected to an Agilent 6224 Time-of-Flight (TOF)

46 A) B) k1 k2 6 6

5 5

4 4

3 3

2 2 Rate [M 4−MUP/M enz/sec] Rate [M 4−MUP/M enz/sec] 1 1

0 0 0 5 10 15 20 0 5 10 15 20 [4−MUP] µM [4−MUP] µM C) D)

Ki(wt) 6 6 k2 x100 5 5 k x10 4 2 4

3 3 original fit

2 2

Rate [M 4−MUP/M enz/sec] T71E model 1 Rate [M 4−MUP/M enz/sec] 1 wtMS2 data T71E data 0 0 0 5 10 15 20 0 2 4 6 8 10 12 14 16 18 20 [4−MUP] µM [4−MUP] µM

Figure 9. Sensitivity analysis for kinetic parameters for encapsulated PhoA-neg. The model of T71E MS2 capsid

kinetics is shown for PhoA-neg with each fitted kinetic parameter A) k1, B) k2, and C) Ki(WT)) increased 2-fold and decreased 2-fold. The area between these two curves is shaded gray to indicate the sensitivity of the model to the value of the given parameter. Greater deviations from the experimental data upon perturbation indicate greater sensitivity to the value of that parameter. The goodness of fit of the model is found not to be sensitive to the substrate diffusion parameter

k1, but is found to be sensitive to the phosphate diffusion parameter k2 and the competitive inhibition parameter Ki(WT).

D) The model of T71E MS2 capsid kinetics is shown for PhoA-neg with the phosphate-transport parameter k2 increased by 10-fold and 100-fold. The model approximates the WT MS2 capsid kinetics upon relaxation of this parameter by 100- fold, indicating that changes in phosphate diffusion alone in this model can account for the difference in apparent kinetics between an enzyme in a WT MS2 capsid and one in a T71E MS2 capsid.

LC/MS system equipped with a Turbospray ion source. For purification, samples were filtered through 0.22 µm filters (Millipore). HPLC-based quantification was performed by comparing -in tegrated fluorescence intensity of tryptophan (ex. 280/ em. 330) over time of untreated MS2 to reassembled samples.

Dynamic light scattering. DLS was performed using a Zetasizer Nano (Malvern Instruments). Be- fore analysis samples were filtered through 0.22 µm filters (Millipore). Measurements were taken in 50 mM Tris pH 7.2, 100 mM NaCl.

Enzyme kinetics. Enzyme assays were performed in 96-well plates 10 wells at a time. Fluorescence was read using a Spectramax M2 plate reader (Molecular Devices). The excitation wavelength was 360 nm, and the emission wavelength was 449 nm. Substrate stock solutions of 4-methylumbel-

47 A) B)

35 k1 35 k2

30 30

25 25

20 20

15 15

10 10 Rate [M 4−MUP/M enz/sec] 5 Rate [M 4−MUP/M enz/sec] 5

0 0 0 10 20 30 40 50 0 10 20 30 40 50 [4−MUP] µM [4−MUP] µM C) D) 35 35 Ki(GN) 30 30 25 k2 x100 25 k x10 20 2 20

15 15 original t 10 10

Rate [M 4−MUP/M enz/sec] T71E model

Rate [M 4−MUP/M enz/sec] 5 5 wtMS2 data T71E data 0 0 0 10 20 30 40 50 0 5 10 15 20 25 30 35 40 45 50 [4−MUP] µM [4−MUP] µM Figure 10. Sensitivity analysis for kinetic parameters for encapsulated PhoA-neg. The model of T71E MS2 capsid

kinetics is shown for PhoA-neg-GN with each fitted kinetic parameter A) k1, B) k2, and C) Ki(WT)) increased 2-fold and decreased 2-fold as indicated by the gray shaded regions. The goodness of fit of the model is again found not to be sensi-

tive to the substrate diffusion parameter k1, but is found to be sensitive to the phosphate diffusion parameter k2 and the

competitive inhibition parameter Ki(GN). D) The model of T71E MS2 capsid kinetics is shown for PhoA-neg-GN with

the phosphate-transport parameter k2 increased by 10-fold and 100-fold. Again, the model approximates the WT MS2 capsid kinetics upon relaxation of this parameter by 100-fold, liferyl phosphate and standards of 4-methylumbelliferone were prepared fresh daily to account for background substrate hydrolysis, bleaching, or variability in plate reader signal. Michaelis-Menten parameters were determined by fitting to the Michaelis-Menten equation in OriginPro 8 software. Modeling of kinetic parameters was performed in MATLAB.

Gel analyses. Sodium dodecylsulfate-poly(acrylamide) gel electrophoresis (SDS-PAGE) analysis of all protein samples was carried out on a Mini Protean apparatus (Bio-Rad, Hercules, CA) using 12.5% poly(acrylamide) gels prepared according to the manufacturer’s specifications. Samples were heated to 95 oC in the presence of Laemmli buffer containing β-mercaptoethanol for 3-5 minutes. Gels were run at 125 V for approximately 75 minutes, stained with Coomassie R-250, and imaged using a Chemidoc imager (Bio-Rad). Molecular masses were estimated by comparison to the EZ-run Prestained Rec Protein Ladder (Fisher). Quantification of PhoA-neg bands was accomplished by densitometry using the Chemidoc software using known quantities of free enzyme as standards.

UV-vis spectroscopy. UV-vis spectroscopy was performed using a Nanodrop 2000C (Thermo Sci- 48 entific). Free enzyme was quantified by absorbance at 280 using an extinction coefficient of ε = 0.71 cm∙ml/mg. Nucleic acid-free MS2 was quantified using ε = 1 cm∙ml/mg.

Transmission electron microscopy. TEM images were taken using an FEI Technai 12 transmission electron microscope with an accelerating voltage of 120 kV. Samples were desalted using NAP-5 desalting columns (GE Healthcare), concentrated to approximately 50 µM using 100 kDa MWCO spin filters (Millipore), incubated on Formvar-coated copper mesh grids for 5 minutes and wicked off using filter paper and dried in air briefly. These grids were then quickly washed with water and immediately wicked again. The samples were stained with 1% UO2(OAc)2 for 2 minutes, and again wicked and dried.

Experimental Procedures

Generation of MS2 mutants. The residues around the pore of the capsid, specifically T71 and V72 (numbering based on crystal structure 1ZDK) were targeted for mutation using the Quikchange method. The mutagenesis primers were found to interfere with ColE1 plasmid origins, so the gene for wild-type MS2 was cloned into pBAD33 with a p15A origin. The gene for wild-type MS2 coat protein was PCR-amplified from pBAD-myc-His-WTMS2, retaining that construct’s 5’ XbaI site and 3’ HindIII site (forward: 5’-AGTCAGCCGTTTTCTAGACTAACAGG-3’; reverse: 5’-AGTCAGCCCAAGCTTAGTAGATGCCG-3’). The gene was digested, ligated into pBAD33, and transformed into DH10B cells. Plasmid pBAD33-WTMS2 was used to generate mutants T71E, T71E/V72D, and T71K/ V72R using the primers in Table 3 below. For optimal expression, the mutant genes were cloned into pTrc99a using the same method outlined above.

Mutant Primer T71E 5’-GGTGCCTAAAGTGGCAACCCAGGAGGTTGGTGGTGTAGAGC-3’ 5’-GCTCTACACCACCAACCTCCTGGGTTGCCACTTTAGGCACC-3’ T71E/V72D 5’-GGTGCCTAAAGTGGCAACCCAGGAGGACGGTGGTGTAGAGC-3’ 5’-GCTCTACACCACCGTCCTCCTGGGTTGCCACTTTAGGCACC-3’ T71K/V72R 5’-GGTGCCTAAAGTGGCAACCCAGAAACGCGGTGGTGTAGAGC-3’ 5’-GCTCTACACCACCGCGTTTCTGGGTTGCCACTTTAGGCACC-3’ Table 3. Mutagenesis primers for MS2 capsid pore mutants.

Expression and purification of MS2. Wild-type MS2 capsid and and pore variants were expressed and purified as previously described. Plasmid pBAD-myc-his containing the gene for the wild-type capsid or pTrc99a containing the desired pore MS2 variant was transformed into DH10B cells and plated on LB + 50 mg/L carbenicillin (Cb). A single colony was grown overnight in 5 mL of LB + Cb and used to inoculate 500 mL 2xYT + Cb at a 1:500 dilution. This culture was shaken at 37 oC until it reached an OD600 of 0.5, at which time it was induced with 0.2 % arabinose (for pBAD) or 1 mM IPTG (for pTrc) and allowed to express overnight at the same temperature. Cells were harvested, resuspended in 10 mM taurine pH 9.0 and lysed by sonication. The clarified lysate was applied to a DEAE column and the flowthrough containing the MS2 mutant was retained. To precipitate the capsid, solid poly(ethylene glycol) 8k and 5 M NaCl were added 49 to final concentrations of 10% w/v and 0.5 M, respectively, and the mixture was incubated at 4o C for 1 hour on an orbital shaker and centrifuged at 14,000 x g for 30 minutes. The pelleted protein was resuspended in a small amount of 10 mM phosphate buffer pH 7.2 and centrifuged again. The re-solubilized capsid was then applied to a S200 size exclusion chromatography column and eluted in 10 mM phosphate buffer. Fractions containing MS2 were re-precipitated and stored at 4 oC.

Generation of PhoA-neg-GN. Plasmid pTrc99a harboring the gene for PhoA-neg was subjected to two successive rounds of Quikchange mutagenesis. First, the D330N mutation was made using forward primer 5’-GTGCGTCAATCGATAAACAGAATCATGCTGCGAATCCTTG-3’ and reverse primer 5’-CAAGGATTCGCAGCATGATTCTGTTTATCGATTGACGCAC-3’ and the resulting plasmid was transformed into DH10B cells, isolated and sequenced. The D153G mutation was then made using forward primer 5’-CGTTTCTACCGCAGAGTTGCAGGGCGCCAC- GCCCGCTGCGCTGGTGGC-3’ and reverse primer 5’-GCCACCAGCGCAGCGGGCGTGGC- GCCCTGCAACTCTGCGGTAGAAACG-3’ and resulting plasmids were again transformed into DH10B cells, isolated and sequenced.

Expression and purification of PhoA-neg and PhoA-neg-GN. PhoA-neg and PhoA-neg-GN were expressed and purified as previously described.37 Plasmid pTrc99a harboring the gene for the desired enzyme was transformed into DH1 cells and plated on LB-Cb. A single colony was grown over-night A single colony was grown overnight in 5 mL of LB + Cb and used to inoculate 1 L

2xYT + Cb supplemented with 1 mM MgSO4 and 0.1 mM ZnSO4 at a 1:500 dilution. This culture o was shaken at 37 C until it reached an OD600 of 0.5, at which time it was induced with 1 mM IPTG and allowed to express overnight at the same temperature. Cells were harvested and washed once with 100 mL of 50 mM Tris HCl pH 8.0, and resuspended in 50 mL of the Tris buffer. Periplasmic proteins were isolated by addition of solid sucrose to 500 mM, 500 mM EDTA to 2.5 mM, and lysozyme to 0.6 mg/ml. The resulting suspension was gently mixed and incubated at 37 oC for 30 minutes, followed by centrifugation at 5000 x g for

20 minutes. Solutions of MgSO4 and ZnSO4 were added to 10 mM and 1 mM final concentration, respectively, and the mixture was heated to 80 oC for 10 minutes, with intermittent mixing. Pre- cipitated protein was pelleted by centrifugation at 10,000 x g for 20 minutes followed by filtration through a 0.22 μm filter. This was then applied to a 5 mL bed of DEAE-sepharose and washed with 20 mL aliquots of 20 mM Tris containing 0 mM, 50 mM, and 100 mM NaCl. The protein was o eluted with 250 mM NaCl and stored at 4 C in the same buffer with MgSO4 and ZnSO4 added to 1 mM and 0.1 mM final concentrations, respectively. This protocol yields approximately 40 mg of PhoA-neg and 25 mg of PhoA-neg-GN per liter of culture.

FITC Labeling of PhoA-neg. PhoA-neg was dialyzed into triethanolamine buffer pH 8.3 for labeling. Fluorescein 5(6)-isothiocyanate (50 mM freshly dissolved in DMF) was added to 1.3 mg/ ml enzyme in 50-fold molar excess and incubated at 4 oC overnight. Excess FITC was removed by dialysis using a 10 kDa MWCO dialysis cassette (Pierce). Attachment of the fluorophore was confirmed by electrospray ionization-time of flight mass spectrometry (ESI-TOF MS) and size exclusion chromatography (SEC) using a Biosep-SEC-4000 column at a flow rate of 1 ml/min.

Encapsulation of enzymes in MS2 mutants. A modified protocol was used to encapsulate enzymes inside MS2 pore mutants. MS2 was disassembled in 66% acetic acid and desalted into 1

50 Kinetic Modeling. Rates are normalized to enzyme concentration [M 4-MU/s/M enz.] Parameters

Species k1 substrate transport

S* substrate in assay solution k2 inhibitor transport S substrate inside capsid kcat catalytic rate constant I inhibitor inside capsid KM Michaelis-Menten constant P product total KI competitive inhibition constant Rates FV volume fraction of assay solution that is capsid lumen (~10-7) dS* =0 Key assumptions dt dS dP Michaelis-Menten kinetics with competitive =- +k S*-S dt dt 1 product inhibition by phosphate dI 1 dP ( ) dS* = -k2I =0 dt FV dt dt

dP kcatS I=0 outside capsid = dt I KM 1+ +S kcat,KM unchanged by encapsulation KI ( )

Figure 11. Inputs to develop kinetic model of encapsulated reactions. mM acetic acid as described. The coat protein was added to a chilled solution of enzyme (0.1 mg/ ml), Bis-Tris buffer (50 mM, pH 6.0), and TMAO (1.8 M). The capsid was allowed to reassemble for 48 hours at 4 oC before analysis and purification.

Purification of Encapsulated Enzymes. Reassembled capsid containing the desired enzyme derivative was purified by precipitation with PEG 8k followed by SEC-HPLC. Solid PEG 8k and 5 M NaCl were added to the reassembly reaction to a final concentration of 10% and 0.5 M, respectively, and the mixture was incubated at 4 oC on an orbital shaker and centrifuged at 21,000 x g for 10 minutes. The precipitated capsids were resuspended in ST buffer (50 mM Tris pH 7.2, 100 mM NaCl) and centrifuged again to pellet any aggregated protein. The supernatant was filtered through a 0.22 μm filter and injected on a Biosep-SEC-4000 column. ST buffer was pumped over the column at 1 ml/min, and reassembled MS2, which elutes at ~6.3 minutes was collected and concentrated using 100 kDa MWCO centrifugal filters (Millipore).

Evaluation of Stability of MS2 encapsulated PhoA-neg. Because the enzyme chosen is reported to be more stable than the MS2 capsid, we examined thermal stability of our system using dynamic light scattering (DLS). Size trends upon change in temperature were recorded for each capsid, starting at 35 oC in 4 oC steps with incubation at each temperature.

51 Enzyme Kinetics. Kinetics of free and encapsulated enzymes were measured by monitoring the increase in fluorescence (λex. 360 nm/λem. 449 nm) upon hydrolysis of the phosphoester bond in 4-methylumbelliferyl phosphate. The enzyme (100μl, 2 nM monomer) in 2x buffer was added to the 100 μl H2O containing the substrate at increasing concentrations, and initial rates were recorded. Standards of 4-methylumbelliferone were used to generate a standard curve in the relevant signal range to convert fluorescent signal to concentration. Data were fit to the Michaelis-Menten equation using OriginPro 8 fitting software. Assays were done in the following buffers: 50 mM

MOPS pH 8.0, 500 mM NaCl, 1 mM MgSO4, 0.1 mM ZnSO4 (high salt); 50 mM MOPS pH 8.0, 1 mM MgSO4, 0.1 mM ZnSO4 (low salt); 10 mM Tris pH 8.0, 4 mM NaCl, 1 mM MgSO4, 0.1 mM

ZnSO4 (low salt); 10 mM Tris pH 8.0, 100 mM NaCl, 1 mM MgSO4, 0.1 mM ZnSO4 (medium salt); 10 mM Tris pH 8.0, 500 mM NaCl, 1 mM MgSO4, 0.1 mM ZnSO4 (high salt).

Determination of inhibition constants (Ki) for free enzymes. Free PhoA-neg and PhoA-neg-GN were incubated with increasing concentrations of phosphate and assayed in medium salt buffer for kinetic parameters as described above. The resulting Km,app values were plotted against phosphate concentrations and fit to theK i equation.

Kinetic modeling. Kinetic modeling was performed in MATLAB using the species, rates, and parameters defined in Figure 11. Key assumptions include approximating the enzyme to follow Michaelis-Mentet kinetics, no initial free phosphate, and that the the change in substrate concentration over the course of the reaction is negligible.

3.7 References 1. Huang, X., Holden, H.M., and Raushel, F.M. (2001). Channeling of Substrates and Intermedi- ates in Enzyme-Catalyzed Reactions. Ann. Rev. Biochem. 70, 149–180. 2. Jandt, U., You, C., Zhang, Y.H.-P., and Zeng, A.-P. (2013). Compartmentalization and Met- abolic Channeling for Multienzymatic Biosynthesis: Practical Strategies and Modeling Ap- proaches. (Springer Berlin Heidelberg), pp. 1–25. 3. Zhang, Y.-H.P. (2011). Substrate Channeling and Enzyme Complexes for Biotechnological Applications. Biotechnol. Adv. 29, 715–725. 4. Agapakis, C.M., Boyle, P.M., and Silver, P.A. (2012). Natural Strategies for the Spatial Opti- mization of Metabolism in Synthetic Biology. Nat. Chem. Biol. 8, 527–535. 5. Idan, O., and Hess, H. (2013). Origins of Activity Enhancement in Enzyme Cascades on Scaf- folds. ACS Nano 7, 8658–8665. 6. Conrado, R.J., Varner, J.D., and DeLisa, M.P. (2008). Engineering the Spatial Organization of Metabolic Enzymes: Mimicking Nature’s Synergy. Curr. Opin. Biotechnol. 19, 492–499. 7. Ovádi, J., and Srere, P.A. (1999). Macromolecular Compartmentation and Channeling. In Int. Rev. Cytol., D.E.B. and P.A.S. Harry Walter, ed. (Academic Press), pp. 255–280. 8. Conrado, R.J., Wu, G.C., Boock, J.T., Xu, H., Chen, S.Y., Lebar, T., Turnšek, J., Tomšič, N., Avbelj, M., Gaber, R., et al. (2012). DNA-guided Assembly of Biosynthetic Pathways Pro- motes Improved Catalytic Efficiency. Nucleic Acids Res. 40, 1879–1889. 9. Delebecque, C.J., Lindner, A.B., Silver, P.A., and Aldaye, F.A. (2011). Organization of Intra- cellular Reactions with Rationally Designed RNA Assemblies. Science 333, 470–474. 10. Dueber, J.E., Wu, G.C., Malmirchegini, G.R., Moon, T.S., Petzold, C.J., Ullal, A.V., Prather, K.L.J., and Keasling, J.D. (2009). Synthetic Protein Scaffolds Provide Modular Control Over 52 Metabolic Flux. Nat. Biotech. 27, 753–759. 11. Lee, H., DeLoache, W.C., and Dueber, J.E. (2012). Spatial Organization of Enzymes for Meta- bolic Engineering. Metab. Engin. 14, 242–251. 12. Moon, T.S., Dueber, J.E., Shiue, E., and Prather, K.L.J. (2010). Use of Modular, Synthetic Scaffolds for Improved Production of Glucaric Acid in Engineered E. coli. Metab. Engin. 12, 298–305. 13. Avalos, J.L., Fink, G.R., and Stephanopoulos, G. (2013). Compartmentalization of Metabolic Pathways in Yeast Mitochondria Improves the Production of Branched-chain Alcohols. Nat. Biotech. 31, 335–341. 14. Heinig, U., Gutensohn, M., Dudareva, N., and Aharoni, A. (2013). The Challenges of Cellular Compartmentalization in Plant Metabolic Engineering. Curr. Opin. Biotechnol. 24, 239–246. 15. You, C., Myung, S., and Zhang, Y.-H.P. (2012). Facilitated Substrate Channeling in a Self- Assembled Trifunctional Enzyme Complex. Angew. Chem. Int. Ed. 51, 8787–8790. 16. Torres Pazmiño, D.E., Riebel, A., de Lange, J., Rudroff, F., Mihovilovic, M.D., and Fraaije, M.W. (2009). Efficient Biooxidations Catalyzed by a New Generation of Self-Sufficient Baey- er-Villiger Monooxygenases. ChemBioChem 10, 2595–2598. 17. You, C., and Zhang, Y.-H.P. (2013). Annexation of a High-Activity Enzyme in a Synthetic Three-Enzyme Complex Greatly Decreases the Degree of Substrate Channeling. ACS Synth. Biol. 18. Bode, S.A., Minten, I.J., Nolte, R.J.M., and Cornelissen, J.J.L.M. (2011). Reactions Inside Nanoscale Protein Cages. Nanoscale 3, 2376. 19. Minten, I.J., Claessen, V.I., Blank, K., Rowan, A.E., Nolte, R.J.M., and Cornelissen, J.J.L.M. (2011). Catalytic Capsids: The Art of Confinement. Chem. Sci. 2, 358. 20. Cardinale, D., Carette, N., and Michon, T. (2012). Virus Scaffolds as Enzyme Nano-carriers. Trends Biotechnol. 30, 369–376. 21. Kim, E.Y., and Tullman-Ercek, D. (2013). Engineering Nanoscale Protein Compartments for Synthetic Organelles. Curr. Opin. Biotechnol. 24, 627–632. 22. Patterson, D.P., Schwarz, B., Waters, R.S., Gedeon, T., and Douglas, T. (2014). Encapsulation of an Enzyme Cascade within the Bacteriophage P22 Virus-Like Particle. ACS Chem. Biol. 9, 359–365. 23. Broz, P., Driamov, S., Ziegler, J., Ben-Haim, N., Marsch, S., Meier, W., and Hunziker, P. (2006). Toward Intelligent Nanosize Bioreactors: A pH-Switchable, Channel-Equipped, Func- tional Polymer Nanocontainer. Nano Lett. 6, 2349–2353. 24. Siti, W., de Hoog, H.-P.M., Fischer, O., Shan, W.Y., Tomczak, N., Nallani, M., and Liedberg, B. (2014). An Intercompartmental Enzymatic Cascade Reaction in Channel-equipped Polymer- some-in-polymersome Architectures. J. Mat. Chem. B 2, 2733. 25. Gaitzsch, J., Appelhans, D., Wang, L., Battaglia, G., and Voit, B. (2012). Synthetic Bio-nanoreactor: Mechanical and Chemical Control of Polymersome Membrane Permeability. Angew. Chem. Int. Edit. 51, 4448–4451. 26. Schoffelen, S., and van Hest, J.C.M. (2012). Multi-enzyme Systems: Bringing Enzymes To- gether In Vitro. Soft Matter 8, 1736. 27. Van Oers, M., Rutjes, F., and van Hest, J. (2014). Cascade Reactions in Nanoreactors. Curr. Opin. Biotechnol. 28, 10–16. 28. Louzao, I., and van Hest, J.C.M. (2013). Permeability Effects on the Efficiency of Antioxidant Nanoreactors. Biomacromolecules 14, 2364–2372.

53 29. Dergunov, S.A., Durbin, J., Pattanaik, S., and Pinkhassik, E. (2014). pH-Mediated Catch and Release of Charged Molecules with Porous Hollow Nanocapsules. J. Am. Chem. Soc. 136, 2212–2215. 30. Kim, K.T., Cornelissen, J.J.L.M., Nolte, R.J.M., and van Hest, J.C.M. (2009). A Polymersome Nanoreactor with Controllable Permeability Induced by Stimuli-Responsive Block Copoly- mers. Adv. Mater. 21, 2787–2791. 31. Kim, J.-K., Lee, E., Lim, Y., and Lee, M. (2008). Supramolecular Capsules with Gated Pores from an Amphiphilic Rod Assembly. Angew. Chem. Int. Edit. 47, 4662–4666. 32. Fornasiero, F., Park, H.G., Holt, J.K., Stadermann, M., Grigoropoulos, C.P., Noy, A., and Ba- kajin, O. (2008). Ion Exclusion by Sub-2-nm Carbon Nanotube Pores. Proc Natl. Acad. Sci. U.S.A. 105, 17250–17255. 33. Theil, E.C. (2011). Ferritin Protein Nanocages use Ion Channels, Catalytic Sites, and Nucle- ation Channels to Manage Iron/oxygen Chemistry. Curr. Opin. Chem. Biol. 15, 304–311. 34. Roux, B., Bernèche, S., Egwolf, B., Lev, B., Noskov, S.Y., Rowley, C.N., and Yu, H. (2011). Ion Selectivity in Channels and Transporters. J. Gen. Physiol. 137, 415–426. 35. Yeates, T.O., Jorda, J., and Bobik, T.A. (2013). The Shells of BMC-Type Microcompartment Organelles in Bacteria. J. Mol. Microbiol. Biotechnol. 23, 290–299. 36. Kinney, J.N., Axen, S.D., and Kerfeld, C.A. (2011). Comparative Analysis of Carboxysome Shell Proteins. Photosynth. Res. 109, 21–32. 37. Glasgow, J.E., Capehart, S.L., Francis, M.B., and Tullman-Ercek, D. (2012). Osmolyte-Medi- ated Encapsulation of Proteins inside MS2 Viral Capsids. ACS Nano 6, 8658–8664. 38. Capehart, S.L., Coyle, M.P., Glasgow, J.E., and Francis, M.B. (2013). Controlled Integration of Gold Nanoparticles and Organic Fluorophores Using Synthetically Modified MS2 Viral Capsids. J. Am. Chem. Soc. 135, 3011–3016. 39. Stonehouse, N.J., Valeg\aard, K., Golmohammadi, R., van den Worm, S., Walton, C., Stockley, P.G., and Liljas, L. (1996). Crystal Structures of MS2 Capsids with Mutations in the Subunit FG Loop. J. Mol. Biol. 256, 330–339. 40. Axblom, C., Tars, K., Fridborg, K., Orna, L., Bundule, M., and Liljas, L. (1998). Structure of Phage fr Capsids with a Deletion in the FG Loop: Implications for Viral Assembly. Virology 249, 80–88. 41. Dayan, J., and Wilson, I.B. (1964). The Phosphorylation of Tris by Alkaline Phosphatase. Bio- chim. Biophys. Acta 81, 620–623. 42. Schoch, R.B., Han, J., and Renaud, P. (2008). Transport Phenomena in Nanofluidics. Rev. Mod. Phys. 80, 839–883. 43. Muller, B.H., Lamoure, C., Le Du, M.-H., Cattolico, L., Lajeunesse, E., Lemaître, F., Pearson, A., Ducancel, F., Ménez, A., and Boulain, J.-C. (2001). Improving Escherichia coli Alkaline Phosphatase Efficacy by Additional Mutations inside and outside the Catalytic Pocket. Chem- BioChem 2, 517–523.

54 Chapter 4

Toward Therapeutic Protein Delivery with MS2-Encapsulated Protein-polymer Conjugates

4.1 Abstract As more knowledge has been gained about the nature of many human diseases, it has become clear that new avenues are needed to treat a number of conditions. This chapter describes work towards building structures using MS2 as a compartment for the encapsulation of therapeutic proteins and endosome-disrupting polymers, decorated on the external surface with targeting groups for delivery to cancer cells. Therapeutic proteins offer several potential advantages over traditional drugs. First, the ability of enzymes to carry out many turnovers of specific reactions enables a cellular response to be elicited from a relatively small number of molecules. Second, proteins are inherently biocompatible, so their breakdown products are unlikely to have off-target effects associated with many unnatural drugs. Delivery of a protein that functions in a known biochemical pathway could allow us to change the fate of a cell type without having to develop a small molecule that specifically acts on that single protein. Unfortunately, protein therapeutics have been hampered by a host of complications. Immunogenicity of many foreign proteins generally limits use to human, humanized, or poly(ethylene glycol)-derivatized proteins. Renal clearance of the proteins from the blood stream makes it necessary to use large proteins or assemblies. Lack of tissue specificity necessitates attachment of targeting groups, which can inhibit binding or activity of the therapeutic. Proteases can degrade the therapeutic before it can carry out its function. Finally, when a protein reaches its target cell, it is generally taken up into an endosome and degraded, rendering it ineffective. As a result, many researchers have sought to develop delivery vehicles that combine size, immune system evasion, protease resistance, scaffolding for targeting groups, and built-in endosome escape mechanisms to carry therapeutic proteins safely from injec- tion to the cytoplasm of their target cell. Here we describe efforts toward such a system by attaching negatively charged polymers to proteins and encapsulating the conjugates into an engineered MS2 variant. We also examine escape mechanisms based on triggered proteolysis of the capsid.

55 4.2 Targeted Delivery of Therapeutic Proteins Protein therapeutics have rapidly gained popularity in recent years for the treatment of a number of diseases.1 However, these drugs have been largely limited to extracellular or lysosmal targets due to several difficulties related to protein pharmacokinetics.2 Issues such as serum stability of proteins, clearance by the kidneys, immunogenicity, and degradation, can often be addressed by chemical or protein engineering methods.3 In particular, attachment of poly(ethylene glycol) chains has become a standard technique in the field of protein therapeutics to mitigate these prob- lems. In addition, immunogenicity can be addressed by computationally predicting highly immu- nogenic sequences on a protein surface and mutating them to more benign sequences; however this requires that the mutation does not affect protein function and requires testing each candidate sequence.4 Tissue- or cell specificity is another issue with many protein drugs which may be addressed chemically by conjugating a targeting group or genetically by fusion to a binding protein, such as an antibody. Finally, when the protein reaches its target cell, it surface hydrophilicity prevents it from crossing the membrane; instead, proteins are generally taken up through various endocytosis mechanisms and eventually degraded.5 Proteins have been delivered into the cytoplasm using genetic fusions of cell-penetrating peptides6,7 or proteins,8,9 membrane fusion of protein-containing liposomes with the cell membrane,10,11 or membrane-active polymers.12,13,14 However these mechanisms are generally not cell-specific, and require careful control to avoid off-target effects.15 Because of these complications, researchers have looked to macromolecular carriers to combine favorable delivery characteristics, environment-responsive release, and surface targeting groups. Many different types of vehicles have been developed for targeted delivery.16 Liposomes, polymer shells, and mixtures of the two are popular choices for their relative ease of production and the availability of biocompatible lipids and polymers. Several types of inorganic nanoparticles may also be used, some of which combine imaging capability along with delivery.17 Protein-based compartments, on the other hand, are relatively under-explored despite their inherent biocompatibility, versatility, and propensity for engineering.18 Viral capsids, in particular, are appealing for their uniform structures and established record in medicine.19 In nature, viruses naturally delivery genetic and protein content to the cytoplasm of target cells, inspiring many small molecule drug- and gene-delivery vehicles.20,21,22,23 A few viral capsids have been successfully engineered for untargeted protein delivery, such as polyoma virus-like particles,24 hybrid HIV-like particles,25 and lentivirus-like particles,26 all using native viral entry mechanisms. HIV-like particles were further engineered to encapsulate more diverse proteins and incorporate protein ligands on the surface, enabling even more effective delivery and opening up the possibility of targeting.27 Particles made from the bacteriophage T4 head and DNA packing machinery were also used to deliver proteins, nucleic acids, or both simultaneously, and were targeted to dendritic cells.28 Bacteriophage MS2 was engineered for targeted delivery of encapsulated RNA-conjugated molecules, including the protein ricin toxin A chain.29,30 The capsid was targeted using either transferrin or short peptides conjugated to the surface, and in the latter case endosome escape peptides were also included to facilitate cytosolic delivery. To date, no approach has combined the advantages of polymeric endosome escape mechanisms with the favorable properties of viral capsids for the targeted delivery of proteins. A large number of model therapeutic proteins have been tested for cytosolic delivery. Many of these proteins come from cell death pathways, particularly apoptosis pathways.31 These proteins are often aberrantly expressed, inhibited, or activated in cancer cells, contributing to these cells’ escape from normal growth control.32 Although several proteins from apoptosis pathways have 56 been engineered for delivery to cancer cells, the most popular candidates are the caspase enzymes, which are the major executioners of the apoptosis cascade.33,34 These enzymes signal for many of the cellular changes leading to apoptotic cell death by proteolytically activating downstream caspases. Some evidence suggests that they can also initiate survival signals to nearby cells.35,36 Caspases have also been a target for gene delivery and have been very well characterized; as a result, several engineered versions exist for therapeutic applications. For example, caspase 7 was engineered to be responsive to its redox environment.37 Caspase 3 has been engineered to be activated by light,38 HIV protease,39 or disulfide reduction,40 or even to be constitutively active and unrecognizable by endogenous regulation.41 As computational studies predict only 30 copies of these enzymes are required for efficient activation of apoptosis, delivery of these proteins is a major therapeutic goal for the treatment of several diseases, particularly cancer.42

4.3 Polymer Conjugates in Cytosolic Protein Delivery Several approaches exist for polymer-aided delivery of proteins to the cytoplasm.12,43 Poly- mers have been developed to be responsive to changes in their environment upon entering a cell, particularly change in pH or reduction potential. Generally, as macromolecule is endocytosed, it is taken first into an early endosome, followed by a late endosome, then on to the lysosome.5 Along this pathway, the pH drops from normal physiological (~7.4) to as low as pH 5. Changes in the redox potential also occur, although there is still debate on if and when the environment becomes reducing.44,45,46 One approach, called the “proton sponge” effect, relies on endocytosis of basic polymers.47 In an endosome, as the pH drops, the protons are absorbed by the polymer, but endosomal ATPases continue pumping protons in until the compartment reaches proper pH resulting in a buildup of salt and eventual osmotic lysis of the endosome. Other polymers rely on a direct, pH-responsive interaction with the cell membrane.48 These polymers are typically amphiphilic and negatively charged at neutral pH. As the pH drops, acidic groups on the polymer are protonated become more hydrophobic, triggering an interaction with the endosomal membrane. This often results in pore formation or lysis.49 If the polymeric delivery vehicle interferes with or blocks the activity of the therapeutic protein, it must also be removed at the proper time. This can be achieved by acid-cleavable linkers,50 redox responsive polymers,51 or enzymatically degradable carriers.52 Caspase 3 has been delivered via a variety of polymer-aided mechanisms.51,53,54,55,56 How- ever, none of these approaches combine cell-specific targeting, pH-responsive endosome escape, and encapsulation to shield both the toxic effects of the polymer and deleterious effects on the enzyme. Here we propose a protein-polymer hybrid drug delivery vehicle, outlined in Figure 1, which utilizes the serum stability,57 targeting capability,58 encapsulation,59 and potential immune system evasion of the MS2 capsid with endosome disrupting polymers60,61 conjugated to therapeutic proteins, such as caspase 3. We explore methods of attaching effective, inexpensive, versatile polymer derivatives to model proteins and demonstrate encapsulation of these in the MS2 shell. We also develop methods to encapsulate caspase 3 and incorporate protease recognition sites into the capsid toward triggered release of the protein cargo upon endocytosis.63

4.4 Conjugation of Negatively Charged Polymer Amphiphiles To facilitate both endosome escape and encapsulation of candidate proteins, we looked to a class of negatively charged, pH responsive polymer amphiphiles previously developed.48 In particular, poly(acrylic acid)-based amphiphiles have been developed for producing protein-sized 57 Figure 1. Protein delivery using MS2-polymer hy- MS2 Vehicle Polymer brid vehicles. In the proposed protein delivery vehicle, MS2 acts as both a protective shell and a scaffold for targeting groups. As the construct is endocytosed, Therapeutic Protein Protease the protein of interest and the polymer are released. The polymer then ruptures the endosomal membrane, releasing the protein to the cytosol. holes in membranes when the pH is decreased to ~6.60 We synthesized a derivative of this polymer, amp-PAA, by amidating ~18% of the acids with dodecylamine and ~2% with Rhodamine B piperazine amide to facilitate imaging using modified published protocols (Figure 2A).64,65 The polymer was then tested for lysis of red blood cells, a common predictor of endosomolytic potential.66 At pH 6.0 the polymer achieved significant lysis over background at low concentrations, while at pH 7.4 significantly higher concentrations were needed for the same effect, indicating the polymer was behaving as expected (Figure 2B). While the parent PAA was very effective in reassembling the capsid, it was unclear if the detergent nature of the amp-PAA would interfere with the reassembly process. Therefore, the polymer was tested for reassembly of MS2 coat proteins (CP) at several concentrations under similar conditions as previously described (Figure 2C).59 At low concentrations the amp-PAA initiated similar reassembly to its parent PAA; however significant inhibition was seen at higher concentrations, suggesting a concentration dependent interaction with the protein or a micellar effect. Moderate polymer concentrations in the reassembly reaction could yield acceptable amounts of encapsulated polymer for further characterization. The free and encapsulated polymers were examined by dynamic light scattering (DLS). As shown in Figure 2D, the hydrodynamic radius of the free polymer increased upon lowering of the pH. The MS2- encapsulated polymer was consistently the size of empty MS2, as expected. These reassemblies were purified and incubated with HCC-1954 and MCF7-clone 18 breast cancer cells. As shown in Figures 2E and 2F, both the free polymer and the untargeted, MS2-encapsulated polymer were able to localize to the cell cytoplasm.

58 A) O B) Hemolysis with amp-PAA C) WT MS2 Reassembly vs. amp-PAA 80 40 N H 10 - 70 35

O O y l 60 b 30 N O m e s f max) s

N o 50 25

a % (

R e

O 40 20 s t i n s e y l 30 c 15 r + o e N m N O P e 20 10 H MS2 pH 7.4 Pol pH 7.4 5 10 MS2 pH 6.0 Pol pH 6.0 0 0 0 0.02 0.04 0.06 0 20 40 60 [Polymer] (mg/ml) [amp-PAA] (μM)

D) E) F) DLS of MS2 and amp-PAA 25 MS2 pol pH 8 MS2 pol pH 5 20 WT MS2 pH 8 WT MS2 pH 5 % 15 Polymer pH8 e

m Polymer pH 5 u l o 10 V

0 4 40 400 d (nm) Figure 2. Amp-PAA characterization. a) The polymer amp-PAA is a random copolymer of acid groups, dodecylamine groups, and rhodamine B piperazine amide groups. b) Hemolysis using free or encapsulated amp-PAA in MOPS pH 7.4 or MES pH 6.0. Hymolysis is higher with free polymer at pH 6.0 and is shielded by MS2. c) Reassembly of WT MS2 using amp-PAA as analyzed by SEC. Reassembly is most efficient at moderate polymer concentrations. d) Dynamic light scattering (DLS) of WT MS2 (untreated or encapsulated amp-PAA) or free amp-PAA. The MS2 samples are indistiguish- able at just below 27 nm (dotted red line). e) and f) MCF-7 clone 18 cells incubated with free (e) or MS2-encapsulated (f) amp-PAA. Red represents polymer rhodamine fluorescence, and blue is Hoescht nuclear stain.

Finally, to demonstrate that we can use this polymer to encapsulate proteins in the capsid, we tried attaching it to a model protein, GFP. Several bioconjugation reactions were attempted, including direct amide formation between the polymer and protein lysines,67 hydrazide formation on the polymer followed by hydrazone formation on transaminated protein,68 and simple co-encapsulation potentially mediated by the 6xHis tag on the protein and the nickel-binding properties of the polymer, but none of these methods proved to be effective. Further complicating the method was the propensity of the polymer to irreversibly adsorb to protein purification resin, including anion exchange, size exclusion, and Ni2+/NTA resins, limiting the amount of molar equivalents of polymer possible. Clearly, a reaction was needed that could attain high levels of modification with minimal equivalents of polymer over biomolecule, such as the oxidative coupling (OC) of anilines and electron-rich arenes developed in the Francis lab.69 Motivated by the successful encapsulation and delivery of the amp-PAA polymer, we synthesized a variant containing a similar amount of pendant dodecylamine amide groups along with a small amount (~2 %) of 3-nitrotyramine amide groups (amp-PAA-NT). The nitrotyramine groups could then be reduced to 3-aminotyramine groups using sodium dithionite in phosphate buffer 59 NH2 pH 6.5 (amp-PAA-AT). This polymer OH O OH was able to similarly reassemble the MS2 OMe O - O capsid under normal conditions. Unfor- -2e, -MeOH + + tunately this route suffered several com- + H2O -H /+H N H plications. First, the carbodiimide-based synthesis of the polymer was difficult to reproduce accurately due to solubility is- OH OH sues, and other methods of activating the taut. OH -2e- O acidic groups, such as treatment with thi- -H+/+H+ onyl chloride or N,N,N′,N′-tetramethyl- N N H O-(N-succinimidyl)uronium tetrafluorob-

Figure 3. Probable mechanism of cxidative coupling be- orate (TSTU) failed to elicit the desired tween anilines and methoxyphenols. reactivity. Second, the dithionite reduction often failed to give amp-PAA-AT, and previously mentioned purification issues made removal of residual reductant tenuous. Finally, amp-PAA-AT was only stable for a few hours, requiring fresh preparation for each reaction at- tempt. These complications led us to explore both a new polymer backbone and a new OC partner. For more facile bioconjugation, we replaced the 3-aminotyramine derivative with a 3-methoxytyramine derivative. Currently under development, o-methoxyphenols undergo a similar periodate-mediated oxidative coupling reaction to the previously developed o-aminophenols, with the added benefit of bench stability (Figure 3). This allows direct use of the oxidative coupling partner, abolishing the need for dithionite reduction before every reaction. The reaction, shown in Figure 4, takes place in a matter of minutes in mild buffer conditions and gives similar yields to the aminophenol OC reaction. To mitigate issues arising from PAA activation and purification, a very similar polymer backbone, poly(isobutylene-alt-maleic anhydride) (PIBMA), was used to create similar random copolymers of dodecylamine- and OC partner-modified amphiphiles (Figure 4A). This backbone, which is even less expensive than PAA and comes pre-activated for amide-coupling reactions, has been used to create similar amphiphiles for coating of quantum dots for solubilization, functionalization, and cellular delivery.61,70 We hypothesized that a hydrolized PIBMA amphiphile would have similar negative charge and endosomolytic properties to the amp-PAA with the added benefits of lower expense and ease of synthesis. Indeed, the PIBMA could be functionalized with dodecylamine and oxidative coupling partners easily and reproducibly without the solubility issues posed by the amp-PAA synthesis. We tested the polymer modified with ~17% dodecylamine groups and ~4% 3-methoxytyramine groups (amp-PIBMA-MT) for the reassembly of T19 p-aminophenylal- anine (pAF) substituted MS2 CP under ideal reassembly conditions, and found that the polymer was highly effective at reassembling the capsid at moderate concentrations but completely inhibited reassembly at higher concentrations, again consistent with concentration-dependent protein binding or micelle formation (Figure 4B). We also tested the hemolytic properties of amp-PIBMA- MT and found no difference between pH 6.0 and pH 7.4 conditions, indicating this polymer could disrupt membranes without regard to pH (Figure 4C). This does not necessarily pose a problem for MS2-mediated delivery as the polymer will be shielded from membranes until it reaches its destination; however, if pH response is desired it might be accomplished by creating polymer variants with fewer dodeclyamine modifications, lowering the overall hydrophobicity of the polymer. Finally, the PIBMA derivatives were tested for periodate-mediated conjugation to ani-

60 A) H B) pAF MS2 Reassembly vs. amp-PIBMA-MT -O O O N 10 120

100 ) % (

y O NH O O- l 80 b m e

s 60 s a e

amp-PIBMA-NT: R = -NO2 R 40

amp-PIBMA-AT: R = -NH2 amp-PIBMA-MT: R = -OMe 20 R OH 0 0 50 100 [amp-PIBMA-MT] (μM) C) amp-PIBMA-MT Hemolysis D) 60 MES pH 6.0 50 MOPS pH 7.4 max)

f o 40 % (

s i

s 30 y l o m

e 20 H

10 AT - + - + - + + - 0 MT - - + - + - + - 0 0.2 0.4 0.6 - - - + + - + + NaIO4 [amp-PIBMA-MT] (mg/ml) - - - - - + - - K3Fe(CN)6 Figure 4. Characterization of amp-PIBMA derivatives. a) amp-PIBMA derivatives are random co-polymers of isobutylene groups, acidic groups, dodecylamine groups, and either 3-nitrotyramine (NT), 3-aminotyramine (AT), or 3-methoxytyramine (MT) groups. b) Reassembly of T19pAF MS2 using amp-PIBMA-MT in Bis Tris pH 6.0. The caspsid re- assembles in very high yields in moderate amounts of polymer, and virtually no reassembly at high or low concentrations. c) Hemolysis using amp-PIBMA-MT. The polymer lyses red blood cells equally well at pH 6.0 and 7.4. d) SDS-PAGE analysis of oxidative coupling of aniline-derivatized GFP and amp-PIBMA derivatives. The polymer (5 equivalents relative to GFP) was incubated with the protein in the presence of 5 mM oxidant for 10 minutes, followed by quenching with loading buffer. line-modified GFP derivatives. The polymer modified with ~18% dodecylamine and ~4% 3-nitrotyramine (amp-PIBMA-NT) was reduced with dithionite to the 3-aminotyramine derivative (amp-PIBMA-AT) and attached to GFP with an N-terminal aniline installed by previously developed methods.68 The modification using either sodium periodate or potassium ferricyanide as the oxidant was highly effective with only 5 equivalents of the polymer as shown in Figure 4D. The polymer amp-PIBMA-MT was directly conjugated to the same modified GFP. When sodium periodate was used as the oxidant, this reaction was also very efficient with only 5 equivalents of polymer to GFP. These polymers can be used interchangeably for reassembly experiments, protein modification, and, potentially, cytoplasmic delivery of cargo. If facile reactions are required the methoxyphenol derivative can be used, and if the cargo is periodate-sensitive, the aminophenol derivative provides a useful alternative.

4.5 Encapsulation of Protein-polymer Conjugates To develop targeted MS2 vehicles containing protein-polymer conjugates, we analyzed the

61 A) B) SEC of pAF-MS2-encapsulated C) DLS of MS2/GFP-amp-PIBMA GFP conjugated to amp-PIBMA-MT 35 )

9 MS2/GFP-polymer GFP-amp-PIBMA 1 5 / GFP-polymer 30 MS2/GFP- 8 8

4 amp-PIBMA

) 25 ( % e ( c 20 e e n c u m l

- - 1x 5x 1x 3x 5x e s 15 o MT r o V

NaIO4 - + - - + + + u l 10 F

F 5 G 0 0 5 10 15 4 40 400 Time (min) Diameter (nm)

Figure 5 Encapsulation of amp-PIBMA-MT. a) Minimum concentrations were determined to efficiently couple the polymer to GFP and proceed directly to encapsulation. The polymer very efficiently reacts with when used in 3-fold excess over GFP. b) SEC analyses of free- and T19pAF MS2-encapsulated GFP-amp-PIBMA-MT. Elution was monitored by GFP fluorescence (ex. 488 nm/em. 519 nm). A significant amount of GFP elutes with the capsid (~5.2 minutes). The free conjugate has significant unmodified GFP flurescence (~8.3 minutes). c) DLS of purified free and encapsulated -poly mer. Encapsulated polymer is approximately 27 nm in diameter, similar to untreated T19pAF MS2. ability of the GFP-amp-PIBMA-MT conjugates to reassemble MS2 under the conditions described for unconjugated amp-PIBMA-MT. It was clear by the color of the conjugation reaction that the polymer interfered with GFP fluorescence. Upon further investigation, it was found that the in- terference was pH dependent, as a mixture of the polymer and protein showed decreased fluorescence and a shift in absorbance at lower pH (~6), while nearing normal at higher pH (~9) (data not shown). This shift appeared to be reversible, indicating the polymer did not completely denature the GFP. This could be caused by a pH-dependent structural change in the polymer, as has been suggested for the similar amp-PAA derivatives, or acid-destabilization of GFP, which irreversibly denatures at ~pH 5.5. Likely a combination of these factors contributes to the decrease in signal from the protein; however as the protein showed significant fluorescence at physiological pH, we proceeded with the encapsulation. For encapsulation studies, the amp-PIBMA-MT variant was attached to GFP modified with 3-(4-aminophenyl)propionic acid as described.68 As shown in Figure 5A, this reaction gave very high yield with only three equivalents of polymer in ten minutes. Because of the difficulties purifying the protein, polymer, and conjugates from each other, the reaction was quenched with excess ethylene glycol and added directly to disassembled pAF CP, using polymer concentrations and buffer conditions optimal for reassembly. After reassembly, the protein was partially purified by precipitation and filtration through a plug of DEAE-sepharose and analyzed by size exclusion chromatograpy (SEC). As shown in Figure 5B, the reassembled MS2 eluted at the expected time, co-eluting with a significant amount of GFP fluorescence, indicating the conjugate was trapped inside. After scaling the reassembly up and purifying by anion exchange and SEC, the reassembled capsid looked identical to untreated 27 nm MS2 by DLS (Figure 5C). Interestingly, the unencap- sulated conjugate showed significant green fluorescence eluting as free GFP, though only a small percentage of free protein was present by Coomassie. This suggests that the fluorescence of the conjugate is indeed inhibited, even at neutral pH. These assemblies were attached to antibodies and incubated with HCC-1954 cells, but very little fluorescence was observed (data not shown). Because of the limited amount of sample, we were unable to characterize the final conjugates

62 A) B) Caspase 3 assay, DEVD-AFC C) 120 O + Na - 100 O O O H O O ) 80 H CF3

N N μ M (

N N N ] 60

H H H C

O O F

O A [ 125 μM DSF Na+ 40 -O O HO 2.5 mM ZnCl2 Enzyme + EDTA/DTT 20 No Enzyme CF3 Caspase 3 Bz-DEVD-COOH + 0 H2O 0 200 400 600 800 1000 O 1 2 3 4 5 6 7 H2N O Time (sec) Figure 6. Analysis of caspase 3 activity. a) The caspase reaction was monitored by release of 7-amino-4-trifluoromethylcoumarin (AFC) from a peptide substrate by fluorescence (ex. 400 nm/em. 505 nm). b) The caspase is highly active when

metal chelators and reducing agents are present, but activity is inhibited by disulfiram (DSF) or ZnCl2. c) SDS-PAGE analysis of caspase 3 degradation of WT and Csub3 MS2. Capsid substrates, either untreated or disassembled into coat protein (CP), were incubated with the caspase under activating conditions. No degradation was observed. Lanes: 1) WT CP + caspase 2) Cusb3 CP + caspase 3) WT capsid + caspase 4) Csub3 caspsid + caspase 5) Csub3 CP + caspase + DSF

6) Csub3 CP + caspase + ZnCl2 40 uM MS2, 800 nM caspase 3. in-depth. Future experiments could explore polymer hydrophobicity and backbone effects on encapsulation and cellular delivery and use fluorophores less sensitive to the detergent effects of the polymer.

4.6 Therapeutic Protein Encapsulation In conjunction with the development of an appropriate polymer for encapsulation and endosome escape, we sought to establish a model therapeutic protein for encapsulation and delivery. As mentioned above, the caspase family has been a popular choice because of its high activity and frequent aberrant expression in cancer. In addition, we hypothesized that the caspase enzymes, as a family of proteases, could provide a triggered release mechanism by activating the proteolytic ac- tivity40 and degrading the capsid. It has also been shown that the MS2 capsid tolerates a variety of insertions and mutations in its primary sequence; therefore we hypothesized that if the wild-type capsid cannot be degraded, mutations could be made to add a proteolytic recognition site. Finally, we wished to determine if the previously developed encapsulation conditions were suitable for loading caspase enzymes into MS2, or if optimal caspase buffer conditions were suitable for MS2 reassembly. We cloned the genes for caspase 8 and caspase 3 from human cDNA into expression vectors. The wild-type proteins are both expressed with N-terminal localization sequences, and caspase 8 has an additional N-terminal binding domain; these were excluded from the final constructs.63 In addition, both of these enzymes require proteolyisis to take their active form—caspase 3 activates itself when expressed in high amounts in E. coli, but caspase 8 was expressed as separate subunits in inclusion bodies, which were solubilized, mixed, and refolded. Caspase 3 expressed in E. coli and purified by either an appended 6xHis tag or a Strep II tag was active on a synthetic peptide substrate (Figures 6A and 6B). Refolded caspase 8, however, was found to be only slightly active; therefore we chose caspase 3 as our model therapeutic protein. Several engineered variants of caspase enzymes are available with potential benefits in cancer therapy. A mutant of the enzyme, D175A/V266E (AEC3), is active in its uncleaved form

63 and is unaffected by X-linked inhibitor of apoptosis (XIAP).41 This could potentially bypass a common apoptosis escape mechanism of cancer, overexpression of XIAP. We made several variations of caspase 3 without its first 29 amino acids (Δ29) and AEC3, incorporating 6xHis or Strep II tags for purification and/or negative tags for downstream encapsulation. While we found that N-terminal negative tags knock expression down to an almost undetectable level, a C-terminal negative tag followed by the Strep II tag (C3NS) allowed acceptable expression in E. coli along with ease of purification using aStrep -tactin column. To examine degradation of MS2 capsid variants, we decided to use 6xHis-tagged caspase 3 Δ29, whereas for encapsulation studies, we chose C3NS. However, the therapeutic potential of AEC3 should not be dismissed, and future studies could further optimize expression of this enzyme for drug delivery purposes. We tested caspase 3 activity in several conditions to verify its activity. Previous reports suggest that caspases can be specifically but reversibly inhibited by disulfide formation using tetraethylthiuram disulfide (disulfiram, DSF) or zinc (II).40,72 We therefore tested the activity in the presence and absence of these inhibitors. Shown in Figure 6B, the enzyme is highly active in the presence of EDTA and DTT, while there is nearly undetectable activity with either ZnCl2 or DSF present. The specific activity was 0.89 μmol/min/mg enzyme. We also tested the activity of the enzyme in the presence of MS2 and found no effect of the capsid on activity. Interestingly, we incubated the enzyme with both intact and disassembled wild-type MS2 and observed no degradation of the capsid (Figure 6C). The only strict rule for caspase cleavage sites is that an aspartate is present followed by a small amino acid; however the presence of such a sequence is not necessarily sufficient for cleavage.73 Enzyme activity on a protein substrate is also increased by flexibility in the cleavage site and a glutamate two residues before the target aspartate; minor contributions of other surrounding amino acids arrive at the consensus sequence DEVD/x, where x is small and proteolysis occurs between D and x. The MS2 sequence has one site that could function as a substrate: D114/G115, which occurs as a loop on the external surface of the capsid. However, the presence of a preceding positive charge (K113), hydrophobic residue (L112), or a stabilizing salt bridge (R56 on adjacent subunit) attenuates caspase activity on this sequence. From one perspective it is good that the caspase has no background activity on the capsid, indicating our vehicle is stable enough to load and carry its cargo without the need of reversible inactivation. Nevertheless, we hypothesized we could engineer the capsid interior surface to contain a cut site for triggered degradation of the capsid by the caspase from the inside out. The MS2 capsid features two loops on its interior surface potentially accessible to an encapsulated caspase: amino acids 26-29 (ANGV) and amino acids 50-55 (QSSAQ). We decided to target the latter “EF loop” because of its larger size, sequence similarity to the caspase 3 consensus (glutamines vs. glutamate residues), and absence of glycines, which can be important for abnormal loop conformations. The EF loop, shown in Figure 7A, extends into the interior of the capsid and forms relatively few contacts with other residues, suggesting it could be both flexible and tolerant to insertions or mutations to caspase cut sequences. We made several mutants of the capsid with potential caspase 3 cut sites in the EF loop, shown in Figure 7A and Table 1. We expressed mutants 1-6 in small scale, isolated the soluble and insoluble cell lysate fractions from each, precipitated the soluble fractions with PEG 8k, resuspended in buffer and spun down any irreversibly precipitated proteins. Soluble expression implied proper folding of the capsid coat protein, while we took reversible precipitation with PEG 8k to imply proper assembly of the capsids. Shown in Figure 7B and 7C, mutants 3,4, and 5 met the criteria for properly assembled capsid. Mutants 3 and 5 were selected because of their similarity to the caspase 3 consensus sequence, and were expressed and

64 A) B)

1C 1IC 1S 1IP 2C 2IC 2S 2IP 3C 3IC 3S 3IP

Csub1 .....QSDEVDSAQ Caspase 3 C) Csub2 ...... QSESDAQ Consensus: Csub3 ...... ESDAQ DEVD/x Csub4 ...... QSDAQ x = G, A, S Csub5 ...... DESDAQ Csub6 ...... ESDAQ

Gel Key: C = Cell pellet IC = Insoluble fraction S = Soluble PEG precipitate IP = Insoluble PEG precipitate

4C 4IC 4S 4IP 5C 5IC 5S 5IP 6C 6IC 6S 6IP

Figure 7. Development of potential caspase substrate mutants. a) View of the interior surface of the MS2 capsid. Residues 50-55 (EF loop, sequence QSSAQ) are colored red. Mutations were made to this loop to produce Csub1-6 capsids. Mutations are indicated with bold formatting. b) and c) SDS-PAGE analysis of Csub capsid expression and partial purification. Csub3, Csub4, and Csub5 expressed soluble, well-formed capsids. purified on large scale. The capsid expressed in somewhat lower yield than the wild-type capsid, but tens of milligrams of protein were readily isolated from 1 L of expression culture. Interestingly, these mutants contained much less E. coli nucleic acid material upon purification, likely due to charge repulsion between the inserted aspartates and glutamates and the negatively charged RNA. We incubated these mutants, either assembled or disassembled by treatment with acetic acid, with purified caspase 3. The assembled caspsid was still resistant to degradation. Surpris- ingly, the disassembled coat protein was also resistant despite having potential caspase cut sites exposed. We wondered if inserting the full capsase consensus sequence would make the coat protein a better substrate; therefore mutant 7 was constructed and purified (Table 1). Similar degradation experiments were performed, and again no proteolysis of the capsid was observed. We then attempted to encapsulated the caspase mutant C3NS in wild-type, Csub3, Csub5, and Csub7 capsids. The negatively tagged enzyme was able to reassemble Csub3 and Csub5, but not Csub7, and again no proteolysis was observed. While these experiments were unsuccessful in creating a proteolysis-triggered escape mechanism for therapeutic cargo, important lessons were learned about the stability and flexibility of the capsid shell. Because similar proteases, cathepsins, are some of the main degradative enzymes in endolysosomal protein degradation, it may be that the stability of capsids slow their

65 degradation, which is important for the delivery of biomolecules of any sort. On the contrary, this stability allows proteolytic cargo to be encapsulated, which could be useful for delivery of many signaling proteases, such as caspases, cathepsins, granzymes, and others. Future studies could use protein engineering approaches to develop triggered disassembly in capsids or use concepts developed herein for delivery of other biomolecules, such as RNA, DNA, or enzymes.

4.7 Materials and Methods General Methods. Unless otherwise noted, reagents were used as received from commercial sources. In all experie- ments, water was deionized using a MilliQ system (Millipore). Chemicals were obtained from Fisher Chemical except PAA, PIBMA, dodecylamine, 3-methoxytyramine, tyramine, Bz-DEVD- AFC, Ac-IETD-AFC, TSTU, DEAE Sepharose FF, and nitric acid (Sigma), and EDC·HCl (Chem- Pep, Inc.). Molecular biology reagents were obtained from New England Biolabs. Instrumentation and Sample Analysis. High Performance Liquid Chromatography (HPLC) was performed using an Agilent 1100 series HPLC using ST buffer (50 mM Tris pH 7.2, 100 mM NaCl) or sodium phosphate buffer (10 mM, pH 7.2) as mobile phase. All size exclusion chromatography was performed using a Biosep SEC-4000 column (Phenomenex). Liquid Chromatography/Mass Spectrometry (LC/MS) was performed using an Agilent 1200 series HPLC connected to a Agilent 6224 Time-of-Flight (TOF) LC/ MS system equipped with a Turbospray ion source. For purification, samples were filtered through 0.22 µm filters (Millipore). HPLC-based quantification was performed by comparing integrated fluorescence intensity of tryptophan (ex. 280/ em. 330) over time of untreated MS2 to reassembled samples. Dynamic light scattering (DLS) was performed using a Zetasizer Nano (Malvern Instru- ments). SDS-PAGE analysis of all protein samples was carried out on a Mini Protean apparatus (Bio-Rad, Hercules, CA) using 12.5% poly(acrylamide) gels prepared according to the manufacturer’s specifications. Samples were heated to 95o C in the presence of Laemmli buffer containing β-mercaptoethanol for 3-5 minutes. Gels were run at 125 V for approximately 75 minutes, stained with Coomassie R-250, and imaged using a Chemidoc imager (Bio-Rad, Hercules, CA). UV-vis spectroscopy was performed using a Nanodrop 2000C (Thermo Scientific). Nucleic acid-free MS2 was quantified using ε = 1 cm∙ml/mg. Experimental Procedures Synthesis of dodecylamine- and rhodamine-modified poly(acrylic acid) (PAA) derivative. PAA derivatives were synthesized using a modification of previously published procedures.64,65 To an aqueous solution of PAA sodium salt (0.125 g, average molecular weight 5100) was added dodecylamine (0.078 g, 0.42 mmol, approximately 0.25 equivalents to acidic groups). Precipitate often formed, which could be mitigated somewhat by addition of NaCl. The pH was adjusted to 6 with HCl and rhodamine B piperazine amide was added (0.035 g, 0.068 mmol, approximately 0.04 equivalents to acidic groups). Finally, N-(3-dimethylaminopropyl)-N’-ethylcarbodiimide hydrochloride (EDC·HCl, 0.194 g, 1 mmol) was added and the mixture was stirred for 4 hours at room temperature. The polymer was then dialyzed twice against 1 L of 50 mM HCl, twice against 1 L

66 of 50 mM NaOH, and several times against deionized water, and finally lyophilized to yield 137 mg of purple solid. Synthesis of 3-nitrotyramine. Synthesis of 3-nitrotyramine was previously reported. To an aquous soluion of 3-hydroxyphenylpropionic acid (tyramine, 900 mg, 6.3 mmol) in 50% acetic acid was added a solution of fuming nitric acid (1.5 ml, 33 mmol). The solution quickly turned bright or- ange. After 15 minutes the reaction was added to ice water, filtered, and dried to yield 392 mg 1 (34%) of yellow powder. H NMR (400 MHz, d6-DMF): 11.14 (br s, 1H), 8.04 (s 1H), 7.97 (d, 1H, J = 2.2), 7.61 (dd, 1H, J = 8.6, 2.3 Hz), 7.23 (d, 1H, J = 8.5 Hz), 3.43 (br sextet, 2H, J = 6.4), 3.11 (t, 2H, J = 7.4). Synthesis of dodecylamine-, methoxyphenol-, and nitrophenol-modified poly(isobutylene-alt- maleic anhydride) (PIBMA) derivatives. A flame dried 500 ml round bottom flask was charged with 200 mg of PIBMA (average molecular weight 6000) in 100 ml dry THF. Dodecylamine (Sigma, 36 mg, 0.195 mmol, approximately 0.15 equivalents to acidic groups) and N,N-diiso- propylethylamine (DIPEA, 0.39 mmol, 68 μl) were dissolved in a small amount of dry THF and added dropwise. The solution was stirred 1 hour at 60 oC. The temperature was reduced and 13.2 mg of 3-methoxytyramine HCl (0.065 mmol, approximately 0.05 equivalents to acidic groups) or 3-nitrotyramine (12 mg, 0.065 mmol) along with 20 μl of DIPEA (0.114 mmol) dissolved in THF were added dropwise. The solution was allowed to stir overnight at room temperature. The solvent and excess DIPEA were removed under reduced pressure and the resulting residue was dissolved in aqueous bicarbonate and the pH was adjusted to 8. The solution was dialzyed against several changes of deionized water over multiple days and lyophilized to yield 112 mg of amp-PIBMA- MT. Yield of the reaction was not determined due to multiple overlapping peaks in the 1H-NMR spectrum. Polymer-induced hemolysis assays. Release of hemoglobin from red blood cells was used to quantify the membrane disrupting properties of hydrophobically modified polymers.66 Defibrin- ized sheep blood cells (Hemostat Laboratories) were washed with normal saline (5 mM sodium phosphate pH 7.8, 154 mm NaCl) three times by centrifugation at 600 x g for 10 minutes. After the final spin the cells were gently resuspended in a 2x buffered solution at the desired pH at approximately 3 times the orignal blood volume (2x MES: pH 6.0, 50 mM MES, 300 mM NaCl; 2x MOPS: pH 7.4, 50 mM MOPS, 300 mM NaCl). Cells were added in 100 μl aliquots to a 96-well plate, and aqueous polymer solutions were added to the desired concentrations and a final volume of 200 μl. The plate was incubated with shaking at 37 oC for two hours. Complete lysis was acheived by addition of Triton X-100 to a final concentration of 0.1% v/v. Cells were pelleted by centrifugation at 600 x g for 10 minutes, and the supernatants were diluted appropriately into a new 96-well plate. Hemolysis was determined as a percentage of Triton X-100-lysed samples by absorbance at 450 nm. Expression and purification of MS2. Wild-type and T19pAF MS2 were expressed and purified as previously described.71 The plasmid pBAD-myc-his containing the gene encoding wild-type capsid in was transformed into DH10B cells and expressed in 2xYT media by induction with 0.2% arabinose. The plasmid pBAD-myc-his containing the gene encoding amber stop codon mutant MS2 T19tag was transformed into DH10B cells already harboring amber suppression plasmid pDULE-pAF and expressed by autoinduction in defined media supplemented with p-aminophe-

67 nylalanine as described.21 The capsids were purified by anion exchange followed by precipitation with PEG 8k, followed by size exclusion chromatography. The final capsid SEC fractions were precipitated with 50% ammonium sulfate, resuspended in 10 mM phosphate buffer pH 7.2, and desalted as needed. Cloning, expression, and purification of GFP. All GFP experiments in this chapter were performed on a variant of the protein featuring an N-terminal sequence AKT followed by a Strep tag (WSHPQFEK) and a C-terminal 6xHis tag. The gene encoding monomeric enhanced GFP (mEGFP) was amplified from pTXB1-mEGFP62 using forward primer 5’-TATCATATGGCTA- AAACCTGGAGCCACCCGCAGTTCGAAAAAGGCGAGGAGCTGTTCACC-3’ and reverse primer 5’-AAATGTAAGCTTTTATTAGTGATGATGATGGTGATGCTTGTACAGCTCGTC- CATGCC-3’, which added a 5’ NdeI site followed by the AKT sequence and Strep tag and a 3’ 6xHis tag followed by two stop codons and a HindIII site. The gene was cloned into pET24b and transformed into DH10B for storage. For expression, the plasmid was transformed into BL21(DE3)* Codon+RIL cells, which were used to start an overnight culture in LB supplemented with 50 mg/L kanamycin and 34 mg/L chloramphenicol. The overnight culture was subcultured o 1:200 into 300 mL of 2xYT with the same antibiotics and grown at 37 C to an OD600 of 0.6, at which time IPTG was added to a final concentration of 1 mM. The induced cells were allowed to express protein overnight at 37 oC. The bright green cells were havested and resuspended in buffer (25 mM sodium phosphate pH 7.2, 300 mM NaCl, 20 mM imidazole) and lysed by sonication on ice. The clarified lysate was then applied to a 1 mL Ni2+/NTA column (GE Healthcare) and washed with 50-75 mLs of buffer containing 20 mM imidazole. The protien was the eluted with the same buffer with 250 mM imidazole and dialyzed against several changes of 25 mM phosphate buffer pH 7.2 in the dark. Conjugation of PAA derivatives to GFP using EDC and NHS. Yields of PAA-protein cojugates using standard amide coupling reagents (EDC/NHS, EDC/HOBT, TSTU etc.) were highly variable; an example conjugation protocol is outlined here.70 The PAA derivative (1 mM in water, 10 μl) was mixed with N-hydroxysuccinimide (NHS, 100 mM, 1ul), 2-(N-morpholino)ethanesulfo- nate buffer (MES, pH 6.0, 0.5 M, 2 μl), and 5 μl H2O. A solution of 0.1 M EDC·HCl was freshly prepared and 2 μl was immediately added to the polymer and mixed. To prevent EDC reaction with protein (particularly reactive thiols on caspases), the reaction was quenched with 1 μl of 0.4 M β-mercaptoethanol. The activated PAA (5 μl) was then immediatly added to GFP (10 μl, 0.1 mg/ ml) in 0.1 M sodium bicarbonate buffer pH 9 and allowed to react several hours. Conjugates were anaylzed by SDS-PAGE. Aniline functionalization of GFP amines. Solutions of 3-(4-aminophenyl)propionic acid (41.3 mg/ml, 250 mM in N,N-dimethylformamide), EDC·HCl (60 mg/ml, 313 mM in 0.5 M phosphate buffer pH 6.5), and N-hydroxysulfosuccinimide sodium salt (sulfo-NHS, 261 mg/ml, 1.2 M in

H2O) were mixed 5:4:1 and allowed to react 15 minutes at room temperature. Unreacted EDC was quenced by addition of excess β-mercaptoethanol, and 10 μl of the reaction mixture was added to 90 μl of GFP (1 mg/ml in 20 mM bicarbonate buffer pH 9.0). The reaction mixture was incubated at room temperature for several hours to overnight and purified by NAP-5 desalting columns (GE Healthcare) followed by several rounds of spin concentration against a 10 kDa MWCO centrifugal filter (Millipore). The number of aniline groups installed was found to be 1-7 per protein by electrospray ionization-time of flight mass spectrometry (ESI-TOF MS).

68 Aniline functionalization of GFP N-terminus. For site-selective oxidative coupling, and aniline group was added to the N-terminus by transamination followed by oxime formation with an aminooxy aniline derivative.68 Briefly, transamination was carried out using pyridoxal 5’ phosphate (PLP) and N-terminal AKT GFP. The PLP was first dissolved to a concentration of 200 mM in 200 mM carbonate buffer pH 9.0, resulting in a 2x stock solution at approximately pH 6.5. This was then mixed in eqaul volume with GFP (2 mg/ml) in 50 mM sodium phosphate buffer pH 6.5 and incubated 1 hour at 37 oC. The transaminated GFP was desalted using a NAP-5 desalting column (GE Healthcare) followed by several rounds of spin concentration against a 10 kDa MWCO centrifugal filter. The transaminated GFP was only slightly soluble at pH 6.5 or lower, so 10 mM carbonate pH 9.0 was used for desalting and concentrating. Oxime formation was performed by adding the aminooxy aniline (pH adjusted using saturated bicarbonate) at a final concentration of approximately 15 mM to N-terminal ketone GFP (1 mg/ml) diluted into MES buffer pH 6.5. The mixture was incubated at room temperature for 48 hours and again desalted by NAP-5 and spin concentration. Conjugation efficiency was ~70% as determined by ESI-TOF MS. Conjugation of aminophenol-modified PIBMA to GFP. To a solution of nitrophenol-modified

PIBMA (1 mM in H2O) was added an equal volume of freshly prepared sodium dithionite (40 mM in 500 mM phosphate buffer pH 6.5). The color of the polymer solution immediately changed from pale yellow to colorless, and the reaction was allowed to proceed for 15 minutes at room temperature. The polymer was desalted by 5 rounds of spin concentration against a 3.5 kDa MWCO centrifugal filter (Millipore). Aminophenol was quantified by absorbance at 280 nm using the -1 -1 extinction coefficient derived from 2-amino-4-methylphenol (ε280= 2.76 mM cm ) and added to aniline-modified GFP (20 μM in 50 mM phosphate buffer pH 6.5) in 5-fold excess aminophenol to aniline. Oxidant (freshly prepared NaIO4 or K3Fe(CN)6, 50 mM in H2O) was added to a final concentration of 5 mM, and the reaction was incubated at room temperature for 10 minutes followed by quenching with Laemmli buffer. Conjugation of methoxyphenol-modified PIBMA to GFP. Methoxyphenol was quantified by ab- -1 -1 sorbance at 280 nm using 3-methoxytyramine HCl as a standard (ε280= 2.25 mM cm ). Polymer was added to aniline-modified GFP in 3-fold excess methoxyphenol to aniline in the desired final concentration. Freshly prepared NaIO4 (50 mM in H2O) was added to a final concentration of 5 mM and the mixture was incubated for 10 minutes at room temperature. The reaction was quenched with Laemmli buffer for SDS-PAGE analysis or ethylene glycol for further use. Encapsulation of PAA derivatives, PIBMA derivatives, and PIBMA-modified GFP. To encapsulate PAA derivatives and PIBMA derivatives, the capsid was disassembled and reassembled as described. Briefly, the capsid (150 μl, 8 mg/ml) was added to glacial acetic acid supercooled on ice (300 μl) and mixed by inversion. White precipitate formed immediately, and the mixture was incubated on ice for 30 minutes, followed by centrifugation at 21,000 x g at 4 oC. The supernatant was then applied to a NAP-5 column (GE Healthcare) pre-equilibrated with 1 mM acetic acid. Frac- tions were collected as the sample was desalted, and each fraction was tested for high acid content by adding 4 μl of the desalting fraction to 1 μl of 0.05% bromophenol blue in water. Fractions containing more than ~10 mM acetic acid turned the dye yellow, and were discarded. Remaining fractions were then checked for protein content by UV-Vis spectroscopy at 280 nm, and those with high protein concentrations were pooled and centrifuged again at 21,000 x g at 4 oC. MS2 CP concentration was determined by A280, which was found to agree well with SDS-PAGE densitometry

69 concentrations. The capsid was added to a mixture of polymer, buffer, salt, and TMAO at the desired concentrations, cooled on ice. Optimal conditions for the amphiphilic polymers, described in the main text, were found to be 5 μM polymer, 50 mM Bis-Tris pH 6.0, and 250 mM TMAO. The capsid allowed to reassemble for 36-48 hours, followed by analysis and purification. To encapsulate polymer-conjugated GFP, the polymer was attached as described above in a 3:1 methoxyphenol:GFP ratio. The crude reaction was added directly to the reassembly conditions, and the capsid was allowed to reassemble 36-48 hours. The reassembled capsids were analyzed by size exclusion chromatography on a Biosep-SEC-4000 column (Phenomenex) with 10 mM phosphate buffer pH 7.2 at a flow rate of 1 mL/min. To purify successful reassemblies, the capsids were precipitated, either with 0.5 M NaCl/PEG 8k as described above, or with 50 % saturated (NH4)2SO4, as PEG 8k was found to interfere with the pAF MS2 capsid. The precipitated protein was reassembled in 10 mM phosphate buffer pH 7.2 and centrifuged again to remove aggregated protein. An equal volume of washed DEAE-Sepharose suspension (Sigma) was added, and the mixture was then incubated briefly at room temperature followed by filtration using a 0.22 μm centrifugal filter. The filtered capsids were the applied to the SEC column as described, collecting fractions. Fractions containing MS2 were pooled and concentrated using a 100 kDa MWCO centrifugal filter (Millipore). Cell Experiments. MCF-7 clone 18 cells or HCC-1954 cells were incubated with 1 mL of media containing 0.4 μM wild-type MS2 capsid carrying approximately 30 amp-PAA molecules per capsid as estimated by UV-Vis spectroscopy and SDS-PAGE/optical densitometry or an equivalent amount (1.2 μM) of free polymer. The cells were imaged at 1 hour and 24 hour time points using confocal microscopy. Cloning, expression, and purification of Human Caspase 3 Δ29, Caspase 3 D175A/V266E (AEC3), 6xHis-Caspase 3-neg, Caspase 3-neg-strep (C3NS), AEC3-neg-strep (AEC3-NS), and AEC3-intein-CBD. The gene for caspase 3 without the first 29 amino acids was amplified from human cDNA with a 5’ 6xHis tag, a 5’ NdeI restriction site, and a 3’ BamHI site added to the primers (forward 5’-ATAGACATATGCATCACCATCATCATCACTCTGGAATATCCCTG- GAC-3’, reverse: 5’-ATAGAGGATCCGTGATGATGATGGTGATGTTAGTGATA- AAAATAGAGTTCTTTTG-3’. The gene was inserted into a pET24b plasmid and transformed into DH10B cells. To make caspase 3-neg, a 3’-negative tag and a 3’ HindIII was added to this by PCR (forward primer: 5’-ATAGACATATGCATCACCATCATCATCACTCTGGAATATCCCTG- GAC-3’, reverse: 5’-ATAGAAAGCTTAATCGTCTTCCTCGTCATCGTCATCTTCGTCAT- CATCCTCTTCTTCCTCGTGATAAAAATAGAGTTCTTTTGTG-3’), inserted into pET24b, and transformed into DH10B cells. To make caspase 3-neg-strep, the 5’ 6xHis tag was removed and a 3’ strep tag was added to the end of the neg tag, followed by a 3’ XhoI site by PCR am- plification (forward primer: 5’-ATAGACATATGTCTGGAATATCCCTGGAC-3’ reverse primer 5’-TTAGACTCGAGTTATTTTTCGAACTGCGGGTGGCTCCAATCGTCTTCCTCGTCATC- GT-3’), inserted into pET24b, and transformed into DH10B cells. The D175A mutation, which prevents self cleavage of the caspase-3 protease, was made by Quick-change (forward: 5’-GAACTGGACTGTGGCATTGAGACAGCCAGTGGTGTTGAT- GATGACATGGCGTGTC-3’, reverse: 5’-GACACGCCATGTCATCATCAACACCACTGGCT- GTCTCAATGCCACAGTCCAGTTC-3’), digested with DpnI overnight, and transformed into 70 DH10B cells.41 The V266E, which increases enzyme activity in the abscence of proteolytic activation, was made by Quick-change (forward: 5’-GCTACTTTTCATGCAAAGAAACAGATTC- CATGTATTGAATCCATGCTCACAAAAGAACTC-3’, reverse: 5’-GAGTTCTTTTGTGAG- CATGGATTCAATACATGGAATCTGTTTCTTTGCATGAAAAGTAGC-3’), digested with DpnI overnight, and transformed into DH10B cells. To install 5’ intein sequence and chitin binding domain for expressed protein ligation and purification, the AEC3 gene was amplified with a 5’ NdeI site and 6xHis tag, and a 3’ Lys-Gly sequence, followed by an intein overlap sequence, followed by an SpeI site (forward: 5’-ATAGACATATGCATCACCATCATCATCACTCTG- GAATATCCCTGGAC-3’, reverse: 5’-GCTATAACTAGTGCATCTCCCGTGATGCATTTGCC- GTGATAAAAATAGAGTTCTTTTGTG-3’). The plasmid pTXB1-mEGFP62 was digested with NdeI and SpeI to remove the mEGFP gene and the AEC3 gene was inserted, followed by transfor- mation into DH10B cells. All caspase 3 plasmid variants were maintained in DH10B cells. All caspase 3 variants were expressed in BL21-AI cells following modified published protocols.63 Briefly, the plasmid harboring the desired caspase construct was transformed in BL21-AI cells, recovered in 1 ml of SOC media, which was used to directly inoculate the desired volume of 2xYT expression media containing 100 mg/L kanamycin (or carbenicillin for AEC3-intein-CBD). The o o cells were grown at 37 C until the OD600 reached 0.8, and the temperature was lowered to 30 C. Expression was induced with 0.2% arabinose and 1 mM IPTG (final) for 4 hours. The cells were harvested, resuspended lysis buffer (100 mM Tris pH 8.0, 100 mM NaCl) and frozen at -80 oC for later use. Caspase 3 variants featuring N- or C- terminal 6xHis tags were purified following published protocols.63 Caspase 3 variants featuring N- or C-terminal Strep tags (WSHPQFEK) were purified using a 1 mL Strep-Tactin Superflow column (IBA BioTAGnology). Cells were resuspended in caspase buffer (20 mM PIPES, pH 7.5, 100 mM NaCl, 1 mM EDTA, 10 mM DTT, 0.1% (w/v) CHAPS, 10% sucrose) and lysed by sonication on ice. Lysate was clarified by spinning at 10,000 xg for 30 minutes, and the supernatant was applied to the column. The bound protein was washed with 20 mL of caspase buffer, and eluted in 8 mL of caspase buffer containing 2.5 mM d-desthiobiotin collecting 1 mL fractions. Fractions were analyzed by SDS-PAGE, and those containing the desired caspase mutant were pooled and stored at -20 oC. AEC3-intein-CBD was purified using chitin resin followed by cleavage with a mercaptoethane sulfonic acid (MESNA) and a cysteine-aniline derivative as described.62 Cloning, expression, purification, and refolding of human Caspase 8. The gene encoding full- length caspase 8 was amplified from human cDNA with a 5’ NdeI site and a 3’ BamHI site added to the primers (forward: 5’-ATAGACATATGGACTTCAGCAGAAATC-3’, reverse: 5’-ATAGAG- GATCCTCAATCAGAAGGGAAGACA-3’), inserted into pET20b, and transformed into DH10B cells. The gene segments encoding the p18 and p12 subunits were each cloned from this. The gene segment for p12 was PCR-amplified with a 5’ NdeI site, which installed a start codon, and a 3’ Bam- HI site (forward: 5’-ATAGACATATGGATTTATCATCACCTCAAAC-3’, reverse: 5’-ATAGAG- GATCCTCAATCAGAAGGGAAGACA-3’), inserted into pET24b, and transformed into DH10B cells. The gene segment for p18 was PCR-amplified with a 5’ NdeI site followed by a 6xHis tag, and a 3’ stop codon followed by a BamHI site (forward: 5’-ATAGACATATGCATCACCAT- CATCATCACGAATCACAGACTTTGGACAAAG-3’, reverse: 5’-ATAGAGGATCCTCAAT- CAGTCTCAACAGGTATACC-3’), inserted into pET24b, and transformed into DH10B cells.

71 Primer Name Loop Sequence Primer Sequence Wild-type QSSAQ - Csub1F 5’-GTAACCTGTAGCGTTCGTCAGAGCGACGAGGTCGACTCTGCG- CAGAATCGCAAATACACC-3’ QSDEVDSAQ Csub1R 5’-GGTGTATTTGCGATTCTGCGCAGAGTCGACCTCGTCGCTCTGAC- GAACGCTACAGGTTAC-3’ Csub2F 5’-CAAAGTAACCTGTAGCGTTCGTCAGAGCGAGTCTGACGCG- CAGAATCGCAAATACACC-3’ QSESDAQ Csub2R 5’-GGTGTATTTGCGATTCTGCGCGTCAGACTCGCTCTGACGAACGC- TACAGGTTACTTTG-3’ Csub3F 5’-AAGTAACCTGTAGCGTTCGTGAGAGCGACGCGCAGAATCG- CAAATACACCATCAAAGTCG-3’ ESDAQ Csub3R 5’-CGACTTTGATGGTGTATTTGCGATTCTGCGCGTCGCTCTCACGAAC- GCTACAGGTTACTT-3’ Csub4F 5’-AAGTAACCTGTAGCGTTCGTCAGAGCGACGCGCAGAATCG- CAAATACACCATCAAAGTCG-3’ QSDAQ Csub4R 5’-CGACTTTGATGGTGTATTTGCGATTCTGCGCGTCGCTCTGAC- GAACGCTACAGGTTACTT-3’ Csub5F 5’-TAACCTGTAGCGTTCGTGACGAGTCTGACGCGCAGAATCG- CAAATACACCATCAAAGTCG-3’ DESDAQ Csub5R 5’-CGACTTTGATGGTGTATTTGCGATTCTGCGCGTCAGACTCGTCAC- GAACGCTACAGGTTA-3’ Csub6F 5’-TAACCTGTAGCGTTCGTGAGAGCGACTCTGCGCAGAATCG- CAAATACACCATCAAAGTCG-3’ ESDSAQ Csub6R 5’-CGACTTTGATGGTGTATTTGCGATTCTGCGCAGAGTCGCTCTCAC- GAACGCTACAGGTTA-3’ Csub7F 5’-TAACCTGTAGCGTTCGTGACGAGGTGGACGCGCAGAATCG- CAAATACACCATCAAAGTCG-3’ DEVDAQ Csub7R 5’-CGACTTTGATGGTGTATTTGCGATTCTGCGCGTCCACCTCGTCAC- GAACGCTACAGGTTA-3’ Table 1. MS2 Caspase 3 substrate mutants based on the Caspase 3 consensus cut site sequence DEVD/X, where X is a small amino acid. Mutations were made to residues 50-54 in the wild-type MS2 structure (PDB 1ZDK).

The caspase 8 subunits were expressed and refolded using a modification of previously described protocols.63 Briefly, the plasmids were transformed into BL21(DE3) Codon+RIL cells and plated on kanamycin. A single colony was used to inoculate a 5 mL culture and grown overnight. This o was subcultured 1:200 into 2xYT media and grown at 37 C to an OD600 of 1.0 and induced for expression with 1 mM IPTG for 4 hours. Cells were harvested and resuspended in lysis buffer (50 mM Tris pH 8.0, 150 mM NaCl) and sonicated. The lysate was centrifuged at 10,000 x g for 20 minutes, and the supernatant was discarded. The resulting insoluble fraction, containing the caspase 8 subunit in inclusion bodies, was washed twice with 25 mL of low concentration guanidinium chloride (50 mM Tris–HCl, pH 8, 300 mM NaCl, and 1 M guanidinium hydrochloride,and 0.1% Triton-X-100), and solubilized overnight in 3 mL of high concentration guanidinium hydrochloride (6.5 M GdnCl, 25 mM Tris, pH 7.5, 5 mM EDTA, and 100 mM DTT). The concentrated, unfolded subunits were mixed together and refolded by rapid dilution (1:100) into refolding buffer (100 mM HEPES pH 7.5, 10% sucrose, 0.1% CHAPS, 10 mM DTT). The resulting solution was clarified by centrifugation at 10,000 x g for 20 minutes and applied to a Ni2+/NTA column (GE 72 Healthcare). The immobilized protein was washed with buffer (100 mM HEPES, 300 mM NaCl, 20 mM imidazole) and eluted with high concentration imidazole (100 mM HEPES, 300 mM NaCl, 300 mM imidazole). Activity assay of caspase enzymes. Activities of caspase 3 and caspase 8 enzymes were evaluated as previously reported.63 Caspase 3 variants were analyzed for hydrolysis of 7-amino-4-trifluoromethylcoumarin (AFC) from N-Benzyl-Asp-Glu-Val-Asp-7-amido-4-trifluoromethylcoumarin (DEVD-AFC) in caspase buffer (20 mM PIPES, pH 7.5, 100 mM NaCl, 1 mM EDTA, 10 mM

DTT, 0.1% (w/v) CHAPS, 10% sucrose) by monitoring increase in fluorescence (λex. 400 nm/λem. 505 nm). Caspase 8 was analyzed for hydrolysis of AFC from N-Acetyl-Ile-Glu-Thr-Asp-AFC (IETD-AFC) in the same buffer by monitoring the increase in fluorescence. Cloning, expression and purification of caspase 3 substrate MS2 mutants. Candidate MS2 mutants for degradation by caspase were chosen based on mutations in flexible loops on the interior surface of the capsid to caspase recognition sequences. In particular, Q50, S51, S52, A53, and Q54 were chosen based on their relative similarity to the caspase consensus sequence, DEVD and their position in the crystal structure (PDB 1ZDK). The following mutants were made by Quick-change with their respective primers, shown in Table 1. Mutations were made to the wild-type MS2 gene inserted in the plasmid pBAD-myc-his and transformed into DH10B cells. To evaluate expression and assembly of the mutants 1-6, overnight cultures of each were grown in LB supplemented with 50 mg/L carbenicillin. These were subcultured 1:100 in 8 mL of fresh

LB with 50 mg/L carbenicillin, grown to OD600 of 0.5, induced with 0.5% arabinose, and allowed to express overnight at 37 oC. Cells were harvested by centrifugation, lysed with BPER and lysozyme, and centrifuged at max speed for 15 minutes. Solid PEG-8k and NaCl were added to the soluble lysate fraction to final concentrations of 10% w/v and 0.5 M, respectively, and incubated at 4 oC for 1 hour. The precipitated protein was isolated by centrifugation and resuspended in 500 μl of 10 mM, and any irreversibly aggregated protein was isolated by centrifugation. Aliquots from each step were saved for each mutant and analyzed by SDS-PAGE. Encapsulation of caspase derivatives in MS2 derivatives. Wild-type MS2, Csub3 MS2, and Csub5 MS2 CPs were reassembled around C3NS as above except the reassembly buffer was replaced with caspase buffer (20 mM PIPES, pH 7.5, 100 mM NaCl, 1 mM EDTA, 10 mM DTT, 0.1% (w/v) CHAPS, 10% sucrose). The reassembled capsids were purified by precipitation and size exclusion chromatography as above. 4.8 References 1. Kariolis, M.S., Kapur, S., and Cochran, J.R. (2013). Beyond Antibodies: Using Biological Principles to Guide the Development of Next-generation Protein Therapeutics. Curr. Opin. Biotechnol. 24, 1072–1077. 2. Lu, Y., Yang, J., and Sega, E. (2006). Issues Related to Targeted Delivery of Proteins and Pep- tides. The AAPS Journal 8, E466–E478. 3. Kontos, S., and Hubbell, J.A. (2012). Drug Development: Longer-lived Proteins. Chem. Soc. Rev. 41, 2686. 4. Chester, K.A., Baker, M., and Mayer, A. (2005). Overcoming the Immunologic Response to Foreign Enzymes in Cancer Therapy. Expert Review of Clinical Immunology 1, 549–559.

73 5. Canton, I., and Battaglia, G. (2012). Endocytosis at the Nanoscale. Chem. Soc. Rev. 41, 2718. 6. Koren, E., and Torchilin, V.P. (2012). Cell-penetrating Peptides: Breaking Through to the Oth- er Side. Trends in Molecular Medicine 18, 385–393. 7. Schwarze, S.R. (1999). In Vivo Protein Transduction: Delivery of a Biologically Active Protein into the Mouse. Science 285, 1569–1572. 8. Cronican, J.J., Beier, K.T., Davis, T.N., Tseng, J.-C., Li, W., Thompson, D.B., Shih, A.F., May, E.M., Cepko, C.L., Kung, A.L., et al. (2011). A Class of Human Proteins that Deliver Func- tional Proteins into Mammalian Cells In Vitro and In Vivo. Chemistry & Biology 18, 833–838. 9. Simeon, R.L., Chamoun, A.M., McMillin, T., and Chen, Z. (2013). Discovery and Character- ization of a New Cell-Penetrating Protein. ACS Chem. Biol. 8, 2678–2687. 10. Wang, M., Alberti, K., Sun, S., Arellano, C.L., and Xu, Q. (2014). Combinatorially Designed Lipid-like Nanoparticles for Intracellular Delivery of Cytotoxic Protein for Cancer Therapy. Angew. Chem. Int. Edit. 53, 2893-2898. 11. Jesorka, A., and Orwar, O. (2008). Liposomes: Technologies and Analytical Applications. Ann. Rev. Anal. Chem. 1, 801–832. 12. Benoit, D.S.W., Gray, W., Murthy, N., Li, H., and Duvall, C.L. (2011). 4.422 - pH-Responsive Polymers for the Intracellular Delivery of Biomolecular Drugs. In Comprehensive Biomateri- als, Paul Ducheyne, ed. (Oxford: Elsevier), pp. 357–375. 13. Sebai, S.C., Milioni, D., Walrant, A., Alves, I.D., Sagan, S., Huin, C., Auvray, L., Massotte, D., Cribier, S., and Tribet, C. (2012). Photocontrol of the Translocation of Molecules, Peptides, and Quantum Dots through Cell and Lipid Membranes Doped with Azobenzene Copolymers. Angew. Chem. 124, 2174–2178. 14. Tang, R., Kim, C.S., Solfiell, D.J., Rana, S., Mout, R., Velázquez-Delgado, E.M., Chompoosor, A., Jeong, Y., Yan, B., Zhu, Z.-J., et al. (2013). Direct Delivery of Functional Proteins and En- zymes to the Cytosol Using Nanoparticle-Stabilized Nanocapsules. ACS Nano 7, 6667–6673. 15. Varkouhi, A.K., Scholte, M., Storm, G., and Haisma, H.J. (2011). Endosomal Escape Pathways for Delivery of Biologicals. J. Control. Release 151, 220–228. 16. Gu, Z., Biswas, A., Zhao, M., and Tang, Y. (2011). Tailoring Nanocarriers for Intracellular Protein Delivery. Chem. Soc. Rev. 40, 3638-3655. 17. Parveen, S., Misra, R., and Sahoo, S.K. (2012). Nanoparticles: a Boon to Drug Delivery, Ther- apeutics, Diagnostics and Imaging. Nanomedicine: Nanotechnology, Biology and Medicine 8, 147–166. 18. King, N.P., and Lai, Y.-T. (2013). Practical Approaches to Designing Novel Protein Assem- blies. Curr. Opin. Struct. Biol. 23, 632–638. 19. Garcea, R.L., and Gissmann, L. (2004). Virus-like Particles as Vaccines and Vessels for the Delivery of Small Molecules. Curr. Opin. Biotechnol. 15, 513–517. 20. Ma, Y., Nolte, R.J.M., and Cornelissen, J.J.L.M. (2012). Virus-based Nanocarriers for Drug Delivery. Advanced Drug Delivery Reviews 64, 811–825. 21. Stephanopoulos, N., Tong, G.J., Hsiao, S.C., and Francis, M.B. (2010). Dual-Surface Modified Virus Capsids for Targeted Delivery of Photodynamic Agents to Cancer Cells. ACS Nano 4, 6014–6020. 22. Ren, Y., Wong, S.M., and Lim, L.Y. (2010). Application of Plant Viruses as Nano Drug Deliv- ery Systems. Pharm. Res. 27, 2509–2513. 23. Teunissen, E.A., de Raad, M., and Mastrobattista, E. Production and Biomedical Applications of Virus-like Particles Derived from Polyomaviruses. J. Control. Release 172, 305-21.

74 24. Abbing, A., Blaschke, U.K., Grein, S., Kretschmar, M., Stark, C.M.B., Thies, M.J.W., Walter, J., Weigand, M., Woith, D.C., Hess, J., et al. (2004). Efficient Intracellular Delivery of a Pro- tein and a Low Molecular Weight Substance via Recombinant Polyomavirus-like Particles. J. Biol. Chem. 279, 27410–27421. 25. Peretti, S., Schiavoni, I., Pugliese, K., and Federico, M. (2005). Cell Death Induced by the Herpes Simplex Virus-1 Thymidine Kinase Delivered by Human Immunodeficiency Virus-1- Based Virus-like Particles. Mol Ther 12, 1185–1196. 26. Aoki, T., Miyauchi, K., Urano, E., Ichikawa, R., and Komano, J. (2011). Protein Transduction by Pseudotyped Lentivirus-like Nanoparticles. Gene Therapy 18, 936–941. 27. Kaczmarczyk, S.J., Sitaraman, K., Young, H.A., Hughes, S.H., and Chatterjee, D.K. (2011). Protein Delivery Using engineered Virus-like Particles. Proc. Nat. Acad. Sci. U.S.A 108, 16998–17003. 28. Tao, P., Mahalingam, M., Marasa, B.S., Zhang, Z., Chopra, A.K., and Rao, V.B. (2013). In Vitro and In Vivo Delivery of Genes and Proteins using the Bacteriophage T4 DNA Packaging Machine. Proc. Nat. Acad. Sci. U.S.A. 110, 5846–5851. 29. Wu, M., Brown, W.L., and Stockley, P.G. (1995). Cell-specific Delivery of Bacteriophage- encapsidated Ricin A Chain. Bioconj. Chem. 6, 587–595. 30. Ashley, C.E., Carnes, E.C., Phillips, G.K., Durfee, P.N., Buley, M.D., Lino, C.A., Padilla, D.P., Phillips, B., Carter, M.B., Willman, C.L., et al. (2011). Cell-Specific Delivery of Diverse Car- gos by Bacteriophage MS2 Virus-like Particles. ACS Nano 5, 5729–5745. 31. Li, H., E. Nelson, C., C. Evans, B., and L. Duvall, C. (2011). Delivery of Intracellular-Acting Biologics in Pro-Apoptotic Therapies. Current Pharmaceutical Design 17, 293–319. 32. Mashima, T., and Tsuruo, T. (2005). Defects of the Apoptotic Pathway as Therapeutic Target Against Cancer. Drug Resistance Updates 8, 339–343. 33. Wen, X., Lin, Z.-Q., Liu, B., and Wei, Y.-Q. (2012). Caspase-mediated Programmed Cell Death Pathways as Potential Therapeutic Targets in Cancer. Cell Proliferation 45, 217–224. 34. Pop, C., and Salvesen, G.S. (2009). Human Caspases: Activation, Specificity, and Regulation. J. Biol. Chem. 284, 21777–21781. 35. Huang, Q., Li, F., Liu, X., Li, W., Shi, W., Liu, F.-F., O’Sullivan, B., He, Z., Peng, Y., Tan, A.- C., et al. (2011). Caspase 3-mediated Stimulation of Tumor Cell Repopulation During Cancer Radiotherapy. Nature Medicine 17, 860–866. 36. Galluzzi, L., Kepp, O., and Kroemer, G. (2012). Caspase-3 and Prostaglandins Signal for Tu- mor Regrowth in Cancer Therapy. Oncogene 31, 2805–2808. 37. Witkowski, W.A., and Hardy, J.A. (2011). A Designed Redox-controlled Caspase. Protein Sci- ence 20, 1421–1431. 38. Endo, M., Nakayama, K., Kaida, Y., and Majima, T. (2004). Design and Synthesis of Photo- chemically Controllable Caspase-3. Angew. Chem. 116, 5761–5763. 39. Vocero-Akbani, A.M., Heyden, N.V., Lissy, N.A., Ratner, L., and Dowdy, S.F. (1999). Killing HIV-infected Cells by Transduction with an HIV Protease-activated Caspase-3 protein. Nat. Med. 5, 29–33. 40. Nobel, C.S.I., Kimland, M., Nicholson, D.W., Orrenius, S., and Slater, A.F.G. (1997). Disulfi- ram is a Potent Inhibitor of Proteases of the Caspase Family. Chemical Research in Toxicology 10, 1319–1324. 41. Walters, J., Pop, C., Scott, F.L., Drag, M., Swartz, P., Mattos, C., Salvesen, G.S., and Clark, A.C. (2009). A Constitutively Active and Uninhibitable Caspase-3 Zymogen Efficiently In-

75 duces Apoptosis. Biochem. J. 424, 335–345. 42. Daub, M., Waldherr, S., Allgöwer, F., Scheurich, P., and Schneider, G. (2012). Death Wins Against Life in a Spatially Extended Model of the Caspase-3/8 Feedback Loop. Biosystems 108, 45–51. 43. Hoffman, A.S., and Stayton, P.S. (2007). Conjugates of Stimuli-responsive Polymers and Pro- teins. Progress in Polymer Science 32, 922–932. 44. Arunachalam, B., Phan, U.T., Geuze, H.J., and Cresswell, P. (2000). Enzymatic Reduction of Disulfide Bonds in Lysosomes: Characterization of a Gamma-interferon-inducible Lysosomal Thiol Reductase (GILT). Proc. Nat. Acad. Sci. U.S.A. 97, 745–750. 45. Yang, J., Chen, H., Vlahov, I.R., Cheng, J.X., and Low, P.S. (2006). Evaluation of Disulfide Reduction During Receptor-mediated Endocytosis by Using FRET Imaging. Proc. Nat. Acad. Sci. U.S.A. 103, 13872–13877. 46. Austin, C.D., Wen, X., Gazzard, L., Nelson, C., Scheller, R.H., and Scales, S.J. (2005). Oxidiz- ing Potential of Endosomes and Lysosomes Limits Intracellular Cleavage of Disulfide-based Antibody–drug Conjugates. Proc. Nat. Acad. Sci. U.S.A. 102, 17987–17992. 47. Behr, J.P. (1997). The Proton Sponge: A Trick to Enter Cells the Viruses did not Exploit. CHI- MIA International Journal for Chemistry 51, 1–2. 48. Yessine, M. (2004). Membrane-destabilizing Polyanions: Interaction with Lipid Bilayers and Endosomal Escape of Biomacromolecules. Advanced Drug Delivery Reviews 56, 999–1021. 49. Binder, W.H. (2008). Polymer-Induced Transient Pores in Lipid Membranes. Angew. Chem. Int. Edit. 47, 3092–3095. 50. Maier, K., and Wagner, E. (2012). Acid-Labile Traceless Click Linker for Protein Transduc- tion. J. Am. Chem. Soc. 134, 10169–10173. 51. Zhao, M., Biswas, A., Hu, B., Joo, K.-I., Wang, P., Gu, Z., and Tang, Y. (2011). Redox-responsive Nanocapsules for Intracellular Protein Delivery. Biomaterials 32, 5223–5230. 52. Biswas, A., Joo, K.-I., Liu, J., Zhao, M., Fan, G., Wang, P., Gu, Z., and Tang, Y. (2011). En- doprotease-Mediated Intracellular Protein Delivery Using Nanocapsules. ACS Nano 5, 1385– 1394. 53. Chul Cho, K., Hoon Jeong, J., Jung Chung, H., O Joe, C., Wan Kim, S., and Gwan Park, T. (2005). Folate Receptor-mediated Intracellular Delivery of Recombinant Caspase-3 for Induc- ing Apoptosis. J. Control. Release 108, 121–131. 54. Gu, Z., Yan, M., Hu, B., Joo, K.-I., Biswas, A., Huang, Y., Lu, Y., Wang, P., and Tang, Y. (2009). Protein Nanocapsule Weaved with Enzymatically Degradable Polymeric Network. Nano Letters 9, 4533–4538. 55. Yan, M., Du, J., Gu, Z., Liang, M., Hu, Y., Zhang, W., Priceman, S., Wu, L., Zhou, Z.H., Liu, Z., et al. (2009). A Novel Intracellular Protein Delivery Platform Based on Single-protein Nanocapsules. Nature Nanotechnology 5, 48–53. 56. Tang, R., Kim, C.S., Solfiell, D.J., Rana, S., Mout, R., Velázquez-Delgado, E.M., Chompoosor, A., Jeong, Y., Yan, B., Zhu, Z.-J., et al. (2013). Direct Delivery of Functional Proteins and En- zymes to the Cytosol Using Nanoparticle-Stabilized Nanocapsules. ACS Nano 7, 6667–6673. 57. Farkas, M.E., Aanei, I.L., Behrens, C.R., Tong, G.J., Murphy, S.T., O’Neil, J.P., and Francis, M.B. (2013). PET Imaging and Biodistribution of Chemically Modified Bacteriophage MS2. Mol. Pharmaceutics 10, 69–76. 58. Stephanopoulos, N., Tong, G.J., Hsiao, S.C., and Francis, M.B. (2010). Dual-Surface Modified Virus Capsids for Targeted Delivery of Photodynamic Agents to Cancer Cells. ACS Nano 4,

76 6014–6020. 59. Glasgow, J.E., Capehart, S.L., Francis, M.B., and Tullman-Ercek, D. (2012). Osmolyte-Medi- ated Encapsulation of Proteins inside MS2 Viral Capsids. ACS Nano 6, 8658–8664. 60. Vial, F., Oukhaled, A.G., Auvray, L., and Tribet, C. (2007). Long-living Channels of Well Defined Radius Opened in Lipid Bilayers by Polydisperse, Hydrophobically-modified Poly- acrylic Acids. Soft Matter 3, 75-78. 61. Lin, C.-A.J., Sperling, R.A., Li, J.K., Yang, T.-Y., Li, P.-Y., Zanella, M., Chang, W.H., and Parak, W.J. (2008). Design of an Amphiphilic Polymer for Nanoparticle Coating and Function- alization. Small 4, 334–341. 62. Hooker, J.M., Esser-Kahn, A.P., and Francis, M.B. (2006). Modification of Aniline Containing Proteins Using an Oxidative Coupling Strategy. J. Am. Chem. Soc. 128, 15558–15559. 63. Roschitzki-Voser, H., Schroeder, T., Lenherr, E.D., Frölich, F., Schweizer, A., Donepudi, M., Ganesan, R., Mittl, P.R.E., Baici, A., and Grütter, M.G. (2012). Human Caspases In Vitro: Expression, Purification and Kinetic Characterization. Protein Expression and Purification 84, 236–246. 64. Wang, K.T., Iliopoulos, I., and Audebert, R. (1988). Viscometric Behaviour of Hydrophobi- cally Modified Poly(sodium acrylate). Polymer Bulletin 20, 577–582. 65. Pelet, J.M., and Putnam, D. (2011). An In-Depth Analysis of Polymer-Analogous Conjugation using DMTMM. Bioconj. Chem. 22, 329–337. 66. Plank, C., Oberhauser, B., Mechtler, K., Koch, C., and Wagner, E. (1994). The Influence of Endosome-disruptive Peptides on Gene Transfer Using Synthetic Virus-like Gene Transfer Systems. J. Biol. Chem. 269, 12918–12924. 67. Sehgal, D., and Vijay, I.K. (1994). A Method for the High Efficiency of Water-Soluble Car- bodiimide-Mediated Amidation. Anal. Biochem. 218, 87–91. 68. Netirojjanakul, C., Witus, L.S., Behrens, C.R., Weng, C.-H., Iavarone, A.T., and Francis, M.B. (2012). Synthetically Modified Fc Domains as Building Blocks for Immunotherapy Applica- tions. Chem. Sci. 4, 266–272. 69. Behrens, C.R., Hooker, J.M., Obermeyer, A.C., Romanini, D.W., Katz, E.M., and Francis, M.B. (2011). Rapid Chemoselective Bioconjugation through Oxidative Coupling of Anilines and Aminophenols. J. Am. Chem. Soc. 133, 16398–16401. 70. Jańczewski, D., Tomczak, N., Han, M.-Y., and Vancso, G.J. (2011). Synthesis of Functional- ized Amphiphilic Polymers for Coating Quantum Dots. Nat. Protocols 6, 1546–1553. 71. Carrico, Z.M., Romanini, D.W., Mehl, R.A., and Francis, M.B. (2008). Oxidative Coupling of Peptides to a Virus Capsid Containing Unnatural Amino Acids. Chem. Comm. 10:1205-1207. 72. Marini, M., Frabetti, F., Canaider, S., Dini, L., Falcieri, E., and Poirier, G.G. (2001). Modula- tion of Caspase-3 Activity by Zinc Ions and by the Cell Redox State. Experimental Cell Re- search 266, 323–332. 73. Pop, C., and Salvesen, G.S. (2009). Human Caspases: Activation, Specificity, and Regulation. J. Biol. Chem. 284, 21777–21781.