Encapsulation and Controlled Release of Protein Guests by the

Subscriber access provided by READING UNIV

Article

Encapsulation and Controlled Release of Protein Guests by the Bacillus subtilis Lumazine Synthase Capsid Xue Han, and Kenneth J. Woycechowsky Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.7b00669 • Publication Date (Web): 31 Oct 2017 Downloaded from http://pubs.acs.org on November 6, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Biochemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Encapsulation and Controlled Release of Protein Guests by the Bacillus subtilis Lumazine Synthase Capsid Xue Han and Kenneth J. Woycechowsky* School of Pharmaceutical Science and Technology, Tianjin University, 300072 Tianjin, China KEYWORDS: Bionanotechnology; Biological Compartmentalization; Protein-Protein Interaction; Protein Self-Assembly; Nanocontainer

ACS Paragon Plus Environment

1

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 42

ABSTRACT: In Bacillus subtilis, the 60-subunit dodecahedral capsid formed by lumazine synthase (BsLS) acts as a container for trimeric riboflavin synthase (BsRS). To test whether the C-terminal sequence of BsRS is responsible for its encapsulation by BsLS, the green fluorescent protein (GFP) was fused to either the last 11 or last 32 amino acids of BsRS, offering variants GFP11 and GFP32, respectively. After purification, BsLS capsids that had been co-produced in bacteria with GFP11 or GFP32 are 15-fold and 6-fold more fluorescent, respectively, than BsLS co-produced with GFP lacking any BsRS fragment, indicating complex formation. ELISA experiments confirm that GFP11 is localized within the BsLS capsid. In addition, fusing the last 11 amino acids of BsRS to the C-terminus of the Abrin A chain also led to its encapsulation by BsLS at a level similar to that of GFP11. Together, these results demonstrate that the C-terminal tail of BsRS can act as an encapsulation tag capable of targeting other proteins to the BsLS capsid interior. As with the natural BsLS+BsRS complex, mild changes in pH and buffer identity trigger dissociation of the GFP11 guest, accompanied by a substantial expansion of the BsLS capsid. This system for protein encapsulation and release provides a novel tool for bionanotechnology.


2

Page 3 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

The spatial organization of biomolecules is a fundamental aspect of life. To aid in this task, some proteins self-assemble into closed hollow shells that form compartments with sizes ranging from nanometers to micrometers 1. Perhaps the most well-known examples of such protein shells are provided by virus capsids, which use their hollow interiors to encapsulate, protect, and transport their genomes. Many non-viral protein shells house other proteins, which has important ramifications for their activities. For example, the protein shell can potentially control access of substrate molecules to guest enzymes inside or block diffusion of intermediates into the bulk environment 2. Similarly, the co-compartmentalization of two or more enzymes that carry out consecutive steps of a biochemical pathway can increase their effective molarities, allowing for more efficient substrate channeling between active sites 3. To date, several subcellular proteinbounded compartments have been described, including encapsulin 4-6 , the carboxysome 7, and bacterial microcompartments (BMCs) for the utilization of propanediol and ethanolamine 8, 9. Increasingly, protein capsids have been repurposed to house non-native guest proteins 10-20, which could be useful for developing novel nanoreactors or drug delivery systems. The lumazine synthase (LS) capsid has attracted much recent attention as a scaffold for bionanotechnology 21-26. In the riboflavin biosynthetic pathway of Bacillus subtilis, LS forms a cage complex with riboflavin synthase (RS) 27-29. Two different forms of B. subtilis RS (BsRS) were initially identified 30. “Light” BsRS was found to be an α3 trimer, whereas “heavy” BsRS was found to contain two different kinds of polypeptide chains, forming an α3β60 complex31. We now know that only the α subunits possess riboflavin synthase activity and that the β subunits are actually LS (BsLS) 32. Self-assembly of 60 BsLS subunits produces a hollow dodecahedron with inner and outer diameters of 8.5 nm and 16 nm, respectively (Figure 1A) 33, 34. Together, BsLS and BsRS form a bifunctional enzyme complex, in which BsLS encapsulates BsRS, and


3

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 42

the two enzymes carry out the last two steps of riboflavin biosynthesis in tandem. The active sites of BsLS are located at the inner capsid surface 34, and substrate channeling has been shown to enhance catalytic efficiency of the BsLS+BsRS complex at low substrate concentrations 35. Despite a great deal of structural and functional studies on BsLS and the BsLS+BsRS complex 31-44, the three-dimensional structure of BsRS remains unknown. This lack of structural information has hampered efforts to understand how BsRS is localized within BsLS. Recently, Azuma et al. 45 reported that the LS and RS homologs from Aquifex aeolicus (AaLS and AaRS, respectively) also form an encapsulation complex, and that the last 12 amino acids of AaRS are necessary for its encapsulation by AaLS. Further, this 12 amino-acid-long tail is sufficient to direct foreign guests into the AaLS capsid. Here, we fuse C-terminal fragments of BsRS to the green fluorescent protein (GFP) and show that these fragments can act as encapsulation tags that promote specific complex formation with BsLS. In contrast to the AaLS-based encapsulation system, guest release can be triggered from the BsLS capsid upon changing the buffer from sodium phosphate at neutral pH to Tris•HCl at slightly alkaline pH.

EXPERIMENTAL PROCEDURES Plasmid construction. To create plasmids encoding GFP fused at its C-terminal end to the last 11 or last 32 amino acids of BsRS (pACYC-GFP-BsRS11, pACYC-Abrin A-BsRS11, and pACYC-GFP-BsRS32), the region encoding the R10 tag in plasmid pACYC-GFP-R10

25

was

replaced with synthetic, double-stranded oligonucleotides encoding the BsRS fragments. The plasmid encoding BsLS (pMG-BsLS) was generated by replacing the sequence encoding the LS from Saccharomyces cerevisiae (and a C-terminal hexahistidine tag) in plasmid pMG-yLS

46

with the BsLS gene. Detailed procedures are described in the Supporting Information.


4

Page 5 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Protein production and purification. Proteins were produced and purified using procedures based on previously reported methods

47

. Detailed procedures are described in the Supporting

Information. Measurement

of

protein

concentration.

Quantitative

determinations

of

protein

concentration were made using the Bradford assay. Briefly, protein sample (1 µL) was combined with PBS buffer (14 µL) and then added to Coomassie Brilliant Blue solution (285 µL). Following a 5 min incubation at room temperature, the A595 was measured using the Eppendorf BioPhotometer. The protein concentration was then estimated based on comparison to a calibration curve obtained using known concentrations of bovine serum albumin. Analytical size-exclusion chromatography. The relative sizes of BsLS, and GFP variants, their complexes, and AaLS were determined by measuring the UV absorbance at 280 nm and the fuorescence emission at 508 nm of fractions eluted from a HiPrep 16/60 Sephacryl S-400 HR size-exclusion column. Detailed procedures are described in the Supporting Information. Fluorescence spectroscopy. The fluorescence emission spectra of various protein samples were measured using an Edinburgh Instruments FLS980 fluorescence spectrometer with an excitation wavelength of 378 nm. Detailed procedures are described in the Supporting Information. Western blotting. The amount of Abrin A-HA-BsRS11 encapsulated by BsLS was estimated by Western blot analysis. A polyclonal mouse anti-HA primary antibody and a monoclonal goat anti-mouse secondary antibody were used to detect Abrin A-HA-BsRS11. The amount of Abrin A-HA-BsRS11 present in purified capsid samples was determined by comparing the band intensity to an internal standard curve made with purified free Abrin A-HA-BsRS11. Detailed procedures are described in the Supporting Information.


5

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 42

Detection of GFP by ELISA. The accessibility of GFP11 in various protein samples was examined using ELISA. A polyclonal rabbit anti-GFP antibody and a monoclonal goat antirabbit antibody conjugated to horseradish peroxidase were used as the primary and secondary antibodies, respectively. Detailed procedures are described in the Supporting Information. Release of GFP11 from BsLS. Previous studies have shown that BsRS can be dissociated from the natural BsLS+BsRS encapsulation complex by changing the buffer conditions

43

. To

see if a similar buffer exchange would cause release of GFP11 from BsLS, we dialyzed a sample of purified BsLS+GFP11 complex (1 mL of a 1 mg/mL solution in size-exclusion buffer) into 2 L of pH 8.0 buffer containing either 0.1 M Tris-HCl and 1 mM EDTA or 0.1 M sodium phosphate and 1 mM EDTA for 2 days, changing the dialysis buffer once after the first day. The sample was then subjected to analytical size-exclusion chromatography, as described above, using a running buffer containing 0.1 M Tris-HCl and 1 mM EDTA at pH 8.0. To check for release of GFP11 with size-exclusion buffer (the standard working and storage buffer for the BsLS+GFP11 complex), 2 mL of freshly purified BsLS+GFP11 complex (3.5 mg/mL) was incubated at 4 °C in the dark for 28 days. An aliquot of this sample (0.57 mL) was mixed with size-exclusion buffer to a final volume of 2 mL, filtered, and then subjected to analytical sizeexclusion chromatography, as described above. Attempts at capsid loading in vitro. In order to know whether GFP11 can bind to intact BsLS capsids, free GFP11 and BsLS (each produced and purified in the absence of the other) were mixed in vitro. BsLS (1 mL of a 3 mg/mL solution) and GFP11 (1 mL of a 1.5 mg/mL solution) were both added to a dialysis bag (14 kDa MWCO) and dialyzed against 2 L of either size-exclusion buffer for 2 days at 4 °C, with a single change of the dialysis buffer after the first


6

Page 7 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

day. The sample was then subjected to analytical size-exclusion chromatography, as described above. Given that certain buffer changes can lead to restructuring of BsLS and cargo release, the same dialysis procedure was carried out using dialysis buffer containing 0.1 M Tris-HCl and 1 mM EDTA at pH 8.0. In this case, the stock solutions of BsLS and GFP11 had concentrations of 0.92 mg/mL and 0.73 mg/mL, respectively. After the second day of dialysis, the samples were dialyzed further for an additional two days, against either the same buffer (0.1 M Tris-HCl, 1 mM EDTA, pH 8.0) or against size-exclusion buffer. The samples were then subjected to analytical size-exclusion chromatography, as described above.

RESULTS Design of fusion proteins. In contrast to BsRS, the three-dimensional structure of riboflavin synthase from Escherichia coli (EcRS) is known 48. Both EcRS and E. coli lumazine synthase (EcLS) are homologous to BsRS and BsLS, sharing 30% and 50% sequence identity, respectively. However, EcLS does not encapsulate EcRS 49. Nonetheless, we looked to the EcRS structure for potential clues to the encapsulation determinants of BsRS. The EcRS homotrimer is shaped like a mushroom, with the N-terminal catalytic domains forming the cap and an α-helical coiled-coil trimerization domain near the C-terminus forming the stem (Figure 1B). In other natural protein encapsulation systems, such as BMCs and encapsulins, peptide tails are used to localize the guest protein within the host cage 5, 50. Notably, the last seven amino acids of EcRS (EcRS207-213) are not visible in the crystal structure 48, suggesting that they may be disordered. Although the amino acid sequences of EcRS and BsRS near their C-termini are not very similar, and BsRS has a four amino-acid extension relative to EcRS (Figure S1), it seems


7

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 42

plausible that BsRS may have a flexible tail as well. The low sequence homology in this region would be consistent with the differing encapsulation abilities of the E. coli and B. subtilis systems. The flexible tail of EcRS immediately follows the α-helical trimerization domain, EcRS187-206 (Figures 1B & S1). It is not clear if the trimeric assembly of BsRS, or the amino acids within this α-helical trimerization domain, might contribute to encapsulation. Therefore, we hypothesized that the last 11 or last 32 amino acids of BsRS (BsRS205-215 and BsRS184-215, respectively, Figure S1) can potentially act as a tag for encapsulation by BsLS. To test the hypothesis that the encapsulation determinants of BsRS are located at its Cterminus, peptide sequences corresponding to BsRS205-215 and BsRS184-215 were each fused to the C-terminus of the green fluorescent protein (GFP), giving variants GFP11 and GFP32, respectively (Figure 1C). To study protein encapsulation by BsLS, monomeric GFP provides a convenient model guest, because it is similar in size to one subunit of BsRS (27 kDa for GFP and 22 kDa for BsRS), and its intrinsic fluorescence can act as a sensitive reporter 51. Analysis of capsid-guest complexes. The GFP variants were each co-produced with BsLS in E. coli BL21 (DE3) cells. As a control, GFP lacking any fusion partner (GFP0) was also coproduced with BsLS. The BsLS capsids were then purified and analyzed by SDS-PAGE (Figure S2). Samples of purified BsLS that had been co-produced with GFP11 or GFP32 (BsLS+GFP11 and BsLS+GFP32, respectively), showed bands with mobilities corresponding to the molecular weights of the co-produced GFP variant. In contrast, no GFP band was apparent for samples of purified BsLS that had been purified with GFP0 (BsLS+GFP0) or that had been produced in the absence of any GFP variant. For the BsLS+GFP11 sample, the assignments of the GFP11 and BsLS bands was confirmed by tryptic digest and LC-MS/MS analysis of protein extracted from


8

Page 9 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

gel slices. Consistent with the presence of additional bands in the gel (which were present in all of the BsLS samples), several other proteins were also identified (Figure S3), the most prominent of which is the chaperonin GroEL. BsLS capsid assembly was unaffected by co-production with the GFP variants, as assessed by transmission electron microscopy (TEM). All samples of BsLS showed circular particles in the TEM images with average diameters around 16 nm (Figure S4), consistent with the T=1 icosahedron previously reported for BsLS 34. Analysis of the purified BsLS samples by sizeexclusion chromatography confirmed that the co-production of BsLS with the GFP variants did not alter the capsid size. In the UV chromatograms of BsLS that had been co-produced with either GFP0, GFP11, or GFP32 (Figure 2A-C, blue traces), the protein elutes as a single peak, at a similar volume to BsLS capsids that were produced in the absence of GFP (Supporting Figure S5). In contrast, the capsid peaks showed varying levels of fluorescence (Figure 2A-C, black traces), depending on the identity of the co-production partner. Both BsLS+GFP11 (Figure 2B) and BsLS+GFP32 (Figure 2C) showed more fluorescence than BsLS+GFP0 (Figure 2A). Further, the fluorescent peaks eluted earlier than the purified free GFP variants that had been produced in the absence of BsLS (Figure 2D-F). The shift in the elution of the GFP-based fluorescence indicates that stable complexes between BsLS and the GFP fusion proteins accumulate during their co-production in the cell. The purified complexes were subjected to more detailed fluorescence analysis. The fluorescence spectra of BsLS+ GFP11 and BsLS+GFP32 both show dramatically higher emission intensities than BsLS+GFP0 or BsLS produced in the absence of GFP (Figure 3A). In contrast, the fluorescence intensities at 508 nm of isolated free GFP11 and GFP32 are 14% and 44% lower, respectively, than that of GFP0, although the general shapes of the spectra are all


9

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 42

similar (Figure S6). For each GFP variant, a calibration curve was constructed, relating fluorescence intensity at 508 nm to the concentration of purified GFP (Figure S7). The amount of GFP variant present in each BsLS capsid sample was estimated by comparing the fluorescence intensity to the corresponding calibration curve. The concentration of 60-subunit BsLS capsids was determined by measuring the total protein concentration and then correcting for the GFP variant concentration. To calculate the number of GFP molecules per capsid, the concentration of GFP variant present in each BsLS sample was divided by the concentration of BsLS capsids. Purified BsLS samples that had been co-produced with GFP11, GFP32, or GFP0 contained an average of 1.3, 0.52, and 0.086 guest molecules per capsid (Figure 3B). To confirm the stoichiometry of the complexes, gel images resulting from SDS-PAGE of purified BsLS+GFP11 and BsLS+GFP32 complexes were analyzed (Figure S8). In these samples, the ratios of the band intensities for BsLS and the corresponding GFP variant indicate the presence of 1.1 and 0.57 GFPs per capsid for BsLS+GFP11 and BsLS+GFP32, respectively. These values are in good agreement with the capsid loading values obtained by fluorescence emission (Figure 3B). Thus, fusing the last 11 or last 32 amino acids of BsRS to GFP increases the number of guest molecules per capsid by about 15-fold and 6-fold, respectively. One possible explanation for the different levels of complex formation seen for BsLS with the GFP variants might be that production levels vary for the different proteins. For BsLS coproduced with GFP11 or GFP32, the yields of purified capsid were about 25% higher and 25% lower, respectively, than that of BsLS co-produced with GFP0. The different efficiencies of complex formation cannot be explained by such minor differences in BsLS production levels. Likewise, the differences in the amounts of cellular GFP present during coproduction of the variants with BsLS (Figure S9) are also too small to explain the elevated amounts of complex


10

Page 11 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

observed with GFP11 or GFP32, relative to GFP0. Further, purified capsids of the homologous Aquifex aeolicus LS (AaLS) that had been co-produced with GFP11 do not contain significant amounts of fluorescence (Figure 4), demonstrating that the 11 amino acid fragment of BsRS interacts specifically with BsLS. Thus, the greater association of GFP11 and GFP32 with BsLS most likely stems from the affinity of the BsRS fragments for the capsid rather than from simple mass transfer effects. While the encapsulation of protein guests bearing the C-terminal tail of BsRS is specific for the BsLS capsid, we hypothesized that the last 11 amino acids of BsRS can act as a general encapsulation tag for targeting guest proteins to the interior of BsLS. To test this hypothesis, we attempted to encapsulate the Abrin A chain, a ribosome-modifying enzyme from the jequirity bean plant (Abrus precatorius), instead of GFP. The Abrin A chain was chosen because it has a similar size (28 kDa) and quaternary structure (monomeric) to GFP, but completely different primary and tertiary structure. Further, the efficient expression of recombinant abrin A chain in E. coli has been reported52, and its ribosome-inactivating activity makes the abrin A chain an attractive candidate for future drug delivery studies53. To promote encapsulation of this new guest, a chimeric peptide sequence consisting of the hemagglutinin (HA) epitope tag (which is recognized by the anti-HA-tag antibody) followed by the BsRS205-215 sequence was fused to the C-terminus of the Abrin A chain, resulting in the variant Abrin A-HA-BsRS11(Figure S10). After coproduction of Abrin A-HA-BsRS11 with BsLS, the capsids were purified and analyzed by Western blot using an anti-HA primary antibody (Figure 5). Comparison of band intensities for samples of purified BsLS that had been co-produced with Abrin A-HA-BsRS11 with samples of free Abrin A-HA-11 gives a loading yield of 1.3 ± 0.5 guests per capsid, which is similar to


11

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 42

the loading yield of GFP11. Thus, the last 11 amino acids of BsRS can provide fairly consistent levels of guest protein encapsulation for two unrelated proteins. The location of the binding site for the BsRS205-215 encapsulation tag on BsLS is not clear. It is reasonable to infer that the tagged guests bind to the interior surface of the BsLS capsid, given that complex formation is driven by a fragment of BsRS, whose encapsulation by BsLS is well established 31, 33, 38, 40. An ELISA experiment was carried out to confirm that GFP11 is localized inside of the BsLS capsid. Briefly, purified protein samples were adsorbed onto a solid phase and polyclonal anti-GFP antibody from rabbit was used as a probe for accessible GFP. BsLS that had been co-produced with GFP11 gave only slightly higher ELISA signal than BsLS that had been produced in the absence of GFP11. In contrast, both free GFP11 and a mixture of free GFP11 and BsLS (that had each been produced independently of the other) gave substantially higher ELISA signals than the purified BsLS+GFP11 complex (Table S1). Thus, the GFP11 in the BsLS+GFP11 complex is hidden from the antibody. The most likely explanation is that the Cterminal fragment of BsRS directs the encapsulation of GFP11 within BsLS. Guest release from the capsid. The association of GFP11 with BsLS is quite long-lived under the standard conditions used during the characterization of the complex. Indeed, four weeks after purification of the BsLS+GFP11 complex, no free GFP11 was detected (Figure 6A). It is not clear whether the persistence of the complex is due to a high affinity of the BsRS205-215 peptide for the BsLS capsid or to the inability of GFP11 to diffuse through the relatively small pores in the capsid shell. Nevertheless, changes in buffer and pH can induce large structural rearrangements in the BsLS capsid54, which have the potential to alter its association with GFP11. To test changes in solution conditions on the stability of the BsLS+GFP11 complex, the pH of the purified complex


12

Page 13 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

was increased from 7.0 to 8.0 by dialysis for two days at 4 °C. When the buffer identity was kept the same (100 mM sodium phosphate), no changes were apparent in the elution behavior of the complex during size-exclusion chromatography (Figure 6B). In contrast, when the pH increase was accompanied by a change in the buffer identity to 100 mM Tris-HCl, both the UV and fluorescent chromatograms showed major changes (Figure 6C). The fluorescence chromatogram showed a new major peak at a much later elution volume, consistent with dissociation of GFP11 from the capsid. The ratio of fluorescence signal associated with the late-eluting peak to that remaining in the original peak indicates that ~90% of the GFP11 escapes from the capsid under these conditions (Figure 6). In the UV chromatogram, a new peak was also observed that eluted earlier, suggesting that a majority of the BsLS capsids increased in size. TEM analysis of BsLS and BsLS+GFP11 samples that had been dialyzed into 100 mM Tris-HCl at pH 8.0 showed that the majority of the capsids had expanded structures with diameters around 26 nm (Figure S11). The dissociation of the BsLS+GFP11 complex and the accompanying changes in the BsLS capsid size under these conditions are consistent with the previously reported behavior of the natural BsLS+BsRS system 31, 38, 54. Thus, the release of protein cargo from the BsLS capsid can also be triggered by a mild switch in the solution conditions. Attempted in vitro loading. Under certain conditions, a large majority of encapsulated GFP11 exits the BsLS capsid, suggesting that the capsid structure is dynamic. Therefore, we speculated that adding a high concentration of free GFP11 to BsLS might lead to accumulation of GFP11 inside the capsid. When an excess of purified free GFP11 is mixed in vitro with BsLS capsids and the proteins are incubated together for two days, no fluorescence is detected in the capsid-containing fractions from the size-exclusion column (Figure S12A). Therefore, GFP11 is


13

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 42

unable to diffuse across the capsid wall under conditions where the encapsulation complex is stable. Under the conditions that promote guest release, GFP11 can certainly cross the BsLS capsid shell. However, the affinity of BsLS for GFP11 is low. After incubation of BsLS with an excess of GFP11 in pH 8.0 buffer containing 100 mM Tris-HCl for 4 days no fluorescence was detected in the BsLS-containing fractions from the size-exclusion column (Figure S12B). These observations suggest that the expanded form of the capsid has little-to-no affinity for GFP11. When a similar mixture is dialyzed from 100 mM sodium phosphate at pH 7.0 to 100 mM TrisHCl at pH 8.0 and then dialyzed back into 100 mM sodium phosphate at pH 7.0, two capsid peaks are seen, but they both lack significant fluorescence (Figure S12C). Expanding the capsid and then returning the buffer to a condition where the encapsulation complex is stable does not lead to accumulation of GFP11 in either form of the capsid. Interestingly, dialysis of BsLS into 100 mM Tris-HCl at pH 8.0 for four days gave quantitative conversion to a single early-eluting peak in the UV chromatogram from the sizeexclusion column, whereas a two-day dialysis gave two major peaks (Figure S12B vs. Figure S12C and Figure 6C). This behavior suggests that the conversion of the original capsid to the fully expanded form is slow. Nevertheless, the structural rearrangements leading to the expanded form of the capsid are irreversible.

DISCUSSION The encapsulation of protein guests within molecular containers has attracted a lot of interest in nanotechnology as a way to build nanoreactors or other functional architectures 55-59. Similar to the BsLS+GFP11 encapsulation complex described above, several other engineered protein encapsulation systems also utilize a fragment of a naturally encapsulated protein to target the


14

Page 15 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

guest to the interior of the Eut bacterial microcompartment 20, the P22 bacteriophage capsid 13, 14, or encapsulin10, 60, 61. Using loading strategies based on in vivo co-production of the capsid and guest, the latter two containers have reported loading yields as high as 350 and 20 guests per capsid, respectively. While the loading yield found in this study is lower than see for those other encapsulation systems, the BsLS capsid is also smaller than those other capsids. Interestingly, the confinement molarity (Mconf) for a single GFP molecule encapsulated by BsLS is around 5 mM, which is quite similar to the Mconf values of the P22 capsid and encapsulin at their highest reported levels of guest loading (7 mM and 6 mM, respectively) 10, 14. It could be that molecular crowding effects impose an effective limit on the number of guest molecules that can be loaded into a single capsid during bacterial co-production, at least for naturally evolved peptide-based encapsulation tags. Steric hindrance does not seem to be the limiting factor for the loading of GFP into BsLS capsids. The observed loading of 1.3 GFP11 molecules per capsid gives a packing density of 0.16, which suggests that there is sufficient space inside the capsid to accommodate the loading of a second (or even a third) GFP11 guest. Further, the trimer of full-length BsRS, which is the natural guest of BsLS, has a molecular weight of 66 kDa, which is more than double that of GFP11 (29 kDa). The pre-association of BsRS trimers may be responsible for its higher loading yield relative to monomeric GFP11. Analogously, non-covalent dimerization of a fluorescent protein guest has been shown to increase encapsulation yield by the CCMV capsid 62. Since a single BsRS-derved encapsulation tag provides a loading yield of >1 guest per capsid for BsLS+GFP11, only one of the three C-terminal tails should be sufficient to encapsulate all three subunits of full-length BsRS. Thus, oligomerization allows multiple protein subunits to act a


15

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 42

single guest, circumventing the apparent molecular crowding limit (Mconf ~7 mM) faced by monomeric guests. It may be that the trimeric assembly of BsRS further helps promote its encapsulation by BsLS by enabling multivalent interactions. The interior surface of BsLS contains 20 three-fold symmetric interfaces. The C-terminal tail of BsRS is likewise located near its three-fold symmetry axis. Thus, symmetry matching between the capsid and its natural cargo might give rise to trivalent interactions between the C-terminal tails of the BsRS guest and their binding sites on the inner surface of BsLS. Surprisingly, GFP32, which includes both the α-helical trimerization domain of BsRS and the C-terminal tail, gets encapsulated about two-fold less efficiently than GFP11. The lower amount of capsid loading can be at least partly explained by the lower production levels of both the capsid and guest during the coproduction of BsLS and GFP32. Since both variants elute similarly, and similar to GFP0, from two different sizeexclusion columns (Sephacryl S-400 and Superdex-75), the putative α-helical segment of BsRS present in GFP32 is insufficient to promote trimerization of GFP under the conditions examined. Therefore, it remains unclear whether the trimeric assembly of wild-type BsRS contributes to its affinity for the BsLS interior. Interestingly, the BsRS-derived encapsulation tag seems to load GFP into the BsLS capsid more efficiently than a recently reported encapsulation system based on the homologous AaLS capsid and a C-terminal fragment of AaRS 45. In the latter system, co-production of AaLS and GFP fused to the last 12 amino acids of AaRS (AaRS196-207) gave a loading yield of 0.47 GFP guests per capsid, which is 2.7 times lower than that of BsLS+GFP11. The basis for this difference is unclear, but might reflect a greater intrinsic affinity of the BsRS205-215 sequence for


16

Page 17 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

the BsLS capsid. In both systems, extending the RS-derived fragment to include the coiled-coil trimerization domain does not improve encapsulation efficiency. It is not immediately obvious which residues of the BsRS encapsulation tag are important for binding to the inner surface of the BsLS capsid. A sequence alignment of 24 RS’s (Figure S13) confirms that a C-terminal extension is preferentially conserved in encapsulated RS’s compared to non-encapsulated varieties. Further, the C-terminal tails of encapsulated RS’s are generally amphipihilic, with a consistent hydrophobic/hydrophilic pattern. Within these tails, the non-polar amino acids show greater conservation than the polar amino acids. Given that GFP11 does not show appreciable affinity for the AaLS capsid, it may be that hydropobic interactions provide a general source of affinity between the capsids and the encapsulation tags, while electrostatic interactions provide specificity. The location of the binding site on the capsid for the BsRS195-205 peptide is not clear. Visual inspection of the BsLS structure reveals both hydrophobic and hydrophilic patches along the inner capsid surface (Figure S14A). Notably, a non-polar cluster surrounds the three-fold symmetry axis (Figure S14B & S14C) while the remaining surface is composed of polar residues (both charged and neutral). A sequence alignment of 24 LS’s (Figure S15) shows that this non-polar cluster, consisting of I121 and I125 in BsLS, is well-conserved among all capsid forming LS’s. Although these residues play an important role in stabilizing the icosahedral capsid assembly47, it is tempting to speculate that these non-polar clusters could also provide an interaction surface for one or more of the conserved hydrophobic residues in the BsRS195-205 peptide. A few other amino acids along the inner capsid surface, A85, T130, K131, and A132 in BsLS, are preferentially conserved among LS’s that encapsulate RS compared to those that do not; however, the sequences of the two capsid-forming LS’s that have been shown not to


17

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 42

encapsulate RS suggest that they might primarily serve as determinants of capsid assembly. Elucidation of the contributions made by the individual amino acids in the BsLS+BsRS195-205 complex will require further study. Most engineered protein encapsulation complexes, including those involving the AaLS capsid, are dead ends whose disassembly requires harsh denaturating conditions. The BsLS+GFP11 complex does not dissociate under the conditions used for its purification. Indeed, the crystal structure of the BsLS capsid reveals that the largest pores in the capsid, which are located at the five-fold symmetry axis, are

Encapsulation and Controlled Release of Protein Guests by the

Recommend Documents