Active Template Synthesis of Protein [ <i>n</i> ]Catenanes Using Engineered Peptide–Peptide Ligation Tools
Fan Zhang, Yajie Liu, Yu Shao, Wenbin Zhang
Abstract
Open AccessCCS ChemistryRESEARCH ARTICLES3 Feb 2024Active Template Synthesis of Protein [n]Catenanes Using Engineered Peptide–Peptide Ligation Tools Fan Zhang, Yajie Liu, Yu Shao and Wen-Bin Zhang Fan Zhang Beijing National Laboratory for Molecular Sciences, Key Laboratory of Polymer Chemistry and Physics of Ministry of Education, Center for Soft Matter Science and Engineering, College of Chemistry and Molecular Engineering, Peking University, Beijing 100871 , Yajie Liu Beijing National Laboratory for Molecular Sciences, Key Laboratory of Polymer Chemistry and Physics of Ministry of Education, Center for Soft Matter Science and Engineering, College of Chemistry and Molecular Engineering, Peking University, Beijing 100871 , Yu Shao Beijing National Laboratory for Molecular Sciences, Key Laboratory of Polymer Chemistry and Physics of Ministry of Education, Center for Soft Matter Science and Engineering, College of Chemistry and Molecular Engineering, Peking University, Beijing 100871 and Wen-Bin Zhang *Corresponding author: E-mail Address: [email protected] Beijing National Laboratory for Molecular Sciences, Key Laboratory of Polymer Chemistry and Physics of Ministry of Education, Center for Soft Matter Science and Engineering, College of Chemistry and Molecular Engineering, Peking University, Beijing 100871 Beijing Academy of Artificial Intelligence, Beijing 100084 https://doi.org/10.31635/ccschem.023.202302762 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail The expansion of protein topological diversity requires new and efficient synthetic tools. Herein, we report the second and third generations of the SpyStapler-mediated SpyTag/BDTag ligation system for the efficient synthesis of 4-arm star proteins and the repurposing of the third generation as an active template to enable the synthesis of higher-order protein [n]catenanes (n = 3, 4, and 5). SpyStapler003 has a higher affinity to its cognate SpyTag and BDTag reactive pairs relative to the original SpyStapler. Hence, it can overcome much more profound steric hindrance in protein ligation and improve the efficiency of the resulting active template tool to facilitate the construction of radial protein [n]catenanes. Various proteins of interest, such as dihydrofolate reductase and the nanobody KN035, can be modularly incorporated into the [n]catenanes with intact activity. Combination of passive and active template strategies gives rise to linear protein [4]catenanes, which further expands the current topological diversity. Moreover, higher-order protein catenation not only leads to enhanced thermal stability and proteolytic resistance but also higher affinity of the nanobody via multivalent effects. Our study provides tools useful for bioconjugation and new topological protein scaffolds for the multivalent display of enzymes and antibodies. Download figure Download PowerPoint Introduction Chemical topology, which concerns the connectivity and spatial relationship of molecules,1 has emerged as a unique dimension in protein engineering.2,3 Natural proteins with nontrivial topologies often display functional benefits associated with enhanced stability and reinforced bioactivities.4–10 Theoretically, topology can impart delicate yet profound effects on the structure-property relationships, leading to useful property improvements such as stability enhancement, dynamic transition, multivalent effects, etc.2 However, the potential of protein topology engineering remains largely untapped, probably due to the grand challenge of the synthesis of various topological proteins. The past few decades have witnessed significant progress in molecular nanotopology, and numerous strategies have been developed for the synthesis of molecules with nonplanar topology.11,12 Two factors have been recognized as essential for precise construction of molecular nanotopology: (1) site-specific ligation to define valid connectivity and (2) entanglement among molecular segments to define spatial relationships. Owing to efficient and specific chemical reactions and supramolecular interactions, a burst of topological diversity has been achieved in synthetic molecules.13–21 Translating these tactics in the context of protein science requires genetically encoded protein tools that are capable of accomplishing the above tasks. The linear nascent protein from ribosomal synthesis can then be converted into branched and/or cyclic structures,22–24 or even knots and links,25–29 which greatly boost the exploration of topological proteins. Genetically encoded protein ligation tools offer reliable control of the protein assembly with covalent bonding toward user-defined properties.30 To date, they have mainly consisted of four categories: (1) ligation based on unnatural amino acid (UAA) incorporation,31 (2) enzyme-mediated ligation,32–36 (3) split-intein technology,37–39 and (4) peptide-protein reactive pairs.40,41 Unlike UAA-mediated ligation, the latter three methods are all based on natural amino acids, which circumvent problems associated with laborious genetic manipulation, low protein yields, and compromised activity. While enzyme-mediated ligation mandates external manipulation, both split-intein technology and peptide-protein reactive pairs are spontaneous and fully compatible with cellular environments. Among them, the peptide-protein reactive pairs, as exemplified by the gold standard SpyTag/SpyCatcher reaction,42 feature fast, irreversible, and stable side-chain crosslinks that are powerful for diverse applications,40,43 including protein bioconjugation,44–47 protein covalent assemblies,48–53 and protein-based biomaterials.54–59 Combining rational design and phage display, Howarth and coworkers60,61 developed the second and third generations of the SpyTag/SpyCatcher reactive pair where the latter exhibits almost infinite affinity approaching the diffusion limit. To reduce the size of the reactive tags, SpyCatcher can be further split into a catalytic domain and a reactive peptide tag, giving rise to the SpyTag/KTag/SpyLigase62 and SpyTag/BDTag/SpyStapler63 ligation systems. Using these tools, H-shaped proteins,22 star proteins22,29,63 and cyclic proteins22,63–67 have been facilely prepared. Combining genetically encoded entangling motifs with these reactive pairs generates considerable topological diversity. Using passive templates such as p53dim entangling dimer, complex topological proteins, such as protein catenanes,29,68–73 lasso proteins,74 and protein pretzelanes75 have been prepared. The emergence of useful tools (such as engineered p53 heterodimer)76 and strategies (such as in situ proteolytic digestion)27,77 has facilitated the cellular synthesis of even more complex protein topologies, such as protein [n]catenanes and protein heterocatenanes. Active template has been recognized as a powerful strategy to construct mechanically interlocked molecules.21,78 By rewiring the SpyTag/BDTag/SpyStapler complex to introduce artificial chain entanglement, we were able to report the first protein active template (AT-Spy) and its application in preparing protein heterocatenanes.28 A cyclic SpyStapler obtained via split-intein-mediated ligation can reconstitute with a telechelic protein bearing BDTag at its N-terminus and SpyTag at its C-terminus to form an entangled structure. Subsequent isopeptide bond formation leads to a protein heterocatenane. Various proteins of interest (POIs) may be incorporated on to the scaffold. Nonetheless, to fully unleash its potential in topological synthesis, the efficiency needs to be further improved, especially for the synthesis of higher-order protein [n]catenanes. In this contribution, we report engineered SpyTag/BDTag/SpyStapler variants with enhanced reactivity to improve active template synthesis of protein higher-order [n]catenanes in high yields (Scheme 1a–c). We show that only a few mutations lead to dramatically improved ligation yield and that higher-order catenation by active template tolerates the insertion of distinct proteins, offering novel scaffolds for multivalent display of enzymes and nanobodies. Scheme 1 | (a) SpyStapler-mediated isopeptide bond formation between SpyTag and BDTag. (b) Synthesis of protein [2]catenane using AT-Spy. (c) Synthesis of protein [n]catenanes (n = 2–5) using AT-Spy003 and schematic representation of the constructed [n]catenanes (n = 2–5). Download figure Download PowerPoint Experimental Methods All reagents and kits were purchased from commercial suppliers. The details about DNA construction, protein sequences, protein expression, and purification can be found in the Supporting Information. Synthesis of protein [n]catenanes After size exclusion chromatography (SEC) purification, c-(SpyStapler003)m was mixed with BDTag003-DHFR/KN035-SpyTag003 (DHFR = dihydrofolate reductase) in a ratio of 1:m. The mixture was allowed to react overnight at 4 °C. For example, 10 μM c-(SpyStapler003)3 was added to 30 μM BDTag003-DHFR-SpyTag003 for overnight incubation at 4 °C to construct [4]cat-DHFR. Tobacco etch virus and tobacco vein mottling virus protease cleavage Proteolytic digestion was carried out at 30 °C. The protein solution (10 μM) was mixed with tobacco etch virus (TEV) protease at a molar ratio of 20:1 and/or tobacco vein mottling virus (TVMV) protease at a molar ratio of 10:1. The resulting products were analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and further confirmed by liquid chromatography-mass spectrometry (LC-MS) (Waters Corp., Milford, Massachusetts, United States). Trypsin digestion Protein substrate (10 μM) was mixed with trypsin working solution (4 μM in phosphate-buffered saline (PBS)) at a final molar ratio of 100:1. Digestion was carried out at 22 °C. After a certain time period, 20 μL mixture was picked up and quenched by 5× loading buffer. All samples were analyzed by SDS-PAGE and relatively quantified using Image J. Thermal shift assay Thermal shift assays of all samples were conducted on the StepOnePlus™ Real-Time polymerase chain reaction system (Applied Biosystems Inc., Foster City, California, United States). 18 μL of each sample (5 μM in PBS) was mixed with 2 μL of 10× SYPRO orange dye and then added into an opaque white 96-well plate. The samples were then heated from 30 to 80 °C with carboxy-X-rhodamine as the fluorescent reporter. The raw fluorescence data were nonlinearly fitted to a Boltzmann sigmoidal curve, and the Tm of protein samples was calculated. Biolayer interferometry analysis The Octet RED96e instrument (Fortebio, Menlo Park, California, United States) was used at 30 °C with a shaking speed of 1000 rpm. Protein samples were diluted with PBS with with Tween 20 (PBST) buffer (PBS, 0.05% Tween-20, 0.1% bovine serum albumin) to 50, 25, 12.5, 6.25, and 3.125 nM. Streptavidin sensors were first immersed in PBST buffer for 60 s, then soaked with 5 μg/mL biotinylated PD-L1 solution for 50 s to load a sufficient amount of PD-L1, then dipped in PBST buffer for 60 s to wash out extra PD-L1. The association phases started by immersing these PD-L1 modified sensors respectively in samples at different concentrations for 120 s, and the dissociation phases started by immersing the above sensors into PBST buffer for 200 s. 100 mM glycine-HCl buffer (pH 2.2) was used for sensor regeneration. Curve fitting was performed using Octet Data Analysis software. DHFR catalytic activity assay Samples of DHFR series in stock solution (PBS, 2 mM dithiothreitol, pH 7.4) were diluted with potassium phosphate buffer (KHP) (40.1 mM K2HPO4, 9.9 mM KH2PO4, 5 mM β-mercaptoethanol, pH 7.5) to 10 μM except for DHFR catenanes (3.33 μM for [4]cat-DHFR, 2.5 μM for [5]cat-DHFR). The diluted samples were heated at the designated temperature for various times, then put into 4 °C for overnight incubation. Then the samples were mixed with 100 μM nicotinamide adenine dinucleotide phosphate (NADPH) (in KHP buffer) and 100 μM dihydrofolic acid (in KHP buffer) at a final concentration of 30 nM in a 96-well transparent microplate, and the absorbance at 340 nm was immediately monitored using an EnSpire multimode plate reader (PerkinElmer Inc., Waltham, Massachusetts, United States) by kinetic mode. The absorbance at 340 nm of NADPH with varied amounts (0, 10, 20, 30, 40, 60, 100 nmol) in 200 μL KHP buffer were also measured for the standard curve. The linear range of sample plots was utilized to determine DHFR activity. Results and Discussion Engineered SpyTag/BDTag/SpyStapler ligation system Compared to SpyTag/SpyCatcher, the SpyTag/BDTag/SpyStapler ligation system reduces the reaction scar and the size of the protein tag but suffers from a slower reaction rate and lower yield. Since their chemical reactivity is encoded in sequences, it is possible to manipulate the reactivity via sequence variation. Inspired by the second and third generations of SpyTag/SpyCatcher reactive pairs,60,61 we envisioned that these mutations should also benefit the SpyTag/BDTag/SpyStapler ligation, generating the corresponding SpyTag/BDTag/SpyStapler 002 and SpyTag/BDTag/SpyStapler 003 (Figure 1). It should be noted that BDTag003 has the same amino acid sequence as BDTag002. The SpyStapler variants were then obtained by expression. Despite similar molecular weights (Figure 2c), these SpyStapler variants showed very different mobility on SDS-PAGE and elution profiles in SEC (Figure 2b and Supporting Information Figure S1), which is attributed to the higher negative charge density and better aggregation resistance of SpyStapler002 and SpyStapler003 than the original SpyStapler. Figure 1 | Sequence alignment of three generations of the SpyTag/BDTag/SpyStapler ligation system. The catalytic Glu, the reactive Asp, and the reactive Lys are underlined. Amino acids appearing in SpyTag/BDTag/SpyStapler are highlighted in yellow, amino acids that first appeared in SpyTag/BDTag/SpyStapler 002 are highlighted in red, and amino acids that first appeared in SpyTag/BDTag/SpyStapler 003 are highlighted in blue. Gray boxes show the negatively charged amino acids in the three generations of SpyStapler. Download figure Download PowerPoint We first compared their difference in reactivity using the model SpyStapler-mediated reaction between YFP-SpyTag (YFP-A) and BDTag-CFP (BD-CFP) whose reactive tags are at relatively exposed terminal regions (Figure 2a and Supporting Information Figure S2, see Supporting Information for sequence information). For simplicity, A stands for SpyTag and BD stands for BDTag. The reactions proceeded smoothly at 4 °C. After only 10 min, the yield was 20% for the original SpyStapler but increased to 30% for SpyStapler002 and further to 56% for SpyStapler003 (Figure 2b). The second-order rate constants were determined to be 153 M−1·s−1 for SpyTag/BDTag/SpyStapler 002 and 522 M−1·s−1 for SpyTag/BDTag/SpyStapler 003 ( Supporting Information Figure S3), which are much higher than the reported 36 M−1·s−1 for the original system. The increased reactivities of 002 and 003 systems might be a result of enhanced electrostatic interactions between SpyStapler variants and SpyTag/BDTag variants, as the mutations bring more negative charges on SpyStapler variants and more positive charges on SpyTag variants ( Supporting Information Table S1). Figure 2 | (a) Schematic representation of SpyStapler-mediated ligation between YFP-A and BD-CFP. (b) SDS-PAGE analysis of reactions in panel (a) after 10 min. (c) LC-MS spectra of three SpyStapler variants. (d) Schematic representation of SpyStapler-mediated ligation between ELP-A-ELP and ELP-BD-ELP. (e) SDS-PAGE analysis of reactions in panel (d) after 6 h. (f) Yields of reactions in panel (d) at different reaction moments. Download figure Download PowerPoint We further challenged the system with internal ligation for star protein synthesis. Previously, the reaction between ELP-A-ELP and ELP-BD-ELP with reactive tags in the middle of a disordered elastin-like polypeptide (ELP) was known to be sluggish for SpyStapler due to huge steric hindrance, especially in the absence of chemical chaperones like trimethylamine N-oxide.63 The yields at 4 °C for 6 h were dramatically improved with SpyStapler003 showing the highest efficiency (Figure 2d–f). The MS spectra of these reactants and coupling products can be found in Supporting Information Figure S4. It seems that the bulky ELP coils near the tags have little influence on the reaction yield of SpyTag/BDTag/SpyStapler 003, except that the reaction rate is slowed down to a certain extent. Using this model reaction, we further examined the effects of temperature and stoichiometry on SpyStapler003-mediated ligation. SpyStapler003 exhibits better activity at lower temperatures (4–16 °C) and an equivalent amount of SpyStapler003 is sufficient to produce decent yield ( Supporting Information Figure S5). Based on the above results, we chose the most efficient SpyTag/BDTag/SpyStapler 003 system for further experiments. Model active template reaction The active template integrates entangling and reaction into one single motif. To examine the effect of SpyStapler003 for cyclization, we applied 2 equiv of SpyStapler003 on DHFR bearing BDTag003 at N-terminus and SpyTag003 at the C-terminus. Cyclization almost reached completion within 10 min at 4 °C ( Supporting Information Figure S6). Encouraged by this result, we brought in the mutations of SpyTag/BDTag/SpyStapler 003 to improve the efficiency of the AT-Spy. In the model reaction with POI being disordered ELP, AT-Spy003 shows better yield than AT-Spy and exhibits a weak temperature dependence (Figure 3a–c and Supporting Information Figure S7). Switching the POI to folded DHFR leads to even higher yields, and the topology of [2]cat-DHFR was proved by orthogonal protease digestion ( Supporting Information Figure S8). To assay whether function could be preserved during active template synthesis, we designed a cyclic protein c-SS003-GFP (GFP = green fluorescent protein) and a telechelic protein BD003-KN035-A003 (Figure 3d). GFP is a superfolder GFP,79 and KN035 is an anti-PD-L1 nanobody.80 The heterocatenane ([2]cat-GFP-KN035) was successfully prepared in 58% yield, as proved by SDS-PAGE (Figure 3e) and LC-MS ( Supporting Information Figure S9). The [2]cat-GFP-KN035 not only exhibits similar fluorescent properties to wt-GFP ( Supporting Information Figure S9), but it also retains the PD-L1 affinity of KN035 (Figure 3f and Table 1). It is evident that heterocatenation by AT-Spy003 does not affect the component proteins' properties. Figure 3 | (a) Active template synthesis of [2]cat-ELP using equivalent amounts of c-ELP-SS003-ELP and BD003-ELP-A003. (b) SDS-PAGE analysis of the reaction in panel (a). (c) Yields of AT-Spy and AT-Spy003 reaction in the synthesis of [2]cat-ELP under different reaction temperatures. The consumption of BD-ELP-A and BD003-ELP-A003 was used to calculate the yields of [2]cat-ELP. (d) Active template synthesis of [2]cat-GFP-KN035 using equivalent amounts of c-SS003-GFP and BD003-KN035-A003. (e) SDS-PAGE analysis of the reaction in panel (d). (f) Biolayer interferometry analysis of [2]cat-GFP-KN035. Download figure Download PowerPoint Table 1 | Summary of Kinetic Binding Constants for wt-KN035, [2]cat-GFP-KN035, and [n]cat-KN035 Samples kon (M−1s−1) koff (s−1) KD (M) wt-KN035 (2.53 ± 0.02) × 105 (5.43 ± 0.09) × 10−4 (2.15 ± 0.02) × 10−9 [2]cat-GFP-KN035 (2.49 ± 0.01) × 105 (4.85 ± 0.06) × 10−4 (1.95 ± 0.04) × 10−9 [3]cat-KN035 (4.03 ± 0.02) × 105 (6.78 ± 0.68) × 10−5 (5.22 ± 0.17) × 10−10 [4]cat-KN035 (4.43 ± 0.02) × 105 (5.06 ± 0.73) × 10−5 (1.14 ± 0.17) × 10−10 [5]cat-KN035 (4.32 ± 0.01) × 105 (4.07 ± 0.35) × 10−5 (9.40 ± 0.80) × 10−11 Higher-order protein [n]catenane synthesis The synthesis of higher-order protein [n]catenanes has been hampered by the relatively low reaction efficiency of AT-Spy. We envisioned that [n]catenanes (n ≥ 3) might be obtained with engineered AT-Spy003 from a series of cyclic protein precursors containing different numbers of SpyStapler003 (Scheme 2). By mixing the c-(SpyStapler)m (m ≥ 2) with m equivalents of BD003-POI-A003, a series of protein [n]catenanes (n ≥ 3) could be obtained. As a proof of concept, we chose two POIs with different 3D structures, namely DHFR and KN035, to demonstrate the general applicability of the method. Scheme 2 | Synthesis of protein [n]catenanes (n = 3–5) using AT-Spy003. TEV and TVMV in the scheme represent the recognition sites of TEV and TVMV proteases, which are designed for topology characterization. Download figure Download PowerPoint We first conducted reactions between c-(SS003)m (m = 2, 3, and 4) and BD003-DHFR-A003 to build [n]cat-DHFR (n = m + 1). The major product was the desired [n]cat-DHFR, and the yields were 62%, 60%, and 40% for = 3, 4, and respectively (Figure are also with lower molecular weights corresponding to catenanes which could be by they form with similar The of [n]cat-DHFR confirmed their as an example, the SEC shows that product at a much lower with a (Figure The LC-MS confirmed that the molecular of was in with the (Figure The SEC and LC-MS of and are in Supporting Information Figure We then the of this higher-order catenation by DHFR with The [n]cat-KN035 samples were also obtained in yields (Figure and their was further confirmed by LC-MS (Figure and Supporting Information Figure Using equivalents of is sufficient for of [n]catenanes. While does not the yield of amounts of shift the products to catenanes ( Supporting Information Figure We that it is possible to construct protein via of different in but this to them, to their similar properties. Figure 4 | (a) SDS-PAGE analysis of the formation of [n]cat-DHFR (n = (b) SDS-PAGE analysis of the formation of [n]cat-KN035 (n = (c) SEC of and [4]cat-DHFR. (d) LC-MS and protein model of [4]cat-DHFR. (e) LC-MS and protein model of Protein are by the ( Schematic representation (f) and SDS-PAGE analysis of orthogonal protease digestion of [4]cat-DHFR. Download figure Download PowerPoint protease digestion were then conducted to their topology (Figure and Supporting Information Since these protein [n]catenanes are containing a mechanically interlocked with they similar of their digestion as an example, it was by TEV protease to cyclic and linear with TVMV linear and cyclic were obtained. with both to two linear proteins, namely and All the products were analyzed by SDS-PAGE (Figure and further proved by LC-MS ( Supporting Information Figure protein by passive and active templates Protein higher-order [n]catenanes can different from the branched construct it is also possible to build the [n]catenane in linear by passive and active templates in the synthesis. We first designed a protein containing one SpyStapler003 in each using p53dim as the passive template and split-intein-mediated covalent ligation. was mixed with BD003-DHFR-A003 to the protein in linear with yield (Figure To the topology, we performed orthogonal protease digestion and the product with the (Figure To the of this is the first protein linear to We that even more complex topological be from these yet synthetic tools. Figure 5 | (a) Schematic representation of the synthesis of via the of AT-Spy003 and passive template and the orthogonal protease digestion of (b) SEC of and the reaction (c) LC-MS of the (d) SDS-PAGE analysis of the orthogonal protease digestion of Download figure Download PowerPoint benefits of higher-order protein catenation has been utilized to the stability of functional proteins in different We first examined whether higher-order catenation could also boost resistance to [4]cat-DHFR, and were for proteins have similar temperatures for the DHFR domain ( Supporting Information Figure which catenation does not or As for the linear and BD003-DHFR-A003 most of their activity after incubation at 60 and catalytic activity after being heated for 30 min (Figure and Supporting Information Figure DHFR catenanes could even being heated for 20 min at 80 °C (Figure and Supporting Information Figure The that catenation the proteins' thermal and to better after which is in with Figure 6 | (a) of [4]cat-DHFR, and after incubation at 60 °C for to min. (b) of and after at different temperatures for 20 min. (c) Kinetic constants of wt-KN035, and (d) of intact protein in trypsin Download figure Download PowerPoint The protein [n]catenane provides an multivalent for on the The resulting [n]cat-KN035 samples can also be as artificial or enhanced PD-L1 affinity due to the multivalent effects with lower dissociation (Figure Table and Supporting Information Figure As the of [n]catenanes the affinity The dissociation of [5]cat-KN035 is lower than that of In catenation also to improved proteolytic resistance (Figure trypsin cyclic were the first to be and even it probably as a complex with as by the of the to BD003-KN035-A003 during proteolytic digestion ( Supporting Information Figure Hence, KN035 in the [n]catenane can a much time than in the linear are with of containing with improved affinity toward the