Bioinspired Spindles using Peptide-DNA Nanotechnology

doi:10.21203/rs.3.rs-1754570/v1

Download PDF

Article

Bioinspired Spindles using Peptide-DNA Nanotechnology

https://doi.org/10.21203/rs.3.rs-1754570/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 23 Apr, 2024

Read the published version in Nature Chemistry →

Version 1

posted

You are reading this latest preprint version

Cytoskeletal assemblies produce functionality through multiscale structural reconfiguration by exploiting a family of protein crosslinkers. Inspired by these principles, we designed an array of complementary peptide-DNA monomers able to form multistranded junctions with different geometries and stabilities and tested their effect on the structural organization of co-polymerized filament-forming peptides. In striking similarity to their biological counterparts, the peptide-DNA crosslinkers organized fibers into bundled architectures with base-pair tunable aspect ratios. Controlling solvent or temperature drove assembly through distinct pathways to form different structural and mechanical states exhibiting multiple levels of hierarchy from the same set of components. The thermal reversibility of DNA crosslinks drove structures to form, disassemble, and rebuild over multiple cycles. Sequential transitioning between bundles and fibers using orthogonal triggers enabled the dosed release of proteins to modulate catalytic reactions. This programmable peptide-DNA nanotechnology framework will pave the way towards the design of hierarchical materials with reconfigurable properties.

The assembly of proteins into multiscale architectures such as filaments, bundles, and networks is vital to many biological functions.¹ Intracellularly, filamentous actin and microtubules are crosslinked by associating proteins to form functional structures such as filopodia, contractile rings, and mitotic spindles that are crucial for cell motility and cell division.^2,3 These protein assemblies reorganize dynamically to alter their morphology and mechanical properties by tuning weak and reversible non-covalent interactions. Engineering materials with dynamic hierarchical assemblies mimicking those of biological proteins is of great interest. Supramolecular self-assembly provides a versatile approach to construct such architectures through rationally designed transient interactions. Numerous studies on protein fibril formation have informed the synthetic engineering of peptides into nanoscale morphologies such as fibers, sheets, spheres, and ribbons.^4–7 While the nanoscale assembly of peptides is extensively explored, strategies to organize them hierarchically^8–10 into functional mesoscale architectures akin to nature are still in their infancy.

A novel tunable strategy involves an emerging class of self-assembling monomers integrating peptides and DNA, where the biological and self-assembly properties of peptides are combined with the programmability and reversibility of DNA base-pairing.^11,12 Stupp et al decorated lipid peptides with complementary DNA strands and co-assembled them with peptide monomers.¹⁰ The DNA hybridization drove large scale monomer exchange to yield networks with DNA enriched bundles and DNA-depleted fibers. Our group subsequently showed that tuning the length of single-stranded DNA arms on short amyloid peptides guided their nanoscale assembly into twisted fibers, and that mixing peptides bearing complementary oligonucleotides drove a fiber to bundle transition.¹³ As DNA crosslinks can be designed to form complex junctions beyond the simplest double helix (or 'duplex'), it allows exploration of a broader parameter space defined by crosslinker binding strength and connective regions.^14–16

Here, we rationally designed a library of DNA-crosslinks which are covalently attached to self-assembling peptides via copper-free click chemistry.^13,17 By modifying various DNA sequences at either 5’ or 3’ ends with peptides, we generated a set of unique DNA-crosslinked peptide arrangements and co-polymerized them with peptide monomers to examine their impact on the evolution of hierarchical structures. Interestingly, the DNA-crosslinked peptides reconstructed many of the morphological and mechanical behaviors consistent with protein-crosslinked microtubules and actin networks. We found that multivalent DNA-peptide junctions generated spindle-like structures that organized further into bundles of tactoids, where the double-stranded DNA density within the bundle correlates with its size. The aspect ratio of the tactoids can be tuned by DNA-crosslinker geometry and stability and can be further modulated by solvent and temperature. Finally, the design of multistranded DNA crosslinkers allowed orthogonal reconfiguration switches (DNA displacers) to control the extent of disassembly, which we used to achieve a dosed release of enzymatic payloads triggering catalytic reactions. This hybrid peptide-DNA nanotechnology approach for programmable material reconfiguration can pave the way towards constructing highly adaptive synthetic cytoskeletal assemblies.

Peptide-DNA crosslinks tune hierarchical formation of spindle-like assemblies

Cytoskeletal actin associates with a myriad of actin-binding crosslinkers with different geometries and flexibilities to form meso-scale structures such as stress fibers, filopodia, and the actin cortex (Fig. 1).^2,3 Inspired by these principles, we generated a library of peptide-DNA monomers^10,13,17 consisting of fiber-forming dipeptide (Fmoc-FF)^18,19 and complementary DNA sequences of various lengths (Fig. S1-3, Table S1). Copolymerization of peptide monomers with interacting peptide-DNAs allows us to rationally tune the resulting double-stranded DNA (dsDNA) junctions, achieving weakly to strongly crosslinked peptide constructs (Fig. 1, Fig. S4).

To evaluate the emergent effect of DNA-crosslinkers on the structural organization of peptides, we prepared solutions of 0.1 wt% peptide (1.87 mM) containing 1 mol% of each peptide-DNA monomer. The various peptide/peptide-DNA mixtures (in 1% DMSO/water) were annealed to optimize DNA hybridization (see Methods) and imaged via confocal microscopy (Fig. 2a,b(i)). In the absence of crosslinkers, the peptides formed nanoscale filaments, as expected.^13,18 Interestingly, as DNA-crosslinkers with linear (A’, A-A’) and branched (A-B-C, and AY-BY-C) junctions were introduced, hierarchical bundles with higher degree of alignment evolved (Fig. 2b(ii)). Quantification of the length, diameter (D_B) and aspect ratio (AR) of the bundles (Fig. 2c-e) revealed two distinct hierarchical states, thin-bundles and large-bundles which are shown as points within the shaded area and outside of the shaded area, respectively (Fig. 2d).

Surprisingly, thin-bundles adopted a spindle-like tactoid structure²⁰ (Fig. 2b(iii), S5) with remarkable similarity to reconstituted actin and microtubule bundles.^21,22 Their ARs decreased from linear to multi-stranded junctions (Fig. 2c,d). This behavior is analogous to crosslinked cytoskeletal assemblies, where the ARs of F-actin and microtubule tactoids were shown to decrease with increasing protein-crosslinker concentration.^21,22 Moreover, TEM revealed that the DNA-crosslinked tactoids consisted of hierarchically-organized and aligned filaments (Fig. S6), suggesting that DNA-crosslinking introduces a preferential inter-filament orientation and alignment.

Structural characterization of large-bundles revealed that increasing the DNA-crosslinker base-pairing from 8 (A’) to 14 (A-A’) yielded thicker bundles with lower ARs (Fig. 2d,e). The trivalent crosslinkers (A-B-C and AY-BY-C) lowered the bundle ARs even further. Intensity line scans across the width of large-bundles in A-A’ (Fig. 2f) and A-B-C (Fig. 2g) from confocal images showed multiple peaks with similar size to the width of thin-bundles (~ 1 µm), suggesting that large-bundles consist of multiple spindle-like thin-bundles organizing in the next level of hierarchy. These results show that crosslinker design using peptide-DNA nanotechnology can recapitulate behaviors arising from natural filaments and associating proteins and program multiple levels of supramolecular hierarchical organization from fibers (1st level) to tactoids (2nd level), to bundled tactoids (3rd level), and finally, networks of bundles (4th level). It also provided access to structural spaces that thus far were not accessible (peptide-DNA tactoids, bundled tactoids, etc.)

Filament length and crosslinker flexibility and density tune hierarchical organization

It was previously shown that actin polymer length tunes the dimensions of spindle-shaped bundles.²¹ We therefore evaluated the effect of peptide-DNA filament length (l_f) on the resulting hierarchical structures. As supramolecular polymerization of peptides is often sensitive to solvents,²³ we explored this strategy to achieve different l_f. We hypothesized that water (versus DMSO/water) would promote longer assemblies due to faster hydrophobic collapse.^24–26 Confocal images confirmed that in the absence of crosslinkers, water assemblies yielded longer fibers (average length of 17.3 versus 11.0 µm in DMSO/water conditions) (Fig. S7). Interestingly, while linear (A-A’) and branched (A-B-C) crosslinkers in water yielded both thin and large-bundles, these bundles exhibited high ARs and did not adopt a spindle shape (Fig. 3a,b). Also, TEM images showed that the bivalent A-A’ crosslinker drove a higher degree of filament ordering within bundles than the trivalent A-B-C crosslinker (Fig. 3c,d, S8b). To isolate the differences in valency and duplex stability between A-A’ and A-B-C, we tested a bivalent crosslinker with a longer duplex similar to A-B-C (C-C’: 46 bp vs. A-A’: 14 bp). Interestingly, the longer crosslinker (C-C’) yielded thin filament networks rather than bundles (Fig. 3a,c). These results recapitulate the behavior of actin crosslinkers, where short antiparallel protein crosslinkers promote tight bundle formation, while longer rigid crosslinkers procure looser filamentous networks,² suggesting the importance of a tight interplay between binding stability and flexibility of crosslinkers. To explore that, we designed an additional crosslinker (A-C) consisting of a short 15 bp-duplex with a flexible single-stranded DNA (ssDNA) overhang (31 bases) (Fig. S8a). We found that the DNA-crosslinker flexibility increases D_B but with a lower degree of alignment than A-A’, similar to actin crosslinkers (Fig. S8b). In striking contrast to DMSO/water conditions, the most rigid, trivalent AY-BY-C crosslinker in water yielded very thin (2–3 fibers across) and long bundles (97.4 µm average) with a high degree of alignment (Fig. 3a-c, S8b-d). We speculate that as l_f increases, their arrangement into large hierarchical bundles requires longer, more flexible crosslinkers to allow for a minimally constrained, long-range attractive interaction between filaments.

As the interplay between l_f and crosslinking density (C_cross) was shown to regulate bundle dimensions of actin,²¹ we wanted to evaluate C_cross in the various structural states. To do this, we stained crosslinked architectures with acridine orange (AO). As AO fluoresces red when bound to ssDNA and green when bound to dsDNA, it allowed us to visualize and estimate the degree of DNA hybridization within bundles from fluorescence intensities in these two spectral regions (green channel: I_green, red channel: I_red) (Fig. 3a). We quantified I_green-βI_red values, which correspond to the difference of fluorescence intensities of dsDNA and ssDNA (I_dsDNA and I_ssDNA), I_green-βI_red=I_dsDNA-βI_ssDNA, accounting for the background non-specific peptide-AO interactions with the average ratio between green and red channels, β = I_green,FF/I_red,FF (Fig. S9, see Methods and SI). Thus, high I_dsDNA-βI_ssDNA values should indicate high C_cross. We generated color maps and histograms of I_dsDNA-βI_ssDNA values from the confocal images (Fig. 3e-h). The map of I_dsDNA-βI_ssDNA for A showed negative values (no dsDNA), in line with ssDNA-peptide fibers (Fig. 3e,h). All DNA-crosslinked architectures showed positive average values of I_dsDNA-βI_ssDNA, confirming the presence of DNA hybridization (Fig. 3h). We observed dsDNA-rich regions in large structures and ssDNA-rich regions in thinner filamentous structures in insets of A-A’ images (Fig. 3f). To evaluate whether C_dsDNA correlates with D_B, we plotted the D_B distribution from A-A’ against I_dsDNA-βI_ssDNA values from the long axis of the bundles (Fig. 3g). Interestingly, we found that I_dsDNA-βI_ssDNA increased with D_B. In addition, AY-BY-C generated structures with the highest C_dsDNA, possibly because of more double-stranded arms within the Y-shape junction. These results show that larger bundles are enriched with dsDNA, and that C_cross can be tuned by the DNA junction design.

Programmable structural hierarchy with tunable mechanical states

The structural polymorphism of actin is of crucial importance to the mechanical behaviors of cells.^3,27 As our biomimetic filament-crosslinker library can reconstruct the co-existence of bundles and networks and generate spindle-shaped bundled networks not realized by natural proteins, we studied the mechanical properties of these various states. We performed frequency-sweep experiments on gels prepared in DMSO/water (Fig. 4a). Peptide-only gels (“FF”) had the lowest shear modulus (G_ω) of all the samples. Interestingly, linearly crosslinked gels (A’ and A-A’) exhibited 10 times higher G_ω than branched-crosslinked gels (A-B-C and AY-BY-C), although they consisted of thinner bundles. Prior studies on cytoskeletal bundle networks showed that their elasticity is determined by two factors, the average axial distance between crosslinks (along a fiber), l_c, and the bending stiffness of bundles, κ_B.^28,29 Bundle networks become stiffer when κ_B increases and l_c decreases, which corresponds to a dense network composed of rigid bundles. From our rheology experiments and confocal imaging, we observed increases in D_B and κ_B but decreases in G_ω from bivalent to trivalent crosslinkers with equal DNA-crosslinker concentrations (Fig. 4a). These results suggest that trivalent crosslinkers produce large-bundles, enriched with DNA-crosslinks, but with looser networks (high l_c) (Fig. 4b).

To further dissect the distinct effect of bundling from network formation on the viscoelastic properties, we monitored G_ω during annealing (Fig. 4c). FF showed no temperature dependence, suggesting that the network was fully formed at elevated temperatures. As DNA hybridization is sensitive to its melting temperature (T_M), we monitored G_ω of A-A' during cooling to 25°C either from above (80°C) or from below the T_M (60°C) at -1°C/min (Fig. 4c,d). Cooling from 80°C (> T_M) gave rise to networks with higher G_ω (3-fold) than cooling from 60°C (< T_M). Imaging revealed that this difference in mechanical states is due to different structural states, as cooling from 60°C yielded higher-AR spindles than the sample cooled from 80°C (Fig. 4d-f). These results suggest that temperature control of DNA-crosslinking can be programmed to give rise to different structural and mechanical states from the same building blocks.

Next, we measured G_ω of the trivalent A-B-C and AY-BY-C crosslinked gels during annealing. We found three distinct regimes (Fig. 4c). G_ω do not monotonically increase at the beginning of annealing (Region I’), followed by monotonic increase from around 65°C for A-B-C and 70°C for AY-BY-C (Region I). They reached plateaus at relatively lower temperatures than A’ and A-A’ (Region II). This rheological behavior can also be understood by the balance between bundling and network formation. As simulated in Fig. S11, B-C and BY-C duplex arms are formed at 80°C before A-B-C and AY-BY-C forms at 70°C and 65°C, respectively. At the beginning of the annealing process, where B-C and BY-C are formed, G_ω do not increase, suggesting that DNA hybridization is first involved in nucleating bundles and not networks. After the bundle formation, unsaturated A-B-C and AY-BY-C start to form crosslinked networks, showing monotonic G_ω increase. Overall, the non-monotonical G_ω increase during annealing is a characteristic rheological behavior of these trivalent crosslinkers. This series of structural and mechanical characterizations showed that, by controlling the architecture of DNA-crosslinkers, we can give rise to distinct structural and mechanical states, in striking analogy to actin crosslinkers.^2,29 For instance, fascin proteins organize actin filaments into bundles to regulate cell migrations, while filamin proteins crosslink actin filaments into a network and are commonly found in the actin cortex in living cells.^30,31

Transient DNA crosslinks impart heat reversible mechanics

As DNA hybridization is thermally reversible, we tested how temperature affects the mechanics of the DNA-crosslinked peptides. We first heated pre-assembled A’ to 40°C (+ 1°C/min) and cooled it to 25°C (-1°C/min) (Fig. 5a, S12). Repeated thermal cycling between 25°C and 40°C showed excellent reproducibility of G_ω (Fig. 5a, S12). Even a third heat-cool cycle of A’ (to 50°C and back to 25°C) re-generated the initial G_ω (Fig. 5b). FF exhibited no temperature dependence (Fig. 5a), suggesting that DNA crosslinkers gave rise to this thermal dependence. Interestingly, the storage moduli (G’) above 30°C in the third cooling cycle (from 50°C) were higher than in the previous cycles (from 40°C), even though all cycles reached the same G’ value (Fig. S13). To investigate this further, we heated pre-formed A’ gels to 60°C and imaged them to examine the bundle dynamics during cooling. While some of the thin-bundles showed dynamic fluctuations that were arrested after cooling (Fig. S14), the bundles did not dissolve, even at temperatures higher than the melting temperature (T_M) of the DNA-crosslinker, suggesting that crosslinks of A’ are further stabilized within thin-bundles.

We compared this interesting rheological behavior of A’ to that of A-A’ and A-B-C crosslinked networks (Fig. 5b, S12). We found that G_ω at 25°C after two heat-cool cycles agreed with the respective initial value in all samples, suggesting that the thermo-reversible properties are characteristic of these peptide-DNA materials. Interestingly, G’ of all materials dropped more than 10-fold within only a 10°C increase from 25°C, well below T_M of the DNA, suggesting that weak crosslinks between thin-bundles consist of partially forming DNA duplexes. We compared the rate of dissociation (from exponential-decay fitting of the reheating curves) (Fig. S15) and the number of weak crosslinks (the difference between G’ at 25°C and 50°C (ΔG)) in each system (Fig. S16). We revealed that A-B-C exhibited the fastest dissociation rate and lowest ΔG. This suggests that A-B-C has more crosslinks stable within the large-bundles than in the network. In contrast, the A-A’ crosslinker showed the largest value of ΔG, suggesting that A-A’ crosslinks filaments into thin-bundles as well as connecting bundles into a network. These heat-cool treatments clearly show that peptide-DNA gels exhibit thermo-reversible properties that allow transitions between different structural and mechanical states.

Reconfiguration of hierarchical structures controls release of protein payloads

Multistranded crosslinkers, such as A-B-C, can be programmed to reversibly transition between structures using toehold-mediated strand displacement as the enabling mechanism.³² We hypothesized that sequentially interfering with each base-paired region within the A-B-C junction would allow us to unravel distinct structures within the gel (Fig. 6a). Using a DNA input to displace A-C binding, we achieved partial dissolution of the bundles as shown by TEM (Fig. S17a,b). Addition of an orthogonal DNA input, designed to displace B-C binding, resulted in complete structural reconfiguration into individual filaments (Fig. S17c). These results demonstrate that through design of orthogonal peptide-DNA arms and respective DNA inputs, we can trigger structures to partially or completely ‘erase’.

We harnessed this programmable structural reconfiguration to controllably release a protein payload (Fig. 6a). We embedded the enzyme amylase as a model cargo in the A-B-C bundled network, using low-temperature heating and cooling (60°C) to prevent protein denaturation. Confocal microscopy images of TAMRA-tagged amylase embedded in Thioflavin T (ThT)-stained peptide-DNA filaments showed colocalization of both colors, suggesting a successful association of amylase within bundles and the surrounding network (Fig. 6b). This was further confirmed from the 2D histograms of the intensities in each color channel. We found two populations in this plot. One group showed large variations of green (peptide) signals with minimum variations of red (amylase) signals as indicated in the 2D histogram, corresponding to ThT-stained fibers without amylase. In the other group, the red signals were highly correlated with green signals, corresponding to proteins incorporated into bundles. After incubation with a DNA invader (A invader, “A-I”) for 1 hour, the bundles began to dissociate and the second group on the 2D histogram, corresponding to proteins within bundles, disappeared (Fig. 6c). This indicated an effective release of a protein dose associated with a partial reconfiguration.

To measure the triggered amylase release, we prepared tubes layered with peptide-DNA/amylase gels on the bottom and amylose/iodine solution on the top (Fig. 6d). The substrate for amylase, amylose, complexes with iodine to generate a blue color.³³ The amylase release can be detected as color fading since amylase cleaves amylose (Fig. 6d, S18-19). We imaged the tubes during the reaction and analyzed the color intensity of the top layer to obtain a relative fraction of amylase released (Fig. S19). Tubes without invader additions showed similar color saturation before and after the reaction (20 min) and were comparable to the standard of only amylose/iodine, indicating that amylase was sufficiently entrapped within the peptide-DNA gel layer. (Fig. 6e, S19). Next, we added the invader strand to release the amylase. Addition of A-I (1 eq.) released a fraction of amylase, achieving slow, controlled release of amylase. Then, we sequentially added two orthogonal invaders (A-I, B-I) (Fig. 6e-f). Each addition of invader released 10–20% of the amylase, demonstrating that a catalytic protein payload could be released from a hydrogel in small doses using a user-defined trigger. This system offers peptide-DNA nanotechnology as a platform for engineering reconfigurable biomaterials able to respond mechanically and biochemically to environmental cues in the biological milieu.

Integrating self-assembled peptides with DNA nanotechnology facilitated a systematic study into the impact of crosslinker strength and geometry on hierarchical self-assembly. We discovered that peptide filaments can organize into tactoid-shaped bundles resembling mitotic spindles. To our knowledge, these constructs are the first example of purely synthetic filaments adopting a spindle-like structure. Varying crosslinker interaction strength and geometry tuned the tactoid ARs and induced 3 distinct regimes of structures and mechanical behavior. Introducing as few as 8 base-pairs increased the bundled network elasticity by an order of magnitude compared to the DNA-free network. Increasing to 14 base-pairs resulted in a more elastic, highly bundled network, and evolving from bivalent to trivalent DNA junctions triggered the tactoids to further organize into bundles of tactoids. Furthermore, the multistranded DNA-crosslinker design enabled structural and mechanical reconfiguration by temperature and orthogonal triggers. Future work can expand the peptide-DNA library to include even more complex DNA junctions such as X-DNA, tetrahedrons, or DNA origami constructs and can be programmed to respond to additional triggers through aptamer-based logic gates, transcription circuits, light or pH.^32,34 Such strategies will further leverage advantages of both peptides and DNA to enable design of responsive and highly programmable materials for drug delivery vehicles, dynamic cell culture scaffolds, intracellular sensors, or synthetic cell machinery.

Materials

All chemicals were used without further purification. Rink amide MBHA resin and Fmoc-Lys(azide)-OH were purchased from Chem-Impex. Fmoc-(PEG)₂-OH was purchased from PurePEG. Fmoc-Phe-OH, DIC, TFA, TIPS and TEAA were purchased from Sigma-Aldrich. Oxyma Pure and piperidine were purchased from VWR. Dibenzocyclooctyne-sulfo-N-hydroxylsuccinimidyl ester (DIBAC-sulfo-NHS) was purchased from Sigma-Aldrich, then dissolved in dry DMSO to a final concentration of 100 mM and stored at -20 °C. All DNA strands were purchased from IDT, then dissolved in water to a final concentration of 1 mM or 100 μM and stored at -20 °C.

Peptide Synthesis

The Fmoc-FF-(PEG)₂-K_az-NH₂ was synthesized using automated fluoren-9-ylmethoxycarbonyl (Fmoc) solid-phase peptide synthesis (Liberty Blue, CEM) on rink amide MBHA resin (100-200 mesh, 0.77 mmol/g). The peptide was cleaved from the resin in a solution of 95% trifluoroacetic acid (TFA), 2.5% triisopropylsilane (TIPS) and 2.5% dH₂O. The cleaved peptide solution was concentrated and re-suspended in a 3:1 ratio of acetonitrile to 0.1% TFA in H₂O. The solution was purified on reverse-phase HPLC (Shimadzu UFLC, Ultra C18 5 µm, 100 X 10 mm column) with a gradient of 0.1% TFA in H₂O (Solvent A) and acetonitrile (Solvent B). The purity of the peptide was confirmed by ESI-Mass Spectrometry (Q-exactive HF-X).

Peptide-DNA Synthesis

Amine-modified oligonucleotides (1 mM or 100 μM in H₂O) were diluted into 20 mM sodium phosphate buffer pH 8.5. DIBAC-sulfo-NHS (100 mM in DMSO) was added to the solution in excess, and the mixture was reacted for 2 hours with vigorous shaking at room temperature. The solution was reactivated with DIBAC-sulfo-NHS and gently shaken overnight at 4 °C. The solution was exchanged with 100 mM sodium phosphate buffer pH 7.5. Fmoc-FF-(PEG)₂-K_az-NH₂ was dissolved in DMSO to a final concentration of 2 mM and added in a 2:1 molar ratio to the DNA-DIBAC solution. Water and DMSO were added to bring the reaction mixture to a total of 20 mM phosphate buffer pH 7.5 with 20% DMSO. The reaction mixture was shaken overnight at room temperature. The solution was exchanged with 50 mM TEAA buffer pH 7.0, concentrated and purified on reverse-phase HPLC (Inspire C18 5µm, 250 x 4.6 mm column) with a 5-80% gradient of 50 mM TEAA buffer pH 7.0 in H₂O (Solvent A) and 90% acetonitrile in water with 50 mM TEAA buffer pH 7.0 (Solvent B). The identity and purity of the peptide-DNAs were confirmed by ESI-MS (ThermoScientific Q Exactive HF-X) and MALDI (AB Sciex 5800 MALDI-TOF/TOF). The peptide-DNAs were quantified using absorbance measurements at 260 nm, lyophilized in aliquots and stored at -20 °C.

Peptide and Peptide-DNA Assembly

For assemblies in DMSO/water, Fmoc-FF-OH (Bachem) was dissolved in dimethyl sulfoxide (DMSO) to prepare a stock solution with a concentration of 10 wt%. The peptide was diluted to 0.1 wt% in 10 mM potassium phosphate buffer pH 7.5 with 150 mM NaCl, then used to dissolve the required amount of lyophilized peptide-DNA.The solutions were heated up to 95 °C and mixed every ~10 min by snapping test tubes for 30 min, then were annealed from 80 °C to 25 °C at a cooling rate of -1 °C per minute. For assemblies in water, Fmoc-FF-OH was dissolved in 10 mM potassium phosphate buffer pH 7.5 with 150 mM NaCl and horn sonicated for 3 x 5 s on ice. For peptide-DNA coassemblies in water, Fmoc-FF was dissolved in 10 mM phosphate buffer pH 7.5 with 150 mM NaCl, then the peptide solution was used to solubilize the lyophilized peptide-DNA. The solutions were horn sonicated for 3 x 5 s on ice to mix the monomers, then annealed from 95°C to 25°C at a cooling rate of -1°C per minute. Final concentrations of Fmoc-FF and pepDNA were 0.1 wt% with 1 mol% of each pepDNA. All samples were stored at 4°C and used up to 2 weeks after assembly.

Confocal Microscopy on Peptide and Peptide-DNA Assembly by DMSO switch method

A NileRed stock solution (10 mM in DMSO) was diluted with 10 mM potassium phosphate buffer pH 7.5 with 150 mM NaCl. For Fmoc-FF gels, Fmoc-FF (dissolved at 10 wt% in anhydrous DMSO) was diluted with the phosphate buffer containing NileRed. For coassemblies, the peptide and NileRed solution was used to dissolve the required amount of peptide-DNA. Solutions of freshly dissolved peptide and peptide-DNA with NileRed in DMSO/water conditions were heated up to 95 °C and were mixed every ~10 min by snapping test tubes for 30 min. The samples were then sandwiched between two coverslips using strips of double-stick tape and sealed with mineral oil to prevent evaporation. The sample chambers were annealed from 80 °C (or 60 °C) to 25 °C at a cooling rate of -1 °C per minute.

Samples were imaged with Andor XD spinning disk confocal microscope with a CSU-X1 Yokogawa head. The system is equipped with a CMOS camera (Hamamatsu Flash4v2 sCMOS, Hamamatsu, Japan). The samples were excited with 561 nm laser and imaged with 20x and 60x objectives (20X/0.75 UPlan S-APO, Olympus, Japan; 60X/1.20 Water UPlan S-APO, Olympus, Japan). z-stacks were obtained with 1 μm steps for 20x and 0.4 μm steps for 60x. For heating experiments on annealed samples, the sample chambers were heated up to 60 °C on the heating plate for 10 min and placed on the microscope at room temperature to monitor the time evolution of the samples.

We found that the samples were inhomogeneously illuminated. To correct the effect of inhomogeneous illumination, a homogenous dye solution was imaged and the image was normalized by the average pixel intensity of the image after background subtraction to create the map of normalization factors on each pixel. The effect of inhomogeneous illumination was corrected by subtracting the background of sample images and dividing them by the map of normalization factors.

Transmission Electron Microscopy

For TEM imaging, 10 μL of sample was spotted on a carbon coated 300-mesh copper grid (Electron Microscopy Sciences). After 5 minutes, the sample was wicked from the grid. The grid was then washed twice with water and stained twice with 2% uranyl acetate for 20 seconds. The samples were imaged on a FEI Tecnai T12 at 120 kV. Images were analyzed and processed with minor adjustments for brightness and contrast using ImageJ software (National Institutes of Health, USA). Alignment fraction analysis was performed using the OrientationJ Analysis plugin on ImageJ.

Acridine Orange Confocal Microscopy

To examine the local effect of DNA binding on bundle formation, we stained the pepDNA materials with acridine orange (AO). As AO emits red fluorescence when bound to single-stranded DNA (ssDNA) (λ_ex = 457 nm, λ_em = 638 nm) and green fluorescence when bound to double-stranded DNA (dsDNA) ((λ_ex = 502 nm, λ_em = 522 nm), it allowed us to estimate the degree of DNA hybridization localized in the bundles by measuring fluorescence intensities in these two spectral regions. To prepare the samples for confocal microscopy, 1.2 μL of 50 μM AO solution was added to 3 μL peptide-DNA solutions. The samples were incubated at room temperature for at least 3 hours. Then, the samples were sandwiched between two coverslips using strips of double-stick tape. The chambers were sealed using nail polish. All experiments were performed at room temperature. Fluorescence images of peptide-DNA bundles were recorded under a Zeiss 880 confocal microscope, which is equipped with a 63x objective (Plan- Apochromat, 63x, NA = 1.4; Zeiss). Fluorescence excitation was done with a 488 nm laser. The excitation light intensity was kept constant through the entire measurement. To collect green and red fluorescence separately, emission ranges of the detector were set to 490-553 nm and 623-680 nm, respectively. Although AO is known as a nucleic acid binding dye, we found that AO also binds to Fmoc-FF and emits both green and red fluorescence (Fig. S9). To see the effect of DNA binding, we subtracted the effect of Fmoc-FF from images and obtained information about DNA base-pairing. Assuming that fluorescence intensities on each channel are a linear combination of fluorescence sources, we defined I_green= I_dsDNA+I_green,FF and I_red= I_ssDNA+I_red,FF, containing the fluorescence intensities in the green (I_green)and red (I_red) channels from each pixel, of dsDNA (I_dsDNA, green channel) and ssDNA (I_ssDNA, red channel), and of the background generated by non-specific peptide-AO interactions (I_green,FF and I_red,FF). To subtract the contributions of the peptide, we defined the ratio of peptide fluorescence intensity between the green and red channels β=I_green,FF/I_red,FF from imaging peptide fibers (Fig. S9) and derived I_green-βI_red=I_dsDNA-βI_ssDNA that relates I_green-βI_red to the difference in local concentrations of dsDNA and ssDNA.

Rheology measurements on Peptide and Peptide-DNA Assemblies

Macrorheology was performed by a commercial rheometer (AR-G2, TA instruments, USA) in an oscillatory mode using a 40-mm aluminum parallel plate. Peptide/peptide-DNA solutions were prepared by a DMSO switch method as described previously. Then, 200 μL of the peptide/peptide-DNA solutions was heated up to 95 °C and was mixed every ~ 10 min by snapping test tubes for 30 min. The samples were placed on the plate at 80 °C (or 60 °C). We applied a thin layer of low-viscosity mineral oil around the sample to minimize evaporation. Annealing process from 80 °C (or 60 °C) to 25 °C at a cooling rate of -1 °C per minute was monitored at the frequency of 1 Hz and 0.5 % strain. Frequency sweep measurements were performed on annealed samples from 0.1 Hz to 50 Hz with 0.5 % strain, which is in the linear viscoelastic range of peptide-DNA as shown in strain sweep measurements (Fig. S10). To examine the temperature dependence of rheological properties, temperature sweep experiments were performed at the frequency of 1 Hz and 0.5 % strain. For A’, the temperature was changed from 25 °C to 40 °C to 25 °C. This procedure was repeated twice. Then, the temperature was increased from 25 °C to 50 °C to 25 °C. For A-A’, the temperature was changed from 25 °C to 50 °C to 25 °C. For A-B-C, the temperature was changed from 25 °C to 40 °C to 25 °C and then from 25 °C to 50 °C to 25 °C.

DNA Strand Displacement

Invader DNA strands at a stock concentration of 1 mM in sterile water were added in 10-fold excess to the peptide-DNA coassemblies. The solution was left to incubate at room temperature for 1 hour, then spotted on a TEM grid for imaging.

Amylase Dye-modification

Amylase was reacted with 8-fold excess TAMRA-NHS in 20 mM potassium phosphate pH 8.5 overnight at 4 °C (ref). The sample was concentrated with a 30 kDA centrifuge filter and re-suspended in 10 mM phosphate pH 7.5 three times. The amylase-embedded peptide-DNA was prepared in the same fashion as the other samples: 15 μL of annealed A-B-C were mixed with 10 μL of amylase-TAMRA then annealed from 60 – 25 °C. Sample was stained with 5 mol% Thioflavin T (1 mM in H₂O) for 30 minutes before imaging.

Amylase Release

A fresh solution of amylase was prepared at 343 μM in 20 mM sodium phosphate buffer pH 6.9. Then, several individual tubes of embedded enzyme in peptide-DNA were prepared by adding 10 μL of the amylase solution to 15 μL of 0.1 wt% A-B-C and annealing from 60 °C to 25 °C at a cooling rate of 1 °C per minute. Amylose solution was prepared by making a slurry of the solid in water, adding 2 M NaOH until the solid was dissolved, then neutralizing the solution to pH 6.9 with 1 M HCl. Iodine solution was prepared at 2 mg/mL KI and 0.2 mg/mL iodine in sterile water. The amylose/iodine solution was prepared by mixing 1 part volume 4.4 mM amylose with 2 parts volume of the iodine solution (solution is blue). To the enzyme/peptide-DNA gel tubes, 25 μL of amylose/iodine was added gently, so that two layers (the clear gel layer and the blue amylose layer) were visible. The samples were arranged upright against a white paper background using double-sided tape, and imaged with a CMOS camera. To perform DNA strand displacement, 1 equivalent of DNA invader was added gently to the tube. Analysis was performed using ImageJ software.

Supporting Information.

Synthesis and purification of molecules; molecular design and DNA sequence details; supporting schemes; TEM of spindle-like structures in DMSO/water; filament length comparison across solvent conditions; Width, Width, percent area covered and alignment comparison across the pepDNA library; confocal of peptide; NuPack simulations of DNA melting temperatures; strain sweeps; Loss moduli from heat-cool cycles; pepDNA spindle fluctuations at elevated temperatures; Fitting analysis of G’ heating curves; Difference in storage moduli before and after heating; scheme of amylose cleavage by amylase; amylase reaction standards; supporting discussion (PDF)

AUTHOR INFORMATION

Corresponding Author

* [email protected]

Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. K.N. and M.L.D contributed equally to the manuscript.

ACKNOWLEDGMENTS

We thank the Department of Applied Physical Sciences at UNC-Chapel Hill for support of this work. Mass spectrometry was performed at the UNC Mass Spectrometry Core Laboratory at CRITCL (Chemical Research Instrumentation Teaching and Core Laboratories) in the Department of Chemistry. TEM was performed at the Chapel Hill Analytical and Nanofabrication Laboratory (CHANL), a member of the North Carolina Research Triangle Nanotechnology Network (RTNN), which is supported by the National Science Foundation (NSF), Grant ECCS-2025064, as part of the National Nanotechnology Coordinated Infrastructure (NNCI). TEM and confocal microscopy were performed at the UNC Hooker Imaging Core Facility and the Microscopy Services Laboratory, supported in part by the P30CA016086 Cancer Center Core Support Grant to the UNC Lineberger Comprehensive Cancer Center. M.L.D. acknowledges support from the NSF Graduate Research Fellowship Program. We thank the R. Superfine lab for the use of the rheometer. We thank Peter Harris and Matt Shelton for the graphic illustrations incorporated into the manuscript.

DNA, deoxyribonucleic acid; ssDNA, single-stranded deoxyribonucleic acid; dsDNA, double-stranded deoxyribonucleic acid; DMSO, dimethylsulfoxide; pepDNA, peptide-DNA; AO, Acridine Orange.

Fratzl, P., and Weinkamer, R. (2007) Nature’s hierarchical materials. Prog. Mater. Sci. 52, 1263–1334.
Winder, S. J., and Ayscough, K. R. (2005) Actin-binding proteins. J. Cell Sci. 118, 651–654.
Fletcher, D. A., and Mullins, R. D. (2010) Cell mechanics and the cytoskeleton. Nature 463, 485–492.
Prince, E., and Kumacheva, E. (2019) Design and applications of man-made biomimetic fibrillar hydrogels. Nat. Rev. Mater. 4, 99–115.
Ekiz, M. S., Cinar, G., Khalily, M. A., and Guler, M. O. (2016) Self-assembled peptide nanostructures for functional materials. Nanotechnology 27.
Acar, H., Srivastava, S., Chung, E. J., Schnorenberg, M. R., Barrett, J. C., LaBelle, J. L., and Tirrell, M. (2017) Self-assembling peptide-based building blocks in medical applications. Adv. Drug Deliv. Rev. 110–111, 65–79.
Gazit, E. (2007) Self-assembled peptide nanostructures: The design of molecular building blocks and their technological utilization. Chem. Soc. Rev. 36, 1263–1269.
Edelbrock, A. N., Clemons, T. D., Chin, S. M., Roan, J. J. W., Bruckner, E. P., Álvarez, Z., Edelbrock, J. F., Wek, K. S., and Stupp, S. I. (2021) Superstructured Biomaterials Formed by Exchange Dynamics and Host–Guest Interactions in Supramolecular Polymers. Adv. Sci. 2004042, 1–11.
Wester, J. R., Lewis, J. A., Freeman, R., Sai, H., Palmer, L. C., Henrich, S. E., and Stupp, S. I. (2020) Supramolecular Exchange Among Assemblies of Opposite Charge Leads to Hierarchical Structures. J. Am. Chem. Soc. 142, 12216–12225.
Freeman, R., Han, M., Álvarez, Z., Lewis, J. A., Wester, J. R., Stephanopoulos, N., McClendon, M. T., Lynsky, C., Godbe, J. M., Sangji, H., Luijten, E., and Stupp, S. I. (2018) Reversible self-assembly of superstructured networks. Science (80-.). 362, 808–813.
Daly, M. L. (2020) Chap. 14: New Functions Emerging from Peptide-DNA Materials, in Peptide-based Biomaterials, pp 459–486.
Macculloch, T., Buchberger, A., and Stephanopoulos, N. (2019) Emerging applications of peptide-oligonucleotide conjugates: Bioactive scaffolds, self-assembling systems, and hybrid nanomaterials. Org. Biomol. Chem. 17, 1668–1682.
Daly, M. L., Gao, Y., and Freeman, R. (2019) Encoding Reversible Hierarchical Structures with Supramolecular Peptide–DNA Materials. Bioconjug. Chem. 30, 1864–1869.
Kaushik, M., Kaushik, S., Roy, K., Singh, A., Mahendru, S., Kumar, M., Chaudhary, S., Ahmed, S., and Kukreti, S. (2016) A bouquet of DNA structures: Emerging diversity. Biochem. Biophys. Reports 5, 388–395.
Wang, P., Meyer, T. A., Pan, V., Dutta, P. K., and Ke, Y. (2017) The Beauty and Utility of DNA Origami. Chem 2, 359–382.
Aldaye, F. A., Palmer, A. L., and Sleiman, H. F. (2008) Assembling materials with DNA as the guide. Science (80-.). 321, 1795–1799.
Stephanopoulos, N., Freeman, R., North, H. A., Sur, S., Jeong, S. J., Tantakitti, F., Kessler, J. A., and Stupp, S. I. (2015) Bioactive DNA-peptide nanotubes enhance the differentiation of neural stem cells into neurons. Nano Lett. 15, 603–609.
Reches, M., and Gazit, E. (2005) Self-assembly of peptide nanotubes and amyloid-like structures by charged-termini-capped diphenylalanine peptide analogues. Isr. J. Chem. 45, 363–371.
Smith, A. M., Williams, R. J., Tang, C., Coppo, P., Collins, R. F., Turner, M. L., Saiani, A., and Ulijn, R. V. (2008) Fmoc-diphenylalanine self assembles to a hydrogel via a novel architecture based on π-π interlocked β-sheets. Adv. Mater. 20, 37–41.
Prinsen, P., and van der Schoot, P. (2003) Shape and director-field transformation of tactoids. Phys. Rev. E - Stat. Physics, Plasmas, Fluids, Relat. Interdiscip. Top. 68, 11.
Weirich, K. L., Banerjee, S., Dasbiswas, K., Witten, T. A., Vaikuntanathan, S., and Gardel, M. L. (2017) Liquid behavior of cross-linked actin bundles. Proc. Natl. Acad. Sci. U. S. A. 114, 2131–2136.
Edozie, B., Sahu, S., Pitta, M., Englert, A., Do Rosario, C. F., and Ross, J. L. (2019) Self-organization of spindle-like microtubule structures. Soft Matter 15, 4797–4807.
Wang, J., Liu, K., Xing, R., and Yan, X. (2016) Peptide self-assembly: thermodynamics and kinetics. Chem. Soc. Rev. 45, 5589.
Mason, T. O., Chirgadze, D. Y., Levin, A., Adler-Abramovich, L., Gazit, E., Knowles, T. P. J., and Buell, A. K. (2014) Expanding the solvent chemical space for self-assembly of dipeptide nanostructures. ACS Nano 8, 1243–1253.
Wang, J., Liu, K., Yan, L., Wang, A., Bai, S., and Yan, X. (2016) Trace Solvent as a Predominant Factor to Tune Dipeptide Self-Assembly. ACS Nano 10, 2138–2143.
Raeburn, J., Mendoza-Cuenca, C., Cattoz, B. N., Little, M. A., Terry, A. E., Zamith Cardoso, A., Griffiths, P. C., and Adams, D. J. (2015) The effect of solvent choice on the gelation and final hydrogel properties of Fmoc-diphenylalanine. Soft Matter 11, 927–935.
Wen, Q., and Janmey, P. A. (2011) Polymer physics of the cytoskeleton. Curr. Opin. Solid State Mater. Sci. 15, 177–182.
Gardel, M. L., Shin, J. H., MacKintosh, F. C., Mahadevan, L., Matsudaira, P., and Weitz, D. A. (2004) Elastic behavior of cross-linked and bundled actin networks. Science (80-.). 304, 1301–1305.
Lieleg, O., Claessens, M. M. A. E., and Bausch, A. R. (2010) Structure and dynamics of cross-linked actin networks. Soft Matter 6, 218–225.
Bartles, J. R. (2000) Parallel actin bundles and their multiple actin-bundling proteins. Curr. Opin. Cell Biol. 12, 72–78.
Stossel, T. P., Condeelis, J., Cooley, L., Hartwig, J. H., Noegel, A., Schleicher, M., and Shapiro, S. S. (2001) Filamins as integrators of cell mechanics and signalling: Article : Nature Reviews Molecular Cell Biology. Nat. Rev. Mol. Cell Biol. 2, 138–145.
Zhang, D. Y., and Seelig, G. (2011) Dynamic DNA nanotechnology using strand-displacement reactions. Nat. Chem. 3, 103–113.
Zhu, Z., Wu, C., Liu, H., Zou, Y., Zhang, X., Kang, H., Yang, C. J., and Tan, W. (2010) An Aptamer Cross-Linked Hydrogel as a Colorimetric Platform for Visual Detection. Angew. Chemie 122, 1070–1074.
Seeman, N. C., and Sleiman, H. F. (2017) DNA nanotechnology. Nat. Rev. Mater. 3.

There is NO Competing Interest.

SIFinalFreeman.pdf

Download PDF

Journal Publication

published 23 Apr, 2024

Read the published version in Nature Chemistry →

Version 1

posted

You are reading this latest preprint version

Bioinspired Spindles using Peptide-DNA Nanotechnology

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results And Discussion

Peptide-DNA crosslinks tune hierarchical formation of spindle-like assemblies

Filament length and crosslinker flexibility and density tune hierarchical organization

Programmable structural hierarchy with tunable mechanical states

Transient DNA crosslinks impart heat reversible mechanics

Reconfiguration of hierarchical structures controls release of protein payloads

Conclusion

Materials And Methods

Declarations

ACKNOWLEDGMENTS

Abbreviations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1