Maturation and Conformational Switching of a DeNovo Designed Phase-Separating Polypeptide

Cellular compartments formed by biomolecular condensation are widespread features of cell biology. These organelle-like assemblies compartmentalize macromolecules dynamically within the crowded intracellular environment. However, the intermolecular interactions that produce condensed droplets may also create arrested states and potentially pathological assemblies such as fibers, aggregates, and gels through droplet maturation. Protein liquid–liquid phase separation is a metastable process, so maturation may be an intrinsic property of phase-separating proteins, where nucleation of different phases or states arises in supersaturated condensates. Here, we describe the formation of both phase-separated droplets and proteinaceous fibers driven by a de novo designed polypeptide. We characterize the formation of supramolecular fibers in vitro and in bacterial cells. We show that client proteins can be targeted to the fibers in cells using a droplet-forming construct. Finally, we explore the interplay between phase separation and fiber formation of the de novo polypeptide, showing that the droplets mature with a post-translational switch to largely β conformations, analogous to models of pathological phase separation.


Computational tools
Protein parameters, including molecular weight and extinction coefficient (ϵ) at 280 nm were calculated from their primary amino acid sequences using ExPASy ProtParam (https://web.expasy.org/protparam). 1 All data analyses except those mentioned explicitly were performed in Python using the Pandas and Numpy libraries, and visualized using the MatPlotLib and Seaborn libraries.

Protein expression and purification
Expression of HERD constructs was performed individually on a pET28a derived plasmid under the control of the T7 promotor and with AMP selection, and dual expression was performed by cotransformation of a second vector with a slightly lower copy number, but also under the control of the T7 promotor and with CMP selection.In brief, 50 μl of chemically competent BL21*(DE3) Escherichia coli (E.coli) were transformed with the plasmid of interest by incubating on ice with DNA (typically 25 ng for a single vector, 100 ng each for two vectors) for 30 minutes, before heat shocking for 30 s at 42 °C, and chilling on ice for 1 minute.200 μl of LB was then added and the cells incubated at 37 °C for 1 hour, before 50 μl of the cell suspension was spread on LB-agar plates containing appropriate antibiotics (AMP -100 μg/ml, CMP -25 μg/ml) and incubated at 37 °C overnight.For expression, a single colony was inoculated into a 5 ml LB supplemented with AMP or CMP, and grown overnight (37 °C, 200 rpm).Expression cultures (50 ml -1 l) supplemented with AMP or CMP were inoculated 1:100 with overnight culture and grown to OD600 = 0.4 -0.6 (37 °C, 200 rpm).Protein expression was then induced to a final concentration of 400 μM IPTG and the culture grown for up to 21 hours (18 °C, 200 rpm).To measure culture growth and protein expression cells were grown after induction in black 96-well plates with optimal bottoms (Thermo Scientific).The plates were sealed with a BEM-1 Breathe Easy (Diversified Biotech) gas permeable membrane to avoid evaporation and incubated at 25 °C with constant shaking at 200 rpm for 24 hours.Both OD600 and GFP fluorescence were recorded every 30 minutes over 24 hours using a Clariostar plate reader (BMG Labtech).To quantify the ratio between the amounts of HERD-2.2 and HERD-2.2-GFPproduced, the cells were grown for expression as described above and aliquots of cultures were collected 5 hours after the induction of protein expression.SDS-PAGE was performed on total cell lysates and serial dilutions of corresponding proteins in the same gels.The gels were imaged and analyzed by densitometry using Fiji.The respective peaks in the profiles of the lanes were integrated.The serial dilutions were used to plot a linear calibration curve that allowed to determine the amount of protein in the lysate.For protein purification, cells were collected by centrifugation (3400 xg, 20 minutes) and resuspended in 40 ml resuspension buffer (20 mM Tris pH 7.5, 50 mM imidazole, 500 mM NaCl, 2 M urea, 1 tablet cOmplete protease inhibitor cocktail).Cells were lysed on ice by sonication (5 s on, 2 s off, 75% amplitude, 15 minutes).The suspension was centrifuged (16000 xg, 20 minutes) and filtered to clarify.The lysate was applied to a 5 ml HisTrap HP IMAC column pre-equilibrated in binding buffer (20 mM Tris pH 7.5, 500 mM NaCl, 50 mM imidazole, 2 M urea).The column was washed in binding buffer for 4 column volumes, and the bound protein eluted using a gradient of elution buffer (0 -100%) (20 mM Tris pH 7.5, 500 mM NaCl, 2 M urea, 500 mM imidazole).The eluted protein was further purified by size exclusion chromatography using a HiLoad Superdex 200 pg column at 1 ml/min flow rate (20 mM Tris pH 7.5, 2 M urea).The purified protein was then buffer exchanged into 20 mM Tris pH 7.5 using a 26/10 desalting column run at 5 ml/min, before concentration using 3 kDa molecular weight cut-off (MWCO) spin concentrators, and snap frozen in liquid nitrogen for storage at -70 °C.Protein concentration was measured by absorbance at 280 nm.

Confocal microscopy
Microscopy was performed using a Leica SP8 confocal microscope using a 65 mW Ar laser (488 nm) with a 63x 1.4 numerical aperture oil immersion objective lens.For in vitro microscopy, 10 μl of sample was applied to a clean glass slide and covered with a coverslip before imaging.For in-cell imaging, E. coli treated as described above for expression of recombinant proteins.Samples were collected 5 hours after induction of protein expression for confocal microscopy. 1 ml of culture was collected by centrifugation (3000 xg, 3 minutes) and washed 3 times with 1 ml PBS, before fixing with 1 ml 2% PFA for 15 minutes.For nucleic acid staining, 2 μL of saturated DAPI solution in DMF (Invitrogen) was added to the fixation solution.Fixed cells were washed 3 more times in PBS, before resuspending in 50 μl PBS. 10 μl was then applied to a clean glass slide, with 1 drop ProLong Diamond Antifade mountant, and covered with a glass coverslip.DAPI and GFP were imaged in blue (λex 405 nm; λem 410-460 nm) and green (λex 488 nm; λem 492-600 nm) channels, respectively.Cell images were analyzed and assembled in FiJi, and are displayed as maximum projections. 2,3ansmission electron microscopy (TEM) Negative stain TEM was performed using a 120 kV Tecnai 12 electron microscope.For purified protein samples, 10 μl was applied to a glow-discharged 300-mesh carbon grid coated with pioloform and incubated for 1 minute, then stained using a 2% uranyl acetate (UA) solution.Grids were incubated for 1 second in a 10 μl drop of UA, then incubated sample-side down into a second drop and incubated for 3 minutes.The grid was then blotted with filter paper to remove excess stain, and swept through a third UA drop, and blotted again.The grid was then swept through two drops of ultrapure water, before blotting and leaving to dry before imaging.For TEM of E. coli, cultures treated as described above for expression of recombinant proteins were collected 21 hours after induction of protein expression. 1 ml of cell suspension was pelleted by centrifugation, and 1 μl of pellet was vitrified using a Leica EM PACT2 high pressure freezer with a rapid transfer system.The vitrified cells were freeze substituted with 0.2% UA, 5% H2O in acetone for 5 hours at -90 °C using a Leica AFS2 automated freeze substitution system.The samples were then warmed to -45 °C, and kept for 2 hours at that temperature before washing in acetone for 30 minutes.Resin (Lowicryl HM20) was then infiltrated into the freeze substituted samples at increasing dilutions (25%, 50%, 75%) for 3 hours each, before embedding in 100% resin for 16 hours, followed by 3 changes of resin, left for 2 hours each.The infiltrated resin was then polymerized using UV for 48 hours.The resin blocks were sectioned using an EM UC6 microtome with a diamond knife at 45 °, and the sections imaged by TEM.

Correlative light electron microscopy (CLEM)
Samples were prepared for CLEM as described for TEM, but sections were first imaged using a Leica SP8 AOBS confocal microscope with a 63x oil immersion objective, before negative staining with 2% UA as described for purified samples, and imaging by TEM.

Circular dichroism (CD) spectroscopy
Circular dichroism (CD) spectra were recorded on a JASCO J-810 spectropolarimeter with a Peltier temperature controller, in a 1 mm path length reduced volume cuvette.Full spectra measured ellipticity between 190 nm and 260 nm in 1 nm intervals, with a 100 nm/min scanning rate, 1 nm bandwidth and 1 s response time.A reference spectrum using the same cuvette, parameters and buffer at 5 °C was subtracted from the measured ellipticity.Measurements of ellipticity with respect to temperature (melting and cooling spectra), were recorded by collecting initial spectra at 5 °C, with ellipticity measured at 222 nm every 1 °C and full spectra measured every 5 °C as the temperature was increased to 90 °C, and then decreased again to 5 °C.Ellipticity (deg) values were converted to mean residue ellipticity (MRE) (deg•cm 2 •dmol -1 •res -1 ) by normalization to the number of peptide bonds in the protein, and the path length using the following equation: ×   ×  ×  Where θ is the difference in absorbed circularly polarized light in millidegrees, c is the protein concentration in mM, l is the path length in cm, and b is the number of amide bonds in the protein.
Fraction helicity was calculated using the MRE at 222 nm (MRE222) using the following equation: Where MREcoil is 640 -45T, T is the temperature in degrees Celsius, and n is the number of amide bonds in the sample. 4Melting temperature (Tm) was calculated as the mean of 3 independent experiments.Deconvolution of CD data was performed using https://bestsel.elte.hu/index.php. 5

X-ray fiber diffraction
Solutions containing purified fibers (16 mg/ml protein, 20 mM Tris pH 7.5) were hung between two glass capillaries with the ends sealed with paraffin, placed approximately 1 cm apart, and left to dry for 18 hours.Fibers were aligned in the detector at 0 and 90° orientations, before diffraction using a Rigaku copper rotating anode X-ray source with a Saturn CCD detector at distances of 50 and 100 mm for 30 -60 s.Diffraction patterns were analyzed using CLEARER to identify distances. 6ioflavin T fluorescence Thioflavin T (ThT) fluorescence (excitation: 449 nm, emission: 493 nm) was measured using a CLARIOstar plate reader in black 96-well plates.Solutions containing 25 μM ThT, 50 mM Tris pH 7.5, and varied protein concentration from 0 -100 μM were made up to 100 μl per well, with 4 replicates per protein concentration.Plates were sealed to prevent evaporation and incubated at 25 °C for 5 hours to equilibrate, before measuring fluorescence emission every 85 seconds for 12 minutes.The mean fluorescence over this period was recorded as the final fluorescence intensity for each well, and plotted as mean and standard deviation over the 4 replicates.

Protease cleavage and droplet maturation
Droplet maturation was triggered by addition of thrombin protease to phase-separated droplets.Phase-separated droplets were formed by mixtures of 10% PEG 3350, 125 mM NaCl, 50 mM Tris pH 7.5, and 0.675 mM HERD-2.2-T-GFP, before 0.06 units of thrombin protease added, and incubated at 25 °C for 30 minutes.