Introduction

Polymer architecture and conformation – the spatial arrangement of the atomic groups in an individual chain – constitute the fundamental molecular basis, which essentially defines or significantly impacts physical and biological behavior of synthetic macromolecules1,2,3,4,5. Although the conformational analysis and direct visualization of naturally assembled rigid and semirigid macromolecular structures, such as proteins and polynucleotides, has been rapidly evolving, flexible synthetic polymers remain somewhat outside of this success story6. The selection of tools for studying, visualizing and controlling conformation of single-chain molecules of flexible synthetic polymers in solutions is still severely limited, mainly due to the propensity of these macromolecules to fold and aggregate under the conditions of the analysis6. Nevertheless, the structural and conformationally dynamic features of flexible chains, as they are exposed to various types of application related activities, can provide important clues to optimizing their performance and understanding their mechanism of action. This is especially important for physiologically active macromolecules and their interactions with living systems.

Poly[di(carboxylatophenoxy)phosphazene], PCPP is a synthetic ionic macromolecule, which stands out from conventional polyelectrolytes due to its potent immunoadjuvant activity, which has been demonstrated in numerous in vivo studies and in the clinical environment7,8,9. The accumulated physico-chemical, biophysical and in vivo data reveal that PCPP realizes its biological activity through spontaneous self-assembly with antigenic proteins in aqueous solutions7. However, the comprehensive structural characterization of PCPP in aqueous solutions and its supramolecular assemblies with proteins has not yet been carried out. It is expected that understanding the mechanism of supramolecular assembly and detailed characterization of the resulting structures remains one of the key prerequisites for the successful development and optimization of this vaccine delivery system. Furthermore, direct measurements of conformational changes from intermolecular interactions between proteins and polyelectrolytes should provide insights on design of materials to modulate biological performance of polyelectrolyte-based biomaterials.

Cryogenic Electron Microscopy (cryo-EM) and Atomic Force Microscopy (AFM) remain some of the most advanced and prevalent techniques for visualizing single biomacromolecules. Cryo-EM provides real-space microscopy images of biomolecules embedded in vitreous, glass-like ice shedding light on the behavior of macromolecules in their native environment10,11,12,13. The technique is rapidly becoming an attractive method in the field of structural biology with molecular structures as small as the 52 kDa streptavidin and the 40 kDa ribonucleic acid (RNA) are within the practical limits of characterization10. More recently, cryo-EM has also become a game-changing tool not only in protein structure determination, but also in protein assemblies, protein ligand–receptor interactions, and, now, assemblies of synthetic peptides14,15. However, being developed as a predominantly biological technique, the use of cryo-EM to study synthetic polymers and soft matter has been much more limited mainly due to insufficient electron-optical image contrast in polymers composed of lighter elements16,17,18 and their inherent flexibility that prevents robust multi-particle averaging that is key to atomic resolution single particle reconstruction of protein molecules. Therefore, most of cryo-EM research on synthetic polymers has been focused not on visualization of individual chains of linear polymers, but on investigations of high mass contrast systems, such as their supramolecular assemblies, polymer-based fibers and micelles. Those structures of higher polymer hierarchy are typically represented by filaments formed by helical polypeptides15,17, dendronized polymers19,20,21,22, spherical23 or cylindrical24,25,26 polyelectrolyte brushes with dense side chains, various micellar assemblies27,28,29,30, non-polymer based nanotubes of an amphiphilic cyanine dye31, aggregates formed by polyelectrolyte complexes32 or stacking of the conjugated polymer in organic solvents33. In contrast, AFM has been successfully employed for imaging single-chain linear synthetic polyelectrolytes, however this technique only allows visualization of macromolecules in their adsorbed state34,35,36,37. It can be expected that the complementary information provided by two methods, which exploit distinctly different experimental techniques, can offer invaluable insights into solution behavior of PCPP and its self-assembly with model protein systems. Alternate approaches, such as small-angle neutron scattering provide average chain conformation and rely on isotopic labeling in order to directly measure the form factor of a chain molecule, which is related by Fourier transform to the pair distribution function and by convolution to real-space density distribution38. Therefore, models must be applied to relate the reciprocal space to the real-space trial structure due to the classical phase problem in scattering39.

Here we successfully visualized and compared images of individual chains of a linear synthetic polymer, PCPP, in aqueous solution in its vitrified and surface adsorbed states. Both cryo-EM and AFM methods demonstrated spontaneous association of PCPP with model antigenic proteins – bovine serum albumin (BSA) and hen egg lysozyme. Two distinct scenarios of complex formation were revealed. While self-assembly with BSA occurred without noticeable changes in polymer conformation, significant compaction of linear chains was observed for lysozyme. This was manifested in the formation of loops and re-crossed chains by cryo-EM and formation of spherical caps of adsorbed complexes by AFM. Given previously demonstrated in vivo immunopotentiating effect of PCPP with either of the studied proteins40,41, the structural insights provided by cryo-EM and AFM can offer important clues in studying the mechanism of biological activity of this important macromolecular immunoadjuvant. Such measurements offer unique data needed to interpret measurements, such as dynamic light scattering and small-angle X-ray and neutron scattering in polyelectrolyte-protein complex solutions.

Results

Direct visualization of individual PCPP chains using cryo-EM and AFM methods

PCPP is a weak synthetic polyelectrolyte with the mass-average molar mass of 800,000 g mol−1, which is fully soluble under neutral and basic conditions42. Its molecular structure includes two carboxylate moieties per repeat unit (Fig. 1a). Cryo-EM images of PCPP samples prepared in a vitrified state from solution in a phosphate buffer at pH 7.4 show clearly visible polymer chains (Fig. 1a, b and Supplementary Fig. 1). Their appearance can be described as random coils of various length with a majority exceeding 100 nm in size. Some loose loops can be noticed in the images, but no apparent re-crossing is visible. The density of macromolecular chains visible in cryo-EM images was dependent on the polymer concentration in the sample. Highly crowded fields observed at 2.5 mg mL−1 PCPP contrasted with images of isolated macromolecules, which were obtained at five-fold dilution (Supplementary Fig. 1a, b).

Fig. 1: Visualization of individual PCPP chains.
figure 1

a Chemical structure of PCPP repeat unit. b Cryo-EM image of a single PCPP chain (defocus 2.12. μm) and (c) projections of representative polymer chain trajectories obtained from cryo-EM images shown in Supplementary Fig. 1. d AFM images of individual PCPP chains.

Full size image

Investigation of PCPP chains by AFM was initially impeded by difficulties in adsorbing the anionic polyphosphazene macromolecule on the surface of freshly-cleaved mica. To obtain sufficient adhesion of PCPP to the inorganic surface, mica was modified with bovine serum albumin (BSA) (Supplementary Fig. 2). AFM images of PCPP adsorbed on such surfaces are shown in Supplementary Fig. 1d, e and in Supplementary Fig. 3a. Although, the dimensions of adsorbed PCPP chains appear to be somewhat larger than those detected in vitrified samples, PCPP coils visualized by both methods appear to be of similar shape. This may reflect that adsorption of PCPP is favored for higher molar mass with the smaller chains rinsed off during the sample preparation step.

Self-assembly of PCPP-BSA complexes as revealed by cryo-EM

To evaluate the potential effect of BSA on PCPP chains visualized by AFM, protein-polymer solutions in their vitrified form were also studied by cryo-EM. BSA is a globular protein with molar mass of 66 kDa, which was characterized with small-angle neutron scattering as a prolate ellipsoid with a = (7.0 ± 0.2) nm and b = (2.0 ± 0.1) nm with the radius of gyration of 3.05 nm, where a and b are the semi-major and minor axes, respectively43. The image of PCPP-BSA mixture shows the assembly of small dots along the curved lines (Fig. 2a). These densely dotted lines resemble chains of free PCPP visualized under the same experimental conditions (Fig. 2b), but show a superior contrast. Essentially, the approach can be considered as ‘decorating’ of PCPP chains through the association of synthetic macromolecule with BSA. These results are in line with previous reports on noncovalent association of BSA and PCPP demonstrated by size exclusion chromatography and asymmetric flow field flow fractionation (AF4)40,44. The comparison of Fig. 2a, b suggests no significant ‘disturbing’ effect of BSA binding on the conformation of the polymer. These results, along with DLS studies on the effect of BSA on the dimensions of PCPP (Supplementary Fig. 3b), indicate that any potential influence of BSA on the conformation of PCPP chains shown in AFM generated images (Fig. 1d and Supplementary Fig. 3a) is minor.

Fig. 2: Contrasting of PCPP with BSA.
figure 2

Cryo-EM images of (a) PCPP-BSA (defocus 2.01 μm) and (b) PCPP (defocus 0.10 μm) (2.5 mg mL−1 PCPP, 0.25 mg mL−1 BSA, 50 mmol phosphate buffer, pH 7.4).

Full size image

Interactions of PCPP with positively charged protein – lysozyme

Hen egg lysozyme, a protein with a large positively charged surface area (Fig. 3a), forms complexes with PCPP at neutral pH45,46. Although such interactions typically result in a severe aggregation and phase separation45, PCPP-lysozyme mixtures prepared at 1:7 polymer-to-protein mole ratio are essentially clear solutions with no aggregates detected by dynamic light scattering (DLS)(Supplementary Fig. 4). Fluorescence quenching data and size exclusion chromatography analysis provide clear indication of binding in the system (Fig. 3b, c), while zeta potential distribution profiles suggest that this is an electrostatically driven process (Fig. 3d). Nevertheless, the above-mentioned data, along with the isothermal titration calorimetry results indicating the dissociation constant of the complex below the micromolar range (Supplementary Fig. 5 and Supplementary Table 1), do not yet constitute a proof of the single-chain complex formation as the co-existence of free PCPP chains and larger aggregates still cannot be ruled out completely. To that end, the analysis of complexes using cryo-EM and AFM methods could provide critical information on the mechanism of complexation in this system, which is important for further understanding of the mechanism of immunoadjuvant activity of PCPP.

Fig. 3: Interactions of PCPP with lysozyme.
figure 3

a Schematic presentations of electrostatic potential surfaces of hen egg lysozyme64 (images generated using PyMOL 2.5.5 (Schrödinger Inc., New York, NY); b Fluorescence of lysozyme – dotted line, PCPP-lysozyme complex (1:7 mole mole−1) – solid line and PCPP – dashed line (0.5 mg mL−1 PCPP, 0.06 mg mL−1 lysozyme, 50 mM phosphate buffer, pH 7.4, excitation wavelength: 280 nm); c size-exclusion chromatography profiles of PCPP-lysozyme – solid line and PCPP – dotted line (0.5 mg mL−1 PCPP, 0.06 mg mL−1 lysozyme, excitation/emission wavelengths: 280 nm/340 nm; 50 mM phosphate buffer, pH 7.4) (d) zeta-potential distribution profiles for PCPP – dashed line, lysozyme – dotted line and their mixture -solid line (0.5 mg mL−1 PCPP, 1 mg mL−1 lysozyme, complex: 0.5 mg mL−1 PCPP and 0.06 mg mL−1 lysozyme, 50 mM phosphate buffer, pH 7.4).

Full size image

Representative cryo-EM images of vitrified samples of PCPP-lysozyme mixtures (1:7 mole ratio) and their chain trajectory are shown in Fig. 4a, b. The absolute majority of coils visible in the field are compact coils with dimensions significantly smaller than for free PCPP chains described above. The conformations show extensive looping and apparent re-crossing of chains. The size distribution obtained by encircling the images of individual coils (pervaded volume), (as shown in Supplementary Fig. 6) indicate that the majority of them can be characterized with diameters in the range of 20 nm to 40 nm (Fig. 4c). The size distribution obtained from cryo-EM measurements was compared with the results of DLS studies, in which the volume distribution was crudely assessed using Mie theory assuming spherical shape and homogeneity of complexes47 (Fig. 4c). The comparison of results shows close similarity between distribution profiles, although DLS assessment revealed that the majority of complexes were in 10 nm to 30 nm diameter range.

Fig. 4: Visualization and characterization of PCPP complexes with lysozyme.
figure 4

a, b Cryo-EM images of PCPP-lysozyme complexes (1:7 mole ratio, 0.5 mg mL−1 PCPP and 0.06 mg mL−1 lysozyme, defoci: 0.10-2.50 μm) and projections of their chain trajectories. c, d Complex size distribution as determined by measuring pervaded diameters of coils in cryo-EM images and DLS.

Full size image

AFM visualization of PCPP-lysozyme complexes was initially attempted using BSA-mediated adsorption on mica surface – the technique described above for free PCPP chains. Although some AFM images of complexes were collected (Supplementary Fig. 7), the best adsorption and image resolution was obtained by using mica surface pretreated with branched polyethyleneimine (bPEI) (Fig. 5a). In either case, images showed spherical caps with an average diameter of around 40 nm and 14 nm average height (Fig. 5b–d) with no individual chains detectable under conditions of the experiment. Although, AFM data provide somewhat more limited information compared to cryo-EM images, the results show a compaction of PCPP chains upon addition of lysozyme. The dimensions of the complexes as visualized by cryo-EM and AFM are comparable, especially if the anisotropy of the spherical cap (larger pervaded diameter) is attributed to the effect of surface adsorption in the sample preparation process. It should also be noted that similarly to protein molecules, PCPP could partially adsorb at the water-air interface48,49,50 resulting in some increase of its apparent size in cryo-EM images. The effect might also be a result of drying out in atmospheric AFM as dehydration might lead to shrinkage of the structure.

Fig. 5: Visualization and characterization of PCPP using AFM.
figure 5

a AFM images at various magnifications. b Schematic presentation of adsorbed complex with average dimensions. c, d Diameter (solid-colored squares) and (open circles) height distribution of complexes (AFM images are taken from bPEI-treated mica substrates).

Full size image

Chain statistics

To quantify the flexibility of PCPP the imaged polymeric chains were traced to provide backbone contours for analysis of conformation. The tracing and conformational analysis was performed according to the SmarTrace algorithm51. Here the traced chains were randomly segmented to nonoverlapped fragments of various contour length. Then for all of the segments squared end-to-end distance <R2(s)> and the tangent vector correlation <cosθ(s)> length-dependent trends were compared with the predictions of the worm-like chain (WLC) model for polymers (Supplementary Figs. 8–10). The persistence lengths for PCPP and PCPP/BSA are presented in Table 1.

Table 1 Persistence lengths values (Ip) of PCPP obtained on the basis of AFM and cryo-EM measurements (the uncertainty (±) in the persistence length is given by standard deviation from the average value)
Full size table

Discussion

The evolving interest in PCPP and its structural analogs is primarily dictated by the potent immunoadjuvant effect demonstrated by this biodegradable macromolecule in multiple animal models and in clinical trials7,8,9. Immunoadjuvants are important components of contemporary vaccines, which enhance, prolong and modulate antigen-specific immune responses52,53,54. Although, the exact mechanisms of action still remain under discussion, the antigen delivery capabilities constitute some of the key features for many representatives of the class55. In particular, dispersed systems, such as aluminium hydroxide or phosphate (Alum or Alhydrogel), nanoemulsions, liposomes, polymer nanoparticulate are known to be capable of effective presentation of the antigen. Their dimensions and well-defined surface, on which the antigen is displayed, provide for virus-mimicking characteristics and largely enabling uptake by immunocompetent cells55. PCPP – a linear flexible macromolecule appears to lack those essential features, but nevertheless outperformed many of particulate adjuvants in vivo. Visualization of macromolecular architecture and conformation of PCPP, as well as their complexes with proteins in solutions constitute important prerequisites for understanding the mechanism of activity of this class of adjuvants.

Cryo-EM of conventional synthetic polymers, except of some highly branched structures and aggregates, has not been sufficiently advanced due to low mass contrast exhibited by most conventional macromolecules with carbon-carbon backbones. To that end, PCPP offered some interesting possibilities since it displays some structural similarities to DNA, which shows high mass contrast in cryo-EM studies due to its phosphorus containing backbone and electron-dense base pairs21. In fact, the phosphorus-nitrogen backbone of PCPP along with two aromatic side groups per repeat unit, enabled direct visualization of its macromolecular coils in samples prepared by vitrification of solutions at near physiological pH. To the best of our knowledge, this is the first report on a direct visualization of single chains of a linear synthetic polymer by cryo-EM. The observed images of individual chains were consistent with those obtained by AFM studies using surface adsorbed PCPP.

Further study on interactions of PCPP with BSA and lysozyme provides a direct proof of polymer association with proteins in aqueous solution. Moreover, in either case, the formation of complexes on the basis of a single polymer chain was observed with the absence of larger multi-chain aggregates. The latter were interpreted from previous studies by DLS and AF4, which are ensemble methods45. The sample processing for cryoEM imaging, such as blotting and vitrification, may remove such large aggregates, although it cannot be ruled out that simply a much larger statistics are needed in order to observe small populations of larger objects within the field of view at the studied stichometry of mixing. Furthermore, both ultrathin carbon support film (cryo-EM) and the BSA treated mica (AFM) may interact with PCPP and its complexes, therefore potentially causing some disturbing effect on their conformational states. However, multivalent cation-mediated adsorption of negatively charged macromolecules to mica is known to be driven by strong ionic interactions56 and PCPP chains were retained on the surface even after rinsing (AFM studies). In contrast, the adsorption of polyacid on carbon was reported to be maximized only under acidic conditions, at which the polymer was weakly charged and coiled, and was greatly suppressed at neutral pH57 – the conditions used in our cryoEM study.

It has to be noted that both proteins selected for the present study were successfully adjuvanted by PCPP in immunogenicity studies in vivo40,41. Therefore, the results provide important information relevant to the ongoing discussion on the mechanism of action of polyphosphazene adjuvants. The study revealed two distinct scenarios of the way PCPP spontaneously self-assembles with proteins in aqueous solution. BSA, a protein with an overall net-negative charge at physiological pH, associates with PCPP without introducing significant conformation changes in the polymer (Fig. 2). In contrast, interactions with lysozyme, which has a net-positive charge at neutral pH, are sufficiently strong to cause compaction of PCPP chains, which is demonstrated by both cryo-EM and AFM visualization (Figs. 4 and 5). Regardless of the resulting complex conformation, PCPP maintains its ability to enhance immunogenicity of either of these proteins40,41. Direct measurements will enable informed interrelationships on polyelectrolyte-protein binding mechanisms driven by electrostatic interactions or hydrophobicity and solvent quality arguments that may occur as in this work at the single chain level versus macrophase separation58,59.

In silico visualization of interactions between PCPP and antigens would be an important future development. However, quantitative atomistic molecular dynamics simulations with validated force fields for PCPP are not available at the present time. Such simulations to study binding between protein and PCPP with explicit water would provide mechanistic insight into the current observations, especially with respect to the roles of electrostatic and Vander Waals or hydrophobic group influences on the binding. However, simulating the high molecular weights would be resource intensive, if not prohibitive. Therefore, in silico studies that consider the binding within a few persistence lengths would inform upon the local energetics, but lack the large-scale conformational changes. Such approaches are along the lines of the extensive research in structural biology where transient protein-protein interactions60 and subsequently docking models61 score the complexation. A development of tools along these lines that engage synthetic polymer and structural biology would be welcome.

Conclusions

The determination of vaccine particle structure represents a modern characterization challenge and world-wide interest. Advances in cryo-EM have enabled characterization of biomacromolecules in the vitrified state due to sufficient mass contrast between the compact protein globule or nucleic acid filaments and surrounding water. Synthetic polymers – flexible linear chains composed of lighter elements typically do not provide for sufficient electron-optical image contrast and have been visualized in their higher hierarchy states, such as supramolecular assemblies or filaments. Here, it was demonstrated that the phosphorus and nitrogen containing linear backbone, intrinsic to polyorganophosphazenes, enables substantial electron microscopy contrast to directly observe individual chains. The spontaneous self-assembly with antigenic proteins in the form of single-chain complexes was observed for the first time under conditions of 1:7 stoichiometry of mixing. The contrasting structures of PCPP complexes with BSA (net-negative) and lysozyme (net-positive) charge offer a more detailed view of the mechanism of self-assembly and the role of protein characteristics, such as their amphoteric character and charge heterogeneity. The stiffening of PCPP upon BSA binding, as revealed by the persistence length estimates, suggests a new mechanism that requires computer simulation modeling on how an already highly charged polyelectrolyte further stiffens. This contrasts with PCPP binding with lysozyme, which leads to some compactization of objects, reminiscent of condensation. In either case, the single chain complexes maintain immunoadjuvant potency due to the flexible and conformational dynamics that allow access to the antigenic protein. In general, the use of polyphosphazene macromolecules, which are well-known for their unprecedented structural diversity, can provide indispensable methodologies for direct visualization of polymer-based supramolecular self-assemblies and polymer-enabled drug delivery systems.

Methods

Certain equipment, instruments or materials are identified in this paper in order to adequately specify the experimental details. Such identification does not imply recommendation by the National Institute of Standards and Technology, nor does it imply the materials are necessarily the best available for the purpose. The symbols M and mM used to represent mol L−1 and mmol L−1, respectively, and are not in SI units, but used to adhere to the conventions of the Journal.

Materials

Lysozyme from chicken egg white, BioUltra, lyophilized powder, ≥98% and bovine serum albumin, lyophilized powder, ≥96% (Sigma-Aldrich, Saint Louis, MO) were used as received. High molecular weight poly[di(carboxylatophenoxy)phosphazene], PCPP (800,000 g mol−1) was synthesized as described previously42,62.

Cryogenic electron microscopy

A sample droplet of 3.0 μL was deposited on holey carbon film TEM grids (Q3100CR1.3-2 nm, Electron Microscopy Sciences, Hatfield, PA), which were plasma-treated beforehand using glow discharge PELCO EasiGlow (Ted Pella Inc., Redding, CA) for 15 s at 10 mA. The atmosphere for glow discharging was a residual air. The grids were then blotted twice for 2 s, each time at 12 °C and 100% humidity on the Vitrobot (Vitrobot Mark IV, FEI, Hillsboro, OR), leaving a thin layer stretched over the grid holes and vitrified in liquid ethane. The temperature was monitored and kept constant in the chamber during all the sample preparation steps. The samples then were transferred to 200 kV Talos Arctica (FEI, Hillsboro, OR) equipped with FEI Falcon3EC direct electron detector. Imaging was performed at temperature about 90 K at an acceleration voltage of 200 kV. The data were collected using EPU software (version 2.9.0) in a counting mode, the pixel size is 0.983 Å, the total electron dose is 48 e Å-2. The total of 90 images of PCPP, 73 images of PCPP-BSA system and 95 images of PCPP-lysozyme system were collected using automated acquisition. The number of frames was 37 for PCPP, 45 for PCPP-BSA system and 30 for PCPP-lysozyme system. Patch motion correction and CTF correction were done in CryoSPARC 4.2.1 (Structural Biotechnology Inc., Toronto, Canada). The data collection statistics is shown in Supplementary Table 2. ImageJ software63 was later used to tune the contrast and brightness of the acquired images. Some persistent background patterns are due to gain reference inconsistencies, but the results are not affected since no structural reconstructions are performed.

Atomic force microscopy

AFM studies were performed using a Bruker Dimension Icon AFM instrument (Bruker, Billerica, MA) in the ScanAsyst mode. The cantilever used for measurements was a ScanAsyst Air probe (Bruker, Billerica, MA) with resonant frequency 70 kHz, spring constant 0.4 N m−1, and with a nominal tip radius of 2.0 nm. The PCPP samples were deposited by dip coating from BSA/PCPP solution (BSA 0.25 mg mL−1, PCPP 0.05 mg mL−1) on freshly cleaved muscovite mica V1 (71856-01, Electron Microscopy Sciences, Hatfield, PA) for 5 min deposition time followed by rinsing and drying in nitrogen flow. The PCPP-lysozyme complexes were deposited by dip coating from (lysozyme 0.006 mg mL−1, PCPP 0.05 mg mL−1) solution on silicon wafers (〈111〉, 0.5 mm, Institute of Electronic Materials Technology, Poland) primed with bPEI for 5 min time followed by drying in nitrogen flow.

Dynamic light scattering and Zeta potential

DLS and Zeta potential measurements were performed using a Malvern Zetasizer Nano ZS (Malvern Instruments Ltd., Worcestershire, UK) using a 532 nm laser with a back-scattering detector an angle of 173° with data recorded and processed in Malvern Zetasizer Software 7.10. Polymer and protein were filtered through poly(vinylidene fluoride) (PVDF) Millex syringe filters (EMD Millipore, Billerica, MA) with a pore size of 0.22 µm prior to measurements. DLS measurements were performed at 25 °C using BRAND (BrandTech Scientific, Inc., Essex, CT) disposable ultra-micro cuvettes with plastic caps for polymer and complex formulations containing 0.5 mg mL−1 PCPP and a protein concentration of 0.5 mg mL−1 in 50 mM phosphate buffer, pH 7.4. Each measurement was averaged from 12 scans. Zeta potential measurements were performed at 25 °C using Malvern DTS1070 cuvettes (Malvern Instruments Ltd., Worcestershire, UK) for the same concentrations with the accumulation and averaging of 100 scans.

Isothermal titration calorimetry

ITC experiments were performed using a Nano ITC SV instrument (TA Instruments, Waters, New Castle, DE, USA) at 25 °C. In a typical experiment, 900 μL of PCPP solution was placed into isothermal chamber and was titrated by 10 μL aliquots of lysozyme from a 250 μL syringe rotating at 36.7 rad s with a 300 s delay between each injection. Each injection generated a heat release curve (microjoules per second versus seconds), which later was processed using NanoAnalyze software, version 3.12.5 (TA Instruments, Waters, New Castle, DE, USA) to yield the heat associated with each injection. Data analysis was performed with the above software and a single set of identical sites (SSIS) binding model was used to calculate thermodynamic parameters: binding constant (Kd), reaction stoichiometry (n), enthalpy (∆H), and entropy (∆S).

Fluorescence

Fluorescence measurements were performed using a BioTek SyNergy neo2 multimode reader (BioTek Instruments, Inc., Winooski, VT). Emission spectra were recorded at a fixed excitation wavelength of 280 nm.

Size-exclusion chromatography

SEC analysis of polymer and polymer/protein complex were conducted using an Agilent 1260 Infinity II Binary LC system (Agilent, Santa Clara, CA) equipped with a G7112B binary pump, G7167A Multisampler, G7116A multicolumn thermostat, G7117C diode array detector, G7121A fluorescence detector, and TSKgel GMPW size-exclusion column (Tosoh Bioscience, LLC, Japan); phosphate buffer (PBS) with 10 % acetonitrile volume fraction used as a mobile phase (0.75 mL min−1 flow rate, injection volume 50 μL).