Integrating Proteins in Layer-by-Layer Assemblies ... - ACS Publications

5 days ago - ces.1−6 For most applications, the immobilization should not change the protein .... at pH 3.0, 5.0, 7.5, or 9.6 and I ≈ 0, 10, 200, ...
0 downloads 0 Views 3MB Size
Integrating Proteins in Layer-by-Layer Assemblies Independently of their Electrical Charge Downloaded via DURHAM UNIV on July 8, 2018 at 12:36:07 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

Aurélien vander Straeten, Anna Bratek-Skicki,† Alain M. Jonas, Charles-André Fustin, and Christine Dupont-Gillain* Institute of Condensed Matter and Nanosciences, Université catholique de Louvain, Place Louis Pasteur, 1 bte L4.01.10, B-1348 Louvain-la-Neuve, Belgium S Supporting Information *

ABSTRACT: Layer-by-layer (LbL) assembly is an attractive method for protein immobilization at interfaces, a much wanted step for biotechnologies and biomedicine. Integrating proteins in LbL thin films is however very challenging due to their low conformational entropy, heterogeneous spatial distribution of charges, and polyampholyte nature. Protein−polyelectrolyte complexes (PPCs) are promising building blocks for LbL construction owing to their standardized charge and polyelectrolyte (PE) corona. In this work, lysozyme was complexed with poly(styrenesulfonate) (PSS) at different ionic strengths and pH values. The PPCs size and electrical properties were investigated, and the forces driving complexation were elucidated, in the light of computations of polyelectrolyte conformation, with a view to further unravel LbL construction mechanisms. Quartz crystal microbalance and atomic force microscopy were used to monitor the integration of PPCs compared to the one of bare protein molecules in LbL assemblies, and colorimetric assays were performed to determine the protein amount in the thin films. Layers built with PPCs show higher protein contents and hydration levels. Very importantly, the results also show that LbL construction with PPCs mainly relies on standard PE−PE interactions, independent of the charge state of the protein, in contrast to classical bare protein assembly with PEs. This considerably simplifies the incorporation of proteins in multilayers, which will be beneficial for biosensing, heterogeneous biocatalysis, biotechnologies, and medical applications that require active proteins at interfaces. KEYWORDS: protein−polyelectrolyte complexes, layer-by-layer, self-assembly, lysozyme, thin film bonding interactions were shown to contribute as well.15 It was found that proteins can be immobilized using that method by replacing one of the PEs by a protein.14 Hence, the exploitation of LbL assembly to create advanced materials coatings offers a wide range of solutions for all abovementioned applications. Multilayer construction by alternate adsorption of a protein and a PE however presents a major drawback. Indeed, PEs generally have a homogeneous charge distribution and adaptable conformation, while proteins are polyampholytes with low conformational entropy and heterogeneous spatial charge distribution.16,17 As a consequence, pH, ionic strength, and PE nature have to be carefully selected in order to achieve a sustainable growth of the multilayer with a given

T

he immobilization of proteins at interfaces is a key challenge for numerous applications in biocatalysis, analytical chemistry, water treatment, chemical manufacturing processes, biotechnology, and medical devices.1−6 For most applications, the immobilization should not change the protein conformation and preferably protect it from denaturation.7 It was found that trapping proteins in highly hydrated polyelectrolyte (PE) layers is a way to meet these expectations.8−13 Such layers can be created by the use of either PE brushes or multilayers obtained by the so-called layer-by-layer (LbL) deposition method. The LbL method presents major benefits compared to PE brushes since it can be performed in soft conditions, it is versatile toward surface geometry and chemistry, and the film architecture can be tuned based on experimental parameters.14 The LbL method usually relies on the alternate adsorption of two oppositely charged PE. This results in a multilayered system that grows through electrostatic interactions, though hydrophobic and hydrogen © XXXX American Chemical Society

Received: May 17, 2018 Accepted: July 2, 2018 Published: July 2, 2018 A

DOI: 10.1021/acsnano.8b03710 ACS Nano XXXX, XXX, XXX−XXX

Article

www.acsnano.org

Cite This: ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

Figure 1. (a) Schematic representation of PPCs formation. Blue circles represent lysozyme molecules, while red lines represent PSSl chains. (b) Hydrodynamic radius (Rh) of the largest population of PPCs measured by dynamic light scattering. (c) The ζ potential of PPCs. In (b) and (c), results are presented as isoresponse maps depending on pH and I1/2. I1/2, used for the y-axis, is inversely proportional to the Debye length κ−1. Black dots show mean values of size and ζ potential measurements (n = 3).

protein.17−20 If care is not taken, then the lack of charge overcompensation of the protein layer, the repulsion between identically charged patches of the protein and the PE, or the protein unfolding resulting from the interaction with the PE layer make either the multilayer growth impossible or useless for applications requiring bioactivity. The successful formation of multilayers was reported, though with a narrow pH range stability that highly depends on the protein−polyelectrolyte interaction.19−21 To increase the range of possible applications, the protein polyampholyte nature and charge anisotropy should be bypassed by providing a building block for the LbL method that standardizes protein charge and favorably interacts with the used counter PE. The successful identification of such a building block will bring many solutions with respect to protein immobilization. In this frame, protein−polyelectrolyte complexes (PPCs) are good candidates as building blocks for LbL assembly. Indeed, the charge of the protein can be standardized by electrostatic complexation with a PE, and the resulting PPCs architecture and charge can be controlled.22 PPCs with a given charge and a PE corona can be obtained, free to interact with the oppositely charged PE used in the LbL assembly. The idea of premixing LbL components prior to assembly has been proposed in the early ages of the LbL method. It was proven to be effective for polyelectrolyte−polyelectrolyte complexes (PECs) assembled in a LbL fashion with a third PE, which gave significantly different film architectures compared to a simpler PE−PE LbL assembly.23 Regarding PPCs, a few studies reported the effect of premixing proteins with PEs prior to their LbL assembly. A loss of the enzymatic activity of the protein was reported by Caruso et al., whereas a possible enhancement, though not directly established, was reported by Onda et al.24,25 With respect to multilayer growth, Lvov et al. improved it by premixing albumin with poly(diallyldimethylammonium chloride).17 Recently, we brought back to the forefront the idea of premixing enzymes with PEs prior to their LbL assembly.26 It was shown that lysozyme charge can be tuned by complexation in solution with poly(styrenesulfonate) (PSS) and then used as a building block for LbL film constructions. The obtained coatings show a higher hydration level and a higher specific enzymatic activity compared to bare protein molecules assembled with PSS. While the diversity that PPCs bring to the array of building blocks used in LbL has been reported, the true advantage that it brings compared to bare protein integration has not been shown yet. Demonstrating that the electrical charge of a

protein can be standardized by a PE and that the resulting PPCs can be used to obtain sustainable multilayer growth, that is, that the multilayer construction using PPCs can be based on PE−PE interactions only, would allow the construction method to be generalized to all proteins, whatever their shape and distribution of amino acid residues, as well as to broader windows of conditions. In the present work, lysozyme is used since it has been widely studied both in PPCs and in multilayers. PPCs are prepared by complexation of lysozyme with low molar mass PSS (PSSl) in different conditions of pH and ionic strength (I). The size and electrical state of the obtained PPCs are investigated in suspension by dynamic light scattering (DLS) and ζ potential measurements. Models are used to compute PSSl flexibility and lysozyme surface potential depending on pH and I. This allows the driving forces for PPCs formation, and the resulting morphology, to be understood. PPCs are then alternately adsorbed with poly(allylamine hydrochloride) (PAH) in multilayers at different pH and I, corresponding to different charge states of lysozyme. For the sake of comparison with the classical way used for protein immobilization via LbL assembly, bare lysozyme molecules are also alternately adsorbed with high molar mass PSS (PSSh). The protein content, multilayer mass, and topography are measured and compared for all constructed multilayers. The generated knowledge with respect to PPCs morphology and lysozyme charge is finally used to further unravel the mechanisms driving LbL construction both for LbL assembly based on bare lysozyme molecules and PPCs.

RESULTS AND DISCUSSION Protein−Polyelectrolyte Complexes (PPCs) Formation and Properties. Lysozyme was complexed with PSSl at pH 3.0, 5.0, 7.5, or 9.6 and I ≈ 0, 10, 200, or 800 mM (Figure 1a) starting from solutions with a (−)/(+net) charge ratio of 2. The charge ratio was computed as the ratio of PSSl charge, that is, −36, to the net structural charge of lysozyme as a function of pH, that is, +6.0, +8.0, +9.5 and +17, respectively, at pH 9.6, 7.5, 5.0, and 3.0.27 When pH and I conditions are mentioned, they are valid for the building of PPCs as well as for the measurement condition. The turbidity of the solution was first measured to confirm the presence of PPCs (see Figure S1). In all conditions above I ≈ 0 mM, the turbidity reached nonzero values and ranged from low to very high, with the highest turbidity measured at pH 7.5 and I = 200 mM. This suggests that PPCs were formed in all conditions above I ≈ 0 B

DOI: 10.1021/acsnano.8b03710 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

Figure 2. (a) Schematic representation of the PPCs as a function of I. Blue circles represent lysozyme molecules, while red lines represent PSSl chains. Computed κ−1 at pH 7.5 (purple line) and RE and lp of PSSl (red and green lines respectively) as a function of I1/2 are presented together with the diameter of lysozyme (dashed blue line) and the Bjerrum length lB (dashed black line). (b) Isoresponse maps as a function of pH and I1/2 of lysozyme electrical surface potential (ψ0). The lysozyme total net charge between pH 3 and 10 is presented in the blue gradient scale.

pH either, except for the contribution of protons and their counterions on ionic strength. The effect of I on PSSl was evaluated by computing the effective persistence length (lp) and the average root-mean-square distance between chain ends (RE) based on the Odijk−Skolnick−Fixman (OSF) theory (see SI for details). The computed values of lp, RE, and κ−1 at pH 7.5 are presented for PSSl as a function of I1/2 in Figure 2a, together with the lysozyme diameter (dashed blue line) and the Bjerrum length lB (dashed black line). PSSl is in a rod-like conformation at low I, that is, I ≈ 0 mM and 10 mM, since the obtained RE values are 8.8 and 6.7 nm, respectively, which is close to the total PSSl contour length of 9 nm. At higher I, the PSSl chains form random coils, with RE decreasing to 4.3 and 4.1 nm. Values of lp are similarly very much increased at low I (lp= 70 nm at I ≈ 0 mM), but are close to lB at I ≥ 200 mM. The Debye screening length κ−1, which is proportional to the reciprocal of I1/2 and is presented in Figure 2a at pH 7.5, reaches values lower than the lysozyme diameter and lB at, respectively, I = 25 mM (i.e., 5 mM1/2) and I = 200 mM (i.e., 14.1 mM1/2). Proteins are polyampholytes, and the number of charges per lysozyme molecule (ql) was computed as a function of pH based on its amino acid sequence and structure (PDB file: 1HEL) using PDB 2PQR.28 Results are presented in blue as a gradient scale (Figure 2b). The lysozyme surface potential (ψ0) was estimated as a function of I assuming that lysozyme is a sphere with a radius of 1.85 nm (Rs), using29,30

mM and that PPCs with a wide variety of size and shape were obtained. At pH 3.0, 5.0, and 9.6 and I ≈ 0 mM, no turbidity was measured, suggesting either that PPCs are not formed, or remain soluble, or do not scatter enough light to be detected. Dynamic light scattering was used to assess the size of PPCs at different pH and I. The hydrodynamic radius (Rh) of the largest detected population is presented in Figure 1b depending on pH and on I1/2. The Debye length (κ−1), which indicates the distance over which electrostatic interactions are significant, is proportional to the inverse of I1/2. Maps and charts built with I1/2 as a scale can thus be interpreted in terms of κ−1. The pH has not much influence on the size of the largest PPCs. At high pH, PPC size tends to decrease, while it tends to increase at medium pH. A nonmonotonic dependence on I1/2 is observed, with Rh values increasing, up to more than 2 μm, from I ≈ 0 to 200 mM (i.e., from 0 to 14.1 mM1/2) and then decreasing from I = 200 to 800 mM (i.e., from 14.1 to 28.3 mM1/2). The ζ potential measurements were performed on PPCs depending on pH and I (Figure 1c). The PPCs are negatively charged over the explored ranges of pH and I. Nevertheless, the surface electrical potential is more negative when the complexes are formed at low pH and low I. For example, the ζ potential measured for PPCs at pH 9.6 and I = 800 mM is only −3.4 ± 0.4 mV, while at pH 3 and I ≈ 0 mM, it is 29.1 ± 0.9 mV. PPCs made 1 week before measurement showed exactly the same charge as when prepared just prior to measurement (Figure S2). This indicates that the PPCs are stable for at least 1 week. As PPCs are negative for all tested pH and I conditions, they can thus be alternately adsorbed with PAH, a polycation, in order to build PPCs-based multilayers. Understanding the molecular forces driving complexation is crucial to shed light on PPCs morphology and therefore on the way they can interact with PEs to form a multilayer. Therefore, PSSl flexibility and lysozyme electrical surface potential were computed as a function of pH and I, to enable the interpretation of the data presented in Figure 1 and understand the way PPCs interact with PAH to form a multilayer. PSSl is a strong polyanion and its degree of charge does not depend on pH. Hence, its conformation is not influenced by

ψ0 =

ql 4πε0εrR s(1 + κR s)

where ε0 is the vacuum permittivity, εr is the relative static permittivity of water, and κ is the reciprocal of the Debye length. ψ0 is presented in Figure 2b as an isoresponse map of pH and I1/2, showing that ψ0 decreases as pH and I increase. For instance, ψ0 = 169 mV at pH 3 and I1/2 ≈ 0 mM1/2, while ψ0 = 6 mV at pH 10 and I1/2 = 14.1 mM1/2, that is, I = 200 mM. The isoelectric point was experimentally found to be reached at pH 11.35, in line with the low ψ0 values predicted at high pH.31 C

DOI: 10.1021/acsnano.8b03710 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

Figure 3. Lysozyme amount within (a.1) [Lys-PSSh]4.5 or (b) [PAH-PPCs]5 multilayers presented under the form of isoresponse maps as a function of pH and I1/2. Black dots show mean values of experimental data (n = 3). Lysozyme net charge (a.1) and PAH ionization degree38 (b) are represented, respectively, in blue and green as a function of pH as a gradient scale. κ−1 as a function of I1/2 is also presented at pH 7.5 for the ease of data interpretation. (a.2) Molecular structure of lysozyme with Coulombic surface coloring. From left to right: aspartic acid, glutamic acid, and histidine residues were successively deprotonated. The electrostatic potential scale ranges from −433 mV (red) to 433 mV (blue). These molecular structures were produced using the UCSF Chimera package and lysozyme structure (P00698).39,40

different pH values, it results in the presence of different counterions that may also influence complexation. The nonmonotonic PPCs size measured as a function of I (Figure 1b) can be explained by a balance between the increasingly compliant conformation of PSSl and less effective electrostatic interactions as I rises. When RE decreases, that is, PSSl flexibility increases, with increasing I (Figure 2a), it leads to a possible better accommodation of PSSl to lysozyme surface charges with therefore a higher number of PSSl monomers per chain interacting with each protein molecule, as sketched in Figure 2a.29 On the other hand, electrostatic

The size variation of PPCs with pH can be explained by the increasingly positive lysozyme charge as pH decreases. At low pH, the high Coulombic attraction between the highly positively charged lysozyme molecules and negatively charged PSSl chains is favorable to PPC formation. Inversely, at a pH close to the lysozyme isoelectric point (pH 11.35), the lower net charge of the protein is less favorable to PPC formation. This is confirmed by simulations that show that at fixed I, that is, at fixed lp since PSSl is a strong polyacid, the probability of binding lysozyme to PSS monomers increases when the net protein charge increases.29 It should be noted that, although a low concentration of the different buffers was used to obtain D

DOI: 10.1021/acsnano.8b03710 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

Figure 4. Total hydrated mass measured by QCM-D after adsorption of a cushion of two PAH-PSSh bilayers [PAH-PSSh]2 (shaded area) followed by either (a) [Lys-PSSh]5 or (b) [PAH-PPCs]5. Multilayers were constructed at pH 3.0 (a.1 and b.1) or 7.5 (a.2 and b.2) and I ≈ 0 mM (dotted line) or 200 mM (solid line).

interactions decrease with increasing I due to charge screening (see κ−1 in Figure 2a and ψ0 in Figure 2b). When κ−1 reaches a value lower than lB, that is, at I > 200 mM (Figure 2a), one can expect that the electrostatic contribution to complexation becomes negligible and that other forces such as van der Waals forces and hydrophobic effects govern protein−PE interactions.32 The Debye−Hückel approximation is actually not valid at high I, and other models than OSF must then be developed. The high PPCs size measured at I = 800 mM and low pH might be due to favorable hydrophobic interactions between PSS l and lysozyme.33 It is worth mentioning that stiffening of PE at high I has been reported.34 This led to a nonmonotonic dependence of lp on I, which could further explain a nonmonotonic protein−polyelectrolyte binding constant and PPCs size as a function of I. Comparison of the two maps describing the surface electrical properties of PPCs (Figure 1c) and of lysozyme (Figure 2b) shows very similar trends, with negative and positive values for PPCs and lysozyme, respectively. When lysozyme surface potential increases, the resulting PPCs take an increasingly negative charge. The PSSl stiffening at low I leads to a higher overcompensation of lysozyme charges since PSSl is less likely to accommodate to the protein surface charge distribution. This explains the higher PPCs negative charge at lower I (Figure 1c). The higher overcompensation of charges at low I is represented by the PPCs sketch in Figure 2a. The accommodation of PSSl to lysozyme can also be interpreted in terms of lp, assuming that the lysozyme is a sphere with a diameter that gives realistic charge−charge interactions, that is, 3.7 nm29 (see Figure 2a and SI for data computation). The lp is higher than the lysozyme diameter at I ≈ 0 mM and 10 mM (lp

= 73.8 and 4.26 nm respectively), while it is lower than the lysozyme diameter at I = 200 mM and 800 mM (lp = 1.16 and 1.04 nm respectively). This is further supported by Monte Carlo simulations that showed a decreasing density of PE in close vicinity of lysozyme as I decreases.29 This is sketched in Figure 2a where lysozyme density in PPCs is lower at low I. The interplay between electrostatic interactions and polymer flexibility results in PPCs with very different surface charge and morphology as a function of I. Based on our data and on the existing literature, it appears that PPCs with very different morphologies can be obtained as a function of pH and, to a larger extent, of I. As sketched (Figure 2a), negative PPCs were always obtained, albeit with a variable protein density and polymer conformation. This will enable us to understand, in the following, how these different PPCs can be alternately adsorbed with PAH in order to construct a multilayered system. PPCs Layer-by-Layer Assembly. The PCCs constructed at pH 3.0, 5.0, 7.5, or 9.6 and I ≈ 0, 10, 200, or 800 mM were alternately adsorbed with PAH on a previously constructed [PAH-PSSh]2 cushion (as represented in Figure 6b). It is worth mentioning that conditions of PPCs formation were kept for multilayer building. For the sake of comparison to the common way used to immobilize proteins using the LbL method, bare lysozyme molecules were alternately adsorbed with PSSh (as represented in Figure 6a). The lysozyme content of the multilayers was quantified by a bicinchoninic acid (BCA) assay after five protein adsorption steps in all pH and I conditions for both [Lys-PSSh]4.5 and [PAH-PPCs]5 systems (Figures 3a.1 and 3b). Multilayer construction at pH 3 or 7.5 and I ≈ 0 or 200 mM was moreover monitored by quartz crystal microbalance with dissipation monitoring (QCM-D), E

DOI: 10.1021/acsnano.8b03710 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

at I ≥ 200 mM, the low immobilized lysozyme amount and the absence of LbL growth of [PAH-PPCs]5 (Figures 3b and 4b) can be attributed to the screening of PE charges at high I. This is in contrast with the [Lys-PSSh]4.5 system that benefits from the better PSSh adaptation to lysozyme surface charge distribution at I = 200 mM (Figure 6a). In multilayers constructed at I = 10 mM, the PSSl shell of the complexes regulates the interaction with PAH, as supported by the negative ζ potential measured for PPCs (Figure 1c). Indeed, if LbL assembly was driven by protein−PE interactions, it would lead to a variation of the immobilized lysozyme amount with pH (as witnessed with the [Lys-PSSh]4.5 system), since the lysozyme charge varies with pH (Figure 3a.1). When multilayers are constructed at I ≈ 0 mM with PPCs, a low lysozyme amount is immobilized between pH 3.0 and 5.0, while it reaches 1.8 μg cm−2 of lysozyme between pH 7.5 and 9.6, which is significantly higher than the amount reached for any condition with the [Lys-PSSh]4.5 system (Figure 3b). Unlike the lysozyme amount, the hydrated film mass is not influenced by pH, and the film weighs ca. 6.5 μg cm−2 when I ≈ 0 mM at both pH 3.0 and 7.5 (Figure 4b). The ratio of lysozyme mass to the total hydrated film mass is thus 0.016 and 0.277 for [PAH-PPCs]5 constructed at I ≈ 0 mM and pH 3.0 or 7.5, respectively, showing that very different layer compositions are obtained. The multilayer constructed at I ≈ 0 mM and pH 3.0 is smooth (Figure 5) and allows the thickness to be measured by ellipsometry. A thickness of only 5 nm was obtained, measured with a Nanofilm EP4 Accurion ellipsometer. Assuming a dry film density of 1.2 g cm−3, the dry mass is only 0.6 μg cm−2, indicating that the multilayer constructed at I ≈ 0 mM and pH 3 contains around 90% of water before drying. This is further confirmed by the high dissipation measured by QCM-D, that is, 6.6 × 10−5. This is the opposite of what is expected for a LbL construction carried out in the same conditions but only with PE. Indeed, the PE stiffening at low I causes lower charge overcompensation, creating usually thinner multilayers,41 as confirmed by the very low hydrated mass of a [PAH-PSSl]5 control multilayer, also constructed on a [PAH-PSSh]2 cushion at I ≈ 0 mM and pH 3 (Figure S4). At low pH and low I, PAH is fully ionized and electrostatic interactions are maximal and effective at long distance (see the green gradient scale for PAH ionization state and κ−1 as a function of I1/2 for the electrostatic screening in Figure 3b). Therefore, when PAH is adsorbed on a PPCs layer, the strong PSSl-PAH interactions may displace the more weakly complexed lysozyme and remove it from the layer, resulting in the observed low lysozyme amount included in the layers. Additionally, the sawtooth pattern observed by QCM upon construction (Figure 4b.1) supports this interpretation: the addition of PAH would induce the release of previously deposited components (and/or the loss of water). This result can also be interpreted as a low PPCs stability in these conditions of LbL construction. Lysozyme release could be facilitated by both its low charge density compared to PAH and the low PSSl accommodation to lysozyme charge that would leave many free negative charges as evidenced in PPCs constructed in these conditions (Figure 2a). Since DLS has shown that PPCs are 3D objects much larger than a monolayer of PE, as sketched in Figure 6b at low pH, their use may artificially create loops of PSSl at the interface. The substitution of lysozyme by the highly charged PAH may keep this 3D structure intact, resulting in a thick film that has a high water content and a very low protein content,

and the masses deposited at each step were computed based on the measured frequency shift and the Sauerbrey equation (Figure 4).35 Finally, topographic images of silicon wafers coated with [Lys-PSSh]4.5 and [PAH-PPCs]5 constructed at pH 3 or 7.5 and I ≈ 0 mM were obtained by atomic force microscopy (AFM) in tapping mode, and the root-meansquare roughness (RRMS) was extracted from these images (Figure 5).

Figure 5. Root-mean-square roughness (RRMS) extracted from AFM images (n > 3) and presented as a function of pH for [LysPSSh]4.5 and [PAH-PPCs]5 constructed at I ≈ 0 mM on a silicon wafer.

For [Lys-PSSh]4.5, both the amount of lysozyme (Figure 3a) and the total hydrated film mass (Figure 4a) increase when pH decreases. This result is explained by an increasing lysozyme net charge with decreasing pH, that is, respectively +6.0, +8.0, +9.5, +17.0 at pH 9.6, 7.5, 5.0, and 3.0 (blue gradient scale in Figure 3a.1 and Figure 2b), which thus favors electrostatic interactions.27 The same effect of pH on protein immobilization was already reported for bovine serum albumin and poly(ethylenimine) multilayers.21 Note that PSS is a strong PE whose charge density is not altered by pH. Higher lysozyme amount is found at midrange of I values, that is, 200 mM. In line with DLS results, the nonmonotonic dependence on I can be explained by the balance between lp, κ−1, van der Waals forces, and hydrophobic effect.36 The lp has indeed been shown to be a critical parameter for LbL assembly.37 As represented in the scheme in Figure 6a, at high I value, the electrostatic screening of intrapolymeric charges favors more loopy conformations of the PE (PSSh, in red). These loops favor charge overcompensation upon build-up, which improves protein integration into multilayers. At low pH, lysozyme is uniformly positively charged (see far left molecular structure in Figure 3a.2), while negatively charged patches are present at high pH (see far right molecular structure in Figure 3a.2). These negative patches are unfavorable for LbL assembly with a strong polyanion, especially at low I due to electrostatic repulsion. This is confirmed by QCM-D data that show almost no film mass increase between the first and the last protein adsorption steps carried out at pH 7.5 and I = 0 mM (Figure 4a.2). Topographic images acquired by AFM (Figure S3) show that [Lys-PSSh]4.5 multilayers are smooth. This is confirmed in Figure 5, with RRMS values of the order of magnitude of lysozyme radius (≈1.85 nm). For multilayers integrating PPCs, that is, [PAH-PPCs]5, the amount of immobilized lysozyme is close to zero when the multilayer is constructed at I ≥ 200 mM. It does moreover not depend on pH at I = 10 mM, and it depends on pH at I ≈ 0 mM (Figure 3b). This result suggests that I is the main parameter that drives LbL assembly. In multilayers constructed F

DOI: 10.1021/acsnano.8b03710 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

higher than a [Lys-PSSh]4.5 multilayer, suggesting that [PAHPPCs]5 multilayers constructed at pH 7.5 and I ≈ 0 mM are more hydrated than similarly constructed films based on bare lysozyme molecule integration. In order to confirm that PE−PE interactions drive the LbL assembly, PPCs with a (−)/(+net) charge ratio of 0.5 (PPCs_0.5) were also assembled with PAH at pH 3, 5, 7.5, or 9.6 and I ≈ 0, 10, 200, or 800 mM. Unlike all other PPCs used in this paper, a lower concentration of PSSl, thus of negative charges, was used to form the PPCs_0.5. This results in the PPCs_0.5 being positively charged and in an increased limitation of PSSl to complex with another PE as is it is already fully complexed with lysozyme.26 Lysozyme content was quantified for multilayers constructed with PPCs_0.5 in all pH and I conditions. The results are presented in Figure S5. Whatever the pH and I used for the LbL construction, no lysozyme is immobilized in the multilayers. This confirms that PE−PE interactions drive the LbL assembly only for PPCs having a core−shell morphology, consisting of a negatively charged PE corona shielding a core made of a protein/PSS complex. Such a core−shell morphology is therefore required for the successful integration of proteins in multilayers. When combined with a counter polyelectrolyte of proper charge density, it allows to circumvent issues related to the charge distribution of protein molecules. Taken together, the results thus show that [Lys-PSSh]4.5 and [PAH-PPCs]5 systems have completely different growth patterns, which enables different LbL coatings with wellcontrolled architecture and properties to be obtained. PPCs offer a wide range of deposition conditions to the LbL method. On the one hand, when the multilayer is constructed at I = 10 mM, it allows a construction that is independent of the protein charge state over a very broad range of pH. On the other hand, when the multilayer is constructed at I = 0 mM, two distinct regimes are obtained, both depending on the counter polycation ionization degree: (1) multilayers made of almost only PE and with a high water content are obtained if the counter polycation is fully ionized (even if constructed at I ≈ 0 mM) and (2) a large protein amount is immobilized in highly hydrated films when the counter polycation is