Abstract

Three different series of amino acid side-chain functionalized homopolypeptides were prepared as variants of previously reported α-helical, coacervate-forming cationic polypeptides. Studies of the physical behavior of these polypeptides in aqueous media in the presence of multivalent counterions enabled a better understanding of the molecular requirements for coacervate formation of side-chain functionalized homopolypeptides. Variation in lengths of side-chain amino acid or linker segments in cationic α-helical polypeptides was found either to prohibit coacervate formation or to allow adjustment of the phase transition temperature. A series of charge-reversed, anionic amino acid side-chain functionalized homopolypeptides were also prepared and found to be α-helical and able to form coacervates similar to analogous cationic homopolypeptides. These results illustrate the ability to predictably tune coacervation properties via molecular adjustment of side-chains in homopolypeptides and show that amino acid side-chain functionalized homopolypeptides can be used as a general platform for development of biomimetic, coacervate-forming polymers.
Keywords: polypeptide, biomimetic, multifunctional, coacervate, polyelectrolyte
Introduction
There is considerable interest in understanding the formation of protein coacervates in living systems, including both their biological functions as well as how their physical properties correlate with molecular features of the proteins.1−3 While considerable advances have been made in correlating protein or peptide sequences with coacervation propensity,4−6 there remains a need for well-defined model systems that are also able to replicate the many functional features of coacervate forming proteins.7−13 These include being able to respond to physiologically relevant stimuli including changes in pH, counterions, temperature, and exposure to oxidizing and reducing agents.14−17 Further, for potential downstream use in applications such as therapeutic delivery and materials science it is important that coacervate properties can be readily and predictably tuned by adjustment of molecular features.18−21
We recently reported the synthesis and coacervation properties of the homopolypeptides 1a−d (Figure 1).22 A variety of features in these polypeptides allow them to form coacervates in water that can respond to changes in physiologically relevant stimuli, similar to natural protein based coacervates. Coacervation can also be switched off and on via addition of mild oxidizing and reducing reagents, respectively.22 These properties are enabled by multifunctional side-chains present in every repeating unit, which contain molecular elements that are sensitive to changes in the solution environment. Unlike natural proteins that rely on complex sequences of different amino acid residues for such multifunctionality, these sequentially uniform homopolypeptides possess multiple, precisely controlled functional side-chain features that allow for a more straightforward investigation and understanding of structure-property relationships.
Figure 1.

Our group recently reported coacervate forming amino-acid functionalized poly(S-alkyl-l-homocysteine)s (1a−d).22 In order to better understand how side-chain molecular structure affects coacervation, we have prepared variants of 1a−d that include homocysteine homologues (2b), linker homologues (3c), and polypeptides with reversal of the side-chain charge (4a−d).
In order to better understand how side-chain molecular features in 1a−d affect coacervate formation and to learn how coacervate properties can be altered for different uses, we have prepared side-chain modified variants of these parent homopolypeptides. Our efforts focused on precise modification of three different regions of the side-chains: the amino acid core (2b), the linker (3c), and the terminal charged group (4a−d, Figure 1). These modifications were chosen to investigate effects of altered side-chain lengths and hydrophobicity (2b, 3c), as well as to compare how charge reversal in anionic 4a−d affects coacervation properties compared to those of cationic 1a−d. Our synthetic method is advantageous for preparation of a variety of different multifunctional homopolypeptides due to the efficiency of the post-polymerization modifications as well as the modular nature of the synthesis, which allows facile substitution of small molecule components to generate diversity.22−24
Experimental Section
Materials and Instrumentation
Tetrahydrofuran (THF) and dichloromethane (DCM) were each degassed with dinitrogen and passed through an alumina column before use. Unless otherwise specified, all post-polymerization modification chemistry was performed in glass vials under ambient atmosphere. Co(PMe3)4, l-methionine N-carboxyanhydride (Met NCA), l-homomethionine (Hmt) NCA, and 6-(methylthio)-l-norleucine (Mtn) NCA monomers were prepared according to literature procedures.25 Poly(l-methionine)60 (M60), poly(l-homomethionine)60 (Hmt60), and poly(6-(methylthio)-l-norleucine)60 (Mtn60) were prepared according to literature procedures.26,27 Poly(S-(3-(2-l-valine amido)ethoxy)-2-hydroxypropyl)-l-homocysteine hydrochloride)60, 1b, poly(S-(3-(2-l-leucine amido)ethoxy)-2-hydroxypropyl)-l-homocysteine hydrochloride)60, 1c, and poly(S-(3-(2-l-leucine amido)ethoxy)-2-hydroxypropyl)-l-homocysteine sulfoxide hydrochloride)60, 5c, were prepared as previously described.22 Reactions of M60,Hmt60, and Mtn60 with epoxides were performed in scintillation vials under ambient atmosphere and temperature, unless otherwise specified. Small molecule chemistry was performed in heat-dried glassware under a nitrogen atmosphere unless otherwise specified. All of the other reagents and solvents were used as received. In-house deionized water was used for all aqueous chemistry and dialysis, unless otherwise specified. Thin-layer chromatography was performed with EMD gel 60 F254 plates (0.25 mm thickness) and spots were visualized using a UV lamp or KMnO4 stain. Silicycle Siliaflash G60 silica (60−200 μm) was used for all column chromatography. Silica used for chromatographic purification of NCA monomers was dried under a vacuum at 250 °C for 48 h and then stored in a dinitrogen-filled glovebox. Compositions of mobile phases used for chromatography are given in volume percentages. Dialysis was performed with regenerated cellulose tubing obtained from Spectrum laboratories with a 2000 Da molecular weight cutoff. NMR spectra of solution samples were recorded on Bruker AV400, Bruker AV300 and Bruker AV500 instruments with chemical shifts reported relative to the deuterated solvent used. Samples for circular dichroism (CD) spectroscopy were prepared by using deionized water. CD spectra for solutions of polypeptides were collected on either a JASCO J-715 or a Chirascan V100 spectrophotometer using a 0.1 cm path length quartz cuvette. Cloud point temperature measurements were recorded at a wavelength of 500 nm on an HP 8453 spectrophotometer equipped with an Agilent 8909A temperature control. Differential interface contrast (DIC) images were taken by using a Zeiss Axiovert 200 DIC/fluorescence inverted optical microscope.
General Preparation of Coacervates
Stock solutions of 2b3-4 and 3c3-4 (all with chloride counterions) and 4a−d (all with ammonium counterions) were prepared at 10 mg/mL in DI water, and separate stock solutions of 4c and 4d were prepared in Milli-Q water at 10 mg/mL containing 300 mM NaCl for mixing with lysozyme. Stock solutions of 1c and 5c were prepared at 10 mg/mL DI water containing 300 mM NaCl for mixing with 4c, and stock solutions of chloride salts of different counterions (26 mM) were prepared in DI water containing NaCl (300 mM). A stock solution of 2x phosphate buffered salne (PBS) buffer (300 mM) was also prepared, as well as a stock solution of lysozyme prepared in 10 mg/mL Milli-Q water. To form coacervates, a stock solution of 1c, 5c, counterions, PBS or lysozyme was then mixed with a stock solution of polypeptide in a 1:1 (v/v) ratio.
Circular Dichroism Spectroscopy
Samples of 2b3-4 and 3c3-4 and 4a−d, were prepared at concentrations between 0.5−0.8 mg/mL in phosphate buffer (12-100 mM). The average degree of polymerization for all polypeptides was 60. The buffers were adjusted to pH of 5.0, 7.0, or 9.0 using HCl (0.1 M) or NaOH (0.1 M) before preparing the polymer solutions, and then the final pH of each solution was confirmed using a pH meter. The data were reported in units of molar ellipticity [θ] (deg cm2 dmol−1), which was calculated using [θ] = (θ × 100 × MW)/(c × l) where θ is the measured ellipticity (millidegrees), MW is the average residue molecular mass (g/mol), c is the polypeptide concentration (mg/mL), and l is the cuvette path length (cm). Percent α-helical content was calculated using % α-helical content = ((−[θ222] + 3000)/39000) × 100, where [θ222] is the molar ellipticity at 222 nm.28
Cloud Point Temperature Measurements
Samples of 3c3-4 were prepared at 2.5 mg/mL in DI water 1x PBS buffer (150 mM). Separate stock solutions of polypeptide (5.0 mg/mL) in DI water and 10x PBS (1500 mM) at pH 5.0, 7.0, and 9.0 were used in sample preparation. Samples were prepared by mixing a polypeptide sample solution and pH 7.0 stock solutions of a sodium counterion (e.g., sodium tripolyphosphate) and NaCl in a 1:0.2:0.15 (v/v/v) ratio and then diluting with DI water to achieve a final polypeptide concentration of 2.5 mg/mL. Samples in PBS buffer were prepared by mixing the polypeptide stock solution and 10× pH 7.0 PBS buffer in a 1:0.2 (v/v) ratio, and the mixture was then diluted with DI water to obtain a final polypeptide concentration of 2.5 mg/mL. The final pH of these samples was adjusted as desired by addition of small volumes of the same sample prepared using either pH 5.0 or 9.0 stock solutions. Transmittance of each sample at 500 nm was measured on an HP 8453 spectrophotometer equipped with an Agilent 8909A temperature control. In initial runs, temperature was increased at a rate of 10 °C/min from 20 to 90 °C, followed by additional experiments within a selected temperature range at a heating rate of 1 °C/min. Cloud point temperatures (Tcp) were determined as the temperature at 50% transmission.
General Procedure for Synthesis of N-TFA-Amino Acid (allyloxy)alkyl Amides with Different Linker Lengths
A round bottom flask was charged with a stirbar, the desired N-trifluoroacetyl-lamino acid (1 equiv, TFA-AA), THF (to give 0.38 M solution), and hydroxybenzotriazole (1 equiv, HOBt). The mixture was cooled to 0 °C in an ice bath; then a 3.6 M solution of N,N′-dicyclohexylcarbodiimide (DCC) (1.2 equiv) in THF was added dropwise to the stirred mixture. The mixture was allowed to stir on ice for 30 min, after which a 3.3 M solution of the desired (allyloxy)alkyl ammonium trifluoroacetate (1.1 equiv) in THF was added to the mixture dropwise, followed by dropwise addition of triethylamine (TEA) (2 equiv). The mixture was allowed to stir in the ice bath, which was allowed to warm to ambient temperature overnight. The next day, the resulting suspension in THF was filtered into a round bottom flask and the THF was removed in vacuo. The residue was dissolved in EtOAc to give a 0.30 M solution with respect to the starting quantity of TFA-AA. This solution was left in a freezer overnight, and any further precipitate was removed by filtration. The resulting solution was purified using flash chromatography at a solvent ratio of 2:1 hexanes (hexane) to EtOAc to afford the product as a white solid.
General Procedure for Synthesis of N-TFA-Amino Acid (Glycidyloxy)Alkyl Amides with Different Linker Lengths
The following procedure from the literature was followed with some modifications.23 The desired TFA-amino acid (allyloxy)alkyl amide (1 equiv) was dissolved in dichloromethane (DCM) (3.3 mL/mmol alkene). Commercial 70 % meta-chloroperoxybenzoic acid (mCPBA) (1.5 equiv) was then added, and the mixture was stirred. After full conversion of alkene was confirmed by thin layer chromatography (TLC) (24−36 h), the suspension was cooled in an ice bath. The mixture was treated with 10% Na2SO3 (aq) (1.5 equiv) followed by 10% Na2CO3 (aq) (1.3 equiv), and stirred for 5 min. The reaction mixture was diluted with EtOAc and washed 2x with sat. NaHCO3 (aq), followed by brine. The organic layer was dried over Na2SO4, concentrated in vacuo, and purified by flash chromatography with a solvent ratio of 1:1 hexane to EtOAc to afford the product as a white solid.
General Procedure for Synthesis of N-(allyloxyacetyl)-Amino Acid Methyl Esters
A round bottom flask was charged with a stirbar, 2-(allyloxy)acetic acid (1 equiv, allyl-acid), and DCM to give a final concentration of 0.83 M. To this mixture was added hydroxybenotriazle (1.1 equiv, HOBt) followed by slow addition of 1-ethyl-3-(3-(dimethylamino)propyl)carbodiimide (1.1 equiv, EDC), and the mixture was permitted to stir for 30 min at ambient temperature. In a separate flask, the desired amino acid methyl ester (1.5 equiv, MeO-AA) was mixed with DCM to give a MeO-AA concentration of 0.62 M, and TEA (2 equiv with respect to allyl-acid). The resulting slurry was added dropwise to the reaction flask containing the allyl-acid, HOBt and EDC. The mixture was allowed to stir overnight before being quenched with sat. NH4Cl, and was then extracted 3× with DCM, 1x NaHCO3, and 1× brine. The organic layer was dried over Na2SO4, concentrated in vacuo, and purified by flash chromatography with a hexane and EtOAc solvent system to afford the product as an oil.
General Procedure for Synthesis of N-(Glycidyloxyacetyl)-Amino Acid Methyl Esters
The following procedure from the literature was followed with some modifications.23 The desired N-(allyloxyacetyl)-amino acid methyl ester (1 equiv) was dissolved in DCM (3.3 mL/mmol alkene). Commercial 70 % mCPBA (1.5 equiv) was then added, and the mixture was stirred. After full conversion of alkene was confirmed by TLC (24−36 h), the suspension was cooled in an ice bath. The mixture was treated with 10% Na2SO3 (aq) (1.5 equiv) followed by 10% Na2CO3 (aq) (1.3 equiv) and stirred for 5 min. The reaction mixture was then diluted with EtOAc and washed 2× with sat. NaHCO3 (aq) followed by brine. The organic layer was dried over Na2SO4, concentrated in vacuo, and purified by flash chromatography with a hexane and EtOAc solvent system to afford the product as an oil.
General Procedure for Alkylation of M60
This procedure was taken directly from the general procedure for alkylation of M60 described in the literature,22 where one of the two dialysis conditions listed below was used for purification.
Acidic Dialysis
The sample was dialyzed against 3 mM HCl (aq) (24 h, 3 changes) followed by DI water (24 h, 3 changes). The retentate was lyophilized to provide the amino acid functionalized sulfonium polypeptide intermediates (used in preparation of 2b3-4 and 3c3-4).
Neutral Dialysis
The sample was dialyzed against 6 mM NaCl (24 h, 3 changes) followed by DI water (24 h, 3 changes). The retentate was lyophilized to provide the amino acid functionalized sulfonium polypeptide intermediates (used in preparation of 4a−d).
General Procedure for Demethylation and Deprotection of Alkylated M60 Sulfonium Salts
This procedure was taken directly from the general procedure for demethylation and deprotection of alkylated M60 sulfonium salts described in the literature,22 where one of the two dialysis conditions listed below was used for purification.
Acidic Dialysis
The sample was dialyzed against 50% MeOH (aq) containing 3 mM HCl (24h, 3 changes), followed by 3 mM HCl (24 h, 3 changes) and then DI H2O (12h, 3 changes). The retentate was lyophilized to provide amino acid functionalized polypeptides 2b3-4 and 3c3-4.
Basic Dialysis
The sample was dialyzed against 50% MeOH (aq) containing 3 mM NH4OH (24h, 3 changes), and then DI H2O (12 h, 3 changes). The retentate was lyophilized to provide amino acid functionalized polypeptides 4a−d.
For polypeptides 2b3-4 and 3c3-4, additional processing was performed to enhance their dissolution in H2O. A minimal amount of TFA was added to lyophilized solids of 2b3-4 and 3c3-4 to obtain complete polypeptide dissolution. The solutions were then diluted by addition of 10 volumes of DI H2O and were then transferred to 2000 Da MWCO dialysis bags, after which they were dialyzed against 1 M NaCl containing 3 mM HCl (24 h, 1 change), followed by 3 mM HCl (24h, 2 changes), and then DI H2O (24 h, 3 changes). Solid samples were then recovered by lyophilization. 19F NMR was used to confirm the complete removal of TFA counterions.
Results and Discussion
A variety of new homopolypeptide variants of 1a−d were prepared in a similar manner using our post-polymerization thioether alkylation and demethylation methodology.23,24 This process utilizes simple homopolypeptide precursors that are selectively functionalized using readily prepared small molecule epoxide reagents (Figure 2).23,24 The alkylation and demethylation reactions both proceed in high yields allowing the preparation of desired functional polypeptides, typically with complete modification of all side-chain residues (see Table S1).23,24 The modular synthesis of amino-acid containing epoxide reagents also allows preparation of different desired epoxides using common reagents and similar reaction conditions.22 To enable meaningful comparisons, all new polypeptides were prepared with the same average degree of polymerization as the parent 1a−d samples.22
Figure 2.
(A) Synthesis of homocysteine homologues 2b3 and 2b4 from poly(l-homomethionine)60, Hmt60, and poly(6-(methylthio)-l-norleucine)60, Mtn60. (B) Synthesis of linker homologues 3c3 and 3c4 from poly(l-methionine)60, M60. AcOH = acetic acid; APDC = ammonium pyrrolidinedithiocarbamate; EtOH = ethanol.
For preparation of polypeptide homologues with different core amino acid residues, the poly(l-methionine)60, M60, precursor used to prepare 1a−d22 was replaced with the homologues poly(l-homomethionine)60, Hmt60, and poly(6-(methylthio)-L-norleucine)60, Mtn60 (Figure 2A). Our group had previously reported Hmt65 and Mtn65 and found that the additional hydrophobic methylene units between the peptide backbones and the sulfur atoms in these higher homologues of M65 resulted in increased stability of α-helical chain conformations for sulfoxide derivatives in water.25 Since 1a−d are all predominantly α-helical in water,22 we wanted to study how alteration of chain conformation stability affects coacervate formation and properties. While l-leucine containing 1c had been previously found to possess an “optimal” level of hydrophobicity for formation of coacervates near physiological temperature,22 we chose to use less hydrophobic l-valine for 2b3 and 2b4. This choice was made in an attempt to balance the increased hydrophobicity of the Hmt60 and Mtn60 backbones26 with the decreased hydrophobicity of l-valine.22 Hence, Hmt60 and Mtn60 were reacted with the same l-valine containing epoxide used to prepare 1b,22 followed by demethylation and deprotection to give polypeptides 2b3 and 2b4, respectively (Figure 2A). Polypeptides were purified by dialysis to give fully protonated 2b3 and 2b4 with chloride counterions. After lyophilization, the resulting 2b3 and 2b4 solids were all found to be soluble in DI water. 1H NMR analysis of 2b3 and 2b4 confirmed >99% functionalization of all residues with the desired side-chains.
Preparation of polypeptide homologues with different linker lengths was accomplished by using the same M60 precursor as that for 1a−d22 in combination with new epoxide reagents with increased alkyl chain lengths between epoxide and amino acid functionalities (Figure 2B). For these samples, l-leucine containing epoxides were used to allow comparisons to 1c. The variation of linker lengths in these samples was used to study the importance of the side-chain amino acid distance from the polypeptide backbone. As the side-chain amino acids are moved further from the α-helical polypeptide backbone, they also become further apart, which may result in weaker interactions between side-chain groups.22 To prepare these homologues, M60 was reacted with new l-leucine containing epoxides with different linker lengths, followed by demethylation and deprotection to give polypeptides 3c3 and 3c4 (Figure 2B). Polypeptides were purified by dialysis to give fully protonated 3c3 and 3c4 with chloride counterions. After lyophilization, the resulting 3c3 and 3c4 white solids were all found to be soluble in DI water. 1H NMR analysis of 3c3 and 3c4 confirmed >99% functionalization of all residues with the desired side-chains.
The preparation of charge-reversed, anionic side-chain amino acid containing polypeptides required a redesign of epoxide synthesis to accommodate the reversed amino acid orientation. This was accomplished by methyl ester protection of the same four amino acids used to prepare 1a−d,22 followed by coupling of their amine groups to 2-(allyloxy)acetic acid and then oxidation to give the desired epoxides (Figure 3A). Since charge reversal is a substantial modification, we thought it best to prepare polypeptides with all different amino acids used in 1a−d in order to best evaluate their coacervate forming properties. To help with comparisons, the epoxide linkers were designed so that the charged atoms in 4a−d would be the same number of bonds from the polypeptide backbone as in 1a−d. The four epoxides were then used to alkylate M60 as above, followed by demethylation and deprotection to give anionic polypeptides 4a−d (Figure 3B).23,24 Polypeptides were purified by dialysis against dilute NH4OH in 50% MeOH, followed by DI water to give fully deprotonated 4a−d with ammonium counterions. After lyophilization, the resulting 4a−d white solids were all found to be soluble in DI water. 1H NMR analysis of 4a−d confirmed >95% functionalization of all methionine residues with the desired side-chains, with the remainder being methionine sulfoxide residues that resulted from oxidation instead of alkylation.29
Figure 3.

(A) Synthesis of ester terminated, amino acid containing epoxides. (B) Synthesis of anionic, charge reversed polypeptides 4a−d from M60. EDC = 1-ethyl-3-(3-(dimethylamino)propyl)carbodiimide; HOBt = 1-hydroxybenzotriazole; DCM = dichloromethane; TEA = triethylamine; mCPBA = meta-chloroperoxybenzoic acid.
An interesting property of 1a−d is the ability of these polypeptides to form disordered coacervate phases, despite their possessing ordered α-helical chain conformations. Disorder in the multifunctional side-chains of 1a−d in aqueous solution was found to be an important feature that drives coacervate formation.22 All the new polypeptides reported here were designed to also be α-helical, and this was confirmed by studying their aqueous solubility and chain conformations over a range of pH. Circular dichroism (CD) spectroscopy of 2b3 and 2b4 in aqueous phosphate buffer at pH 5, 7, and 9 revealed that these polypeptides were soluble and predominately α-helical across this range of pH (Figure 4A,B, see Table S2).28 CD spectroscopy of 3c3 and 3c4 in aqueous phosphate buffer at pH 5, 7, and 9 revealed that these polypeptides were also predominately α-helical across this range of pH (Figure 4C,D, see Table S2), although both were found to partially phase separate at pH 9. Based on studies of 1a−d, the side-chain terminal α-amine groups in 2b3, 2b4, 3c3, and 3c4 are expected to have pKa values in the range of ca. 7 to 8,22 and thus would transition from being mostly protonated to mostly non-protonated as pH is raised from 5 to 9. The α-helical conformations of 2b3 and 2b4 were found to be minimally affected by the protonation state. However, the samples of 3c3 and 3c4 did show a significant decrease in the intensity of CD bands and calculated α-helical content at pH 9, which was likely due to partial phase separation observed as the chains were neutralized.
Figure 4.

Circular dichroism spectra of polypeptides (A) 2b3, (B) 2b4, (C) 3c3, and (D) 3c4 at different solution pH. Samples of 2b3 and 2b4 (0.5 mg/mL) and 3c3 and 3c4 (0.83 mg/mL) were prepared in pH 5.0, 7.0, or 9.0 phosphate buffer (12 mM) at 20 °C. (E) Comparison of circular dichroism spectra of 2b3 and 2b4 at 0.5 mg/mL in pH 7.0 phosphate buffer (12 mM) with 1b at 0.5 mg/mL in pH 7.0 phosphate buffer (100 mM), all at 20 °C. Polypeptide 1b was found to be 77% α-helical under these conditions. (F) Comparison of circular dichroism spectra of 3c3 and 3c4 at 0.83 mg/mL in pH 7.0 phosphate buffer (12 mM) with 1c at 0.5 mg/mL in pH 7.0 phosphate buffer (100 mM), all at 20 °C. Polypeptide 1c was found to be 96% α-helical under these conditions.
In terms of overall aqueous solubility and chain conformation, the amino acid homologues 2b3 and 2b4 were found to be quite similar to the parent polymer 1b. In phosphate buffer at pH 7, subtle differences in α-helical content can be observed, where the stepwise increases in amino acid side-chain hydrocarbon length from 1b to 2b3 to 2b4 resulted in small increases in α-helical content (Figure 4E, see Table S2). These observations are consistent with previous findings on M65, Hmt65 and Mtn65 derivatives that showed α-helical stability increased with longer hydrocarbon lengths.26 The polypeptides with increased hydrocarbon linker lengths 3c3 and 3c4 were also found to be similar in terms of chain conformation to the parent polymer 1c. In phosphate buffer at pH 7, only very slight differences in α-helical content can be observed for 1c, 3c3 and 3c4 (Figure 4F, see Table S2). The main difference among these samples is the decreased aqueous solubility of 3c3 and 3c4 relative to 1c at pH 9, which is likely due to the stepwise increases in the hydrocarbon linker length from 1c to 3c3 to 3c4. Overall, the differences in aqueous solubility and α-helical content in all these polypeptides are small. However, increasing amino acid side-chain length was found to provide an increase in conformational order, and increasing linker length was found to diminish polypeptide solubility at higher pH.
The aqueous solubility and chain conformations of anionic 4a−d were also studied over a range of pH. CD spectroscopy of 4a−d in aqueous phosphate buffer at pH 5, 7, and 9 revealed that these polypeptides were soluble and predominately α-helical across this pH range (see Figure S1 and Table S2). Within this pH range, there was no significant change in α-helical content of these polypeptides, which was expected since all chains would be highly charged under these conditions due to the low pKa values (ca. 4) of carboxylate groups.30 There was a slight trend of increasing α-helical content with increased amino acid hydrophobicity (see Table S2), similar to results obtained for 1a−c.22 Sample 4d possessed a lower α-helical content than 4c, which may be due to the CD spectra of 4d being affected by UV absorption of the phenyl rings (see Figure S1).31 It is worth noting that sample 4d is soluble at elevated pH since it remains highly charged, while 1d becomes insoluble in water at pH > 5.5 as the chains become neutralized. Overall, the solubility and stable α-helical conformations of 4a−d in water validated their design as anionic variants of 1a−d.
We previously reported that mixture of solutions of sodium tripolyphosphate (TPP) and 1a−d in aqueous media resulted in immediate phase separation at 20 °C.22 While 1a and TPP formed precipitates, 1b−d and TPP all formed stable liquid coacervate phases. To compare their properties to those of 1b and 1c, aqueous solutions of the amino acid and linker homologues 2b3, 2b4, 3c3, and 3c4 were mixed with solutions of TPP under the same conditions (Figure 5). Immediate phase separation occurred for all samples, but the nature of modification gave distinctly different results. Both amino acid homologues 2b3 and 2b4 when mixed with TPP gave precipitates that possessed jagged edges and irregular shapes, which contrasted with 1b and TPP that had been found to form coacervates under identical conditions (Figure 5A,B).22 The length of alkyl chains between the polypeptide backbone and sulfur atoms was found to be a critical feature for coacervate formation, where the addition of just one or two methylene units in each residue resulted in precipitation instead of coacervation. The increased core hydrophobicity of 2b3 and 2b4 and resulting increased stability of their α-helical conformations in water relative to 1b appeared to be significant enough to drive formation of precipitates, possibly due to overall reduced flexibility of the α-helical chains.26
Figure 5.
Optical micrographs of mixtures of polypeptides with sodium tripolyphosphate in aqueous 150 mM NaCl at pH 7.0 and 20 °C. (A) 2b3 at 3.0 mg/mL and (B) 2b4 at 5.0 mg/mL mixed with sodium tripolyphosphate (12 mM final concentration). (C) 3c3 and (D) 3c4 both at 5.0 mg/mL mixed with sodium tripolyphosphate (13 mM final concentration). After mixing, the resulting turbid suspensions were allowed to settle onto the glass slides before imaging. Scale bars: 20 μm.
Alternatively, increasing polypeptide hydrophobicity by lengthening alkyl linker chains in 3c3 and 3c4 was found to not adversely affect coacervate formation. Mixture of solutions of 3c3 and 3c4 with TPP gave liquid coacervate droplets that spread and coalesced on glass slides and had smooth surfaces (Figure 5C,D), similar to coacervates of 1c and TPP formed under identical conditions.22 Also similar to 1c, solutions of 3c3 and 3c4 in 150 mM phosphate buffered saline (PBS) were found to form coacervates at elevated temperature in a pH dependent manner (Figure 6, see Table S3). As the pH was increased, 3c3 and 3c4 became less charged and more hydrophobic, leading to phase separation at their cloud point temperatures (Tcp). This temperature-dependent phase separation allowed coacervation properties to be directly compared for the linker homologues 1c, 3c3 and 3c4 by measuring Tcp for each at the same pH. With the assumption that pKa values of the three polypeptides are essentially the same, differences in Tcp should reflect changes in properties due solely to differences in linker length. At pH 7.4, Tcp were found to decrease in the order of 1c to 3c3 to 3c4, or in the order of shortest to longest linker length (Figure 6C). This is sensible since 3c4 would be expected to have the lowest Tcp, as it has the longest alkyl linker and is therefore the most hydrophobic. Based on these results, variation in linker lengths was found to be tolerated in the design of coacervate forming polypeptides. This adjustable molecular parameter may be useful for fine-tuning coacervation properties, especially in terms of controlling the phase transition temperature.
Figure 6.

Temperature-dependent coacervate formation in solutions of polypeptides 3c3 and 3c4 at different pH. Panels show optical transmittance at 500 nm for polypeptides at 2.5 mg/mL in 150 mM PBS buffer at different pH measured as a function of temperature. (A) 3c3 and (B) 3c4. (C) Comparison of optical transmittance at 500 nm of 3c3 and 3c4 at 2.5 mg/mL with polypeptide 1c at 3.0 mg/mL, all in 150 mM PBS buffer at pH 7.4. Polypeptide 1c was found to have a Tcp value of 49 °C under these conditions.
More substantial alterations to the structures of 1a−d were created in the preparation of anionic, α-helical polypeptides 4a−d. To test if 4a−d are able to form coacervates in aqueous media, we needed to identify suitable multivalent anions for complex formation with the polypeptides. For initial evaluation, we chose water-soluble CaCl2 (Ca2+) and Co(NH3)6Cl3 (Co3+) as model divalent and trivalent cations, respectively. Aqueous solutions of 4a−d were mixed with solutions of either Ca2+ or Co3+ in the presence of 150 mM NaCl at 20 °C and pH 7.0. All four anionic polypeptides were found to form liquid coacervates with either multivalent cation under these conditions, as confirmed by optical imaging (Figure 7) as well as centrifugation that gave coalesced liquid coacervate phases. Divalent Ca2+ was found to produce smaller quantities of coacervate compared to Co3+, suggesting that a significant fraction of polypeptide remained soluble in the presence of the divalent cations under these conditions (Figure 7A,C,E,G). Also, only small quantities of coacervate were observed under all conditions for 4a and 4b, which are the least hydrophobic of the 4a−d series (Figure 7A-D). In general, coacervation was favored when both polypeptide hydrophobicity and counterion valency were increased, similar to observations of 1a−d.22 Consequently, mixtures of more hydrophobic 4c and 4d with Co3+ resulted in the most robust coacervate formation (Figure 7F,H).
Figure 7.
Optical micrographs of polypeptides 4a−d mixed with different multivalent cations in aqueous media. Solutions of polypeptides at 5 mg/mL in 150 mM NaCl at 20 °C and pH 7.0 were mixed with either CaCl2 or [Co(NH3)6]Cl3 (13 mM final concentration for both) resulting in formation of turbid suspensions that were allowed to settle on glass slides before imaging. (A) 4a with CaCl2, (B) 4a with [Co(NH3)6]Cl3, (C) 4b with CaCl2, (D) 4b with [Co(NH3)6]Cl3, (E) 4c with CaCl2, (F) 4c with [Co(NH3)6]Cl3, (G) 4d with CaCl2, (H) 4d with [Co(NH3)6]Cl3. Scale bars = 20 μm.
The best anionic coacervate formers, 4c and 4d, were selected for further studies aimed at the preparation of complex coacervates with other biopolymers. Solutions of 4c and 4d were mixed with solutions of lysozyme, used as a model cationic protein with pI ∼ 11, in the presence of 150 mM NaCl at 20 °C and pH 7.0. Immediate phase separation occurred for both samples, but both were found to form precipitates, as determined by observation of particles with jagged edges and irregular shapes (see Figure S2). It is possible that the compact, globular structure of lysozyme32 combined with the α-helical conformations of 4c and 4d presented too much order to form liquid coacervates. To further investigate this hypothesis, we mixed an aqueous solution of anionic, α-helical 4c with a solution of cationic, α-helical 1c in the presence of 150 mM NaCl at 20 °C and pH 7.0. For comparison, we also mixed an aqueous solution of anionic, α-helical 4d with a solution of cationic 5c, which is the oxidized, sulfoxide derivative of 1c (see Figure S3),22 in the presence of 150 mM NaCl at 20 °C and pH 7.0. The oxidized polypeptide 5c possesses a disordered chain conformation and is considerably more hydrophilic than 1c, due to solvation of the side-chain sulfoxide groups by water.22 For both mixtures, phase separation occurred, indicating rapid polyion complex formation. The combination of α-helical 4c with α-helical 1c was found to form precipitates as determined by optical microscopy (Figure S3). However, the combination of α-helical 4d with disordered 5c was found to form a viscous, liquid coacervate phase (see Figure S3). These results support the hypothesis that there must be a significant degree of conformational disorder in polypeptides that favors coacervate formation. While α-helical homopolypeptides are able to form coacervates with small multivalent counterions due to side-chain conformational disorder,22 complex coacervates of oppositely charged polypeptides appear to require at least one polypeptide to be conformationally disordered. Our results are consistent with those of Tirrell,33 who showed that a cationic, α-helical polypeptide was able to form a coacervate when mixed with a disordered anionic polypeptide.
Conclusions
We prepared three different series of amino acid side-chain functionalized polypeptides designed as variants of the previously reported α-helical, coacervate forming cationic polypeptides 1a−d. Studies of the physical properties of new polypeptides 2b3 and 2b4, 3c3 and 3c4, and 4a−d provided insights into how adjustment of molecular features in the polypeptide side-chains affects their ability to form coacervates. We found that lengthening the alkyl segments from the polypeptide backbone to the sulfur atoms resulted in precipitation instead of coacervation, and that the linker lengths between side-chain sulfur atoms and amino acids could be used to adjust coacervation transition temperatures. New α-helical, amino acid side-chain functionalized anionic homopolypeptides 4a−d were found to form coacervates similar to cationic 1a−d. This result showed that the design principles of 1a−d can also be successfully applied to anionic polypeptides, thus, expanding the range of properties and potential applications of these materials. The ability to predictably tune coacervation properties via molecular adjustment of side-chains34 in these polypeptides is expected to be advantageous for their use as multifunctional, well-defined models of coacervate forming proteins.
Acknowledgments
This work was supported by NSF (MSN-2202743).
Supporting Information Available
The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/polymscitech.4c00003.
Sample characterization data, additional tables and figures, and spectral data (PDF)
The authors declare no competing financial interest.
Supplementary Material
References
- Brangwynne C. P. Soft active aggregates: mechanics, dynamics and self-assembly of liquid-like intracellular protein bodies. Soft Matter 2011, 7, 3052–3059. 10.1039/c0sm00981d. [DOI] [Google Scholar]
- Waite J. H. Mussel adhesion−essential footwork. J. Exp. Biol. 2017, 220, 517–530. 10.1242/jeb.134056. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Banani S. F.; Lee H. O.; Hyman A. A.; Rosen M. K. Biomolecular condensates: organizers of cellular biochemistry. Nat. Rev. Mol. Cell. Biol. 2017, 18, 285–298. 10.1038/nrm.2017.7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Dzuricky M.; Roberts S.; Chilkoti A. Convergence of artificial protein polymers and intrinsically disordered proteins. Biochemistry 2018, 57, 2405–2414. 10.1021/acs.biochem.8b00056. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kapelner R. A.; Yeong V.; Obermeyer A. C. Molecular determinants of protein-based coacervates. Curr. Opin. Coll. Int. Sci. 2021, 52, 101407. 10.1016/j.cocis.2020.101407. [DOI] [Google Scholar]
- Quiroz F. G.; Chilkoti A. Sequence heuristics to encode phase behaviour in intrinsically disordered protein polymers. Nat. Mater. 2015, 14, 1164–1171. 10.1038/nmat4418. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yewdall N. A.; André A. A. M.; Lu T.; Spruijt E. Coacervates as models of membraneless organelles. Curr. Opin. Coll. Int. Sci. 2021, 52, 101416. 10.1016/j.cocis.2020.101416. [DOI] [Google Scholar]
- Roberts S.; Dzuricky M.; Chilkoti A. Elastin-like polypeptides as models of intrinsically disordered proteins. FEBS Lett. 2015, 589, 2477–2486. 10.1016/j.febslet.2015.08.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Abbas M.; Lipiński W. P.; Wang J.; Spruijt E. Peptide-based coacervates as biomimetic protocells. Chem. Soc. Rev. 2021, 50, 3690–3705. 10.1039/D0CS00307G. [DOI] [PubMed] [Google Scholar]
- Perry S. L.; Leon L.; Hoffmann K. Q.; Kade M. J.; Priftis D.; Black K. A.; Wong D.; Klein R. A.; Pierce C. F.; Margossian K. O.; et al. Chirality-selected phase behaviour in ionic polypeptide complexes. Nat. Commun. 2015, 6, 6052. 10.1038/ncomms7052. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chang L. W.; Lytle T. K.; Radhakrishna M.; Madinya J. J.; Vélez J.; Sing C. E.; Perry S. L. Sequence and entropy-based control of complex coacervates. Nat. Commun. 2017, 8, 1273. 10.1038/s41467-017-01249-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Abbas M.; Lipiński W. P.; Nakashima K. K.; Huck W. T. S.; Spruijt E. A short peptide synthon for liquid-liquid phase separation. Nat. Chem. 2021, 13, 1046. 10.1038/s41557-021-00788-x. [DOI] [PubMed] [Google Scholar]
- Priftis D.; Tirrell M. V. Phase behaviour and complex coacervation of aqueous polypeptide solutions. Soft Matter 2012, 8, 9396–9405. 10.1039/C2SM25604E. [DOI] [Google Scholar]
- Schuster B. S.; Reed E. H.; Parthasarathy R.; Jahnke C. N.; Caldwell R. M.; Bermudez J. G.; Ramage H.; Good M. C.; Hammer D. A. Controllable protein phase separation and modular recruitment to form responsive membraneless organelles. Nat. Commun. 2018, 9, 2985. 10.1038/s41467-018-05403-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Aumiller W. M.; Keating C. D. Phosphorylation-mediated RNA/peptide complex coacervation as a model for intracellular liquid organelles. Nat. Chem. 2016, 8, 129–137. 10.1038/nchem.2414. [DOI] [PubMed] [Google Scholar]
- Nakashima K. K.; Baaij J. F.; Spruijt E. Reversible generation of coacervate droplets in an enzymatic network. Soft Matter 2018, 14, 361–367. 10.1039/C7SM01897E. [DOI] [PubMed] [Google Scholar]
- Lu T.; Nakashima K. K.; Spruijt E. Temperature-responsive peptide-nucleotide coacervates. J. Phys. Chem. B 2021, 125, 3080–3091. 10.1021/acs.jpcb.0c10839. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wei W.; Tan Y.; Martinez Rodriguez N. R.; Yu J.; Israelachvili J. N.; Waite J. H. A mussel-derived one component adhesive coacervate. Acta Biomater. 2014, 10, 1663–1670. 10.1016/j.actbio.2013.09.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Reed E. H.; Hammer D. A. Redox sensitive protein droplets from recombinant oleosin. Soft Matter 2018, 14, 6506–6513. 10.1039/C8SM01047A. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kaminker I.; Wei W.; Schrader A. M.; Talmon Y.; Valentine M. T.; Israelachvili J. N.; Waite J. H.; Han S. Simple peptide coacervates adapted for rapid pressure-sensitive wet adhesion. Soft Matter 2017, 13, 9122–9131. 10.1039/C7SM01915G. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fraccia T. P.; Jia T. Z. Liquid crystal coacervates composed of short double-stranded DNA and cationic peptides. ACS Nano 2020, 14, 15071–15082. 10.1021/acsnano.0c05083. [DOI] [PubMed] [Google Scholar]
- Scott W. A.; Gharakhanian E. G.; Bell A. G.; Evans D.; Barun E.; Houk K. N.; Deming T. J. Active controlled and tunable coacervation using side-chain functional α-helical homopolypeptides. J. Am. Chem. Soc. 2021, 143, 18196–18203. 10.1021/jacs.1c07925. [DOI] [PubMed] [Google Scholar]
- Gharakhanian E. G.; Deming T. J. Versatile synthesis of stable, functional polypeptides via reaction with epoxides. Biomacromolecules 2015, 16, 1802–1806. 10.1021/acs.biomac.5b00372. [DOI] [PubMed] [Google Scholar]
- Gharakhanian E. G.; Deming T. J. Chemoselective synthesis of functional homocysteine residues in polypeptides and peptides. Chem. Commun. 2016, 52, 5336–5339. 10.1039/C6CC01253A. [DOI] [PubMed] [Google Scholar]
- Kramer J. R.; Deming T. J. Biomacromolecules 2012, 13, 1719–1723. 10.1021/bm300807b. [DOI] [PubMed] [Google Scholar]
- Gharakhanian E. G.; Bahrun E.; Deming T. J. Influence of sulfoxide group placement on polypeptide conformational stability. J. Am. Chem. Soc. 2019, 141, 14530–14533. 10.1021/jacs.9b07223. [DOI] [PubMed] [Google Scholar]
- Kramer J. R.; Deming T. J. Biomacromolecules 2010, 11, 3668–3672. 10.1021/bm101123k. [DOI] [PubMed] [Google Scholar]
- Morrow J. A.; Segall M. L.; Lund-Katz S.; Phillips M. C.; Knapp M.; Rupp B.; Weisgraber K. H. Differences in stability among the human apolipoprotein E isoforms determined by the amino-terminal domain. Biochemistry 2000, 39, 11657–11666. 10.1021/bi000099m. [DOI] [PubMed] [Google Scholar]
- Petitdemange R.; Garanger E.; Bataille L.; Bathany K.; Garbay B.; Deming T. J.; Lecommandoux S. Tuning thermoresponsive properties of cationic elastin-like polypeptides by varying counterions and side-chains. Bioconjugate Chem. 2017, 28, 1403–1412. 10.1021/acs.bioconjchem.7b00082. [DOI] [PubMed] [Google Scholar]
- Appel P.; Yang J. T. Helix-coil transition of poly-L-glutamic acid and poly-L-lysine in D2O. Biochemistry 1965, 4, 1244–1249. 10.1021/bi00883a005. [DOI] [PubMed] [Google Scholar]
- Sage H. J.; Fasman G. D. Conformation studies on poly-L-glutamic acid and copolymers of L-glutamic acid and L-phenylalanine. Biochemistry 1966, 5, 286–296. 10.1021/bi00865a037. [DOI] [PubMed] [Google Scholar]
- Matthews B. W. Studies on Protein Stability With T4 Lysozyme. Adv. Protein Chem. 1995, 46, 249–278. 10.1016/S0065-3233(08)60337-X. [DOI] [PubMed] [Google Scholar]
- Priftis D.; Leon L.; Song Z.; Perry S. L.; Margossian K. O.; Tropnikova A.; Cheng J.; Tirrell M. Self-assembly of α-helical polypeptides driven by complex coacervation. Angew. Chem. Int. Ed. 2015, 54, 11128–11132. 10.1002/anie.201504861. [DOI] [PubMed] [Google Scholar]
- Chowdhury P.; Saha B.; Bauri K.; Sumerlin B. S.; De P. Hydrogen bonding-driven self-coacervation of nonionic homopolymers for stimuli-triggered therapeutic release. J. Am. Chem. Soc. 2024, 146, 21664–21676. 10.1021/jacs.4c05624. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.



