Enzyme-Instructed Self-assembly of the Stereoisomers of Pentapeptides to Form Biocompatible Supramolecular Hydrogels

Abstract

This article reports enzyme-instructed self-assembly (EISA) of stereoisomers of pentapeptides as a simple approach for generating biocompatible supramolecular hydrogels as potential soft bionanomaterials. Peptide-based supramolecular hydrogels are emerging as a new type of biomaterials. The use of tyrosine phosphate offers a trigger for enzymatic hydrogelation, and the incorporation of D-amino acids can increase the proteolytic stability of peptides. This work compared four phosphorpeptides that are stereoisomers in terms of rate of dephosphorylation, proteolytic stability, and cell compatibility. The results show that the naphthyl (Nap) capped pentapeptides, containing the amino acid sequence of Phe-Phe-Gly-Glu-_pTyr, are able to undergo EISA to form the hydrogels consisting the nanofibers of the dephosphorylated pentapeptides. The napthyl capped D-phosphopentpeptides, contrasting to a naphthyl capped D-phosphotripeptide (Nap-D-Phe-D-Phe-D-_pTyr), are largely cell compatible. This result, suggesting that the sequence of phophopeptides also dedicates the cell compatibility of the peptide assemblies resulted from EISA, provides useful insights for developing supramolecular hydrogels as potential biomaterials with tailored properties.

This article describes the evaluation of enzyme-instructed self-assembly (EISA) of the stereoisomers of pentapeptides to form supramolecular hydrogels and the evaluation of the cell compatibility of the supramolecular hydrogelators. EISA[1–5], an effective and facile way that integrates enzymatic transformation and molecular self-assembly, has served as a powerful approach for generating supramolecular hydrogels[6,7] in cellular environment for a variety of applications.[8–10] For example, EISA generates molecular nanofibers inside cells[11–13] as well as in the pericellular space[14]. Ulijn et al. also reported the thermodynamically controlled EISA and elucidated the chemical composition/nanostructure relationships of four closely related esters which formed hydrogels[5]. Yang et al. and others reported self-assembled molecules forming nanofibers to inhibit tumor growth[15–18]. Pires and Ulijn also used EISA to generate pericellular nanofibers[19], and Maruyama demonstrated the use of EISA for generating intracellular nanofibers that inhibit cancer cells[20]. One common consequence of EISA of small molecules is the formation of supramolecular hydrogels, resulted from that the three-dimensional network, formed by molecular self-assemblies, traps the water in the interstitial spaces. Supramolecular hydrogels have received increasing attention because of their potential for developing new classes of biomaterials[21,22]. Multiple studies have shown the applications of supramolecular hydrogels in biomedicine, such as drug delivery[23–26]. In addition, hydrogels have become a useful scaffold for tissue engineering[27–32] because cells in a three dimensional hydrogel behave differently from being in two dimensional cell culture[33]. In the use of EISA to generate hydrogels, generally, peptide derivatives (the precursors) act as the substrates of enzymes for generating the corresponding hydrogelators that self-assemble in aqueous solution[1] to form the hydrogels. In order to serve as the substrates of the enzymes, most peptide derivatives explored so far are made of L-amino acids. Despite the natural suitability of L-amino acids as the substrates for the enzyme-instructed hydrogelation, L-amino acids are easily degraded by various proteinases in cellular environment, which may decrease the stability of the peptides and limit the applications of the hydrogels in vivo for long-term. Thus, it is logical to use D-amino acids to replace L-amino acids for increasing the stability of the peptide derivatives in cellular environment or in vivo[34–36]. However, EISA of heterochiral peptides[37–39] and D-peptides receives little attention, and its exploration is still at the beginning[35,40].

Based on the fact that D-peptides are less likely to undergo proteolytic degradation and a naphthyl (Nap) capped pentapeptide (Nap-L-Phe-L-Phe-L-Gly-L-Glu-L-Tyr, 1) known to act as a hydrogelator[41], this work designed and synthesized three stereoisomers of 1 (Scheme 1) to explore the gelation properties, proteolytic stability, and biocompatibility of these stereoisomers in the process of EISA. This study shows that the precursors (2-P, 3-P, and 4-P) of the hydrogelators are biocompatible and result in hydrogels after EISA. In addition, these results indicate that the combination of L- and D-amino acids can control the proteolytic stability of the precursors/hydrogelators, and as expected, the precursor/hydrogelator consisting of all D-amino acid residues is the most resistant to proteolysis. Particularly, contrasting to a naphthyl capped D-phosphotripeptide (Nap-D-Phe-D-Phe-D-_pTyr), which is cytotoxic[14], 4-P, having the sequence of Nap-D-Phe-D-Phe-Gly-D-Glu-D-_pTyr, is cell compatible. This work, suggesting that the sequence of phophopeptides also dedicates the cell compatibility of the peptide assemblies resulted from EISA, provides useful insights for developing EISA to generate supramolecular hydrogels as biomaterials with tailored properties.

Scheme 1.

Enzymatic transformation of the precursors to the corresponding hydrogelators.

Results

As shown in Scheme 1, the stereoisomers are based on a known EISA substrate that has the sequences of Nap-L-Phe-L-Phe-Gly-L-Glu-L-_pTyr (Nap-FFGE_pY, 1-P). According to the previous report,[41] after dephosphorylation, 1-P becomes 1, which self-assembles in water to form nanofibers. The entanglement of the nanofibers results in a hydrogel. In the structure of 1-P, Nap-FF acts as an excellent self-assembly motif[42] due to its ability to provide multiple aromatic-aromatic interactions and hydrogen bonds, and tyrosine phosphate (Y_p) renders the precursor to be a substrate of alkaline phosphatase. Moreover, a previous study reported the use of kinase/phosphatase to regulate the supramolecular hydrogel formed by cell compatible hydrogelator 1.[41] Despite that, 1-P has illustrated the development of EISA for generating supramolecular hydrogels, 1 is susceptible to proteolysis catalysed by proteases, thus its applications is limited. To increase the stability of the precursors and hydrogelators, one approach is to introduce D-amino acid into 1-P or 1. Thus, this work aims to generate a series of stereoisomers of 1 and to examine the properties of the stereoisomers. According to the structure of 1-P, replacing L-Phe by D-Phe generates 2-P, exchanging L-Glu and L-Tyr to D-Glu and D-Tyr produces 3-P, and converting all the L-amino acid residues in 1-P to D-amino acid residues leads to 4-P. In such a design, it is expected 2-P, 3-P and 4-P act as the precursors for hydrogelators 2, 3, and 4, respectively. Based on the synthetic procedure used to make 1-P, solid phase peptide synthesis (SPPS) was used to produce 2-P, 3-P, and 4-P. Briefly, the protected phosphotyrosine (Fmoc-_pTyr (D or L-enantiomer)) was added to 2-chlorotrityl chloride resin. After capping any open sites on the resin, 20% piperidine was then added to deprotect the amino acid. After several washes with dimethylformamide, the Fmoc protected amino acids Glu, Gly, and Phe (D or L-enantiomers) were subsequently added followed by deprotection with 20% piperidine and several washes between each addition of the amino acid. The peptide was capped with 2-naphthaleneacetic acid and cleaved from the resin with trifluoroacetic acid (TFA), creating the peptides shown in Scheme 1. The same general procedure was used for producing the hydrogelator peptides (1, 2, 3, and 4).

After the synthesis of the precursors, the gelation properties of the precursors with the addition of alkaline phosphatase (ALP) were evaluated. 1-P is included as the reference for comparison. As shown in Fig. 1A, 0.6 wt% of 1-P with the addition of ALP (10 U/mL) results in a transparent hydrogel of 1. Transmission electron microscopy (TEM) shows that the hydrogel of 1 contains entangled nanofibers with a diameter of 7 ± 2 nm. 2-P, being different from 1-P, is unable to result in a hydrogel of 2 after ALP (10 U/mL) is added into the solution of 0.6 wt% of 2-P (Fig. S2). TEM image shows that the solution of 2, however, contains nanofibers with a diameter of 10 ± 2 nm, confirming the self-assembly of 2 at the concentration of 0.6 wt%. The self-assembly of 2 in the solution suggests the increase of the concentration of 2 should lead to hydrogelation, which turns out to be the case: Doubling the concentration of 2-P (to 1.2 wt%) and decreasing the concentration of ALP to 5 U/mL affords a hydrogel containing nanofibers with a diameter of 8 ± 2 nm after EISA (Fig. 1B). This result indicates that both the concentrations and the dephosphorylation rate contribute to the production of 2 and the subsequent self-assembly for hydrogelation. Also, mixing a 1:1 ratio of 1-P and 2-P (0.6 wt%) with 10 U/mL ALP resulted in the formation of nanofibers before a weak hydrogel with small aggregates was observed (Fig. S3) about one week after the addition of ALP. As shown in Fig. 1C, 3 forms a transparent hydrogel containing nanofibers with a diameter of 7 ± 2 nm within 24 hours after the addition of ALP into the solution of 3-P (0.6 wt%). TEM also reveals that the nanofibers in the hydrogel of 3 form and entangle extensively. The addition of ALP in the solution of 4-P results in the hydrogel of 4 in 1 h. The hydrogel, however, undergo syneresis after 24 hours (Fig. 1D), suggesting strong interfibrillar hydrophobic interactions. TEM shows that the nanofibers in the hydrogel of 4 has a diameter of 9 ± 2 nm and exhibit a relatively low density. The low density of the nanofibers in the hydrogel of 4 at 0.6 wt% agrees with that the hydrogel of 4 is weaker that the hydrogel of 3. This observation indicates that the chirality of the amino acid residues in the heterochiral peptides dictates the self-assembly of the peptides, thus affecting the mechanical behaviour of the hydrogels made of the peptides.

Figure 1.

TEM images of the hydrogel (inset: optical images) formed by the addition of ALP to the solution of (A) 1-P (0.6 wt%; 10 U/ml ALP), (B) 2-P (1.2 wt%; 5 U/mL ALP), (C) 3-P (0.6 wt%; 10 U/mL ALP) or (D) 4-P (0.6 wt%; 10 U/mL ALP) in PBS buffer at pH = 7.4 (Scale bar = 100 nm).

To further explore this behaviour, rheological analysis was done on select gels. The strain and frequency dependence of dynamic storage modulus G’ and loss modulus G” of gels formed by 1-P and 3-P (Figure S4) show that G’ is more than G” which supports the strong hydrogels shown in the inset images of Figure 1. Figure S5 shows the instability of the hydrogel formed by the mixing of 1-P and 2-P. Figure S6 supports our earlier findings in which the increase in concentration of 2-P to 1.2 wt% contributes to more stable gel formation. Another important factor, as mentioned before, is the interactions between two peptides which can contribute to hydrogelation. Hydrogen bonding, aromatic interactions and the effect of chirality on enzyme interaction play different roles in the strength of resulting hydrogels. Figure S7 shows the difference in gel formation between 1-P and 4-P alone versus a mixture. From the data shown the addition of 1U/mL of enzyme is not enough to form a gel, but in the case of 4-P, the gelation point is evident and the mixture (Fig. S7C) starts to form a gel almost immediately after the addition of enzyme. This result indicates that mixing of two or more peptides may display certain properties that are not achievable with singular peptide solutions.

After the examination of the hydrogels formed by the enzymatic transformation of the precursors, the cell compatibilities of the precursors were examined by evaluating the cell viability of the HeLa cells incubated with the precursors. HeLa is chosen as a model of mammalian cells because its availability allows others to reproduce the study (if interested)._For instance, in Fig. 2A, all the precursors show little cytotoxicity to the HeLa cells in 48 h at the concentration of 500 μM. Interestingly, 4-P, contrasting to a naphthyl capped D-phosphotripeptide (Nap-D-Phe-D-Phe-D-_pTyr) that is cytotoxic at 400 μM,[14] is cell compatible. This result indicates that the sequence of the phosphopentapeptides also dedicates the cell compatibility of the peptide assemblies resulted from EISA. Moreover, precursor 2-P and 3-P show cell viability higher than 100 %, which means there is slight enhancement of cell proliferations. These results (Fig. 2A and Fig. S8) indicate that these four pairs of precursors/hydrogelators (Fig. S9) are cell compatible within 48 h. The inherent stability of the precursors, being important for their potential applications, is examined by treating them with proteinase K for 32 h (Fig. 2B). Among the four precursors, 4-P, which consists of all D-amino acid residues, is the most stable one: more than 93 % of 4-P remains after 32 h. The second most stable precursor is 2-P, which contains two D-Phe, and there are about 70% of 2-P remaining after 32 h. 1-P and 3-P are less stable. Both undergo complete proteolysis after being incubated with proteinase K for 12 h. Proteinase K preferentially hydrolyzes near aromatic and hydrophobic residues[43] and its activity is promoted by long N-terminal regions.[44] Compounds 1P and 3P from Scheme 1 contain l-phenylalanine amino acids which are hydrolyzed almost completely within the first 5 hours (Figure 2B), while 2P and 4P are relatively stable. Wu and colleagues saw a similar result when incubating proteinase K with d and l peptoid-peptide conjugates.[45] As mentioned in the introduction D-peptides are more resistant to peptidase cleavage than its L-enantiomers, lending evidence to the results shown above. The difference of the precursors in proteolysis catalysed by proteinase K suggests that these precursors, as stereoisomers, may also differ in their interactions with other enzymes, such as ALP. Thus, it is necessary to evaluate the effect of D-amino acids on the dephosphorylation rate of the precursors.

Figure 2.

(A) Cell viability of HeLa cells incubated with the precursors at the concentration of 500 μM for 48 h. (B) Digestion curves of precursors remaining upon treatment with proteinase K (3 U/mL) for 32 h. All precursors are at the concentration of 200 μM.

The dephosphorylation rates were evaluated by treating the precursors with ALP at 0.2 U/mL and analyzing the composition of the resulted mixtures. As shown in Fig. 3A, almost all of the precursor 1-P and 2-P can be dephosphorylated to the corresponding hydrogelator 1 and 2 by ALP in 24 hours. While dephosphorylation of 1-P and 2-P proceeds at the same rate, the hydrogelation of 2-P takes much longer time. This result indicates that, in this case, the self-assembly kinetics of the heterochiral peptides likely is slower than the homochiral peptides. About 11% of 3-P remained after 24 h incubation together with ALP (0.2 U/mL), and 29% of 4-P remained after 24 h treatment of ALP (0.2 U/mL). The dephosphorylation rates show difference in the first hour that more than 70% of 1-P or 2-P turns into 1 or 2, respectively, in 1h. About 60% of 3-P is converted to 3 while only 40% of 4-P is converted to 4 in the first hour. The dephosphorylation rate of 4-P is much slower than those of the other three precursors, as indicated by the initial rates of the dephosphorylation processes shown in Fig. 3B. Precursor 4-P shows the slowest initial rate and in comparison, 2-P has the fastest initial rate which is 1380 μM/h. The initial rates of precursor 1-P and 3-P are similar that both are around 1100 μM/h. These results indicate that the weaker hydrogel from the EISA of 2-P unlikely originates from incomplete or slow dephosphorylation, but more likely from the self-assembly of 2. On the contrary, the weaker hydrogel from the EISA of 4-P likely originates from incomplete dephosphorylation. Overall, the incomplete dephosphorylation in both figures may originate from the self-assembly of peptides to reduce the diffusion of enzymes, as observed previously.[13]

Figure 3.

(A) Increase of hydrogelators with time shows the dephosphorylation process of the precursors after incubation with ALP (0.2 U/mL) at 37 °C. The precursors dissolve in pH 7.4 PBS buffer at a concentration of 500 μM. (B) The calculated initial rate of the precursors with the addition of ALP (0.2 U/mL).

Conclusion

In summary, this study illustrates EISA as a versatile approach for generating supramolecular hydrogels[46] from several related and cell compatible stereoisomers of pentapeptide precursors,[47] thus providing insights to create smart biomaterials with tailored properties. The difference of the morphologies of the nanofibers imply the interaction of ALP, as a protein, with the hydrogelators, which agrees with the reported interactions between proteins and low molecular weight hydrogelators[48,49], Although all the four precursors/hydrogelators are cell compatible, the origin of their cell compatibility is different, thus highlighting the versatility of EISA and the important of molecular design. Although this study focuses on pentapeptides, the knowledge obtained here is useful for designing supramolecular hydrogels from longer peptides[50–54]. In the future, replacing Nap-FF with another assembling motif, or using the D-amino acid sequence with another cleavable functional group will yield more EISA generated supramolecular hydrogels for more broad applications, including cancer cytotoxicity or tissue engineering.