Phosphobisaromatic motifs enable rapid enzymatic self-assembly and hydrogelation of short peptides

Abstract

Enzyme-instructed self-assembly (EISA) and hydrogelation is a versatile approach for generating soft materials. Most of the substrates for alkaline phosphatase catalysed EISA utilizes phosphotyrosine (_pTyr) as the enzymatic trigger for EISA and hydrogelation. Here we show the first example of phosphonaphthyl (_pNP) and phosphobiphenyl (_pBP) motifs acting as faster enzymatic triggers than phosphotyrosine for EISA and hydrogelation. This work illustrates novel enzyme triggers for rapid enzymatic self-assembly and hydrogelation.

Phosphonaphthyl (_pNP) or phosphobiphenyl (_pBP), acting as both the N-terminal capping group and and the enzymatic trigger, of short peptides results in alkaline phosphatase (ALP)-catalyzed gelation (almost) instantly.

Graphical Abstract:

Enzyme-instructed self-assembly (EISA) or enzymatic noncovalent synthesis,¹ as a versatile approach for mimicking the regulation of noncovalent interactions of biomolecules in a living cell, has emerged as a useful bottom-up strategy for controlling functional supramolecular peptide assemblies,² which promise a wide range of potential applications of soft materials in biomedicine, such as tissue engineering,^3–4 molecular imaging,^5–7 drug delivery,^8–10 multimolecular crowding in biosystems,¹¹ and cancer therapy.^12–16 Because enzymatic reactions provide a fast and specific transformation of supramolecular peptide assemblies, EISA is particularly attractive for generating non-diffusive supramolecular assemblies^17–19 in cellular environment²⁰ for modulating cellular activities,²¹ such as apoptosis,²² morphogenesis,²³ and protein trafficking.^24–25 Most of these studies related to short peptides utilize phosphotyrosine (_pTyr) as an enzymatic trigger²⁶ being activated by alkaline phosphatase (ALP) for initiating self-assembly of the peptides because ALP plays important roles in cell biology and is overexpressed in certain tumours.^27–28 Particularly, most of these peptides contain an aromatic capping motif, such as naphthyl,²⁹ fluorenyl,³⁰ or pyrenyl³¹ group, at the N-terminal and L- or D-_pTyr at the C-terminal or in the middle of the peptides.^31–34 A representative example of EISA substrates is Nap-ff_py (1P), which carries a D-_pTyr (_py) at the C-terminal of the peptide. 1P has revealed many key features of EISA and led to unexpected formation of pericellular nanofibers that selectively kill cancer cells.³⁵ On the other hand, the phosphate trigger at N-terminal of peptide is much less explored, except the work of Ye et al. that employs a phosphorylated dye at the N-terminal of a peptide,⁷ but the rate of dephosphorylation is relatively slow.

Therefore, we decide to examine the dephosphorylation of N-terminal aromatic capping motif of short peptides for enzymatic self-assembly and hydrogelation. As shown in Scheme 1, we attach phosphohydroxybenzoic acid (_pB), phosphohydroxynaphthoic acid (_pNP), or phosphohydroxybiphenyl-carboxylic acid (_pBP) at the N-terminal of D-diphenylalanine (ff)³⁶ or D-tri-phenylalanine peptide (fff) to generate phosphorylated peptide derivatives (2-7P) as the substrates of ALP. Our results show that while the ff derivatives are unable to form hydrogels, the fff derivatives result in hydrogels after ALP-catalysed dephosphorylation converting the nanoparticles made of the precursors to the nanofibers consisted of the corresponding hydrogelators (i.e., 3, 5, or 7). Rheological evaluation shows that the resulting three hydrogels have relatively high storage moduli, up to 10⁴ Pa, when the concentrations of the hydrogelators are about 8 mM (about 0.5 wt%). Moreover, _pBP and _pNP act as faster enzyme triggers than _py and _pB motif for hydrogelation. As the first example to show the dephosphorylation of _pBP and _pNP for rapid enzymatic self-assembly and hydrogelation, this work offers a novel molecular platform and identifies fast triggers for EISA catalysed by ALP.

Scheme 1.

Structures of the peptide derivatives with phosphate at the C-terminal (previously explored) and N-terminal (this work).

Scheme 1 shows the structures of the designed substrates of ALP. Based on the structure of 1P, we use _pB, _pNP or _pBP as the enzymatic trigger of ALP to replace _py in 1P. We also move the ALP trigger to the N-terminal of the peptides. That is, _pB, _pNP or _pBP acts as the N-terminal capping group for ff or fff. Such a combination leads to six substrates of ALP: _pB-ff (2P), _pB-fff (3P), _pNP-ff (4P), _pNP-fff (5P), _pBP-ff (6P), and _pBP-fff (7P). After being dephosphorylated by ALP, these substrates would result in six peptide derivatives: B-ff (2), B-fff (3), NP-ff (4), NP-fff (5), BP-ff (6), and BP-fff (7). According to this design, we first produce _pB, _pNP and _pBP according to previously reported procedures.³⁷ Then, we, using Fmoc-based solid-phase peptide synthesis,³⁸ synthesize the six substrates of ALP.

After obtaining the precursors, we examine enzymatic gelation of these substrates upon the addition of ALP. Each of precursor dissolves in PBS buffer to form a clear solution with the concentrations of 0.5 wt%. As shown in Figure 1, enzymatic dephosphorylation of 3P, 5P or 7P results in a hydrogel 24 h after adding ALP. While the dephosphorylation of 2P and 4P affords a solution, the dephosphorylation of 6P results in a suspension. This result indicates that 2 or 4 is more water soluble than 6, agreeing with the higher hydrophobicity of biphenyl (logP = 3.71) than those of naphthyl (logP = 3.03) and phenyl (logP = 2.03) groups. These results also suggest that the tri-phenylalanine enhances intermolecular interactions to favour hydrogelation of 3, 5, or 7.

Fig. 1.

The optical images of 2P, 3P, 4P, 5P, 6P and 7P (0.5 wt % in PBS buffer, pH7.4) before and after incubation with ALP (1 UmL⁻¹) for 24 h.

We use dynamic time sweep to characterize the rheological properties of the hydrogels resulted from EISA of 3P, 5P, and 7P by measuring their storage and loss moduli (G’ and G”) (Fig. 2). We make the solutions of 3P, 5P, and 7P with the concentrations of 8 mM (about 0.5 wt%). After the addition of ALP to the solution of 3P, a crossover of G’ and G” occurs at about 1 h or 11 h of incubation when the concentration of ALP is 1.0 or 0.1 UmL⁻¹, respectively. With the treatment of 1.0 UmL⁻¹ or 0.1 UmL⁻¹ ALP, the solution of 5P shows the crossover of G’ and G” around 2 minutes or at 20 minutes, respectively. Being incubated with ALP at 1.0 or 0.1 UmL⁻¹, the solution of 7P exhibits the crossover of G’ and G” less than one minutes or at about 13 minutes, respectively. These results indicate that _pBP or _pNP, as an enzyme trigger, enables enzymatic hydrogelation about 50–60 times faster than _pB does. Moreover, the times of 5P and 7P to reach the gelation point are two and three times shorter than that of 1P (Fig. S1), respectively. This result confirms that _pBP and _pNP are faster enzyme triggers for EISA than _py does. After 12 hours, the G’ values of the hydrogels by EISA of 3P, 5P, and 7P reach their plateau values when the concentration of ALP is 1.0 UmL⁻¹.

Fig. 2.

Dynamic time sweeps of 3P (8 mM), 5P (8 mM), and 7P (8 mM) incubated with ALP at 1 and 0.1 UmL⁻¹ and at a strain of 1% and frequency of 6.28 rads⁻¹.

We also conduct frequency and strain sweeps (Fig. 3) of the hydrogels. The strain and frequency applied in time sweep fall within the linear viscoelastic range of the gels, indicating that the time-dependent strain sweeps are carried out on appropriate conditions. Being incubated with 1 UmL⁻¹ and 0.1 UmL⁻¹ of ALP, the hydrogels, being made of 3, 5, or 7, obtained by the dephosphorylation of 3P, 5P, or 7P, exhibit frequency-independent G’ (0.1 rads⁻¹ to 200 rads⁻¹), suggesting that gels behave solid-like. For the gel resulted from dephosphorylation of 3P by 1 UmL⁻¹ of ALP, G’ and G” are independent of strain below 1% and show the existence of linear viscoelastic region (LVR). Within the LVR, G′ (up to 10⁴ Pa) is significantly greater than G′′, reflecting their dominant elastic nature. For the gel resulted from dephosphorylation of 3P by 0.1 UmL⁻¹ of ALP, though G’> G” below 1% strain, G” fluctuates and increases with the increase of strain, which fails to show LVR and indicates the hydrogel being relatively weak. Unlike the case of 3P, the strain sweeps of the hydrogels resulted from dephosphorylation of 5P or 7P by 1 or 0.1 UmL⁻¹ of ALP show that G’ and G” are independent of strain below 2%. Both show the existence of linear viscoelastic region (LVR), suggesting that NP or BP enhances the viscoelasticity of the hydrogels made of 5 or 7.

Fig. 3.

(A) Frequency sweeps of 3P (8 mM), 5P (8mM), and 7P (8 mM) conducted after 24 h incubation with ALP at 1.0 and 0.1 UmL⁻¹ and at the strain of 1%. (B) Dynamic strain sweeps of 3P (8mM), 5P (8mM), and 7P (8mM) conducted after 24 h incubation with ALP at 1.0 and 0.1 UmL⁻¹ and at the frequency of 6.28 rads⁻¹.

To investigate the morphological properties, we use the transmission electron microscopy (TEM) to image these precursors without or with the addition of ALP. As shown in Fig. 4, TEM of the solutions of 2P, 4P, 6P, and 7P show aggregated nanoparticles, with the diameters about 10 nm. While the TEM of the solution of 3P reveals the existence of short nanofibers with the diameters of 4 nm, TEM of the solutions of 5P show the coexistence of short nanofibers with diameters of 4 nm and nanoparticles. After the addition of 0.1 UmL⁻¹ of ALP into the solutions of 2P and 6P for 24 h, the resulting solutions of 2 and the suspension of 6 contain the nanoparticles with the diameters around 12 nm. The resulting solution of 4 showed coexistence of nanoparticles and nanosheets. The hydrogels of 3 show extended and entangled nanofibers with the diameters of 4 nm, and some of the nanofibers form bundles with the diameter of 14 nm; the hydrogel of 5 shows uniformed nanofibers with the width of 8 nm; and the hydrogel of 7 shows uniformed bundles with a diameter of 13 ± 2 nm. The formation of the nanofibers of 3, 5, or 7 likely contributes to the formation of the hydrogel of 3, 5, or 7. Notably, while the scanning electron microscopic (SEM) image of the dried gel of 3 display nanofiber networks, the dried gel of 7 is largely amorphous with a few thick fibres (Fig. S2), agreeing with that the high hydrophobicity of biphenyl group significantly enhances the intermolecular interactions of 7.

Fig. 4.

The TEM images of 2P, 3P, 4P, 5P, 6P and 7P at 8 mM before and after ALP treatment. The concentration of ALP is 0.1 UmL⁻¹. The duration time is 24 h. The scale bar is 100 nm.

In summary, we design six short peptides containing phosphoaromatic both the capping groups and the enzyme trigger at the N-terminal as novel ALP substrates for EISA and hydrogelation. The ability to form the hydrogels indicate that the tripeptide backbone having aromatic groups (i.e., Phe and/or Tyr) enhances self-assembly and leads to hydrogelation. The result that _pBP is a faster substrate than _pTyr for ALP agrees with the report that _pBP is a faster substrate than _pTyr for protein tyrosine phosphatase 1 (PTP1B).³⁹ The rates of the enzymatic hydrogelation catalysed by ALP follows the trend of 7P > 5P > 1P > 3P, implying that distancing the phosphate trigger away from the peptide backbone likely favours fast enzymatic self-assembly, a design principle that may help combine EISA with other self-assembling molecules.^40–49 This work also suggests that it is worthwhile to examine other phosphobisaromatic capping groups for exploring the anticancer drug candidates that act via EISA catalysed by phosphatases.