This communication reports that conjugating simple diglycine-activated succinic esters to self-assembling molecules or drugs results in autohydrolysis to enable instant cellular uptake of the supramolecular assemblies or drugs. Peptides are emerging as an important class of building blocks to generate supramolecular assemblies, [1] and the conjugation of peptides to drugs has received increasing attention for designing molecules to deliver drugs into cells. [1i,2] In a recent study that aimed at enzymatic hydrolysis of succinic ester prodrugs, we conjugated dipeptides to chloramphenicol succinate to generate enzymatically controlled assemblies. We found that conjugating diglycine to succinate lowers the cytotoxicity of chloramphenicol to bone marrow stromal cells while maintaining the activity of the conjugates against Gram-negative bacteria. [3] In our investigation into the alternative amino acids to glycine, we discovered a significant decrease in the antibacterial activity of other chloramphenicol conjugates, particularly when glycine was not directly attached to the succinic acid. [4] This finding challenges the assumption that esterases, which are typically nonselective, catalyze the hydrolysis of the ester bond in the conjugates.

The above question stimulates us to further investigate the uniqueness of incorporating glycine in succinic ester. We found that conjugating diglycine motif (GG) to hemisuccinate leads to the rapid autohydrolysis of the succinate ester under physiological pH, such as in phosphate-buffered saline (PBS). Removing GG or replacing GG with mono-glycine (G), dialanine (AA), or di-β-alanine (BABA) motif abolishes the hydrolysis in PBS. When the compound bears selfassembly motifs, the hydrolysis results in a morphological transformation from nanofibers to nanoparticles, and thus enhancing cellular uptake of the peptide assemblies. The generality of autohydrolysis was supported by pyrenyl (Py) and biphenyl (BP) conjugated diglycine succinic ester. When replacing the self-assembling motifs with drugs, conjugating GG to succinate prodrugs recover the efficacy of the drugs (Figure 1A). 2D NMR of the diglycine-activated succinic esters s indicates a "U-turn" of diglycine, suggesting that intramolecular base-catalyzed hydrolysis promotes the autohydrolysis of succinic ester and subsequently boosts cell uptake of the products. This work represents a unique example of rapid autohydrolysis of esters for instant cellular uptake. This work also illustrates the modulation of the rate of nonenzymatic bond breaking for intracellular targeting.

Figure 1B illustrates the rationale for molecular design. A fluorescent probe, nitrobenzoxadiazole (NBD), was conjugated to a rigid-rod motif, BP, to generate NBD-BP as a self-assembling motif because it is easy to image NBD-BP assemblies in cells. [5] Subsequent conjugation of a succinic acid to NBD-BP generates the molecule NBD-BP-suc ( ). To examine the impact of glycine on the rate of hydrolysis, we added one or two glycine to the C-terminus of , generating NBD-BP-G (2) or NBD-BP-GG (3), respectively. The hydrolysis of 3 (Figure 1C) leads to NBD-BP-OH (4). To investigate the roles of GG in autohydrolysis, we replaced it with BABA, which bears an extra methylene group in the amino acids, or AA, which has an additional branched methyl side chain, to produce NBD-BP-BABA (5) or NBD-BP-AA (6), respectively. To investigate the generality of the diglycine-promoted autohydrolysis, we introduced other self-assembling motifs or drugs. Specifically, substituting NBD-BP with Py yields Py-suc (7) and Py-GG (8), while removing the NBD motif leads to BP-suc (9) and BP-GG ( 0). Moreover, we selected two representative anticancer drugs, paclitaxel (Taxol) and camptothecin (CPT), both of which are hydrophobic and contain hydroxy groups for making succinate prodrugs. Thus, we synthesized Taxol-suc ( ), Taxol-GG ( 2), CPT-suc ( 3) and CPT-GG ( 4) for evaluating the recovery of drug efficacy in the absence or presence of diglycine conjugation. According to the results of 2D NMR (see below), we substituted the carboxy-terminal glycine in 3 with sarcosine to generate NBD-BP-GSAR ( 5) that bears a tertiary amide. We also introduced O-methylation of the carboxylic acid of 3 to generate NBD-BP-GG-OMe ( 6). 5 and 6 act as the controls to eliminate the intramolecular hydrogen bonding at two possible sites in 3. Combining liquid-phase and solidphase synthesis, we synthesized the aforementioned compounds and used liquid chromatography-mass spectrometry to confirm their identities (Scheme S1 and Figure S1-S17).

These peptide conjugates, prior to hydrolysis, comprise a hydrophobic capping motif and a hydrophilic carboxyterminus. The hydrophilic carboxy-terminus may form disordered peripheries [6] to allow hydrolysis. Given that hydrolysis results in the conversion of the molecule's hydrophilic region from carboxylic acid to hydroxy groups, we hypothesized that this transformation would impact the molecule's self-assembly behavior. Therefore, we examined the morphology of SA motif-capped -4 using TEM. The results revealed that -3 form nanofibers, while 4 generates amorphous aggregates in PBS (Figure 2A). Based on these findings, we hypothesized that the hydrolysis of -3 to 4 would lead to morphological transformation from nanofibers to aggregates. This hypothesis is supported by the 96 hincubation TEM results of 3 in PBS (Figure S18). We then sought to investigate the rate of hydrolysis with glycine conjugation. Bearing one glycine, the hydrolysis of 2 is pH sensitive. While 2 hardly hydrolyzes in water (pH = 6.0), the hydrolysis initiates at pH 7.4 in PBS, and accelerates in Minimum Essential Medium (MEM) (pH = 7.6) and saturated sodium bicarbonate (NaHCO 3 ) (pH = 8.5), reaching 90 % hydrolysis after 48 h (Figure 2B and Figure S19). With diglycine conjugation, the hydrolysis of 3 remains pH sensitive but occurs at much faster rates. While 3 remains intact in water, it hydrolyzes quickly in PBS, reaching 60 % conversion in 48 h (Figure 2C). 3 hydrolyzes faster in MEM and NaHCO 3 , approaching complete hydrolysis in 8 and 2 h, respectively (Figure S20). Diluted samples of 3 hydrolyze faster than the concentrated ones, indicating that selfassembly slows down the hydrolysis (Figure S21). Moreover, 3 hardly hydrolyzes in acidic conditions after 7 days (Figure S22). Therefore, 3 likely undergoes base-catalyzed autohydrolysis in aqueous solutions.

Removing the glycine conjugation, however, abolishes the hydrolysis. , 7, 9, , and 3 do not hydrolyze in PBS (Figure 2D). Further investigation into the hydrolysis of in MEM and NaHCO 3 shows no conversion after 48 h (Figure S23). We then determined the hydrolysis of other analogs. Elongation of the molecule (5) renders more flexibility, while introducing amino acid side chains ( 6) makes the molecule more rigid. Both 5 and 6 remain unhydrolyzed in PBS (Figure 2E and Figure S24), indicating that the hydrolysis relies on the unique conformation of diglycine. However, the hydrolysis is less sensitive to the hydrophobic part of the molecule, where Py (8), BP ( 0), Taxol ( 2), and CPT ( 4) capped diglycine conjugates all readily hydrolyze in PBS (Figure 2E and Figure S24).

Based on the increasing speed of hydrolysis of -3 at cell-free conditions, we sought to evaluate their cellular uptake. Confocal images show that hardly enters cells, while 2 and 3 are readily taken up by the cells, with more intense intracellular fluorescence for 3 (Figure 3A). 4 also quickly enters cells with similar cellular distribution to that of 2 and 3 (Figure S25). Based on these observations, we hypothesized that the autohydrolysis happens extracellularly to initiate the morphological transformation of 2-3 from nanofibers to aggregates. Furthermore, the charge removal of the molecule after hydrolysis facilitates its entry into the cells. Additionally, the cellular uptake of 2-3 is nonselective across the tested cells, agreeing with the autohydrolysis in physiological buffers under cell-free conditions. Quantification of the cellular uptake over 20 minutes reveals similar results, where 3 instantly enters cells and the difference between analogs is more significant than that between cell lines (Figure 3B). While 5 and 6 also enter cells, their rates of entry are lower than that of 3, corresponding to their reduced autohydrolysis (Figure 3C).

Considering that succinic esterification is a common prodrug strategy to enhance the water solubility of hydrophobic drugs, we replaced the self-assembling motifs with anticancer drugs and evaluated the efficacy of the diglycine prodrugs. Due to the inertness of the ester bonds in the conjugate of succinate and drugs, the resulting prodrugs often suffer from incomplete bioconversion at the therapeutic targets and undergo rapid excretion from bodies before exerting intended therapeutic effects, therefore loss of efficacy. Therefore, we hypothesized that the hydrolysis of diglycine conjugates, which is independent to esterase and hydrolyzes sufficiently fast, would recover drug efficacy. 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) assays and half-maximal inhibitory concentration (IC 50 ) determination of Taxol and CPT derivatives support this hypothesis. The efficacies of succinate prodrugs and 3 are compromised, while diglycine conjugation to generate prodrug 2 and 4 recovers the efficacy in both HeLa and Saos-2 cell lines (Figure 3D-E and Figure S26). Using succinate diglycine conjugation, the recovery of Taxol and CPT provides insights into prodrug strategy to retain efficacy at the same time increase water solubility.

After characterizing the analogs mentioned above, we aimed to investigate the hydrolysis mechanism. For this purpose, we used 1 H-1 H Nuclear Overhauser Effect Spectroscopy (NOESY) to compare the spatial proximity within the molecules of compounds 3, 5, and 6. Following the peak assignments based on their 1 H-NMR spectra (Figure S27), we utilized Correlation Spectroscopy (COSY) to investigate the through-bond correlations (Figure S28), and NOESY to determine the through-space correlations. The NOESY signals were assigned by excluding the COSY signals (Figure 4A and Figure S29-S34). Based on the chemical structures, the observed NOE signals suggest a possible conformation of 3, 5, and 6 (Figure 4B): The carboxy-terminus in compound 3 bends over (or a "U-turn"), enabling proximity between H p and H l , thereby implying the possibility of intramolecular base-catalyzed ester hydrolysis. However, no NOE signal was observed near the ester bond in 5 and 6, indicating that these molecules adopt different conformations that disfavor intramolecular hydrolysis, which corresponds to their inability to undergo hydrolysis in PBS (Figure 2E) and reduced cellular uptake (Figure 3C).

Based on the molecular bending observed in 3, we investigated the importance of two carboxy-terminal nucleophiles: the secondary amide and carboxylic acid derived from the terminal glycine. To evaluate their contribution to hydrolysis, we synthesized compounds 5 and 6 by methylating the amide and methyl esterifying the carboxylic acid, respectively. The results show that 5 hardly hydrolyzes in PBS, similar to 2, while 6 readily hydrolyzes in PBS, achieving nearly 90 % conversion after 48 h, similar to 3 (Figure 4C). Both compounds are able to undergo hydrolysis in NaHCO 3 (Figure S35). The hydrolysis tendency correlates with cellular uptake, where 6 enters cells more quickly than 5 (Figure 4D and Figure S36). These results suggest the significance of the amide group in the carboxy-terminal diglycine motif in facilitating ester hydrolysis and provide a strategy to modulate the rate of autohydrolysis.

In conclusion, we have demonstrated that the conjugation of diglycine to a succinate enhances the hydrolysis of ester at physiological pH (in PBS). By employing this strategy to conjugate succinic diglycine with a self-assembling motif, cellular uptake can be significantly boosted. Additionally, conjugation with hydrophobic drugs restores their efficacy, highlighting the generality of this approach. Besides, this prodrug design strategy may also be applied to the pharmacophores with cell selectivity (e.g. 4-Hydroxytamoxifen), thereby minimizing off-target cytotoxicity. The hydrolysis of the analogs underscores the importance of the carboxy-terminal amide of the diglycine motif in promoting hydrolysis. This work emphasizes that modulating rate of nonenzymatic reactions through molecular design is an alternative means, in comparison to enzymatic reactions [7] or pH response [8] , to boost cellular uptake. Overall, this enzyme-independent, mild base-catalyzed ester hydrolysis approach reveals the broad potential of short peptides in prodrug design [2a,9] , drug delivery, [1i,2b,10] and intracellular targeting [2c,d, 11] .