Enzyme Responsive Peptide Thioesters for Targeting Golgi Apparatus

Abstract

Golgi apparatus (GA) is the hub of intracellular trafficking, but selectively targeting GA remains a challenge. We show an unconventional type of peptide thioesters, consisting of an aminoethyl thioester and being substrates of thioesterases, for instantly targeting GA of cells. The peptide thioesters, above or below their critical micelle concentrations, enter cells mainly via caveolin-mediated endocytosis or macropinocytosis, respectively. After being hydrolyzed by GA-associated thioesterases, the resulting thiopeptides form dimers and accumulate in the GA. After saturating GA, the thiopeptides enrich in endoplasmic reticulum (ER). Their buildup in ER and GA disrupts protein trafficking, thus lead to cell death via multiple pathways. The peptide thioesters target Golgi of wide variety of cells, including human, murine, and Drosophila cells. Changing D-diphenylalanine to L-diphenylalanine in the peptide maintains the GA-targeting ability. In addition, targeting GA redirects protein (e.g., NRAS) distribution. This work illustrates a thioesterase responsive and redox active molecular platform for targeting GA and controlling cell fates.

Graphical Abstract

Golgi apparatus (GA)^1-2 is an important hub for different signaling pathways.^3-5 Being an essential component of the secretory pathway used by all eukaryotic cells to distribute membrane and secretory proteins,⁶ GA is emerging as important target for understanding and treating illnesses such as cancer⁷ and Alzheimer’s diseases⁸. The increasing understanding of the protein trafficking to and from Golgi, such as the sophisticated delivery of proteins from endoplasmic reticulum (ER) to GA,⁹ also highlights a unique opportunity to develop molecules to target GA for modulating cell functions. Although there are several existing molecular probes for imaging GA,^10-11 they still require pretreatment,¹⁰ take relatively long incubation time,^11-12 or only response to certain cells.¹¹ Thus, there is still an unmet need for developing molecules for targeting GA.

Our recent work of thiophosphopeptides instantly targeting GA and selectively killing cancer cells¹³ reveals two important insights: (i) enzymatic reaction enables the instant and efficient GA targeting; and (ii) redox reaction enables proper thiopeptides to accumulate in GA of cells by self-assembly, dimerization, and the plausible reaction with cysteine rich proteins (CRPs). Replacing thiophosphate group with other sulfur containing moieties and, we unexpectedly found that fluorescent peptide thioesters instantly target GA of a wide variety of cells. Consisting of an aminoethyl thioester as the enzyme responsive cap at a terminal of self-assembling D-diphenylalanine (ff)¹⁴ and nitrobenzoxadiazole (NBD) as the fluorophore at the other end of ff, the peptide thioesters (1 and 3, Scheme 1) fail undergoing S-to-N acyl shift¹⁵ of the conventional peptide thioesters used for protein synthesis.^16-18 The peptide thioester (1 or 3), above or below its critical micelle concentrations (CMC), enters cells mainly via caveolin-mediated endocytosis or macropinocytosis, respectively (Scheme 1). After being hydrolyzed by Golgi-associated thioesterases, such as PPT1,¹⁹ LYPLA1,²⁰ or LYPLA2,²¹ the resulting thiopeptides (2 or 4) (Schemes 1 and S4) form dimers and likely react with cysteine-rich proteins in GA, thus accumulating in the GA. After saturating GA, the peptide thiols enrich in ER. At about 5-10 times of the concentrations used for imaging GA, the peptide thioesters buildup in ER and GA, result in ER-stress, and disrupt protein trafficking (e.g., NRAS trafficking), thus resulting cell death via multiple pathways. While changing ff to L-diphenylalanine (FF) maintains the GA-targeting ability of the peptide thioester (5), additional controls (7-13, Scheme S5), such as replacing the thioester by methylsulfonyl group or thioether, switching the thioester to carboxylester or amide, or mutating the D-phenylalanine to D-alanine, fails to target GA. 1 instantly targets Golgi of a variety of cells, including human, murine, Drosophila cells. In addition, 1 redirects the distribution of NRAS. This work illustrates the unconventional type of self-assembling peptide thioesters as a new and facile molecular platform for targeting GA and controlling cell fates via enzymatic activation and redox reactions in cells.

Scheme 1.

Enzyme responsive peptide thioesters targeting GAs.

1 instantly enters the cells, and the fluorescence resulted from 1 overlaps with Golgi-RFP (Figure 1A). Fluorescence appears in cytosolic region and at GA immediately after the addition of 1 and GA becomes distinguishable in less than 1 min (Figure S1, S2, video S1). Single cell imaging analysis (Figure 1B) confirms faster fluorescence increase inside than outside GA: the fluorescence plateaus in less than 15 min or longer than 45 min at GA or outside GA, respectively; the plateaus of the fluorescence are higher at GA than outside GA. LC/MS analysis of the lysate of the HeLa cells incubated with 1 for 24 h reveals that about 60% of the peptide thioester (1) turns into the corresponding thiol (2) and around 55% of 2 forms a dimer (Figure S3). The dimers, resulted from intracellular oxidants to oxidize 2, enhance the accumulation of 2 at GA. These results indicate that 1, being deacetylated to give 2, accumulates primarily at GA.

Figure 1.

(A) CLSM of Golgi-RFP transfected HeLa treated with 1 (10 μM, 30 min); (B) Single cell analysis of fluorescence inside and outside GA of HeLa treated with 1 (10 μM, 60 min); (C) CMC of 1; (D) CLSM of HeLa treated with 1 (10 μM or 2 μM) for 30 min. Bar = 20 μm.

1 enters HeLa cells differently below and above its critical micelle concentration (CMC = 2.1 μM, Figure 1C). mβCD,²² a caveolin-mediated endocytosis (CME) inhibitor, significantly slows down the fluorescence build-up at GA of the cell treated with 10 μM of 1; CytD,²³ a macropinocytosis inhibitor, dramatically decreases the fluorescence at GA of the cells incubated with 2 μM of 1 (Figure S4). Single cell analysis (Figure S5, S6, S7) confirms that mβCD is more potent to inhibit the GA accumulation than CytD when HeLa cells being treated with 10 μM of 1, while CytD is more potent in the case of 2 μM of 1. Treating the HeLa with chlorpromazine suggests that 1 hardly or only slightly depends on clathrin-mediated endocytosis (Figure S8) at 2 and 10 μM, respectively. These results suggest that the assemblies and monomers of 1 enter cells mainly via CME and micropinocytosis, respectively.

Thioesterase inhibitors, ML211^24-26 or DC661²⁷, by itself or in combination, efficiently block 1 accumulating at GA (Figures S4, S9, and S10). Meanwhile, both inhibitors cannot block the GA accumulation of commercial Golgi dye, C6-NBD-Ceramide²⁸ (Figures S11 and S12). Although carboxylesterase (CES1) is able to catalyze the hydrolysis of 1 in a cell free assay (Figure S13), pretreating HeLa cells with a nonspecific esterase inhibitor bis(p-nitrophenyl) phosphate²⁹ hardly decrease the rate of fluorescence increase from 1 at GA (Figure S14). These results suggest that thioesterases (i.e., LYPLA1, LYPLA2, and PPT1) contribute to the deacetylation of 1 to form 2 and act as main control for 1 to target GA.

Fluorescence appears instantly inside cells in less than 1 minute after adding 1 in the culture of HeLa cells (Figure 2A). The fluorescence at GA increases significantly with the time of incubation of 1, and GA can be easily distinguished from the background in less than 4 minutes (Figure 2A). There are at least 3 times enhancement of fluorescence intensity at GA from 1 minute to 8 minutes (Figure S15). The fluorescence outside GA increases significantly after 16 min, likely due to the retrograde Golgi to ER trafficking.³⁰ While C6-NBD-ceramide²⁸ shows dim fluorescence over 16 minutes (Figure S16), 1 exhibits a superior ability for targeting GA . Unlike the thiophosphopeptide,¹³ 1 is able to target GA of various cells from different organisms (Figure 2B), such as Homo sapiens (HeLa, Saos-2, HEK293 , MCF-7, HS-5, SJSA-1 and HepG2), Mus musculus (B16F10), and Drosophila (S2). Single cell analysis of fluorescence at GA (Figure S17) and CLSM images (Figures S18-S23) of different cell lines confirm that 1 accumulates at the GA of these cells swiftly. 1 also rapidly enters primary cells, such as neutrophils (polymorphonuclear leukocytes), and localizes at GA (Figure S24). These results establish that 1 is able to target GA of a wide variety of cells, likely due to that thioesterases are essential and ubiquitous.³¹

Figure 2.

CLSM images of (A) HeLa treated with 1 for 1, 4, 8 and 16 min; (B) different cells (Saos-2, HEK293, MCF-7, HS-5, SJSA-1, HepG2, B16F10, S2) treated with 1 for 8 min. [1] = 10 μM, Bar = 20 μm.

At above 20 μM, 1 significantly inhibits HeLa, HEK293, Saos-2, MCF-7 and HS-5 cells (Figure 3A) with the IC₅₀ values of 15.0, 10.9, 9.5, 14.5 and 14.9 μM, respectively. Several commonly used inhibitors (Z-VAD-FMK, NAc, Nec-1, DFO, disulfiram and PD150606)³² of cell death hardly rescue these four cell lines (Figure S25), indicating that 1 results in cell death via multiple pathways or possibly a new mechanism. 1 shows slightly mild cytotoxicity against SJSA-1 cells (IC₅₀ = 22.5 μM) and exhibits rather low inhibitory activity against HepG2 cells, with the IC₅₀ values above 100 μM (Figure 3B). These results coincide with the high level of glutathione (GSH) in hepatocytes³³ and in SJSA-1³⁴ as GSH compromises the cytotoxicity of the redox-active 2. Depletion of GSH in HepG2 cells by L-buthionine-sulfoximine (BSO)³⁵ significantly increase of cytotoxicity of 1 against HepG2 cells (IC₅₀ drops to 13.45 μM, Figure S26), which supports that GSH antagonizes the peptide thioesters.

Figure 3.

(A) Cytotoxicity and (B) IC₅₀ of 1 against different cell lines (HeLa, HEK293, Saos-2, MCF-7, HS-5, SJSA-1, HepG2); (C) ICC staining of MCF-7 with NRAS antibody after the pretreatment with or without 1 (10 μM, 30 min). Bar = 20 μm.

Because the trafficking of oncogenic NRAS between GA and plasma membrane is essential for tumor growth,³⁶ we determined the intracellular distribution of NRAS of MCF-7 cells. NRAS, being overexpressed in MCF-7,³⁷ relocates from mainly plasma membrane to overwhelmingly GA after the pretreatment of 1 for 30 minutes (Figure 3C left, middle). No fluorescence in the NBD channel (green, Figure 3C) after the standard immunocytochemistry staining excludes that the re-distribution of NRAS is due to bleed-through. As S-palmitoylation is a critical process for protein trafficking,³⁸ this result suggests that 1 disrupts GA functions and interferes with the secretory pathway, which likely contributes to its cytotoxicity when the accumulation of 1 at GA reaches certain threshold.

3, behaving similarly as 1, enters cells in a concentration-dependent manner (Figures S27, S28, and S29) and accumulates at GA swiftly (Figure S30A, Video S2). Thioesterase inhibitors slow the GA accumulation of 3 (Figures S31, S32). The cytotoxicity of 3 is slightly lower than that of 1, but with the similar trend (Figure S30B), indicating that (i) acetyl thioester is enzymatic responsive on either C- or N-terminal of the peptide thioester and (ii) 1 and 3 likely enter and accumulate at GA via similar mechanisms. 5, having FF instead of ff, accumulates at GA with the similar rate as 1 does (Figures S30A, S33, Video S3). These results suggest that the cellular uptake of 1, 3, or 5 and the rapid GA accumulation of 2, 4, or 6 depend rather on the enzyme response of the thioester group than on the configurations of them. The cytotoxicity of 5 is much lower than that of 1 (Figure S30B), agreeing with the proteolysis of 5. Replacing the thioester with methylsulfonyl group or methyl thioether, respectively (Scheme S5), and resists enzymatic cleavage (Figure S34), 7 or 8 fails to accumulate at GA (Figures S35, S36, and S37). The analogues replace thioester with ester (9) or amide (10) (Scheme S5) are unable to accumulate at GA (Figures S35, S38, and S39). Although cells cleave the ester bond of 9 (Figure S40A), there is little fluorescence at GA, confirming that redox active thiol is crucial for GA targeting. With D-dialanine replacing diphenylalanine, 11 (Scheme S5), though being enzymatically responsive (Figure S40B), hardly enters cells and accumulates at GA (Figures S35 and S41). Omitting ff motif from 1, 12 (Scheme S5) exhibits the similar intracellular behavior as 11, not targeting GA (Figures S35, S41, and S42). TEM shows that 11 hardly self-assembles while 1 self-assembles into nanoparticles and nanofibers at 20 μM (Figure S43). These results indicate that self-assembly is essential for the peptide thioesters to target GA. 13, with thioester on a cysteine residue (Scheme S5) as the enzymatically responsive thioester bond (Figure S40C), fails targeting GA (Figures S35 and S44). 7-13 exhibits low cytotoxicity against HeLa cells (Figure S30C), agreeing with their inability to target GA.

In summary, this work illustrates that several peptide thioesters undergo enzymatic hydrolysis and instantly accumulate at GA. Although 2 also accumulates at GA slowly,¹³ the thioester group, acting as a pre-form of thiol and one type of high energy bond in vivo,³⁹ is a unique and useful building block for enzymatic response. This instant GA targeting process not only allows the imaging of ER and GA trafficking,⁴⁰ but also provides new insights for developing supramolecular assemblies^41-44 for disrupting protein trafficking and further controls cell fates. Although the detailed mechanism remains to be further elucidated, dimerization of the thiopeptides and the likely reaction with CRPs in GA⁴⁵ contributes to the GA accumulation, providing a new way for modulating GA signaling cascades.