Use of Multiple Carboxylates to Increase Intracellular Retention of

Nov 20, 2007 - Xiaoxu Li,†,§ Ryuji Higashikubo,‡ and John-Stephen Taylor*,†. Chemistry Department, Washington University, St. Louis, Missouri 6...
0 downloads 0 Views 446KB Size
50

Bioconjugate Chem. 2008, 19, 50–56

Use of Multiple Carboxylates to Increase Intracellular Retention of Fluorescent Probes Following Release From Cell Penetrating Fluorogenic Conjugates Xiaoxu Li,†,§ Ryuji Higashikubo,‡ and John-Stephen Taylor*,† Chemistry Department, Washington University, St. Louis, Missouri 63130, and Department of Radiation Oncology, Washington University School of Medicine, St. Louis, Missouri 63108. Received March 8, 2007; Revised Manuscript Received September 18, 2007

Fluorogenic reporter systems for use inside cells require that the fluorophore be retained inside the cell following activation to ensure accumulation of an observable signal. In the process of developing ester-based nucleic acidtriggered probe activation systems for use in cells, we found that simple O-alkylated fluorescein esters coupled to cell-penetrating peptides led to very poor signals, presumably because the released fluorophore was too membrane permeable and rapidly exited the cell. To circumvent this problem, we have examined the effect of adding one or two carboxylates to the fluorescein to reduce its membrane permeability. N-maleimido D-valine and R-methylβ-L-alanine esters of fluorescein, in which the second phenolic hydroxyl group was derivatized with a carboxymethyl group and then further conjugated with glutamate, were linked to the cell-penetrating peptide Arg9Cys through conjugate addition of the thiol group to the maleimido group. HeLa cells were incubated with these conjugates, washed, and then further incubated for various times prior to analysis by flow cytometry. Quantitative analysis of the data by a simplified kinetic scheme showed that the fluorescein with two appended carboxylic acid groups effluxed with a rate constant of about 0.00113 min-1, corresponding to a half-life of 8.8 h. The dicarboxylated fluorescein effluxed about 6.1 times more slowly than the fluorescein with a single carboxylic acid group and led to a fairly stable signal. The analysis also showed that the D-Val ester was hydrolyzed about 4.6 times more slowly than the β-alanine ester and had a half-life of about 31 min. These data indicate that the fluorescein with two appended carboxylates may be a useful membrane-impermeant fluorophore for fluorogenic probe applications inside living cells.

INTRODUCTION Fluorogenic compounds have found extensive use as probes for a wide range of intracellular biomolecules and processes by becoming fluorescent only after activation by a specific chemical or enzymatic event (1–7). Fluorogenic compounds can also be used to assist the development and optimization of prodrugs by enabling the lability of the prodrug-activating reaction to be monitored in target and nontarget cells (8–10). Recently, we have proposed a new approach to chemotherapy and diagnostics called NATPA for nucleic acid-triggered prodrug and probe release (11–15). This new approach depends on a disease-specific mRNA sequence to template the association of a complementary prodrug or probe with a catalytic component that then catalyzes the release of the prodrug or probe. For this system to work efficiently, the prodrug or probe must be both stable to the diverse biological environments it will encounter and labile to the catalyst. Selecting an optimal prodrug or probe for NATPA requires screening a large number of candidates in serum and inside cells. An ideal fluorogenic probe would be one that enters the cell faster than it is activated and one for which the fluorescent product effluxes from the cell more slowly than it is produced (Figure 1). Recently, we developed a general fluorogenic system 1 (Figure 2) that could be used for monitoring the stability of * To whom correspondence should be addressed. Department of Chemistry, Campus Box 1134, Washington University, One Brookings Drive, St. Louis, MO 63130. E-mail: [email protected]. † Washington University. ‡ Washington University School of Medicine. § Present address: Department of Chemical Engineering and Columbia Genome Center, Columbia University, New York, NY 10032.

ester-based prodrugs and probes inside cells that consists of a fluorogenic ester of a monoalkylated fluorescein attached to a cell-penetrating Arg9 peptide (16). The advantage of using the cell-penetrating peptide, Arg9, to deliver a fluorogenic substrate is that this short peptide has been shown to rapidly deliver attached cargo into cells, irrespective of their inherent membrane permeability or size. The Arg9 peptide, which has charge and amino acid composition similar to that of HIV Tat peptide 46–57, promotes rapid cellular uptake of diverse types of molecules, such as drugs, peptides, and enzymes into cells, in many different cell lines (17–20). A strategy based on using a cell-penetrating peptide for delivering fluorogenic substrates is therefore much more flexible and general than one that requires that the fluorogenic agent itself be membrane permeable and that relies on the activation reaction to render the fluorophore less membrane permeable (21). This latter strategy is severely limited by how permeable the fluorogenic substrate can be made as well as how much less permeable the fluorescent probe can be made by the activating reaction. However, attachment of a cell-penetrating carrier such as Arg9 makes it possible to transport a variety of different fluorogenic probes into cells rapidly so that the activation rates of various probes can be easily compared and quantified. This method of cellular delivery has recently been applied to the design of a number of fluorogenic reagents (22). Esters of mono-O-alkylated fluoresceins were chosen as the fluorogenic ester component because monoalkylated fluoresceins have fluorescence properties similar to those of fluorescein and are easy to prepare, thereby making it easy to tailor the efflux properties of the fluorescein (16). When the remaining phenolic oxygen of a monoalkylated fluorescein is esterified, the fluorescein derivative becomes trapped in its nonfluorescent lactone

10.1021/bc070079y CCC: $40.75  2008 American Chemical Society Published on Web 11/20/2007

Cell-Penetrating Fluorogenic Esters

Figure 1. General mechanistic scheme for the entry and activation of a fluorogenic ester conjugate of a cell-penetrating peptide. In the first step, the conjugate enters the cell with rate constant kin, presumably by an endocytotic mechanism mediated by the cell-penetrating peptide (CPP), and exits by a slower, more complex mechanism. In the second step, the ester linkage is cleaved by endogenous enzymes, thereby activating the fluorophore with rate constant kact. Depending on the nature of the permeability factor (PF), the fluorophore will efflux the cell with rate constant keff.

form. As a result, di-O-substituted fluoresceins have been widely used as fluorogenic substrates for in Vitro enzyme activity studies. For example, mono-O-alkylated fluorescein phosphate has been used as a fluorogenic substrate for the continuous assay of phosphoinositide-specific phospholipase C, (23, 24) and N-acetyl-β-D-glucosaminides of 6-O′-alkylated fluorescein derivatives have been used as substrates for a kinetic assay of N-acetyl-β-D-glucosaminidase (25). In these studies, enzyme activities were quantified by analysis of the fluorescence resulting from enzymatic formation of the fluorescent monoalkylated fluorescein. As part of our program to develop fluorogenic probes for in ViVo gene expression, we wanted to examine fluorescent activation of the N-maleimido R-D,L-methyl-β-alanine and D-valine esters of monoalkylated fluoresceins. We have previously shown that nitrophenyl and fluorescein esters tethered to an oligodeoxynucleotide (ODN) or peptide nucleic acid (PNA) that is complementary to a target nucleic acid sequence can be hydrolyzed by imidazole tethered to another ODN or PNA that is complementary to the adjacent site on that sequence (11–13, 15). For our system to work in ViVo, however, requires that the background spontaneous or enzyme-catalyzed hydrolysis of the fluorescein esters occurs much more slowly than the mRNA templated hydrolysis of the esters by imidazole. We had already shown that increasing substitution of the ester slows down the background hydrolysis in human serum (12) but did not know what the effect would be inside human cells. In our previous study, we synthesized conjugates of Nmaleimido β-L-alanine and D-valine esters of a series of monoalkylated fluoresceins with Arg9Cys for the purpose of studying the stability of these ester linkages in cells (16). To render the monoalkylated fluorescein component more water

Bioconjugate Chem., Vol. 19, No. 1, 2008 51

Figure 2. Design and retrosynthesis of the fluorogenic ester conjugate 1. Cleavage of the ester linkage results in the formation of a fluorescent monoalkylated fluorescein derivative whose efflux rate depends on the properties of the attached permeability factor. Linkage of the cellpenetrating peptide Arg9Cys is carried out through conjugate addition of the thiol group on the peptide to the maleimido group on the fluorogenic ester. The fluorogenic ester is prepared by reacting the acid chloride of an N-maleoyl amino acid with a monoalkylated fluorescein that is prepared from fluorescein.

soluble and less membrane permeable, we had replaced the hydrophobic butyl substituent of a reported fluorescein derivative with methoxyethoxymethyl (MEM) and ethylmethylethene (EME) groups. Unfortunately, in subsequent studies, we found that these conjugates showed very little fluorescence inside cells in spite of other data indicating that these esters should have hydrolyzed rapidly. This suggested to us that the released fluorescein derivatives were effluxing faster than they were being produced. This is consistent with previous findings that fluorescein and some of its derivatives quickly exit numerous types of cell lines either via passive cell membrane permeation or via multidrug resistance protein 1(MRP1)-mediated transport (26–28). However, carboxylate-derivatized forms of fluorescein are not substrates for export by MRP1, and the presence of additional charges lowers their membrane permeability (28–30). In this article, we report the synthesis of Arg9 conjugates of fluorogenic esters of monoalkylated fluoresceins bearing one or two carboxylic groups. We also show that the fluorescein with two carboxylates effluxes much more slowly than the one with one carboxylate and that both lead to intracellular fluorescence signals in HeLa cells that can be quantified by flow cytometry. By using these conjugates, we were able to estimate the rates of background hydrolysis of the N-maleimido R-D,Lmethyl-β-alanine and D-valine ester linkages and the efflux rates of the two fluorophore derivatives.

EXPERIMENTAL PROCEDURES General Procedures. Dichloromethane (DCM) and triethylamine (TEA) were dried by refluxing with CaH2 overnight followed by distillation. Fluorescein, t-butyl-bromoacetate, ethylbromoacetate, thionyl chloride, D-valine, D,L-3-amino-2-methylpropionic acid, DCC, DIPEA, and TFA were purchased from

52 Bioconjugate Chem., Vol. 19, No. 1, 2008

Aldrich. Fmoc amino acids, t-butyl glutamic acid diester, resin, and ByPOP for solid phase peptide synthesis were purchased from NovaBiochem. N-maleoyl-D,L-R-methyl-β-alanine and N-maleoyl-D-valine were prepared as previously described (12). 1 H NMR spectra were obtained on a 300 MHz Varian UnityPlus-300 or Varian Mercury-300 spectrometer. Proton chemical shifts are expressed in ppm from tetramethylsilane. Flash chromatography was performed on Selecto Scientific silica gel. TLC and preparative TLC were run on precoated 254-nm fluorescent silica gel sheets manufactured by Alltech Associates. Fluorescence spectra were recorded on a SPEX Fluoromax instrument. UV spectral data were acquired on a Bausch and Lomb Spectronic 1001 spectrophotometer or Varian Cary 100 Bio UV–visible Spectrophotometer. Matrix-assisted laser desorption–ionization (MALDI) mass spectra of peptide conjugates were measured on a PerSeptive Voyager RP MALDI-time of flight (TOF) mass spectrometer. HR-FAB mass spectra were obtained on a MS-50TA (Ion Spec B126). HPLC purifications were carried out on Beckman Coulter System Gold 126 with RP C18 semi preparative columns. Compound 3a. t-Butyl-bromoacetate (1.07 g, 1.57 mL, 10.6 mmol) was added dropwise with stirring to a suspension of the disodium salt of fluorescein 2 (12) (1 g, 2.65 mmol) in DMF (30 mL) containing catalytic NaI at RT and then heated to 90 °C with stirring for 5 h. The DMF was removed from the mixture by evaporation under reduced pressure. The residue was redissolved in 50 mL of CH2Cl2 and washed with H2O (2 × 40 mL), then dried over Na2SO4, and evaporated to give the crude product. Further purification by chromatography on silica gel using hexane/ethyl acetate (2:1) gave 0.325 g (27.5% yield) of 3a as a yellow powder. 1H NMR (CD3COCD3): δ 1.47 (s, 9H), 4.71 (s, 2H), 6.64–6.66 (m, 2H), 6.69–6.77 (m, 3H), 6.84 (d, J)2.1 Hz 1H), 7.30 (d, J ) 7.5 Hz, 1H), 7.74 (td, J ) 7.2, 1.2 Hz, 1H), 7.81 (td, J ) 7.5, 1.2 Hz, 1H), 7.98 (d, J ) 7.5 Hz, 1H), 9.02 (s, 1H); HR FAB MS (m/z), calculated for C26H23O7 (M + H+) 447.1444; found, 447.1442. Compound 3b. Compound 5 (50 mg, 0.128 mmol) and t-butyl glutamic acid diester (free base) (35 mg, 0.134 mmol) were dissolved in 10 mL of DCM/DMF (50/50), and DCC (52 mg, 0.252 mmol) was added. The reaction mixture was stirred at RT for 1 h, after which 20 mL of DCM was added. After removal of the precipitate, the organic phase was washed with 1% KHSO4 aqueous solution (2 × 20 mL) and H2O (1 × 20 mL), dried over Na2SO4, and evaporated. The residue was purified by Flash silica gel chromatography using ethyl acetate/ hexane (1:1.3) to give 70 mg of compound 3b as a yellow syrup (87% yield). 1H NMR (CD3COCD3): δ 1.41 (s, 9H), 1.44 (s, 9H), 1.69–1.79 (m, 2H), 2.32 (t, J ) 7.5 Hz, 2H), 4.44–4.49 (m, 1H), 4.65 (d, J ) 2.7 Hz, 2H), 6.61–6.67 (m, 2H), 6.69–6.83 (m, 3H), 6.95 (d, J ) 2.1 Hz, 1H), 7.29 (d, J ) 7.5 Hz, 1H), 7.74 (t, J ) 7.2 Hz, 1H), 7.82 (t, J ) 7.2 Hz, 1H), 7.98 (d, J ) 7.5 Hz, 1H); HR FAB MS (m/z), calculated for C35H38O10N (M + H+) 632.2496; found, 632.2500. Compound 5. Ethyl-bromoacetate (1.18 mL, 1.77 g, 10.6 mmol) was added dropwise with stirring to a suspension of the disodium salt of fluorescein 2 (1.00 g, 2.65 mmol) in DMF (30 mL) at RT and then heated at 90–100 °C with stirring for 5 h. The DMF was removed from the mixture by evaporation under reduced pressure and the residue treated with 25 mL of 4% NaOH in MeOH/water (3:1, v/v) and stirred for 2 h. After the removal of MeOH in vacuum with a rotatory evaporator, the solution was then acidified by the dropwise addition of 1 M aqueous HCl until a pH of about 2.0. The aqueous solution was extracted with ethyl acetate (3 × 20 mL), and the combined ethyl acetate fractions were dried over Na2SO4, and evaporated to give the crude product. Further purification by flash chromatography on silica gel using ethyl acetate/hexane (3:1) then

Li et al.

ethyl acetate/methanol (10:1) gave 0.53 g (51% yield) of 5 as a yellow powder. 1H NMR (CD3OD): δ 4.73 (s, 2H), 6.55–6.58 (m, 2H), 6.68–6.73 (m, 3H), 6.85–6.86 (m, 1H), 7.20 (d, J ) 7.2 Hz, 1H), 7.70 (td, J ) 7.5 Hz, 1.2 Hz, 1H), 7.78 (td, J ) 7.2, 1.2 Hz, 1H), 8.01 (d, J ) 7.5 Hz, 1H); HR FAB MS (m/z), calculated for C22H15O7 (M + H+) 391.0818; found, 391.0818. Compound 8a. This compound was prepared as described for compound 9a. N-maleoyl-D,L-3-amino-2-methylpropionic acid (36 mg, 0.18 mmol) was treated with 5 mL of thionyl chloride, and the resulting mixture was slowly added to a stirred mixture of 3a (40 mg, 0.090 mmol) and triethylamine (TEA) (25 µL, 18 mg, 0.18 mmol) in 10 mL of CH2Cl2 at 0 °C and allowed to warm to room temperature for 3 h. Following workup, the residue was purified by preparative silica gel TLC (1:2 ethyl acetate–hexane) to give 42 mg (77% yield) of compound 8a as a white powder. 1H NMR (CD3OCD3): δ 1.26 (d, J ) 7.2 Hz, 3H), 1.47 (s, 9H), 3.08–3.16 (m, 1H), 3.76 (dd, J ) 14.1, 6 Hz, 1H), 3.92 (dd, J ) 14.1, 7.5 Hz, 1H), 4.73 (s, 2H), 6.74–6.83 (m, 2H), 6.88–6.97 (m, 5H), 7.23–7.25 (m, 1H), 7.36 (d, J ) 7.2 Hz, 1H), 7.77 (td, J ) 7.5 Hz, 1.2 Hz, 1H), 7.84 (td, J ) 7.5, 1.2 Hz, 1H), 8.03 (d, J ) 7.2 Hz, 1H); HR FAB MS (m/z), calculated for C34H30NO10 (M + H+) 612.1869; found, 612.1871. Compound 8b. Compound 8b was prepared from N-maleoylD,L-R-methyl-β-alanine (22 mg, 0.11 mmol) and 3b (34 mg, 0.054 mmol) following the same procedure described for the preparation of 8a. Preparative silica gel TLC (1:2 ethyl acetate–hexane) gave 24 mg (56% yield) of 8b as a white powder. 1H NMR (CD3COCD3): δ 1.26 (d, J ) 7.2 Hz, 3H), 1.41 (s, 9H), 1.44 (s, 9H), 1.70–1.80 (m, 2H), 2.32 (t, J ) 7.5 Hz, 2H), 3.08–3.17(m, 1H), 3.76 (dd, J ) 14.1, 6 Hz, 1H), 3.92 (dd, J ) 13.5, 7.5 Hz, 1H), 4.43–4.51 (m, 1H), 4.67 (d, J ) 2.1 Hz, 2H), 6.72–6.86 (m, 2H), 6.87–7.02 (m, 5H), 7.23–7.25 (m, 1H), 7.35 (d, J ) 7.2 Hz, 1H), 7.77 (t, J ) 7.5 Hz, 1H), 7.84 (t, J ) 7.5, 1H), 8.03 (d, J ) 7.5 Hz, 1H); HR FAB MS (m/z), calculated for C43H45O13N2 (M + H+) 797.2921; found, 797.2922. Compound 9a. N-Maleoyl-D-valine (29 mg, 0.14 mmol) (11) was dissolved in 5 mL of thionyl chloride and heated at reflux until gas evolution had ceased; at which point, the excess thionyl chloride was removed under reduced pressure. The residual thionyl chloride was removed by dissolving the residue in carbon tetrachloride and then removing the solvent under reduced pressure. The crude acid chloride 7 was dissolved in 5 mL of CH2Cl2 and was slowly added to a stirred mixture of 3′-Obutyloxycarbonylmethyl-fluorescein (3a) (30 mg, 0.067 mmol) and triethylamine (19 µL, 14 mg, 0.14 mmol) in 10 mL of CH2Cl2 at 0 °C. The reaction mixture was allowed to warm to RT. After stirring for 3 h, the reaction mixture was diluted with 20 mL of CH2Cl2, washed with brine and water (2 × 25 mL), dried over Na2SO4, and concentrated in vacuo. The residue was purified by silica gel TLC (1:2 ethyl acetate–hexane) to afford 22.3 mg (53% yield) of 9a as a white powder. 1H NMR (CD3OCD3): δ 0.94 (d, J ) 6.9 Hz, 3H), 1.14 (d, J ) 6.9 Hz, 3H), 1.47 (s, 9H), 2.58–2.70 (m, 1H), 4.73 (s, 2H), 4.83 (d, J ) 7.2 Hz, 1H), 6.74–6.83 (m, 2H), 6.87–6.95 (m, 3H), 7.05–7.14 (m, 1H), 7.34 (d, J ) 7.2 Hz, 1H), 7.76 (t, J ) 7.2 Hz, 1H), 7.83 (t, J ) 7.5, 1H), 8.02 (d, J ) 7.2 Hz, 1H); HR FAB MS (m/z), calculated for C35H32NO10 (M + H+) 626.2026; found, 626.2010. Compound 9b. Compound 9b was prepared from N-maleoylD-valine (23 mg, 0.11 mmol) and 3b (34 mg, 0.054 mmol) following the same procedure described for the preparation of 9a. Preparative silica gel TLC (1:2 ethyl acetate–hexane) gave 15 mg (34% yield) of 9b as a white powder. 1H NMR (CD3COCD3): δ 0.94 (d, J ) 6.9 Hz, 3H), 1.12 (d, J ) 6.9 Hz,

Cell-Penetrating Fluorogenic Esters

3H), 1.41 (s, 9H), 1.44 (s, 9H), 1.69–1.79 (m, 2H), 2.32 (t, J ) 7.5 Hz, 2H), 2.58–2.69 (m, 1H), 4.43–4.49 (m, 1H), 4.67 (d, J ) 2.1 Hz, 2H), 4.78 (d, J ) 7.2 Hz, 1H), 6.71–7.00 (m, 5H), 7.04–7.14 (m, 3H), 7.34 (d, J ) 7.5 Hz, 1H), 7.76 (t, d, J ) 7.5, 1.2 Hz, 1H), 7.83 (t, d, J ) 7.2, 1.2 Hz, 1H), 8.02(d, J ) 7.2 Hz, 1H). HR FAB MS (m/z), calculated for C44H47O13N2 (M + H+) 811.3078; found, 811.3072. General Procedure for the Preparation of Compounds 10a,b and 11a,b. Compounds 8a,b and 9a,b were dissolved in TFA-DCM (50/50, v/v) and stirred under N2 at RT for 4 h. Complete removal of TFA and DCM under reduced pressure afforded the corresponding deprotected compounds 10a,b and 11a,b as white solids. Compound 10a. 1H NMR (CD3OCD3): δ 1.25 (d, J ) 7.2 Hz, 3H), 3.06–3.17 (m, 1H), 3.76 (dd, J ) 14.1, 5.7 Hz, 1H), 3.92 (dd, J ) 14.1 Hz, 7.8 Hz, 1H), 4.83 (s, 2H), 6.76–6.82 (m, 2H), 6.89–6.96 (m, 5H), 7.23–7.25 (m, 1H), 7.35 (d, J ) 7.5 Hz, 1H), 7.76 (t, J ) 7.5 Hz, 1H), 7.83 (t, J ) 7.2 Hz, 1H), 8.02 (d, J ) 7.2 Hz, 1H); HR FAB MS (m/z), calculated for C30H22NO10 (M + H+) 556.1243; found, 556.1223. Compound 10b. 1H NMR (CD3COCD3): δ 1.26 (d, J ) 7.2 Hz, 3H), 1.70–1.80 (m, 2H), 2.45 (t, J ) 7.5 Hz, 2H), 3.08–3.18 (m, 1H), 3.76 (dd, J ) 14.1, 6 Hz, 1H), 3.92 (dd, J ) 13.8, 7.5 Hz, 1H), 4.58–4.65 (m, 1H), 4.67 (s, 1H), 4.91 (s, 1H), 6.72–6.85 (m, 2H), 6.87–7.02 (m, 5H), 7.23–7.25 (m, 1H), 7.35 (d, J ) 7.2 Hz, 1H), 7.77 (t, J ) 7.2 Hz, 1H), 7.84 (t, J ) 7.5, 1H), 8.03 (d, J ) 7.2 Hz, 1H); HR ESI MS (m/z), calculated for C35H28O13N2Na (M+Na+) 707.1513; found, 707.1498. Compound 11a. 1H NMR (CD3OCD3): δ 0.89 (d, J ) 7.2 Hz, 3H), 1.10 (d, J)7.2 Hz, 3H), 2.54–2.66 (m, 1H), 4.58 (s, 2H), 4.81 (d, J ) 6.9 Hz, 1H), 6.55–7.06 (m, 6H), 7.34 (m, 1H), 7.59–7.75 (m, 2H), 7.95 (d, J ) 6.9 Hz, 1H); HR FAB MS (m/z), calculated for C31H24NO10 (M + H+) 570.1400; found, 570.1404. Compound 11b. 1H NMR (CD3COCD3): δ 0.93 (d, J ) 6.9 Hz, 3H), 1.12 (d, J ) 6.9 Hz, 3H), 1.69–1.81 (m, 2H), 2.42 (t, J ) 7.5 Hz, 2H), 2.58–2.69 (m, 1H), 4.47–4.54 (m, 1H), 4.65 (s, 1H), 4.78 (d, J ) 7.2 Hz, 1H), 4.91 (s, 1H), 6.71–7.00 (m, 5H), 7.04–7.14 (m, 3H), 7.34 (d, J ) 7.2 Hz, 1H), 7.76 (t,d J ) 7.5, 1.2 Hz, 1H), 7.83 (t,d J ) 7.5, 1.2 Hz, 1H), 8.02 (d, J ) 7.5 Hz, 1H). HR FAB MS (m/z), calculated for C36H31O13N2 (M + H+) 699.1826; found, 699.1829. Synthesis of TyrArg9Cys. The oligopeptide H2N-Tyr-(Arg)9Cys-CONH2 was synthesized by standard solid phase Fmoc peptide synthesis methodology as described previously (16). The oligopeptide was purified by semipreparative RP-HPLC with a 30 min 0–40% gradient of A (acetonitrile containing 0.1% TFA) in B (water containing 0.1% TFA) at a flow rate of 1.0 mL/ min and detection at 270 nm. MALDI-TOF MS (m/z), calculated (M + 1) 1689; found, (M + 1): 1689.8. General Method for the Preparation of Arg9Cys Fluorogenic Ester Conjugates. Compounds 10a,b and 11a,b (4 equiv) were each dissolved in 50 µL of acetonitrile, added separately to TyrArg9Cys in 0.5 mL of deoxygenated MeOH, and then shaken under N2 for 2–4 h at RT. Ethyl ether (10-fold excess) was then added to precipitate the products. Purification was carried out by semipreparative RP-HPLC with a 1.0 mL/ min 30 min 0–40% gradient of B (acetonitrile containing 0.1% TFA) in A (water containing 0.1% TFA) with detection at 270 nm. MALDI-TOF MS (m/z), calculated (M + 1); observed, 12a, 2244, 2245.73; 12b, 2373, 2374.37; 13a, 2258, 2259.51; 13b, 2387, 2389.41. Flow Cytometry. Exponentially growing HeLa cells were subcultured 48 h before the studies at a concentration that would ensure the exponential growth of cells at the time of experimentation. The cells were monodispersed by trypsinization and washed twice with 5 mL of Hanks’ balanced salt solution

Bioconjugate Chem., Vol. 19, No. 1, 2008 53

containing 10 mM Hepes (pH 7.3), 1 mM MgCl, 1 mM CaCl, and 0.1% bovine serum albumin (HBSS buffer). An aliquant of the samples were examined by flow cytometry to establish the background fluorescence intensity level. To 1 mL of cell suspension containing approximately 0.75 × 106 cells, an appropriate volume of a probe was added, and flow cytometric observation was conducted at 1 min after the addition. Cells were incubated for 5 min with 40 µM probe, washed twice with 5 mL of HBSS, and resuspended in 1 mL of HBSS. Cells were kept at room temperature in the dark and analyzed by flow cytometry at the times indicated. Flow cytometry was carried out on a Becton Dickinson FACS 440 (San Jose, CA). The cells were excited with 300 mW of 488 nm argon ion laser, and fluorescence was detected through a 525 nm band-pass filter. A minimum of 20,000 events, gated by forward and side-scatter intensities, were recorded for each time point in list mode in a CYTOMATION CICERO/CYCLOPS data acquisition/analysis system (Ft. Collins, CO). The mean intensity of green fluorescence for intact cells was calculated and normalized to the background fluorescence intensity of unlabeled cells. Three independent experiments were conducted for each of the probes, and their means are presented in the data (except for 12a, for which only two of the data sets were used because of a data collection problem). Following flow cytometric examination, samples containing fluorescent cells were examined by fluorescence microscopy (Olympus BX40, Olympus, Center Valley, PA).

RESULTS AND DISCUSSION Design. The fluorogenic esters were designed so that it would be easy to systematically vary the cell-penetrating peptide, ester substrate, and the permeability and other properties of the fluorophore (Figure 2). The peptide component can be synthesized by solid phase Fmoc synthesis, while ester linkage can be made directly from a variety of natural and unnatural amino acids through a simple two-step procedure. The permeability properties of fluorescein can be tailored through a side chain attached to one of the phenolic oxygens. Though the fluorogenic esters prepared in this article were made by solution-phase synthesis, it would be quite easy to adapt the synthetic scheme to a solid supported route, as we have previously found that the phenyl esters are stable to the deprotection conditions for the peptide and carboxylate groups. Synthesis of the Monoalkylated Fluoresceins. Fluorescein was monoalkylated as illustrated in Scheme 1 by a method we previously reported for the synthesis of other monoalkylated derivatives (16). The disodium salt of fluorescein 2 was prepared by treatment of fluorescein 1 with 2 equiv of NaOH and then alkylated with t-butyl R-bromoacetate to afford t-butylcarboxymethyl fluorescein (3a) and dialkylated product 4. The dialkylated product could be converted back to 3a by hydrolysis with NaOH followed by acidification but was never carried out in practice. Removal of the t-butyl group with TFA then provided carboxylic acid 5, which is a more versatile intermediate than fluorescein for adding functionality, as it can be easily coupled to a wide variety of functionalized amines and amino acids by amide bond formation. Compound 5 could be obtained in a better overall yield of 52%, however, by basic hydrolysis of the mixture of monoalkylated and dialkylated fluoresceins that is produced by treatment of the disodium salt of fluorescein 2 with ethylbromoacetate instead of t-butylbromoacetate. To prepare an acid labile precursor to a dicarboxylate, compound 5 was coupled to the amino group of the di-t-butyl ester of L-glutamic acid with DCC to produce compound 3b in 53% yield. Synthesis of N-Maleoyl Amino Acid Esters of the Monoalkylated Fluoresceins. The monoalkylated fluoresceins 3a and 3b were then converted to the N-maleoyl-R-methyl-β-

54 Bioconjugate Chem., Vol. 19, No. 1, 2008 Scheme 1. Synthesis of the Monoalkylated Fluorescein Building Blocks 3a and 3b

Scheme 2. Synthesis of the N-Maleoyl Amino Acid Esters of the Monoalkylated Fluoresceins, 10a,b and 11a,b

amino butyric acid and N-maleoyl-D-Val esters 8a,b and 9a,b via acid chlorides 6 and 7 in the presence of triethylamine (Scheme 2). These acids chlorides are readily prepared in two steps from the corresponding amino acids (11). The t-butyl groups were then removed by TFA to afford the maleimido fluorogenic esters 10a,b and 11a,b bearing one and two carboxylates.

Li et al. Scheme 3. Synthesis of the Arg9Cys Conjugates of the Fluorogenic Esters, 12a,b and 13a,b

Conjugation of the N-Maleimido Amino Acid Esters of the Monoalkylated Fluoresceins to Arg9Cys. The cellpenetrating peptide Arg9 with a cysteine at the carboxy terminus was synthesized by standard Fmoc solid phase synthesis. The cysteine was introduced to provide a thiol group at the carboxy terminus of the peptide for conjugating to the maleimide group of the N-maleoyl amino acid esters 10a,b, and 11a,b. The resulting fluorogenic ester conjugates 12a,b and 13a,b were obtained by incubating the peptide with the maleimido esters in a 1:4 ratio for 2 h (Scheme 3). Methanol was preferred as a solvent because of its ability to dissolve both reactants and from which the conjugated products could be easily isolated by ether precipitation. All of the conjugates were purified by RP-HPLC and characterized by MALDI-TOF MS. Flow Cytometry. To determine the stability of the esters inside living cells, the Arg9 peptide fluorescein ester conjugates 12a,b and 13a,b were incubated with Hela cells (one million/ mL) in media for 5 min at a concentration of 40 µM, after which

Figure 3. Fluorescence activation of the fluorogenic ester conjugates 12a,b and 13a,b in HeLa cells. A plot of the average fluorescence observed in HeLa cells as a function of time and fitted to an analytical expression for two sequential irreversible first order steps is shown. The error bars represent the standard deviation for three separate experiments (except for 12b, which is for two experiments). All values for the plot are given in Table 1.

Cell-Penetrating Fluorogenic Esters

Bioconjugate Chem., Vol. 19, No. 1, 2008 55

Table 1. Kinetic Parameters for the Fluorescence Activation of the Ar9 Fluorogenic Ester Conjugates 12a,b and 13a,b by Fitting Each Curve Individually To an Expression for Two Sequential Irreversible Steps

a

cmpd

amino acid

no. of carboxylates

Ao

kact (min-1)

keff (min-1)

12a 12b 13a 13b

R-methyl-β-Ala R-methyl-β-Ala D-Val D-Val

1 2 1 2

29.7 ( 1.0 22.9 ( 0.6 14.1 ( 0.9a 21.7 ( 0.5b

0.12 ( 0.016 0.087 ( 0.007 0.032 ( 0.005a 0.017 ( 0.0008b

0.008 ( 0.0006 0.0013 ( 0.0003 0.008a 0.0013b

Values from fitting when keff is fixed at 0.008 min-1. b Values from fitting when keff is fixed at 0.0013 min-1.

the cells were washed twice to remove agents that had not penetrated the cells. The cells were then incubated under standard conditions (37 °C, 5% CO2), for various times, after which the average fluorescence intensity of the cells was determined by flow cytometry for three separate experiments and then averaged except for 12b for which two data sets were averaged (Figure 3). Fluorescence from the R-methyl-β-alanine ester conjugate bearing a single carboxylate, 12a, increased more rapidly than that from D-valine conjugate 13a, but both lost fluorescence at longer incubation times. This behavior is what would be expected for a fluorescent probe that is able to escape from the cell. Fluorescence from the R-methyl-β-alanine ester conjugate bearing two carboxylates, 12b, also increased more rapidly than for D-valine conjugate 13b, but unlike the conjugates bearing a single carboxylate, 12a and 13a, the fluorescence did not appreciably decrease at longer incubation times. Thus, it would appear that having two additional negative charges on the fluorophore was sufficient to render it practically membrane impermeable. Attempts to visualize the location of the released probes by fluorescence microscopy 30 and 120 min after treatment revealed only faint images, possibly due to low concentration and diffuse, nonlocalized distribution of the probes. Semi-Quantitative Analysis of the Flow Cytometry Data. Because of the limited data, it was not possible to reliably fit all the rate constants in Scheme 1 for the four different probes. Therefore, we decided to carry out a semiquantitative assessment of the rates of fluorescent probe activation and efflux from the cells by fitting the fluorescence data to an analytical expression (31) for two sequential irreversible first order steps, Ain f Bin f Bout, with rate constants kact and keff as follows: [Bin] ) [Ain]o * (τκ – τ)/(1 – κ), where τ ) exp(–kactt) and κ ) keff/kact. In this expression, kact corresponds to the rate constant for activating the fluorogenic probe A inside the cell, and keff corresponds to the rate constant for efflux of the resulting fluorophore B from inside the cell as diagrammed in Figure 1. This simplified analysis is based on the assumption that there is no external fluorogenic probe that can enter the cell (kin) following washing of the cells and that efflux of the fluorogenic probe from the cell (kout) is negligible compared to kact. This would appear to be a reasonable set of assumptions given the results of many studies indicating that cationically charged cellpenetrating peptides enter cells rapidly through macropinocytosis and become trapped in endosomes (32–34). Furthermore, any activation of the fluorogenic probe bound to the surface would not contribute to the signal observed by flow cytometry, as it would be released into the medium. Fits of the fluorescence data for the two R-methyl-β-alanine conjugates to the two-step model gave a rate constant of 0.12 min-1 for fluorescence activation of the conjugate with one negative charge (12a), which was very similar to that of 0.087 min-1 for the conjugate with two negative charges (12b). The average half-life for the activation of these two conjugates was about 6.8 min. The efflux rates of the released fluoresceins, however, differ by a factor of about 6.1, with the fluorescein bearing two carboxylates (14b, Scheme 3) effluxing with a rate constant of about 0.0013 min-1, corresponding to a half-life of 8.8 h. The fluorescein bearing one carboxylate (14a), however,

effluxes much more rapidly with a rate constant of 0.008 min-1 corresponding to a half-life of 87 min. The amplitudes, [Ain]o for the fits to 12a and 12b differed slightly (29.7 vs 22.9) and might have been due to differing amounts of material absorbed by the cells and/or to differing experimental conditions. Attempts to fit the data for the two D-valine ester conjugates 13a and 13b in a similar fashion resulted in rate constants and amplitudes with large standard deviations, indicating that the fit for these curves was underdetermined. To reduce the number of variables to be fit, we fixed the efflux rate constants for the activated probes 14a and 14b to be the same as those obtained from fitting the data for 12a and 12b, which release the same probes. Fixing the efflux rates in this manner led to good fits with small standard deviations. The activation rate constants and amplitudes differed by almost a factor of two, however, suggesting that the kinetic scheme may be more complicated for this set of probes and warrants further study. The average of the activation rate constants for the D-valine ester (0.025 min-1) determined in this fashion is about 4.6 times less than the average for the alanine esters of 0.11 min-1, indicating that the D-valine ester is more stable than the alanine ester inside HeLa cells. This is consistent with previous studies on the stability of D-valine esters compared to glycine esters in human serum (12). The calculated half-life of about 31 min for the D-valine ester, though much longer than that of about 7 min for the alanine ester, is probably too short to be useful for NATPA in ViVo. We propose, however, to use this assay system to screen for a more stable ester.

CONCLUSIONS In our studies, we found that the Arg9 peptide can rapidly transport a sufficient amount of fluorogenic ester into cells to lead to quantifiable signals in flow cytometry that can be analyzed by a simple two-step kinetic scheme involving fluorescence activation followed by efflux. These conjugates would therefore appear to be useful for evaluating libraries of fluorogenic esters in which the structure and properties of the ester and the permeability factor are systematically varied. In this way, we should be able to identify fluorogenic esters suitable for NATPA and possibly fluorogenic esters capable of distinguishing between various cell types and conditions. The ability to tune the hydrolysis and efflux rates could be very useful for optimizing fluorogenic compounds for particular assays.

ACKNOWLEDGMENT This work was financially supported by an NIH grant (RO1CA92477). Mass spectrometry was provided by the Washington University Mass Spectrometry Resource (Grant No. P41RR0954).

LITERATURE CITED (1) Boonacker, E., and Van Noorden, C. J. (2001) Enzyme cytochemical techniques for metabolic mapping in living cells, with special reference to proteolysis. J. Histochem. Cytochem. 49, 1473–1486. (2) Wahler, D., and Reymond, J. L. (2001) High-throughput screening for biocatalysts. Curr. Opin. Biotechnol. 12, 535–544.

56 Bioconjugate Chem., Vol. 19, No. 1, 2008 (3) Fang, X., Mi, Y., Li, J. J., Beck, T., Schuster, S., and Tan, W. (2002) Molecular beacons: fluorogenic probes for living cell study. Cell Biochem. Biophys. 37, 71–81. (4) Becker, S., Schmoldt, H. U., Adams, T. M., Wilhelm, S., and Kolmar, H. (2004) Ultra-high-throughput screening based on cellsurface display and fluorescence-activated cell sorting for the identification of novel biocatalysts. Curr. Opin. Biotechnol. 15, 323–329. (5) Martinez-Manez, R., and Sancenon, F. (2005) New advances in fluorogenic anion chemosensors. J. Fluoresc. 15, 267–285. (6) Mohr, G. J. (2006) New chromogenic and fluorogenic reagents and sensors for neutral and ionic analytes based on covalent bond formation: a review of recent developments. Anal. Bioanal. Chem. 386, 1201–1214. (7) Lavis, L. D., Chao, T. Y., and Raines, R. T. (2006) Fluorogenic label for biomolecular imaging. ACS Chem. Biol. 1, 252–260. (8) Stoeckel, K., Hofheinz, W., Laneury, J. P., Duchene, P., Shedlofsky, S., and Blouin, R. A. (1998) Stability of cephalosporin prodrug esters in human intestinal juice: implications for oral bioavailability. Antimicrob. Agents Chemother. 42, 2602–2606. (9) Oliyai, R., Shaw, J. P., Sueoka-Lennen, C. M., Cundy, K. C., Arimilli, M. N., Jones, R. J., and Lee, W. A. (1999) Aryl ester prodrugs of cyclic HPMPC. I: Physicochemical characterization and in vitro biological stability. Pharm. Res. 16, 1687–1693. (10) Yang, C., Gao, H., and Mitra, A. K. (2001) Chemical stability, enzymatic hydrolysis, and nasal uptake of amino acid ester prodrugs of acyclovir. J. Pharm. Sci. 90, 617–624. (11) Ma, Z., and Taylor, J.-S. (2000) Nucleic acid triggered catalytic drug release. Proc. Natl. Acad. Sci. U.S.A. 97, 11159–11163. (12) Ma, Z., and Taylor, J. S. (2001) Nucleic acid triggered catalytic drug and probe release: a new concept for the design of chemotherapeutic and diagnostic agents. Bioorg. Med. Chem. 9, 2501–2510. (13) Ma, Z., and Taylor, J. S. (2003) PNA-based RNA-triggered drug-releasing system. Bioconjugate Chem. 14, 679–683. (14) Cai, J., Li, X., Yue, X., and Taylor, J. S. (2004) Nucleic acidtriggered fluorescent probe activation by the Staudinger reaction. J. Am. Chem. Soc. 126, 16324–16325. (15) Cai, J., Li, X., and Taylor, J. S. (2005) Improved nucleic acid triggered probe activation through the use of a 5-thiomethyluracil peptide nucleic acid building block. Org. Lett. 7, 751–754. (16) Li, X., and Taylor, J. S. (2004) General strategy for the preparation of membrane permeable fluorogenic peptide ester conjugates for in vivo studies of ester prodrug stability. Bioorg. Med. Chem. 12, 545–552. (17) Rothbard, J. B., Garlington, S., Lin, Q., Kirschberg, T., Kreider, E., McGrane, P. L., Wender, P. A., and Khavari, P. A. (2000) Conjugation of arginine oligomers to cyclosporin A facilitates topical delivery and inhibition of inflammation. Nat. Med. 6, 1253–1257. (18) Chen, L., Wright, L. R., Chen, C. H., Oliver, S. F., Wender, P. A., and Mochly-Rosen, D. (2001) Molecular transporters for peptides: delivery of a cardioprotective epsilonPKC agonist peptide into cells and intact ischemic heart using a transport system, R(7). Chem. Biol. 8, 1123–1129. (19) Jin, L. H., Bahn, J. H., Eum, W. S., Kwon, H. Y., Jang, S. H., Han, K. H., Kang, T. C., Won, M. H., Kang, J. H., Cho, S. W., Park, J., and Choi, S. Y. (2001) Transduction of human catalase mediated by an HIV-1 TAT protein basic domain and arginine-

Li et al. rich peptides into mammalian cells. Free Radical Biol. Med. 31, 1509–1519. (20) Futaki, S. (2006) Oligoarginine vectors for intracellular delivery: design and cellular-uptake mechanisms. Biopolymers 84, 241–249. (21) Haugland, R. P. (1999) Handbook of Fluorescent Probes and Research Chemicals, Molecular Probes, Eugene, OR. (22) Bullok, K. E., Gammon, S. T., Violini, S., Prantner, A. M., Villalobos, V. M., Sharma, V., and Piwnica-Worms, D. (2006) Permeation peptide conjugates for in vivo molecular imaging applications. Mol. Imaging 5, 1–15. (23) Rukavishnikov, A. V., Smith, M. P., Birrell, G. B., Keana, J. F., and Griffith, O. H. (1998) Synthesis of a new fluorogenic substrate for the assay of phosphoinositide-specific phospholipase. Tetrahedron Lett. 39, 6637–6640. (24) Zaikova, T. O., Rukavishnikov, A. V., Birrell, G. B., Griffith, O. H., and Keana, J.F.W. (2001) Synthesis of fluorogenic substrates for continuous assay of phosphatidylinositol-specific phospholipase C. Bioconjugate Chem. 12, 307–313. (25) Kasai, K., Okada, K., and Yamaji, N. (1993) Glycosides having chromophores as substrates for sensitive enzyme analysis. V. Synthesis of 6′-O-substituted 2′,7′-dichlorofluorescein Nacetyl-b-D-glucosaminides as substrates for the rate-assay of N-acetyl-b-D-glucosaminidase. Chem. Pharm. Bull. 41, 1513– 1520. (26) Rotman, B., and Papermaster, B. W. (1966) Membrane properties of living mammalian cells as studied by enzymatic hydrolysis of fluorogenic esters. Proc. Natl. Acad. Sci. U.S.A. 55, 134–141. (27) Prosperi, E., Croce, A. C., Bottiroli, G., and Supino, R. (1986) Flow cytometric analysis of membrane permeability properties influencing intracellular accumulation and efflux of fluorescein. Cytometry 7, 70–75. (28) Rychlik, B., Balcerczyk, A., Klimczak, A., and Bartosz, G. (2003) The role of multidrug resistance protein 1 (MRP1) in transport of fluorescent anions across the human erythrocyte membrane. J. Membr. Biol. 193, 79–90. (29) Dive, C., Cox, H., Watson, J. V., and Workman, P. (1988) Polar fluorescein derivatives as improved substrate probes for flow cytoenzymological assay of cellular esterases. Mol. Cell. Probes 2, 131–145. (30) Homolya, L., Hollo, Z., Germann, U. A., Pastan, I., Gottesman, M. M., and Sarkadi, B. (1993) Fluorescent cellular indicators are extruded by the multidrug resistance protein. J. Biol. Chem. 268, 21493–21496. (31) Chien, J.-Y. (1948) Kinetic analysis of irreversible consecutive reactions. J. Am. Chem. Soc. 70, 2256–22561. (32) Potocky, T. B., Menon, A. K., and Gellman, S. H. (2003) Cytoplasmic and nuclear delivery of a TAT-derived peptide and a beta-peptide after endocytic uptake into HeLa cells. J. Biol. Chem. 278, 50188–50194. (33) Kaplan, I. M., Wadia, J. S., and Dowdy, S. F. (2005) Cationic TAT peptide transduction domain enters cells by macropinocytosis. J. Controlled Release 102, 247–253. (34) Fotin-Mleczek, M., Fischer, R., and Brock, R. (2005) Endocytosis and cationic cell-penetrating peptides: a merger of concepts and methods. Curr. Pharm. Des. 11, 3613–3628. BC070079Y