|Home | About | Journals | Submit | Contact Us | Français|
Our goal is to define the properties of cell-impermeable cargo molecules that can be delivered into cells by pHLIP (pH (Low) Insertion Peptide), which can selectively target tumors in vivo based on their acidity. Using biophysical methods and fluorescence microscopy, we show that pHLIP can successfully deliver polar and membrane-impermeable cyclic peptides linked to its C-terminus through the membranes of lipid vesicles and cancer cells. Our results also indicate that the translocation of these cargo molecules is pH-dependent and mediated by transmembrane helix formation. Since a broad range of cell-impermeable molecules is excluded from discovery efforts because they cannot traverse membranes on their own, we believe that pHLIP has the potential to expand therapeutic options for acidic tissues such as tumors and sites of inflammation.
Since many anticancer drugs have off-target side effects that severely limit the efficacy of chemotherapy (Fennelly, 1995; Ross and Small, 2002; Sehouli, et al., 2002), the use of targeted drug delivery systems has emerged as a way to diminish such cytotoxic effects on healthy organs (Minko, et al., 2004). Some, such as liposomes (Fenske and Cullis, 2005), polymers (Duncan, 2006; Haag and Kratz, 2006) and other nanoparticles (Torchilin, 2007; Yezhelyev, et al., 2006), can passively target tumors due to the enhanced permeation and retention effect from the increased permeability and minimal lymphatic drainage of many solid tumors (Maeda, et al., 2001; Maeda, et al., 2000). However, this effect is small for certain tumors, including those that are pre-angiogenic, necrotic, or poorly vascularized (Allen and Cullis, 2004). Therefore, other specific targeting strategies have been developed, usually involving the addition of ligands that target a particular type of binding site on cancer cells. Most cancer cells over-express certain receptors and enzymes (cancer bio-markers) due to the change in biology required for tumor progression (Hanrahan, et al., 2005). For example, the receptors for luteinizing hormone-releasing hormone are over-expressed in breast, ovarian, and prostate cancer cells (Dharap and Minko, 2003; Dharap, et al., 2003; Minko, et al., 2004). Unfortunately, an important limitation of targeting specific binding sites is the heterogeneity of human tumors. Tumor cells are genetically and phenotypically complex, due in part to the accumulation of a large number of somatic mutations in the cancer genome and to the genetic variations in individual patients, so that only a subset of the cells in a tumor may be targeted and the rest continue to proliferate (Bild, et al., 2006; Bild, et al., 2006; Murphy, et al., 2005). Cancer cells inside the same tumor or in different tumors from the same origin might then have different receptors on their surface. Thus, it may not be enough to rely on a single cancer biomarker even for one type of cancer or a single tumor, so it is important to ask which other features of tumors may be exploited for targeting.
Rapidly expanding cancer cells have abnormal nutritional requirements and metabolic traits, creating unique micro-environmental features that distinguish almost all malignant solid tumors from surrounding normal tissues. One such characteristic is the lower extracellular pH of tumors when compared with healthy tissues (pH 6.5-7.0 vs. pH 7.2-7.4) (Cairns, et al., 2006; Gerweck and Seetharaman, 1996). Acidosis results in part from the specially evolved metabolism of cancer cells, of which the most notable is the Warburg effect (i.e. the preference to use glycolysis to generate ATP, even in the presence of oxygen) (Kim and Dang, 2006), which leads to the accumulation of lactic acid in the cytoplasm and subsequent excretion of protons out of the cell. Other factors also contribute to extracellular acidity, such as the release of carbon dioxide from cancer cells (Helmlinger, et al., 2002), poor clearance of extracellular protons by intra-tumor microvessels (Gatenby and Gillies, 2004), and induction of carbonic anhydrase IX in solid tumor tissues (Gatenby and Gillies, 2004; Swietach, et al., 2007). For these reasons, acidosis is a hallmark of tumor progression from early to advanced stages and may provide an opportunity for tumor detection and targeted therapy (Junior, et al., 2007; Mrkvan, et al., 2005; Sethuraman, et al., 2006).
Many pharmaceutical agents must reach specific organelles in the interior of the cell to exert their therapeutic action, e.g. agents of gene therapy need to reach the nucleus and pro-apoptotic anticancer drugs need access to mitochondria (Torchilin, 2008). However, the hydrophobic barrier of cell membranes prevents drug carriers from entering cells unless an active transport mechanism is involved. In a number of cases, the strategy is to rely on the endocytotic pathways to internalize the drug carrier and its payload. Various cell-penetrating proteins or peptides have shown the ability to deliver therapeutic agents into cells, with their cargo payload ranging from small-molecules, peptides, and proteins to RNA and DNA polymers (Fischer, et al., 2005; Futaki, 2006; Henriques, et al., 2006; Martin and Rice, 2007; Wagstaff and Jans, 2006). Many of the these peptides contain fewer than twenty amino acids and are highly enriched in Arg or Lys, such as TAT (Vives, et al., 1997), penetratin (Derossi, et al., 1994), pVEC (Elmquist, et al., 2001) and other artificially designed peptides (Futaki, et al., 2001). However, when taken up by some form of endocytosis, peptide drug conjugates and other delivery nanoparticles are still topologically outside of the cell, with endosomal membranes separating the cargo molecules from the cytoplasm. Thus, additional devices, such as GALA peptides or the N-terminal domain of the influenza virus hemagglutinin protein HA2 (Li, et al., 2004; Stegmann, 2000), are employed to facilitate the endosomal escape. These efforts highlight the need for delivery systems that directly transport drug molecules into the cell cytoplasm.
As an alternative, our method is based on pHLIP (pH Low Insertion Peptide) - a peptide that inserts into lipid bilayers at acidic pH. Unlike other strategies, which may consist of a carrier component, a targeting agent and an endosomal escape moiety, the pHLIP peptide can simultaneously target tumors, carry the cargo and translocate the payload across the plasma membrane at low pH. pHLIP is a 36 amino acid peptide derived from the C-helix of bacteriorhodopsin (Hunt, et al., 1997). At pH above 7, pHLIP peptides are soluble as monomers in aqueous buffers and associate with lipid bilayer surfaces largely as unstructured peptides. Under acidic conditions, pHLIP inserts across a lipid bilayer with an apparent pK of 6 in vitro, forming a transmembrane helix (Hunt, et al., 1997). The mechanism and thermodynamics of pHLIP-membrane interaction have been studied extensively (Hunt, et al., 1997; Reshetnyak, et al., 2008; Tang and Gai, 2008; Zoonens, et al., 2008). At peptide concentrations below 7 μM, pHLIP is predominantly monomeric in all three states—in solution, on the membrane, and in the membrane (Reshetnyak, et al., 2007). The pH dependent insertion process is coupled to the protonation of one or both of the Asp residues located in the transmembrane region of the peptide (Hunt, et al., 1997; Reshetnyak, et al., 2006). Furthermore, pHLIP insertion seems to be reversible and unidirectional, with its C-terminus translocated across the membrane (Reshetnyak, et al., 2006; Reshetnyak, et al., 2007). Once inserted, pHLIP helices cause minimal disturbance to phospholipid bilayers: they do not form pores or induce membrane permeabilization (Reshetnyak, et al., 2006; Zoonens, et al., 2008).
We have shown that pHLIP can accumulate and persist in vivo in tissues with mildly acidic environments (pH 6.5-7.0), including solid tumors in mice and inflammatory foci in rats. By comparison, pHLIP accumulation is much less favorable in healthy tissues, with their extracellular milieu at a physiological pH of 7.2-7.4. In a mouse breast adenocarcinoma model, fluorescently labeled pHLIP constructs found solid acidic tumors with a high degree of accuracy and accumulated in them, even at very early stages of tumor development (Andreev, et al., 2007). In addition, pHLIP shows no toxicity to cells (Reshetnyak, et al., 2006) or mice (Andreev, et al., 2007) at concentrations tested (1-50 μM).
The insertion of pHLIP into lipid bilayers, leading to the formation of stable transmembrane helices, is associated with an energy release. We have established that this energy can be used to move cargo molecules across a membrane, raising the possibility of targeted drug delivery (Reshetnyak, et al., 2006). It is also worth noting that pHLIP-mediated translocation of cargo molecules into cells is not mediated by endocytosis, interactions with cell surface receptors, or formation of pores in cell membranes (Reshetnyak, et al., 2006).
However, there has not been any systematic exploration of the properties of cargo molecules that can be delivered. To this point we have only a scattering of data on the molecular characteristics that define what constitutes useable cargo molecules. Indeed, the properties of the cargos that have been previously shown to be translocated by pHLIP into the cytoplasm of cells, were not well known or carefully controlled (Reshetnyak, et al., 2006). For instance, when we described the phalloidin cargo as “cell-impermeable” we neglected the presence of the Rhodamine (TRITC) moiety (Log Po/w = +2.9; MW = 516) and the photocrosslinker (Log Po/w = +0.8; MW = 212, used to attached phalloidin-TRITC to pHLIP), resulting in a cargo molecule much less polar than the phalloidin alone, which has a Log Po/w of −2.9 (MW = 788). Similarly, the dansyl dye cargo has a calculated Log Po/w of +3, making it potentially cell-permeable on its own.
The goal of the present study is to define the useful range of chemical properties for cargo molecules deliverable by pHLIP-mediated translocation. To accomplish this objective, we take advantage of the fact that a wide range of properties, such as polarity, charge, size and shape, can be varied systematically in a host-guest model cyclic peptide. We show that pHLIP can successfully translocate cell-impermeable, polar cyclic peptides (Log Po/w = − 3; MW ~ 800) through the membranes of lipid vesicles and cancer cells. Hence, we suggest that pHLIP can deliver membrane impermeable molecules that are usually excluded from being tested as potential therapeutic agents, expanding the range of molecular alternatives.
We chose cyclic hexa-peptides as model cargo molecules in this study, to avoid ambiguities that might arise from alternative conformations that linear peptides might adopt in the various environments found in solution, at a membrane surface, or in a membrane interior. The cargo structures are shown in Figure 1A. These NBD-Cyclic-(X)4 peptides (NC(X)4) contain four X positions that are varied as Ser, Asp, Asn, or Arg residues, allowing adjustments of cargo polarity, as shown by their octanol/water partition coefficients (Log Po/w) calculated using QikProp (Jorgensen, Yale University) (Table 1). Overall, these cargo molecules were designed to be quite hydrophilic (Log Po/w < −2.5), with the objective of making them membrane-impermeable on their own. To follow the cargo and test for topology, we included a Lys residue carrying the NBD (7-Nitrobenz-2-Oxa-1,3-Diazole) fluorescence probe, and to serve as the point of attachment to the carrier pHLIP peptide via a S-S disulfide bond we have a SH group.
We utilized the dithionite (S2O4−2) quenching of NBD fluorescence to confirm that the NC(X)4 cargo peptides behave in accordance with our design and cannot cross a lipid bilayer on their own at the pH used to activate pHLIP translocation. For each cargo, NBD fluorescence was monitored at 530 nm (excited at 480 nm) when incubated with POPC (1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine) vesicles at pH 4.0. Figure 1B shows that addition of the membrane-impermeable dithionite ion leads to rapid, complete quenching of NBD fluorescence of all four NC(X)4 peptides, whereas no change of signal is observed when NBD is conjugated to the C-terminus of pHLIP in the control construct pHLIP-NBD (100% protection). These results indicate that the C-terminus of pHLIP is located inside the vesicles at pH 4, thus protecting NBD from dithionite quenching, while the NC(X)4 cyclic cargo peptides stay outside. Slow quenching of NBD fluorescence is sometimes observed on a time scale of minutes, as can be seen for the pHLIP-NBD conjugate, perhaps due to leakage of dithionite or its decomposition products (Lem and Wayman, 1970; Wayman and Lem, 1970) into the liposomes at pH 4. As mentioned in the Materials and Methods section, protection is evaluated from the first data point collected immediately after addition of dithionite (denoted by a circle in Figure 1B).
We also tested how NC(X)4 peptides would behave in the presence of human cancer cells in an acidic environment by monitoring NBD fluorescence via confocal microscopy. Figure 1C shows that these cargo peptides do not cross the membrane of HeLa cells when incubated at pH 6.2. Indeed, with the exception of a few dead cells, no fluorescence is observed in the cytoplasm of the cells. Thus, these cargo molecules cannot cross a lipid bilayer or a cell membrane on their own.
Prior to evaluating pHLIP-mediated translocation of cyclic peptides into lipid vesicles and cancer cells, each NC(X)4 cargo peptide was conjugated to the C-terminus region of pHLIP-Cys via a disulfide bond. The interactions of these pHLIP-cargo constructs with lipid bilayers were then studied using circular dichroism and tryptophan fluorescence.
Far-UV CD spectroscopy was used to determine the secondary structure of various pHLIP constructs in the presence of POPC liposomes at pH 8 and 4 (Figure 2A and 2B). Figure 2A shows that every pHLIP conjugate adopts an unstructured configuration at pH 8, whereas Figure 2B shows that all pHLIP conjugates form α-helices under acidic conditions similar to that formed by pHLIP at pH 4, except for pHLIP-NC(Arg)4. The latter remains unstructured when the pH is lowered from 8 to 4.
There are two Trp residues present in the sequence of pHLIP, and at least one of them is in the postulated transbilayer region of the peptide. Thus, since the Trp fluorescence emission is sensitive to the polarity of the environment, they serve as reporter groups for monitoring the interactions of pHLIP (or pHLIP conjugates) with lipid bilayers in POPC vesicles. At pH 8, the Trp fluorescence emission maxima of all pHLIP conjugates (except pHLIP-NC(Arg)4) are centered around 350 nm (Figure 2C), reflecting the exposure of Trp residues to polar, aqueous environments (Burstein, et al., 1973). Lowering the pH from 8 to 4 results in significant λmax blue shifts (≥ 10 nm, except for pHLIP-NC(Arg)4) and increases in fluorescence intensity (except for pHLIP-NC(Asn)4). These changes are characteristic for Trp residues buried in hydrophobic environments and suggest that at least pHLIP-NC(Asp)4 and pHLIP-NC(Ser)4 conjugates insert into vesicles at pH 4 (Burstein, et al., 1973; Reshetnyak, et al., 2007).
These findings agree with the pH-responsive changes in secondary structure revealed by CD. Taken together, notwithstanding the NC(Arg)4 exception, these results indicate that conjugating a NC(X)4 cargo peptide (or a NBD reporter group) to pHLIP does not disturb its characteristic shape-shifting behavior. Moreover, in very recent tryptophan fluorescence experiments, we measured the pK of insertion of pHLIP-NC(Asp)4 into the bilayer of POPC vesicles to be 5.9 ± 0.1 (data not shown), which is very close to the value of 6.0 measured for pHLIP alone (Hunt, et al., 1997). In other words, acidic environments can induce pHLIP conjugates to insert into lipid bilayers to form stable transmembrane α-helices in a fashion similar to that of pHLIP.
As observed in CD, pHLIP-NC(Arg)4 does not behave like the other pHLIP-conjugates: its Trp λmax is already centered around 340 nm at pH 8 (Figure 2C), reflecting sequestration of Trp residues away from water—a state possibly mediated by aggregation. Furthermore, the Trp λmax does not change significantly when the pH is lowered, consistent with the unstructured CD spectra observed at both pH 8 and pH 4 for this conjugate. In short, the pHLIP-NC(Arg)4 construct seems to aggregate.
We assessed the ability of pHLIP to translocate NC(X)4 cargos across lipid bilayers and cell membranes. The conjugates were first tested with lipid vesicles. Figure 3A shows that when pHLIP-NC(X)4 conjugates are incubated at pH 8 with POPC vesicles, addition of dithionite quenches NBD fluorescence in a rapid and complete fashion (0 % protection) for all four pHLIP-NC(X)4 conjugates, i.e. cargo molecules stay outside of the liposomes and are not translocated across lipid bilayers at neutral pH. However, when the pH is lowered to 4, different NBD fluorescence quenching traces are observed among the conjugates (Figure 3B). For pHLIP-NC(Asn)4 and pHLIP-NC(Arg)4 conjugates, NBD fluorescence are completely quenched by dithionite even at pH 4, with 0.4 ± 1.8 % and 3 ± 5 % protection observed respectively (see Table 2), suggesting that these two conjugates leave their cargos outside of the liposomes under acidic conditions. On the other hand, near complete protection from dithionite quenching is observed for pHLIP-NBD (91 ± 9 %, positive control), pHLIP-NC(Asp)4 (85 ± 7 %) and pHLIP-NC(Ser)4 (76 ± 8 %) conjugates, showing that these three C-terminal cargos, with Log Po/w values in the range of −0.8 to −2.9, are successfully translocated into or across lipid bilayers by pHLIP. The quantitative results for these translocation assays are summarized in Table 2. As already shown in figure 1B for pHLIP-NBD, slow quenching of NBD after addition of dithionite is also observed for pHLIP-NC(Ser)4 and pHLIP-NC(Asp)4, indicating that this effect is not specific to the (Ser)4 and (Asp)4 cargos and that these two cargos do not cause membrane distortion resulting in major liposome leakage.
We also evaluated the capability of pHLIP to deliver NC(X)4 cargos across membranes and into the cytoplasm of HeLa cells using NBD fluorescence microscopy. The results are in complete agreement with that observed in liposomes. Figure Figure3C3C and and44 show that NC(Asp)4 and NC(Ser)4 cargos are translocated and released into cell cytoplasms (since the S-S bond that conjugates the cargo to pHLIP is cleaved in cells) only when the cells are incubated at pH 6.2. On the other hand, pHLIP-NC(Asn)4 and pHLIP-NC(Arg)4 do not deliver their cargos into cell cytoplasms, regardless of the pH. In the case of pHLIP-NC(Arg)4, the conjugate seems to stick to cell plasma membranes and aggregate on the glass window at both pHs, further confirming what we observed by CD (Figure 2A and 2B) and Trp fluorescence (Figure 2C). This behavior is perhaps caused by interactions between the multiple positive charges on the cargo peptide NC(Arg)4 and the multiple negative charges on pHLIP (which has six carboxyl side chains) at neutral pH. In the case of pHLIP-NC(Asn)4, we suggest that it is the polarity of the cargo (Log Po/w ~ −8.5) that prevents pHLIP constructs from inserting across lipid bilayers (at an appreciable rate), since its CD and Trp fluorescence data are consistent with the notion that this pHLIP conjugate undergoes a clear transition from random coil to helix when pH is decreased from 8 to 4. Furthermore, the pHLIP-NC(Asn)4 construct seems to stick to the surface of HeLa cells at pH 6.2 as revealed by fluorescence microscopy (Figure 3C). Perhaps, this conjugate aggregates at or binds to the membrane surface as a α-helices, as CD data suggest (Figure 2B). It is also interesting that even though NC(Asp)4 is more hydrophilic than NC(Asn)4 at neutral pH (calculated Log Po/w of −11.9 vs. −8.5, see Table 1), the aspartic acid cargo is translocated into cells at pH 6.2 while the asparagine cargo is not, in agreement with what is observed in liposome quenching assays carried out at pH 4. Such behavior can be explained by the fact that the carboxylate side chains of NC(Asp)4 can be protonated (by cell surface protons), leading to a significant decrease in cargo polarity (from a calculated Log Po/w of nearly −12 to −2.9), making the crossing of the hydrophobic bilayer interior less formidable. The local environment at the cell surface (i.e. the phospholipid head group region) could significantly increase the pKa of the carboxylate side chains of the Asp residue from its canonical values of ~ 3.9, making the protonation more favorable at pH 6.2.
Moreover, even though NC(Ser)4 and NC(Asp)4 share similar Log Po/w values (about −3), they differ slightly in their molecular weights (742.7 vs. 854.8, respectively), yet pHLIP translocates them with the same efficiency (76 ± 8 % vs. 85 ± 7 %, respectively). This suggests that a Log Po/w of −3 (and MW of ~ 850) may not represent absolute set limits and that pHLIP may be able to translocate even more polar or larger molecules.
Remarkably, the results obtained with the lipid vesicles correlate very well with that found in the presence of cancer cells. This finding supports the idea that liposomes are appropriate models to study pHLIP-mediated translocations in cells and that pHLIP peptides traverse cell membranes via direct insertion into the bilayer, reinforcing the view that pHLIP-assisted translocation is not mediated by membrane proteins or endocytosis.
Two variants of pHLIP were used in this study: pHLIP-Cys with a cysteine residue at its C-terminus and Cys-pHLIP with a cystein at its N-terminus. They have the following sequences:
Both variants were prepared by solid-phase peptide synthesis and purified via reverse-phase HPLC at the W.M. Keck Foundation Biotechnology Resource Laboratory (Yale University).
The cargo molecules are synthetic cyclic peptides following a host-guest model, called NBD-Cyclic-(X)4 or NC(X)4, where X is Arg, Asn, Asp or Ser (Figure 1A). These cyclic hexa-peptides also contain a Lys where the fluorescence probe 7-nitrobenz-2-oxa-1,3-diazole (NBD) is attached, and a Cys residue to allow conjugation to pHLIP via a disulfide bond. These cyclic peptides were purchased from Anaspec, Inc. (San Jose, CA) and subsequently conjugated to pHLIP-Cys using methods described below. The Log Po/w (octanol/water partition coefficient) of these cyclic cargo peptides was calculated using QikProp 3.0 for Windows (W. Jorgensen – Dept. of Chemistry, Yale University) from the structures with deprotonated or protonated side-chain carboxyl groups.
Solvent mixture A consists of 1:1:3:5 of DMF, DMSO, methanol, and aqueous NH4HCO3 buffer (200 mM, pH 8). To a solution of TCEP.HCl (tris-(2-carboxyethyl)-phosphine, hydrochloride salt, 0.765 mg, 2.67 μmole, 3.38 eq.) in 85 μL of solvent mixture A was added a solution of cyclic peptide NC(Ser)4 (0.585 mg, 0.79 μmole, 1 eq.) in 340 μL of A, followed by a solution of pHLIP-Cys (3.6 mg, 0.85 μmole, 1.08 eq.) in 340 μL of A. This mixture was stirred at room temperature and in the dark for 20 min. To this reduced mixture, an oxidizing solution of K3Fe(III)CN6 (6 mg, 18.2 μmole, 23.1 eq.) in 60 μL of aqueous NH4HCO3 buffer (200 mM, pH 8) was added. This reaction mixture was stirred at room temperature for 5 h, and then allowed to sit at 0°C for 16 h. The desired product was isolated via reverse phase HPLC (Hewlett Packard Zorbax semi-prep 9.4 × 250 mm SB-C18 column; flow rate: 2 mL/min; phase A: water + 0.01% TFA; phase B: acetonitrile + 0.01% TFA; gradient: 70 min from 99:1 A/B to 1:99 A/B). Lyophilization provided the desired pHLIP-S-S-NC(Ser)4 adduct (0.4 μmole) in ~ 50% yield. The product was quantified using UV-vis absorbance of the NBD group (ε 280nm ~ 2,080 M−1cm−1, ε 348nm ~ 9,800 M−1cm−1, ε 480nm ~ 26,000 M−1cm−1) and pHLIP peptide (ε 280nm ~ 13,940 M−1cm−1). MS (MALDI-TOF MS+): Molecular weight calculated for pHLIP-S-S-NC(Ser)4 (C223H329N55O71S2): 4980.5. Found (MH+): 4980.6.
Other pHLIP-S-S-NC(X)4 conjugates were prepared in similar fashions but often employing a simplified, milder procedure in which the TCEP reduction step is omitted, thus allowing the use of less oxidizing reagent (K3Fe(III)CN6, 1-1.5 eq.). The desired adducts pHLIP-S-S-NC(X)4 were isolated in 30-40 % yields. In the case of NC(Arg)4, the reaction proceeded as a heterogeneous mixture, (perhaps due to aggregation between the multiple positive charges on NC(Arg)4 and the multiple negative charges on pHLIP). MS (MALDI-TOF MS+): Molecular weight calculated for pHLIP-S-S-NC(Asp)4 (C227H329N55O75S2): 5092.5. Found (MH+): 5100.6. Molecular weight calculated for pHLIP-S-S-NC(Asn)4 (C227H333N59O71S2): 5088.6. Found (MH+): 5084.5.
To a solution of pHLIP-Cys (3 mg, 0.7 μmole, 1 eq.) or Cys-pHLIP in 900 μL of argon saturated aqueous potassium phosphate buffer (100 mM, pH 7.25) was added a solution of IANBD amide (N,N′-dimethyl-N-(iodoacetyl)-N′-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)ethylenediamine, from Invitrogen, 0.3 mg, 0.7 μmole, 1 eq.) in 300 μL of DMF. This reaction mixture was stirred at room temperature in the dark under argon for 20 h. The desired adducts were isolated in > 85% yield via C18 HPLC using methods described previously. The products were quantified using the 498 nm absorbance of the secondary amine NBD group (ε 498nm ~ 23,500 M−1cm−1). MS (MALDI-TOF MS+): Molecular weight calculated for pHLIP-NBD (C208H306N50O62S): 4531.0. Found (MH+): 4531.4. Molecular weight calculated for NBD-pHLIP (C205H301N49O61S): 4459.9. Found (MH+): 4460.6.
Ten milligrams of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC; Avanti Polar Lipids; Alabaster, AL) in chloroform were dried under a stream of argon and then held under vacuum overnight. The dried lipid film was re-hydrated with 1 mL of pH 8.0, 100 mM sodium phosphate and 50 mM NaCl (buffer B) and vortexed. The resulting multilamellar vesicles were freeze-thawed in liquid nitrogen for seven cycles and were extruded through a polycarbonate membrane with 100 μm diameter pores using a mini extruder (Avanti Polar Lipids) to obtain unilamellar vesicles (LUVs). Liposomes were used immediately following their preparation. Lipid concentration was checked using Marshall's assay (Stewart, 1980).
Prior to circular dichroism (CD) or fluorescence measurements, stock solutions of pHLIP constructs (0.2-0.5 mM) were prepared in acetonitrile/water (1:1). The concentration of the stock solution was determined by UV-vis absorption in 6M guanidine-HCl (ε280 = 13,940 M−1 cm−1; ε480 = 25,000 M−1 cm−1). This stock solution was diluted to the working pHLIP concentration (of 2 or 7 μM) with buffer B, and incubated overnight at room temperature in the dark. On the next day, when appropriate, solutions of POPC liposomes (prepared in buffer B) were added to the diluted pHLIP solution at a 1:400 peptide-to-lipid molar ratio and incubated for 3 h at room temperature in the dark at pH 8.0. Subsequently, attempts at pHLIP insertion into the liposome lipid bilayer were triggered by reducing the pH from 8.0 to 4.0 with the addition of aliquots of a HCl solution. This mixture was allowed to equilibrate at pH 4 for 30 minutes prior to fluorescence and CD measurements.
Measurements were carried out using a SLM-Aminco 8000C spectrofluorimeter (ISS, Champaign, IL) equipped with a thermo-bath RTE-111 (Neslab). All measurements were performed at 25°C and with pHLIP peptide concentrations equal to 2 μM. To minimize light scattering effects, the emission polarizer was oriented at 0° and the excitation polarizer at 90°. The widths of the emission and excitation slits were set to 4 nm. When tryptophan emission was monitored (pHLIP has two Trp residues), the samples were excited at 295 nm and the emission spectra were taken from 300 to 500 nm. When monitoring NBD fluorescence, the samples were excited at 480 nm and the fluorescence signal was monitored at 530 nm in slow kinetic mode.
pHLIP-mediated cargo translocation into liposomes was determined by using the NBD-dithionite quenching reaction: Membrane-impermeable dithionite ion (S2O42−) can chemically modify the NBD fluorophore and quench its fluorescence (McIntyre and Sleight, 1991). Changes in the fluorescence signal of NBD were monitored upon addition of 5 μL of sodium dithionite (Na2S2O4) (pH 4.0, 1M sodium phosphate; final dithionite concentration in the cuvette = 0.5 – 2 mM) to a 95 μL sample of pHLIP construct in buffer B that had been incubated in the presence of liposomes at pH 4.0 for 30 min. Fresh dithionite stock solution was prepared for each measurement because dithionite decomposes rapidly at low pH (Lem and Wayman, 1970; Wayman and Lem, 1970). The percentages of protection are calculated from the first data point collected after addition of dithionite.
Far-UV CD spectra of pHLIP constructs were recorded on an Aviv model 215 spectrometer equipped with a Peltier thermal-controlled cuvette holder. All measurements were performed at 25°C and with pHLIP constructs prepared as mentioned above and at concentrations equal to 7 μM. CD intensities are expressed in Mean Residue Molar ellipticity [θ] calculated from the equation:
where θobs is the observed ellipticity in millidegrees, l is the optical path length in centimeters, c is the final molar concentration of the peptide, and n is the number of amino acid residues. Samples were measured in a 0.1 cm path length quartz cuvette and raw data were acquired from 260 nm to 190 nm at 1 nm interval using a 2 s averaging time, and at least two scans were averaged for each sample.
HeLa cells (kind gift from Dr. Mooseker, Yale University) were grown in 35-mm dishes with 12-mm glass-bottom window (WillCo Wells) in DMEM supplemented with 10% FBS, 100 units/mL penicillin and 0.1 mg/mL streptomycin in a humidified atmosphere of 5% CO2 at 37°C. After 48 h of culture, cells were washed twice with PBS buffer at the desired experimental pH and then incubated for 1h in PBS buffer at pH 7.4 or 6.2 in the presence of NC(X)4 or pHLIP-S-S-NC(X)4 (5 μM). After incubation the cells were washed with PBS buffer at the experimental pH and fresh DMEM media was then added. Fluorescent images were taken by using a Zeiss LSM 510 confocal microscope equipped with a 20x Plan-Apochromat (NA: 0.8) objective and with excitation at 488 nm. A series of optical planes (z-stack) were acquired using a 63x oil objective. Images were acquired by Zeiss' ZEN software and processed using ImageJ (Abramoff, et al., 2004) and the OrtView plug-in (MedNuc).
The authors thank Dr. Mark S. Mooseker and Dr. Joseph Wolenski (Department of Molecular Cellular and Developmental Biology, Yale University) for the kind gift of HeLa cells and help with confocal microscopy, respectively. We also would like to thank Dr. Yana K. Reshetnyak and Dr. Oleg A. Andreev (Physics Department, University of Rhode Island) for their insights and comments. This work was supported by the National Institute of Health grants GM073857 and CA133890, and by an Anna Fuller Fellowship to MA.
In conventional drug design aimed at intracellular targets, a constraint has been to create molecules that are sufficiently hydrophobic and/or small to traverse membranes on their own (Log Po/w from −0.4 to +5.6 and MW of 160 to 480 g.mol−1). As a result, cell-impermeable and larger molecules are usually excluded. Furthermore, many cancer drugs have off-target side effects that severely limit the efficacy of chemotherapy. In this work, we have explored the properties of pHLIP as a targeting and delivery system by defining some of the properties of molecules that can be translocated into cells, thus establishing boundary conditions for a new arena of drug design using this approach. Our method is based on pHLIP, a 36 amino-acid peptide that is soluble at physiological pH, and binds to membrane lipid bilayer surfaces if they are present. Unlike other strategies, it can simultaneously target tumors, carry the cargo, insert as a transmembrane helix and translocate the payload across the plasma membrane at low pH. We find that cargo molecules with Log Po/w of ~ −3 and molecular weights of ~ 800 are efficiently delivered into cancer cells. As a comparison, there is, to the best of our knowledge, no example of an existing anti-cancer drug with a Log Po/w < −1 and MW > 500 and that has intracellular targets. Thus, pHLIP may help to expand the chemical space for anti-cancer (and anti-inflammatory) drugs to include polar, cell-impermeable molecules aimed at cytoplasmic targets. Because such drug molecules will not easily cross cell membranes in normal tissues, certain of the side effects associated with conventional chemotherapy may also be mitigated.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.