European Journal of Pharmaceutical Sciences 67 (2015) 76–84
Contents lists available at ScienceDirect
European Journal of Pharmaceutical Sciences journal homepage: www.elsevier.com/locate/ejps
Selenium as an alternative peptide label – comparison to fluorophore-labelled penetratin Laura Hyrup Møller, Jesper Søborg Bahnsen, Hanne Mørck Nielsen, Jesper Østergaard, Stefan Stürup, Bente Gammelgaard ⇑ Department of Pharmacy, University of Copenhagen, Denmark
a r t i c l e
i n f o
Article history: Received 28 August 2014 Accepted 5 November 2014 Available online 14 November 2014 Keywords: Peptide quantification Cell-penetrating peptides Labelling techniques
a b s t r a c t In the present study, the impact on peptide properties of labelling peptides with the fluorophore TAMRA or the selenium (Se) containing amino acid SeMet was evaluated. Three differently labelled variants of the cell-penetrating peptide (CPP) penetratin (Pen) were synthesized, PenMSe, TAMRA-PenMSe and TAMRA-Pen. The labelled peptides were characterized in terms of hydrodynamic radius, secondary structure during peptide–membrane interaction, effect on membrane leakage induction, uptake efficiency in HeLa cells. Furthermore, stability of peptides and identities of degradation products in cell media and cell lysate were evaluated. TAMRA-labelling increased the hydrodynamic radius of Pen and PenMSe significantly. Labelling with Se caused no or minimal changes in size, secondary structure and membrane leakage induction in concentration levels relevant for cellular uptake studies. Similar degradation patterns of all labelled peptides were observed in HBSS media; degradation was mainly due to oxidation. Cellular uptake was significantly higher for the TAMRA labelled peptides as compared to Se-labelled Pen. Extensive degradation was observed in media during cellular uptake studies, however, in all cell lysates, primarily the intact peptide (PenMSe, TAMRA-PenMSe or TAMRA-Pen) was observed. Selenium labelling caused minimal alteration of the physicochemical properties of the peptide and allowed for absolute quantitative determination of cellular uptake by inductively coupled plasma mass spectrometry. Selenium is thus proposed as a promising alternative label for quantification of peptides in general, altering the properties of the peptide to a minor extent as compared to commonly used peptide labels. Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction Biopharmaceuticals are increasingly used as drug candidates, as they most often are highly potent and biocompatible due to their similarities to natural occurring biomolecules (Vlieghe et al., 2010). To elicit the desired pharmacologic response their target sites, which may be in the cellular plasma membrane or in the intracellular environment must be reached. However, due to their relatively large molecular size and structure, these drugs are poorly absorbed across biological membranes and most often labile to enzymatic degradation once dosed. To improve the therapeutic potential of the peptide drugs, drug carriers such as cell penetrating peptides (CPPs), e.g. penetratin, have been applied to facilitate delivery across eukaryotic cell membranes (Liu et al., 2014; Khafagy and Morishita, 2012; Farkhani et al., 2014). Unravelling the mechanism of how the CPP acts and how potential peptide drug ⇑ Corresponding author. E-mail address: [email protected] (B. Gammelgaard). http://dx.doi.org/10.1016/j.ejps.2014.11.001 0928-0987/Ó 2014 Elsevier B.V. All rights reserved.
cargoes are delivered is of great interest and tracking the peptides is crucial for design and development of future peptide-based drugs, but often difficult. Peptide quantification is often performed by high performance liquid chromatography (HPLC) and UV or fluorescence detection (FLD), exploiting cromophores and fluorophores within the peptide structure (Grohganz et al., 2004; Farin et al., 1998; Backes et al., 1998). However, these methods exhibit limited sensitivity and especially selectivity for analysis of peptides in complex biological sample matrices. In order to enhance sensitivity and selectivity, a variety of different labelling strategies have been employed. One of the currently most used approaches to study peptide delivery at a cellular level is to track the peptide labelled with a fluorophore (Liu et al., 2014; Bahnsen et al., 2013; Amand et al., 2008; Lundberg and Langel, 2006; Fischer et al., 2000; Elmquist et al., 2001; Lindgren et al., 2004). Fluorophore-labelled peptides are easily accessible and allow for direct measurement of the fluorophore with high sensitivity compared to the non-labelled peptide detection. However, labelling of relatively small peptides with relatively large and
L. Hyrup Møller et al. / European Journal of Pharmaceutical Sciences 67 (2015) 76–84
hydrophobic moieties affects the physicochemical properties of the parent peptide and thereby possibly its pharmacokinetic (PK) profile Jones and Sayers, 2012. In order to monitor parameters such as stability, kinetics and metabolism of the peptide and the peptide delivery system, other quantitative methods are thus needed. In addition to fluorescence detection, liquid chromatography mass spectrometry (LC–MS) has become the method of choice for analysis of peptides. LC–MS provides detailed information of the masses of peptides and possible degradation products. However, quantitative analysis demands for costly species specific standards (John et al., 2004; Burlina et al., 2005). Recently, LC hyphenated to inductively coupled plasma mass spectrometry (ICP-MS) has been applied for biomolecule analysis (Bettmer et al., 2009; Kretschy et al., 2012). The ICP–MS provides low limits of detection and a wide dynamic range. Using a hightemperature plasma, any compound is decomposed to single elemental ions providing in principle matrix-independent sensitivity to a variety of natural elements, and thus making quantification using a simple inorganic standard possible. Detection of the naturally occurring elements such as C, N, and O is, however, challenged due to interferences from the air, while analysis of the hetero element sulfur (S) is challenged by low ionization efficiency and spectral interferences in the ICP–MS (Prange and Profrock, 2008). Thus, a number of other elemental labels has been suggested for peptide analysis. Studies employing element-tagged immunoassays using lanthanide labels and subsequent analysis by ICP–MS have been reported (Careri et al., 2009; Baranov et al., 2002; Quinn et al., 2002; Mueller et al., 2014). Furthermore, element-tagged affinity labelling techniques have been pursued in means of associating organic mercury or metallocene tags to thiol-groups and iodination of tyrosine residues (He et al., 2013; Bomke et al., 2010). Recently, metal or element chelating agents functionalized with a reactive moiety targeting a specific group in the peptide, either the thiol group of Cys or the N- or C-terminal have been used for peptide labelling prior to ICP–MS analysis (Kretschy et al., 2012; Prange and Profrock, 2008; Koellensperger et al., 2009; Esteban-Fernandez et al., 2011). These elemental labels increase analytical selectivity and sensitivity considerably compared to natural occurring hetero element tags, however, association of a metal, lanthanide or a relatively large chelating moiety to the peptide structure may induce changes in physicochemical properties of the peptide as described for fluorescent labels. Recently, we introduced selenium as an internal detection label in peptides by exchange of a natural occurring methionine (Met) residue by the selenium containing analogue selenomethionine (SeMet) Moller et al., 2014. Selenium (Se) and S belong to the group of chalcogens in the periodic table and thus share similar properties (Wessjohann et al., 2007). Furthermore, SeMet is incorporated in competition with Met in human proteins (Whanger, 1986). In our previous work, Se-containing penetratin (PenMSe) was successfully synthesized with 1:1 labelling stoichiometry and the label allowed for peptide detection by ICP–MS providing detailed information of the fate of the peptide during cellular uptake study (Moller et al., 2014). The main advantage of labelling with Se is the very small change of the peptide as the Se-labelling is performed only by exchanging the S-containing amino acid Met with the Se-containing amino acid SeMet. It is therefore expected that the physicochemical properties are only insignificantly altered. Furthermore, the labelling procedure is highly controllable, as the SeMet is introduced in the synthesis of the new CPP candidates; thus ensuring 1:1 labelling stoichiometry, which is not always the case in metal-based tagging techniques, as these processes are more difficult to control (Kretschy et al., 2012). At the same time, Se is not as abundant as S in biological systems and the Se-labelled peptides are easily distinguished from reagents and endogenous cell components in cellular uptake studies.
77
The primary aim of the current work was to evaluate the possibility of using Se-exchange of S-containing amino acids as an approach for labelling peptides using penetratin (Pen) as a model compound. Furthermore, the objective was to compare the behaviour of the fluorophore 5(6)-carboxy-tetraemethyl-rhodamine (TAMRA) labelled Pen with the Se-labelled Pen and evaluate how the resulting different changes of physicochemical properties of the labelled peptides influence cell uptake studies. We present here a comparison of hydrodynamic radii, secondary structures, peptide–membrane interactions including membrane leakage induction, cellular uptake and peptide stability of three Pen peptide derivatives: selenium labelled Pen with SeMet incorporated instead of the Met-residue (PenMSe), Pen with TAMRA at the N-terminal (TAMRA-Pen), and an analogue labelled with SeMet as well as TAMRA (TAMRA-PenMSe) (Table 1). 2. Material and methods 2.1. Reagents Rink amide resin, coupling reagents and amino acid building blocks for solid-phase peptide synthesis as well as supplementary solvents and chemicals were acquired from Iris Biotech (Merktredwitz, Germany). L-Selenomethionine (SeMet) was obtained from TCI Europe (Zwijndrecht, Belgium). 5(6)-Carboxy-tetraemethylrhodamine (TAMRA) was obtained from Novabiochem (Darmstadt, Germany). Calcein was purchased from Sigma–Aldrich (Buchs, Switzerland) and the lipids: 1-palmitoyl-2-oleoyl-sn-glycero-3phosphocholine (POPC), cholesterol, 1-palmitoyl-2-oleoyl-snglycero-3-phospho-(10 -rac-glycerol) (POPG) were obtained from Avanti Lipids (Alabaster AL, USA). For mobile phases, glacial acetic acid, trifluoroacetic acid (TFA) and methanol were obtained from BDH Prolabo (VWR, Bie & Berntsen, Søborg, Denmark). 2.2. Peptide synthesis and labelling All peptides were synthesized by Fmoc-based solid phase peptide analysis (SPPS) employing a microwave-assisted automated synthesizer (CEM, Matthews, NC, USA). N-protection of H-SeMetOH with N-(9-fluorenylmethoxycarbonyloxy)succinimide (FmocOSU) and synthesis of Pen and PenMSe by use of fully automated synthesizer were carried out as previously described (Moller et al., 2014). N-terminal labelling was carried out manually with a mixture of 5(6)-carboxy-tetraemethyl-rhodamine (TAMRA; 2.5 eq.), diisopropylcarbodiimid (DIC; 2.5 eq.) and 1-hydroxybenzotriazole (HOBt; 2.5 eq.) in dimethylformamide (DMF) overnight at room temperature in a Teflon reactor (10 mL) with a polypropylene filter and kept under nitrogen. Peptides were purified at room temperature by preparative HPLC (Luna C18(2) column, 5 lm, 250 21 mm; Phenomenex, SupWare, Denmark) with UV detection at 280 nm. A linear gradient of 0–35% eluent B (H2O:MeCN 95:5 with 0.1% (v/v) TFA) in eluent A (H2O:MeCN 5:95 with 0.1% (v/v) TFA) within 20 min was applied. Flow rate was 20 mL min1. Amino acid sequences of the synthesized peptides are shown in Table 1. 2.3. Taylor dispersion analysis Taylor dispersion analysis (TDA) was performed on an Agilent 3D-capillary electrophoresis instrument (Agilent Technologies, Waldbronn, Germany) coupled to an Actipix D100 UV imaging detector (Paraytec, York, UK) as described by Ostergaard and Jensen (2009). The UV traces were recorded with Actipix control software version 1.4 (Paraytec, York, UK) at a wavelength of 214 nm. A fused silica capillary (Polymicro Technologies, Phoenix,
78
L. Hyrup Møller et al. / European Journal of Pharmaceutical Sciences 67 (2015) 76–84
Table 1 Structures, molecular weights and hydrodynamic radii (Rh) of the labelled penetratin peptide derivatives. Rh determined by Taylor dispersion analysis using 80 lM peptide solutions and given as mean ± SD; Pen and PenMSe n = 6, TAMRA-Pen n = 7 and TAMRA-PenMSe n = 10. Abbreviation
Sequence
Maverage
Mmonoisotopic
Rh (nm)
Pen PenMSe TAMRA-Pen TAMRA-PenMSe
RQIKIWFQNRRMKWKK-NH2 RQIKIWFQNRR(MSe)KWKK-NH2 TAMRA-RQIKIWFQNRRMKWKK-NH2 TAMRA-RQIKIWFQNRR(MSe)KWKK-NH2
2245.74 2292.64 2659.19 2706.08
2244.31 2292.25 2657.46 2705.40
1.25 ± 0.03 1.23 ± 0.03 1.36 ± 0.03 1.33 ± 0.02
AZ, USA) coated with polydiallyldimethylammonium chloride (PDMAC) was prepared for the dispersion measurements. The total length of the capillary was 100 cm with 75 lm ID, and with two detection windows along the capillary, located after 30 and 50 cm, respectively. The running buffer was 67 mM phosphate buffer pH 7.4. Peptide samples were introduced into the capillary by applying a pressure of 50 mbar for 7 s (34 nL) followed by a mobilization pressure of 50 mbar (flow rate 1.1 mm/s) to force the sample plug through the capillary. The temperature of the cassette was 25 °C and each sample was analyzed in six to ten replicates. Apparent diffusion coefficients of the peptides were determined from the peak appearance times and the variances of the Gaussian shaped peaks according to Eq. (1) Ye et al., 2012:
D¼
R2C ðt 2 t1 Þ 24ðr22 r21 Þ
ð1Þ
where Rc is the radius of the capillary, t1 and t2 are the peak appearance time at the first and second detection window, respectively, and r21 and r22 are the temporal variances of the concentration profile at the first and second detection window, respectively. The hydro-dynamic radii (Rh) of the peptides were subsequently calculated from the Stokes–Einstein Eq. (2):
Rh ¼
kB T 6pgD
2.6. Calcein release assay To measure leakage from calcein-loaded liposomes, 180 lL suspension of these in Tris buffer (10 mM, pH 7.4) containing 150 mM NaCl, 0.1 mM EDTA and 70 mM calcein were added to black 96well plates (MicroWell 96 optical bottom plates, NUNC, Roskilde, Denmark) at a total lipid concentration of 20 lM. Fluorescence was monitored with a FLUOstar OPTIMA plate reader (GMB Labtech, Offenburg, Germany) at an excitation wavelength of 485 nm and an emission wavelength of 520 nm to determine initial fluorescence (F0). A volume of 20 lL of peptide solution in Tris buffer (10 mM, pH 7.4) containing 150 mM NaCl, 0.1 mM EDTA and 70 mM calcein (0.005–20 lM) was added and measurement of calcein leakage was initiated within 10 s from the first addition of peptide, and the peptide-induced calcein leakage was recorded for a period of 840 s in order to determine the fluorescence (F). Finally, the liposomes were disrupted completely by addition of 20 lL 20% (v/v) Triton X-100 (Sigma–Aldrich, Buchs, Switzerland), and the signal from the total amount of calcein was recorded (Ft). The relative release was calculated according to Eq. (3):
Calcein leakage ð%Þ ¼
ðF F 0 Þ 100% ðF t F o Þ
ð3Þ
ð2Þ
where kB is Boltzmann’s constant, T is the absolute temperature, and g is the solution viscosity. 2.4. Liposome preparation Liposomes were prepared by using a procedure described previously (Bahnsen et al., 2013). In brief, liposomes were prepared from POPC, POPG and cholesterol in a molar ratio of 5:3:2. For liposomes used for circular dichroism (CD) the lipid film was dispersed in Tris buffer (10 mM, pH 7.4), whereas calcein-loaded liposomes were formed in Tris buffer (10 mM, pH 7.4) containing 150 mM NaCl, 0.1 mM EDTA and 70 mM calcein. Excess calcein was removed by passing the liposomes through a Sephadex G-50 column (GE Healthcare, Little Chalfont, UK). The size of the liposomes was measured on a Zetasizer Nano ZS (Malvern, Worcestershire, UK) as described previously (Bahnsen et al., 2013). The total phosphor concentration of liposomes was determined by ICP–MS as previously described by Nguyen et al. (2011). 2.5. Circular dichroism (CD) spectroscopy CD spectra were measured as described previously (Bahnsen et al., 2013). In brief, peptide and liposomes were scanned 5 consecutive times in the range 190–260 nm in a quartz cell with 1 mm light path. The scans were averaged and the background contribution from solvent and/or liposomes was subtracted. An Olis DSM 10 spectrophotometer (Olis, Bogart, GA, USA) was used for all measurements. CD spectra were recorded of solutions containing 20 lM of peptide in Tris buffer (10 mM, pH 7.4), 95% trifluoroethanol (TFE) (Acros organics, Geel, Belgium) and in the presence of POPC:POPG:cholesterol (molar ratio 5:3:2) liposomes.
2.7. Cell culture Cell culturing was performed as described previously (Moller et al., 2014). In brief, the human cervical cancer cell line HeLa WT, from American Type Culture Collection (ATCC, LGC Standards AB, Boras, Sweden) was used. The cells were maintained in supplemented Eagle’s minimal essential medium (EMEM) and grown in an atmosphere of 5% CO2 at 37 °C. One day prior to the uptake experiments, 7.6 104 cells/well were seeded on the bottom of 24 well plates (Fischer Scientific, Slangerup, Denmark). 2.8. Cell uptake studies Experiments were performed 22–24 h after seeding of the HeLa WT cells; when a sub-confluent monolayer (80–90% confluence) was achieved. Peptide solutions in Hanks Balanced Salt Solution (HBSS; Gibco, Paisley, UK) supplemented with 10 mM of 4-(2hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES, Sigma Aldrich, St. Lois, US) were diluted immediately prior to use in cell uptake studies from aqueous 200 lM peptide stock solutions stored at 4 °C. Initiation of the uptake experiments was carried out by washing the cells once with 37 °C phosphate buffered saline (PBS, Sigma Aldrich, St. Louis, MO, US) and subsequent addition of 400 lL of peptide solution. Cells were incubated with the peptide solution for 2 h on an orbital shaker (90 rpm) at 37 °C. After exactly 2 h, medium was collected and immediately acidified with acetic acid to a final concentration of 2% (v/v). Cell uptake was terminated by washing the cells four times with ice-cold PBS. The cells were lysed by addition of 100 lL 0.1% (v/v) Triton-X 100 + 2% (v/v) acetic acid and kept on ice for 10 min. The lysis mixture was transferred to low-binding Eppendorf tubes (Alpha Laboratories, Hampshire, UK), centrifuged at 25,000g for 10 min and the supernatant
L. Hyrup Møller et al. / European Journal of Pharmaceutical Sciences 67 (2015) 76–84
79
collected for analysis. All samples were stored at 18 °C until analysis.
areas of TAMRA-PenMSe solution standardized from the Se-content by FI–ICP–MS.
2.9. Determination of total selenium by flow injection inductively coupled plasma mass spectrometry (FI–ICP–MS)
2.12. Identification of species by liquid chromatography electrospray ionization mass spectrometry (LC–ESI–MS)
The total amount of selenium uptake in HeLa cells was determined by flow injection (FI) analysis and ICP–MS detection. The measurements were performed on a PE Sciex ELAN 6000 ICP–MS (Perkin Elmer, Norwalk, CT, USA) equipped with a MicroMist glass nebulizer (AHF, Lab Support Hillerød, Denmark) and a PC3 cyclonic spraychamber (Elemental Scientific, Omaha, NE, USA) operated at 4 °C. The sampler and skimmer cones were made of nickel. On a daily basis, the nebulizer gas flow, lens voltage and RF power were optimized with a solution containing 100 lg L1 Se in mobile phase. The data acquisition settings were dwell time 50 ms and sweeps per reading 1. The carrier fluid for FI analysis was an aqueous solution containing 0.02% (v/v) TFA, 0.1% (v/v) acetic acid and 5% (v/v) MeOH. The flow rate was 200 lL min1 and the injection volume was 10 lL. Quantification was performed by single standard calibration based on peak areas of a 50 lg L1 solution of certified selenium element standard (1010 mg Se L1, PlasmaCAL, SCP Science, Quebec, Canada). This method was validated by numerous linear standard curves of the standard. 77Se+, 78Se+ and 82Se+ isotopes were monitored; 82Se+ was used for quantification.
LC–ESI–MS analysis was performed on a Dionex Ultimate 3000 UHPLC (Thermo Scientific, Waltham, MA) equipped with a degasser, quaternary pump, autosampler, thermostated column compartment and a diode array detector (DAD) and hyphenated to a Q Exactive Orbitrap mass spectrometer (Thermo Scientific, San Jose, CA). The column and mobile phase prepared as described for LC–ICP–MS analysis were applied for this system as well. MS data were collected between 2 and 15 min run time. The Q Exactive Orbitrap mass spectrometer was run with 250 °C heated electrospray ionization. The sheath gas was set to 30 arbitrary units; auxiliary gas to 10 arbitrary units; capillary temperature to 350 °C; and spray voltage to 3300 V. The MS was run in full scan positive mode. Scan range was 300–3000 m/z, with lock mass 445.120 m/ z, and the mass accuracy was below 5 ppm. The instrument was mass calibrated prior to analysis.
2.10. Speciation analysis by liquid chromatography inductively coupled plasma mass spectrometry (LC–ICP–MS) LC–ICP–MS analysis was carried out using an Agilent 1100 series HPLC system equipped with a degasser, a quaternary pump, an auto sampler and a column oven. The column was an Aeris PEPTIDE XB-C18, 3.6 lm, 100 2.1 mm ID column protected by a C18-peptide SecurityGuard ULTRA (Phenomenex, SupWare, Torrance, CA). A linear gradient of 20–80% (v/v) MeOH, (with 0.1% (v/v) acetic acid and 0.05% (v/v) TFA added) within 10 min was applied with a flow rate of 200 lL min1, a column temperature of 60 °C, and injection volume of 5 lL. The ICP–MS instrument applied was similar to the one described for FI–ICP–MS analysis. In order to remove MeOH prior to introduction to the ICP–MS, the membrane desolvation system Aridus II (CETAC Technologies, Omaha, NE, USA) was applied in front of the ICP–MS. Aridus II was equipped with a 200 lL min1 nebulizer and the settings were: spray chamber temperature 110 °C; desolvator temperature 160 °C; sweep gas flow 7 mL min1, and nitrogen gas flow 6 mL min1. ICP–MS nebulizer gas flow was 0.9 mL min1. Lens voltage and ICP RF power were optimized regularly using a solution of PenMSe containing 10 lg L1 Se in 50% MeOH (v/v). Data acquisition settings for speciation analysis: dwell time 200 ms; sweeps per reading 1; and readings per replicate 1325. 77Se+, 78Se+ and 82Se+ isotopes were monitored; 82Se+ was used for data analysis. 2.11. Speciation analysis by liquid chromatography fluorescence detection (LC–FLD) LC–FLD analysis was performed using the same HPLC system as described for LC–ICP–MS connected to an Agilent 1200 series fluorescence detector (G1321A FLD). The chromatographic settings were equivalent to the method described under 2.10 and the settings of the fluorescence detector were extinction wavelength 543 nm and emission wavelength 572 nm. Total amount of TAMRA taken up by cells after exposure to TAMRA-PenMSe or TAMRA-Pen were determined by single standard calibration based on peak
3. Results and discussion In the present study, the impact on peptide properties of labelling peptides with the fluorophore TAMRA or the selenium containing amino acid SeMet was evaluated. Three differently labelled variants of the cell-penetrating peptide (CPP) penetratin (Pen) were synthesized, PenMSe, TAMRA-PenMSe and TAMRA-Pen (Table 1). The labelled peptides were characterized in terms of hydrodynamic radius, secondary structure during peptide– membrane interaction, effect on membrane leakage induction, uptake efficiency in HeLa cells and evaluation of the species taken up by the cells including stability in the media during the cell uptake study. 3.1. Taylor dispersion analysis In order to determine the influence of Se- and/or TAMRA-labelling on the size of the peptide, the hydrodynamic radii (Rh) of native Pen, PenMSe, TAMRA-Pen and TAMRA-PenMSe were determined by Taylor dispersion analysis (TDA). The results are presented in Table 1. Association of TAMRA to Pen and PenMSe increased Rh significantly. Increase of radius due to labelling with TAMRA was expected, as TAMRA (Mw = 431.2 Da) provide a relatively large addition of weight and thus size to Pen and PenMSe (Mw in Table 1). On the other hand, no significant difference in Rh was observed by substitution of Met with SeMet, neither between Pen and PenMSe nor between TAMRA-Pen and TAMRA-PenMSe. Labelling with selenium only increased the molecular weight of the peptide with 46.9 Da. Furthermore, the covalent radius of selenium (116 pm) Wessjohann et al., 2007 is only slightly larger than for sulfur (102 pm) Wessjohann et al., 2007, and was not expected to change the size of the peptide significantly. The results confirmed that incorporation of SeMet instead of Met cause minimal disturbance according to size of the peptides. 3.2. Peptide–membrane interactions Circular dichroism (CD) experiments were performed in order to evaluate the influence of selenium labelling on the secondary structure of Pen in solution and when interacting with liposomes of POPC:POPG:cholesterol (molar ratio 5:3:2), a membrane composition similar to the eukaryotic membrane (Zhang et al., 2010). The average size (diameter) of the liposomes was 90 nm (PDI = 0.14).
80
L. Hyrup Møller et al. / European Journal of Pharmaceutical Sciences 67 (2015) 76–84
CD spectra of Pen and PenMSe measured in Tris buffer, trifluorethanol (TFE) and when interacting with liposomes POPC:POPG:cholesterol (5:3:2) are shown in Fig. 1. In buffer both of the peptides showed unstructured conformation, whereas an a-helical structure for both peptides was observed in 95% (v/v) TFE as indicated by the minima observed at 208 nm and 222 nm. These observations are in agreement with previously reported CD spectra for Pen (Czajlik et al., 2002). Overall, no considerable structural difference between Pen and PenMSe was detected, which indicates that the exchange of Met with SeMet did not cause any changes in the ability of the peptide to form a-helical structure. The observations for native Pen is in accordance with previous reports on the folding of Pen in an apolar environment (Bahnsen et al., 2013; Persson et al., 2004; Caesar et al., 2006). The native Pen showed a pronounced a-helix forming propensity in the presence of liposomes. This was slightly higher as compared to the folding of PenMSe in the presence of liposomes as illustrated by the lower minima at 208 nm and 222 nm for Pen. The differences observed between Pen and PenMSe may indicate, that the interaction with liposomes and thus possibly with cellular membranes are slightly different for PenMSe as compared to Pen. Differences in the degree of folding may also be explained by differences in the bond length Se–C (196 pm) and S–C (180 pm) Wessjohann et al., 2007. For proteins, the structure most often retains sufficient flexibility and plasticity to accommodate the changes from SeMet without any changes in geometry (Moroder, 2005) and the differences observed in the present work may be explained by less flexibility in the relative small peptide structure of Pen compared to proteins.
3.3. Effect of PenMSe and Pen on membrane leakage induction The effect of Pen and PenMSe on membrane leakage was investigated by the release of dye from calcein-loaded POPC:POPG:cholesterol (5:3:2) liposomes similar to those used for CD. The average size of the calcein-loaded liposomes was 129 nm (PDI < 0.1). The calcein-release was investigated in a lipid concentration of 20 lM and peptide concentrations in the range 0.005–20 lM (peptide:lipid ratios of 1:4000 to 1:1) (only data for 1.25 lM and 20 lM are shown). For peptide concentrations up to 1.25 lM (peptide:lipid ratio 1:16), no difference between Pen and PenMSe in calcein release and the kinetic release profile of calcein from liposomes was observed. Above 1.25 lM the release of calcein from liposomes increased remarkably when exposed to PenMSe while Pen caused no distinct increase of calcein release (Fig. 2). These results corroborate with the fact that Pen is known not to be pore-forming and thus, no release of calcein was expected (Caesar et al., 2006). Further studies will be necessary, in order to explain the increased release by higher concentration of the Se-containing Pen. Peptide:lipid ratios corresponding to higher concentrations than 1.25 lM of peptide (peptide:lipid 1:16), are assumed to be irrelevant in relation to cellular in vitro studies. Thus, the observed difference in membrane leakage induction from PenMSe and Pen are not expected to contribute to any disturbances in the cellular uptake studies of this work. Hence, labelling with selenium seems to constitute a promising method of labelling Pen, however, further studies on membrane interactions would be valuable. 3.4. Purity and stability of synthesized peptides Purity of the synthesized peptides was determined by LC separation and either fluorescence or ICP–MS detection. Chemical stability of the synthesized peptides was evaluated in aqueous stock solution and after 2 h incubation in HBSS at ambient temperature. The Se-containing peptides (PenMSe and TAMRA-PenMSe) were analyzed by LC–ICP–MS, while the TAMRA-labelled peptides (TAMRA-Pen and TAMRA-PenMSe) were measured by LC–FLD. The chromatograms of TAMRA-PenMSe from FLD and ICP–MS detection are shown in Fig. 3. Similar peak patterns were observed by the two complementary detection methods comprising two main peaks in each chromatogram. LC–FLD analysis of TAMRA-Pen in water and HBSS showed similar peak pattern as for TAMRA-PenMSe, however, TAMRA-Pen was not degraded to the same extent as TAMRAPenMSe (Supplementary S1). The compounds corresponding to the smaller peaks may be a result of the synthesis procedure (impurities) or due to degradation during incubation with HBSS (degradation products). By LC–ESI–MS, identities of the peptides corresponding to the main peaks in the chromatograms were determined. The highest peaks corresponded to the intact peptides, TAMRA-PenMSe and
Fig. 1. Circular dichroism spectroscopy of Pen (solid line) and PenMSe (dashed line) in presence of (a) buffer, (b) TFE or (c) POPC:POPG:Cholesterol (molar ratio 5:3:2) liposomes. Measurements were performed with a lipid concentration of 2 mM and a peptide concentration of 20 lM resulting in peptide to lipid molar ratio of 1:100.
Fig. 2. Release kinetics of calcein from POPC:POPG:cholesterol (molar ratio 5:3:2) liposomes in presence of 1.25 lM and 20 lM Pen and PenMSe and control without peptide (grey). Results are given as mean ± SD, n = 3.
81
L. Hyrup Møller et al. / European Journal of Pharmaceutical Sciences 67 (2015) 76–84
Fig. 3. LC–ICP–MS and LC–FLD chromatograms of TAMRA-PenMSe in diluted aqueous stock solution (water) and after incubation in HBSS at room temperature for 2 h.
TAMRA-Pen, respectively. The smaller peaks corresponded to a compound with a mass increase of 15.995 Da (MS-data in supplementary S2) indicating oxidation of the peptide. This oxidation probably occurred for Se (SeO) in the SeMet residue of TAMRAPenMSe and for S in the Met residue of TAMRA-Pen as those amino acids are easily oxidized to selenooxide and sulfoxide when exposed to oxygen (Moroder, 2005). The chemical degradation patterns of TAMRA-PenMSe and TAMRA-Pen may thus be considered similar under the investigated conditions. Furthermore, oxidation of PenMSe in cell culture media (HBSS) was recently reported (Moller et al., 2014). The distribution in percentages of intact peptide, major impurity/degradation product and minor impurities/ degradation products of PenMSe, TAMRA-PenMSe and TAMRA-Pen in aqueous solution and after incubation in HBSS calculated from the areas in the chromatograms are shown in Table 2. While the ICP–MS detection reveals Se-containing impurities (Fig. 4a), the FLD detection reveals unspecific fluorescent impurities (Fig. 4b). The initial purity of the synthesized labelled peptides, was similar for TAMRA-Pen (92 ± 3.0%) and PenMSe (90 ± 1.7%), indicating that labelling with TAMRA introduced about 8% fluorescent impurities while incorporation of SeMet introduced about 10% Se-containing impurities (similar impurities due to oxidation of S may be present in Pen, but this has to our knowledge never been examined). Thus, the TAMRA-PenMSe contained only 84% Se-intact peptide. Addition of SeMet to the peptide sequence will increase the propensity of oxidation of the peptide, as selenium is more readily oxidized as compared to sulfur (Wessjohann et al., 2007). This was reflected by the 9% oxidized peptide in PenMSe. However, the native TAMRA-labelled peptide TAMRA-Pen also contained about 5% impurity corresponding to an oxidation product. The oxidation may have occurred during the automated synthesis owing to exposure to high temperatures and microwave-irradiation in the presence of air. Peptide purity above 90% is often used for studies and should be expected for the final peptide products, however, lower purity of TAMRA-PenMSe was accepted as this study was
Fig. 4. Chromatograms of lysates of HeLa WT cells incubated with TAMRA-PenMSe or TAMRA-Pen in HBSS for 2 h. (a) LC–ICP–MS of TAMRA-PenMSe, (b) LC–FLD of TAMRA-PenMSe (black) and TAMRA-Pen (grey).
performed as a proof of concept study and thus, only a limited amount of peptide was synthesized. The purity of the final peptides may be improved by improving the resolution of the preparative
Table 2 Percentage of intact peptide, major and minor impurities/degradation products. Results were obtained by LC–ICP–MS and LC–FLD of peptides in aqueous solution and HBSS medium, both incubated at room temperature for 2 h. Percentages given as mean ± SD of three replicates unless otherwise stated. ICP–MS
PenMSe standard* PenMSe in HBSS TAMRA-PenMSe standard** TAMRA-PenMSe in HBSS TAMRA-Pen standard** TAMRA-Pen in HBSS** * **
n = 9, three different dilutions. n = 2.
FLD
Peptide (%)
Major (%)
Minor (%)
Peptide (%)
Major (%)
Minor (%)
90 ± 1.7 82 ± 0.8 84 ± 2.2 70 ± 1.5 – –
9 ± 1.4 17 ± 0.3 16 ± 0.7 30 ± 1.8 – –
1 ± 0.1 1 ± 0.1 – – – –
– – 74 ± 6.9 52 ± 0.3 92 ± 3.0 78 ± 2.1
– – 19 ± 6.2 41 ± 1.5 5 ± 0.7 17 ± 0.5
– – 6 ± 0.8 6 ± 1.8 3 ± 2.3 5 ± 1.5
82
L. Hyrup Møller et al. / European Journal of Pharmaceutical Sciences 67 (2015) 76–84
HPLC method, and thus separate the peptide and the major impurity. However, complete avoidance of oxidation will probably demand for absolute oxygen free environments. The chemical degradation observed for all peptides after 2 h incubation in HBSS was mainly due to further oxidation. This may have been caused by a combination of the exposure to oxygen from the air during incubation and the increased ionic strength of HBSS solution as compared to water. The amount of degradation product observed subsequent to incubation was similar (about 17%) for PenMSe and TAMRA-Pen, whereas the amount of major degradation product for TAMRA-PenMSe increased from 16% to 30% determined by ICP–MS and 19% to 41% determined by FLD. Degradation of fluorophore-labelled Pen (5(6)-carboxyfluorescein) in HBSS buffer was previously reported by Tréhin and colleagues; however, the identity of the degradation product was not described (Trehin et al., 2004). Thus, chemical degradation apparently occurs no matter which type of label is associated to Pen. However, from the current knowledge it is not possible to determine if the degradation is influenced by the label or solely due to transformation in the peptide structure. It must be noted, that this study only addresses the chemical degradation of the peptide and does not consider the stability of the fluorophore in terms of excited-state lifetime, sensitivity to photo bleaching and pH sensitivity. 3.5. Determination of total selenium in cell lysates by FI–ICP–MS and total fluorescence by LC–FLD In order to compare the impact of TAMRA labelling on the cellular uptake of PenMSe in HeLa WT cells, the total uptake of selenium was determined by FI–ICP–MS. SeMet was included as control in the experimental setup, as SeMet has formerly been reported to be taken extensively up by e.g. Caco-2 cells (Gammelgaard et al., 2012). The cellular uptake in HeLa-cells was calculated as pmol pr. 104 cells (Table 3). Significantly more selenium from TAMRA-PenMSe (52.0 ± 0.4 pmol/104 cells) was taken up by the HeLa-cells as compared to PenMSe (21.7 ± 2.5 pmol/104 cells). These results were obtained even though the initial selenium concentration of TAMRA-PenMSe was only 352 pmol/104 cells compared to 503 pmol/104 cells for PenMSe, thus demonstrating increased uptake efficiency of TAMRA-PenMSe. This indicates, that association of the fluorophore TAMRA to PenMSe significantly increased the uptake compared to the uptake of PenMSe itself. Normalization of the uptake amount to the initial amounts of peptide under the given conditions emphasizes this observation. In the cell lysate, 4% of the initial amount of selenium from PenMSe was detected whereas 15% of the initial amount of selenium from TAMRA-PenMSe was observed. In order to determine if Se-labelling influenced the uptake of TAMRA-labelled Pen, the total uptake of fluorophore from TAMRA-PenMSe and TAMRA-Pen was determined by integration of the total area of LC–FLD chromatograms of cell lysates. The total amount of fluorophore detected in lysates exposed to TAMRAPenMSe was consistent with selenium uptake from TAMRA-PenMSe
measured by FI–ICP–MS as presented in Table 3. This indicates that the intact doubly-labelled peptide is taken up by the cells, and not only the TAMRA-label or a selenium containing degradation product. For TAMRA-Pen, the total fluorescence was 90.0 ± 15.5 pmol/ 104 cells. Normalization of the FLD results for uptake of TAMRAPen to the initial concentration of peptide under the given conditions showed, however, that 18% of TAMRA-Pen was taken up by the cells. This result is not significantly different from the results obtained for TAMRA-PenMSe corresponding to 15% based on Se as well as fluorophore measurements (Table 3). In summary, the results showed that TAMRA labelling increased the uptake of labelled Pen significantly compared to Se-labelled Pen. Labelling with different fluorophores on different positions of the peptides has been the subject of several studies and it is generally accepted that addition of fluorophore may change the cellular uptake (Jones and Sayers, 2012). Fischer et al. investigated the influence of cellular uptake by labelling with either the anionic carboxyfluoroscein (CF) or the zwitter-ionic TAMRA, N-terminally or by addition of an extra Lys-residue within the peptide sequence linked to the fluorophore. The cellular uptake of fluorophore was determined by flowcytometry (Fischer et al., 2002). In that study, the total uptake efficiency of N-terminal-labelled Pen was similar regardless of fluorophore, while considerable difference in uptake was observed depending on the location of the fluorophore, suggesting, that the uptake efficiency was more related to the overall structure of the peptide than the fluorophores (Fischer et al., 2002). However, both fluorophores are large molecules and may both have influenced uptake considerably. These results support our findings, that cellular uptake might be dependent on the label and likely also its location. The association of a label to the peptide structure may thus, alter the peptide function and membrane interaction, whereas exchange of single amino acid residues in fluorophore-labelled peptides may be less relevant for cellular uptake. For example, exchange of the Met residue in Pen with e.g. Leu has previously been shown by measurement of total fluorescence not to affect the amount of fluorophore-labelled peptide taken up by cells (Bahnsen et al., 2013; Fischer et al., 2000). This supports the findings from the CD and calcein release experiments conducted in the present study that PenMSe only altered the secondary structure and membrane interaction with liposomes to a minor extent compared to Pen. In addition, studies have shown that decreasing the hydrophobicity of the N-terminal of CPPs may decrease the cellular uptake (Elmquist et al., 2006). As the TAMRA-labelled peptides appear later in reversed phase HPLC chromatograms than PenMSe and Pen, the TAMRA-labelled peptides are more hydrofobic than the native Pen and PenMSe. Thus, the increase hydrofobicity of the TAMRA-labelled peptides may explain the increased uptake of these Pen peptide derivatives. 3.6. Speciation analysis of cell lysates by LC–ICP–MS and LC–FLD Determination of the total selenium uptake in HeLa cells showed that considerable more selenium penetrated the HeLa cells when exposed to TAMRA-PenMSe compared to PenMSe. In order to
Table 3 Cellular uptake in HeLa WT cells after 2 h incubation with peptide. Total selenium was determined by FI–ICP–MS and total fluorophore was determined by LC–FLD. The amounts are given as mean ± SD (PenMSe and TAMRA-PenMSe n = 9 lysates, SeMet n = 6 lysates). FLD: uptake amounts given as mean ± SD (TAMRA-PenMSe n = 3 lysates, TAMRA-Pen n = 4 lysates). ICP–MS
PenMSe TAMRA-PenMSe TAMRA-Pen SeMet
FLD
Applied pmol/104 cells
Uptake pmol/104 cells
Normalized %
Applied pmol/104 cells
Uptake pmol/104 cells
Normalized %
503 352 – 508
21.8 ± 4.1 52.0 ± 1.4 – 57.0 ± 2.4
4 ± 0.8 15 ± 1.0 – 11 ± 1.7
– – 494 –
– 51.7 ± 6.4 90.0 ± 15.5 –
– 15 ± 1.8 18 ± 3.1 –
L. Hyrup Møller et al. / European Journal of Pharmaceutical Sciences 67 (2015) 76–84
83
describe the compounds taken up by the cells, cell lysates were analyzed by LC–ICP–MS and LC–FLD (Fig. 4). For both of the TAMRA-labelled peptides, the major compounds in the lysates were the intact peptides. LC–ESI–MS analysis confirmed the masses of intact peptide at the given retention times. LC–ICP–MS chromatograms (Fig. 4a) of cell lysate from cells exposed to TAMRA-PenMSe showed the intact peptide and a minor compound (tr = 11.1 min). LC–FLD revealed several other minor peaks (Fig. 4b). However, for both peptides the major peak corresponds to the intact peptide (TAMRA-PenMSe 58 ± 2.0% and TAMRA-Pen 69 ± 1.8% of the total peak area of the HPLC–FLD chromatograms), confirming that the total uptake of selenium determined from FI–ICP–MS may be considered primarily due to intact peptide. The pattern of peaks observed in the LC–FLD chromatogram of TAMRA-Pen and TAMRA-PenMSe were similar, indicating that the presence of selenium in TAMRA-labelled Pen causes none or minimal influence on the cellular uptake. From the present data, it is not possible to conclude if the compounds additional to the intact peptides penetrated the cell membrane or were formed in the cell lysate. Only a few studies with LC–FLD measurements of lysates exposed to fluorophore-labelled Pen have been reported; extensive intracellular degradation of 5(6)-carboxyfluorescein labelled Pen was shown in yeast and non-mammalian cells after 1 h incubation (Holm et al., 2005; Palm et al., 2006). However, these results are not directly comparable to results of this study, as different cell-lines and conditions were applied. 3.7. Stability of TAMRA-PenMSe and TAMRA-Pen during cell uptake studies In order to evaluate the stability of the TAMRA-labelled and Se-labelled peptides in the media during the cell uptake study, TAMRA-PenMSe and TAMRA-Pen incubated in HBSS medium with HeLa cells were analyzed (Fig. 5). Degradation of the peptides was extensive in the media incubated with the presence of cells as compared to the degradation observed in plain HBSS media (Fig. 3 and S 1). Furthermore, a variety of degradation products was observed. Rapid and extensive degradation of CF-labelled Pen during cellular uptake studies has previously been reported and suggested as a critical determinant of cellular uptake and thus corroborate the results of this study (Lindgren et al., 2004; Trehin et al., 2004; Palm et al., 2007). The degradation products from PenMSe and TAMRA-PenMSe appeared as similar peak patterns in the beginning of the LC–ICP– MS chromatogram (Fig. 5a). This indicated that the Se-containing degradation products of PenMSe to some extent were similar, independent of the association of TAMRA. The LC–FLD chromatograms of media samples of TAMRA-PenMSe and TAMRA-Pen (Fig. 5b) shows similar degradation of the two TAMRA-labelled peptides but also reveal several fluorescent degradation products that do not contain Se. Intact peptides and oxidation products of the peptides were identified by LC–ESI–MS (MS-data in S2). Furthermore, cleaved TAMRA ([M + 1H]1+ 431.160, mass accuracy
Contents lists available at ScienceDirect
European Journal of Pharmaceutical Sciences journal homepage: www.elsevier.com/locate/ejps
Selenium as an alternative peptide label – comparison to fluorophore-labelled penetratin Laura Hyrup Møller, Jesper Søborg Bahnsen, Hanne Mørck Nielsen, Jesper Østergaard, Stefan Stürup, Bente Gammelgaard ⇑ Department of Pharmacy, University of Copenhagen, Denmark
a r t i c l e
i n f o
Article history: Received 28 August 2014 Accepted 5 November 2014 Available online 14 November 2014 Keywords: Peptide quantification Cell-penetrating peptides Labelling techniques
a b s t r a c t In the present study, the impact on peptide properties of labelling peptides with the fluorophore TAMRA or the selenium (Se) containing amino acid SeMet was evaluated. Three differently labelled variants of the cell-penetrating peptide (CPP) penetratin (Pen) were synthesized, PenMSe, TAMRA-PenMSe and TAMRA-Pen. The labelled peptides were characterized in terms of hydrodynamic radius, secondary structure during peptide–membrane interaction, effect on membrane leakage induction, uptake efficiency in HeLa cells. Furthermore, stability of peptides and identities of degradation products in cell media and cell lysate were evaluated. TAMRA-labelling increased the hydrodynamic radius of Pen and PenMSe significantly. Labelling with Se caused no or minimal changes in size, secondary structure and membrane leakage induction in concentration levels relevant for cellular uptake studies. Similar degradation patterns of all labelled peptides were observed in HBSS media; degradation was mainly due to oxidation. Cellular uptake was significantly higher for the TAMRA labelled peptides as compared to Se-labelled Pen. Extensive degradation was observed in media during cellular uptake studies, however, in all cell lysates, primarily the intact peptide (PenMSe, TAMRA-PenMSe or TAMRA-Pen) was observed. Selenium labelling caused minimal alteration of the physicochemical properties of the peptide and allowed for absolute quantitative determination of cellular uptake by inductively coupled plasma mass spectrometry. Selenium is thus proposed as a promising alternative label for quantification of peptides in general, altering the properties of the peptide to a minor extent as compared to commonly used peptide labels. Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction Biopharmaceuticals are increasingly used as drug candidates, as they most often are highly potent and biocompatible due to their similarities to natural occurring biomolecules (Vlieghe et al., 2010). To elicit the desired pharmacologic response their target sites, which may be in the cellular plasma membrane or in the intracellular environment must be reached. However, due to their relatively large molecular size and structure, these drugs are poorly absorbed across biological membranes and most often labile to enzymatic degradation once dosed. To improve the therapeutic potential of the peptide drugs, drug carriers such as cell penetrating peptides (CPPs), e.g. penetratin, have been applied to facilitate delivery across eukaryotic cell membranes (Liu et al., 2014; Khafagy and Morishita, 2012; Farkhani et al., 2014). Unravelling the mechanism of how the CPP acts and how potential peptide drug ⇑ Corresponding author. E-mail address: [email protected] (B. Gammelgaard). http://dx.doi.org/10.1016/j.ejps.2014.11.001 0928-0987/Ó 2014 Elsevier B.V. All rights reserved.
cargoes are delivered is of great interest and tracking the peptides is crucial for design and development of future peptide-based drugs, but often difficult. Peptide quantification is often performed by high performance liquid chromatography (HPLC) and UV or fluorescence detection (FLD), exploiting cromophores and fluorophores within the peptide structure (Grohganz et al., 2004; Farin et al., 1998; Backes et al., 1998). However, these methods exhibit limited sensitivity and especially selectivity for analysis of peptides in complex biological sample matrices. In order to enhance sensitivity and selectivity, a variety of different labelling strategies have been employed. One of the currently most used approaches to study peptide delivery at a cellular level is to track the peptide labelled with a fluorophore (Liu et al., 2014; Bahnsen et al., 2013; Amand et al., 2008; Lundberg and Langel, 2006; Fischer et al., 2000; Elmquist et al., 2001; Lindgren et al., 2004). Fluorophore-labelled peptides are easily accessible and allow for direct measurement of the fluorophore with high sensitivity compared to the non-labelled peptide detection. However, labelling of relatively small peptides with relatively large and
L. Hyrup Møller et al. / European Journal of Pharmaceutical Sciences 67 (2015) 76–84
hydrophobic moieties affects the physicochemical properties of the parent peptide and thereby possibly its pharmacokinetic (PK) profile Jones and Sayers, 2012. In order to monitor parameters such as stability, kinetics and metabolism of the peptide and the peptide delivery system, other quantitative methods are thus needed. In addition to fluorescence detection, liquid chromatography mass spectrometry (LC–MS) has become the method of choice for analysis of peptides. LC–MS provides detailed information of the masses of peptides and possible degradation products. However, quantitative analysis demands for costly species specific standards (John et al., 2004; Burlina et al., 2005). Recently, LC hyphenated to inductively coupled plasma mass spectrometry (ICP-MS) has been applied for biomolecule analysis (Bettmer et al., 2009; Kretschy et al., 2012). The ICP–MS provides low limits of detection and a wide dynamic range. Using a hightemperature plasma, any compound is decomposed to single elemental ions providing in principle matrix-independent sensitivity to a variety of natural elements, and thus making quantification using a simple inorganic standard possible. Detection of the naturally occurring elements such as C, N, and O is, however, challenged due to interferences from the air, while analysis of the hetero element sulfur (S) is challenged by low ionization efficiency and spectral interferences in the ICP–MS (Prange and Profrock, 2008). Thus, a number of other elemental labels has been suggested for peptide analysis. Studies employing element-tagged immunoassays using lanthanide labels and subsequent analysis by ICP–MS have been reported (Careri et al., 2009; Baranov et al., 2002; Quinn et al., 2002; Mueller et al., 2014). Furthermore, element-tagged affinity labelling techniques have been pursued in means of associating organic mercury or metallocene tags to thiol-groups and iodination of tyrosine residues (He et al., 2013; Bomke et al., 2010). Recently, metal or element chelating agents functionalized with a reactive moiety targeting a specific group in the peptide, either the thiol group of Cys or the N- or C-terminal have been used for peptide labelling prior to ICP–MS analysis (Kretschy et al., 2012; Prange and Profrock, 2008; Koellensperger et al., 2009; Esteban-Fernandez et al., 2011). These elemental labels increase analytical selectivity and sensitivity considerably compared to natural occurring hetero element tags, however, association of a metal, lanthanide or a relatively large chelating moiety to the peptide structure may induce changes in physicochemical properties of the peptide as described for fluorescent labels. Recently, we introduced selenium as an internal detection label in peptides by exchange of a natural occurring methionine (Met) residue by the selenium containing analogue selenomethionine (SeMet) Moller et al., 2014. Selenium (Se) and S belong to the group of chalcogens in the periodic table and thus share similar properties (Wessjohann et al., 2007). Furthermore, SeMet is incorporated in competition with Met in human proteins (Whanger, 1986). In our previous work, Se-containing penetratin (PenMSe) was successfully synthesized with 1:1 labelling stoichiometry and the label allowed for peptide detection by ICP–MS providing detailed information of the fate of the peptide during cellular uptake study (Moller et al., 2014). The main advantage of labelling with Se is the very small change of the peptide as the Se-labelling is performed only by exchanging the S-containing amino acid Met with the Se-containing amino acid SeMet. It is therefore expected that the physicochemical properties are only insignificantly altered. Furthermore, the labelling procedure is highly controllable, as the SeMet is introduced in the synthesis of the new CPP candidates; thus ensuring 1:1 labelling stoichiometry, which is not always the case in metal-based tagging techniques, as these processes are more difficult to control (Kretschy et al., 2012). At the same time, Se is not as abundant as S in biological systems and the Se-labelled peptides are easily distinguished from reagents and endogenous cell components in cellular uptake studies.
77
The primary aim of the current work was to evaluate the possibility of using Se-exchange of S-containing amino acids as an approach for labelling peptides using penetratin (Pen) as a model compound. Furthermore, the objective was to compare the behaviour of the fluorophore 5(6)-carboxy-tetraemethyl-rhodamine (TAMRA) labelled Pen with the Se-labelled Pen and evaluate how the resulting different changes of physicochemical properties of the labelled peptides influence cell uptake studies. We present here a comparison of hydrodynamic radii, secondary structures, peptide–membrane interactions including membrane leakage induction, cellular uptake and peptide stability of three Pen peptide derivatives: selenium labelled Pen with SeMet incorporated instead of the Met-residue (PenMSe), Pen with TAMRA at the N-terminal (TAMRA-Pen), and an analogue labelled with SeMet as well as TAMRA (TAMRA-PenMSe) (Table 1). 2. Material and methods 2.1. Reagents Rink amide resin, coupling reagents and amino acid building blocks for solid-phase peptide synthesis as well as supplementary solvents and chemicals were acquired from Iris Biotech (Merktredwitz, Germany). L-Selenomethionine (SeMet) was obtained from TCI Europe (Zwijndrecht, Belgium). 5(6)-Carboxy-tetraemethylrhodamine (TAMRA) was obtained from Novabiochem (Darmstadt, Germany). Calcein was purchased from Sigma–Aldrich (Buchs, Switzerland) and the lipids: 1-palmitoyl-2-oleoyl-sn-glycero-3phosphocholine (POPC), cholesterol, 1-palmitoyl-2-oleoyl-snglycero-3-phospho-(10 -rac-glycerol) (POPG) were obtained from Avanti Lipids (Alabaster AL, USA). For mobile phases, glacial acetic acid, trifluoroacetic acid (TFA) and methanol were obtained from BDH Prolabo (VWR, Bie & Berntsen, Søborg, Denmark). 2.2. Peptide synthesis and labelling All peptides were synthesized by Fmoc-based solid phase peptide analysis (SPPS) employing a microwave-assisted automated synthesizer (CEM, Matthews, NC, USA). N-protection of H-SeMetOH with N-(9-fluorenylmethoxycarbonyloxy)succinimide (FmocOSU) and synthesis of Pen and PenMSe by use of fully automated synthesizer were carried out as previously described (Moller et al., 2014). N-terminal labelling was carried out manually with a mixture of 5(6)-carboxy-tetraemethyl-rhodamine (TAMRA; 2.5 eq.), diisopropylcarbodiimid (DIC; 2.5 eq.) and 1-hydroxybenzotriazole (HOBt; 2.5 eq.) in dimethylformamide (DMF) overnight at room temperature in a Teflon reactor (10 mL) with a polypropylene filter and kept under nitrogen. Peptides were purified at room temperature by preparative HPLC (Luna C18(2) column, 5 lm, 250 21 mm; Phenomenex, SupWare, Denmark) with UV detection at 280 nm. A linear gradient of 0–35% eluent B (H2O:MeCN 95:5 with 0.1% (v/v) TFA) in eluent A (H2O:MeCN 5:95 with 0.1% (v/v) TFA) within 20 min was applied. Flow rate was 20 mL min1. Amino acid sequences of the synthesized peptides are shown in Table 1. 2.3. Taylor dispersion analysis Taylor dispersion analysis (TDA) was performed on an Agilent 3D-capillary electrophoresis instrument (Agilent Technologies, Waldbronn, Germany) coupled to an Actipix D100 UV imaging detector (Paraytec, York, UK) as described by Ostergaard and Jensen (2009). The UV traces were recorded with Actipix control software version 1.4 (Paraytec, York, UK) at a wavelength of 214 nm. A fused silica capillary (Polymicro Technologies, Phoenix,
78
L. Hyrup Møller et al. / European Journal of Pharmaceutical Sciences 67 (2015) 76–84
Table 1 Structures, molecular weights and hydrodynamic radii (Rh) of the labelled penetratin peptide derivatives. Rh determined by Taylor dispersion analysis using 80 lM peptide solutions and given as mean ± SD; Pen and PenMSe n = 6, TAMRA-Pen n = 7 and TAMRA-PenMSe n = 10. Abbreviation
Sequence
Maverage
Mmonoisotopic
Rh (nm)
Pen PenMSe TAMRA-Pen TAMRA-PenMSe
RQIKIWFQNRRMKWKK-NH2 RQIKIWFQNRR(MSe)KWKK-NH2 TAMRA-RQIKIWFQNRRMKWKK-NH2 TAMRA-RQIKIWFQNRR(MSe)KWKK-NH2
2245.74 2292.64 2659.19 2706.08
2244.31 2292.25 2657.46 2705.40
1.25 ± 0.03 1.23 ± 0.03 1.36 ± 0.03 1.33 ± 0.02
AZ, USA) coated with polydiallyldimethylammonium chloride (PDMAC) was prepared for the dispersion measurements. The total length of the capillary was 100 cm with 75 lm ID, and with two detection windows along the capillary, located after 30 and 50 cm, respectively. The running buffer was 67 mM phosphate buffer pH 7.4. Peptide samples were introduced into the capillary by applying a pressure of 50 mbar for 7 s (34 nL) followed by a mobilization pressure of 50 mbar (flow rate 1.1 mm/s) to force the sample plug through the capillary. The temperature of the cassette was 25 °C and each sample was analyzed in six to ten replicates. Apparent diffusion coefficients of the peptides were determined from the peak appearance times and the variances of the Gaussian shaped peaks according to Eq. (1) Ye et al., 2012:
D¼
R2C ðt 2 t1 Þ 24ðr22 r21 Þ
ð1Þ
where Rc is the radius of the capillary, t1 and t2 are the peak appearance time at the first and second detection window, respectively, and r21 and r22 are the temporal variances of the concentration profile at the first and second detection window, respectively. The hydro-dynamic radii (Rh) of the peptides were subsequently calculated from the Stokes–Einstein Eq. (2):
Rh ¼
kB T 6pgD
2.6. Calcein release assay To measure leakage from calcein-loaded liposomes, 180 lL suspension of these in Tris buffer (10 mM, pH 7.4) containing 150 mM NaCl, 0.1 mM EDTA and 70 mM calcein were added to black 96well plates (MicroWell 96 optical bottom plates, NUNC, Roskilde, Denmark) at a total lipid concentration of 20 lM. Fluorescence was monitored with a FLUOstar OPTIMA plate reader (GMB Labtech, Offenburg, Germany) at an excitation wavelength of 485 nm and an emission wavelength of 520 nm to determine initial fluorescence (F0). A volume of 20 lL of peptide solution in Tris buffer (10 mM, pH 7.4) containing 150 mM NaCl, 0.1 mM EDTA and 70 mM calcein (0.005–20 lM) was added and measurement of calcein leakage was initiated within 10 s from the first addition of peptide, and the peptide-induced calcein leakage was recorded for a period of 840 s in order to determine the fluorescence (F). Finally, the liposomes were disrupted completely by addition of 20 lL 20% (v/v) Triton X-100 (Sigma–Aldrich, Buchs, Switzerland), and the signal from the total amount of calcein was recorded (Ft). The relative release was calculated according to Eq. (3):
Calcein leakage ð%Þ ¼
ðF F 0 Þ 100% ðF t F o Þ
ð3Þ
ð2Þ
where kB is Boltzmann’s constant, T is the absolute temperature, and g is the solution viscosity. 2.4. Liposome preparation Liposomes were prepared by using a procedure described previously (Bahnsen et al., 2013). In brief, liposomes were prepared from POPC, POPG and cholesterol in a molar ratio of 5:3:2. For liposomes used for circular dichroism (CD) the lipid film was dispersed in Tris buffer (10 mM, pH 7.4), whereas calcein-loaded liposomes were formed in Tris buffer (10 mM, pH 7.4) containing 150 mM NaCl, 0.1 mM EDTA and 70 mM calcein. Excess calcein was removed by passing the liposomes through a Sephadex G-50 column (GE Healthcare, Little Chalfont, UK). The size of the liposomes was measured on a Zetasizer Nano ZS (Malvern, Worcestershire, UK) as described previously (Bahnsen et al., 2013). The total phosphor concentration of liposomes was determined by ICP–MS as previously described by Nguyen et al. (2011). 2.5. Circular dichroism (CD) spectroscopy CD spectra were measured as described previously (Bahnsen et al., 2013). In brief, peptide and liposomes were scanned 5 consecutive times in the range 190–260 nm in a quartz cell with 1 mm light path. The scans were averaged and the background contribution from solvent and/or liposomes was subtracted. An Olis DSM 10 spectrophotometer (Olis, Bogart, GA, USA) was used for all measurements. CD spectra were recorded of solutions containing 20 lM of peptide in Tris buffer (10 mM, pH 7.4), 95% trifluoroethanol (TFE) (Acros organics, Geel, Belgium) and in the presence of POPC:POPG:cholesterol (molar ratio 5:3:2) liposomes.
2.7. Cell culture Cell culturing was performed as described previously (Moller et al., 2014). In brief, the human cervical cancer cell line HeLa WT, from American Type Culture Collection (ATCC, LGC Standards AB, Boras, Sweden) was used. The cells were maintained in supplemented Eagle’s minimal essential medium (EMEM) and grown in an atmosphere of 5% CO2 at 37 °C. One day prior to the uptake experiments, 7.6 104 cells/well were seeded on the bottom of 24 well plates (Fischer Scientific, Slangerup, Denmark). 2.8. Cell uptake studies Experiments were performed 22–24 h after seeding of the HeLa WT cells; when a sub-confluent monolayer (80–90% confluence) was achieved. Peptide solutions in Hanks Balanced Salt Solution (HBSS; Gibco, Paisley, UK) supplemented with 10 mM of 4-(2hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES, Sigma Aldrich, St. Lois, US) were diluted immediately prior to use in cell uptake studies from aqueous 200 lM peptide stock solutions stored at 4 °C. Initiation of the uptake experiments was carried out by washing the cells once with 37 °C phosphate buffered saline (PBS, Sigma Aldrich, St. Louis, MO, US) and subsequent addition of 400 lL of peptide solution. Cells were incubated with the peptide solution for 2 h on an orbital shaker (90 rpm) at 37 °C. After exactly 2 h, medium was collected and immediately acidified with acetic acid to a final concentration of 2% (v/v). Cell uptake was terminated by washing the cells four times with ice-cold PBS. The cells were lysed by addition of 100 lL 0.1% (v/v) Triton-X 100 + 2% (v/v) acetic acid and kept on ice for 10 min. The lysis mixture was transferred to low-binding Eppendorf tubes (Alpha Laboratories, Hampshire, UK), centrifuged at 25,000g for 10 min and the supernatant
L. Hyrup Møller et al. / European Journal of Pharmaceutical Sciences 67 (2015) 76–84
79
collected for analysis. All samples were stored at 18 °C until analysis.
areas of TAMRA-PenMSe solution standardized from the Se-content by FI–ICP–MS.
2.9. Determination of total selenium by flow injection inductively coupled plasma mass spectrometry (FI–ICP–MS)
2.12. Identification of species by liquid chromatography electrospray ionization mass spectrometry (LC–ESI–MS)
The total amount of selenium uptake in HeLa cells was determined by flow injection (FI) analysis and ICP–MS detection. The measurements were performed on a PE Sciex ELAN 6000 ICP–MS (Perkin Elmer, Norwalk, CT, USA) equipped with a MicroMist glass nebulizer (AHF, Lab Support Hillerød, Denmark) and a PC3 cyclonic spraychamber (Elemental Scientific, Omaha, NE, USA) operated at 4 °C. The sampler and skimmer cones were made of nickel. On a daily basis, the nebulizer gas flow, lens voltage and RF power were optimized with a solution containing 100 lg L1 Se in mobile phase. The data acquisition settings were dwell time 50 ms and sweeps per reading 1. The carrier fluid for FI analysis was an aqueous solution containing 0.02% (v/v) TFA, 0.1% (v/v) acetic acid and 5% (v/v) MeOH. The flow rate was 200 lL min1 and the injection volume was 10 lL. Quantification was performed by single standard calibration based on peak areas of a 50 lg L1 solution of certified selenium element standard (1010 mg Se L1, PlasmaCAL, SCP Science, Quebec, Canada). This method was validated by numerous linear standard curves of the standard. 77Se+, 78Se+ and 82Se+ isotopes were monitored; 82Se+ was used for quantification.
LC–ESI–MS analysis was performed on a Dionex Ultimate 3000 UHPLC (Thermo Scientific, Waltham, MA) equipped with a degasser, quaternary pump, autosampler, thermostated column compartment and a diode array detector (DAD) and hyphenated to a Q Exactive Orbitrap mass spectrometer (Thermo Scientific, San Jose, CA). The column and mobile phase prepared as described for LC–ICP–MS analysis were applied for this system as well. MS data were collected between 2 and 15 min run time. The Q Exactive Orbitrap mass spectrometer was run with 250 °C heated electrospray ionization. The sheath gas was set to 30 arbitrary units; auxiliary gas to 10 arbitrary units; capillary temperature to 350 °C; and spray voltage to 3300 V. The MS was run in full scan positive mode. Scan range was 300–3000 m/z, with lock mass 445.120 m/ z, and the mass accuracy was below 5 ppm. The instrument was mass calibrated prior to analysis.
2.10. Speciation analysis by liquid chromatography inductively coupled plasma mass spectrometry (LC–ICP–MS) LC–ICP–MS analysis was carried out using an Agilent 1100 series HPLC system equipped with a degasser, a quaternary pump, an auto sampler and a column oven. The column was an Aeris PEPTIDE XB-C18, 3.6 lm, 100 2.1 mm ID column protected by a C18-peptide SecurityGuard ULTRA (Phenomenex, SupWare, Torrance, CA). A linear gradient of 20–80% (v/v) MeOH, (with 0.1% (v/v) acetic acid and 0.05% (v/v) TFA added) within 10 min was applied with a flow rate of 200 lL min1, a column temperature of 60 °C, and injection volume of 5 lL. The ICP–MS instrument applied was similar to the one described for FI–ICP–MS analysis. In order to remove MeOH prior to introduction to the ICP–MS, the membrane desolvation system Aridus II (CETAC Technologies, Omaha, NE, USA) was applied in front of the ICP–MS. Aridus II was equipped with a 200 lL min1 nebulizer and the settings were: spray chamber temperature 110 °C; desolvator temperature 160 °C; sweep gas flow 7 mL min1, and nitrogen gas flow 6 mL min1. ICP–MS nebulizer gas flow was 0.9 mL min1. Lens voltage and ICP RF power were optimized regularly using a solution of PenMSe containing 10 lg L1 Se in 50% MeOH (v/v). Data acquisition settings for speciation analysis: dwell time 200 ms; sweeps per reading 1; and readings per replicate 1325. 77Se+, 78Se+ and 82Se+ isotopes were monitored; 82Se+ was used for data analysis. 2.11. Speciation analysis by liquid chromatography fluorescence detection (LC–FLD) LC–FLD analysis was performed using the same HPLC system as described for LC–ICP–MS connected to an Agilent 1200 series fluorescence detector (G1321A FLD). The chromatographic settings were equivalent to the method described under 2.10 and the settings of the fluorescence detector were extinction wavelength 543 nm and emission wavelength 572 nm. Total amount of TAMRA taken up by cells after exposure to TAMRA-PenMSe or TAMRA-Pen were determined by single standard calibration based on peak
3. Results and discussion In the present study, the impact on peptide properties of labelling peptides with the fluorophore TAMRA or the selenium containing amino acid SeMet was evaluated. Three differently labelled variants of the cell-penetrating peptide (CPP) penetratin (Pen) were synthesized, PenMSe, TAMRA-PenMSe and TAMRA-Pen (Table 1). The labelled peptides were characterized in terms of hydrodynamic radius, secondary structure during peptide– membrane interaction, effect on membrane leakage induction, uptake efficiency in HeLa cells and evaluation of the species taken up by the cells including stability in the media during the cell uptake study. 3.1. Taylor dispersion analysis In order to determine the influence of Se- and/or TAMRA-labelling on the size of the peptide, the hydrodynamic radii (Rh) of native Pen, PenMSe, TAMRA-Pen and TAMRA-PenMSe were determined by Taylor dispersion analysis (TDA). The results are presented in Table 1. Association of TAMRA to Pen and PenMSe increased Rh significantly. Increase of radius due to labelling with TAMRA was expected, as TAMRA (Mw = 431.2 Da) provide a relatively large addition of weight and thus size to Pen and PenMSe (Mw in Table 1). On the other hand, no significant difference in Rh was observed by substitution of Met with SeMet, neither between Pen and PenMSe nor between TAMRA-Pen and TAMRA-PenMSe. Labelling with selenium only increased the molecular weight of the peptide with 46.9 Da. Furthermore, the covalent radius of selenium (116 pm) Wessjohann et al., 2007 is only slightly larger than for sulfur (102 pm) Wessjohann et al., 2007, and was not expected to change the size of the peptide significantly. The results confirmed that incorporation of SeMet instead of Met cause minimal disturbance according to size of the peptides. 3.2. Peptide–membrane interactions Circular dichroism (CD) experiments were performed in order to evaluate the influence of selenium labelling on the secondary structure of Pen in solution and when interacting with liposomes of POPC:POPG:cholesterol (molar ratio 5:3:2), a membrane composition similar to the eukaryotic membrane (Zhang et al., 2010). The average size (diameter) of the liposomes was 90 nm (PDI = 0.14).
80
L. Hyrup Møller et al. / European Journal of Pharmaceutical Sciences 67 (2015) 76–84
CD spectra of Pen and PenMSe measured in Tris buffer, trifluorethanol (TFE) and when interacting with liposomes POPC:POPG:cholesterol (5:3:2) are shown in Fig. 1. In buffer both of the peptides showed unstructured conformation, whereas an a-helical structure for both peptides was observed in 95% (v/v) TFE as indicated by the minima observed at 208 nm and 222 nm. These observations are in agreement with previously reported CD spectra for Pen (Czajlik et al., 2002). Overall, no considerable structural difference between Pen and PenMSe was detected, which indicates that the exchange of Met with SeMet did not cause any changes in the ability of the peptide to form a-helical structure. The observations for native Pen is in accordance with previous reports on the folding of Pen in an apolar environment (Bahnsen et al., 2013; Persson et al., 2004; Caesar et al., 2006). The native Pen showed a pronounced a-helix forming propensity in the presence of liposomes. This was slightly higher as compared to the folding of PenMSe in the presence of liposomes as illustrated by the lower minima at 208 nm and 222 nm for Pen. The differences observed between Pen and PenMSe may indicate, that the interaction with liposomes and thus possibly with cellular membranes are slightly different for PenMSe as compared to Pen. Differences in the degree of folding may also be explained by differences in the bond length Se–C (196 pm) and S–C (180 pm) Wessjohann et al., 2007. For proteins, the structure most often retains sufficient flexibility and plasticity to accommodate the changes from SeMet without any changes in geometry (Moroder, 2005) and the differences observed in the present work may be explained by less flexibility in the relative small peptide structure of Pen compared to proteins.
3.3. Effect of PenMSe and Pen on membrane leakage induction The effect of Pen and PenMSe on membrane leakage was investigated by the release of dye from calcein-loaded POPC:POPG:cholesterol (5:3:2) liposomes similar to those used for CD. The average size of the calcein-loaded liposomes was 129 nm (PDI < 0.1). The calcein-release was investigated in a lipid concentration of 20 lM and peptide concentrations in the range 0.005–20 lM (peptide:lipid ratios of 1:4000 to 1:1) (only data for 1.25 lM and 20 lM are shown). For peptide concentrations up to 1.25 lM (peptide:lipid ratio 1:16), no difference between Pen and PenMSe in calcein release and the kinetic release profile of calcein from liposomes was observed. Above 1.25 lM the release of calcein from liposomes increased remarkably when exposed to PenMSe while Pen caused no distinct increase of calcein release (Fig. 2). These results corroborate with the fact that Pen is known not to be pore-forming and thus, no release of calcein was expected (Caesar et al., 2006). Further studies will be necessary, in order to explain the increased release by higher concentration of the Se-containing Pen. Peptide:lipid ratios corresponding to higher concentrations than 1.25 lM of peptide (peptide:lipid 1:16), are assumed to be irrelevant in relation to cellular in vitro studies. Thus, the observed difference in membrane leakage induction from PenMSe and Pen are not expected to contribute to any disturbances in the cellular uptake studies of this work. Hence, labelling with selenium seems to constitute a promising method of labelling Pen, however, further studies on membrane interactions would be valuable. 3.4. Purity and stability of synthesized peptides Purity of the synthesized peptides was determined by LC separation and either fluorescence or ICP–MS detection. Chemical stability of the synthesized peptides was evaluated in aqueous stock solution and after 2 h incubation in HBSS at ambient temperature. The Se-containing peptides (PenMSe and TAMRA-PenMSe) were analyzed by LC–ICP–MS, while the TAMRA-labelled peptides (TAMRA-Pen and TAMRA-PenMSe) were measured by LC–FLD. The chromatograms of TAMRA-PenMSe from FLD and ICP–MS detection are shown in Fig. 3. Similar peak patterns were observed by the two complementary detection methods comprising two main peaks in each chromatogram. LC–FLD analysis of TAMRA-Pen in water and HBSS showed similar peak pattern as for TAMRA-PenMSe, however, TAMRA-Pen was not degraded to the same extent as TAMRAPenMSe (Supplementary S1). The compounds corresponding to the smaller peaks may be a result of the synthesis procedure (impurities) or due to degradation during incubation with HBSS (degradation products). By LC–ESI–MS, identities of the peptides corresponding to the main peaks in the chromatograms were determined. The highest peaks corresponded to the intact peptides, TAMRA-PenMSe and
Fig. 1. Circular dichroism spectroscopy of Pen (solid line) and PenMSe (dashed line) in presence of (a) buffer, (b) TFE or (c) POPC:POPG:Cholesterol (molar ratio 5:3:2) liposomes. Measurements were performed with a lipid concentration of 2 mM and a peptide concentration of 20 lM resulting in peptide to lipid molar ratio of 1:100.
Fig. 2. Release kinetics of calcein from POPC:POPG:cholesterol (molar ratio 5:3:2) liposomes in presence of 1.25 lM and 20 lM Pen and PenMSe and control without peptide (grey). Results are given as mean ± SD, n = 3.
81
L. Hyrup Møller et al. / European Journal of Pharmaceutical Sciences 67 (2015) 76–84
Fig. 3. LC–ICP–MS and LC–FLD chromatograms of TAMRA-PenMSe in diluted aqueous stock solution (water) and after incubation in HBSS at room temperature for 2 h.
TAMRA-Pen, respectively. The smaller peaks corresponded to a compound with a mass increase of 15.995 Da (MS-data in supplementary S2) indicating oxidation of the peptide. This oxidation probably occurred for Se (SeO) in the SeMet residue of TAMRAPenMSe and for S in the Met residue of TAMRA-Pen as those amino acids are easily oxidized to selenooxide and sulfoxide when exposed to oxygen (Moroder, 2005). The chemical degradation patterns of TAMRA-PenMSe and TAMRA-Pen may thus be considered similar under the investigated conditions. Furthermore, oxidation of PenMSe in cell culture media (HBSS) was recently reported (Moller et al., 2014). The distribution in percentages of intact peptide, major impurity/degradation product and minor impurities/ degradation products of PenMSe, TAMRA-PenMSe and TAMRA-Pen in aqueous solution and after incubation in HBSS calculated from the areas in the chromatograms are shown in Table 2. While the ICP–MS detection reveals Se-containing impurities (Fig. 4a), the FLD detection reveals unspecific fluorescent impurities (Fig. 4b). The initial purity of the synthesized labelled peptides, was similar for TAMRA-Pen (92 ± 3.0%) and PenMSe (90 ± 1.7%), indicating that labelling with TAMRA introduced about 8% fluorescent impurities while incorporation of SeMet introduced about 10% Se-containing impurities (similar impurities due to oxidation of S may be present in Pen, but this has to our knowledge never been examined). Thus, the TAMRA-PenMSe contained only 84% Se-intact peptide. Addition of SeMet to the peptide sequence will increase the propensity of oxidation of the peptide, as selenium is more readily oxidized as compared to sulfur (Wessjohann et al., 2007). This was reflected by the 9% oxidized peptide in PenMSe. However, the native TAMRA-labelled peptide TAMRA-Pen also contained about 5% impurity corresponding to an oxidation product. The oxidation may have occurred during the automated synthesis owing to exposure to high temperatures and microwave-irradiation in the presence of air. Peptide purity above 90% is often used for studies and should be expected for the final peptide products, however, lower purity of TAMRA-PenMSe was accepted as this study was
Fig. 4. Chromatograms of lysates of HeLa WT cells incubated with TAMRA-PenMSe or TAMRA-Pen in HBSS for 2 h. (a) LC–ICP–MS of TAMRA-PenMSe, (b) LC–FLD of TAMRA-PenMSe (black) and TAMRA-Pen (grey).
performed as a proof of concept study and thus, only a limited amount of peptide was synthesized. The purity of the final peptides may be improved by improving the resolution of the preparative
Table 2 Percentage of intact peptide, major and minor impurities/degradation products. Results were obtained by LC–ICP–MS and LC–FLD of peptides in aqueous solution and HBSS medium, both incubated at room temperature for 2 h. Percentages given as mean ± SD of three replicates unless otherwise stated. ICP–MS
PenMSe standard* PenMSe in HBSS TAMRA-PenMSe standard** TAMRA-PenMSe in HBSS TAMRA-Pen standard** TAMRA-Pen in HBSS** * **
n = 9, three different dilutions. n = 2.
FLD
Peptide (%)
Major (%)
Minor (%)
Peptide (%)
Major (%)
Minor (%)
90 ± 1.7 82 ± 0.8 84 ± 2.2 70 ± 1.5 – –
9 ± 1.4 17 ± 0.3 16 ± 0.7 30 ± 1.8 – –
1 ± 0.1 1 ± 0.1 – – – –
– – 74 ± 6.9 52 ± 0.3 92 ± 3.0 78 ± 2.1
– – 19 ± 6.2 41 ± 1.5 5 ± 0.7 17 ± 0.5
– – 6 ± 0.8 6 ± 1.8 3 ± 2.3 5 ± 1.5
82
L. Hyrup Møller et al. / European Journal of Pharmaceutical Sciences 67 (2015) 76–84
HPLC method, and thus separate the peptide and the major impurity. However, complete avoidance of oxidation will probably demand for absolute oxygen free environments. The chemical degradation observed for all peptides after 2 h incubation in HBSS was mainly due to further oxidation. This may have been caused by a combination of the exposure to oxygen from the air during incubation and the increased ionic strength of HBSS solution as compared to water. The amount of degradation product observed subsequent to incubation was similar (about 17%) for PenMSe and TAMRA-Pen, whereas the amount of major degradation product for TAMRA-PenMSe increased from 16% to 30% determined by ICP–MS and 19% to 41% determined by FLD. Degradation of fluorophore-labelled Pen (5(6)-carboxyfluorescein) in HBSS buffer was previously reported by Tréhin and colleagues; however, the identity of the degradation product was not described (Trehin et al., 2004). Thus, chemical degradation apparently occurs no matter which type of label is associated to Pen. However, from the current knowledge it is not possible to determine if the degradation is influenced by the label or solely due to transformation in the peptide structure. It must be noted, that this study only addresses the chemical degradation of the peptide and does not consider the stability of the fluorophore in terms of excited-state lifetime, sensitivity to photo bleaching and pH sensitivity. 3.5. Determination of total selenium in cell lysates by FI–ICP–MS and total fluorescence by LC–FLD In order to compare the impact of TAMRA labelling on the cellular uptake of PenMSe in HeLa WT cells, the total uptake of selenium was determined by FI–ICP–MS. SeMet was included as control in the experimental setup, as SeMet has formerly been reported to be taken extensively up by e.g. Caco-2 cells (Gammelgaard et al., 2012). The cellular uptake in HeLa-cells was calculated as pmol pr. 104 cells (Table 3). Significantly more selenium from TAMRA-PenMSe (52.0 ± 0.4 pmol/104 cells) was taken up by the HeLa-cells as compared to PenMSe (21.7 ± 2.5 pmol/104 cells). These results were obtained even though the initial selenium concentration of TAMRA-PenMSe was only 352 pmol/104 cells compared to 503 pmol/104 cells for PenMSe, thus demonstrating increased uptake efficiency of TAMRA-PenMSe. This indicates, that association of the fluorophore TAMRA to PenMSe significantly increased the uptake compared to the uptake of PenMSe itself. Normalization of the uptake amount to the initial amounts of peptide under the given conditions emphasizes this observation. In the cell lysate, 4% of the initial amount of selenium from PenMSe was detected whereas 15% of the initial amount of selenium from TAMRA-PenMSe was observed. In order to determine if Se-labelling influenced the uptake of TAMRA-labelled Pen, the total uptake of fluorophore from TAMRA-PenMSe and TAMRA-Pen was determined by integration of the total area of LC–FLD chromatograms of cell lysates. The total amount of fluorophore detected in lysates exposed to TAMRAPenMSe was consistent with selenium uptake from TAMRA-PenMSe
measured by FI–ICP–MS as presented in Table 3. This indicates that the intact doubly-labelled peptide is taken up by the cells, and not only the TAMRA-label or a selenium containing degradation product. For TAMRA-Pen, the total fluorescence was 90.0 ± 15.5 pmol/ 104 cells. Normalization of the FLD results for uptake of TAMRAPen to the initial concentration of peptide under the given conditions showed, however, that 18% of TAMRA-Pen was taken up by the cells. This result is not significantly different from the results obtained for TAMRA-PenMSe corresponding to 15% based on Se as well as fluorophore measurements (Table 3). In summary, the results showed that TAMRA labelling increased the uptake of labelled Pen significantly compared to Se-labelled Pen. Labelling with different fluorophores on different positions of the peptides has been the subject of several studies and it is generally accepted that addition of fluorophore may change the cellular uptake (Jones and Sayers, 2012). Fischer et al. investigated the influence of cellular uptake by labelling with either the anionic carboxyfluoroscein (CF) or the zwitter-ionic TAMRA, N-terminally or by addition of an extra Lys-residue within the peptide sequence linked to the fluorophore. The cellular uptake of fluorophore was determined by flowcytometry (Fischer et al., 2002). In that study, the total uptake efficiency of N-terminal-labelled Pen was similar regardless of fluorophore, while considerable difference in uptake was observed depending on the location of the fluorophore, suggesting, that the uptake efficiency was more related to the overall structure of the peptide than the fluorophores (Fischer et al., 2002). However, both fluorophores are large molecules and may both have influenced uptake considerably. These results support our findings, that cellular uptake might be dependent on the label and likely also its location. The association of a label to the peptide structure may thus, alter the peptide function and membrane interaction, whereas exchange of single amino acid residues in fluorophore-labelled peptides may be less relevant for cellular uptake. For example, exchange of the Met residue in Pen with e.g. Leu has previously been shown by measurement of total fluorescence not to affect the amount of fluorophore-labelled peptide taken up by cells (Bahnsen et al., 2013; Fischer et al., 2000). This supports the findings from the CD and calcein release experiments conducted in the present study that PenMSe only altered the secondary structure and membrane interaction with liposomes to a minor extent compared to Pen. In addition, studies have shown that decreasing the hydrophobicity of the N-terminal of CPPs may decrease the cellular uptake (Elmquist et al., 2006). As the TAMRA-labelled peptides appear later in reversed phase HPLC chromatograms than PenMSe and Pen, the TAMRA-labelled peptides are more hydrofobic than the native Pen and PenMSe. Thus, the increase hydrofobicity of the TAMRA-labelled peptides may explain the increased uptake of these Pen peptide derivatives. 3.6. Speciation analysis of cell lysates by LC–ICP–MS and LC–FLD Determination of the total selenium uptake in HeLa cells showed that considerable more selenium penetrated the HeLa cells when exposed to TAMRA-PenMSe compared to PenMSe. In order to
Table 3 Cellular uptake in HeLa WT cells after 2 h incubation with peptide. Total selenium was determined by FI–ICP–MS and total fluorophore was determined by LC–FLD. The amounts are given as mean ± SD (PenMSe and TAMRA-PenMSe n = 9 lysates, SeMet n = 6 lysates). FLD: uptake amounts given as mean ± SD (TAMRA-PenMSe n = 3 lysates, TAMRA-Pen n = 4 lysates). ICP–MS
PenMSe TAMRA-PenMSe TAMRA-Pen SeMet
FLD
Applied pmol/104 cells
Uptake pmol/104 cells
Normalized %
Applied pmol/104 cells
Uptake pmol/104 cells
Normalized %
503 352 – 508
21.8 ± 4.1 52.0 ± 1.4 – 57.0 ± 2.4
4 ± 0.8 15 ± 1.0 – 11 ± 1.7
– – 494 –
– 51.7 ± 6.4 90.0 ± 15.5 –
– 15 ± 1.8 18 ± 3.1 –
L. Hyrup Møller et al. / European Journal of Pharmaceutical Sciences 67 (2015) 76–84
83
describe the compounds taken up by the cells, cell lysates were analyzed by LC–ICP–MS and LC–FLD (Fig. 4). For both of the TAMRA-labelled peptides, the major compounds in the lysates were the intact peptides. LC–ESI–MS analysis confirmed the masses of intact peptide at the given retention times. LC–ICP–MS chromatograms (Fig. 4a) of cell lysate from cells exposed to TAMRA-PenMSe showed the intact peptide and a minor compound (tr = 11.1 min). LC–FLD revealed several other minor peaks (Fig. 4b). However, for both peptides the major peak corresponds to the intact peptide (TAMRA-PenMSe 58 ± 2.0% and TAMRA-Pen 69 ± 1.8% of the total peak area of the HPLC–FLD chromatograms), confirming that the total uptake of selenium determined from FI–ICP–MS may be considered primarily due to intact peptide. The pattern of peaks observed in the LC–FLD chromatogram of TAMRA-Pen and TAMRA-PenMSe were similar, indicating that the presence of selenium in TAMRA-labelled Pen causes none or minimal influence on the cellular uptake. From the present data, it is not possible to conclude if the compounds additional to the intact peptides penetrated the cell membrane or were formed in the cell lysate. Only a few studies with LC–FLD measurements of lysates exposed to fluorophore-labelled Pen have been reported; extensive intracellular degradation of 5(6)-carboxyfluorescein labelled Pen was shown in yeast and non-mammalian cells after 1 h incubation (Holm et al., 2005; Palm et al., 2006). However, these results are not directly comparable to results of this study, as different cell-lines and conditions were applied. 3.7. Stability of TAMRA-PenMSe and TAMRA-Pen during cell uptake studies In order to evaluate the stability of the TAMRA-labelled and Se-labelled peptides in the media during the cell uptake study, TAMRA-PenMSe and TAMRA-Pen incubated in HBSS medium with HeLa cells were analyzed (Fig. 5). Degradation of the peptides was extensive in the media incubated with the presence of cells as compared to the degradation observed in plain HBSS media (Fig. 3 and S 1). Furthermore, a variety of degradation products was observed. Rapid and extensive degradation of CF-labelled Pen during cellular uptake studies has previously been reported and suggested as a critical determinant of cellular uptake and thus corroborate the results of this study (Lindgren et al., 2004; Trehin et al., 2004; Palm et al., 2007). The degradation products from PenMSe and TAMRA-PenMSe appeared as similar peak patterns in the beginning of the LC–ICP– MS chromatogram (Fig. 5a). This indicated that the Se-containing degradation products of PenMSe to some extent were similar, independent of the association of TAMRA. The LC–FLD chromatograms of media samples of TAMRA-PenMSe and TAMRA-Pen (Fig. 5b) shows similar degradation of the two TAMRA-labelled peptides but also reveal several fluorescent degradation products that do not contain Se. Intact peptides and oxidation products of the peptides were identified by LC–ESI–MS (MS-data in S2). Furthermore, cleaved TAMRA ([M + 1H]1+ 431.160, mass accuracy
