Available online at www.sciencedirect.com
Biochimica et Biophysica Acta 1778 (2008) 1197 – 1205 www.elsevier.com/locate/bbamem
Structural polymorphism of two CPP: An important parameter of activity Sébastien Deshayes 1 , Marc Decaffmeyer 1 , Robert Brasseur, Annick Thomas ⁎ Centre de Biophysique Moléculaire Numérique (CBMN), Faculté Universitaire des Sciences Agronomiques de Gembloux, 2, Passage des Déportés, 5030 Gembloux, Belgium Received 10 May 2007; received in revised form 27 December 2007; accepted 28 January 2008 Available online 14 February 2008
Abstract Despite numerous investigations, the important structural features of Cell Penetrating Peptides (CPPs) remain unclear as demonstrated by the difficulties encountered in designing new molecules. In this study, we focused our interest on Penetratin and Transportan and several of their variants. Penetratin W48F and Penetratin W48F/W56F exhibit a reduced and a complete lack of cellular uptake, respectively; TP07 and TP10 present a similar cellular uptake as Transportan and TP08, TP13 and TP15 display no or weak internalization capacity. We applied the algorithmic method named PepLook to analyze the peptide polymorphism. The study reveals common conformational characteristics for the CPPs and their permeable variants: they all are polymorphic. Negative, non permeable, mutants share the opposite feature since they are monomorphic. Finally, we support the hypothesis that structural polymorphism may be crucial since it provides peptides with the possibility of adapting their conformation to medium hydrophobicity and or to partner diversity. © 2008 Elsevier B.V. All rights reserved. Keywords: Cell penetrating peptides; Carrier peptides; Transportan; Penetratin; PepLook; Structure; Polymorphism
1. Introduction Plasma membranes are impermeable barriers for a large variety of therapeutic agents. In order to circumvent permeability problems, several delivery strategies have been conceived including the development of Carrier Peptides. These short peptides, less than 30 amino acids are often amphipathic, possess a net positive charge and are in most cases Cell Penetrating Peptides (CPP) since they are able to self-translocate across biological membranes [1]. Carrier Peptides can in addition deliver various kinds of molecules, from small particles to peptides, from proteins to nucleic acids into a wide variety of cells and organisms [2,3]. Despite the numerous studies carried out, the mechanisms of nude peptide and of carrier–cargo complexes uptake by cells are unclear. Several mechanisms such as reverse micelle model [4], local electroporation [5], induced-membrane fusion [6], transient pore-like structure [7] and endocytotic pathways [8–10] have ⁎ Corresponding author. Tel.: +32 81 62 25 21; fax: +32 81 62 25 22. E-mail address:
[email protected] (A. Thomas). 1 Authors have contributed equally to this work. 0005-2736/$ - see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.bbamem.2008.01.027
been proposed individually or in combination [11]. A large variety of biological and biophysical approaches have been applied to elucidate translocation properties [12] and relationships between mechanism of internalization and peptide 3D structure remain to be clarified [13]. In the present work we have analyzed the possibilities of structure polymorphism which could be important to explain influence of the environment [14–16]. We have analyzed two CPPs: Penetratin and Transportan, as well as several of their permeable and non permeable variants, named positive and negative mutants, respectively. Penetratin is a 16 residue peptide corresponding to the third helix of the Antennapedia homeodomain (segment 43–58) [17], generally used to translocate peptides and oligonucleotides [18]. Transportan is a 27 amino acid chimeric CPP derived from the 12 first residues of the N-terminal part of the neuropeptide galanin linked to the 14 amino acids of the wasp venom mastoparan [19], which mediates transfer of proteins, PNAs (Peptide Nucleic Acids) and siRNA [18]. We have compared the conformational features of native and mutant peptides using PepLook [20,21], a Boltzman–Stochastic algorithm that calculates large series of random peptide conformations and selects the best 3D models on energy criteria.
1198
S. Deshayes et al. / Biochimica et Biophysica Acta 1778 (2008) 1197–1205 Finally the force field allows calculation of the energy of the structure with respect to the solvent. Solvent contribution is calculated via the external hydrophobicity energy as described in Eq. (4):
2. Materials and methods 2.1. Peptide sequences
Epho out ¼
Penetratin: RQIKIWFQNRRMKWKK-NH2 Penetratin W48F RQIKIFFQNRRMKWKK-NH2 Penetratin W48F/W56F RQIKIFFQNRRMKFKK-NH2 Transportan: GWTLNSAGYLLGKINLKALAALAKKIL-NH2 TP07: LNSAGYLLGKINLKALAALAKKIL-NH2 TP08: LLGKINLKALAALAKKIL-NH2 TP10: AGYLLGKINLKALAALAKKIL-NH2 TP13: LNSAGYLLGKALAALAKKIL-NH2 TP15: LNSAGYLLGKLKALAALAK-NH2
N X
Si EtrSi
ð4Þ
i¼1
Where S is the solvent-accessible surface of atoms calculated using the method described by Shrake and Rupley with a surface precision of 162 pts [28] as previously used to compute hydrophobic and hydrophilic surface of residues in soluble proteins [29]. EtrSi is the energy of transfer of atom i from a hydrophobic to a hydrophilic phase by surface area unit [30].
3. Results 2.2. The PepLook method In order to explore the conformational space of peptides we used the Boltzmann– Stochastic method, PepLook [20,21]. Stochastic methods are used to explore large combinatorial systems in different scientific domains [22,23]. In our case, first, a random population of 10,000 peptide conformations is generated using a set of 64 Φ/ Ψ couples of angles (Supplementary material Table 1) derived from the structural alphabet for protein structures of Etchebest et al. [24]. This alphabet was built from 1407 PDB structures determined by X-Ray (resolution less than 2 Å) so as to describe all possible fragments of protein structures. The energy of all peptide conformations is computed using the force field previously described [20]. In the next steps, the algorithm iteratively varies (decreases and increases) the probability of randomly selecting each combination of Φ/Ψ couples of angles if they contribute in exclusively poor or, in exclusively good energy structural solutions, respectively. Step after step, the probability of Φ/Ψ value for each residue varies. When the mean probability of peptide Φ/Ψ couples of angles remains constant, the process is stopped and 99 structures of lower energy are further minimized using the Simplex [25] method with a precision of 5 degrees and a maximal number of 1000 steps.
2.3. Structure energy The energy of peptides was calculated as the sum of four contributions: van der Waals energy, electrostatic energy, internal hydrophobic energy and external hydrophobic energy. The van der Waals contribution is calculated using the 6–12 Lennard–Jones description of interaction energy between unbonded atoms (Eq. (1)): 2 !12 !6 3 X ri0 þ rj0 ri0 þ rj0 4Aij 5 ð1Þ Bij EvdW ¼ dij dij ij
Structures of PepLook models of Penetratin, Transportan and of their variants were studied. The two Penetratin variants i.e. Penetratin W48F, in which residue Trp48 is substituted by a Phe, and Penetratin W48F/W56F in which both Trp residues (Trp48 and Trp56) are replaced by Phe [4,11,17] display less internalization capacities than the original molecule [4,17,31]. For Transportan, the variants are either taken up by cells (TP07 and TP10) or display low or no internalization (TP08, TP13 and TP15) [32]. For each peptide, PepLook was used to generate 99 models i.e. a Prime and 98 low energy structures, firstly in a hydrophilic, secondly in a hydrophobic environment. Variations of energy between the Prime and the 98 other models for Transportan and Penetratin are reported in Fig. 1. Maximal difference of energy between the 99th model and the Prime are 4.9% for Transportan and 7% for Penetratin. Each calculation was run in triplicate in each environment. 3.1. Penetratin and mutants 3.1.1. Prime structure analysis The three Primes of Penetratin and of its mutants, Penetratin W48F and Penetratin W48F/W56F are presented in Fig. 2a.
Aij and Bij are coefficients assigned to atoms pairs, roi and roj are the van der Waals radii of atoms i and j, and dij is the distance between i and j. The Coulomb's equation (Eq. (2)) was used for the calculation of electrostatic interaction energy between unbonded atoms: N 1 X N X Eelec ¼ k i¼1 j¼i1
qi qj eij ðzÞdij
ð2Þ
λ is the electronic density unit conversion factor, dij is the distance between atoms i and j. εij(Z) is the medium dielectric constant varying from 1 to 80 with a sigmoid function [26] of dij between 2 and 10 Å. qi and qj are the FCPAC charges of atoms i and j [27]. The intra molecular hydrophobicity contribution was calculated using Eq. (3). In this equation the hydrophobicity energy decreases as an exponential function of the distance between atoms: Epho
intra
¼
N 1 X N X
dij Etri fij þ Etrj fji exp ri0 þ ri0 dij =2rsol
ð3Þ
i¼1 j¼iþ1
δij = − 1 if atoms in interaction are both hydrophobic or both hydrophilic and δij = +1 if atoms i and j are of opposite type. Etri and Etrj are the energy for transfering atoms i and j from a hydrophobic to a hydrophilic phase; fij and fji are the ratio of atomic surface covered by its partner in the interaction; r0i and r0j are the van der Waals radii of atoms i and j and, dij is the distance between i and j and rsol is the radius of a water molecule.
Fig. 1. Variation of energy between the Prime and the 98 low energy structures generated by PepLook. Penetratin (triangle) and Transportan (square) in both hydrophobic (open labels) and hydrophilic (bold labels) environments, respectively. Negative numbers are for hydrophobic models and positive numbers are for hydrophilic models. Zero stand for the Primes.
S. Deshayes et al. / Biochimica et Biophysica Acta 1778 (2008) 1197–1205
Fig. 2. Prime structures of cell-penetrating peptides and variants. (a) PepLook Penetratin, Penetratin W48F and Penetratin W48F/W56F models. The three Prime structures in hydrophilic (Phi) medium (Cyan) and in hydrophobic (Pho) medium (Magenta). (b) The three Prime structures of Transportan, TP07, TP08, TP10, TP13 and TP15 in hydrophilic (Phi) medium (Cyan) and in hydrophobic (Pho) medium (Magenta). Leu16, 13 and 10 are shown for Transportan, TP07 and TP10, respectively.
Analyses of the secondary structure per residue by Pex [33] reveal that Penetratin's Primes are structurally loose with 30% helix contributions. Residues Arg53 to Lys55 of the C-terminal moiety are always helical. Hydrophilic Primes of mutants present a higher helical propensity: Penetratin W48F and Penetratin W48F/W56F are 55% and 75% helical, respectively (Fig. 2a). Helical residues are Ile47 to Lys55 and Ile47 to Lys58, respectively. Mutant hydrophobic Primes are 31% and 44% helical for Penetratin W48F and W48F/W56F, respectively. Helical conformations involve two short stretches: Ile47 to Phe49 and Arg53 to Lys55 for Penetratin W48F and Ile47 to Phe49 and Arg53 to Lys57 for Penetratin W48F/W56F. Finally, mutations result in an increased helicity, especially in the hydrophilic medium. 3.1.2. Analysis of polymorphism In addition to the Prime models, The three PepLook runs generated three times 98 models of low energy for each sequence (Supplementary material Figure 1). Diversity of models, i.e. polymorphism [20], was calculated via two kinds of Root Mean Square deviation (RMSd) values: global and window RMSd. Each of the 98 structures was compared to its Prime, by the global RMSd to detect subpopulations (Fig. 3, insets) and, by the RMSd within a sliding window of nine residues (RMSd [9]) to map the structural divergence in sequence (Fig. 3). For Penetratin, mean global RMSd values are 1.7 and 2.3 in the hydrophilic and hydrophobic conditions, respectively (Table 1).
1199
Distributions of global RMSd reveal at least three subpopulations in the hydrophilic medium and two in the hydrophobic condition (Fig. 3a, inset). In all cases, the major subpopulation includes the Prime and accounts for 64% and 56% of the 98 structures in hydrophilic and hydrophobic conditions, respectively. Models of these major subpopulations have low RMSd [9] (below 1.5 Å) suggesting they are similar to the Prime throughout the sequence (Fig. 3a). For the other subpopulations, the RMSd [9] values throughout the sequence map the structural variability in the peptide center (WFQNR) in both conditions. For the third subpopulation of the hydrophilic medium, structural differences are optimal in the C-terminal end of the peptide (RMKWKK). For mutants, mean global RMSd of the 98 structures are 1.5, 2.1 and 0.8, 1.8 for Penetratin W48F and W48F/W56F in hydrophilic and hydrophobic media respectively (Table 1). The slight decrease as compared to native Penetratin goes with a reduced number of subpopulations in the hydrophilic environment and still at least three of them in the hydrophobic conditions (Fig. 3b and c). The Prime's subpopulation accounts for 88% of the hydrophilic structures of Penetratin W48F and, for all structures of Penetratin W48F/W56F. Window RMSd [9] indicates that models of the Prime's subpopulations are monomorphic throughout the sequence with RMSd [9] values below 1.5 Å (Fig. 3b and c). The small additional subpopulation identified for Penetratin W48F in hydrophilic conditions diverges from native Penetratin in the C-terminal part of the sequence. In the hydrophobic medium, two small subpopulations are identified in addition to the Prime subpopulation (Fig. 3b and c). The RMSd [9] plots show higher values than 1.5 Å throughout the sequence indicating a disperse mapping of polymorphism. Penetratin W48F/W56F has a different pattern since a major polymorphismconferring fragment appears to be the IFFQ motif. 3.1.3. Comparison with experimental data Various studies have investigated the structure of native and variant Penetratin (Table 2): CD spectroscopy indicates that native Penetratin is random coiled in water or in phosphate buffer [34,35], that it displays a propensity to fold into an α-helix in TFE, in SDS-containing media or in the presence of phospholipid vesicles [11,34], and that it adopts an antiparallel β-oligomeric conformation at a lipid containing air–water interface [36]. NMR investigations in bicelles reveal an α-helical contribution involving residues Lys46 to Met54 [31]. Other studies reveal that Penetratin undergoes an α→β transition in relation with the surface charge density of phospholipid vesicles [37] and, that it adopts a β structure with a short helix in the presence of phospholipid vesicles at low concentrations [14]. Finally, Penetratin is a “chameleon-like” peptide generally random coiled in water that tends to fold into an α-helix or to adopt a β-fold depending on the conditions [14]. CD analyses of both mutants indicate that they are mostly random coiled in water or in a buffer [34,37]. In the presence of a membrane mimicking environment, CD and NMR investigations reveal an α-helix [11,31,37]. However, in the presence of negatively-charged phospholipid vesicles, partial β-structures have been observed [37]. Finally helical contributions are higher for the two mutants than for Penetratin [11].
1200
S. Deshayes et al. / Biochimica et Biophysica Acta 1778 (2008) 1197–1205
Fig. 3. Analysis of all PepLook models for Penetratin (a) Penetratin W48F (b) and Penetratin W48F/W56F (c). Right histograms: the 98 PepLook structures ranked according to their global RMSd with respect to the Prime in hydrophilic (Phi) and hydrophobic (Pho) environments. Left plots: RMSd on [9aa] window with respect to the Prime along sequence. Each plot is the mean of subpopulation models: Prime subpopulation (diamond) and other subpopulations (triangle, circle and square) in both hydrophilic (bold labels) and hydrophobic (open labels) environments.
The Peplook Penetratin Prime is mainly random coiled (70%) in water (Fig. 2a) and α-helical propensity is identified for residues Arg53 to Lys55. These residues were also shown by NMR Table 1 RMS deviation of the PepLook structures Peptide
Mean global RMSd a on 98 structures
Penetratin Penetratin W48F Penetratin W48F/W56F Transportan TP07 TP08 TP10 TP13 TP15
1.7 1.5 0.8 5.1 4.2 1.9 2.3 0.8 0.9
Hydrophilic environment Hydrophobic environment 2.3 2.1 1.8 5.0 3.7 0.8 3.3 0.8 0.9
a RMSd values are obtained by fitting backbone atoms of each of the 98 PepLook structures to the Prime. Mean values (in Å) are provided.
investigations to adopt a helical conformation in bicelles [31]. When the mutants are examined, the PepLook models do not reveal the random coil contribution that was described in CD experiment but rather correlate in the enrichment of helical contribution as reported by Letoha et al. [11]. PepLook does not identify significant beta structures but, the poor self-stability of the peptide and the presence of aromatic residues support the possibility that structural rearrangement might occur depending on the environment [20,38]. 3.2. Transportan and deletion analogues 3.2.1. Prime structure analysis Transportan's Primes are composed of two distinct structural domains as already described by Thomas et al. [20] (Fig. 2b). The C-terminal domain of Transportan is helical between residues Leu16 and Leu27 with both α- and 310-helices. Helices cover 50% of the residues irrespective of medium hydrophobicity. The
S. Deshayes et al. / Biochimica et Biophysica Acta 1778 (2008) 1197–1205
1201
Table 2 Summary of experimental data from the literature Peptide
Experimental data
Structure
References
Penetratin
CD in buffer CD in different liposome PM-IRRAS at the interface CD in POPG/POPC L/P = 100 CD in POPG L/P = 100 CD in PC/PS L/P = 50 NMR in DMPG/DMPC bicelles NMR in TFE CD in buffer CD in POPG/POPC L/P = 100 CD in POPG L/P = 100 CD in buffer CD in POPG/POPC L/P = 100 CD in POPG L/P = 100 NMR in DMPG/DMPC bicelles NMR in TFE CD in buffer CD in SDS NMR in SDS NMR in DMPC/DHPC bicelles NMR in DMPC/DMPG bicelles CD in POPG/POPC L/P = 100 CD in POPG L/P = 100
Random coils β-structures + helical contributions Antiparallel β-sheet 55% of helix 45% of β-structures 46% of helix α-helix (46–54) α-helix Random coil 40% of helix 50% of β-structures Random coil 25% of helix 55% of β-structures α-helix (46–54) α-helix (46–54) 24–30% of helix 60% of helix α-helix (14–26) α-helices (5–8 + 15–26) α-helices (5–8 + 15–26) 60% of helix 60% –
[31,34,35,37] [14] [36] [37] [37] [34] [31] [11] [34,37] [37] [37] [34,37] [37] [37] [31] [11] [35,37,39,40] [35,39] [35,39] [40] [41] [37] [37]
Penetratin W48F
Penetratin W48F/W56F
Transportan
N-terminal region of Transportan is structurally more variable. A short α-helix between residues Ser6 and Gly8 is observed, specifically in the hydrophobic environment. The PepLook Primes of positive mutants, TP07 and TP10 are 50% helical at the C-terminus and variable in the N-terminal moiety (Fig. 2b). The helix extends from Leu13 to Leu24 and from Leu10 to Leu21 for TP07 and TP10, respectively. In the flexible N-terminal region, a short β-conformation is characterized between Gly5 and Leu7 of TP07. For the negative analogues, PepLook Primes are helical all throughout the sequence (Fig. 2b). α-helix accounts for 70% of TP08 and about 80% of TP13 and TP15. 3.2.2. Analysis of polymorphism We analyzed both global RMSd and window RMSd [9] values for the 98 PepLook additional low energy structures for Transportan (Supplementary material Figure 2) and its positive and negative analogues (Fig. 4). For Transportan, global RMSd indicate that the Prime model is unique (hydrophobic condition) or included in a very small subpopulation (hydrophilic condition) and that the other models are spread in a large distribution of least three subpopulations (Fig. 4a, insets). The RMSd [9] values throughout the sequence, (Fig. 4a) indicate that the polymorphism of the non-Prime subpopulations is mapped in the N moiety. Distributions of RMSd values of PepLook 98 models of Transportan's positive analogues TP07 and TP10 are not as dispersed as for Transportan but still suggest several subpopulations (Table 1, Fig. 4b and c, insets) i.e. an isolated Prime model and at least two subpopulations for TP07 in the two environments. The origin of TP07 polymorphism is mapped on the N-side of the sequence (Fig. 4b). At least three subpopulations are characterized for TP10 in the hydrophilic and hydrophobic conditions
(Fig. 4c). The non-Prime subpopulations are polymorphic in the N-terminus. For the negative analogues of Transportan, reduced numbers of subpopulations are observed (Fig. 4d, e and f, insets). For TP13 and TP15, only one population including the Prime is characterized in both environments. The RMSd [9] of all models with respect to the Prime is low throughout the sequence (Fig. 4e and f). For TP08, a second population is characterized in the hydrophilic medium only. This subpopulation displays high RMSd [9] values for the N-moiety suggesting that TP08 has some residual Transportan-like polymorphism. (Fig. 4d). 3.2.3. Comparison with experimental data Most structural investigations on Transportan are reported in Table 2. CD studies have demonstrated α-helical contributions of 30% in water and around 60% in the presence of lipid vesicles or SDS micelles [35,37,39]. NMR experiments have revealed that the C-terminal domain is an α-helix in SDS and bicelles (residues Leu16 to Ile26) and that the N-terminal structure is much less stable [39]. It folds into a partly helical conformation (residues Asn5 to Gly8) in phospholipid bicelles [40,41] and NMR investigations indicate that both (N and C) domains are separated by a bend centered at Asn15. For deletion analogues, few experimental investigations have been reported. Most data were from molecular modeling [32]. Features of Transportan's PepLook Primes fit with experimental data. As in CD experiments, PepLook shows that the helical contribution is weakly increased in a hydrophobic environment. NMR and PepLook Transportan models are similar with a stable C-terminal α-helix separated from an N-terminal polymorphic domain by a little bend around Asn15. For the analogues, PepLook points out a clear difference between positive and negative analogues: positive mutants present a Transportan-like
1202 S. Deshayes et al. / Biochimica et Biophysica Acta 1778 (2008) 1197–1205 Fig. 4. Analysis of PepLook models for Transportan (a) and for its deleted analogues TP07 (b), TP10 (c), TP08 (d), TP13 (e) and TP15 (f). Right histograms: the 98 PepLook structures are classified according to their global RMSd with respect to the Prime in both hydrophilic (Phi) and hydrophobic (Pho) environments. Left plots: RMSd on [9aa] window with respect to the Prime along sequence. Each plot is the mean of subpopulation models: Prime subpopulation (diamond) and other subpopulations (triangle, circle and square) in both hydrophilic (bold labels) and hydrophobic (open labels) environments.
S. Deshayes et al. / Biochimica et Biophysica Acta 1778 (2008) 1197–1205
structure with N-terminal polymorphism while negative analogues are essentially monomorphic throughout the sequence. 4. Discussion Transportan and Penetratin are both Cell Penetrating Peptides (CPP) [18]. In this study, we have used the algorithm PepLook recently calibrated on a series of peptides by comparison to the NMR structures [20,21] to investigate the structural polymorphism of these two molecules and of some of their mutants. PepLook, with its description of the best model (The Prime) and of a series of low energy models, is an important mean to decipher the structural polymorphism of peptides simply from sequences [20,42]. Besides their activity as carriers, Transportan and Penetratin are cell-penetrating peptides. We here demonstrate that both peptides are polymorphic and, that their non-permeant mutants [11,17,32] have lost this polymorphism. We suggest that polymorphism is somehow related to activity. This might occur for permeability. Several mechanisms of transport have been evoked for these CPP. Among them, endocytosis, and especially macropinocytosis, seem to be routes of cellular internalization [12]. We might here simply refer to peptide self-permeability. Indeed, any molecule that traverses a membrane must somehow be able to solubilize in water as in a lipid phase. This may occur by different means as the binding to a polar partner to mask polar moieties and to increase the apparent hydrophobicity. Indeed, we previously demonstrated that membrane permeability of the calcium ionophore A23187 [43] and of the leukotriene–calcium complexes [44] involves such structural modifications triggered by calcium. The membrane–water interface should also be the layer where the peptide will transfer from the pure hydrophilic to the pure hydrophobic environment. Polar groups of lipids and of sugar could be important in that transition as long as charged groups of CPP could bind to them since it would result in an increase of the peptide apparent hydrophobicity. More simply here, we suggest that the structural diversity may help for permeability by increasing the molecule capacities to adapt to different media or partners. This could be the privilege of peptides as compared to smaller chemical drugs. For the two carriers characterized in this study, Prime models of PepLook display a mixture of mostly random coiled and helical moieties. Analysis of the 98 lower energy models reveals two to three subpopulations of conformations. The sequence fragments responsible for Penetratin polymorphism map both to the center and to the C-terminus. The existence of several subpopulations is in agreement with the description that Penetratin has a “chameleon-like” conformational behavior [14,45]. For Transportan, the Prime conformations correspond to a poorly defined N-moiety and a helical C-terminal region. This fits with experimental data [20,39–41]. Analysis of the 98 PepLook models also reveals several subpopulations, all with a helical C-terminal domain and variable in the N-terminal region. This supports the conclusion that Transportan is polymorphic but only in its N-terminus moiety. Thus, both Penetratin and Transportan are polymorphic, at least for a part of their structure. Two kinds of mutants were derived from Penetratin and Transportan; positive analogues whose cellular uptake activity
1203
resembles that of the wild type carrier and negative variants displaying no or weak cellular internalization [11,17,32]. For Penetratin, mutants W48F and W48F/W56F are weak and strong negative variants, respectively. Primes of Penetratin W48F and Penetratin W48F/W56F are more helical than the Primes of Penetratin. The 98 PepLook models reveal a reduced structural diversity in the hydrophilic environment. Some polymorphism is still observed in hydrophobic conditions. Polymorphism is more restricted for the di-Trp-substituted Penetratin W48F/W56F than for the mono-Trp-substituted W48F. This suggests a relationship between polymorphism capability and the presence of Trp. Substitution of Trp by Phe does not decrease the peptide hydrophobicity but does affect its polymorphism. Trp is the bulkiest amino acid and, as such it is the most restrictive for its sequence neighbour in search of stability. Changing Trp to Phe favors an helical structure where residues find intramolecular partners for interactions [11,31]. Finally, structural monomorphism is more pronounced for Penetratin W48F/W56F than for Penetratin W48F, Penetratin being the most polymorphic. Analysis of Transportan analogues reveals a more helical structure for the negative mutants TP08, TP13 and TP15 than for the positive ones TP07 and TP10. Primes of TP07 and TP10 display the same α-helical C-terminal segment as Transportan whereas their N-terminal regions adopt random coiled conformations. These structures are close to those experimentally described for the native Transportan. In contrast, Prime models of TP08, TP13 and TP15 are helical almost all along the sequence. Analysis of the 98 PepLook models indicates that Transportan and its positive mutants are polymorphic in the N-terminal region; negative mutants are monomorphic. One might argue that, rather than polymorphism, variation of Transportan size is responsible for the loss of activity. Indeed, Transportan, TP07 and TP10 are 27, 24 and 21 residues, respectively. TP13 is 20 residues only, i.e. one amino acid less than TP 10. Hence, the loss of activity concords with a reduced peptide sequence. However sequence length cannot be the only reason: Penetratin is only 16 amino acids and is a CPP. Hence, peptide polymorphism rather than length might be associated with activity. Interestingly, we have previously demonstrated that hCT(9–32), another Cell Penetrating Peptide is polymorphic too [20]. It is clear that, to cross a cell membrane, a peptide must elicit a minimal hydrophobicity. Too large hydrophobicity would lead to permanent insertion into the membrane and to aggregation in water, whereas a too little hydrophobicity would lead to incapacity to insert into membranes. In this study, we propose that structural polymorphism can help modulating the apparent hydrophobicity. But, the consequences of structural polymorphism should not be restricted to membrane permeability. Besides being CPPs, Transportan and Penetratin are also carriers. Polymorphism might be required to explain the diversity of cargoes that they can translocate into cells. Polymorphism would in that case refer to a “partner adaptability”; both “partner” and “medium” adaptabilities resulting from a sequence of amino acids encoding for 3D structures with low capacities of self-stabilization i.e. for polymorphism [20].
1204
S. Deshayes et al. / Biochimica et Biophysica Acta 1778 (2008) 1197–1205
Acknowledgments The authors acknowledge Biosiris-Peptides SA (Belgium) for the use of PepLook and Dr. M.C. Morris for her help with the manuscript. S.D. and M.D. were supported by the Fonds de la Recherche Scientifique Médicale (FRSM, Belgium) and by the Ministère de la Région Wallonne (DGTRE; contract PEPSEIN, Belgium), respectively. A.T. is Research Director at the Institut National de la Santé et de la Recherche Médicale (INSERM, France). R.B. is Research Director at the National Fund for Scientific Research (FNRS, Belgium).
[16]
[17]
[18]
[19] [20]
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bbamem.2008.01.027. References [1] M. Zorko, U. Langel, Cell-penetrating peptides: mechanism and kinetics of cargo delivery, Adv. Drug Deliv. Rev. 57 (2005) 529–545. [2] R. Tréhin, H.P. Merkle, Chances and pitfalls of cell penetrating peptides for cellular drug delivery, Eur. J. Pharm. Biopharm. 58 (2004) 209–223. [3] G.P. Dietz, M. Bahr, Delivery of bioactive molecules into the cell: the Trojan horse approach, Mol. Cell. Neurosci. 27 (2004) 85–131. [4] D. Derossi, S. Calvet, A. Trembleau, A. Brunissen, G. Chassaing, A. Prochiantz, Cell internalization of the third helix of the Antennapedia homeodomain is receptor-independent, J. Biol. Chem. 271 (1996) 18188–18193. [5] H. Binder, G. Lindblom, Charge-dependent translocation of the Trojan peptide penetratin across lipid membranes, Biophys. J. 85 (2003) 982–995. [6] S.T. Henriques, M.A. Castanho, Consequences of nonlytic membrane perturbation to the translocation of the cell penetrating peptide pep-1 in lipidic vesicles, Biochemistry 43 (2004) 9716–9724. [7] S. Deshayes, T. Plenat, P. Charnet, G. Divita, G. Molle, F. Heitz, Formation of transmembrane ionic channels of primary amphipathic cell-penetrating peptides. Consequences on the mechanism of cell penetration, Biochim. Biophys. Acta 1758 (2006) 1846–1851. [8] I.M. Kaplan, J.S. Wadia, S.F. Dowdy, Cationic Tat peptide transduction domain enters cells by macropinocytosis, J. Control. Release 102 (2005) 247–253. [9] J.P. Richard, K. Melikov, H. Brooks, P. Prevot, B. Lebleu, L.V. Chernomordik, Cellular uptake of unconjugated TAT peptide involves clathrin dependent endocytosis and heparan sulfate receptors, J. Biol. Chem. 280 (2005) 15300–15306. [10] A. Fittipaldi, A. Ferrari, M. Zoppe, C. Arcangeli, V. Pellegrini, F. Beltram, M. Giacca, Cell membrane lipid rafts mediate caveolar endocytosis of HIV-1 Tat fusion proteins, J. Biol. Chem. 278 (2003) 34141–34149. [11] T. Letoha, S. Gaal, C. Somlai, A. Czajlik, A. Perczel, B. Penke, Membrane translocation of penetratin and its derivatives in different cell lines, J. Mol. Recognit. 16 (2003) 272–279. [12] R. Fischer, M. Fotin-Mleczek, H. Hufnagel, R. Brock, Break on through to the other side-biophysics and cell biology shed light on cell-penetrating peptides, Chembiochem 6 (2005) 2126–2142. [13] S. Deshayes, M.C. Morris, G. Divita, F. Heitz, Cell penetrating peptides: a tool for therapeutics delivery, Cell. Mol. Life Sci. 62 (2005) 1839–1849. [14] M. Magzoub, L.E. Eriksson, A. Graslund, Conformational states of the cell-penetrating peptide penetratin when interacting with phospholipid vesicles: effects of surface charge and peptide concentration, Biochim. Biophys. Acta 1563 (2002) 53–63. [15] P.A. Wender, D.J. Mitchell, K. Pattabiraman, E.T. Pelkey, L. Steinman, J.B. Rothbard, The design, synthesis, and evaluation of molecules that enable or
[21] [22]
[23]
[24]
[25] [26] [27] [28] [29] [30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
enhance cellular uptake: peptoid molecular transporters, Proc. Natl. Acad. Sci. U. S. A. 97 (2000) 13003–13008. L. Chaloin, P. Vidal, A. Heitz, N. Van Mau, J. Mery, G. Divita, F. Heitz, Conformations of primary amphipathic carrier peptides in membrane mimicking environments, Biochemistry 36 (1997) 11179–11187. D. Derossi, A.H. Joliot, G. Chassaing, A. Prochiantz, The third helix of the Antennapedia homeodomain translocates through biological membranes, J. Biol. Chem. 269 (1994) 10444–10450. M. Mae, U. Langel, Cell-penetrating peptides as vectors for peptide, protein and oligonucleotide delivery, Curr. Opin. Pharmacol. 6 (2006) 509–514. M. Pooga, M. Hallbrink, M. Zorko, U. Langel, Cell penetration by transportan, FASEB J. 12 (1998) 67–77. A. Thomas, S. Deshayes, M. Decaffmeyer, M.H. Van Eyck, B. Charloteaux, R. Brasseur, Prediction of peptide structure: how far are we? Proteins 65 (2006) 889–897. http://www.biosiris.com/02prod_serv/PepLook/PepLook_Info.php. M. Glick, A. Rayan, A. Goldblum, A stochastic algorithm for global optimization and for best populations: A test case of side chains in proteins, Proc. Natl. Acad. Sci. U. S. A. 99 (2002) 703–708. A. Rayan, H. Senderowitz, A. Goldblum, Exploring the conformational space of cyclic peptides by a stochastic search method, J. Mol. Graph. 22 (2004) 319–333. C. Etchebest, C. Benros, S. Hazout, A.G. de Brevern, A structural alphabet for local protein structure: improved prediction methods, Proteins 59 (2005) 810–827. J.A. Nelder, R.A. Mead, A simplex method for function minimization, Comput. J. 7 (1965) 303–313. P.E. Smith, B.M. Pettitt, Modeling solvent in biomolecular systems, J. Phys. Chem. 98 (1994) 9700–9711. A. Thomas, A. Milon, R. Brasseur, Partial atomic charges of amino acids in proteins, Proteins 56 (2004) 102–109. A. Shrake, J. Rupley, Environment and exposure to solvent of protein atoms, J. Mol. Biol. 79 (1973) 351–371. L. Lins, A. Thomas, R. Brasseur, Analysis of accessible surface of residues in proteins, Prot. Sc. 12 (2003) 1406–1417. R. Brasseur, Differentiation of lipid-associating helices by use of threedimensional molecular hydrophobicity potential calculations, J. Biol. Chem. 266 (1991) 16120–16127. M. Lindberg, H. Biverstahl, A. Graslund, L. Maler, Structure and positioning comparison of two variants of penetratin in two different membrane mimicking systems by NMR, Eur. J. Biochem. 270 (2003) 3055–3063. U. Soomets, M. Lindgren, X. Gallet, M. Hallbrink, A. Elmquist, L. Balaspiri, M. Zorko, M. Pooga, R. Brasseur, U. Langel, Deletion analogues of transportan, Biochim. Biophys. Acta 1467 (2000) 165–176. A. Thomas, O. Bouffioux, D. Geeurickx, R. Brasseur, Pex, analytical tools for PDB files. I. GF-Pex: basic file to describe a protein, Proteins 43 (2001) 28–36. B. Christiaens, J. Grooten, M. Reusens, A. Joliot, M. Goethals, J. Vandekerckhove, A. Prochiantz, M. Rosseneu, Membrane interaction and cellular internalization of penetratin peptides, Eur. J. Biochem. 271 (2004) 1187–1197. M. Magzoub, K. Kilk, L.E. Eriksson, U. Langel, A. Graslund, Interaction and structure induction of cell-penetrating peptides in the presence of phospholipid vesicles, Biochim. Biophys. Acta 1512 (2001) 77–89. E. Bellet-Amalric, D. Blaudez, B. Desbat, F. Graner, F. Gauthier, A. Renault, Interaction of the third helix of Antennapedia homeodomain and a phospholipid monolayer, studied by ellipsometry and PM-IRRAS at the air–water interface, Biochim. Biophys. Acta 1467 (2000) 131–143. M. Magzoub, L.E. Eriksson, A. Graslund, Comparison of the interaction, positioning, structure induction and membrane perturbation of cell-penetrating peptides and non-translocating variants with phospholipid vesicles, Biophys. Chem. 103 (2003) 271–288. A. Thomas, R. Meurisse, B. Charloteaux, R. Brasseur, Aromatic side-chain interactions in proteins. I. Main structural features, Proteins 48 (2002) 628–634.
S. Deshayes et al. / Biochimica et Biophysica Acta 1778 (2008) 1197–1205 [39] M. Lindberg, J. Jarvet, U. Langel, A. Graslund, Secondary structure and position of the cell-penetrating peptide transportan in SDS micelles as determined by NMR, Biochemistry 40 (2001) 3141–3149. [40] E. Barany-Wallje, A. Andersson, A. Graslund, L. Maler, NMR solution structure and position of transportan in neutral phospholipid bicelles, FEBS Lett. 567 (2004) 265–269. [41] E. Barany-Wallje, A. Andersson, A. Graslund, L. Maler, Dynamics of transportan in bicelles is surface charge dependent, J. Biomol. NMR 35 (2006) 137–147. [42] C. El Amri, F. Bruston, P. Joanne, C. Lacombe, P. Nicolas, Intrinsic flexibility and structural adaptability of Plasticins membrane-damaging
1205
peptides as a strategy for functional versatility, Eur. Biophys. J. (2007) (Electronic publication ahead of print). [43] R. Brasseur, M. Deleers, W.J. Malaisse, J.M. Ruysschaert, Conformational analysis of the calcium-A23187 complex at the lipid–water interface, Proc. Natl. Acad. Sci. U. S. A. 79 (1982) 2895–2897. [44] R. Brasseur, M. Deleers, Conformational analysis of 6-cis- and 6-trans leukotriene B4- Calcium complexes, Proc. Natl. Acad. Sci. U. S. A. 81 (1984) 3370–3374. [45] D. Persson, P.E. Thoren, B. Norden, Penetratin-induced aggregation and subsequent dissociation of negatively charged phospholipid vesicles, FEBS Lett. 505 (2001) 307–312.
