Slowing down single-molecule trafficking through a protein nanopore reveals intermediates for peptide translocation

Transcription

1 Supplementary information for Slowing down single-molecule trafficking through a protein nanopore reveals intermediates for peptide translocation Loredana Mereuta 1,*, Mahua Roy 2,*, Alina Asandei 3, Jong Kook Lee 4, Yoonkyung Park 4, Ioan Andricioaei 2,#, Tudor Luchian 1,# 1 Department of Physics, Alexandru I. Cuza University, Iasi, Romania 2 Department of Chemistry, University of California, Irvine CA 92697, USA 3 Department of Interdisciplinary Research, Alexandru I. Cuza University, Iasi, Romania 4 Research Center for Proteineous Materials, Chosun University, Gwangju, South Korea * These authors contributed equally to this work # Corresponding authors (andricio@uci.edu, luchian@uaic.ro) Uni-dimensional formalism for the derivation of the drift velocity Assuming negligible contributions from pressure gradients, the uni-dimensional drift velocity of a peptide in the electric field along the protein vestibule is the vector sum of the electrophoretic and the electro-osmotic components, whose lumped modulus can be expressed as (1, 2, 3): v drift = v electrophoretic + v electroosmotic = ε η ζ peptide ζ vestibule E (1) where ζ peptide and ζ vestibule refer to the zeta potential of the peptide and inner vestibule, E represents the electric field intensity along the movement pathway, parallel to the pore s axial direction, ε and η represent the absolute dielectric constant of water and its viscosity, respectively. Due to the fact that in practice, the zeta potential is usually evaluated by using electrophoretic light scattering, a more convenient and practical approach toward evaluating v electroosmotic proceeds indirectly, via the estimation of the water flux from the ion current through the pore. Qualitatively, by neglecting the interaction of the applied electric field with the counterions in the electrical double layer around the peptide, and within the simplifying assumptions that underlie the Goldman Hodgkin Katz formalism for the independent movement of K + and Cl - ions with no concentration gradient, under an applied potential ( V) that creates a uniform electric field inside the nanopore, and considering that the monovalent cations and anions 1

2 (z K + = 1 and z Cl - = -1) migrate in opposite directions along the electric field lines with distinct permeability (P K+ <P Cl ), the net number of particles flowing towards the positively biased region, through a cross-sectional area (S) in unit time is, dn = F N dt RT A VCS P Cl- -P K+ (2) where F represents the Faraday constant, R represents the molar gas constant, T represents the absolute temperature, N A represents the Avogadro s number and C is the molar concentration of particles. Within the same formalism, the net electric current (I) transported through a cross-sectional area (S) reads: I = e- dn dt = F RT F VCS P Cl- + P K+ (3) where e - represents the electronic charge, and the rest of the terms have the same meaning as above. By combining equation (2) and (3), one obtains: dn dt = I P Cl- -P K+ e - P Cl- +P K+ (4) Further on, knowing the number of water molecules that are associated with each ion which moves through the pore (N h ), the net number of water molecules (N water ) passing through can be inferred: dn water dt = N hi e - P Cl- -P K+ P Cl- +P K+ (5) and the resulting electro-osmotic velocity of water through the nanopore (v electroosmotic ) becomes: v electroosmotic = N hi P Cl- -P K+ 1 e - P Cl- +P K+ S pore [H 2 O] (6) and when applied to the vestibule region, S pore represents the average cross-sectional area of the α -HL vestibule, and [H 2 O] the concentration of water. Given that the amount of water molecules carried by each ion (N h ) is being estimated herein as a number of particles, for further numerical estimations (vide infra) the water concentration ([H 2 O]) number of particles number of moles value shall be expressed as instead of. In deriving equation m 3 m 3 (6), used the relation J = Cv derived from the continuity equation, describing quantitatively the 2

3 flux of particles (J) of local concentration (C), moving with the velocity (v); for a comprehensive reference source, see (1). If the peptides were to move along the protein vestibule only under the influence of the applied electric field, its average, electrophoretic velocity (v electrophoretic ) would be: v electrophoretic = µe (7) where µ represents the electrophoretic mobility of the peptide within the vestibule, and E the intensity of the electric field. By combining relations (6) and (7), and taking into account that in our particular situation the vector representing electro-osmotic water velocity is oriented opposite to the peptide electrophoretic velocity vector - this is caused by the fact that the protein is anion selective, so that the positively charged peptide moves trans-to-cis, opposite to the net flow of water carried by cis-to-trans moving anions - one may estimate the modulus of the uni-dimensional drift velocity of a peptide moving along the electric field lines, in the trans to cis direction within the protein vestibule, as: v drift = v electrophoretic -v electroosmotic = µe- N hi P Cl- -P K+ 1 e - P Cl- +P K+ S pore [H 2 O] (8) If we disregard any drop of the potential outside the protein pore (β-barrel and vestibule combined) of length (l pore ), the electric field will be non-zero only within the protein pore subjected to the potential ( V), and the relation above reads: v drift = µ V l pore - N hi e - P Cl- -1 P K+ P Cl- +1 P K+ 1 S pore [H 2 O] (9) 3

4 Figure S1. Selected electrophysiology segments showing the reversible interaction between a single α-hl protein and a CAMA P6 peptide added on the grounded, cis side of the membrane, at a bulk concentration of 30 µm. The clamping trans-membrane potential was -50 mv, and the buffer ph was set to 4.5. The protein and peptide cartoon-like representations illustrate the peptide entering the protein vestibule from the cis side of the membrane associated to transition O B 2 (panels a and b), the peptide exiting the pore through the trans side of the membrane via protein s β-barrel (transition B 1 O) (panel a), or the peptide exiting the vestibule by moving backwards to the cis side of the membrane, respectively (transition B 2 O) (panel b). The peptide interaction with the protein pore is denoted by the upwardly-pointing current fluctuations from the base current assigned to the peptide-free, open (O) state of the protein, and current amplitude histograms shown on the right-hand-side of the traces in both panels illustrate the absolute values of current associated to distinct substates of the pore, i.e. in the open (O), B 1 and B 2 substates. As opposed to the cases when the peptide was added to the trans side of the membrane (see main text), in all instances, the initial blockage state induced by the electrophoretic driven peptide towards the protein corresponds to the lower blockage level (denoted by B 2 ), whose relative amplitude (ΔI B 2 = I B2 - I O ) equals that measured with the peptide added on the trans side of the membrane, and related to the same substate. This substantiates the fact that the B 2 substate corresponds to the temporarily occlusion α-hl vestibule, first accessible to an incoming peptide, which thereafter moves along the protein pore and reaches its β-barrel, thus giving rise to the higher blockage level (denoted by B 1 ). The reversible transitions between the B 1 and B 2 blockage substates constitute direct observations of 4

5 the sequence of spatio-temporal events whereby the peptide, initially captured in the pore s vestibule, migrates along the protein inner space and dwells temporarily in the pore s vestibule or its β-barrel region. As it is visible from such representative trace segments, in certain instances peptide release from the pore (substate O) proceeds through substate B 1 (Fig. S1, panel a) or substate B 2 (Fig. S1, panel b), and this allows for a direct indication regarding the direction in which the initially vestibule-trapped peptide get released, namely to the trans (Fig. S1, panel a) or cis (Fig. S1, panel b) side of the membrane. Figure S2. Selected electrophysiology segments recorded at + 50 mv, showing the transient current blockages - the downwardly pointing electric current departures from the baseline measured through the peptide-free (O) protein-induced by a CAMA P6 peptide, added on the trans side of the membrane at a bulk concentration of 30 µm. The buffer ph was set to 3.3 (panel a) and 2.2, respectively (panel b). Peptide capture by the β-barrel gives rise to the B 1 blockage substate, and this is followed sequentially by the B 2 substate, denoting migration of the peptide to the protein vestibule. Notably, and taking into considerations all arguments presented herein, in both panels a and b we were able to ascertain that upon peptide capture by the protein β-barrel (blockage substate B 1 ) on the trans side, and subsequent transition of the peptide to the protein s vestibule (blockage substate B 2 ), the current level corresponding to the open pore (substate O) is associated to the peptide release to the cis side of the membrane. Unlike experiments performed at less acidic ph s (see main text), a peptide confined within the protein vestibule (i.e., associated to the B 2 substate) did not get released at once from the protein pore, as the B 1 blockage substate ensued again, indicating the backwards movement of the peptide to the protein β-barrel (panel a). Only after several such reversible transitions between the B 1 and B 2 substates, the pore enters the open substate (O), associated to the peptide release from the protein. At even lower ph values (ph = 2.2; panel b), our data show that although following its capture by the protein β-barrel, the trans-added peptide manages to exit the protein vestibule at once (i.e., the very first sequence of transitions O B 1 B 2 O, red-colored), it re-enters the protein vestibule, getting trapped again in the B 2 substate (panel b, the blue-colored sequence). Inset pictures were used to draw the reactions schemes corresponding to traces shown in panels a and b, emphasizing the microscopic correspondence between various substates (O, B 1 5

6 and B 2 respectively) and position of the peptide relative to the α-hl pore and the reversible reactions taking place between various substates (vide supra also). Figure S3. Representative snapshots of peptide translocation when peptide is caught within the maximum construction of the pore. As observed, the peptide assumes either a compact, folded shape or a partial unfolded shape depending on the initial configuration of the peptide. The figures represent the initial configurations shown in Figure S5 - (1) s1f (2) s2u (3) s2f (4) s2u. Excluded Cross-sectional area and Electric Current The excluded cross-sectional area (S'(z)) of the pore along the pore axis (z), being proportional to the flux of ions and hence ionic current detected at an instant was calculated using CHARMM simulation package. For each angstrom along the pore axis, z, the free volume was computed (using CORMAN module) which corresponds to excluded cross sectional area. The ratio of open pore current intensity I c (t) vs. clogged pore current intensity I o (t) (Equation 10 and Fig. S4) gives the residual current, I ( t) c = I ( t) o L dz d d + ( 1 2 ) 0 S( z) L dz d ' d ' + ( 1 2 ) 0 S'( z) 2 (10) where, L is the extent of the pore along the pore axis, z, with d 1 and d 2 the "entrance diameters" (effective diameters) of the β-barrel and vestibule facing the solution. d 1 'and d 2 ' are the effective entrance diameter of the clogged pore. 6

7 Figure S4. Graphical representations of cross-sectional diameter for the (a) open pore and clogged pore in the (b) β-barrel and the (c) vestibule respectively. Figure S5. (a) Residual current calculated using Equation (10) for the complete translocation trajectory for four starting conformation of the peptide within the pore. (b) Two folded states (s1f and s2f) and two unfolded states (s1u and s2u) were taken as the starting conformation for initiating translocation. Orange and purple VDW representations show the C-terminus and N-terminus respectively. The initial state s1f when unfolded within the pore using HQBM (described later) unfolds with C-terminal facing up towards 7

8 the trans-end whereas the state s2f unfolds with N-terminus up. Table S1: The residual current values tabulated from experiment and simulation for both states B1 and B2. The experimental values correspond to different ph whereas the simulation values to the four starting conformations identified in Figure S5 (b). The simulation values correspond to an average of an ensemble of states captured during the translocation trajectory. We attribute the slightly higher values in case of state 2 for both folded and unfolded state to the opening of N-terminus and the peptide translocating with its N-terminus first whereas in case of state 1, the peptide translocates with its C-terminus up (details included in the SI movies). Effective diffusion coefficient estimate in the vestibule The measured voltage-dependence of the duration of the lower blockage level (B 2 ), i.e., of the average peptide sojourn times within the vestibule, depends on peptide drift velocity. Knowing that the electrophoretic mobility of the peptide (µ) can be expressed through its diffusion coefficient (D) as μ = (Z e D)/kT, by replacing in Eq. (1), main text, numerical values for all constants involved (Z = +8, T=296 K, l pore =10*10-9 m, N h = 10, P Cl /P K+ =2.27 at ph=4.5, S pore = 16.61*10-18 m 2, [H 2 O] = 3.4*10 28 m -3, l vestibule = 5*10-9 m), using I = g V, where the experimentally estimated unitary conductance value corresponding to the B 2 substate, while a 9 peptide resides within the α-hl vestibule at ph=4.5, reads g = S, we further employed Eq. (2), main text, to make a numerical estimation of the diffusion coefficient value of the peptide within the protein vestibule (D) at ph = 4.5. By non-linear fitting of data shown in Fig. (5) panel d, with the function given by Eq. (2), main text, which expresses the rate off B 2 vs. V dependency, we arrived at D = 1.5*10-12 m 2 s -1. Estimates of peptide diffusion in vestibule from current blockage For the sake of quantitative comparison, by regarding the peptide within the vestibule as a sphere, the expression ΔI B 2 = γσ V 2 δ volume 2, where in order to take into consideration the access l vestibule resistance of the nanopore to the vestibule, its length (l vestibule ) is replaced by (l vestibule + 0.8d vestibule ; l vestibule = 5 nm and d vestibule = 4.6 nm)), σ = 169 ms cm -1 and the shape factor γ = 1.5, 8

9 allows to estimate the volume of the vestibule-confined peptide at ph = 4.5 and an applied potential V = +50 mv, from the measured ΔI B 2 (ph = 4.5) = 34.7 pa, at a value of δ volume = 4.13 nm 3. In doing so, one may further invoke the Stokes-Einstein equation (D = kt 6πηr ) for the diffusion of a spherical particle of radius r (r = 3δ volume ) through bulk water of viscosity η =10-3 Pa s, resulting in the diffusion coefficient of the CAMA P6 peptide at T = 296 K, of D = 2.2*10-10 m 2 s -1. This indirect result is about two orders of magnitude greater than the value obtained experimentally under similar conditions (D = 1.5*10-12 m 2 s -1, see above and discussion in main text), strengthening the fact that friction forces manifested between the peptide and protein s inner walls and a modified water viscosity within the protein pore as opposed to bulk solution greatly affect peptide diffusion across the protein vestibule. Based on the Stokes- Einstein formalism applied to the peptide movement across the protein vestibule at ph = 4.5 (vide supra), we estimated that in order for the peptide to attain a diffusion coefficient comparable in order of magnitude with that inferred experimentally (D ~ m 2 s -1 ), the apparent viscosity of water within the protein vestibule should equal η ~ 0.2 Pa s. This result comes in agreement with previous research establishing that the manifestation of the electrical double layer, particularly in nanoscopic volumes, leads to an increase in the apparent viscosity of a fluid in nano-channels (4), and for water films with a thickness smaller than 2 nm, studied on hydrophilic surfaces, it was found that the apparent viscosity was ~ 4 orders of magnitude higher than in bulk (5). Although the precise evaluation of the viscosity of water within the α-hl protein needs further refinement and validation, the presented rationale highlights a novel experimental approach towards evaluating solvent properties in confined nano-volumes. Furthermore, as it is visible from data shown in Fig. 6, panel f (main text), the ph decrease alters also the value of the current blockade amplitude ( I B 2 ) associated to the peptide presence within the α-hl vestibule. By resorting to a similar formalism as described above, and using characteristic geometric values describing the α-hl vestibule (l vestibule = 5 nm, d vestibule = 4.6 nm), the amplitude of ion current blockages mediated by a peptide inside the α-hl vestibule ( I B 2 ) measured at an externally applied potential difference V = +50 mv, at ph 3.3 and 5.1, resulted in δ volume; ph = 3.3 = 4.6 nm 3 and δ volume; ph = 5.1 = 2.5 nm 3. It is thus apparent that in addition to the electro-osmotic effect described herein, a still elusive low ph-mediated increase in the peptide excluded volume within the protein vestibule may also account partially for the slowing-down of the CAMA P6 peptide transit through the α-hl vestibule with decreasing ph. As a word of caution, we note that such estimations on the peptide excluded volume were based on the assumption that the peptide assumed a spatial topology of a sphere under the externally applied potential difference V, so that the shape factor (γ) equals 1.5. However, various orientations relative to the pore s axis of the peptide during its translocation, or peptides shapes for that matter, result in distinct values of the shape factor (6). 3 4π 9

10 Figure S6. Protonated residues D13, D2, D4, D227 on each of the seven monomers of the α - HL homoheptamer (shown as a cross section) at acidic ph in-turn leads to positive charge on residues K8, R56, R104, K154 making the α-hl vestibule positively charged at acidic ph. Figure S7. Voltage dependence of the association rate reflecting the interaction of a single CAMA P6 peptide with the α-hl protein (O B 1, rate on, ) (panel a) and the rate characterizing the B 1 B 2 transition (rate off B 1, ) (panel b), estimated at ph = 4.5. CAMA P6 peptide was added to the trans chamber at a concentration of 30 μm, and α-hl protein was present on the cis (grounded) chamber. Dotted lines in (a) and (b) represent the 95 % confidence domain for rate values. 10

11 Figure S8. Representative segments from electrophysiology experiments undertaken at ph = 9.17, showing the interaction, reflected by the downwardly oriented spikes, of a CAMA P6 peptide with a single α-hl protein, measured at V = + 50 mv. Dotted lines in panels (a, b) show level of open-pore current before (the open state, O) and after a peptide partitioned within the β-barrel (blockage B 1 ). Downward spikes reflect reduction of pore current induced by the reversible association of a peptide with an open protein pore (blockage B 1 ). The zoomed-in trace segments in panel (b) reveal the distinct blockage sub-state B 1 and lack of blockage sub-state B 2 associated with a single peptide in the pore (see also text). The distinct B 1 sub-state is visualized in panel (c) scatter plot of dwell time vs. relative blockage amplitude of blockage events (ΔI block = I B1 - I O ). Estimates of ph dependence of peptide excluded volume in the β-barrel (see Fig. 6 in main text) The effective blocking peptide volume within pores may be found from analyzing current blockage by varying several working conditions (voltage, ph, etc.). Such measurements at constant ph but variable voltage implied peptide unfolding by stretching caused by the applied potential (7). We analyzed the dependence of the ionic current blockade amplitude ( I B 1 ) induced by the presence of a peptide inside the pore β-barrel as a function of ph at constant applied voltage. Within the framework of a similar mechanistic interpretation as the study above, the rationalization of our results on the dynamics of the CAMA P6 peptide through the protein β- barrel and blockage induced, as a function of the ph, indicates that elevated values of the electro- 11

12 osmotic flux acting collectively and oppositely to the electrophoretic force, create local forces that can induce gradual changes of the folding of a single confined peptide, along the axis of the protein pore. With the help of the data shown in Fig. 6, panel f (main text), the excluded volumes (δ volume ) associated to the peptide presence within the protein β-barrel at various ph values, estimated from the amplitude of ion current blockages (ΔI B 1 ), recorded at an applied potential difference V = +50 mv, in a 2 M KCl buffer of electrical conductivity σ = 169 ms cm -1 (ΔI B1 = γσ V 2 δ volume l β-barrel 2 ), where the shape factor γ = 1 (based on the assumption that the β-barrel - residing peptide can be viewed as a cylinder aligned parallel to the electric field lines, mainly due to the restrictive topology of the β-barrel) and within over simplifying physical and geometrical considerations (e.g. the potential varies linearly across the α-hl pore, so we assumed that the potential difference drop across the protein β-barrel is half of the potential difference across the whole protein pore, V), resulted in δ volume; ph = 3.3 = 6.5 nm 3, δ volume; ph = 4.5 = 7.1 nm 3, δ volume; ph = 5.1 = 7.2 nm 3, δ volume; ph = 7.1 = 7.5 nm 3. To take into consideration the access resistance of the nanopore, in the calculations above the value of the β-barrel length (l β- barrel) was replaced by (l β-barrel + 0.8d β-barrel ), where d β-barrel represents the average diameter of the α-hl s β-barrel (d β-barrel = 2 nm) and l β-barrel = 5.2 nm. By comparison, the theoretical volume value of the CAMA P6 peptide in the extended conformation, estimated with a protein property calculator (e.g., (8)) equals δ volume; theory = 2.9 nm 3. It is apparent that the higher volume of the peptide within the protein β-barrel at neutral ph (δ volume; ph = 7.1 = 7.5 nm 3 ) reflects the partially folded configuration assumed by the peptide, as it is also indicated by molecular simulation studies. As a word of caution, we note that such estimations on the peptide excluded volume were based on the assumption that the peptide assumed a spatial topology of a cylinder aligned parallel to the electric field lines through the pore, under the externally applied potential difference V, so that the shape factor (γ) equals unity. However, various orientations relative to the pore s axis of the peptide during its translocation, or peptides shapes for that matter, result in greater than one values of the shape factor (6). 12

13 Figure S9. Higher charged peptides move faster through vestibule. Representative segments from electrophysiology experiments undertaken at ph = 4.5, showing the interaction, reflected by the downwardly oriented spikes, of a CAMA P6 (effective charge ~ 8 e - ; panel a), CAMA P1 (effective charge ~ 9 e - ; panel b) and CAMA P5 (effective charge ~ 10 e - ; panel c) peptide with the α-hl protein pore immobilized in a planar lipid membrane, at an applied potential of + 50 mv. The bulk concentration of all peptides added on the positively biased, trans side of the membrane, was 30 µm. In panels below are displayed the scatter-plots of the dwell time - relative blockage amplitude joint distributions of various peptide-induced blockage events (panel d, peptide CAMA P6; panel e, peptide CAMA P1; panel f, peptide CAMA P5), capturing the distinct fingerprint of the blockages induced by such peptide construct on the open-state α-hl current. For brevity, the distinct blockage events associated to B 1 and B 2 blockage substates are shown for the CAMA P6 peptide alone (panel d; the ellipse-like domains), although the corresponding peaks are visible for the other peptides, as well. 13

14 Peptides Sequence pk a (K) = 10.5 pk a (H) = 6 Net charge ph = 4.5 ph = 7.1 ph = 9.17pH = 14 CAMA P6 KWKLFKKIGIGKFLQSAKKF-NH CAMA P1 KWKLFKKIGIGKHFLSAKKF-NH CAMA P5 KWKHLKKIGIGKHFLSAKKF-NH Table S2. The net charge of the peptides used in this study (CAMA P6, CAMA P1 and CAMA P5), at various ph values, calculated with the Innovagen's peptide calculator (9). Peptides Sequence Retention time Molecular Weight Observed Calculated CAMA P1 KWKLFKKIGIGKHFLSAKKF-NH CAMA P5 KWKHLKKIGIGKHFLSAKKF-NH CAMA P6 KWKLFKKIGIGKFLQSAKKF-NH Table S3. The primary sequence and molecular weight of peptides used in this study. 14

15 15

16 Figure S10. Mass analysis of peptides used in this study. Molecular masses of peptides were determined by using MALDI-MS spectrometry. 16

17 Figure S11. Representative normalized histograms of time intervals associated to the B 1 B 2 transition (τ off B 1 ) and B 2 O transition (τ off B 2 ), respectively, collected at ph = 7.1 (panel a), ph = 4.5 (panels b and c), and ph = 3.3 (panels d and e). The applied transmembrane potential was V = + 50 mv. The dotted line represent the non-linear fit with a decaying mono-exponential function (y ~ exp(-t/τ off )). Biased Molecular Dynamics. Unfolding of the peptides within the pore constriction is effected by adding a half quadratic perturbation term to the Hamiltonian, dependent on both time and reaction coordinate, forcing the peptide to above away from its initial configuration using a half harmonic force constant. The time-dependent reaction coordinate is chosen to be the square of deviation from the reference structure for all atoms averaged over the number of pairs. (11) where, r i j ( t) = r ( t) r ( t) (12) i j for, j i and r ij (0) is the pairwise distance in the native or reference structure. To force the system away from the native structure, a perturbation dependent on time and position, W(r,t) is added to the potential energy function, to prevent its backward movement and to prevent the reaction coordinate from decreasing. 17

18 if (13) where, ρ m (t) is the maximum value of the reaction coordinate between time 0 and t. The magnitude of allowed backward fluctuation depends on α, the force constant for the half harmonic potential. A high value of α acts as a stiff force constant allowing no backward fluctuation whereas a low value can allow flexibility. If the reaction coordinate exceeds the maximum value then the perturbation is zero, perturbation being time-dependent acts only when the reaction coordinates decreases than the maximum allowed value. If the reaction coordinate ρ (t) exceeds the maximum value ρ m (t) the perturbation W(r, t) is zero. A large value of perturbation W (ρ(t)) indicates barrier crossing during which spontaneous thermal motion is incapable of carrying the reaction forward along the reaction coordinate. Hence the force due to perturbation is applied only when the reaction coordinate is less than the maximum. Supplementary Movies. Complete translocation movies are included, with their description as follows: (a) pore-l.gif Identifies the pore-lining residues during the translocation trajectory. All pore residues except the pore-lining residues are fixed in coordinate space during a translocation event. The pore residues are color coded according to the residue type. (b) s1f.gif Translocation trajectory for the folded peptide in state 1 with N- and C-terminus pointed towards the cis end (c) s1u.gif Translocation trajectory for the unfolded peptide in state 1. Unfolding is achieved using the biased MD method (see SI text) followed by translocation using electric field. The unfolding trajectory has not been shown. (d) s2f.gif Translocation trajectory for the folded peptide in state 2 with N- and C-terminus pointed towards the trans end (e) s2u.gif Translocation trajectory for the unfolded peptide in state 2. For states 1, the unfolding and translocation takes with C-term going head up whereas for states 2, the unfolding and translocation takes with N-term going head up. We would like to emphasize the fact that the unfolding (not shown) and translocation in both states are completely random following Langevin dynamics and has not been forced with either N-term or C-terminus heads up. The pore lining residues though not shown in motion in the translocation movies are in the same state as shown in the movie pore-l.gif throughout the trajectory. Supplementary References 1. Schoch, R. B., Han, J. & Renaud, P. Transport phenomena in nanofluidics. Rev. Mod. Phys. 80, (2008). 2. Davenport, M. et al. The Role of Pore Geometry in Single Nanoparticle Detection. ACS Nano 6, (2012). 3. Bard, A. J. & Faulkner, L. R. Electrochemical Methods: Fundamentals and Application (2nd ed. New York, Wiley 2001). 18

19 4. Yang, C. & Li, D. Electrokinetic effects on pressure-driven liquid flows in rectangular microchannels. J. Colloid Interface Sci.194, (1997). 5. Li, T.-D. et al. Structured and viscous water in subnanometer gaps. Phys. Rev. B. 75, (2007). 6. Yusko, E. C. et al. Controlling Protein Translocation Through Nanopores with Bio-Inspired Fluid Walls. Nat. Nanotechnol. 6, (2011). 7. Freedman, K. J. et al. Single molecule unfolding and stretching of protein domains inside a solid-state nanopore by electric field. Sci Rep. 3, 1638 (2013) ( 19