Electrostatic Map Of Proteasome α-Rings Encodes The Design of Allosteric Porphyrin-Based Inhibitors Able To Affect 20S Conformation By Cooperative Binding

The importance of allosteric proteasome inhibition in the treatment of cancer is becoming increasingly evident. Motivated by this urgent therapeutic need, we have recently identified cationic porphyrins as a highly versatile class of molecules able to regulate proteasome activity by interfering with gating mechanisms. In the present study, the mapping of electrostatic contacts bridging the regulatory particles with the α-rings of the human 20S proteasome led us to the identification of (meso-tetrakis(4-N-methylphenyl pyridyl)-porphyrin (pTMPyPP4) as a novel non-competitive inhibitor of human 20S proteasome. pTMPyPP4 inhibition mechanism implies a positive cooperative binding to proteasome, which disappears when a permanently open proteasome mutant (α-3ΔN) is used, supporting the hypothesis that the events associated with allosteric proteasome inhibition by pTMPyPP4 interfere with 20S gating and affect its “open-closed” equilibrium. Therefore, we propose that the spatial distribution of the negatively charged residues responsible for the interaction with regulatory particles at the α-ring surface of human 20S may be exploited as a blueprint for the design of allosteric proteasome regulators.


S3
diethyl ether, then with hexane to yield the desired N-alkylated porphyrin (2) as a purple solid (168 mg, 95% yield). Mp > 300 °C. UV-vis (DMSO): λ max , nm (log ε): 428 (5.30), 520 (4.28), 555 (4.13), 594 (3.89) 644 (3.85). 1  Proteasome activity assay. 2 nM of 20S proteasome was incubated with increasing concentrations of inhibitor (ranging from 0.5 to 3 µM) for 30 min at 37°C in the assay buffer (50 mM HEPES, 1mM EDTA, pH7.5) and, subsequently, 100 µM of AMC-labeled substrate peptide was added Latent Core Particles (CP). Proteasome activity was monitored by measuring the AMC fluorescence at 440 nm (excitation at 360nm) for 20 min, using a fluorescence plate reader (Multiskan, Thermo) in a 384 multiwell black plate. A minimum of three replicates were performed for each data point. Fluorescent substrate cleavage by the 20S proteasome was linear during this incubation time frame. Data are expressed as normalized percentages of residual activity considering the slope of the control (fluorogenic peptide/proteasome in the absence of inhibitors) as 100% of proteasome activity. Dose-response plots of the residual proteasome activity in the presence of increasing concentration of inhibitor provides a quantitative estimate of its potency. The IC 50 is defined as the concentration of the inhibitor which causes 50% reduction of activity and it is thus calculated from the x-axis value on a semilog scale of the dose−response plot occurring at a fractional activity of 50%. The estimation of the IC 50 is based on a nonlinear fit with a sigmoidal functions expressed by the equation: The midpoint of this sigmoidal function occurs at a fractional velocity value of 50%, corresponding to half inhibition of the target enzyme.
Data analysis. To visualize the effect of the inhibitor concentration the conversion rate (normalized for the enzyme concentration, v/[E tot ]) was measured as a function of substrate concentration (between 6 µM and 100 µM) for each inhibitor concentration (I) and analyzed as a double-reciprocal Lineweaver-Burk plot where [E 0 ] is the total enzyme concentration, v (moles of substrate / (volume x time)) is the enzymatic rate, [S] is the substrate concentration, k cat is the speed of the rate-limiting step and K m is the Michaelis-Menten constant, corresponding to the enzyme:substrate affinity. In order to distinguish among the different inhibition mechanisms, data were fitted according to Eqs. (S1-S3).
All curve fitting and statistical analysis were carried out using the Non Linear Fitting Tool (NLFit) in Origin7. The parametric data fitting was based on nonlinear regression and the method of least squares. Model discrimination and choice was based on the goodness of fit. The goodness of fit was evaluated by visual examination of the fitted curves, 95% confidence bounds for the fitted coefficients and statistical analysis for determining the square of the multiple correlation coefficient (R 2 ).
Data analysis. Data analysis of enzymatic assays has been accomplished according to the nonlinear least-squares fitting of Lineweaver-Burk equations assuming either a competitive inhibition equation: in which k cat is the rate-limiting step of the reaction, K i is the dissociation constant for inhibitor I, K M is the Michaelis-Menten constant for substrate binding, α is an interaction parameter (reflecting the difference of inhibitor affinity between the free enzyme E and the enzyme-substrate ES complex), [S] and [I] are the substrate and inhibitor concentrations, respectively.
The three different mechanisms reflect three different ways by which an inhibitor acts on the enzyme activity, namely (i) a competitive inhibitor interacts only with the free enzyme, (ii) an uncompetitive inhibitor binds only to the enzyme-substrate complex, and (iii) a mixed inhibitor reacts with both the free enzyme and the enzyme-substrate complex, the α parameter indicating the different affinity for the two forms. When α is very large, binding of inhibitor impairs binding of the substrate and the mixed-model becomes identical to competitive inhibition. When α = 1 the mixed model becomes a purely non-competitive model.

Cooperativity for Porphyrin Binding.
In the presence of (at least) two states (i.e., "open" and "closed" in equilibrium) in 20S proteasome, the observed cooperative binding by 2 (see Fig. 7 where "n" is the total number of porphyrins which can bind to that cluster in a concerted way. It is important to remark that application of these thermodynamic equations to the kinetic behaviour implies that the rate of the "open"-"closed" conformational transition is faster than the bimolecular S5 porphyrin binding. Employing values of K c , K o and L, reported in Table S13, it is possible to describe the dependence of k obs on porphyrin 2 concentration according to Eq. (1b) (see Figs. 8 and 9), applying values of k c and k o , reported in Table S13.
Stopped-flow kinetic experiments. Experiments have been carried out by rapid-mixing experiments using the SX18.MV stopped-flow apparatus (Applied Photophysics, Salisbury, UK) equipped with a diode array for spectra acquisition over a 1 ms time range. Kinetics has been investigated by mixing 1 nM 20S proteasome with different concentrations of porphyrin(s) and following optical density changes as a function of time. Kinetic progress curves at selected wavelengths were analyzed according to the following equation: where OD obs is the observed optical density at a selected wavelength and at a given time interval, OD 0 is the optical density at t = 0, n is the number of exponentials, DOD i is the optical density change associated to the exponential i, i k obs is the observed rate constant of the exponential, and t is the time.
Native Gel Electrophoresis. The native gel analysis was performed according to the procedure described elsewhere. 3 1 µg of the purified 20S proteasome was resolved under native conditions after a pre-incubation with the indicated concentration of 2, ranging from 0.3 to 10 µM in 25 mM Tris-HCl (pH 7.5). Thereafter, the particles were probed with 100 µm Suc-LLVY-AMC. It is worth recalling that while at 1 and 10 µM of 2 we observed an inhibition of the proteolytic activity of the 20S in the fluorimetric approach, these concentrations appear to be linked to an activation effect if the extent of degradation of the fluorogenic peptide is compared to that of the 20S in the absence of 2 by native gel. This behaviour is likely attributable to the highly different experimental conditions between the fluorimetric assays and the native gel. The hypothesis concerning the existence of two clusters of sites on the 20S proteasome finds further correspondence with the data obtained by native gel electrophoresis. (see Figure S10), where a significant activation of the proteolytic activity of the purified 20S proteasome on the Suc-LLVY-AMC substrates was seen at 0.3 µM 2 followed by a progressive decrease of the proteolytic activity at higher p-TMPyPP4 concentrations.

Modeling of human 20S proteasome in the open conformation.
According to the reference structure (PDB ID: 5T0J), it must be highlighted that the molecular model of 20S in the open conformation was built considering only two packed rings (i.e., one α subunit ring and one β subunit ring). The molecular model of human 20S proteasome in the open conformation was built starting from the experimentally determined structure of human 20S proteasome (PDB ID: 5T0J), which lacks the Nterminal regions (α1: aa1-4; α2: aa1-2; α3: aa1; α4: aa1-2; α5: aa1-8; α6: aa1-3; α7: aa1-5), Cterminal regions (α1: aa245-246; α3: aa252-261; α4: aa241-248; α6: aa242-263; α7: aa246-255; β1: aa192-205; β2: aa221-234; β4: aa200-201; β5: aa202-204; β7: aa216-219), one loop region (α5: aa128-133), and the side chains of several residues (α1: The sequence of 5T0J was aligned with the sequences of human 20S proteasome downloaded from the UniProtKB/Swiss-Prot Data Bank (http://www.uniprot.org; entry P60900 (α1); P25787 (α2); P25789 (α3); O14818 (α4); P28066 (α5); P25786 (α6); P25788 (α7); P28072 (β1); Q99436 (β2); P49720 (β3); P49721 (β4); P28074 (β5); P20618 (β6), and P28070 (β7)) by using the Multiple_Alignment algorithm (Homology module, Accelrys, San Diego). Subsequently, the secondary structural prediction of human 20S proteasome was performed using the Structure Prediction and Sequence Analysis server PredictProtein (http://www.predictprotein.org/). The coordinates of the structurally conserved regions (α1: aa5-244; α2: aa3-234; α3: aa2-251; α4: aa3-240; α5: aa9-127 and aa134-241; α6: aa4-241; α7: aa6-245; β1: aa1-191; β2: aa1-220; β3: aa1-204; β4: aa1-199; β5: aa1-201; β6: aa1-213; β3: aa1-215) were accordingly assigned by the Structurally Conserved Regions (SCRs)-AssignCoords procedure (Homology Module, Insight 2005) using 5T0J as template structure. The lacking loop segment in the α5 subunit (aa128-133) was inserted by using the Generate Loops procedure. With the Generate Loops procedure, a peptide backbone chain is built between two conserved peptide segments using randomly generated values for all the loops' φ's and ψ's. The chain was defined starting from the N-terminal end of the loop being built; the Start and Stop Residues were defined as the SCR residues of the model protein at either end of the loop itself. The geometry about the base was described by the four distances between Cα and Ntermini of the Start residue and the Cα and C-termini of the Stop Residues. In the process of closing the loop, the values for the generated φ's and ψ's are adjusted until the four distance criteria are met. Specifically, a function was defined for the distances in terms of the dihedral angles (Scale Torsions: 60). The differences between the desired distances and their current values were minimized using a linearized Lagrange multiplier method. After a series of 1000 iterations, the loop was closed, except in the case where the distances between the ends of the loop were not respected (Convergence = 0.05). The geometry at the base of the loop is then checked for proper chirality. Finally, the loops were screened on the basis of steric overlap violations. All loops that are found to have unacceptable contacts were rejected. Since successive calculations can correct some bad contacts, a fairly large overlap factor was used (Internal and External overlap = 0.6). A bump check of the 10 generated loops together with the evaluation of their conformational energy were used as selection criteria. The lowest conformational energy loop presenting no steric overlap with the rest of the protein, was selected. Finally, the coordinates of the lacking N-terminal and C-terminal amino acids were assigned using the EndRepair command (Homology Module, Insight 2005). The obtained homology model was completed inserting the missing residue side chains by using the Replace command (Biopolymer module, Accelrys, San Diego). The obtained homology model was subjected to the same full energy minimization and structural check procedure previously described for the homology model of human 20S proteasome in the closed conformation, and, then, used for the subsequent dynamic docking studies. During the minimization, the whole disordered N-and Cterminals, the inserted loop region and the SCRs side chains were left free to move, whereas the SCRs backbone were fixed to avoid unrealistic results. Calculation of the chemical-physical properties of 2. The apparent pKa values were estimated by using the ACD/Percepta software. 7 The compound was considered in its cationic form in all calculations performed, as a consequence of the estimation of percentage of neutral/ionized forms computed at pH 7.4 (physiological value) and pH 7.2 (cytoplasmic value) using the Handerson−Hasselbalch equation. Atomic potentials were assigned using the CVFF force field, while the partial charges were assigned using the partial charges estimated by MNDO semiempirical 1 SCF calculations. 8 The conformational space of compound was sampled through 200 cycles of simulated annealing (SA; ε =80*r) followed by molecular mechanics (MM) energy minimization. During the SA procedure, the temperature is altered in time increments from an initial temperature to a final temperature by adjusting the kinetic energy of the structure (by rescaling the velocities of the atoms). The following protocol was applied: the system was heated to 1000 K over 2000 fs (time step of 1.0 fs); a temperature of 1000 K was applied to the system for 2000 fs (time step of 1.0 fs) to surmount torsional barriers; successively, temperature was linearly reduced to 300 K in 1000 fs with a decrement of 0.5 K/fs (time step of 1.0 fs). Resulting conformations were then subjected to MM energy minimization within Insight 2005 Discover 3 module (CVFF force field; ε = 80*r) until the maximum rms was less than 0.001 kcal/Å, using conjugate gradient 9 as the minimization algorithm. The resulting MM conformers were subsequently ranked by: i) conformational energy (ΔE from the global energy minimum < 5 kcal/mol), ii) interatomic distances between the charged nitrogen atoms, and iii) conformation of porphyrin ring. In order to properly analyze the electronic properties, the conformers, obtained from molecular dynamics and mechanics calculations, were subjected to a full geometry optimization through semiempirical calculations, using the quantum mechanical method PM7 10 in the Mopac2012 package 11 . The EF (Eigenvector Following routine) algorithm of geometry optimization was used, 12 with a GNORM value set to 0.01. To reach a full geometry optimization, the criterion for terminating all optimizations was increased by a factor of 100, using the keyword PRECISE. The resulting PM7 conformers were subsequently ranked as reported above for MM conformers.

S9
Docking studies on human 20S proteasome in complex with 2. According to the bioinformatic analysis performed on 20S proteasome and the conformational analysis of 2, we selected the α4-α5 groove as starting point for the docking studies of 2. The putative starting complex was subjected to dynamic docking studies (Affinity, SA_Docking; Insight2005, Accelrys, San Diego). In particular, a docking methodology, which considers all the systems flexible (i.e., ligand and protein), was used. Flexible docking was achieved using the Affinity module in the Insight 2005 suite, setting the SA_Docking procedure 13 and using the Cell Multipole method for nonbond interactions. 14 The binding domain area was defined as a subset including all residues of human 20S proteasome. All atoms included in the binding domain area were left free to move during the entire course of docking calculations, whereas, in order to avoid unrealistic results, a tethering restraint was applied on the SCRs of protein.

Site
Distances ( Table S13. Kinetic parameters for the reaction of 2 and 1 with the human 20S proteasome.