Unveiling the Catalytic Mechanism of a Processive Metalloaminopeptidase

Intracellular leucine aminopeptidases (PepA) are metalloproteases from the family M17. These enzymes catalyze peptide bond cleavage, removing N-terminal residues from peptide and protein substrates, with consequences for protein homeostasis and quality control. While general mechanistic studies using model substrates have been conducted on PepA enzymes from various organisms, specific information about their substrate preferences and promiscuity, choice of metal, activation mechanisms, and the steps that limit steady-state turnover remain unexplored. Here, we dissected the catalytic and chemical mechanisms of PaPepA: a leucine aminopeptidase from Pseudomonas aeruginosa. Cleavage assays using peptides and small-molecule substrate mimics allowed us to propose a mechanism for catalysis. Steady-state and pre-steady-state kinetics, pH rate profiles, solvent kinetic isotope effects, and biophysical techniques were used to evaluate metal binding and activation. This revealed that metal binding to a tight affinity site is insufficient for enzyme activity; binding to a weaker affinity site is essential for catalysis. Progress curves for peptide hydrolysis and crystal structures of free and inhibitor-bound PaPepA revealed that PaPepA cleaves peptide substrates in a processive manner. We propose three distinct modes for activity regulation: tight packing of PaPepA in a hexameric assembly controls substrate length and reaction processivity; the product leucine acts as an inhibitor, and the high concentration of metal ions required for activation limits catalytic turnover. Our work uncovers catalysis by a metalloaminopeptidase, revealing the intricacies of metal activation and substrate selection. This will pave the way for a deeper understanding of metalloenzymes and processive peptidases/proteases.


Supporting information note 2: Simulations using Kintek Global Explorer for Mn x Mg utilization
The following model was used, where E is free enzyme, no metals bound, EMn1 or EMg1 are enzyme with one metal ion bound, EMn2 or EMg2 are enzyme with two metal ions bound, LeuMn is the product formed when Mn was the activating metal, LeuMg is the product formed when Mg was the activating metal.
For this simulation, the following conditions were used: [Mn] = 0.32 mM [Mg] = 150 mM [Leu-pNA] = 100mM [PaPepA] = 0.01mM Constraints were used as indicated in the image below, based on values for KD and kcat/KACT (used to set a lower limit to association rate constants to metal ions).The product endpoints in the end of the simulation were 98.87% LeuMg, 1.12% LeuMn.
Supporting information note 3: Derivation of kinetic constants kcat, kcat/KM-Leu-pNA and kcat/KACT.We used the net rate constants method 1 to derive equations defining kcat, kcat/KM-Leu-pNA and kcat/KACT.This assumes that under steady state conditions chemistry and product release are irreversible.We also assumed PepA exists in a "1 bound metal" (EMn) during steady-state turnover due to the tight binding affinity to the first metal binding site.Binding to the weak site, however, needs to take place again between catalytic cycles, giving rise to the high value for KACT.In summary, the net rate constants for each enzyme species are: For comparison, the affinity to the first metal site is defined by For kcat/KM with saturating Mn: Supporting information note 4: Kintek Global Explorer fitting of stopped flow data Both MTO and STO pre-steady-state data were fit according to the mode described in Figure 4A with scaling factors for each fluorescent species (EP and P) to account for differences in AMC fluorescence whilst in complex with PaPepA versus free in solution.Scaling factors were different across STO and MTO experiments as both were carried out using different detector voltages.
The model that best fitted acquired data was as below, where F is metal free enzyme, E1 is PaPepA with one Mn 2+ ion, E is PepA with two Mn 2+ ions, S is Leu-AMC and P is Leucine: For fitting, k1 and k-1 ratios were constrained (linked) based KD for Mn 2+ binding as measured by ITC (12nM).Initially, k2 and k-2 were allowed to vary without constraints, and values best fitted as well as Fitspace lower and upper boundaries are reported in Table 2.The value for k1 was best fitted as diffusion limited, and therefore was fixed in further fitting cycles.The best fitted value for k-2 was high and unconstrained, and therefore also fixed at 1000 s -1 for further fitting cycles.k3, k-3, k4, and k5 were not constrained.Multiple iterations of fitting were performed, and best fit values recorded.k1, k2, k3, k-3, k4, and k5 were evaluated using FitSpace Explorer to generate confidence intervals for each of these rate constants.Figure 4B shows the 3-dimensional surfaces generated with Fitspace data analysis, displaying the extent to which any two parameters can co-vary and produce a fit that is well constrained by data.Fitting data to this 5-step mechanism and evaluating the fit showed a very low rate constant for k3.
Following metal binding, PaPepA must bind its substrate-Leu AMC or Leu-pNAwhich occurs quickly but with a fast k-3, likely causing the lack of a plateau in observed rate constants for the STO experiment.The lack of a visible burst in the MTO experiments suggests that product release is not rate limiting, which is supported by the calculated values for k4 and k5 (Figure 4B).Prior to fitting, substrate peaks were integrated and slope of calibration curve used as a scaling factor for fitting.For data fitting using Kintek Global Explorer, the following model and assumptions were made:      S3: ADH standard peptides following digestion with Trypsin.Singly, doubly, triply and quadruply protonated charge states were searched for.In bold = peptides which would result from full trypsin cleavage of ADH.However, up to three missed cleavages were allowed in our search parameters.Peptides in bold contain 0 missed cleavages following trypsin digestion.

Molecule
, it should be mentioned that this calculated SKIE is an intrinsic solvent kinetic isotope effect, calculated on Supporting note 4. Additionally, models accounting for reactant state protons did not converge to fitted values.Proton inventories on D2O kcat/KM-LeupNA-Mn and D2O kcat/KACT Had errors that exceeded 20% and therefore were not interpreted mechanistically.

Calculation of intrinsic Solvent kinetic isotope effects
With the scheme above and the fitted rate constants according to the table below: value k1 (M -1 s -1 ) 86.5 ± 0.1 k-1 (s -1 ) 1.18* k2 (M -1 s -1 ) 18.9 ± 0.1 k-2 (s -1 ) 1.89* k3 (M -1 s - is used to simplify nomenclature, but the rate constant k4 is a macroscopic rate constant , likely including several microscopic steps with some contributing to kcat/KM-LeupNA-Mn and others to kcat, hence different values for calculated intrinsic solvent kinetic isotope effects.

Figure S1 :
Figure S1: PaPepA mediated cleavage of Leu-pNA.Representative raw data for an aa-pNA substrate.Left: Raw data for cleavage of Leu-pNA when [Mn 2+ ] is varied and Leu-pNA is present in excess.Right: Raw data when [Leu-pNA] is varied and MnCl2 is present in excess.

Figure S3 :
Figure S3: Sequence logos for peptide sequences.Top: Logo resulting from analysis of sequences from the P. aeruginosa PA14 proteome containing N-terminal methionine and leucine residues.Bottom: Logo from ADH peptide cleavage showing the first six residues of (left) PaPepA-cleaved peptides, (right) uncleaved peptides.

Figure S4 :Figure S5 :
Figure S4: Degradation of AVLQSGFRKK-NH2 to a tripeptide following incubation with 1 µM PaPepA.(A) and (B) -SIRs monitoring AVLQSGFRKK-NH2 and its degradation products when incubated in the absence (A) and presence (B) of PaPepA for 6 hours, respectively.(C) Extracted ion chromatogram of RKK-NH2 degradation product (+H + and +Na + ) showing increased abundance in the sample incubated with PaPepA.

Figure
Figure S12: A) Multiple turnover stopped flow experiment with excess Leu-pNA.(Leu-pNA present at 250 M).B) Calibration curve of fluorescence readout as a function of AMC concentration for plate reader-based assays evaluating Leu-AMC as a PaPepA substrate.C) Initial velocity plotted as a function of substrate concentration fit to a Michaelis-Menten curve.(left) MnCl2 added in excess, (right) Leu-AMC added in excess.

Figure S15 :
Figure S15: Mn2+ binding site.2Fo-Fc map (blue) and Fo-Fc map (orange) depicting two Mn binding sites and the bridging water molecule.Mn ions are coloured in grey, water molecule is red.

Table S6 :
Exponential fitted data for time courses with AVLQSGFRKK-NH2

Table S7 :
Average binding constants, concentrations, and stoichiometries from ITC Data for Mn 2+ and Mg 2+ Binding to PaPepA.

Table S9 :
SKIEs calculated by fitting proton inventory data (Figure5C) to distinct models.
SKIE calculated based on the equation below, where φ  stands for reactant state fractionation factor, φ  stands for transition state fractionation factor: