Fig 1.
Overview of the PSA structure.
The backbone of PSA shown in a ribbons representation with domains in different colors: N-terminal domain I (red), catalytic domain II (gold), linker domain III (green), C-terminal domain IV (cyan). Secondary structure elements are labeled, and the active site zinc ion is shown as a pink sphere.
Fig 2.
PSA sequence and secondary structure.
Sequence and corresponding secondary structure of human PSA are illustrated. Vertical red lines indicate domain boundaries.
Table 1.
Crystallographic statistics.
Fig 3.
PSA domain structure and domain interactions.
Overview of the individual domain structures and interfaces for the (A) N-terminal (Domain I), (B) catalytic (Domain II), (C) linker (Domain III), and (D) C-terminal (Domain IV) domains. Secondary structural elements are labeled, as are residues mentioned in the text.
Fig 4.
The active site of PSA is shown in a stick representation with the catalytic zinc ion and water molecule indicated by the pink and red spheres, respectively. Electron density (2Fo-Fc, 1.2 σ contour) is shown in blue. Zinc coordinating residues and the catalytic glutamate are labeled. Coordination interactions with the zinc ion are shown as dashed lines. In the crystal structure, the coordination distances to the zinc ion are: E375, 1.96 Å; H352, 2.12 Å; H356, 2.11 Å; and the coordinating solvent, 2.03 Å.
Fig 5.
Microtubule associated protein sequences in PSA.
The location of the microtubule associated protein sequences are shown in red on a ribbons trace of the PSA structure. One sequence is in the catalytic domain (domain II) and the other in the C-terminal domain (domain IV). The two views are related by a 90° rotation.
Fig 6.
Open and closed conformations.
A ribbons view of PSA (cyan) is shown superimposed on ERAP2 [49] (gray; PDB ID: 3SE6). The more open conformation of PSA, with domains II and IV separated by a wide channel, results in a larger and more accessible active site region.
Fig 7.
(A) Active site region of PSA showing side chains of zinc coordinating residues and other residues that may participate in catalysis as discussed in the text. The active site zinc ion is shown as a pink sphere. (B) View of the active site with elements that close off one end to promote aminopeptidase activity. Side chains for the conserved Glu-Ala-Met-Glu-Asn-Trp (GAMENW) sequence are shown in a stick representation. The Glu180 side chain from the N-terminal domain (red) is also shown. Glutamates 180 and 319 likely interact with the N terminus of bound peptide substrates. (C) Molecular surface (semi-transparent) and ribbons view of the active site region. Glu319 of the GAMENW sequence is indicate, as is nearby Glu180. A likely path for bound peptide substrate is indicated by the curved cylinder. The active site zinc ion is in pink.
Fig 8.
Extended peptide-binding surface in PSA.
Side chains of hydrophobic and aromatic residues that form parts of the molecular surface surrounding the active site of PSA are shown in a gold stick representation. The active site zinc ion (pink sphere) and side chains of active site residues are also shown.
Fig 9.
Polyglutamine peptide binding by PSA.
(A) Weighted difference electron density (green mesh, 1.5 σ cutoff) in the PSA active site region was calculated with 3.6 Å resolution data obtained from a crystal soaked in a solution containing 200 μM polyglutamine peptide (Lys2Gln15Lys2). A polyalanine peptide (cyan carbons) is shown modeled into the density to illustrate the path of the bound peptide. Nearby elements of the catalytic domain are shown in gold with the side chains of active site and potentially interacting residues shown in stick representation (gold carbons). (B) Refined peptide complex with the bound peptide built as polyglutamine. The peptide is shown with cyan carbons and the protein in gold with all side chains in stick representation. 2Fo-Fc electron density for the peptide is in light blue mesh (0.26 sigma cutoff). The active site zinc ion is shown as a pink sphere in both panels.
Fig 10.
Interaction with polyglutamine peptide (PQ) and dynorphin A(1–17).
Binding of the two peptides was followed by their ability to inhibit hydrolysis of a standard fluorogenic peptide substrate, alanine 4-methoxy-β-naphthylamide. Double reciprocal plots for inhibition by (A) dyn A(1–17) and (B) PQ indicate the peptides act as competitive inhibitors of the fluorogenic substrate. Single reciprocal plots (Dixon plots) for dyn A(1–17) interacting with (C) PSAwt and (D) PSAF433A are shown along with PQ interacting with (E) PSAwt and (F) PSAF433A. In these plots, the abscissa intercept is–Ki(1+[S]/Km), where [S] is the concentration and Km the Michaelis constant of the fluorogenic substrate. Error bars represent standard error of the mean for triplicate measurements.
Table 2.
Apparent Ki values for polyglutamine and dynorphin A(1–17) with PSAwt and PSAF433A.
Fig 11.
(A) Positions of substrates and substrate analogs over the binding surfaces of M1 family peptidases. Selected peptides or peptide analogs from M1 aminopeptidase structures superimposed on PSA are shown in stick format with different carbon colors. The active site zinc ion is shown as a pink sphere and elements of the catalytic domain are in gold. The polyalanine model built into the PSA-polyglutamine structure is shown with carbons colored in cyan. Other structures shown are APA with bestatin [74] (PDB ID 4KXB, green carbons), APA with amastatin [74] (4KX8, magenta), APN with bestatin [75] (4FYR, yellow), APN with amastatin [75] (4FYT, salmon), APN with angiotensin IV [75] (4FYS, gray), AnAPN1 with 5-mer peptide [78] (4WZ9, slate), PfA-M1 with bestatin [45] (3EBH, orange), PfA-M1 with phosphinic dipeptide analog [45] (3EBI, lime), LTA4H with bestatin [46] (1HS6, dark teal), LTA4H with a 3-mer peptide [109] (3B7S, hot pink), ePepN with actinonin [110] (4Q4E, marine), ePepN with amastatin [110] (4Q4I, olive), porcine APN with substance P [76] (4HOM, split pea), ERAP1 with a 10-mer phosphinic peptide [59] (6RQX, teal), and M1dr with a 3-mer peptide [84] (6IFG, dark violet). (B) Hinge motion at the interface between PSA domains 1 and 2. Cα traces of domain 1 (red) and domain 2 (gold) from the crystal structure are shown superimposed on the trace of a structure from normal mode analysis (gray) using the NOMAD-Ref server [111]. Movement of domain 1 in the normal mode analysis relative to its position in the crystal structure can be seen as a shift of the gray trace toward the top of the figure.
Fig 12.
Modeling puromycin interaction with a closed form of PSA.
The closed form model for PSA was constructed by superimposing the open PSA crystal structure reported here on the closed form AlphaFold model domain by domain. (A) Docking of puromycin with the closed form PSA using ROSIE Ligand_docking [113–115]. Puromycin shown with yellow carbon atoms is from superposition of an inactive ePepN-puromycin complex reported by Addlagatta and colleagues [112] on the closed PSA model. That puromycin pose was used as the starting point for docking with the closed PSA model. The three lowest energy complexes are show with green, purple, orange carbons, corresponding protein side chains, and the lowest energy model backbone in green. The zinc ion is show as a pink sphere. (B) The lowest energy docked PSA-puromycin model. Side chains positioned to possibly interact with the ligand are shown. (C) Hydrolysis fragments of puromycin superimposed on the closed PSA model. The fragments from the active ePepN-puromycin crystal structure [112] are shown with magenta carbons based on the superposition of the ePepN complex on the closed PSA model. Side chains from the PSA model are shown.