Single fluorescence probes along the reactive center loop reveal site-specific changes during the latency transition of PAI-1

Abstract

The serine protease inhibitor (serpin), plasminogen activator inhibitor-1 (PAI-1), is an important biomarker for cardiovascular disease and many cancers. It is therefore a desirable target for pharmaceutical intervention. However, to date, no PAI-1 inhibitor has successfully reached clinical trial, indicating the necessity to learn more about the mechanics of the serpin. Although its kinetics of inhibition have been extensively studied, less is known about the latency transition of PAI-1, in which the solvent-exposed reactive center loop (RCL) inserts into its central β-sheet, rendering the inhibitor inactive. This spontaneous transition is concomitant with a large translocation of the RCL, but no change in covalent structure. Here, we conjugated the fluorescent probe, NBD, to single positions along the RCL (P13-P5') to detect changes in solvent exposure that occur during the latency transition. The results support a mousetrap-like RCL-insertion that occurs with a half-life of 1-2 h in accordance with previous reports. Importantly, this study exposes unique transitions during latency that occur with a half-life of ∼5 and 25 min at the P5' and P8 RCL positions, respectively. We hypothesize that the process detected at P5' represents s1C detachment, while that at P8 results from a steric barrier to RCL insertion. Together, these findings provide new insights by characterizing multiple steps in the latency transition.

Keywords: PAI-1; RCL; latency transition; serpin.

Figure 1

The active and latent structures of PAI‐1. The RCL (red) in the active structure of PAI‐1 [PDB: 3Q02 (W175F)54] is mostly unresolved (dotted line), while it is inserted as the fourth strand in the central β‐sheet A (s4A) in the latent conformation [PDB: 1C5G30]. Important functional regions in self‐regulation by latency transition, including the gate (green), shutter (s5A/s3A, cyan), helix F (and the loop connecting it to s3A) of the flexible joint regions (orange), and helix A (magenta) are also indicated. Residues of the RCL are assigned by their P designation according to its distance from the P1–P1′ scissile bond, with C‐terminal residues also indicated by a prime (′). The side chains of certain RCL residues are circled, and the molecular surface of the latent structure included is colored according to the legend (bottom right).

Figure 2

Half‐life of inactivation via protease inhibition. Unlabeled (white bars) and NBD‐labeled (black bars) PAI‐1 proteins were incubated at 37°C in MOPS buffer (0.05M, 0.1M ammonium sulfate, 1 mM EDTA, pH 7.4) and their activity assayed by the addition of tPA and excess Spectrozyme tPA (a chromogenic substrate for tPA) at various time points. Initial rates were plotted over time, and the inhibitory half‐life determined according to a one‐phase decay model. Wild‐type (wt) and single‐cysteine PAI‐1 constructs according to their P designation are indicated. Error bars are plotted as standard deviations of triplicate experiments.

Figure 3

Active and latent fluorescence spectra for NBD along the RCL. NBD‐PAI‐1 (0.5 μM) was added to phosphate buffer (0.05M NaH₂PO₄, 0.3M NaCl, 1 mM EDTA, 0.1% PEG8000, pH 7.4) at 37°C, excited at 480 nm, and the emission spectra from 500 to 600 nm collected. The initial active (solid line) and final latent (dashed line) emission spectra are shown for several representative variants of PAI‐1 with NBD labels at different positions in the RCL.

Figure 4

Loss of PAI‐1 activity and changes in NBD fluorescence. The residual activity of NBD‐PAI‐1 via tPA inhibition (left axis), and the change in fluorescence emission at 530 nm over time after excitation at 480 nm (right axis) were measured at 37°C in buffer with a pH of 7.4. The activity (solid circles) and fluorescence change (open squares) were measured in MOPS buffer (0.05M, 0.1M ammonium sulfate, 1 mM EDTA) and phosphate buffer (0.05M NaH₂PO₄, 0.3M NaCl, 1 mM EDTA, 0.1% PEG 8000), respectively, for PAI‐1 labeled with NBD at the (A) P11, (B) P1′, and (C) P4′ positions. Data were normalized, averaged, and then fit to a one‐phase decay model. Error bars are plotted as standard deviations.

Figure 5

Stabilization of PAI‐1. Buffer effects on NBD‐PAI‐1 activity were measured by incubation at 37°C in either MOPS (0.05M, 0.1M ammonium sulfate, 1 mM EDTA, pH 7.4) or phosphate buffer (0.05M NaH₂PO₄, 0.3M NaCl, 1 mM EDTA, 0.1% PEG8000, pH 7.4). Samples were excited at 480 nm, the fluorescence emission at 530 nm collected over time, and fit to a one‐phase decay model from which the half‐life was determined. Representative data are for NBD P9 (S339C) PAI‐1. Error bars are plotted as standard deviations of duplicate experiments.

Figure 6

Half‐life of RCL insertion via fluorescence. A. NBD‐labeled PAI‐1 was added at 0.3–0.5 µM in phosphate buffer (0.05M NaH₂PO₄, 0.3M NaCl, 1 mM EDTA, 0.1% PEG 8000, pH 7.4) at 37°C. Samples were excited at 480 nm and fluorescence emission at 530 nm was plotted against time. Except for P8 and P5′, data were fit to a one‐phase decay model from which the half‐life of the latency transition was obtained. For P8 and P5′, the half‐life of the slow process according to a two‐phase decay model is plotted. B. The overall change in fluorescence [(F _final − F _initial)/F _initial] upon completion of the latency transition is plotted for NBD at all RCL positions. Values are normalized to the initial fluorescence at 530 nm. Experiments were performed in at least duplicate and plotted with standard deviation error bars.

Figure 7

Changes in NBD fluorescence at the P5′, P8 and P13 RCL positions. PAI‐1 (0.5 μM) with NBD conjugated to the (A) P5′, (B) P8, or (C) P13 residues was incubated in phosphate buffer (0.05M NaH₂PO₄, 0.3M NaCl, 1 mM EDTA, 0.1% PEG 8000, pH 7.4) at 37°C, excited at 480 nm, and emission at 530 nm plotted over time. Experiments were performed in at least triplicate, but only a representative curve for each is shown. Data were fit to a one‐phase (top) and two‐phase (bottom) decay model. Residual plots for each are included below their curves.

Figure 8

Temporal model of PAI‐1 latency transition. The RCL is shown in red, with s1C depicted as a red arrow, the gate loops in green, the shutter β‐strands as blue arrows, and hF as an orange cylinder. Additional strands of the central β‐sheet A are shown as grey arrows. Starting from the native, metastable state (A), the RCL is extended, but close to the protein core. In the steps toward the latent conformation, the shutter partially opens (B). This allows the RCL to partially insert and s1C to partially detach, reversibly, in the prelatent conformation (C). For further RCL insertion, the gate must widen (D) and hF must be displaced (E). The C‐terminal RCL may then pass through the gate and consecutive N‐terminal RCL residues insert as s4A (F). When the N‐terminal RCL is fully inserted as s4A and C‐terminal RCL extended along the surface of the serpin (G), hF returns to its position over the shutter and the latency process is completed (H).