Probing neuroserpin polymerization and interaction with amyloid-beta peptides using single molecule fluorescence

Abstract

Neuroserpin is a member of the serine proteinase inhibitor superfamily. It can undergo a conformational transition to form polymers that are associated with the dementia familial encephalopathy with neuroserpin inclusion bodies and the wild-type protein can inhibit the toxicity of amyloid-beta peptides in Alzheimer's disease. We have used a single molecule fluorescence method, two color coincidence detection, to determine the rate-limiting steps of the early stages of the polymerization of fluorophore-labeled neuroserpin and have assessed how this process is altered in the presence of A beta(1-40.) Our data show that neuroserpin polymerization proceeds first by the unimolecular formation of an active monomer, followed by competing processes of both polymerization and formation of a latent monomer from the activated species. These data are not in keeping with the recently proposed domain swap model of polymer formation in which the latent species and activated monomer are likely to be formed by competing pathways directly from the unactivated monomeric serpin. Moreover, the A beta(1-40) peptide forms a weak complex with neuroserpin (dissociation constant of 10 +/- 5 nM) that increases the amount of active monomer thereby increasing the rate of polymerization. The A beta(1-40) is displaced from the complex so that it acts as a catalyst and is not incorporated into neuroserpin polymers.

Figure 1

Characterization of Ser²²⁹Cys neuroserpin. (A) Purified Ser²²⁹Cys neuroserpin was labeled with Alexa Fluor 647 and assessed by 7.5–15% w/v nondenaturing PAGE (upper panel); 229 represents labeled Ser²²⁹Cys neuroserpin and wt is the wild-type protein. (Lower panel) Fluorophore-labeled Ser²²⁹Cys formed a complex with tPA on 7.5–15% w/v/ SDS-PAGE. The protein was visualized by either Coomassie staining or by fluorescence. (Left to right): Ser²²⁹Cys labeled with Alexa 647, Ser²²⁹Cys labeled with Alexa 647 and addition of tPA, Ser²²⁹Cys labeled with Alexa 647, Ser²²⁹Cys labeled with Alexa 647 and addition of tPA, wt neuroserpin, wt neuroserpin and addition of tPA, tPA, molecular mass standard. (B) A 7.5–15% w/v nondenaturing PAGE to assess the polymerization of Alexa 488- and 647-labeled neuroserpin incubated at 0.1 mg/mL and 45°C in PBS buffer pH 7.4. Each lane represents a different time of incubation (min). Two micrograms of protein were loaded in each lane. (C) Polymerization of Ser²²⁹Cys neuroserpin labeled with Alexa 488 and 647 in the presence of different ratios of Aβ_1–40 to neuroserpin. Incubation was at 45°C in PBS at pH 7.4, Aβ_1–40 was added after 30 min of incubation. The samples were visualized on 7.5–15% w/v nondenaturing PAGE.

Figure 2

(A) Association quotient time traces for labeled Nsp-Alexa 488 and Nsp-Alexa 647 incubation at a total concentration of 0.1 mg/mL. For incubations 1 and 2 (circles and squares), each data point represents the average value of three repetitions at the same timepoint. The last data point for incubation 2 is 240 min. (B) Average brightness of the coincident events in the blue (squares) and red (circles) channels from an incubation of labeled Nsp-Alexa 488 and Nsp-Alexa 647 at a total concentration of 0.1 mg/mL. The average standard error in the measurements is ∼4%. (C) Association quotient time traces for labeled Nsp-Alexa 488 and Nsp-Alexa 647 incubation at a total concentration of 0.1 (circles), 0.2 (squares), and 0.3 mg/ml (triangles) using the same stock of labeled neuroserpin. The fit to exponential decay functions with τ = 44, 171, and 486 min, respectively are also shown (lines). (D) Zoom of the early part of C showing the linear increase in association quotient with time.

Figure 3

Neuroserpin polymerization kinetic schemes. Neuroserpin monomer M is activated to M^∗ with an activation rate k_A; depending on the model, either a normal (M) or an active monomer (M^∗) can undergo a transformation into the latent form (M_L) or form a dimer M₂ by adding either M or M^∗ that we assume to occur at the same rate k₁; the polymers can grow by addition of a monomer (M or M^∗) with a rate k₂, or by addition of an oligomer of size i (M_i) with any other oligomer of size j to form an oligomer M_i+j at a rate k₃ that we assume to be the same for whatever value of i and j. In A the latent species is formed from the activated monomer, M^∗. In B the dimer formation and monomer addition only occur through activated monomers M^∗; whereas in C the transformation to nonpolymerizing latent form is a parallel pathway of the monomer M that may transform either to the active M^∗ or to the latent monomer (M_L).

Figure 4

Global fit of the association quotient time trace and the coincidence histograms to the kinetic model for neuroserpin polymerization. (A) Experimental association quotient (symbols) and the fit (line) for a 0.2 mg/mL neuroserpin incubation. (B) Experimental TCCD histogram after 240 min of incubation and a simulated histogram based on the global fit (red line). (C) Calculated distribution of oligomers based on the kinetic model at the three aliquot times analyzed: 30 (first bar), 90 (middle bar), and 240 min (last bar). (D) Experimental association quotient (symbols) and simulated (lines) using the rate constants in Table 1 for a 0.1 (circles) and 0.3 mg/mL (triangles) neuroserpin incubation. (E) Predicted association quotient time traces by kinetic model in Fig. 3C for neuroserpin polymerization at a total concentration of 0.1 (dashed), 0.2 (solid), and 0.3 (dotted) mg/mL.

Figure 5

(A) Effect on 0.1 mg/mL neuroserpin polymerization of the addition of Aβ_1–40 spiked in after 30 min of incubation (black dotted line). [Aβ_1–40]/[Nsp] = 0 (squares), 0.1 (circles), 1 (triangles), 10 (diamonds), and 100 (inverted triangles). (B) Effect on 0.1 mg/mL neuroserpin polymerization of the addition of equimolar concentration of Aβ_1–40 at the beginning of the incubation (open triangles), and preincubation with Aβ_1–40 for 24 h at 37°C (solid circles). Both curves are compared to that obtained in the absence of Aβ_1–40 (open squares).

Figure 6

(A) Association quotient for the complex Nsp/Aβ_1–40 under preincubation conditions. The solid line indicates the limit of detection of the TCCD technique. (B) Fraction of Alexa647-Nsp oligomeric events from Nsp incubations in the presence of HL488-Aβ_1–40 with (open symbols, dashed line) and without (solid symbols, solid lines) 24 h of preincubation.

Scheme 1

Scheme for the interaction of Amyloid-β_1–40 with neuroserpin. A fast transient interaction creates a complex where the neuroserpin is activated and able to polymerize. If the neuroserpin and Aβ_1–40 are incubated long enough, they form a stable inactive complex.