Role of conformational change in the C-terminus of β 2 -microglobulin in dialysis-related amyloidosis

Abstract

Background

β₂-Microglobulin (β₂m) has been identified as the precursor protein of dialysis-related amyloidosis (DRA), which is a serious complication for haemodialysis (HD) patients. However, mechanisms underlying β₂m amyloid fibril formation remains to be elucidated. We previously demonstrated, in amyloid deposits from HD patients, a conformational isoform of β₂m with an unfolded C-terminus. However, no direct experiments have previously been performed to address whether unfolded β₂m in the C-terminus may be prone to form amyloid fibrils.

Methods

To evaluate roles of C-terminal amino acids in β₂m-induced amyloid formation, we generated six types of recombinant β₂m with amino acid substitutions in the C-terminal region. To investigate their conformational change and amyloidogenicity, we measured circular dichroism spectra, the fluorescence intensity of tryptophan and thioflavin-T (ThT) of the recombinant β₂m. To analyse morphological change of β₂m, we performed electron microscopy (EM) on the samples with elevated ThT fluorescence intensity. We used ultrasonication to enhance β₂m destabilization of the protein.

Results

β₂M Trp95Leu and Arg97Ala showed conformational changes and increased their amyloidgenicity compared with β₂m wild-type (WT). With ultrasonication, β₂m Trp95Leu and Arg97Ala generated more amyloid fibrils than did β₂m WT even in physiological solution. EM showed that β₂m formed amorphous debris containing typical amyloid fibrils at 24 hours, when ThT fluorescence intensity was three-fold lower than that at six hours.

Conclusions

Conformational changes in the C-terminus of β₂m may play an important role in DRA and that ultrasonication is useful for analysis of β₂m amyloidogenesis.

Introduction

In 1985, Gejyo et al. ¹ identified β₂-microglobulin (β₂m) as the precursor protein of dialysis-related amyloidosis (DRA). DRA, a serious complication occurring in patients receiving long-term haemodialysis (HD), causes various clinical manifestations, such as carpal tunnel syndrome, polyarthralgia, destructive spondyloarthropathy and bone cysts. More than 2,40,000 Japanese patients are treated with chronic HD, and of those, 6.4% have a history of HD for more than 20 years. Up to 50% of patients had developed DRA after 20 years and the percentage was even higher after 25 years. ¹ However, to date, no established treatments of DRA have been developed, and the mechanism of β₂m amyloid fibril formation remains to be elucidated.

β₂M consists of 99 amino acids and appears to be simple in the plane dimension, but in 1985, ² Becker and Reeke identified a three-dimensional configuration with a complex conformation involving seven β strands. Several variants of β₂m have been found in clinical specimens, including amyloid-containing tissues from HD patients. ^3–6 However, no reports of unfolded variants in a clinical setting, except for an abstract by Rampino et al., ⁷ have been published. Stoppini et al. ⁸ reported that carpal amyloid-containing tissues from HD patients showed a positive reaction for a monoclonal antibody specific to the C-terminal octapeptide (i.e. residues 92–99 of β₂m) as well as to a commercial polyclonal anti-β₂m serum under physiological conditions. This C-terminal portion is believed to be normally hidden in the core of the quaternary structure of the light chain of class I major histocompatibility complex ⁹ and therefore is regarded as being non-reactive with commercial polyclonal anti-β₂m serum.

That misfolding of the amyloid precursor protein initiates amyloid formation in several types of amyloidosis, such as transthyretin amyloidosis, has been well studied, and therapeutic approaches have been proposed. ^10–12 In 1998, Bellotti et al. ¹³ first demonstrated, in amyloid deposits obtained from HD patients, the presence of a partially unfolded β₂m that could refold reversibly into native β₂m.

We previously reported preparation of a monoclonal antibody against the C-terminal octapeptide of β₂m and identified, in amyloid deposits obtained from DRA patients, a conformational isoform of β₂m with an unfolded C-terminus. ¹⁴ However, no experiments have been performed to specifically show that β₂m that has a misfolded C-terminus may be more likely to produce amyloid fibrils than native β₂m. Trp95 and Arg97 are believed to be important residues for stability of β₂m (Figure 1). In this study, to evaluate the role of C-terminal amino acids in β₂m amyloid formation, we generated six types of recombinant β₂m with amino acid substitutions in the C-terminal region (Trp95Leu, Trp95Glu, Trp95Ala, Trp95Val, Trp95Ser and Arg97Ala).

Figure 1

Structure of the C-terminus of human β₂m drawn using the programs BobScript ²¹ and Raster3D. ²² Trp95 is hydrophobic; Arg97 can form a hydrogen bond with Asn17 and Thr73. Arrows indicate the mutated positions of Trp95 and Arg97 of the C-terminus of β₂m in our experiment

Amyloid formation can also be enhanced by destabilizing or unfolding proteins by increased temperature, high pressure and chemical agents in vitro. ^15–17 Stathopulos et al. ¹⁸ reported that ultrasonication of various proteins induced amyloid formation. Ultrasonication is also frequently used to break aggregates into smaller pieces for seeding new aggregate growth in laboratory studies of proteins associated with various conformational disorders, such as prion diseases. ¹⁸ Saborio et al. ¹⁹ developed a new method using ultrasonication, termed protein misfolding cyclic amplification (PMCA), for early diagnosis of prion disorders. In PMCA, a small amount of prion aggregates can be detected with high sensitivity by amplifying the prion aggregates through cycles of ultrasonication which breaks small amount of prion aggregates into smaller pieces acting as seeds for the formation of new aggregates from soluble prion protein.

In this report, we also employed ultrasonication to enhance destabilizing of those recombinant β₂m and investigated amyloid fibril formation in those β₂m.

Methods

Protein expression and purification of recombinant β₂microglobulin

β₂M wild-type (WT) and mutated human β₂m were produced in Escherichia coli with six histidine residues and a Factor Xa cleavage site attached to the N-terminus. The pET28-hβ₂m was produced as follows. A culture of E. coli was grown in luria broth (LB) medium supplemented with 100 mg/L ampicillin, with agitation, at 37°C overnight. At a bacterial density of OD₆₀₀ 0.6–1.0, isopropyl thio-β-d-galactoside (Sigma-Aldrich Japan K.K., Tokyo, Japan) was added to achieve a final concentration of 1 mmol/L. After four hours of incubation at 22°C, cells were collected by centrifugation at 8000 g for 30 minutes at 4°C, and the cell pellet was resuspended in BugBuster plus Benzonase Nuclease (Novagen, Madison, WI, USA). To obtain human β₂m protein, cell lysate was directly injected onto the His·Bind column. For removal of the six histidine residues, fractions were digested with the Factor Xa kit (Novagen) according to the manufacturer's instructions and were injected onto the His·Bind column to remove undigested protein. Human β₂m without the six histidine residues was collected in the flow-through fraction. A 50 mL sample of the β₂m fraction was passed through a 1.0 mL Resource Q column (Amersham Pharmacia Biotech, Uppsala, Sweden), equilibrated with 20 mmol/L Tris–HCl, pH 8.0. The partially purified β₂m was concentrated to 2 mL by ultrafiltration with a Centriplus-10 concentrator (Amicon, Bedford, MA, USA). The material was applied at a flow rate of 5 mL/min to a TSK gel G3000SW column (10 mm × 60 cm) (Tosoh, Tokyo, Japan), equilibrated with 50 mmol/L sodium acetate buffer, pH 6.7, containing 0.3 mol/L NaCl. The β₂m peak was collected at the elution time of approximately 30 minutes. This sample was then concentrated to 2 mL by ultrafiltration with a Centricon-10 concentrator (Amicon), after which the concentrated sample was dialysed against 0.15 mol/L NaCl. Sample purity was evaluated by means of density analysis of silver-stained protein bands on sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE), performed as previously described by Hermansen et al. ²⁰ Matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) (Applied Biosystems, Foster City, CA, USA) confirmed the presence of β₂m.

Structural analysis

The structure of the C-terminus was drawn by using MolScript ²¹ and Raster3D. ²²

Measurements of circular dichroism spectra

Circular dichroism (CD) spectra of β₂m WT, β₂m Trp95Leu and β₂m Arg97Ala were measured at 25°C with a J-720-type spectropolarimeter (JASCO, Tokyo, Japan). The standard buffer for spectroscopic measurements was 20 mmol/L Tris–HCl (pH 7.4). For far-UV measurements (200–250 nm), the CD signal was recorded in a 1 mm path length cell. Data for far-UV CD were recorded with a 1 nm bandwidth, 1 nm resolution, scan speed of 20 nm/min and response time of two seconds. Spectra were recorded at a protein concentration of 10 μmol/L. Results were expressed as mean residue ellipticity (θ).

Tryptophan fluorescence intensity

The fluorescence intensity of trptophan (trp) was measured using a Hitachi F-2000 spectrofluorimeter (Hitachi, Tokyo, Japan) at 25°C. All assays used an excitation wavelength of 295 nm and an emission wavelength of 340 nm. Excitation and emission slits were set at 5 nm. Trp fluorescence intensity was recorded at a protein concentration of 10 μmol/L in 67 mmol/L sodium phosphate (pH 7.4).

Thioflavin-T-based fluorimetric assays

β₂M was dissolved in distilled water, syringe-filtered (0.25 μm) to remove aggregates, and diluted in Eppendorf tubes to a final concentration of 30 μmol/L in 50 mmol/L glycine–HCl (pH 2.5) containing 100 mmol/L NaCl. Samples (1.5 mg/mL) were then incubated at 37°C for 10 days. The resulting suspensions were analysed using the ThT fluorescence test. ^23,24 Fluorescence spectra were obtained by means of a Hitachi F-2000 spectrofluorimeter (Hitachi) with an assay volume of 1 mL. All assays used an excitation wavelength of 448 nm and an emission wavelength of 489 nm. Excitation and emission slits were set at 5 nm. The reaction mixture contained 5 μmol/L ThT and 50 mmol/L Gly–NaOH buffer, pH 10.0. ^23,24 Spectra were obtained at 25°C within minutes after addition of the sample to the reaction mixture.

Ultrasonication

Directional ultrasonic treatment was performed with a water bath-type ultrasonic transmitter (Model SP070-PG-M035SP, Elekon, Tokyo, Japan), which focused 20 kHz ultrasonic waves from two sides and the bottom on the central part of the water bath where the reaction samples were placed. The temperature of the water bath was maintained at 37°C using a thermostatic water bath (Model NP035-08, Elekon). This ultrasonication proceeded for one minute, followed by incubation for nine minutes without ultrasonication. Aliquots were collected to measure the degree of ThT fluorescence.

Ultrasonication-induced fibril formation

For seeded growth in an acidic extension reaction, fibrils of β₂m were first induced to form by incubating 2 mg/mL β₂m at 37°C with ultrasonic treatment in 50 mmol/l glycine–HCl (pH 2.5) containing 100 mmol/L NaCl. Fibrils used as seeds were collected by centrifugation at 14,000 g for 10 minutes, and supernatant was removed. Then, 30 μmol/L monomeric β₂m was mixed with 1 μmol/L seed fibrils with or without ultrasonication in 50 mmol/L glycine–HCl (pH 2.5) containing 100 mmol/L NaCl. For seeded growth at neutral pH, seed fibrils were formed by incubating 2 mg/mL β₂m at 37°C, with ultrasonication using 1 mmol/L SDS in 50 mmol/L sodium phosphate (pH 7.0) containing 100 mmol/L NaCl. Then, 30 μmol/L monomeric β₂m was incubated with 5 μmol/L of seeds with or without ultrasonic treatment in 50 mmol/L sodium phosphate (pH 7.0) containing 100 mmol/L NaCl. Measurements were performed in triplicates, and values are expressed as mean ± standard deviation (SD) (plotted as error bars).

Electron microscopy

Pellets (5 μL) from each extension reaction were diluted with 95 μL of distilled water. Samples were placed on carbon/collodion-coated grids and were allowed to adsorb for one minute. Samples were then stained with a drop of 0.2% uranyl acetate for 30 seconds. Excess stain was drained off, after which grids were air-dried. All images were taken with a Hitachi H-7500 electron microscope (Hitachi) operating at 80 keV.

Statistical analysis

Student's t-test (two-tailed, for independent samples) was used to determine significant differences (P < 0.05).

Results

Production of recombinant β₂m with amino acid substitutions in the C-terminal region

To evaluate the effects of conformational changes in the β₂m C-terminal amino acids on amyloid formation, we produced six different kinds of recombinant β₂m with an amino acid substitution in strand G of the C-terminal region—Trp95Leu, Trp95Glu, Trp95Ala, Trp95Val, Trp95Ser and Arg97Ala (Figure 1).

We obtained and collected recombinant β₂m WT, β₂m Trp95Leu and β₂m Arg97Ala as pure protein. MALDI-TOF MS of the purified proteins revealed the molecular mass of β₂m WT, β₂m Trp95Leu and β₂m Arg97Ala to be 11,731, 11,655 and 11,645 Da, respectively. However, we could not collect and purify other recombinant proteins—β₂m Trp95Glu, β₂m Trp95Ala, β₂m Trp95Val and β₂m Trp95Ser—because of their aggregative nature.

Changes in conformation of β₂m Trp95Leu and β₂m Arg97Ala

We utilized CD analysis to investigate conformational changes of β₂m Trp95Leu and Arg97Ala. The CD spectra of β₂m Trp95Leu and Arg97Ala near 218–220 nm were broader and decreased than that of β₂m WT, which was consistent with the β-sheet structure of the native protein (Figure 2a). To investigate conformational changes around the Trp residues, we also analysed Trp fluorescence for β₂m WT, Trp95Leu and Arg97Ala. The Trp fluorescence spectrum of β₂m WT had a maximum at 334 nm, but the maximum for the β₂m Trp95Leu and Arg97Ala spectrums had shifted to 349 and 342 nm, respectively, which indicated exposure of the Trp residues. Compared with β₂m WT, β₂m Trp95Leu and Arg97Ala had lower intensity in Trp fluorescence (Figure 2b).

Figure 2

Circular dichroism (CD) spectra and tryptophan (Trp) fluorescence intensity for β₂m wild-type (WT), Trp95Leu and Arg97Ala. (a) Far-ultraviolet CD spectra of β₂m WT, Trp95Leu, and Arg97Ala. Samples of 10 μmol/L β₂m WT, Trp95Leu and Arg97Ala (in 67 mmol/L sodium phosphate, pH 7.4) were applied to a JASCO J-720 spectropolarimeter. (b) Trp fluorescence intensity for β₂m WT, Trp95Leu and Arg97Ala. Fluorescence intensity of 2 mL samples of 10 μmol/L β₂m WT, Trp95Leu and Arg97Ala was measured using a Hitachi F-2000 spectrofluorimeter. Solid line, β₂m WT; dashed lines, β₂m Trp95Leu and Arg97Ala

Amyloid fibril formation of β₂m WT, Trp95Leu and Arg97Ala at acidic pH

To examine the differences in amyloidogenesis of β₂m WT, Trp95Leu and Arg97Ala, samples were incubated at pH 2.5. The ThT fluorescence of β₂m Trp95Leu and Arg97Ala were significantly greater than that of β₂m WT (Figure 3).

Figure 3

Amyloid fibril formation of β₂m wild-type (WT), Trp95Leu and Arg97Ala at an acidic pH (pH 2.5), monitored by thioflavin-T fluorescence. β₂m WT (○), Trp95Leu (▪) and Arg97Ala (•). Data are expressed as mean ± standard deviation; n = 3. *P < 0.05 compared with β₂m WT

Ultrasonication-induced amyloid fibril formation

To enhance destabilization of β₂m, we used ultrasonication as described in Materials and Methods. Incubation of samples of β₂m WT, Trp95Leu and Arg97Ala at pH 2.5 with ultrasonication dramatically increased the ThT fluorescence intensity of β₂m Trp95Leu and Arg97Ala compared with that of β₂m WT (Figures 4a and b). With ultrasonication, the ThT fluorescence intensity peaked earlier than those without ultrasonication (Figures 3, 4a and b). Interestingly, the ThT fluorescence intensity of the β₂m samples both with and without ultrasonication, gradually decreased after 6 and 12 hours, respectively. In addition, at physiological pH, we also incubated β₂m Trp95Leu, Arg97Ala and WT with or without ultrasonication. With ultrasonication at neutral pH, the ThT fluorescence intensity of β₂m Trp95Leu and Arg97Ala increased more markedly than did that of β₂m WT, with the maximal value achieved at five days (Figure 5), although those β₂m did not form amyloid fibrils without ultrasonication (data not shown). The ThT fluorescence intensity of β₂m at neutral pH with ultrasonication peaked later than those at acidic pH with ultrasonication (Figures 4a, b and 5).

Figure 4

Time course of amyloid fibril formation induced by β₂m wild-type (WT), Trp95Leu and Arg97Ala as monitored by thioflavin-T (ThT) fluorescence. (a) ThT fluorescence intensity of ultrasonicated samples of β₂m WT, Trp95Leu and Arg97Ala at pH 2.5. (b) Early changes (0–6 hours) of β₂m fibril formation at pH 2.5 with ultrasonication. β₂m WT (○), Trp95Leu (▪) and Arg97Ala (•). Data are expressed as mean ± standard deviation; n = 3. *P < 0.05 compared with β₂m WT

Figure 5

At neutral pH (pH 7.0), ultrasonication inducing amyloid fibril formation of β₂m wild-type (WT), Trp95Leu and Arg97Ala, monitored by thioflavin-T fluorescence. β₂m WT (○), Trp95Leu (▪) and Arg97Ala (•). Data are expressed as mean ± standard deviation; n = 3. *P < 0.05 compared with β₂m WT

Time-dependent morphological changes of β₂m Arg97Ala with ultrasonication

To evaluate morphological changes in amyloid fibrils at different times (6 and 24 hours), we obtained electron microscopic (EM) images of β₂m Arg97Ala incubated with ultrasonication at acidic pH. At six hours, β₂m Arg97Ala formed long, straight fibrils and also partially curved fibrils (with diameters of approximately 10 nm) (Figure 6a). At 24 hours, when ThT fluorescence intensity of β₂m Arg97Ala was three-fold lower than that at six hours, EM images showed amorphous debris that contained long, straight fibrils (also with diameters of approximately 10 nm) (Figure 6b). The morphology of fibrils in the debris was that of typical amyloid fibrils.

Figure 6

Negative-stain electron microscopic images of amyloid fibrils of β₂m Arg97Ala incubated with seeds in 50 mmol/l glycine–HCl containing 100 mmol/L NaCl at pH 2.5 with ultrasonication. Images were obtained at (a) six hours and (b) 24 hours of incubation. Arrows indicate amyloid fibrils in the debris

Discussion

In this study, we demonstrated that the recombinant β₂m Trp95Leu and Arg97Ala, with conformational changes in the C-terminus, formed more amyloid fibrils than did β₂m WT. With ultrasonication to enhance destabilization of the protein and amyloidogenicity, β₂m Trp95Leu and Arg97Ala generated amyloid fibrils even at physiological pH. Results of the present study provided direct evidence that C-terminal amino acids of β₂m may play an important role in amyloid fibril formation.

Conformational analysis of β₂m revealed that the Trp95 of β₂m is hydrophobic and the most important residue in the C-terminus of β₂m for stability of this globular protein. In our present study, we attempted to produce recombinant β₂m with a mutation at Arg97 or Trp95. We were able to purify β₂m WT, β₂m Trp95Leu and β₂m Arg97Ala. In contrast, the other four types of recombinant β₂m with an amino acid substitution at the Trp95 residue formed aggregates and could not be purified. These results clearly suggest that Trp of β₂m at position 95 may play a critical role in the stability of the molecule by influencing folding of the native form, and that β₂m with those mutations at this amino acid may not be able to maintain stability.

To investigate conformational changes of β₂m Trp95Leu and β₂m Arg97Ala, we utilized CD and Trp fluorescence spectrum analyses. In the far-UV CD spectra, a protein with well-defined β-sheets shows a negative peak at 220 nm. ²⁵ Our data indicated that the signal intensity of the β₂m Trp95Leu and Arg97Ala spectrum decreased compared with that of β₂m WT, although the β₂m Trp95Leu and Arg97Ala spectra had negative peaks at about 220 nm. These data suggest that β₂m Trp95Leu and Arg97Ala may possess a different conformation, and as the result, it may reduce β-sheet content because of the change of the C-terminal amino acids. The positive peaks of the Trp fluorescence spectrum for β₂m Trp95Leu and Arg97Ala shifted from 334 to 349 and 342 nm, respectively, which indicates conformational changes around the Trp residues. In general, the red shift and decrease in the peak fluorescence intensity indicate the exposure of Trp residues. β₂M has two Trp residues at positions 60 and 95. In native β₂m, Trp60 connects the D and E strands and is exposed to the outside of the molecule from the beginning, whereas Trp95 lies initially in the middle of the molecule. ²⁶ Thus, the change in the Trp fluorescence spectrum (Figure 2b) may indicate the exposure of Trp95, the partial unfolding of the C-terminus.

Other researchers have also elucidated various aspects of the mechanisms of amyloid fibril formation. McParland et al. ²⁷ reported that an unfolding of both the C-terminal and N-terminal regions of β₂m is a crucial step in the initiation phase of amyloid fibril formation. Hoshino et al. ²⁸ reported that the hydrogen molecule in the A and G strands is most likely to be exchanged with deuterium. Stoppini et al. ⁸ reported that a monoclonal antibody specific for C-terminal peptides could inhibit β₂m amyloid formation. We also previously reported that a monoclonal antibody against the C-terminal residues of β₂m at positions 92–99 could detect the conformational intermediate of β₂m in amyloid deposits obtained from DRA patients. ¹⁴

Our results clearly indicated that β₂m Trp95Leu and Arg97Ala were unstable and had a greater amyloidogenic potential than did β₂m WT, although mutated β₂m had previously been investigated along with 12 other mutated types of β₂m and the authors concluded that β₂m Arg97Ala had little effect on protein stability. ^25,29 Structural prediction suggests that β₂m Arg97Ala cannot maintain the hydrogen bond with both Asn17 and Thr73, and thus hydrophobicity of the G strand is reduced (data not shown). This destabilization of the C-terminus may expose the G strand to the outside of the molecule and may produce a state of the β₂m that allows assembly of amyloid fibrils.

In this study, we also demonstrated that ultrasonication enhanced amyloid formation of β₂m at acidic and physiological pH. With ultrasonication, β₂m Trp95Leu and Arg97Ala generated more amyloid fibrils than did β₂m WT; similar relationship of the result on the two proteins was also obtained without ultrasonication. These findings confirmed that β₂m Trp95Leu and Arg97Ala had more amyloidogenic potential than did β₂m WT. Stathopulos et al. ¹⁸ previously reported that ultrasonication induced amyloid fibril formation in various proteins. More recently, Ohhashi et al. ³⁰ demonstrated that ultrasonication enhanced WT-β₂m amyloid fibril formation. Ultrasonication produces gas bubbles, high temperature and mechanical forces. ³¹ Those factors may induce destabilization of β₂m and may thus enhance amyloid fibril formation.

For the diagnosis of prion diseases, one of the amyloid forming disorders, ultrasonication was employed in PMCA in which cycles of ultrasonication broke small amount of prion fibrils into smaller pieces acting as seeds for the formation of new fibrils from soluble prion protein. ¹⁹ By the same token, using ultrasonication, we could make the diagnosis for β₂m amyloidosis earlier, if we could amplify and detect small amount of β₂m fibrils in the early stage of DRA patients. Further investigations are needed.

It is interesting that ThT fluorescence intensity gradually decreased with or without ultrasonication after 6 and 12 hours, respectively. At 24 hours, when the ThT fluorescence intensity of β₂m was three-fold lower than that at six hours, EM images of β₂m incubated with ultrasonication showed amorphous debris containing long, straight fibrils, that is, typical amyloid fibrils (Figure 6b). This debris was not observed at six hours (Figure 6a). These results suggest that ThT may have difficulty to bind amyloid fibrils in debris and that the decreasing ThT fluorescence intensity was caused simply by this binding difficulty.

In conclusion, conformational changes in the C-terminus of β₂m may play an important role in amyloid fibril formation of DRA. Ultrasonication is a useful technique to enhance β₂m amyloid fibril formation even in the physiological solution.