Mass and physical dimensions of two distinct populations of paired helical filaments

Neurobiologyof Aging, Vol. 15, No. 1, pp. 11-19, 1994 Copyright © 1994 ElsevierScience Ltd Printed in the USA. All rights reserved 0197-4580/94 $6.00 + .00

Pergamon

Mass and Physical Dimensions of Two Distinct Populations of Paired Helical Filaments H A N N A K S I E Z A K - R E D I N G *l A N D J O S E P H S. W A L L i -

*Department of Pathology, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461 ~Biology Department, Brookhaven National Laboratory, Upton, Long Island, NY 11973 R e c e i v e d 15 February 1993; R e v i s e d 16 A u g u s t 1993; A c c e p t e d 26 A u g u s t 1993 KSIEZAK-REDING, H. AND J. S. WALL. Mass and physical dimensions of two distinctpopulations of paired helicalfilaments. NEUROBIOL AGING 15(1) 11-19, 1994.--We studied the ultrastructure of two fractions of paired helical filaments (PHF) from Alzheimer brains separated on sucrose density gradient. Fraction A2 (IM sucrose) contained filaments which were short in length and did not aggregate while those in fraction AL2 (1/1.5 M sucrose interface) were mostly aggregated. By scanning transmission electron microscopy, PHF in fraction A2 had significantly more mass per nm length of filament (107-120 kD/nm) than those in fraction AL2 (79-85 kD/nm), and they were also wider in their maximum and minimum widths but did not differ in their periodicity. Differences in mass and dimensions between two morphologically distinct populations of PHF suggest that a partial proteolysis may be involved in the generation of the aggregated population of PHF. The results suggest that a similar process may be active in the formation of neurofibrillary tangles. STEM Mass of PHF Proteolysis

Dimensions of PHF

Microtubule associated proteins tau

Alzheimer's disease

Comparisons of various populations of PHF at the ultrastructural level with respect to their mass or physical dimensions have been limited (7). The mass of tangle-derived PHF has been examined by scanning transmission electron microscopy (STEM) and estimated to be in the range of 60-100 kD with an average of 79 kD per nm length of filament (42). The mass of SDS-soluble PHF however, has not yet been determined. Such measurements will be necessary to determine the size of a structural subunit in a model of PHF developed by Crowther and Wischik (9). In this model, based on computer image reconstruction techniques, PHF are described as helically twisted ribbon composed of a double stack of subunits. According to the model and the STEM data, the mass of a single structural subunit in tangle-derived PHF was estimated to be 110 kD (40,42). It is not known whether other populations of PHF, e.g., SDS-soluble PHF, are composed of subunits of similar size. In the present study we examined the ultrastructure of PHF in two fractions obtained from Alzheimer brain. These fractions sedimented in different layers of a sucrose density gradient and contained PHF in different aggregation state. Other differences such as SDS-solubility and tau immunoreactivity have been reported elsewhere (27,28). By STEM analysis, we have demonstrated that fractions of less aggregated PHF contained filaments larger in mass per unit length and wider in physical dimensions than those in fractions of more aggregated PHF. The maximum size of a structural subunit in less aggregated PHF was estimated to be 173 kD. It was 71% larger than that for more aggregated PHF and 53% larger than that previously reported for tangle-derived PHF.

PAIRED helical filaments (PHF) accumulate in neurofibrillary tangles, neuritic plaques and neuropil threads in Alzheimer disease brain (2,21,35). Several immunocytochemical studies have demonstrated that PHF in these lesions share epitopes with tau (3,16, 23), a family of microtubule associated proteins of 55~52 kD (6,38). Analysis of PHF accumulated in tangles with EM reveals characteristic parallel bundles of highly aggregated filaments. These filaments measure approximately 18-20 nm in maximum width, narrowing to 8-10 nm every 70-90 nm (21,35,43). Preparations of tangle-derived PHF have limited solubility in SDS (18,32,45) and only with repeated extractions or other treatments can peptide fragments of 45-62 kD and smaller be obtained (19). From tangle-enriched preparations peptide fragments of 9 kD and 12 kD tightly bound to PHF were isolated and found to contain the microtubule binding region of the tau molecule (13,22,41,42). This region has been reported to be important for the structural stability of PHF (29) and has recently been shown to be capable of self-assembly into Alzheimer-like PHF (8,39). Another population of PHF has been isolated from Alzheimer brain and found soluble in SDS (14,29,30). This SDS-soluble population of PHF differed from tangle-derived PHF in morphology and biochemical composition. With EM examination, SDSsoluble PHF were shorter in length and less aggregated than tangle-derived PHF. Western blot analysis revealed that SDS-soluble PHF were primarily composed of three polypeptides of 60, 64, and 68 kD, which were immunoreactive with antibodies to tau (24,31). These polypeptides were characterized as modified or hyperphosphorylated tau and named PHF-tau or A68 (I 0,15,24,26,30).

i Requests for reprints should be addressed to Hanna Ksiezak-Reding, Department of Pathology, Rm. F-538, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461. 11

i2

KSII;ZAK-RE[)[N~} /\N!} WA~; METHOD

Material Brain tissue was obtained from five patients with neuropathologically confirmed Alzheimer disease using criteria of Khachaturian (20). Data on individual patients and status of brain tissue is listed in Table 1. Tissue was stored at - 7 0 ° C until use.

Isolation of PHF Our isolation procedure was based on that described previously (26,29,31). Brain tissue was homogenized in 5 volumes of 20 mM 2-(N-morpholino)ethanesulfonic acid (MES)/NaOH, pH 6.8, 80 mM NaCI, 1 mM MgCI2, 2 mM EGTA, 0.1 mM EDTA. The homogenate was separated by centrifugation (20 min at 27,000 × g) into a supernatant which contained most of normal tau proteins and a pellet which contained most of PHF. The pellet was rehomogenized in 10 volumes of 0.8 M NaCI, 10% sucrose, 10 mM MES/NaOH, pH 7.4, and 1 mM EGTA and centrifuged as above. The supernatant was incubated with Sarcosyl (1% final concentration) for 1 h at room temperature or overnight at 4°C. The mixture was then centrifuged (2 h at 100,000 × g; Beckman rotor SW41, 24 000 rpm) and the crude PHF pellet was resuspended in 10 mM MES/NaOH, pH 7.0 (0.2 ml/g tissue). The suspension, containing numerous PHF, was subjected to further separation on a sucrose density gradient (2 h at 100,000 x g). The discontinuous gradient was constructed from 2.5 ml layers of 1 M, 1.5 M, and 2 M sucrose in 10 mM MES/NaOH, pH 7. Several fractions of PHF were obtained and examined by EM. Most of the fractions were characterized by a small number of PHF and the presence of amorphous material. Two fractions however, contained most of PHF. These fractions, the 1 M sucrose layer (fraction A2) and the 1 M/1.5 M sucrose interface (fraction AL2) were used for further analysis by STEM.

Electron Microscopy Copper grids (200 mesh) coated with Formvar and carbon were used (Fullam, Inc., N J). Grids were placed on top of 25-50 ml aliquots of PHF fractions for 1-5 min, washed with phosphate buffered saline and water, and stained with 2% uranyl acetate in water.

STEM For the STEM analysis at the Brookhaven National Laboratory (Upton, NY), PHF fractions were treated according to methods described earlier (37). Briefly, samples of PHF in 1 M or 1.5 M sucrose with 10 mM MES/NaOH, pH 7 were deposited on grids precoated with an internal mass standard, tobacco mosaic virus

TABLE l DATA ON ALZHEIMER PATIENTS

Case Number

1 2 3 4 5

Sex/Age

Duration of AD (y)

Post Mortem (h)

Brain Region

NFI"

M62 F72 F89 F82 M79

2 11 25 16 6

4.5 7.5 6 12 7

Frontal Occipital Frontal Temporal Parietal

7-10 3-5 9-12 8-15 12-14

Tissue sections were stained with histopathological methods at the time of autopsy and the number of neurofibrillary tangles (NFT) was determined under x40 magnification. Cases 1, 2, 3, and 5 were used also in other studies for phosphate analysis (26; Cases l, 7, 5, and 8, respectively).

II31 kD/nm). Grids were then washed lt! tm~es with 20 mM ammonium acetate, freeze-dried, and subjected .,_oexantinati,.m b) STEM. STEM images were recorded digitally as well as on mi crographs. The recorded data and a computerized program werc used to determine the mass of filaments in reference to a standard mass. For computations, areas with minimum interference of ~c sidual salt and sucrose were selected as judged by the mass per urm length and radial profile of the tobacco mosak virus. Statistical analysis of STEM measurements was performed using Student'> ; test, one-way analysis of variance tANOVA ~. linear regressiot~ and correlation coefficient ~46)

Measurements q[ Dimension,~ The minimum and maximum widths of PHF were determined on STEM micrographs using measuring microscopes of 10x and 20x (Ted Pella, Inc.). In addition, the half-periodicity, defined as a distance between consecutive constrictions or helical twists in filaments, was also measured. RESUI/FS

Characterization of PHF Fractions Crude fractions of PHF contained morphologically mixed population of PHF as demonstrated by fractional separation on sucrose density gradients. EM and STEM examinations revealed that fraction A2, separated in 1 M sucrose, contained predominantly short and nonaggregated filaments (Fig. 1 A and C). The minimum length of these filaments was a half of helical twist (approximately 4 0 4 5 nm) and the maximum length rarely exceeded 4 twists (approximately 310--360 nm). In contrast, fraction AL2, separated at the interface of 1 M/1.5 M sucrose, was enriched for aggregated filaments (Fig. 1 B and D). These aggregates were composed of a dense mass of filaments which could be distinguished as individual filaments at the periphery. Some of the aggregates were made of parallel bundles of filaments as shown in Figs. l B and 2B and reported elsewhere (27). A limited amount of filaments was not aggregated and formed no clumps. A detail biochemical characterization of filaments in fraction A2 presented in several studies have demonstrated that filaments sedimented in l M sucrose were readily soluble in SDS and primarily composed of three polypeptides of PHF-tau (14,15,29,30). Similar characterization of filaments in fraction AL2 has only recently been performed (27). In that study, direct comparisons with fraction A2 have indicated that fraction AL2 contained filaments with reduced SDS solubility and decreased immunogold labelling with antibodies to the N-terminus of tau. In the present paper, the differences in the intensity and pattern of immunogold labelling of filaments in these fractions have been illustrated using Alz 50. The monoclonal antibody, directed to an epitope in the first 10 amino acid residues of tau (11,25), immunolabelled short and dispersed filaments in fraction A2 (Fig. 2A) and some of the short filaments present in fraction AL2 (Fig. 2B, arrowheads). It did not label, however, or labelled only sparsely the aggregated bundles of filaments characteristic of fraction AL2 (Fig. 2B). The results suggested the possibility that aggregated population of PHF was composed of filaments that were partially digested or were digested more than those found in nonaggregated population. Therefore, we performed direct comparisons of physical mass of filaments in two fractions of PHF.

Mass Measurements by STEM: Maximum Versus Minimum Width Measurements of mass per nm length of PHF were performed in various regions of filaments to determine whether distribution of

MASS AND PHYSICAL DIMENSIONS OF PHF

13

/ .......

4)

FIG. 1. Comparison of morphology of PHF by EM and STEM. PHF from Case 3 were separated on a sucrose gradient in fraction A2 (1 M sucrose) and fraction AL2 (1 M/I.5 M sucrose interface) and examined by EM (A and B, respectively) or STEM (C and D, respectively). Samples were used without fixation and were either stained with uranyl acetate (EM) or unstained (STEM). In STEM analysis, the mosaic tobacco virus served as a reference mass (in C, arrowhead). In samples of AL2 fraction (B and D) note the presence of PHF aggregated in parallel bundles of filaments and forming a clump. Due to a dark field background both aggregated and nonaggregated filaments appear delineated better in STEM (C and D) than in EM (A and B). Scale bar corresponds to 100 nm.

mass was uniform along the length of filament. Regions of minimum and maximum widths were of particular interest (Fig. 3). STEM analysis indicated that both of these regions had similar values of mass per unit length of filaments (Table 2). For example, in fraction A2 from two Alzheimer cases, the mass at the minimum width of filament was 106-114 kD/nm while the maximum width was 110-116 kD/nm. In fraction AL2, masses at the minimum and maximum width of filaments were similar to each other in one case (Case 3) while in another (Case 5), the difference in mass was at the border of statistical significance (p = 0.05). Similarity in mass values determined at the maximum and minimum widths suggests that effects of residual salts and sucrose in our mass measurements were minimal. Additional comparisons of mass values obtained at ends of filaments with that in the middle assured that there was no edge effect in our measurements (not shown). The results indicate that the distribution of mass is uniform along the length of filaments and is not affected by variations in widths.

Frequency Distribution of Mass Individual measurements of mass at the minimum or maximum width of filaments in each fraction were plotted as frequency distribution. In fractions A2 from four Alzheimer cases, the majority

of filaments ( > 5 4 % ) were characterized by a mass of 90-130 kD/nm (Fig. 4 A-D). In two of these cases (Case 3 and 4), the frequency distribution was fairly symmetrical suggestive of relative homogeneity of PHF (Fig. 4 C and D). In other cases (Case 1 and 2), distribution was asymmetrical due to the presence of a small number of filaments with a mass of 170 kD/nm or higher (Fig. 4 A and B), perhaps aggregates of two filaments. This asymmetry was unilateral and observed only in the higher mass range. The majority of filaments in these two cases displayed a symmetrical distribution similar to that previously shown (Fig. 4 C and D). In fractions AL2 from two cases, the frequency distribution revealed that the majority of filaments ( > 7 2 % ) had a mass in a range of 60-100 kD/nm (Fig. 4 E and F). In one case (Case 3), the mass of filaments was relatively uniform and this resulted in a more symmetrical distribution than in another (Case 5). In fractions AL2 from three other cases, mass measurements were performed on few filaments only, insufficient for the statistical analysis. It was due to the extensive aggregation of filaments in these fractions and difficulties in obtaining a relatively unobstructed view of filaments suitable for measurements as in Fig. 1D. In fraction A2, the average mass values were 107-120 kD/nm and they were 26%-52% higher than that in fraction AL2, which ranged from 79-85 kD/nm (Table 3). Comparisons of the average

14

KSIEZAK-REI)iNG ~ND ~ A i .

FIG. 2. lmmunogold electron microscopy of fraction A2 (A) and fraction AL2 (B) with the monoclonal antibody Alz 50. Note a decoration of short and nonaggregated filaments in fraction A2. In fraction AL2, a sparse decoration of bundles of aggregated filaments is noticed in contrast to a heavy decoration of occasional short filaments (arrowheads). lmmunolabelling with 10-nm colloidal gold was performed essentially as described earlier (29). Scale bar corresponds to 100 nm.

mass values and patterns of the frequency distribution indicated that filaments in fractions A2 and AL2 distinctly differed from each other in mass per unit length. The conclusion that filaments in these fractions represent two different populations of PHF with either higher (fraction A2) or lower mass (fraction AL2) was confirmed by A N O V A (Table 3).

Physical Dimensions of PHF: Correlation With Mass In fractions A2 from four Alzheimer cases, the maximum width of filaments was 22-26 nm while the minimum width was 11-16 nm (Table 3). In fractions AL2 from two Alzheimer cases, the maximum and minimum widths of filaments were smaller and were 16-20 nm and 8-11 nm, respectively. Overall, the physical dimensions of filaments in fraction A2 were larger by approximately 33-45% than that observed in fraction AL2. ANOVA fur-

ther confirmed that filaments in one fraction AL2 (Case 3) differed significantly in width from filaments in all four fractions A2 examined. Filaments in another fraction AL2 (Case 5) significantly differed in width from filaments in two fractions A2 (Cases 1 and 2). The half periodicity was similar in all filaments and ranged from 7 % 9 0 nm. In all fractions, filaments displayed a general trend in that filaments with higher mass were also wider in dimensions. Therefore, we examined whether there was any correlation between variations in mass and the maximum or minimum width of filaments. The existence of such a direct relationship was conf'mned in PHF from all fractions A2 and AL2 combined (Fig. 5). The relationship between mass and dimensions of filaments was highly significant because the respective correlation coefficients were r = 0.908 (p < 0.02, n = 6) for the mass and maximum width of filaments and r = 0.834 (p < 0.05, n = 6) for the mass and the

MASS AND PHYSICAL DIMENSIONS

OF PHF

15

FIG. 3. Determination of mass of PHF by STEM. Samples of fraction A2 (A and C) and fraction AL2 (B and D) were deposited on grids precoated with a reference tobacco mosaic virus (asterisk) and examined by STEM. In C and D, identical to A and B, respectively, various areas of PHF were selected at the maximum or the minimum width (in C and D: pixels numbered 1-8 and 1-11, respectively) to compare their mass with that of the virus (in D: pixels numbered 1-4). The average mass of filaments in fraction A2 (A and C) was determined as 115 kD/nm and that in fraction AL2 (B and D) as 79 kD/nm (see Table 3, Case 3) in comparison to that of the virus (131 kD/nm). The STEM micrographs were also used to determine physical dimensions of PHF presented in Tables 2 and 3. In E, a higher magnification of a filament is shown (fraction AL2). Scale bars correspond to 100 nm. m i n i m u m width (Fig. 5). For c o m p a r i s o n , there w a s no relationship b e t w e e n the m a s s and half-periodicity as the respective correlation coefficient w a s r = 0 . 0 7 (ns; not s h o w n ) . Data obtained for m a s s and the physical d i m e n s i o n s o f filam e n t s allowed us to estimate the average m a s s density o f f i l a m e n t s in both fractions o f PHF. If we define a m a s s density o f P H F as m a s s per n m cube o f filament [length × width ( = m a x i m u m width) x t h i c k n e s s ( = m i n i m u m width)], fraction A 2 contained f i l a m e n t s with an a v e r a g e m a s s density o f 0 . 2 9 - 0 . 4 4 k D / n m 3 (4 cases). In c o m p a r i s o n , the m a s s density o f f i l a m e n t s in fraction A L 2 from two c a s e s w a s either within the range (0.39 k D / n m 3) or higher (0.62 k D / n m 3) than that in fraction A2. In two different fractions o f P H F obtained f r o m the s a m e brain tissue (Case 3), the difference in m a s s d e n s i t y o f f i l a m e n t s w a s m o r e evident since P H F in fraction A2 were m u c h less d e n s e (0.40 k D / n m 3) than that in fraction A L 2 (0.62 kD/nm3). DISCUSSION O u r studies d e m o n s t r a t e that P H F separated on s u c r o s e density gradient in two fractions, A 2 a n d A L 2 , differ in their physical d i m e n s i o n s a n d m a s s . P H F collected in fraction A 2 were approximately 3 2 % - 4 2 % wider in m i n i m u m and m a x i m u m width than those in fraction A L 2 . T h e y did not differ in half-periodicity. By

S T E M e x a m i n a t i o n , P H F in fraction A2 also carried approximately 2 5 % - 5 2 % more m a s s per n m length than those in fraction A L 2 . A N O V A of data collected f r o m 5 c a s e s o f A l z h e i m e r d i s e a s e indicated that P H F separated into two fractions represented two distinct populations of PHF. P H F obtained in fraction A 2 were the heaviest and widest TABLE 2 COMPARISON OF MASS OF FILAMENTS AT THE MAXIMUM AND MINIMUM WIDTHS Mass. kD/nm length Fraction A2 AL2

Case Number

Maximum Width

Minimum Width

p

3 4 3 5

116-+4(29) 110 -+ 3 (46) 75 -+ 2 (16) 90 -+ 4 (19)

114-+4(27) 106 -+ 3 (48) 81 -+ 3 (10) 78 -4- 3 (33)

ns ns ns = 0.05

Values represent means -+ SEM for the number of individual measurements in the parentheses. Student's t test was employed to determine if the differences in mass between maximum and minimum widths were statistically significant.

I(~

KSIEZAK-RH)I~'~(~ :\NI~ V<'~i i -]

B[

A /

~

~

,,

Ii

Case#t I

i,

l

t

n

I{

t Fraction A2

E

5

,

C

4

t

j

040

DI

~0., Case # 3

80 100 120 140 150 ~80 200 Z~0 Mass, kD/nm

I t

Case # 4

II

Fraction

l

[]

[]

8 6 4 z

!

3~, 3o.I

I I

[

,

Z

60 80 tOO IZO 140 160 180 ZOO Z.ZO MASS, kD/nm

4 z

I Fraction A2! J

6

i I

r-r=

0

! Case # 2 t

Fraction

AZ 1

%

% 6/~00J 100 Joo lzo ,,o 16o lao zoo ZZO 2 Mass, kD/nm

E Case

F

# 3

60 O0 100 120 t40 160 1110ZOO 220 Mass, kO/nm

I i 1

FractionAL2

40 GO

I~

IZO 140 150

160 2OO 220

Mass, kOlnm

o

Fraction

AL2

40 60 80 ?~ 1ZO 14~D160 100 ZOO Z20 Mass, kD/nm

FIG. 4. Histograms showing the frequency distribution of mass of filaments in PHF fractions from 5 cases. Mass per umt length of filaments was estimated in samples of fraction A2 (A, B, C, and D; open bars) in four cases (Case 14) and in samples of fraction AL2 (E and F; filled bars) in two cases (Case 3 and 5). Number of measurements varied from 44 to 115 for each case. The distribution has a tail toward higher mass values in some fractions (A, B. and F).

filaments that have yet been obtained from Alzheimer brain tissue (14,29,30). The mass of PHF in fraction A2 was found to be 107-120 kD/nm length with the maximum and the minimum widths of 22-26 nm and 11-16 rim, respectively. In comparison, the mass of filaments in fraction AL2 (79-85 kD/nm) and dimensions at the maximum (16-20 nm) or minimum (8-11 nm) width were similar to those reported previously for tangle-derived PHF (42,43). The mass of filaments from both fractions seemed to be distributed uniformly along the length of filaments since the mass at the maximum and minimum widths was either similar (fraction A2) or within -+ 8 % - 1 5 % from one another (fraction AL2). STEM analysis has been used successfully in ultrastructural studies of enzyme complexes such as dynein (36), pyruvate dehydrogenase (5), or alpha-keto acid dehydrogenase (17). It made possible to determine the molecular weight as well as the number and arrangement of the component enzymes in these multisubunit complexes. In the present study, STEM analysis and physical measurement of filaments permitted us to estimate the size of the structural subunit in both populations of PHF defined in the model of PHF by Crowther and Wischik (9). According to that model, the structural subunit of PHF is a C-shaped molecule that is arranged in dimers and stacked vertically to form a double-helical

strand of subunits (40). Based on a detected 3 nm axial spacing, the structural PHF subunit is expected to have a thickness of 3 nm along the axis of filament and a mass equal to half of mass/nm length x 3 nm. In the present studies, considering PHF in fraction A2, the calculated mass of the structural subunit was 173 kD (1/2 x 115 kD/nm × 3 nm). In fraction AL2, the mass of the structural subunit was smaller, 123 kD (1/2 × 82 kD/nm × 3 nm), and close to the value of 110 kD estimated previously for tangle-derived PHF using similar approach (42). Assuming that PHF-tau is a major component of PHF, and its average molecular weight is 41.3 kD based on the primary sequence of variants of normal tau (12), the structural subunit of PHF may consist from as many as four molecules of PHF-tau (fraction A 2 : 1 7 3 kD/41.3 kD = 4.2 molecules). Similar calculations may not be valid for PHF in fraction AL2 due to the possibility that these PHF are composed of digested fragments rather than full molecules of PHF-tau. Somewhat surprising was the fact that filaments with more intact structure and more mass per nm length were separated in a less dense sucrose layer (fraction A2) than those with 21-34% less mass (fraction AL2). It indicated that factors other than mass were responsible for the separation. One of these factors could be the extent of aggregation of filaments since nonaggregated filaments

MASS AND PHYSICAL DIMENSIONS OF PHF

17

TABLE 3 MASS AND PHYSICAL DIMENSIONS OF PHF Dimensions, nm

Fraction

A2

Case Number

Mass kD/nm

Maximum Width

Minimum Width

Half Periodicity

1

120 z 2 (73) 113 ± 3 (83) 115 ~ 2 (103) 107 ± 2* (115) 79 ± 1"** (44) 85 ± 3*** (69)

26 ± 1 (18) 25 ± 1 (23) 22 ~ 0.4* (41) 22 ~ 1" (24) 16 ~ 0.3*** (27) 20 ~ 0.4** (28)

16 ± 1 (17) 15 ± 1 (24) 12 ~ 0.3* (44) 11 ± 0.4* (29) 8 ± 0.3*** (24) 11 ¢ 0.5** (26)

84 ± 3 (13) 83 ± 2 (10) 80 ± 2 (29) 89 ± 3 (16) 79 ± 2 (20) 90 ± 4 (10)

2 3 4 AL2

3 5

Values represent means +-- SEM for the number of individual measurements in the parentheses. Fractions A2 and AL2 were obtained at different layers of discontinuous sucrose gradient according to Methods. Statistical analysis of multiple comparisons by one-way ANOVA (46) was performed using StatView program. Values that significantly differ from those in other cases in fraction A2 are marked *(0.025 < p < 0.05). Values in fraction AL2 that significantly differ from the corresponding values in fraction A2 are marked **(0.025 < p < 0.05 vs. Case 1 and 2) or ***(/9 < 0.001 vs. Case 1-4).

sedimented in less dense sucrose (1 M sucrose) than those heavily aggregated (1 M/1.5 M sucrose). Other factors, such as a difference in mass density of PHF, could also affect the separation of PHF. This was most evident in fractions A2 and AL2, isolated from the same brain (Case 3). The mass density of PHF in these fractions was estimated as 0.40 and 0.62 kD/nm 3, respectively, indicating that more intact filaments were less dense than those partially degraded. Additional studies are needed to confirm this finding. Other differences in PHF obtained from various sucrose gradient fractions have been reported. PHF in fraction A2 were found more soluble in SDS as compared with those in fraction AL2 (27) or similar fractions (30). Moreover, immunolabelling of SDS soluble PHF with antibodies to the N-terminus of tau (e.g., with Alz 50 or Tau 14) has recently been reported to decrease in less SDS soluble PHF (27). Together with the present studies, it strongly suggests that PHF undergo partial proteolytic degradation at the N-terminus of PHF-tau molecule and that at least one third of the molecule is digested. It is very likely that these two fractions of PHF which differ by several biochemical, morphological, and physical criteria originate from different Alzheimer lesions, Considering PHF in fraction AL2, these filaments share some properties with neurofibrillary tangles: they are aggregated and have physical dimensions and mass per nm length similar to that of tangle-derived PHF. In contrast, PHF in fraction A2 share less similarities with tangles and more similarities with nontangle lesions such as abnormal neurites (33) or pretangle inclusions described in the cytoplasm of affected neurons (1,34,44). By their larger dimensions and mass, filaments in fraction A2 are likely to be a precursor of PHF accumulating in neurofibrillary tangles. In the process of tangle formation, however, a partial proteolysis of PHF would be required to reduce the mass and width of filaments by 33%-39%. It is then conceivable that accumulation of PHF in neurofibrillary tangles in Alzheimer brain may be a result of the proteolytic activities taking place during the progress of the disease.

In our previous studies, proteolytic digestion of SDS-soluble PHF in vitro resulted in the removal of 60% of the mass based on changes in electrophoretic mobility of PHF-tau fragments (29). Digestion was accompanied by a disproportionally small 10%20% decrease in the width of the filaments. In the present studies,

30

E C

20

C

.g W =.O

E ~3

10

0

5O

100

150

Mass, kD/nm FIG. 5. A linear regression between mass and physical dimensions of PHF: maximum (MAX) or minimum (MIN) width. Regression was obtained using values for fraction A2 (open symbols) and fraction AL2 (filled symbols) presented in Table 3. Correlation coefficients were r = 0.908 (p = 0.0124, n = 6) and r = 0.834 (p = 0.0389, n = 6) for the relationship between mass and maximum or minimum width, respectively. There was no relationship between mass and half periodicity of filaments (not shown) as the respective correlation coefficient was r = 0.068 (p = 0.9, n = 6).

18

KSIEZAK-REI)ING ANI'~ WAi,,

differences in mass (by 39%) and dimensions (by 33%} of two fractions of PHF were observed at the same comparable level. Moreover, by simultaneous measurements of both parameters in unfixed and unstained preparations, a strong, linear correlation between mass and width of PHF has been described (r - 0.8340.908; p < 0.05). Such a discrepancy may be due to differences in the techniques used to determine the mass in both studies (SDS-gel electrophoresis vs. STEM) and the fact that dimensions have previously been measured in fixed and negatively stained preparations. The use of fixatives and stains has been shown to affect some of the PHF dimensions (29.43). In studies in vitro, proteolytic degradation of SDS-soluble PHF was found to be responsible for the aggregation of filaments and proteolysis was proposed to be involved in the formation of neurofibrillary tangles (29). Here wc confirm this view and describe the existence of two morphologically different populations of PHF in Alzheimer brain, one nonaggregated and relatively intact and another aggregated and partially degraded. Although the presence of two populations of PHF has previously been indicated, the

differences between them as duc to proteolyti~ degradation wt:~c not well understood. Proteolytic digestion and self-aggregation oi proteolytic fragments have been implicated in the formation ,'.f other abnormal inclusions accumtdating in Aizheimer braim ~t,~ example deposits of amyloid (41. Moreover. fragments ol normai tau encompassing 3-repeat microtubule binding domain have bec~:~ reported to self-assemble into PHF-like structures (8.39}. ht the present paper, we provide an additional evidence to suggest that similar processes, proteolytic digestion and aggregation ,)f digested filaments, may be active in the formation of neurofibrillar\ tangles from non aggregated population of PHI: ACKNOWIEDGMENTS We thank Dr. Dennis W. Dickson for providing autopsy material and commenting on the manuscript, Dr. Peter Davies for a generous supply of Alz 50, and Dr. Martha N. Simon for her help in the STEM operation and collection of data. Studies were supported by National Institute of Health Grants AG06803 and NS30027.

REFERENCES 1. Bancher, C.; Brunner, C.; Lassmann, H.; Budka, H.; Jellinger, K.: Wiche, G.; Seitelberger, F.; Grundke-lqbal, 1.; lqbal, K.: Wisniewski, H. M. Accumulation of abnormally phosphorylated tau precedes the formation of neurofibrillary tangles in Alzheimer's disease. Brain Res. 477:90-99; 1989. 2. Braak, H.; Braak, E.; Grundke-Iqbal, I.; lqbal, K. Occurence of neuropil threads in the senile human brain and in Alzheimer's disease: A third location of paired helical filaments outside of neurofibrillary tangles and neuritic plaques. Neurosci. Lett. 65:351-355; 1986. 3. Brion, J. P.; Passareiro, H.; Nunez, J.; Flament-Durand, J. Mise en evidence immunologique de la proteine tau au niveau des lesions de degenerescence neurofibrillaire de la maladie d'Alzheimer. Arch. Biol. (Bruxelles). 95:229-235; 1985. 4. Burdick, D.; Soreghan, B.; Kwon, M.; Kosmoski, J.; Knauer, M.: Henschen, A.; Yates, J.; Cotman, C.; Glabe, C. Assembly and aggregation properties of synthetic Alzheimer's A4/Beta amyloid peptide analogs. J. Biol. Chem. 267:546-554; 1992. 5. CaJacob, C. A.; Frey, P. A.; Hainfeld, J. F.; Wall, 1. S.; Yang, H. Escherichia coil pyruvate dehydrogenase complex: particle masses of the complex and component enzymes measured by scanning transmission electron microscopy. Biochemistry 24:2425-2431: 1985. 6. Cleveland, D. W.; Hwo, S.-Y.; Kirschner, M. W. Physical and chemical properties of purified tan factor and the role of tan in microtubule assembly. J. Mol. Biol, 116:227-247; 1977. 7. Crowther, R. A. Structural aspects of pathology in Alzheimer's disease. Biochim. Biophys. Acta 1096:1-9; 1991. 8. Crowther, R. A.; Olesen O. F.; Jakes, R.; Goedert, M. The microtubule binding repeats of tau protein assemble into filaments like those found in Alzheimer's disease. FEBS Lett. 309:199-202; 1992. 9. Crowther, R. A.; Wischik, C. M. Image reconstruction of the Alzheimer paired helical filament. EMBO J. 4:3661-3665; 1985. 10. Goedert, M.; Spillantini, M. G.; Cairns, N. J.: Crowther, R. A. Tau proteins of Alzheimer paired helical filaments: abnormal phosphorylation of all six brain isoforms. Neuron 8:159-168; 1992. 11. Goedert, M.; Spillantini, M. G.; Jakes, R. Localization of the Alz 50 epitope in recombinant human microtubule-associated protein tau. Neurosci. Lett. 126:149-154; 1991. 12. Goedert, M.; Spillantini, M. G.; Jakes, R.; Rutherford, D. ; Crowther, R. A. Multiple isoforms of human microtubule-associated protein tau: sequences and localization in neurofibrillary tangles of Alzheimer's disease. Neuron 3:519-526; 1989. 13. Goedert, M.; Wischik, C. M.; Crowther, R. A.; Walker, 1. E.; Klug, A. Cloning and sequencing of the cDNA encoding a core protein of the paired helical filament of Alzheimer disease: Identification as the microtubule-associated protein tau. Proc. Natl. Acad. Sci. USA 85: 4051-4055; 1988. 14. Greenberg, S. G.; Davies, P. A preparation of Alzheimer paired he-

15. 16. 17.

18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30.

lical filaments that displays distinct tau proteins by polyacrylamide gel electrophoresis. Proc. Natl. Acad. Sci. USA 87:5827-5831; 1990. Greenberg, S. G.; Davies, P.; Schein, J. D.; Binder, L. I. Hydrofluoric acid-treated tau-PHF proteins display the same biochemical properties as normal tan. J. Biol. Chem. 267:564-569; t992. Gmndke-lqbal, I.; Iqbal, K.; Quinlan, M.; Tung, Y-C.; Zaidi, M. S.: Wisniewski, H. M. Microtubule protein tan: A component of Alzheimer paired helical filaments. J. Biol. Chem. 261:6084--6089; 1986. Hackert, M. L.; Xu, W.-X.; Oliver, R. M.; Wall, J. S,; Hainfeld, J. F.; Mullinax, T. R.; Reed, L. J. Branched-chain alpha-keto acid dehydrogenase complex from bovine kidney: radial distribution of mass determined from dark-field electron micrographs. Biochemistry 28:6816-6821; 1989. lqbal, K.; Grundke-Iqbat, I.; lqbal, K.; Zaidi, M. S.; All, N.: Wisniewski, H. M. Are Alzheimer neurofibrillary tangles insoluble polymers? Life Sci. 38:1695-1700; 1986. Iqbal, K.; Zaidi, T.; Thompson, C. H.; Merz, P. A.; Wisniewski, H. M. Alzheimer paired helical filaments: Bulk isolation, solubility and protein composition. Acta Neuropathol. (Bed) 62:167-177; 1984. Khachaturian, Z. S. Diagnosis of Alzheimer's disease. Arch. Neurol. 42:1097-1105; 1985. Kidd, M. Alzheimer's disease - an electron microscopical study. Brain 87:307-320; 1964. Kondo, J.; Honda, T.; Mori, H.; Hamada, Y.; Miura, R.; Ogawara, M.; Ihara, Y. The carboxyl third of tau is tightly bound to paired helical filaments. Neuron 1:827-834; 1988. Kosik, K. Minireview: The molecular and cellular pathology of Alzheimer neurofibrillary lesions. J. Gerontol. 4:55-58; 1989. Ksiezak-Reding, H.; Binder, L.; Yen, S-H. Alzheimer disease proteins (A68) share epitopes with tan but show distinct biochemical properties. J. Neurosci. Res. 25:420-430; 1990. Ksiezak-Reding, H.; Chien, C.-H.; Lee, V. M. Y.; Yen, S.-H, Mapping of the Alz 50 epitope in microtubule-associated proteins tan. J. Neurosci. Res. 25:412-419; 1990. Ksiezak-Reding, H.; Liu, W.-K.; Yen, S.-H. Phosphate analysis and dephosphorylation of modified tau associated with paired helical filaments. Brain Res. 579:209-219; 1992. Ksiezak-Reding, H.; Morgan, K. A.; Dickson, D. W. Tau immunoreactivity and SDS solubility of two populations of paired helical filaments that differ in morphology (submitted). Ksiezak-Reding, H.; Morgan, K. A.; Wall, J. S. Two distinct populations of paired helical filaments differ in mass, solubility, and tau immunoreactivity. J. Cell Biol. 115:383a; 1991. Ksiezak-Reding, H.; Yen, S.-H. Structural stability of paired helical filaments requires microtubule-binding domains of tau: A model for self-association. Neuron 6:717-728; 1991. Lee, V. M.-Y.; Balin, B. J.; Otvos, L.; Trojanowski, J. Q. A68: A

MASS AND PHYSICAL DIMENSIONS OF PHF

31.

32.

33.

34.

35. 36.

37.

38.

39.

major subunit of paired helical filaments and derivatized forms of normal tau. Science 251:675-678; 1991. Liu, W.-K.; Ksiezak-Reding, H.; Yen, S.-H. Abnormal proteins from Alzheimer's disease brain. Purification and amino acid analysis. J. Biol. Chem. 266:21723-21727; 1991. Selkoe, D. J.; Ihara, Y.; Salazar, F. J. Alzheimer's disease: Insolubility of partially purified paired helical filaments in sodium dodecyl sulfate and urea. Science 215:1243-1245; 1982. Shaw, G.; Chau, V. Ubiquitin and microtubule-associated protein tau immunoreactivity each define distinct structure with differing distributions and solubility properties in Alzheimer brain. Proc. Natl. Acad. Sci. USA 85:2854-2858; 1988. Tabaton, M.; Whitehouse, P. J.; Perry, G.; Davies, P.; AutilioGambetti, L.; Gambetti, P. Alz 50 recognizes abnormal filaments in Alzheimer's disease and progressive supranuclear palsy. Ann. Neurol. 24:407-413; 1988. Terry, R. D.; Gonatas, N. F.; Weiss, M. Ultrastruetural studies in Alzheimer's presenile dementia. Am. J. Pathol. 44:269-297; 1964. Vallee, R. B.; Wall, J. S.; Paschal, M. B.; Shpetner, H. S. Microtubule-associated protein 1C from brain is a two-headed cytosolic dynein. Nature 332:561-563; 1988. Wall, J. S.; Hainfeld, J. F. Mass mapping with the scanning transmission electron microscope. Ann. Rev. Biophys. Biophys. Chem. 15:355-376; 1986. Wiche, G.; Oberkanins, C.; Himmler, A. Molecular structure and function of microtubule-associated proteins. Int. Rev. Cytol. 124: 217-273; 1991. Wille, H.; Drewes, G.; Biernat, J.; Mandelkow, E.-M.; Mandelkow,

19

40.

41.

42.

43.

44.

45.

46.

E. Alzheimer-like paired helical filaments and antiparallel dimers formed from microtubule-associated protein tau in vitro. J. Cell Biol. 118:573-584; 1992. Wischik, C. M.; Crowther, R. A.; Stewart, M.; Roth, M. Subunit structure of paired helical filaments in Alzheimer's disease. J. Cell Biol. 100:1905-1912; 1985. Wischik, C. M.; Novak, M.; Thogersen, H. C.; Edwards, P. C.; Runswick, M. J.; Jakes, R.; Walker, J. E.; Milstein, C.,; Roth, M.; Klug, A. Isolation of a fragment of tau derived from the core of the paired helical filaments of Alzheimer disease. Proc. Natl. Acad. Sci. USA 85:4506~.510; 1988. Wischik, C. M.; Novak, M.; Edwards, P. C.; Klug, A.; Tichelaar, W.; Crowhter, R. A. Structural characterization of the core of the paired helical filament of Alzheimer disease. Proc. Natl. Acad. Sci. USA 85:4884-4888; 1988. Wisniewski, H. M.; Merz, P. A.; Iqbal, K. Ultrastructure of paired helical filaments of Alzheimer's neurofibrillary tangle, J. Neuropathol. Exp. Neurol. 43:643-656; 1984. Wolozin, B. L.; Pruchnicki, A.; Dickson, D. W.; Davies, P. A neuronal antigen in the brains of Alzheimer patients. Science 232:648650; 1986. Yen, S.-H.; Kress, Y, The effect of chemical reagents or proteases on the ultrastructure of paired helical filaments. In: Katzman R., ed. Banbury report, vol. 15, Cold Spring Harbor Laboratory: 155-165; 1983. Zar, J. H. Biostatistical analysis, 2rid ed. Englewood Cliffs, NJ: Prentice Hall; 1984.