Effects of disulfide bonds formed during isolation process on the structure of the postsynaptic density

Brain Research 873 (2000) 268–273 www.elsevier.com / locate / bres

Research report

Effects of disulfide bonds formed during isolation process on the structure of the postsynaptic density Chia-Wen Sui, Wei-Yuan Chow, Yen-Chung Chang* Department of Life Science, National Tsing Hua University, Hsinchu 30043, Taiwan, ROC Accepted 30 May 2000

Abstract The biochemical, morphological and structural properties of rat postsynaptic densities (PSDs) isolated under conditions where disulfide bond formation was allowed or curtailed were studied here. Biochemical analyses revealed that the isolated PSDs were composed by a similar set of proteins regardless of the differences in their isolation processes. The PSDs isolated under the conditions where disulfide bond formation was curtailed were more easily dissociated by treatments with urea, guanidine hydrochloride and deoxycholate than the PSDs isolated under conditions where disulfide bond formation was allowed. Consistently, the structure of the PSDs isolated under the former condition appeared to be more fragmented than those isolated under the latter condition, as revealed by electron microscopy. The results indicate that the disulfide bonds formed during the isolation process significantly tighten the PSD structure and further suggest that the PSD in vivo is a protein aggregate whose constituent proteins be held together primarily by non-covalent interactions.  2000 Elsevier Science B.V. All rights reserved. Theme: Excitable membranes and synaptic transmission Topic: Postsynaptic mechanisms Keywords: Postsynaptic density; Disulfide bond; Structure

1. Introduction The postsynaptic density (PSD) is a layer of densely packed protean material subjacent to the postsynaptic membrane of asymmetric synapses in the mammalian CNS [5,8]. By three-dimensional electron microscopy reconstruction, disk-shaped, segmented and perforated PSDs have been reported [9]. PSDs could be isolated from brain tissues by biochemical methods, and the isolated PSDs consisted of more than 30 different proteins [4,6,12]. The structure of the isolated PSDs was found exceptionally tight; so that the dissociation of isolated PSDs by various chaotropic reagents or detergents (except for strong de-

Abbreviations: PSD, postsynaptic density; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; PVDF, polyvinylidene difluoride; DOC, deoxycholate; PMSF, phenylmethanesulfonyl fluoride; DTE, 1,49-dithioerythritol *Corresponding author. Fax: 1886-3-571-5934. E-mail address: [email protected] (Y.-C. Chang).

tergents like SDS) into subunits or individual proteins was very difficult [1,2]. As a result, our understanding of the molecular compositions of various structural components of the PSD, e.g., the filamentous and granular components observed by electron microscopy [19], is limited. Sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) analyses performed under non-reducing conditions revealed that the isolated PSDs contained extensive inter-molecular disulfide bonds [11,20]. The tight structure of the isolated PSDs has thus been proposed to, at least partly, result from disulfide bonds that cross-link various PSD proteins together. Most recently, analyses of the PSD isolated from pig brains have led us to conclude that the disulfide bonds formed during the isolation process tighten the structure of the PSDs [15]. It is, however, uncertain if this conclusion is also true for the PSDs of animals other than pig. Furthermore, it was reported that the freshness of brain tissues influenced the protein composition of the isolated PSD [3]. It is hence uncertain if the conclusion drawn from studies of pig PSDs, which are

0006-8993 / 00 / $ – see front matter  2000 Elsevier Science B.V. All rights reserved. PII: S0006-8993( 00 )02544-0

C.-W. Sui et al. / Brain Research 873 (2000) 268 – 273

prepared from |3-h-old brain tissues, is also true for the PSDs isolated from freshly dissected brains. To answer these questions, we have isolated the PSDs from freshly dissected rat brains by the procedure of Lai et al. [14] and by the same procedure except for adding iodoacetamide or 1,49-dithiothreitol (DTE) to all buffer solutions to curtail disulfide bond formation during the isolation process. The protein composition, the strength of protein organization, and the morphology of the PSDs isolated under these conditions are analyzed in this study.

2. Materials and methods Iodoacetamide, phenylmethanesulfonyl fluoride (PMSF), 1,49-dithioerythritol (DTE), sodium deoxycholate (Na-DOC), urea, EGTA, EDTA, Triton X-100, pepstatin A, leupeptin, benzamidine and HEPES were obtained from Sigma. PVDF (polyvinyldiene difluoride) membrane was purchased from Millipore. ProtoGold protein detection kit was purchase from Bio-Cell. Antibodies to actin, b-tubulin subunit, a-subunit of calcium, calmodulin-dependent protein kinase II (CaMK II), rat GluR1 subunit, and rat GluR2 / 3 subunits were purchased from Chemicon (USA). Other chemicals were obtained from Merck-Suchardt. Brains of lightly ether-anesthetized adult Sprague–Dawley rats (2–3 months old) were dissected, frozen rapidly and kept in liquid nitrogen until use. PSDs were isolated from rat brains (|50 g) by the procedure described in Lai et al. [14] and by the same procedure except for adding 2 mM iodoacetamide or 5 mM DTE in all buffer solutions, and these PSD samples were named as the PSD, carbamoyl methylated (CM)-PSD and DTE-PSD samples, respectively. All buffer solutions used for the isolation of PSD, DTE-PSD and CM-PSD samples also contained EGTA (0.1 mM), EDTA (0.1 mM), PMSF (0.25 mM), leupeptin (1 mg / ml), benzamidine (1 mM) and pepstatin A (1 mg / ml).

2.1. Electrophoresis and immunoblotting Electrophoresis was carried out with a mini-gel apparatus (Protean II, Bio-Rad). SDS–PAGE analysis was performed by the method of Laemmli [13]. After electrophoresis, the proteins in the gel were transferred to a PVDF membrane and detected by ProtoGold (BioCell). Immunoblotting analysis was done following the procedure of Wu et al. [24]. Protein concentrations were determined by the bicinchoninic acid method [22].

2.2. Sucrose density gradient centrifugation Sucrose density gradient centrifugation was performed with a SW-41 rotor (Beckman). Sucrose gradients between 10% and 70% sucrose were prepared by making step gradients consisting of five layers of 70%, 55%, 40%, 25%

269

and 10% sucrose in 10 mM HEPES at pH 7.4, from bottom to top, and the resultant gradients were kept at 58C for 24 h before use. Samples were then applied on top of the gradients and centrifuged at 100 0003g and 58C for 1 h. After centrifugation, samples were removed from the bottom of gradients in fractions of 1 ml. Sucrose concentrations were determined by a refractometer (Atago Co., Japan).

2.3. Electron microscopy The PSD samples (containing approx. 0.6 mg / ml protein) were applied to the glow-discharged carbon-coated collodion film on nickel grids (200-mesh) as 2-ml drops. After 10 min, the liquid was drained off by contact with filter paper (Whatman [1). The samples were incubated with 2% (w / v) phosphotungstic acid at pH 7.4 for 30 s and air-dried. Electron microscopy was performed on a Hitachi (H-600-3) transmission electron microscope.

3. Results Two CM-PSD samples, two DTE-PSD samples and a PSD sample were prepared from rat brains with the yields of 16.0 and 15.4, 15.3 and 18.7, and 22 mg protein per 100 g (wet weight) brain tissues, respectively. SDS–PAGE analyses of these samples indicated that each of these samples consisted of three major protein bands of |46, |52 and |56 kDa (as indicated by A, C and T, respectively in Fig. 1A). Western blotting revealed that the major proteins of |46, |52 and |56 kDa of these samples were recognized by antibodies to actin, a-subunit of calcium / calmodulin-dependent protein kinase II (CaMK II) and b-tubulin (Fig. 1B). The percentages of these major proteins relative to the total proteins of these three PSD samples were calculated from densitometric scans of the SDS–PAGE results (Table 1). The SDS–PAGE and immunoblotting results indicated regardless of how the PSDs were prepared they all consisted of a similar set of major proteins although differences in the intensities of some minor proteins were also found. The isolated PSDs also contained proteins recognized by antibodies to AMPA receptor subunits. Anti-GluR1 and 2 / 3 antibodies were found to recognize proteins with sizes corresponding to the reported sizes of these subunits, |108 kDa (Fig. 1B). Sucrose density sedimentation analyses of PSD, CMPSD and DTE-PSD samples indicated that they consisted primarily of large protein aggregates that migrated rapidly to a position near the bottom of gradients where the sucrose density matched that of proteins. To examine the structural strengths of the large aggregates in PSD, CMPSD and DTE-PSD samples, these samples were treated with 1 M glycine / phosphoric acid (at pH 2.0), 6 M urea (in 10 mM Hepes at pH 7.4), 4 M guanidine hydrochloride (in 10 mM Hepes at pH 7.4) or Na-DOC (3% (w / v), |pH

C.-W. Sui et al. / Brain Research 873 (2000) 268 – 273

270

Fig. 1. SDS–PAGE and immunoblotting analyses of the PSD, DTE-PSD and CM-PSD samples. (A) Two micrograms of the PSD, DTE-PSD and CM-PSD samples were subjected SDS–PAGE analysis. The proteins on the SDS gels were transferred to a PVDF membrane and stained by the ProtoGold method. Protein bands of tubulin, a-subunit of CaMKII and actin were labeled by T, C and A, respectively. (B) One microgram of the PSD, DTE-PSD and CM-PSD proteins were separated by 9% SDS–PAGE, transferred to a PVDF membrane, and immunostained by antibodies to rat GluR1, rat GluR 2 / 3, b-tubulin subunit, a-subunit of CaMK II and actin.

10) on ice for 1 h. The samples were then subjected to sucrose density sedimentation analysis to separate the dissociated smaller components (open arrows) from the large aggregates (closed arrows) following these treatments

Table 1 Percentages of the major proteins of the PSD, DTE-PSD and CM-PSD samples a Protein b

PSD

DTE-PSD

CM-PSD

Tubulin (Mr |56 000)c CaMKII (Mr |52 000) Actin (Mr |46 000)

9.361.0 8.161.2 5.360.4

9.461.8 7.661.8 5.561.1

11.662.5 10.962.6 4.260.8

b

Immunostained by antibodies to b-tubulin, a-subunit of CaMKII or actin. c Apparent relative molecular weights estimated from the gel. a Data were mean6S.D. of the intensities of protein bands from densitometric scans of 8, 5 and 5 ProtoGold-stained polyacrylamide gels of the PSD, DTE-PSD and CM-PSD samples, respectively.

(Fig. 2). A comparison of the protein distributions between the dissociated smaller components and large aggregates (percentages shown in Fig. 2) indicated that glycine / phosphoric acid treatment did not significantly dissociate the large aggregates of the PSD or CM-PSD samples. The glycine / phosphoric acid treatment of DTE-PSD however produced a peak, containing 35% of the total protein, in the middle of the gradient. Urea, guanidine hydrochloride or Na-DOC treatments resulted in the dissociation of more than 60% of large aggregates of the CM-PSD and DTEPSD samples into smaller ones. On the other hand, the majority of the proteins of the PSD sample remained as large aggregates following these treatments. These results indicated that the large protein aggregates of the CM-PSD and DTE-PSD samples were more susceptible to the dissociation effects of urea, guanidine hydrochloride and Na-DOC than those of the PSD sample. The morphology of PSDs, CM-PSDs and DTE-PSDs was examined by negative-staining electron microscopy

C.-W. Sui et al. / Brain Research 873 (2000) 268 – 273

271

Fig. 2. Fragmentation of PSDs, DTE-PSDs and CM-PSDs by different treatments. The PSD, DTE-PSD and CM-PSD samples (500 mg protein each) were incubated with either 10 mM Hepes at pH 7.4 (control), 1 M glycine / H 3 PO 4 at pH 2.0, 3% (w / v), Na-DOC (at |pH 10), 6 M urea (in 10 mM Hepes at pH 7.4) or 4 M guanidine hydrochloride (in 10 mM Hepes at pH 7.4) in a final volume of 500 ml on ice for 1 h and then subjected sucrose density gradient (10–70%) sedimentation analysis. Fractions of 1 ml were collected from the bottom of the gradients. Proteins in the resultant fractions were determined by the bicinchoninic acid method [17] except for the experiment of glycine / phosphoric acid treatment. Aliquots of 10 ml were removed from each fraction collected from the glycine / phosphoric acid treatment experiments and subjected to SDS–PAGE analysis. The proteins in the SDS-gel were then transferred to a PVDF membrane, and stained by ProtoGold. Relative protein concentrations of different fractions were determined by the densitometric scans of the resultant membrane, and the protein concentration of the peak fraction was set as 100. Percentages were the sums of the protein concentrations of the fractions under different peaks relative to the sum of the protein concentrations of all fractions. (j) protein concentration; (h) sucrose concentration.

(Fig. 3). It was found that all three samples consisted of objects with outside diameters in the range between 100 and 1000 nm. The objects found in the PSD sample (Fig. 3A,B) resembled closely the PSDs that were isolated from pig and rat brains by similar procedures and studied by the same technique [15,17]. The constituent proteins of the PSDs appeared to associate to each other very closely into dense aggregates, and no subunit with a regular shape was readily discerned. In comparison to the compact structure of the PSDs, the objects found in the DTE-PSD (Fig. 3C–E) and CM-PSD samples (Fig. 3F–G) appeared to be more fragmented, and holes of varied size were found in the CM-PSDs and DTE-PSDs. The DTE-PSD sample contained more small objects with outside diameters around 100–300 nm, as shown in Fig. 3C and D, than the remaining two samples. This observation is consistent with our sedimentation analyses that the DTE-PSD sample contained more small components than the remaining two

samples (Fig. 2). The morphology of CM-PSDs and DTEPSDs resembled closely that of the pig PSDs isolated in buffer solutions containing iodoacetic acid [15].

4. Discussion Iodoacetamide is known to efficiently modify free thiol groups of proteins and hence prevent random formation of disulfide bonds [18], and DTE can efficiently break those disulfide bonds that are exposed to the solvent. In this study, we found that the PSDs isolated in DTE- or iodoacetamide-containing solutions were more susceptible to the dissociation effects of urea, guanidine hydrochloride and Na-DOC than the PSDs isolated in solutions lacking DTE or iodoacetamide. Electron microscopic studies also indicated that the structure of the former PSDs were more fragmented than that of the latter ones. Parallel to our

272

C.-W. Sui et al. / Brain Research 873 (2000) 268 – 273

Fig. 3. Morphology of PSDs (A,B), DTE-PSDs (C–E) and CM-PSDs (E,F). Scale bar5100 nm.

previous study of the pig PSDs [15], these observations suggest that the structure of the PSD isolated from freshly dissected rat brains could also be tightened by disulfide bonds formed during the isolation process. It appears that the various constituent proteins of the PSD, regardless of the sources and freshness of the brain from which the PSD is prepared, tend to form extensive disulfide bonds if the environment allows. Because iodoacetamide and DTE curtail the random formation of disulfide bonds, it is deduced that the disulfide bonds of the PSD isolated in solution containing either one of these reagents are in a state similar to that the original PSD under in vivo conditions. Our sucrose density sedimentation analyses indicate that the CM-PSD and DTE-PSD behave like most of the protein aggregates that are dissociated by chaotropic reagents like urea and guanidine hydrochloride and by ionic detergents like deoxycholate. As a result, it is likely that the original PSD in neurons is also an aggregate of proteins held together primarily by non-covalent protein– protein interactions. However, both electron microscopic and chemical-treatment studies indicated that the structure of DTE-PSD is more fragmented than that of CM-PSD (Figs. 2 and 3). These results imply that under in vivo conditions some disulfide bonds are also required to help maintain the PSD structure. Since the redox potential of cytoplasm greatly favors the breakdown of disulfide bonds,

these disulfide bonds may be buried in the interior of tightly packed PSDs in neurons [7]. The physical property of the PSD under in vivo conditions deduced here is consistent with the results of numerous in vivo studies suggesting that the PSD is made with plastic materials and undergoes rapid structural changes when exposed to stimuli (see for reviews Refs. [16,21]). The deduced physical property is also parallel to the observation that the constituent proteins of the PSD in vivo are in equilibrium with their counterparts in the cytosol. For example, CaMK II has been reported to translocate to the PSD region when the hippocampal slices is subjected to stimuli capable of inducing long-term potentiation of synapses [23]. Significant increases in many cytosolic proteins and intracellular signaling molecules co-purified with the PSD that is prepared from transiently ischemia-treated rats [10]. The results of this study suggest that the constituent proteins of the PSD under in vivo conditions are held together primarily by non-covalent protein–protein interactions. Furthermore, the structure of the isolated PSD is tightened if the isolation process is performed under conditions where the random formation of disulfide bonds is allowed. Including reagents in buffer solutions for blocking disulfide formation appears to be crucial for preparing PSDs, with a structure mimicking that of the PSDs in the native state, for further analyses.

C.-W. Sui et al. / Brain Research 873 (2000) 268 – 273

Acknowledgements This work was supported by Grants NSC 88-2311-B007-026 (to Y.-C.C.) and NSC 82-0203-B007-026 (to W.Y.C.) from the National Science Council of Republic of China.

References [1] G. Adam, A. Matus, Role of actin in the organization of brain postsynaptic densities, Mol. Brain Res. 43 (1996) 246–250. [2] F. Blomberg, R.S. Cohen, P. Siekevitz, The structure of postsynaptic densities isolated from dog cerebral cortex, J. Cell Biol. 74 (1977) 204–225. [3] R.D. Carlin, D.J. Grab, P. Siekevitz, Postmortem accumulation of tubulin in postsynaptic density preparations, J. Neurochem. 38 (1982) 94–100. [4] R.S. Cohen, F. Blomberg, K. Berzins, P. Siekevitz, The structure of postsynaptic densities isolated from dog cerebral cortex, 1. Overall morphology and protein composition, J. Cell Biol. 74 (1977) 181– 203. [5] M. Colonnier, Synaptic patterns on different cell types in the different laminae of the cat visual cortex: an electron microscope study, Brain Res. 9 (1968) 268–287. [6] C.W. Cotman, G. Banker, L. Churchill, D. Taylor, Isolation of postsynaptic densities from rat brain, J. Cell Biol. 63 (1974) 441– 445. [7] T.E. Creighton, Disulphide bonds between cysteine residues, in: T.E. Creighton (Ed.), Protein Structure, A Practical Approach, 1990, pp. 155–167. [8] E.G. Gray, Axo-somatic and axo-dendritic synapses of the cerebral cortex: an electron microscope study, J. Anat. 93 (1959) 420–433. [9] K.M. Harris, S.B. Kater, Dendritic spines: cellular specializations imparting both stability and flexibility to synaptic function, Annu. Rev. Neurosci. 17 (1994) 341–371. [10] B.-R. Hu, M. Park, M.E. Martone, W.H. Fischer, M.H. Ellisman, J.A. Zivin, Assembly of proteins to postsynaptic densities after transient cerebral ischemia, J. Neurosci. 18 (1998) 625–633. [11] P.T. Kelly, C.W. Cotman, Intermolecular disulfide bonds at central

[12] [13] [14]

[15]

[16]

[17]

[18] [19]

[20]

[21]

[22]

[23]

[24]

273

nervous system synaptic junctions, Biochem. Biophys. Research Commun. 73 (1976) 858–865. M.B. Kennedy, The postsynaptic density at glutamatergic synapses, Trends Neurosci. 20 (1997) 261–268. U.K. Laemmli, Cleavage of structural proteins during the assembly of the head of bacteriophage T4, Nature 227 (1970) 680–685. S.-L. Lai, S.-C. Ling, L.-H. Kuo, Y.-C. Shu, W.-Y. Chow, Y.-C. Chang, Characterization of granular particles isolated from postsynaptic densities, J. Neurochem. 71 (1998) 1694–1701. S.-L. Lai, S.-F. Chiang, I.-T. Chen, W.-Y. Chow, Y.-C. Chang, Interprotein disulfide bonds formed during isolation process tighten the structure of the postsynaptic density, J. Neurochem. 73 (1999) 2130–2138. J.E. Lisman, K.M. Harris, Quantal analysis and synaptic anatomyintegrating two views of hippocampal plasticity, Trends Neurosci. 16 (1993) 141–147. A.I. Matus, D.H. Taff-Jones, Morphology and molecular composition of isolated postsynaptic junctional structures, Proc. R. Soc. Lond. B 203 (1978) 135–151. G.E. Means, R.T. Feeney, Chemical Modification of Proteins, Holden-Day, San Francisco, 1971, pp. 218–219. A. Peters, S.L. Palay, H.D.E.F. Webster, The Fine Structure of the Nervous System, 3, Oxford University Press, New York / Oxford, 1991. N. Ratner, H.R. Mahler, Structural organization of filamentous proteins in postsynaptic density, Biochemistry 22 (1983) 2446– 2453. P. Siekevitz, The postsynaptic density: a possible role in long-lasting effects in the central nervous system, Proc. Natl. Acad. Sci. USA 82 (1985) 3494–3498. P.K. Smith, R.J. Krohn, G.T. Hermanson, A.K. Mallia, F.H. Gartner, M.D. Provenzano, E.K. Fujimoto, N.M. Gorke, B.J. Olson, D.C. Klenk, Measurement of protein using bicinchoninic acid, Anal. Biochem. 150 (1985) 76–85. S. Strack, S. Choi, D.M. Lovinger, R.J. Colbran, Translocation of autophosphorylated calcium / calmodulin-dependent protein kinase II to the postsynaptic density, J. Biol. Chem. 272 (1997) 13467– 13470. T.-Y. Wu, C.-I. Liu, Y.-C. Chang, A study of the oligomeric state of the a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid-preferring glutamate receptors in the synaptic junctions of porcine brain, Biochem. J. 319 (1996) 731–739.