Advertisement

Identification of a Common Protein Association Region in the Neuronal Cdk5 Activator*

Open AccessPublished:October 13, 2000DOI:https://doi.org/10.1074/jbc.M004358200
      Cyclin-dependent protein kinase 5 (Cdk5) depends on the association with neuronal Cdk5 activator (Nck5a) for kinase activity. A variety of cellular proteins have been shown to undergo high affinity association with Nck5a, including three novel proteins, C42, C48, and C53 found by a yeast two-hybrid screen (Ching, Y. P., Qi, Z., and Wang, J. H. (2000) Gene 242, 285–294). The three proteins show competitive binding to Nck5a suggesting that they bind at a common site. The binding site has been mapped to a region of 26 amino acid residues (residues 145 to 170) at the N-terminal boundary of the kinase activation domain of Nck5a. This region of Nck5a contains an amphipathic α-helix whose hydrophobic face is involved in Cdk5 activation (Chin, K. T., Ohki, S, Tang, D., Cheng, H. C., Wang, J. H., and Zhang, M. (1999)J. Biol. Chem. 274, 7120–7127). Several lines of evidence suggest that Nck5a interacts with the binding proteins at the hydrophilic face of the amphipathic α-helix. First, the Nck5a-(145–170) peptide can bind Cdk5 and Nck5a-binding proteins simultaneously. Second, the association of Nck5a-(145–170) to C48 can be markedly reduced by high ionic strength whereas the interaction between Nck5a and Cdk5 is not affected. Third, substitution of Glu157 by glutamine in Nck5a-(145–170) abolishes the peptide's ability to bind to the three Nck5a-binding proteins without diminishing its Cdk5 binding activity.
      Cdk
      cyclin-dependent kinase
      GSH
      glutathione
      GST
      glutathioneS-transferase
      Nck5a
      neuronal Cdk5 activator
      Nck5ai
      neuronal Cdk5 activator isoform
      PCR
      polymerase chain reaction
      Cyclin-dependent protein kinase 5, Cdk5,1 was originally discovered by virtue of its sequence homology to cell cycle regulatory Cdks (
      • Lew J.
      • Beaudette K.
      • Litwin C.M.
      • Wang J.H.
      ,
      • Meyerson M.
      • Enders G.H.
      • Wu C.L.
      • Su L.K.
      • Gorka C.
      • Nelson C.
      • Harlow E.
      • Tsai L.H.
      ,
      • Hellmich M.R.
      • Pant H.C.
      • Wada E.
      • Battey J.F.
      ,
      • Xiong Y.
      • Zhang H.
      • Beach D.
      ). Like cell cycle regulatory Cdks that depend on association with a cyclin for kinase activity (
      • Morgan D.O.
      ), Cdk5 is dependent on association with a protein activator for its kinase activity. The first active form of Cdk5 was isolated from mammalian brain and shown to be a heterodimer of Cdk5 and a 25-kDa regulatory protein (
      • Lew J.
      • Huang Q.-Q.
      • Qi Z.
      • Winkfein R.J.
      • Aebersold R.
      • Hunt T.
      • Wang J.H.
      ,
      • Ishiguro K.
      • Takamatsu M.
      • Tomizawa K.
      • Omori A.
      • Takahashi M.
      • Arioka M.
      • Uchida T.
      • Imahori K.
      ,
      • Ishiguro K.
      • Kobayashi S.
      • Omori A.
      • Takamatsu M.
      • Yonekura S.
      • Anzai K.
      • Imahori K.
      • Uchia T.
      ). The regulatory subunit was subsequently shown to be a truncated form of a 35-kDa protein now known as neuronal Cdk5 activator, Nck5a (
      • Tsai L.-H.
      • Delalle I.
      • Caviness Jr., V.S.
      • Chae T.
      • Harlow E.
      ). Two forms of mammalian Cdk5 activators have been identified, the 35-kDa protein and a 39-kDa protein that is called neuronal Cdk5 activator isoform, Nck5ai (
      • Tang D.
      • Yeung J.
      • Lee K.-Y.
      • Matsushita M.
      • Matsui H.
      • Tomizawa K.
      • Hatase O.
      • Wang J.H.
      ). Among the mammalian tissues examined, Cdk5 kinase activity was readily demonstrated only in brain extracts (
      • Lew J.
      • Beaudette K.
      • Litwin C.M.
      • Wang J.H.
      ,
      • Ishiguro K.
      • Takamatsu M.
      • Tomizawa K.
      • Omori A.
      • Takahashi M.
      • Arioka M.
      • Uchida T.
      • Imahori K.
      ,
      • Hayes T.E.
      • Valtz N.L.
      • McKay R.D.
      ), and the two activators, p35 nck5a and p39 nck5ai , were detected virtually exclusively in neurons of the central nervous system (
      • Lew J.
      • Huang Q.-Q.
      • Qi Z.
      • Winkfein R.J.
      • Aebersold R.
      • Hunt T.
      • Wang J.H.
      ,
      • Tsai L.-H.
      • Delalle I.
      • Caviness Jr., V.S.
      • Chae T.
      • Harlow E.
      ,
      • Tang D.
      • Yeung J.
      • Lee K.-Y.
      • Matsushita M.
      • Matsui H.
      • Tomizawa K.
      • Hatase O.
      • Wang J.H.
      ). Cdk5 and its activators have been suggested to play important regulatory functions in mammalian brain development as well as neuronal activities in mature brains (
      • Nikolic M.
      • Dudek H.
      • Kwon Y.T.
      • Ramos Y.F.
      • Tsai L.H.
      ,
      • Paglini G.
      • Pagino G.
      • Kunda P.
      • Morfini G.
      • Maccioni R.
      • Quiroga S.
      • Ferreira A.
      • Caseres A.
      ). Evidence has been accumulating to suggest that aberrant regulation of Cdk5 may lead to cell death and/or neurodegeneration, thus contributing to various neurodegenerative diseases including Alzheimer's Disease (
      • Nikolic M.
      • Dudek H.
      • Kwon Y.T.
      • Ramos Y.F.
      • Tsai L.H.
      ,
      • Paglini G.
      • Pagino G.
      • Kunda P.
      • Morfini G.
      • Maccioni R.
      • Quiroga S.
      • Ferreira A.
      • Caseres A.
      ,
      • Kwon Y.T.
      • Tsai L.H.
      ,
      • Kwon Y.T.
      • Tsai L.H.
      • Crandall J.E.
      ,
      • Chae T.
      • Kown Y.T.
      • Broson R.
      • Dikkes P.
      • Li E.
      • Tsai L.H.
      ,
      • Ohshima T.
      • Ward J.M.
      • Huh C.G.
      • Longenecker G.
      • Veeranna
      • Pant H.C.
      • Brady R.O.
      • Martin L.J.
      • Kulkarni A.B.
      ,
      • Gilmore E.C.
      • Ohshima T.
      • Goffinet A.M.
      • Kulkarni A.B.
      • Herrup K.
      ,
      • Patrick G.N.
      • Zukerberg L.
      • Nikolic M.
      • de la Monte S.
      • Dikkes P.
      • Tsai L.H.
      ).
      Previously we reported the existence of three protein complex forms of Cdk5 in bovine brain: the monomeric Cdk5, a heterodimer of Cdk5 and p25Nck5a, and a 670-kDa macromolecular protein complex containing Cdk5 and p35Nck5a complex (
      • Lee K.Y.
      • Rosales J.L.
      • Tang D.
      • Wang J.H.
      ). The three forms of Cdk5 have different kinase activities (
      • Lee K.Y.
      • Rosales J.L.
      • Tang D.
      • Wang J.H.
      ). The monomeric Cdk5 has no endogenous kinase activity but can be activated by the addition of bacterial expressed Nck5a. While the heterodimer of Cdk5 and p25 nck5a displays high kinase activity, and the macromolecular protein complex containing Cdk5 and p35 nck5a has, surprisingly, no kinase activity, nor can it be activated by the addition of its activator (
      • Lee K.Y.
      • Rosales J.L.
      • Tang D.
      • Wang J.H.
      ). This observation suggests that the proteolytic conversion of p35 nck5a to p25 nck5a may be a mechanism of Cdk5 regulation in neurons. This suggestion is supported by recent studies showing various biological and functional differences between the two forms of Nck5a such as, intracellular localization, protein turnover rate, and the ability to induce cell apoptosis (
      • Patrick G.N.
      • Zukerberg L.
      • Nikolic M.
      • de la Monte S.
      • Dikkes P.
      • Tsai L.H.
      ).
      The fact that Cdk5-p25 nck5a exists as a heterodimer whereas Cdk5-p35 nck5a is part of a macromolecular protein complex suggests that Nck5a may show high affinity binding to specific cellular proteins (
      • Lee K.Y.
      • Rosales J.L.
      • Tang D.
      • Wang J.H.
      ). Over last few years, several laboratories have reported the identification of specific Nck5a-binding proteins, including neurofilament proteins, retinoblastoma protein, a small GTPase Rac, and β-catenin (
      • Qi Z.
      • Tang D.
      • Zhu X.
      • Fujita D.J.
      • Wang J.H.
      ,
      • Lee K.Y.
      • Helbing C.C.
      • Choi K.S.
      • Johnston R.N.
      • Wang J.H.
      ,
      • Kwon Y.T.
      • Gupta A.
      • Zhou Y.
      • Nikolic M.
      • Tsai L.H.
      ,
      • Nikolic M.
      • Chou M.M.
      • Lu W.
      • Mayer B.J.
      • Tsai L.H.
      ). We have used a yeast two-hybrid system to screen for Nck5a-binding proteins, resulting in the identification in a human brain library of 7 Nck5a-binding proteins, including three novel proteins. Full-length clones of these novel Nck5a-binding proteins, called C42, C48, and C53, have subsequently been isolated from a rat brain cDNA library (
      • Ching Y.P.
      • Qi Z.
      • Wang J.H.
      ).
      Although Nck5a and Nck5ai activate a cyclin-dependent protein kinase, they do not appear to contain a cyclin box, a conserved region of the protein's primary structure characteristic of members of the cyclin protein family (
      • Tang D.
      • Chun A.C.S.
      • Zhang M.
      • Wang J.H.
      ,
      • Chou K.C.
      • Watenpaugh K.D.
      • Heinrikson R.L.
      ,
      • Nugent J.H.
      • Alfa C.E.
      • Young T.
      • Hyams J.S.
      ). On the other hand, D-type and E-type cyclins have been shown to bind to Cdk5, but the heterodimeric proteins have no observable kinase activity (
      • Xiong Y.
      • Zhang H.
      • Beach D.
      ,
      • Miyajima M.
      • Nornes H.O.
      • Neuman T.
      ). Thus, it has been suggested that the unique function and regulatory properties of Cdk5 arise, to a large extent, from the structure of Nck5a and Nck5ai. The present study is concerned mainly with the examination of the structural basis of the interactions between Nck5a- and Nck5a-binding proteins. We have found that the three novel Nck5a-binding proteins either share a common binding site or have overlapping binding sites. Using C48 as the model binding protein, a region of Nck5a spanning 26 amino acid residues, has been found to be the minimal sequence required for C48 binding. This region of Nck5a is proximal to the N-terminal boundary of the kinase activation domain (
      • Tang D.
      • Chun A.C.S.
      • Zhang M.
      • Wang J.H.
      ). We have also characterized the interaction between Nck5a and C48 in detail.

      DISCUSSION

      During the last few years a variety of cellular proteins have been found that undergo specific and high affinity association with Nck5a (
      • Qi Z.
      • Tang D.
      • Zhu X.
      • Fujita D.J.
      • Wang J.H.
      ,
      • Lee K.Y.
      • Helbing C.C.
      • Choi K.S.
      • Johnston R.N.
      • Wang J.H.
      ,
      • Kwon Y.T.
      • Gupta A.
      • Zhou Y.
      • Nikolic M.
      • Tsai L.H.
      ,
      • Nikolic M.
      • Chou M.M.
      • Lu W.
      • Mayer B.J.
      • Tsai L.H.
      ,
      • Ching Y.P.
      • Qi Z.
      • Wang J.H.
      ). Characterizations of some of these protein-protein interactions have shed light on the function, regulation, and the mechanism of action of Nck5a (
      • Nikolic M.
      • Chou M.M.
      • Lu W.
      • Mayer B.J.
      • Tsai L.H.
      ,
      • Moorthamer M.
      • Zumstein-Mecker S.
      • Chaudhuri B.
      ). However, to date there is no study on the biochemical mechanisms of interactions between Nck5a and its binding proteins. The present study is concerned mainly with the molecular basis of the interactions between Nck5a and three novel Nck5a-binding proteins, designated C42, C48, and C53 (
      • Ching Y.P.
      • Qi Z.
      • Wang J.H.
      ). Upon analyses of the binding characteristics of a large number of Nck5a deletion mutants, a region of 26 amino acid residues, proximal to the N-terminal boundary of the kinase activation domain of Nck5a, was identified as containing the binding site(s) of these Nck5a-binding proteins. The peptide corresponding to this region of Nck5a (designated as Nck5a-(145–170)) can associate with these Nck5a-binding proteins with affinities in the same order of magnitude as those of the full-length Nck5a. Interestingly, this region of Nck5a has also been suggested to contain structural elements essential for the kinase activation (
      • Tang D.
      • Chun A.C.S.
      • Zhang M.
      • Wang J.H.
      ,
      • Chou K.C.
      • Watenpaugh K.D.
      • Heinrikson R.L.
      ). This suggestion has been substantiated in the present study by the observation that Nck5a-(145–170) binds Cdk5 with high affinity.
      Due to its small size and relative ease of purification of the recombinant protein, C48 was used initially as the model Nck5a-binding protein in these studies. When full-length C48 was used to examine the binding properties of the various Nck5a deletion mutants, the results suggested the existence of two binding sites, one within the 26-residue region (a-site), and another in the region of residues 214–240 (b-site). Subsequently, the Nck5a-binding site was mapped to within the first α-helix of C48 and a C48 deletion mutant corresponding to this α-helix region, C48α1, could bind only to thea-binding site. This observation suggests that only thea-binding site can bind C48 in intact Nck5a. This suggestion was further supported by the observation that the binding of C53 and C42 to p25 nck5a can be effectively be blocked by both C48 and C48α1. It is a common practice to use deletion protein mutants to map specific binding sites on a protein molecule. The observation that binding site b demonstrated on deletion mutants may not function in the intact protein suggests caution against such artifacts.
      The three Nck5a-binding proteins (C42, C48, and C53) used in this study to characterize interactions between Nck5a and its binding proteins show common binding characteristics. They all display high affinity and specific binding to the peptide Nck5a-(145–170) and they compete with each other for bindings to the peptide or intact Nck5a, suggesting that they bind to Nck5a at a common site. The suggestion is further supported by the observation that the interactions of Nck5a with C42, C48, and C53 are similarly affected by the ionic strength and detergent contents of the reaction medium, as well as by site-directed mutation of Nck5a (see below). However, the possibility that C42, C48, and C53 have distinctive but overlapping sites in this region cannot be completely excluded.
      In an earlier study (
      • Chin K.T.
      • Ohki S.
      • Tang D.
      • Cheng H.C.
      • Wang J.H.
      • Zhang M.
      ), we showed that a 29-residue peptide derived from Nck5a displayed potent inhibitory activity toward Cdk5 and Cdk2. This Nck5a-derived inhibitor that spans residues Gln145 to Asp173 encompasses the amino acid residues of the peptide Nck5a-(145–170). Secondary structure prediction and analysis of the Cdk inhibitory peptide by circular dichroism and two-dimensional1H NMR spectroscopy have identified an amphipathic α-helix that spans amino acid residues Ser149 to Arg162 (
      • Chin K.T.
      • Ohki S.
      • Tang D.
      • Cheng H.C.
      • Wang J.H.
      • Zhang M.
      ). A number of amino acid side chains at the hydrophobic face of this amphipathic α-helix have been suggested to play important roles in the interaction between Nck5a and Cdk5 (
      • Tang D.
      • Chun A.C.S.
      • Zhang M.
      • Wang J.H.
      ,
      • Chou K.C.
      • Watenpaugh K.D.
      • Heinrikson R.L.
      ). Our observation that 1% Triton X-100 completely abolishes the binding of the peptide Nck5a-(145–170) to Cdk5 is consistent with this suggestion (Fig. 5 B). While the interaction between Cdk5 and Nck5a is dominated by hydrophobic interactions, several lines of evidence suggests that the hydrophilic face of the amphipathic α-helix plays a major role in the association of Nck5a with the Nck5a-binding proteins. First, Cdk5 and the Nck5a-binding proteins do not compete in their interactions with Nck5a-(145–170), and Nck5a-binding proteins can bind to both monomeric Nck5a and Nck5a in the heterodimer Cdk5/Nck5a (
      • Ching Y.P.
      • Qi Z.
      • Wang J.H.
      ). These observations indicate that Nck5a uses distinct binding sites to interact with Nck5a-binding proteins and with Cdk5. Second, Nck5a-(145–170) loses its ability to associate with the Nck5a-binding proteins in high concentrations of NaCl, whereas the interaction between the peptide and Cdk5 is not adversely affected. On the other hand, while the interactions of the peptide with the Nck5a-binding proteins are totally refractory to the presence of 1% Triton X-100, the peptide-Cdk5 interaction is negated in the presence of the nonionic detergent. These results suggest that, in contrast to the Nck5a-Cdk5 interaction that depends on hydrophobic interactions, the association of Nck5a-binding proteins with Nck5a-(145–170) is dependent on electrostatic interactions. Lastly, substitution of a glutamate, Glu157, that is situated at the center of the hydrophilic face of the α-helix, by glutamine results in almost complete elimination of the interaction of Nck5a-(145–170) peptide with the Nck5a-binding proteins, without interfering with the binding of the peptide to Cdk5. On the other hand, substitution of two leucine residues, Leu151 and Leu152, in the hydrophobic face of the amphipathic α-helix by asparagines, previously shown to significantly decrease the interaction between Nck5a and Cdk5, has little effect on the association of Nck5a with the binding proteins.
      As the three Nck5a-binding proteins appear to bind at a common site on Nck5a, structural comparison of the three proteins may be expected to reveal a structural motif that is specific for the Nck5a-binding site. However, amino acid sequence alignments have failed to reveal such a structural motif. Perhaps the binding motif depends on structural features other than those in the primary structure. The smallest C48 fragment displaying high affinity binding to Nck5a, C48α1, contains an amphipathic α-helix. It is possible that the binding of C48 to Nck5a involves the hydrophilic face of this protein. Work is in progress to isolate the smallest Nck5a-binding fragments of the three binding proteins so as to define the binding motif.
      Cyclin-dependent kinase 5 has many distinct functional and regulatory properties among members of Cdk family. It has been suggested that many of the distinct properties of Cdk5 arise from the unique structure of Nck5a. While all the other known activators of Cdks belong to the cyclin protein family, Nck5a does not contain in its structure a conserved cyclin-box characteristic of cyclins. Structure and function analysis of Nck5a, however, has localized the kinase activation domain of Nck5a to a region of 142 residues, Glu150 to Asn291 (which is similar in size to cyclin fold of other cyclins) and this region of Nck5a appears to assume a cyclin-box structure (
      • Chou K.C.
      • Watenpaugh K.D.
      • Heinrikson R.L.
      ). These results suggest that the unique structure of Nck5a has evolved to support a number of other functions in addition to the kinase activation. Presumably, some of these functions are manifested in the specific interactions of Nck5a with the various cellular proteins to which it binds. The identification of a specific protein-binding site in Nck5a represents the first attempt in the elucidation of the structural basis of interactions between Nck5a and Nck5a-binding proteins. The binding site is within a subdomain of Nck5a that is important for Cdk5 activation. An amphipathic α-helix in this domain uses its hydrophobic and hydrophilic phases to interact with Cdk5 and the Nck5a-binding proteins, respectively. It should be reiterated that although all three binding proteins studied in the present work appear to bind to this site, it does not indicate that this is a common binding site for all Nck5a-binding proteins as preliminary results have shown that certain other Nck5a-binding proteins bind to Nck5a at distinct sites.
      Z. Qi, unpublished results.

      Acknowledgement

      We thank Dr. David Banfield for careful reading and critical comments of the manuscript.

      REFERENCES

        • Lew J.
        • Beaudette K.
        • Litwin C.M.
        • Wang J.H.
        J. Biol. Chem. 1992; 267: 13383-13390
        • Meyerson M.
        • Enders G.H.
        • Wu C.L.
        • Su L.K.
        • Gorka C.
        • Nelson C.
        • Harlow E.
        • Tsai L.H.
        EMBO J. 1992; 11: 2909-2917
        • Hellmich M.R.
        • Pant H.C.
        • Wada E.
        • Battey J.F.
        Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 10867-10871
        • Xiong Y.
        • Zhang H.
        • Beach D.
        Cell. 1992; 71: 505-514
        • Morgan D.O.
        Nature. 1995; 374: 131-134
        • Lew J.
        • Huang Q.-Q.
        • Qi Z.
        • Winkfein R.J.
        • Aebersold R.
        • Hunt T.
        • Wang J.H.
        Nature. 1994; 371: 423-426
        • Ishiguro K.
        • Takamatsu M.
        • Tomizawa K.
        • Omori A.
        • Takahashi M.
        • Arioka M.
        • Uchida T.
        • Imahori K.
        J. Biol. Chem. 1992; 267: 10897-10901
        • Ishiguro K.
        • Kobayashi S.
        • Omori A.
        • Takamatsu M.
        • Yonekura S.
        • Anzai K.
        • Imahori K.
        • Uchia T.
        FEBS Lett. 1994; 342: 203-208
        • Tsai L.-H.
        • Delalle I.
        • Caviness Jr., V.S.
        • Chae T.
        • Harlow E.
        Nature. 1994; 371: 419-423
        • Tang D.
        • Yeung J.
        • Lee K.-Y.
        • Matsushita M.
        • Matsui H.
        • Tomizawa K.
        • Hatase O.
        • Wang J.H.
        J. Biol. Chem. 1995; 270: 26897-26903
        • Hayes T.E.
        • Valtz N.L.
        • McKay R.D.
        New Biol. 1991; 3: 259-269
        • Nikolic M.
        • Dudek H.
        • Kwon Y.T.
        • Ramos Y.F.
        • Tsai L.H.
        Genes Dev. 1996; 10: 816-825
        • Paglini G.
        • Pagino G.
        • Kunda P.
        • Morfini G.
        • Maccioni R.
        • Quiroga S.
        • Ferreira A.
        • Caseres A.
        J. Neurosci. 1998; 18: 9858-9869
        • Kwon Y.T.
        • Tsai L.H.
        J. Comp. Neurol. 1998; 395: 510-522
        • Kwon Y.T.
        • Tsai L.H.
        • Crandall J.E.
        J. Comp. Neurol. 1999; 415: 218-229
        • Chae T.
        • Kown Y.T.
        • Broson R.
        • Dikkes P.
        • Li E.
        • Tsai L.H.
        Neuron. 1997; 18: 29-42
        • Ohshima T.
        • Ward J.M.
        • Huh C.G.
        • Longenecker G.
        • Veeranna
        • Pant H.C.
        • Brady R.O.
        • Martin L.J.
        • Kulkarni A.B.
        Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 11173-11178
        • Gilmore E.C.
        • Ohshima T.
        • Goffinet A.M.
        • Kulkarni A.B.
        • Herrup K.
        J. Neurosci. 1998; 18: 6370-6377
        • Patrick G.N.
        • Zukerberg L.
        • Nikolic M.
        • de la Monte S.
        • Dikkes P.
        • Tsai L.H.
        Nature. 1999; 402: 615-622
        • Lee K.Y.
        • Rosales J.L.
        • Tang D.
        • Wang J.H.
        J. Biol. Chem. 1996; 271: 1538-1543
        • Qi Z.
        • Tang D.
        • Zhu X.
        • Fujita D.J.
        • Wang J.H.
        J. Biol. Chem. 1998; 273: 2329-2335
        • Lee K.Y.
        • Helbing C.C.
        • Choi K.S.
        • Johnston R.N.
        • Wang J.H.
        J. Biol. Chem. 1997; 272: 5622-5626
        • Kwon Y.T.
        • Gupta A.
        • Zhou Y.
        • Nikolic M.
        • Tsai L.H.
        Curr. Biol. 2000; 10: 363-372
        • Nikolic M.
        • Chou M.M.
        • Lu W.
        • Mayer B.J.
        • Tsai L.H.
        Nature. 1998; 395: 194-198
        • Ching Y.P.
        • Qi Z.
        • Wang J.H.
        Gene. 2000; 242: 285-294
        • Tang D.
        • Chun A.C.S.
        • Zhang M.
        • Wang J.H.
        J. Biol. Chem. 1997; 272: 12318-12327
        • Chou K.C.
        • Watenpaugh K.D.
        • Heinrikson R.L.
        Biochem. Biophys. Res. Commun. 1999; 259: 420-428
        • Nugent J.H.
        • Alfa C.E.
        • Young T.
        • Hyams J.S.
        J. Cell Sci. 1991; 99: 669-674
        • Miyajima M.
        • Nornes H.O.
        • Neuman T.
        Neuroreport. 1995; 6: 1130-1132
        • Chin K.T.
        • Ohki S.
        • Tang D.
        • Cheng H.C.
        • Wang J.H.
        • Zhang M.
        J. Biol. Chem. 1999; 274: 7120-7127
        • Moorthamer M.
        • Zumstein-Mecker S.
        • Chaudhuri B.
        FEBS Lett. 1999; 446: 343-350
        • Rost B.
        • Sander C.
        J. Mol. Biol. 1993; 232: 584-599