The alpha- and beta-tubulin folding pathways.

The alpha-beta tubulin heterodimer is the subunit from which microtubules are assembled. The pathway leading to correctly folded alpha- and beta-tubulins is unusually complex: it involves cycles of ATP-dependent interaction of newly synthesized tubulin subunits with cytosolic chaperonin, resulting in the production of quasi-native folding intermediates, which must then be acted upon by additional protein cofactors. These cofactors form a supercomplex containing both alpha- and beta-tubulin polypeptides, from which native heterodimer is released in a GTP-dependent reaction. Here, we discuss the current state of our understanding of the function of cytosolic chaperonin and cofactors in tubulin folding.

Generation of antisera that discriminate among mammalian alpha-tubulins: introduction of specialized isotypes into cultured cells results in their coassembly without disruption of normal microtubule function.

To assay the functional significance of the multiple but closely related alpha-tubulin polypeptides that are expressed in mammalian cells, we generated three specific immune sera, each of which uniquely recognizes a distinct alpha-tubulin isotype. All three isotypes are expressed in a tissue-restricted manner: one (M alpha 3/7) only in mature testis, one (M alpha 4) mainly in muscle and brain, and the third (M alpha 6) in several tissues at a very low level. A fourth specific antiserum was also generated that distinguishes between the tyrosinated and nontyrosinated form of a single alpha-tubulin isotype. Because individual tubulin isotypes cannot be purified biochemically, these sera were raised using cloned fusion proteins purified from host Escherichia coli cells. To suppress the immune response to shared epitopes, animals were first rendered tolerant to fusion proteins encoding all but one of the known mammalian alpha-tubulin isotypes. Subsequent challenge with the remaining fusion protein then resulted in the elicitation of an immune response to unique epitopes. Three criteria were used to establish the specificity of the resulting sera: (a) their ability to discriminate among cloned fusion proteins representing all the known mammalian alpha-tubulin isotypes; (b) their ability to uniquely detect alpha-tubulin in whole extracts of tissues; and (c) their capacity to stain microtubules in fixed preparations of cells transfected with sequences encoding the corresponding isotype. The transfection experiments served to demonstrate (a) the coassembly of M alpha 3/7, M alpha 4, and M alpha 6 into both interphase and spindle microtubules in HeLa cells and NIH 3T3 cells, and (b) that the M alpha 4 isotype, which is unique among mammalian alpha-tubulins in that it lacks an encoded carboxy-terminal tyrosine residue, behaves like other alpha-tubulin isotypes with respect to the cycle of tyrosination/detyrosination that occurs in most cultured cells.

Complex regulation and functional versatility of mammalian alpha- and beta-tubulin isotypes during the differentiation of testis and muscle cells.

In the accompanying paper (Gu, W., S. A. Lewis, and N. J. Cowan. 1988. J. Cell Biol. 106: 2011-2022), we report the generation of three antisera, each of which uniquely recognizes a different mammalian alpha-tubulin isotype, plus a fourth antibody that distinguishes between microtubules containing the tyrosinated and nontyrosinated form of the only known mammalian alpha-tubulin gene product that lacks an encoded carboxy-terminal tyrosine residue. These sera, together with five sera we raised that distinguish among the known mammalian beta-tubulin isotypes, have been used to study patterns of tubulin isotype-specific expression in muscle and testis, two tissues in which characteristic developmental changes are accompanied by dramatic rearrangements in microtubule structures. As in the case of cells in culture, there is no evidence to suggest that there is subcellular sorting of different tubulin isotypes among different kinds of microtubule, even in a cell type (the developing spermatid) that simultaneously contains such functionally distinct structures as the manchette and the flagellum. On the other hand, the patterns of expression of the various tubulin isotypes show marked and distinctive differences in different cell types and, in at least one case, evidence is presented for regulation at the translational or posttranslational level. The significance of these observations is discussed in terms of the existence of the mammalian alpha- and beta-tubulin multigene families.

The microtubule binding domain of microtubule-associated protein MAP1B contains a repeated sequence motif unrelated to that of MAP2 and tau.

We report the complete sequence of the microtubule-associated protein MAP1B, deduced from a series of overlapping genomic and cDNA clones. The encoded protein has a predicted molecular mass of 255,534 D and contains two unusual sequences. The first is a highly basic region that includes multiple copies of a short motif of the form KKEE or KKEVI that are repeated, but not at exact intervals. The second is a set of 12 imperfect repeats, each of 15 amino acids and each spaced by two amino acids. Subcloned fragments spanning these two distinctive regions were expressed as labeled polypeptides by translation in a cell-free system in vitro. These polypeptides were tested for their ability to copurify with unlabeled brain microtubules through successive cycles of polymerization and depolymerization. The peptide corresponding to the region containing the KKEE and KKEVI motifs cycled with brain microtubules, whereas the peptide corresponding to the set of 12 imperfect repeats did not. To define the microtubule binding domain in vivo, full-length and deletion constructs encoding MAP1B were assembled and introduced into cultured cells by transfection. The expression of transfected polypeptides was monitored by indirect immunofluorescence using anti-MAP1B-specific antisera. These experiments showed that the basic region containing the KKEE and KKEVI motifs is responsible for the interaction between MAP1B and microtubules in vivo. This region bears no sequence relationship to the microtubule binding domains of kinesin, MAP2, or tau.

ADP ribosylation factor-like protein 2 (Arl2) regulates the interaction of tubulin-folding cofactor D with native tubulin.

The ADP ribosylation factor-like proteins (Arls) are a family of small monomeric G proteins of unknown function. Here, we show that Arl2 interacts with the tubulin-specific chaperone protein known as cofactor D. Cofactors C, D, and E assemble the alpha/beta- tubulin heterodimer and also interact with native tubulin, stimulating it to hydrolyze GTP and thus acting together as a beta-tubulin GTPase activating protein (GAP). We find that Arl2 downregulates the tubulin GAP activity of C, D, and E, and inhibits the binding of D to native tubulin in vitro. We also find that overexpression of cofactors D or E in cultured cells results in the destruction of the tubulin heterodimer and of microtubules. Arl2 specifically prevents destruction of tubulin and microtubules by cofactor D, but not by cofactor E. We generated mutant forms of Arl2 based on the known properties of classical Ras-family mutations. Experiments using these altered forms of Arl2 in vitro and in vivo demonstrate that it is GDP-bound Arl2 that interacts with cofactor D, thereby averting tubulin and microtubule destruction. These data establish a role for Arl2 in modulating the interaction of tubulin-folding cofactors with native tubulin in vivo.

Genetics, evolution, and expression of the 68,000-mol-wt neurofilament protein: isolation of a cloned cDNA probe.

A 1.2-kilobase (kb) cDNA clone (NF68) encoding the mouse 68,000-mol-wt neurofilament protein is described. The clone was isolated from a mouse brain cDNA library by low-stringency cross-hybridization with a cDNA probe encoding mouse glial fibrillary acidic protein (Lewis et al., 1984, Proc. Natl. Acad. Sci. USA., 81:2743-2746). The identity of NF68 was established by hybrid selection using mouse brain polyA+ mRNA, and cell-free translation of the selected mRNA species. The cell-free translation product co-migrated with authentic 68,000-mol-wt neurofilament protein on an SDS/polyacrylamide gel, and was immunoprecipitable with a monospecific rabbit anti-bovine neurofilament antiserum. In addition, DNA sequence analysis of NF68 showed 90% homology at the amino acid level compared with the sequence of the porcine 68,000-mol-wt neurofilament protein. At high stringency, NF68 detects a single genomic sequence encoding the mouse 68,000-mol-wt neurofilament protein. Two mRNA species of 2.5 kb and 4.0 kb are transcribed from the single gene in mouse brain. The level of expression of these mRNAs remains almost constant in postnatal mouse brains of all ages and, indeed, in the adult. At reduced stringency, NF68 detects a number of mRNAs that are expressed in mouse brain, one of which encodes the 150,000-mol-wt neurofilament protein. The NF68 probe cross-hybridizes at high stringency with genomic sequences in species as diverse as human, chicken, and (weakly) frog, but not with DNA from Drosophila or sea urchin.

A cloned cDNA encoding MAP1 detects a single copy gene in mouse and a brain-abundant RNA whose level decreases during development.

Screening of a bacteriophage lambda gt11 cDNA expression library with a polyclonal anti-microtubule associated protein (MAP) antiserum resulted in the isolation of two non-cross-hybridizing sets of cDNA clones. One set was shown to encode MAP2 (Lewis, S. A., A. Villasante, P. Sherline, and N. J. Cowan, 1986, J. Cell Biol., 102:2098-2105). To determine the specificity of the second set, three non-overlapping fragments cloned from the same mRNA molecule via a series of "walking" experiments were separately subcloned into inducible plasmid expression vectors in the appropriate orientation and reading frame. Upon induction and analysis by immunoblotting, two of the fusion proteins synthesized were shown to be immunoreactive with an anti-MAP1-specific antibody, but not with an anti-MAP2-specific antibody. Since these MAP1-specific epitopes are encoded in non-overlapping cDNAs cloned from a single contiguous mRNA, these clones cannot encode polypeptides that contain adventitiously cross-reactive epitopes. Furthermore, these cDNA clones detected an abundant mRNA species of greater than 10 kb in mouse brain, consistent with the coding requirement of a 350,000-D polypeptide and the known abundance of MAP1 in that tissue. The MAP1-specific cDNA probes were used in blot transfer experiments with RNA prepared from brain, liver, kidney, stomach, spleen, and thymus. While detectable quantities of MAP1-specific mRNA were observed in these tissues, the level of MAP1 expression was approximately 500-fold lower than in brain. The levels of both MAP1-specific and MAP2-specific mRNAs decline in the postnatal developing brain; the level of MAP1-specific mRNA also increases slightly in rat PC12 cells upon exposure to nerve growth factor. These surprising results contrast sharply with reported dramatic developmental increases in the amount of MAP1 in brain and in nerve growth factor-induced PC12 cells. The cDNA clones encoding MAP1 detect a single copy sequence in mouse DNA, even under conditions of low stringency that would allow the detection of related but mismatched sequences. The cDNAs cross-hybridize with genomic sequences in rat, human, and chicken DNA, but not with DNA from frog, Drosophila, or sea urchin. These data are discussed in terms of the evolution and possible biological role of MAP1.

The mammalian beta-tubulin repertoire: hematopoietic expression of a novel, heterologous beta-tubulin isotype.

We describe the structure of a novel and unusually heterologous beta-tubulin isotype (M beta 1) isolated from a mouse bone marrow cDNA library, and a second isotype (M beta 3) isolated from a mouse testis cDNA library. Comparison of M beta 1 and M beta 3 with the completed (M beta 4, M beta 5) or extended (M beta 2) sequence of three previously described beta-tubulin isotypes shows that each includes a distinctive carboxy-terminal region, in addition to multiple amino acid substitutions throughout the polypeptide chain. In every case where a mammalian interspecies comparison can be made, both the carboxy-terminal and internal amino acid substitutions that distinguish one isotype from another are absolutely conserved. We conclude that these characteristic differences are important in determining functional distinctions between different kinds of microtubule. The amino acid homologies between M beta 2, M beta 3, M beta 4, and M beta 5 are in the range of 95-97%; however the homology between M beta 1 and all the other isotypes is very much less (78%). The dramatic divergence in M beta 1 is due to multiple changes that occur throughout the polypeptide chain. The overall level of expression of M beta 1 is low, and is restricted to those tissues (bone marrow, spleen, developing liver and lung) that are active in hematopoiesis in the mouse. We predict that the M beta 1 isotype is functionally specialized for assembly into the mammalian marginal band.

Prefoldin-nascent chain complexes in the folding of cytoskeletal proteins.

In vitro transcription/translation of actin cDNA and analysis of the translation products by native-PAGE was used to study the maturation pathway of actin. During the course of actin synthesis, several distinct actin-containing species were observed and the composition of each determined by immunological procedures. After synthesis of the first approximately 145 amino acids, the nascent ribosome-associated actin chain binds to the recently identified heteromeric chaperone protein, prefoldin (PFD). PFD remains bound to the relatively unfolded actin polypeptide until its posttranslational delivery to cytosolic chaperonin (CCT). We show that alpha- and beta-tubulin follow a similar maturation pathway, but to date find no evidence for an interaction between PFD and several noncytoskeletal proteins. We conclude that PFD functions by selectively targeting nascent actin and tubulin chains pending their transfer to CCT for final folding and/or assembly.

Brain-specific expression of MAP2 detected using a cloned cDNA probe.

We describe the isolation of a set of overlapping cDNAs encoding mouse microtubule associated protein 2 (MAP2), using an anti-MAP antiserum to screen a mouse brain cDNA expression library cloned in bacteriophage lambda gt11. The authenticity of these clones was established by the following criteria: (a) three non-identical clones each expressing a MAP2 immunoreactive fusion protein were independently isolated from the expression library; each of these clones cross-hybridized at the nucleic acid level; (b) anti-MAP antiserum was affinity purified using nitrocellulose-bound fusion protein; these antibodies detected only MAP2 in an immunoblot experiment of whole brain microtubule protein; (c) a series of cDNA "walking" experiments was done so as to obtain a non-overlapping cloned fragment corresponding to a different part of the same mRNA molecule. Upon subcloning this non-overlapping fragment into plasmid expression vectors, a fusion protein was synthesized that was immunoreactive with an anti-MAP2 specific antiserum. Thus, a single contiguous cloned mRNA molecule encodes at least two MAP2-specific epitopes; (d) the cloned cDNA probes detect an mRNA species in mouse brain that is of a size (approximately 9 kb) consistent with the coding capacity required by a 250,000-D protein. The MAP2-specific cloned cDNA probes were used in RNA blot transfer experiments to assay for the presence of MAP2 mRNA in a variety of mouse tissues. Though brain contained abundant quantities of MAP2 mRNA, no corresponding sequences were detectable in RNA prepared from liver, kidney, spleen, stomach, or thymus. We conclude that the expression of MAP2 is brain-specific. Use of the MAP2 specific cDNA probes in genomic Southern blot transfer experiments showed the presence of a single gene encoding MAP2 in mouse. The microheterogeneity of MAP2 is therefore ascribable either to alternative splicing within a single gene, or to posttranslational modification(s), or both. Under conditions of low stringency, the mouse MAP2 cDNA probe cross-hybridizes with genomic sequences from rat, human, and (weakly) chicken, but not with sequences in frog, Drosophila, or sea urchin DNA. Thus, there is significant interspecies divergence of MAP2 sequences. The implications of the above observations are discussed in relationship to the potential biological function of MAP2.

Five mouse tubulin isotypes and their regulated expression during development.

We describe five mouse tubulin cloned cDNAs, two (M alpha 1 and M alpha 2) that encode alpha-tubulin and three (M beta 2, M beta 4, and M beta 5) that encode beta-tubulin. The sequence of these clones reveals that each represents a distinct gene product. Within the sequence common to the two alpha-tubulin cDNAs, the encoded amino acids are identical, though the 3' noncoding regions are wholly dissimilar. In contrast, the three beta-tubulin cDNAs show considerable carboxy-terminal heterogeneity. Two of the beta-tubulin isotypes defined by the cloned sequences are absolutely conserved between mouse and human, and all three beta-tubulin isotypes are conserved between mouse and rat. This result implies the existence of selective constraints that have maintained sequence identity after species divergence. This conclusion is reinforced by the near identity between a third mouse beta-tubulin isotype and a chicken beta-tubulin isotype. The significance of the interspecies conservation of tubulin isotypes is discussed in relationship to microtubule function. We have used non-cross-hybridizing 3' noncoding region probes from the five cloned mouse tubulin cDNAs to study the developmental expression of each isotype in various mouse tissues. M alpha 1 and M beta 2 are expressed in an approximately coordinate fashion, and their transcripts are most abundant in brain and lung. M alpha 2 and M beta 5 are ubiquitously expressed and to a similar extent in each tissue, with the greatest abundance in spleen, thymus, and immature brain. In contrast, M beta 4 is expressed exclusively in brain. Whereas the expression of the latter isotype increases dramatically during postnatal development, transcripts from all four other tubulin genes decline from maximum levels at or before birth. Tissue-specific development changes in the abundance of tubulin isotype-specific mRNAs are discussed in relationship to organogenesis in the mouse.

Chaperonin-mediated folding of vertebrate actin-related protein and gamma-tubulin.

The folding of actin and tubulin is mediated via interaction with a heteromeric toroidal complex (cytoplasmic chaperonin) that hydrolyzes ATP as part of the reaction whereby native proteins are ultimately released. Vertebrate actin-related protein (actin-RPV) (also termed centractin) and gamma-tubulin are two proteins that are distantly related to actin and tubulin, respectively: gamma-tubulin is exclusively located at the centrosome, while actin-RPV is conspicuously abundant at the same site. Here we show that actin-RPV and gamma-tubulin are both folded via interaction with the same chaperonin that mediates the folding of beta-actin and alpha- and beta-tubulin. In each case, the unfolded polypeptide forms a binary complex with cytoplasmic chaperonin and is released as a soluble, monomeric protein in the presence of Mg-ATP and the presence or absence of Mg-GTP. In contrast to alpha- and beta-tubulin, the folding of gamma-tubulin does not require the presence of cofactors in addition to chaperonin itself. Monomeric actin-RPV produced in in vitro folding reactions cocycles efficiently with native brain actin, while in vitro folded gamma-tubulin binds to polymerized microtubules in a manner consistent with interaction with microtubule ends. Both monomeric actin-RPV and gamma-tubulin bind to columns of immobilized nucleotide: monomeric actin-RPV has no marked preference for ATP or GTP, while gamma-tubulin shows some preference for GTP binding. We show that actin-RPV and gamma-tubulin compete with one another, and with beta-actin or alpha-tubulin, for binary complex formation with cytoplasmic chaperonin.

Nerve growth factor regulates both the phosphorylation and steady-state levels of microtubule-associated protein 1.2 (MAP1.2).

This study characterizes effects of nerve growth factor (NGF) on the steady-state level and phosphorylation of a high molecular mass microtubule-associated protein in PC12 rat pheochromocytoma cells. Past work showed that NGF significantly raises the relative levels of this phosphoprotein, designated MAP1.2, with a time course similar to that of neurite outgrowth. To study this in greater detail, MAP1.2 in PC12 cell lysates was resolved by SDS-PAGE in gels containing 3.25% acrylamide/4 M urea and identified by comigration with material immunoprecipitated from the lysates by MAP1 antibodies. Quantification by metabolic radiolabeling with [35S]methionine or by silver staining revealed a 3.0-3.5-fold increase in MAP1.2 levels relative to total cell protein after NGF treatment for 2 wk or longer. A partial increase was detectable after 3 d, but not after 2 h of NGF exposure. As measured by incorporation of [32P]phosphate, NGF had a dual effect on MAP1.2. Within 15 min to 2 h, NGF enhanced the incorporation of phosphate into MAP1.2 by two- to threefold relative to total cell phosphoproteins. This value slowly increased thereafter so that by 2 wk or more of NGF exposure, the average enhancement of phosphate incorporation per MAP1.2 molecule was over fourfold. The rapid action of NGF on MAP1.2 could not be mimicked by either epidermal growth factor, a permeant cAMP derivative, phorbol ester, or elevated K+, each of which alters phosphorylation of other PC12 cell proteins. SDS-PAGE revealed multiple forms of MAP1.2 which, based on the effects of alkaline phosphatase on their electrophoretic mobilities, differ, at least in part, in extent of phosphorylation. Before NGF treatment, most PC12 cell MAP1.2 is in more rapidly migrating, relatively poorly phosphorylated forms. After long-term NGF exposure, most is in more slowly migrating, more highly phosphorylated forms. The effects of NGF on the rapid phosphorylation of MAP1.2 and on the long-term large increase in highly phosphorylated MAP1.2 forms could play major functional roles in NGF-mediated neuronal differentiation. Such roles may include effects on microtubule assembly, stability, and cross-linking and, possibly for the rapid effects, nuclear signaling.

A novel cochaperonin that modulates the ATPase activity of cytoplasmic chaperonin.

The folding of alpha- and beta-tubulin requires three proteins: the heteromeric TCP-1-containing cytoplasmic chaperonin and two additional protein cofactors (A and B). We show that these cofactors participate in the folding process and do not merely trigger release, since in the presence of Mg-ATP alone, alpha- and beta-tubulin target proteins are discharged from cytoplasmic chaperonin in a nonnative form. Like the prokaryotic cochaperonin GroES, which interacts with the prototypical Escherichia coli chaperonin GroEL and regulates its ATPase activity, cofactor A modulates the ATPase activity of its cognate chaperonin. However, the sequence of cofactor A derived from a cloned cDNA defines a 13-kD polypeptide with no significant homology to other known proteins. Moreover, while GroES functions as a heptameric ring, cofactor A behaves as a dimer. Thus, cofactor A is a novel cochaperonin that is structurally unrelated to GroES.

Tubulin subunits exist in an activated conformational state generated and maintained by protein cofactors.

The production of native alpha/beta tubulin heterodimer in vitro depends on the action of cytosolic chaperonin and several protein cofactors. We previously showed that four such cofactors (termed A, C, D, and E) together with native tubulin act on beta-tubulin folding intermediates generated by the chaperonin to produce polymerizable tubulin heterodimers. However, this set of cofactors generates native heterodimers only very inefficiently from alpha-tubulin folding intermediates produced by the same chaperonin. Here we describe the isolation, characterization, and genetic analysis of a novel tubulin folding cofactor (cofactor B) that greatly enhances the efficiency of alpha-tubulin folding in vitro. This enabled an integrated study of alpha- and beta-tubulin folding: we find that the pathways leading to the formation of native alpha- and beta-tubulin converge in that the folding of the alpha subunit requires the participation of cofactor complexes containing the beta subunit and vice versa. We also show that sequestration of native alpha-or beta-tubulins by complex formation with cofactors results in the destabilization and decay of the remaining free subunit. These data demonstrate that tubulin folding cofactors function by placing and/or maintaining alpha-and beta-tubulin polypeptides in an activated conformational state required for the formation of native alpha/beta heterodimers, and imply that each subunit provides information necessary for the proper folding of the other.

Microtubule-associated protein MAP2 shares a microtubule binding motif with tau protein.

The microtubule-associated protein MAP2 is a prominent large-sized component of purified brain microtubules that, like the 36- to 38-kilodalton tau proteins, bears antigenic determinants found in association with the neurofibrillary tangles of Alzheimer's disease. The complete sequence of mouse brain MAP2 was determined from a series of overlapping cloned complementary DNAs. The sequence of the carboxyl-terminal 185 amino acids is very similar (67 percent) to a corresponding region of tau protein, and includes a series of three imperfect repeats, each 18 amino acids long and separated by 13 or 14 amino acids. A subcloned fragment spanning the first two of the 18-amino acid repeats was expressed as a polypeptide by translation in vitro. This polypeptide copurified with microtubules through two successive cycles of polymerization and depolymerization, whereas a control polypeptide derived from the amino-terminal region of MAP2 completely failed to copurify. These data imply that the carboxyl-terminal domain containing the 18-amino acid repeats constitutes the microtubule binding site in MAP2. The occurrence of these repeats in tau protein suggests that these may be a general feature of microtubule binding proteins.

Diverse mechanisms in the generation of human beta-tubulin pseudogenes.

The sequence of two human beta-tubulin pseudogenes is described. One contains an intervening sequence but lacks sequences encoding the 55 N-terminal amino acids of the polypeptide chain. A second has no introns but has a polyadenylate signal and an oligoadenylate tract at its 3' end, and it is flanked by a short direct repeat. These sequences have arisen by different mechanisms, including one that probably involves reverse transcription of a processed messenger RNA and reintegration of the complementary DNA copy into the genome.

GFAP is necessary for the integrity of CNS white matter architecture and long-term maintenance of myelination.

To investigate the structural role of glial fibrillary acidic protein (GFAP) in vivo, mice carrying a null mutation in GFAP were generated. In 7/14 mutant animals older than 18 months of age, hydrocephalus associated with white matter loss was detected. Mutant mice displayed abnormal myelination including the presence of actively myelinating oligodendrocytes in adults, nonmyelinated axons in optic nerve, and reduced myelin thickness in spinal cord. White matter was poorly vascularized and the blood-brain barrier was structurally and functionally impaired. Astrocytic structure and function were abnormal, consisting of shortened astrocytic cell processes, decreased septation of white matter, and increased CNS extracellular space. Thus, GFAP expression is essential for normal white matter architecture and blood-brain barrier integrity, and its absence leads to late-onset CNS dysmyelination.

Anomalous placement of introns in a member of the intermediate filament multigene family: an evolutionary conundrum.

The origin of introns and their role (if any) in gene expression, in the evolution of the genome, and in the generation of new expressed sequences are issues that are understood poorly, if at all. Multigene families provide a favorable opportunity for examining the evolutionary history of introns because it is possible to identify changes in intron placement and content since the divergence of family members from a common ancestral sequence. Here we report the complete sequence of the gene encoding the 68-kilodalton (kDa) neurofilament protein; the gene is a member of the intermediate filament multigene family that diverged over 600 million years ago. Five other members of this family (desmin, vimentin, glial fibrillary acidic protein, and type I and type II keratins) are encoded by genes with six or more introns at homologous positions. To our surprise, the number and placement of introns in the 68-kDa neurofilament protein gene were completely anomalous, with only three introns, none of which corresponded in position to introns in any characterized intermediate filament gene. This finding was all the more unexpected because comparative amino acid sequence data suggest a closer relationship of the 68-kDa neurofilament protein to desmin, vimentin, and glial fibrillary acidic protein than between any of these three proteins and the keratins. It appears likely that an mRNA-mediated transposition event was involved in the evolution of the 68-kDa neurofilament protein gene and that subsequent events led to the acquisition of at least two of the three introns present in the contemporary sequence.

A novel 205-kilodalton testis-specific serine/threonine protein kinase associated with microtubules of the spermatid manchette.

To identify proteins which interact with and potentially modulate the function of microtubules during spermatogenesis, we prepared a total testis MAP (microtubule-associated protein) antiserum and used it to isolate cDNA clones from a mouse testis cDNA expression library. Antibodies affinity purified by using one expression clone recognized a 205-kDa protein, termed MAST205, which colocalizes with the spermatid manchette. Sequencing of full-length cDNA clones encoding MAST205 revealed it to be a novel serine/threonine kinase with a catalytic domain related to those of the A and C families. The testis-specific MAST205 RNA increases in abundance during prepuberal testis development, peaking at the spermatid stage. The microtubule-binding region of MAST205 occupies a central region of the molecule including the kinase domain and sequences C terminal to this domain. Binding of MAST205 to microtubules requires interaction with other MAPs, since it does not bind to MAP-free tubulin. A 75-kDa protein associated with immunoprecipitates of MAST205 from extracts of both whole testis and testis microtubules becomes phosphorylated in in vitro kinase assays. This 75-kDa substrate of the MAST205 kinase may form part of the MAST205 protein complex which binds microtubules. The MAST205 protein complex may function to link the signal transduction pathway with the organization of manchette microtubules.