BIOLOGICAL OVERVIEW

Cloning and functional analysis of the Drosophila Lissencephaly-1 (Lis1) gene have led to further understanding of the roles of the oocytic cytoskeletal network in the nuclear migration and the development of the oocyte (Liu, 1999; Swan, 1999, and Lei, 2000). Lis1 has seven WD repeats and is a protein that resembles the beta-subunit of trimeric G proteins: several proteins with such repeats have been shown to interact with other polypeptides, giving rise to multiprotein complexes. Lis1 is a component of the Dynein complex, a set of polypeptides involved in the function of the Dynein molecular motor (see Dynein heavy chain 64C). This overview will first briefly describe the consequences of Lis1 mutation in mammals, the role of Lis1 homologs in nuclear migration of model organisms, and then describe the role of Drosophila Lis1 in oocyte development.

Lissencephaly is a congenital brain malformation manifested by a smooth cerebral surface due to incomplete neuronal migration (Dobyns, 1989). Type I lissencephaly occurs either as an isolated abnormality or in association with dysmorphic facial appearance in Miller-Dieker syndrome (MDS). Lissencephaly patients usually die young; they evince profound mental retardation, muscle weakness and seizures. One causal gene for lissencephaly, lissencephaly 1 (LIS1), encodes a protein containing seven WD-40 repeats (Reiner, 1993). Heterozygous deletions or mutations in the LIS1 gene have been identified in patients with MDS and isolated lissencephaly (Reiner, 1993; Lo Nigro, 1997), indicating the haplo-insufficiency of LIS1. Mice homozygous for a null Lis1 mutation die early in embryogenesis soon after implantation. Heterozygous and compound heterozygous mice have dosage-dependent defects in neuronal migration and neurogenesis (Hirotsune, 1998). Lissencephaly is also caused by mutations in an X-linked gene encoding a potentially phosphorylated protein that may play a role in a signal transduction pathway overlapping with the function of LIS1 (des Portes, 1998; Gleeson, 1998; Reiner, 2000; Liu, 1999 and references therein).

LIS1 has also been identified biochemically in bovine brain extracts as the regulatory subunit alpha of platelet-activating factor acetylhydrolase (PAFAH; Hattori, 1994). PAFAH contains two additional catalytic subunits (beta and gamma): the crystal structure of the beta subunit closely resembles that of GTPases (Ho, 1997). PAFAH catalyzes the removal of the acetyl group at the sn-2 position and produces biologically inactive lyso-PAF. Thus, PAFAH is important for regulating levels of the active PAF, which has potent biological functions in diverse organs, including the central nervous system (Liu, 1999 and references therein).

A function for PAFAH in the brain is also suggested by the observation that LIS1 and the two catalytic subunits are highly expressed in the same developing brain tissues (Mizuguchi, 1995; Albrecht, 1996). Since PAF is membrane localized, it is possible that the role of LIS1 is to localize the beta and gamma catalytic subunits of PAFAH to the plasma membrane. LIS1 has been shown to associate with tubulin (Sapir, 1997) and beta-spectrin (Wang, 1995); to co-localize with microtubules in tissue culture cells, and to reduce microtubule catastrophe frequency in vitro (Sapir, 1997; Liu, 1999 and references therein).

A LIS1 homolog, nudF (Xiang, 1995), was isolated in a screen for nuclear migration mutants in Aspergillus nidulans. NudF interacts genetically with nudC, another nuclear migration gene in A. nidulans (Osmani, 1990; Xiang, 1995). NUDC is required to maintain a normal concentration of NUDF protein (Xiang, 1995). The murine NUDC homolog and Lis1 are co-expressed in the ventricular zone of the forebrain and in the cortical plate, and they also interact in a two-hybrid system (S. M. Morris, 1998), suggesting that nuclear migration may play an important role for neuronal or cell migration (N. R. Morris, 1998). Components of the dynein complex have also been identified as nuclear migration mutants in the filamentous fungi A. nidulans and Neurospora crassa. These components include cytoplasmic dynein heavy chain (NUDA and RO-1); cytoplasmic dynein light chain (NUDG); p150^Glued (RO-3), which is the largest polypeptide in the dynactin complex; Glued stimulates vesicle movement by dynein, and centractin (RO-4), the most abundant component in the dynactin complex. PAC1, sharing significant identity with LIS1, is one of the components of the cytoplasmic dynein pathway in Saccharyomyces cerevisiae (Geiser, 1997). These results support the notion that LIS1 and the dynein complex function together to regulate nuclear movement in fungi (Liu, 1999 and references therein).

Analysis of Lis1 ovarian mutant clones demonstrates that Lis1 is required in the germline for synchronized cystocyte division and oocyte differentiation. In the Drosophila ovary, two or three germline stem cells are located at the tip of each ovariole, the germarium. Each stem cell divides asymmetrically to produce a new stem cell and a cystoblast that undergoes four rounds of synchronized divisions with incomplete cytokinesis to form a 16-cell cyst. The cystocytes in one cyst are interconnected in a stereotypical pattern by stable intercellular bridges termed ring canals. Following cyst formation, oocyte-specific transcripts and proteins accumulate in one cell that contains four ring canals and then develops as the oocyte. The other fifteen cells become polyploid nurse cells supplying nutrients for oocyte maturation and subsequent embryonic development (Liu, 1999).

A germline-specific cytoplasmic vesicle, the fusome, grows from a small sphere in a cystoblast into a large branched structure that extends through the ring canals into every cell of a developing cyst. The fusome is composed of membrane skeletal proteins alpha-spectrin, beta-spectrin, ankyrin and Hu-li tai shao (Hts), an adducin-like protein. Mutations in the genes encoding Hts or alpha-spectrin apparently eliminate the fusomes and cause the formation of cysts with fewer than 16 cells, usually without an oocyte. Cytoplasmic dynein associates with the fusome in a cell-cycle-dependent manner. Mutations in the cytoplasmic dynein heavy chain gene, Dhc64C, cause disruption of spindle orientation in the dividing cyst and block oocyte differentiation. The mutant cysts contain fewer than 16 cells, suggesting that dynein is also involved in cystocyte division. In Drosophila, dynein forms a protein complex with the p150 subunit of dynactin encoded by glued. The gene bag-of-marbles (bam) encodes another component of the fusome required for cystoblast differentiation and is thought to be involved in vesicle recruitment for fusome assembly. Taken together, these observations indicate that the fusome has a critical role in regulating the synchronous cystocyte division and establishing or maintaining cyst polarity (Liu, 1999 and references therein).

The rapid, synchronous divisions of the cystoblast are the basis for the formation of a normal germline cyst and the ultimate differentiation of the oocyte and 15 nurse cells. It was found that, in Lis1 mutant germline clones, these critical cell divisions are disrupted, resulting in cysts with a significantly reduced number of cells that almost never contain an oocyte. In these mutant cysts, the fusomes are aberrantly formed and do not branch normally: they are either fragmented or grow to larger than normal spheres that eventually disintegrate (Liu, 1999).

What is the connection between the cystoblast divisions and the fusome formation and branching, and how does Lis1 function impact these processes? At least two hypotheses can be advanced to explain the observed phenotypes. The first hypothesis suggests that fusome formation may be more sensitive than cystoblast divisions to the levels of Lis1. If this is the case, the observed disruption of the rapid and synchronized germline cell divisions would result from the defects in the fusome. This hypothesis is supported by the similarity of the Lis1 cyst phenotypes to those observed in hts and alpha spectrin mutants. In such mutants, the abnormal cysts are apparently caused by the disruption of the fusome. Both these genes encode membrane skeletal proteins that are integral parts of the fusome and, in mutants, the fusome is not formed (Liu, 1999).

The second hypothesis is that cystoblast division is directly affected by the absence of Lis1 function, resulting in arrest of division in early stages. The abnormality in cell division would result in the formation of the large spherical fusomes and their ultimate disintegration. Four observations support this second hypothesis. (1) In a germarium that contained only Lis1 mutant cysts, the number of developing cysts was dramatically reduced. Fusomes are not essential for the division of germline stem cells and cystoblasts. Thus, the reduction in division rates of the stem cells or cystoblasts appears to be due to a lack of Lis1. (2) In some mutant cysts cell division is arrested and the nuclei apparently undergo endoreplication and become prematurely polyploid. The fusomes in these cysts often grow in size without branching and some fusomes later degenerate. (3) When the function of Lis1 was investigated by clonal analysis in the eye, mutant clones were never observed, a result consistent with Lis1 also functioning in the zygote to control cell division and differentiation. (4) In agreement with a function of Lis1 in cell division and differentiation, homozygous Lis1 knock-out mice die early in embryogenesis (Liu, 1999 and references therein).

The fusome grows from a small sphere in a cystoblast into a large branched structure that extends through the ring canals into every cell of a developing cyst. During stem cell and cyst cell division, one pole of the mitotic spindle associates with the fusome. Following mitosis, newly synthesized fusome plugs move toward and fuse with the pre-existing fusome. Their associated ring canals also move, changing the geometry of the cyst and resulting in the formation of a rosette. The fusome remains asymmetrically distributed within the cyst after the first division, thereby generating a polarity in the developing cyst. Although it is as yet impossible to pinpoint the precise role of Lis1 in early oogenesis, it is tempting to speculate that Lis1 is involved in the movement of the ring canals and the newly formed fusome plugs to the original fusomes to form continuous fusomes and the rosette structure. Lis1 may also be involved in the constant transport of vesicles into fusomes to maintain the dynamic fusome structure. The failure of Lis1 mutant cysts to differentiate an oocyte is not surprising. The disruption in germline cell division and fusome formation leads to the disruption of polarity of the cyst and consequently the differentiation of the oocyte. Even when the fusome and cysts appear normal, oocyte differentiation is affected, suggesting a more stringent and constant requirement for Lis1 function in oocyte differentiation (Liu, 1999).

In Aspergillus nidulans, the Lis1 homolog nudF interacts genetically with the cytoplasmic dynein heavy chain gene nudA (Willins, 1997). Other components of the dynein motor complex, cytoplasmic dynein light chain, p150, and centractin subunits of dynein receptor dynactin have also been implicated in the same pathway in filamentous fungi (S. M. Morris, 1998). Drosophila Lis1 and Dhc64C mutants have overlapping phenotypes in cyst formation, fusome integrity and oocyte differentiation, suggesting that the two genes also function in Drosophila in the same processes. An overlapping function is also suggested by biochemical experiments. Lis1 is present in microtubule-associated proteins (MAPs) prepared from Drosophila embryo extracts. However, unlike dynein, Lis1 is not enriched in MAP preparations and a substantial pool of Lis1 remains in the supernatant. Moreover, a significant amount of Lis1 is released when the MT pellet is washed in buffer and MTs are repelleted. The significance of the association of Lis1 with MTs in vitro remains unclear but may reflect a low affinity interaction (Liu, 1999).

Several observations suggest that Lis1 not only functions in maintaining the integrity of fusomes, but also in maintaining membrane skeletons in general. (1) Vertebrate Lis1 was shown to associate in vitro with the pleckstrin homology domain of beta-spectrin (Wang, 1995). (2) Platelet-activating factor, the substrate of platelet-activating factor acetylhydrolase (PAFAH, an enzyme that contains Lis1 functioning as a regulatory subunit), is localized to the plasma membrane, suggesting a role for Lis1 in the localization of the enzyme complex to the plasma membrane. (3) In the heterotrimeric G protein, Gbeta serves to anchor Gs alpha to the cytoplasmic face of the plasma membrane. PAFAH is a unique G-protein-like trimer (Ho, 1997). By analogy, Lis1 could have a similar function as Gs (Liu, 1999 and references therein).

The finding that centractin, a component of the dynactin complex, is associated with spectrin (Holleran, 1996) provides a link of the dynein motor complex with membrane skeletal proteins, thus dynein and Lis1 may interact at the plasma membrane. In support of the idea that Lis1 functions in maintaining the integrity of membrane skeletons in general, it was observed that mutant Lis1 cysts were mispackaged, a phenotype that has not been described for mutants in any of the genes that function in cyst formation in early germarium. This phenotype may be due to a loss of organization and integrity of cyst cytoskeletons, such that migrating follicle cells separate one cyst into two parts (Liu, 1999).

The function of microtubules in germline cell division and in fusome formation remains to be elucidated. A prominent MT organization center is established only after the formation of the 16-cell cyst. Lis1 and dynein are likely to interact, and to function throughout oogenesis to affect the transport of determinants to the oocyte and the migration of the oocyte nucleus. Indeed, both processes are affected in Lis1 hypomorphic mutations. The integrity of the membrane skeleton is essential for the organization of cells and the establishment of the MT network. Their disruption would affect both nuclear and cellular migration. Lis1 is likely to function in these two processes through the interactions with the dynein motor complex and membrane skeletons (Liu, 1999 and references therein).

Studies of Lis-1 mutants reveal that Lis-1 acts together with Bicaudal-D (Bic-D), Egalitarian (Egl), dynein and microtubules to determine oocyte identity. Lis-1 is further required for nurse-cell-to-oocyte transport during oocyte growth, and for the positioning of the nucleus in the oocyte. To identify genes required for oocyte determination, a screen was carried out for dominant genetic enhancers of Bic-D, and one strong enhancer, E415, was chosen for further characterization. In a Bic-D⁺ background, females hemizygous for E415 are sterile. Half of the egg chambers contain 16 polyploid nuclei, indicating that each cell within the chamber adopted a nurse-cell fate. In 20% of the egg chambers an oocyte is detectable by the presence of an under-replicated nucleus, but this cell is often mispositioned and oocyte-specific factors either fail to accumulate or accumulate only transiently in a single cell; such defective accumulation occurs for Oskar mRNA. In wild-type egg chambers, a microtubule-organizing center develops in the oocyte as early as region 2 of the germarium. This is not detectable in Bic-D and egl mutants, which fail to make an oocyte, and is also not observed in most hemizygous E415 mutant egg chambers. Therefore, E415 is essential for oocyte determination and interacts with Bic-D in this process. The remaining 30% of hemizygous E415 egg chambers exhibit a later defect in cyst encapsulation, a phenotype not seen in Bic-D-null or egl mutants. Therefore, E415 also has functions within the ovary that appear to be independent of Bic-D and egl (Swan, 1999).

Genes required for oocyte determination block this process at either of two steps. One set of genes is required for differentiation of a germline-specific membranous organelle called the fusome and for proper division of the germ line. Mutations in the second set of genes comprised of Bic-D and egl, do not affect cystocyte division and fusome formation, and appear to function at a later step in oocyte determination. E415 hemizygous egg chambers rarely contain fewer than 16 cells, and they appear to develop a normal fusome that is correctly associated with the mitotic spindle. Disruption of this association is seen rarely, possibly reflecting the low percentage of cysts with defects in cystocyte divisions. Therefore, it is likely that the E415 gene is involved in fusome/spindle anchoring, but that the hypomorphic E415 allele belongs to the second group of mutants that affect oocyte determination at a later step (Swan, 1999).

E415 homozygous mutants exhibit a milder range of phenotypes than hemizygous mutants, allowing a study of the function of E415 during later patterning of the egg chamber. These females show a striking reduction in oocyte growth. This phenotype is also seen in Bic-D^mom females, indicating that these genes may function together in nurse-cell-to-oocyte transport. As in Bic-D^mom mutants, this defect in E415 homozygotes appears to be specific to the early stages of oogenesis, since later egg chambers are almost wild-type in size (Swan, 1999).

Disruption of microtubules also results in a block in oocyte determination and a reduction in oocyte growth, suggesting the possibility that E415 is involved in the organization of microtubules. However, E415/E415 mutant ovaries show normal distribution of the microtubule subunit alpha-tubulin, including the strong focus of tubulin expression in the oocyte that reflects the asymmetric organization of microtubules in the egg chamber. In addition, oocyte-specific factors (osk, orb and Bic-D mRNAs and Orb, Bic-D and Egl proteins), all of which depend on microtubules for their localization, still accumulate in the E415/E415 oocyte at or near wild-type levels. Thus the effect of E415 on oocyte determination and on oocyte growth is not due to a general disruption of microtubule organization (Swan, 1999).

Patterning of the Drosophila embryo depends on the correct localization of patterning determinants within the oocyte, beginning in mid-oogenesis. Starting in stage 8, specific mRNAs begin to accumulate either at the posterior end of the oocyte or along the anterior cortex; this pattern is dependent on Bic-D and microtubules. The distributions of anteriorly localized factors (orb, egl, nos and bcd mRNAs and Orb protein) and posteriorly localized factors (osk mRNA and Osk protein) were studied in E415 ovaries. All of these factors show disruptions in their normal subcellular localization within the oocyte. These effects on localization could be due to effects on transport, anchoring or stability of transcripts in the mutant (Swan, 1999).

In stage 7 of oogenesis, the oocyte nucleus migrates from the posterior to the future dorsal-anterior corner of the oocyte. In about half of stage-9 and stage-10 E415/E415 mutant egg chambers, the oocyte nucleus is positioned incorrectly. This could be due to a failure in nuclear migration or anchoring. Since the oocyte nucleus takes up almost the entire oocyte in stage-8 and earlier mutant egg chambers, it is not possible to determine whether nuclear positioning is affected before stage 9. Bic-D is also required for positioning of the oocyte nucleus, and therefore E415 could function in nuclear positioning through the Bic-D-Egl complex. Consistent with this view, Bic-D and Egl proteins accumulate at high levels in a region between the oocyte cortex and nucleus, and this localization is abolished in E415 mutants. Microtubules are also required for oocyte nuclear positioning, and whereas microtubule organization in E415 homozygous mutants appears to be unaffected in early oogenesis, some alterations are observed in microtubules after stage 7. In wild-type oocytes, microtubules appear to be concentrated at the anterior cortex, apparently reflecting their nucleation from anterior sites. Microtubules still appear to be organized in this way in E415 mutants, but are often less focused. Also, a marker for microtubule plus ends, kinesin-beta-galactosidase, still accumulates at the posterior end in E415 mutants but at reduced levels (Swan, 1999).

Lis1 shares with the dynein heavy chain gene, Dhc, a requirement in oocyte determination, indicating that it may function like its fungal homologs in a pathway with dynein/dynactin. A specific allele of Dhc, Dhc^6-6, dominantly suppresses the rough-eye phenotype produced by a mutation in the dynactin component Glued (Gl¹ mutation) whereas a deficiency removing Dhc has no effect. This allele-specific interaction is evidence that Gl and Dhc may act in the same pathway. Similarly, Dhc^6-6 behaves as a strong dominant suppressor of the Lis1^E415 homozygous phenotype, resulting in fertility, proper nuclear positioning and near normal oocyte growth. A deficiency of Dhc and the point mutation Dhc ^6-12 have no effect on the Lis1^E415 phenotype. Therefore Dhc shows the same allele-specific interaction with Lis1 as it does with Gl, indicating that Lis1 may function in a genetic pathway with Dhc, and implicating dynein in nurse-cell-to-oocyte transport and nuclear positioning in the oocyte. Genetic interactions between Gl and Lis1 and between Gl and Bic-D were also examined. The antimorphic Gl¹ mutation confers lethality in a Lis1^E415 homozygote and in a Bic-D^PA66/Df(2L)TW119 background, indicating that both Lis1 and Bic-D may function in the same essential process as dynactin (Swan, 1999).

To determine how Lis1 could interact with dynein/dynactin, the localization of Lis1 and Dhc proteins was examined in wild-type oocytes and in oocytes with mutations in either gene. Lis1 signal is concentrated along the cortex of wild-type oocytes from as early as stage 5 of oogenesis. To determine whether Lis1 localization is dependent on dynein function, Lis1 distribution was examined in Dhc^6-6/ Dhc^6-12 mutants. This hypomorphic allelic combination has no effect on the cortical accumulation of Lis1 protein. Lis1 localization was examined in egg chambers treated with the microtubule-destabilizing drug colchicine. Under conditions that disrupt microtubules and dynein localization, cortical Lis1 signal is still present, indicating that its cortical localization or maintenance does not depend on localized dynein or microtubules (Swan, 1999).

Dynein localization, in contrast, is dependent on Lis1. In wild-type egg chambers, Dhc localizes to the presumptive oocyte in region 2 of the germarium and remains enriched in the oocyte throughout oogenesis. In Lis1^E415 mutants, this specific accumulation is completely abolished. This strong effect of Lis1^E415 on Dhc localization contrasts with its subtle effect on the accumulation of other oocyte-specific factors in the oocyte, indicating that Lis1 may specifically regulate Dhc localization and that the localization of most oocyte-specific factors to the oocyte is independent of Lis1 and Dhc function. Thus there appear to be two distinct microtubule-dependent nurse-cell-to-oocyte transport mechanisms at work during oogenesis. One involves Bic-D, Lis1 and Dhc, is involved in bringing oocyte determinants into the presumptive oocyte, and, subsequently, is responsible for oocyte growth in stages 1-7 of oogenesis. A second microtubule-based transport mechanism could function during these stages in the transport of oocyte-specific mRNAs and proteins into the oocyte, possibly using other dyneins (Swan, 1999).

In wild-type egg chambers at stages 7-9, Dhc accumulates along the oocyte cortex and around the oocyte nucleus. Later in stage 9, Dhc accumulates mainly at the posterior and anterior oocyte margins. These aspects of Dhc localization are disrupted in Lis1 mutants and, therefore, Lis1 is necessary for most or all aspects of dynein localization within the female germ line (Swan, 1999).

Given the role of Lis-1 homologs in fungi, it has been suggested that the requirement for human Lis-1 in neuronal migration could also reflect a role in nuclear migration. In the developing cerebral cortex and cerebellum, migrating neurons project out a cytoplasmic extension toward their target, and then their nucleus translocates along this extension. An analogous process occurs during neural development in Drosophila. In the third-instar eye imaginal disc, undifferentiated cells lie at the basal surface and extend processes apically. As photoreceptor cells differentiate posterior to the morphogenetic furrow, their nuclei translocate to the apical surface of the eye disc. In serial confocal sections, nuclei start to appear ~4 µm below the apical surface and are no longer visible beyond 8 µm. In flies homozygous for a pupal-lethal allele of Lis1, Lis1^K13209, nuclei also start to appear 4 µm below the apical surface, but many nuclei are also found more basally. This phenotype is identical to that of mutants for Glued, suggesting that Lis1 functions with dynein/dynactin in nuclear migration in these neural cells. Given that Lis1 is 70% identical to the human Lis-1, these findings strongly support the possibility that the failure in neuronal migration in Miller-Dieker syndrome also results from a failure in dynein/dynactin-dependent nuclear migration (Swan, 1999).

As in the ovary, Lis1 appears to function with Bic-D in nuclear migration during eye development. Bic-D^r5 eye discs also exhibit a severe defect in nuclear positioning, with photoreceptor nuclei being found at all levels basal to 4 µm and, frequently, in the axons that project basally from these cells (Swan, 1999).

The requirement for Lis1 and Bic-D in nuclear positioning in the developing eye imaginal disc could reflect a role for these genes in microtubule organization. Microtubules in the eye disc extend longitudinally along the apical-basal axis, but the polarity of these microtubules is not known. To establish the orientation of these microtubules, third-instar eye imaginal discs were stained with antibodies to gamma-tubulin. In wild-type imaginal discs, gamma-tubulin is found at high levels along the apical cortex and within this region in a single strong focus about 2 µm below the apical cortex and 2 µm apical to the photoreceptor nuclei. In Lis1^K13209 mutants, gamma-tubulin still accumulates at the apical surface, indicating that microtubules are oriented normally within the mutant photoreceptors. However, the subapical focus of gamma-tubulin is more diffuse and is undetectable in many photoreceptors, indicating a requirement for Lis1 in focusing microtubule minus ends. A similar effect is observed in Gl mutants, suggesting that Lis1 and dynein/dynactin also function in the same pathway in focusing microtubule minus ends in the eye. Interestingly, whereas Bic-D mutants consistently exhibit a more marked defect in nuclear positioning, localization of gamma-tubulin is normal in these mutants. This result correlates with the finding that microtubule organization is not affected in Bic-D^mom ovaries (Swan, 1999).

Several models for nuclear localization have been advanced. An analysis of Lis1 allows for the proposal of a model for nuclear migration in the Drosophila oocyte in which a cortical protein (Lis1) anchors microtubules via the minus-end-directed microtubule motor dynein: Lis1 appears to function at the cortex, and its localization to the cortex is independent of microtubules and dynein. The dynein heavy chain, Dhc, also associates with the oocyte cortex and this localization requires Lis1. Association of the nucleus with these microtubules would then allow it to be anchored to the cell cortex. The cortical localization of Dhc is also microtubule dependent, and this could indicate that dynein uses microtubules to reach the cortex. Bic-D-Egl could mediate the interaction between Lis1 and dynein or the interaction between microtubules and the oocyte nucleus. As well as localizing to the oocyte cortex, Dhc also associates with the oocyte nucleus throughout oogenesis, indicating that it may function as well in linking the nucleus to microtubules (Swan, 1999).

The ability of Lis1 to associate with the cell cortex could reflect an interaction between Lis1 and the membrane-associated cortical cytoskeleton. Similarly, the role of Lis1 in nurse-cell-to-oocyte transport could reflect an ability of this protein to interact with vesicles coated with membrane cytoskeletal proteins. Interestingly, mammalian Lis-1 interacts physically with the membrane cytoskeletal protein spectrin in vitro, while the dynactin complex protein Arp1 appears to associate with Golgi spectrin during vesicle transport (Swan, 1999 and references therein).

Lis1 seems to have several other functions in Drosophila. The lethality of Lis1^K13209 indicates that Lis1 has one or more essential zygotic functions. The synthetic lethality observed between Lis1 and Gl, and between Bic-D and Gl, suggests that Lis1 acts through Bic-D and dynein/dynactin in mediating at least some of its essential functions. However, there is good evidence that Lis1 can act through dynein/dynactin in a Bic-D-independent way. Defects are found in spindle/fusome alignment and in mitotic divisions in the germ line of Lis1 but not Bic-D mutants. Dynein/dynactin is implicated in focusing of microtubule minus ends during mitotic spindle assembly and the results indicate that Lis-1 could function with dynein/dynactin in this essential process. In addition, the microtubule-bundling defects in oocytes and photoreceptors of Lis1 (but not Bic-D) mutants could also reflect a role of Lis1 with dynein/dynactin in focusing microtubule minus ends in postmitotic cells (Swan, 1999).

GENE STRUCTURE

There are three alternatively spliced Lis1 transcripts. Two of the splice variants originate in a common 122-nucleotide 59 noncoding exon and are spliced either directly or by way of an optional 27-nucleotide exon to the 109-nucleotide exon that contains the translational start site. The initiation site for the third transcript lies 500 nucleotides 3' of the other two splice variants in a 42-nucleotide noncoding exon and is spliced directly to the exon containing the translation start signal. The six exons that encode the protein are common to all three mRNA isoforms. Six of the EST clones were obtained from the genome project and after preliminary analysis by restriction mapping, 4 representative cDNA clones were selected for sequencing. Analysis of these clones shows that they differ in the length of the 3' untranslated trailer sequences due to the utilization of alternative polyadenylation sites. In this context it may be significant that the mouse Lis1 locus also undergoes alternative splicing of upstream noncoding exons, and Lis1 genes in humans and mice use at least four alternative polyadenylation sites (Lei, 2000).

PROTEIN STRUCTURE

Amino Acids - 411

Structural Domains

The amino acid sequence of Drosophila Lis1 is 70% identical to mammalian LIS1, 41% identical to Aspergillus NUDF and 26% identical to yeast PAC1. The similarities extend along the whole proteins. They share an N-terminal coiled-coil motif and seven WD-40 repeats and are likely to form similar secondary structures. The high level of conservation suggests that the new gene is an ortholog of human LIS1 (Liu, 1999).

Drosophila Lis1 encodes a polypeptide with marked identity (70%) to the product of the human Lissencephaly-1, the causative gene for Miller-Dieker syndrome. Homologs of Lis1 have also been identified in the filamentous fungus Aspergillus nidulans and in Saccharomyces cerevisiae, and both function with the microtubule minus-end-directed motor dynein/dynactin in nuclear migration. Both mammalian Lis-1 and Drosophila Lis1 contain seven WD repeats, which confer on proteins a globular structure and are implicated in protein-protein interactions. Mammalian and Drosophila Lis1 proteins contain a region predicted to form a coiled-coil, as well as a highly conserved amino-terminal domain (Swan, 1999).

As in the human protein, the fourth and fifth repeats in Lis1 are separated by a large spacer sequence. In the G-protein subunit beta-transducin, for which the crystal structure is known, the seven WD repeats fold into a circular propeller-like shape with seven blades. Each blade of the beta-propeller consists of four anti-parallel beta-sheets and is thought to provide a surface for protein-protein interaction. Manual alignment of Drosophila Lis1 and beta-transducin sequences followed by secondary structure prediction using the Insight II molecular modeling program suggest that Lis1 can also form a seven-bladed beta-propeller structure. Data from biophysical and biochemical characterization of human LIS1 protein are consistent with the idea that it folds into a beta-propeller structure in vivo (Lei, 2000).

date revised: 23 June 2023

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.