The PAR-4 and PAR-1 kinases are necessary for the formation of the anterior-posterior (A-P) axis in Caenorhabditis elegans (Kemphues, 1988; Watts, 2000; Guo, 1995). PAR-1 is also required for A-P axis determination in Drosophila. The Drosophila par-4 homologue, lkb1, is required for the early A-P polarity of the oocyte, and for the repolarization of the oocyte cytoskeleton that defines the embryonic A-P axis. LKB1 is phosphorylated by PAR-1 in vitro, and overexpression of LKB1 partially rescues the par-1 phenotype. These two kinases therefore function in a conserved pathway for axis formation in flies and worms. lkb1 mutant clones also disrupt apical-basal epithelial polarity, suggesting a general role in cell polarization. The human homologue, LKB1, is mutated in Peutz-Jeghers syndrome (Hemminki, 1998; Jenne, 1998) and is regulated by prenylation and by phosphorylation by protein kinase A (Collins, 2000; Sapkota, 2001). Protein kinase A phosphorylates Drosophila LKB1 on a conserved site that is important for its activity. Thus, Drosophila and human LKB1 may be functional homologues, suggesting that loss of cell polarity may contribute to tumour formation in individuals with Peutz-Jeghers syndrome (Martin, 2003).

The A-P axis of Drosophila is specified during oogenesis when a signal from the posterior follicle cells induces the formation of a polarized oocyte microtubule cytoskeleton, in which most minus ends are nucleated from the anterior cortex, with the plus ends extending towards the posterior pole. These polarized microtubules direct both the localization of bicoid messenger RNA to the anterior of the oocyte to specify where the head and thorax will develop, and the transport of oskar mRNA to the posterior, where it induces the formation of polar granules that contain the abdominal and germline determinants (Martin, 2003).

To identify other genes required for formation of the A-P axis, a genetic screen was carried out in germline clones for mutants that disrupt the localization of Staufen tagged with green fluorescent protein (GFP), which colocalizes with bicoid and oskar mRNAs. Mutants in one lethal complementation group of two alleles abolish the posterior localization of both Staufen and oskar mRNA in 80%-90% of oocytes. In over 40% of mutant oocytes, both bicoid and K10 mRNAs are detected around the whole oocyte cortex, rather than only at the anterior pole. In contrast to wild-type oocytes, which have an anterior-to-posterior gradient of microtubules, mutant oocytes have a high density of microtubules all around the cortex, with the lowest concentration in the centre. Kinesin-ß-galactosidase, a microtubule plus-end marker, also fails to concentrate at the posterior of the oocyte and accumulates instead in the centre. Thus, these mutants disrupt bicoid and oskar mRNA localization by preventing the A-P polarization of the microtubules (Martin, 2003).

Hypomorphic mutants in Drosophila par-1 show defects in the polarization of the oocyte microtubule cytoskeleton and in the localization of bicoid and oskar mRNA that are very similar to the defects of lkb1 mutants. The PAR-1 kinase is also required much earlier in oogenesis for the determination of the oocyte. The oocyte is selected from a cyst of 16 germline cells in the germarium and forms a microtubule-organizing centre, which directs the microtubule-dependent localization of oocyte-specific factors, such as ORB, to this cell. The microtubule-organizing centre then moves from the anterior to the posterior of the oocyte in region 3, the most posterior region of the germarium, and oocyte-specific factors accumulate posteriorly. This anterior-to-posterior switch is disrupted in par-1 null mutants, and the oocyte exits meiosis and becomes a nurse cell (Martin, 2003).

To study the role of lkb1 in this process, mutant germline clones marked by the loss of GFP were induced. ORB still accumulates in the presumptive oocyte in lkb1 cysts but often fails to move to the posterior in region 3; instead, it disperses throughout the cyst as the oocyte exits meiosis and adopts the nurse cell fate. Thus, lkb1 and par-1 share very similar phenotypes in both oocyte determination and polarization, suggesting that they function together. Because the penetrance of the early lkb1 phenotype increases with age (67% of mutant clones in 5-day-old females, rising to 85% after 12 d), wild-type LKB1 activity seems to perdure for several days after the clones are induced, and this presumably accounts for the escapers that show the later oocyte polarity phenotype (Martin, 2003).

Since PAR-1 and LKB1 seem to act in a common pathway, whether either kinase could phosphorylate the other was tested. Although recombinant PAR-1 is not a substrate for LKB1, immunoprecipitated GFP-PAR-1 phosphorylates recombinant LKB1 in vitro. It was found that the phosphorylation site or sites are located in the amino-terminal half of the protein, although it has not been possible to map them precisely. Also a strong genetic interaction is observed between the two genes. The oskar mRNA localization defects in hypomorphic combinations of par-1 alleles cause loss of abdominal segments in the embryo, which can be strongly enhanced by removing one copy of lkb1. In addition, overexpression of GFP-LKB1 partially rescues the A-P polarity phenotype of par-1⁶³²³/par-1^W3, the strongest allelic combination that produces late-stage oocytes. Whereas Staufen localizes to the posterior normally in only 12% of these oocytes, 80% have wild-type amounts of Staufen at the posterior when LKB1 is overexpressed. This phenotype is not rescued by a kinase-dead form of LKB1, indicating that this suppression requires kinase activity. By contrast, overexpression of PAR-1 does not rescue the phenotype of lkb1 mutant germline clones. These results are consistent with a model in which LKB1 is a direct target of PAR-1 regulation in vivo and functions as a downstream effector in the polarization of the oocyte microtubule cytoskeleton (Martin, 2003).

A GFP-LKB1 fusion construct under the control of the endogenous promoter rescues both the lethality and oogenesis phenotypes of lkb1 mutants and is expressed in very low amounts in both the germline and somatic follicle cells of the ovary. The highest expression was observed in the germarium, where GFP-LKB1 colocalizes with PAR-1 on the fusome, a branched membranous organelle that connects the germ cells in a cyst. This localization presumably reflects their common function in early oocyte polarity and determination, because the cell that inherits most fusome is selected to become the oocyte. During the rest of oogenesis, GFP-LKB1 shows a uniform cortical localization in both the germline and the follicle cells. It is enriched in the oocyte from stage 7, when the A-P axis is polarized, and colocalizes with cortical actin, but not with pole plasm components (Martin, 2003).

Follicle cell clones mutant for lkb1 also show a defect in polarity. In severely affected clones, the follicular monolayer is disorganized, with mutant cells rounding up and sorting out of the epithelium. Morphologically wild-type clones show defects in the apical localization of atypical protein kinase C (aPKC) and Armadillo, which become either diffuse or ectopically localized along lateral membranes. In less severely affected cells, the apical localization is discontinuous. These phenotypes are penetrant in large stem-cell clones but not in small clones, indicating that LKB1 activity perdures. Expression of the wild-type or S535E transgenes with arm-GAL4 in lkb1 mutants rescues these follicle cell phenotypes, whereas expression of S535A rescues lethality but gives rise to a completely disorganized follicular epithelium, in which most cells appear unpolarized. Thus, LKB1 is required for cell polarity in the germ line and the follicle cells, and is probably regulated by phosphorylation on the conserved C-terminal serine in both processes (Martin, 2003).

In Drosophila, par-1 and lkb1, the homologue of C. elegans par-4, show very similar phenotypes. In addition, LKB1 is an in vitro substrate for PAR-1 and can suppress the polarity phenotype of par-1 mutants when overexpressed. These results suggest that LKB1 functions downstream of PAR-1. This conclusion is consistent with genetic data in C. elegans that show that par-4 mutants display only a subset of the par-1 A-P polarity phenotypes. Notably, mutants in par-4 and par-1, but not in other par genes, show a disappearance of P granules in the one-cell zygote (Kemphues, 1988). Thus, LKB1 and PAR-1 function in a conserved pathway that is required for the polarization of the A-P axis in both worms and flies (Martin, 2003).

Drosophila LKB1 is closely related to human LKB1, and conserved prenylation and PKA phosphorylation sites are essential for the in vivo function of both proteins, indicating that the two are likely to be functional homologues. Mutants in LKB1 cause Peutz-Jeghers syndrome (Hemminki, 1998; Jenne, 1998), which is characterized by the formation of intestinal polyps and a high incidence of adeno-carcinomas (tumours of epithelial origin) (Hemminki, 1999). In addition, mutations in lkb1 have been identified in several sporadic epithelial cancers (Sanchdz-Cdspedes, 2002). The role of LKB1 as a tumour suppressor is not well understood, however, and LKB1 has been proposed to regulate apoptosis, the cell cycle or angiogenesis (Yoo, 2002; Bardeesy, 2002). In addition, LKB1 seems to function in a context-dependent manner that is different from classical tumour-suppressor genes such as ras or p53. Given that Drosophila lkb1 is required to polarize the epithelial follicle cells, an alternative model is proposed, that loss of LKB1 leads to polyp and tumour formation by disrupting epithelial polarity (Martin, 2003).

GENE STRUCTURE

cDNA clone length - 2573

Bases in 5' UTR - 490

Exons - 6

Bases in 3' UTR - 379

PROTEIN STRUCTURE

Amino Acids - 567

Structural Domains

Using a mapping strategy based on single-nucleotide polymorphisms, a complementation group that abolished the posterior localization of both Staufen and oskar was localized to a 46-kilobase (kb) fragment in 87F. The sequencing of candidate genes showed that both alleles contain lesions in CG9374 that are predicted to be null mutations. A genomic rescue construct rescues both lethality and localization of oskar mRNA to the posterior of the oocyte, confirming that these phenotypes are caused by mutations in this gene. CG9374 encodes a serine/threonine kinase with homology to C. elegans PAR-4, a kinase involved in the localization of P granules to the posterior of the one-cell zygote and to the human tumour-suppressor LKB1 (Hemminki, 1998; Jenne, 1998). Given the stronger homology to the latter, the Drosophila gene was named lkb1 (Martin, 2003).

The centrally located kinase domain of LKB1 has 44% and 66% amino-acid identity with PAR-4 and human LKB1, respectively, but the amino- and carboxy-terminal regions are not well conserved. Drosophila and human LKB1 share a C-terminal prenylation motif, however, and the latter is prenylated in vivo, which is necessary for its localization to the plasma membrane. Mutation of the prenyl-acceptor cysteine to alanine (C537A) in Drosophila GFP-LKB1 results in a cytoplasmic accumulation of the protein, showing that Drosophila LKB1 is targeted to the cell cortex through prenylation. This cortical localization is essential for LKB1 function, because UAS-GFP-LKB1^C537A does not rescue the lkb1 phenotype when expressed in near-endogenous amounts. This construct can, however, partially rescue the posterior localization of Staufen when it is expressed in tenfold higher amounts. In agreement with this observation, the LKB1^C537A mutant shows a weak residual membrane localization, which may be sufficient to mediate LKB1 function when overexpressed. Thus, cortical localization is important for LKB1 activity, consistent with its role in organizing microtubules along the lateral and posterior cortex (Martin, 2003).

date revised: 22 June 2006

Home page: The Interactive Fly © 2006 Thomas Brody, Ph.D.

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