BIOLOGICAL OVERVIEW

The Drosophila ovarian tumor gene (otu) encodes a novel cytoplasmic protein crucial to a variety of processes and cells active during oogenesis. Otu ensures the survival of female germ cells in pupae, cyst formation in germ-line cells, the attainment of mature chromosome structure in nurse cells, and egg maturation. Otu function is thought to involve cytoskeletal function, but it is unknown whether Otu is part of the function of the tubulin based microtubule cytoskeleton or the actin based microfilament cytoskeleton. This essay will outline the various functions of Otu and consider evidence of Otu's involvement in either microtubular or microfilament function.

During oogenesis, germ-line cells divide four times to give rise to germ-line cysts, each containing 16 interconnected cells known as cystocytes. 15 of the cells differentiate into nurse cells, which synthesize and transport products required for the development of the remaining cell, the oocyte. The otu gene has long been thought to play a key role in cyst development and growth based on the analysis of three classes of mutant alleles (Storto, 1988). In the presence of quiescent (QUI) alleles, germ-line cells are absent; this has a negative impact on germ cell survival. In contrast, germ cells overproliferate to form benign tumors in the presence of oncogenic (ONC) alleles. When differentiating (DIF) alleles are present, nurse cells display abnormal chromosome condensation, fail to grow normally and do not fully transfer their contents to the oocyte.

The earliest developmental defect associated with otu-null mutants is manifest during the pupal stage in lower survival rates for female germ cells. This otu function is shared with another gene, the transcripition factor ovo, which, when missing, produces ovarian tumors. Mutant phenotypes for females lacking ovo function show up even earlier in development, at the end of the first larval instar. At this stage, a minority of mutant gonads have lost their germ cells. This is in contrast to the phenotype for females lacking otu. Such females have a wild type number of germ cells through all larval stages. In these animals most of the germ cells die during pupariation, although some do survive and divide, but do not differentiate within the adult ovary. Since ovo and otu are required for the survival of XX germ cells, they must control a vital, sex-specific process in these cells. Do the two genes control the same process? Maternal ovo transcripts are present in pole cells (which are prospective germ cells) until embryonic stage 9. In the case of otu, maternal transcripts are also found in 0-4 hour embryos (Geyer, 1993). If required this early in development, maternal ovo and otu products might be thought to assure the survival of XX germ cells before these genes are transcribed zygotically. In this scenario, a vital process served by maternal ovo has already taken place by the end of the first larval instar (Staab, 1996).

With respect to the ovarian tumor phenotype, several genes have been characterized in Drosophila that carry a common defect in oogenesis. In addition to ovarian tumor (otu), these include ovo, benign gonial cell neoplasm, sans fille, Sex lethal, fused and, bag of marbles. Mutations at these loci result in the absence of mature germ cells, and in the overproliferation of small cells with the morphological characteristics of undifferentiated germ cells. Keeping this in mind, the possibility that Otu affects the actin based cytoskeleton will be considered.

A prominent feature of germ cell differentiation is the formation of the cytoplasmic bridges connecting the cystocytes. In oogenesis, germ line stem cells contain a spectrin-rich body known as the spectrosome. As the cystoblast divides, the spectrosome gives rise to an elongated structure known as the fusome, which extends through the ring canals that connect cystocytes to form branching connections between daughter cystocytes. (Cystocytes normally do not undergo complete fission but remain connected via ring canals). Some components of the fusome have been identified, including adducin (the product of the hu-li tai shao gene, a protein that promotes actin assembly), alpha-spectrin and Bag-of-marbles (see alpha Spectrin for a discussion of fusome function).

A comparison was made between the structure of wild-type germaria and the germaria of flies mutant for either strong or weak alleles of otu. The otuPdelta1 allele constitutes a deletion of the otu coding region. Mutant flies are female-sterile and produce either agametic ovarioles or tumorous egg chambers. The otuPdelta1 mutation causes enlargement of germarial region I, resulting from an increased number of germ cells arrested at stages prior to when they would normally interact with follicle cells. Severe hypomorphic alleles of otu form primarily tumorous egg chambers that superficailly resemble the tumors produced by the otuPdelta1 null mutations. However, both hypomorphic alleles otuPdelta3 and otu13 mutant germ cells can undergo further differentiation than do otuPdelta1 mutants; the former two appear to reach stages of development associated with region II of the ovarium, a stage when germ cells normally interact with the migrating follicle cell layer. In normal ovaries, no actin, as revealed by phalloidin staining is apparent in the fusome. The presence of detectable actin filaments in spectrosomes but not fusomes suggests that changes in actin filament localization or accessability normally occur during the spectrosome-to-fusome transition. As with wild type, otuPdelta3 mutants contain spectrin; the cystocytes still have high levels of cortical actin filaments. However, in contrast to wild type, substantial levels of actin filaments were detected in all mutant fusomes examined. These data indicate that otu is required for normal fusome structure, in particular actin filament organization or levels during the conversion of a spectrosome to a fusome (Rodesch, 1997).

Differentiate (DIF) otu alleles disrupt actin organization in stage 10 egg chambers. The cytoplasmic dumping that occurs during stage 10 is believed to require subcortical actin for nurse cell contraction and a complex array of cytoplasmic actin filaments, which anchor the nurse cell nucleus to the plasma membrane. Nurse cells mutant for the DIF allele otu7 fail to form this cytoplasmic actin array. Even in hypomorphic alleles there is a major disruption in the cytoplasmic actin cytoskeleton, while cortical actin appears unaffected. However, in addition to the loss of actin fibers connecting the nucleus to the plasma membrane, actin filaments are observed around the nucleus and partially extending to the plasma membrane. It therefore appears as if otu mutations cause actin polymerization to initiate from the nucleus (Rodesch, 1997).

A different picture of otu function is apparent when considering the gene cup and its effects on meiotic chromosome segregation. In a screen for female sterile P element insertions that effect nurse cell chromosomes, one mutation in the collection (fs(2)04506) proved to be allelic with a previously described female sterile gene, fs(2)cup. In most cup alleles nurse cell chromosomes fail to decondense completely; therefore, individual chromosomal masses remain distinct. In many respects cup phenotypes appear to be remarkably similar to otu phenotypes. As with otu, cup alleles can be grouped into three general classes based on the stage at which their oogenesis is arrested. Class I, the largest set of alleles, causes egg chambers to arrest prior to vitellogenesis. The strongest class I alleles produce enlarged and misshapen germaria where the egg chambers sometimes bud abnormally. Egg chambers from class II females grow larger than class I egg chambers, taking up yolk and sometimes supporting follicle cell migration. However, nurse cell nuclei display abnormal chromatin configurations. Females from class III, the weakest group of cup alleles, produce defective mature eggs, characteristically shaped like cups. Egg chambers that will develop into cup-shaped eggs have normal proportions during early stages, but during stages 9 and 10, their oocytes reach only a quarter to half the size of corresponding wild-type oocytes. The decreased size of the cup oocyte relative to its nurse cell suggests that the transport of materials from the nurse cells into the oocyte is reduced during stages 9 and 10. cup codes for a novel cytoplasmic protein, which accumulates early in the future oocyte within 16-cell cysts of the germarium but which later, during oogensis, is found in large aggregates in nurse cells, around the periphery of the nuclei. Still later these aggregates are found to disperse throughout the nurse cell cytoplasm, still in large aggregates, which eventually move toward the cellular periphery to form particulate structures along the subcortical surface of the nurse cells (Keyes, 1997).

Several observations suggest that cup and otu might interact. Late arresting alleles of otu affect condensation of the nurse cell chromosomes and partially prevent nurse cells from transfering their contents into the oocyte. The distribution of Otu protein also parallels Cup expression: Otu p104 becomes enriched in early oocytes and both p104 and Otu p98 move toward the periphery of the nurse cells during oocytic stages 9-10. There is a strong genetic interaction between cup and otu. In otu;cup double homozygous combinations, neither gene is epistatic to the other. Instead, the double mutant combinations produce defects that are much more severe than either alone. One finds further evidence for an interaction between cup and otu in females homozygous for cup but heterozygous for otu. A single copy of otu11 had a strong dominant effect on three of the four cup alleles tested. In the presence of one copy of otu11, the fertility of females homozygous for cup15 is restored. The ability of a reduced dose of otu to rescue an intermediate cup allele suggests that the balance of these gene products is important (Keyes, 1997).

Despite the diversity of the effects resulting from perturbations in cup function, many of them are known to involve microtubules. The distribution and relocalization of Cup protein during egg chamber maturation suggests that it is directly or indirectly associated with microtubules, at least during the previtellogenic stage. The ability of cup to disrupt chromosome segregation suggests that cup may also affect the ability of chromsomes to associate with the meiotic spindle (Keyes, 1997).

Neither of the above studies provides definitive proof of an interaction between Otu and either the actin based or the tubulin based cytoskeleton, but they do suggest an association. Because of the intimate associations of the two cytoskeletal systems during oogeneis (see cytoskeleton), a complete understanding of Otu's roles in oogenesis awaits a more thorough understanding of Otu's protein interactions.

GENE STRUCTURE

Sequence analysis of otu cDNAs suggests that the two Otu proteins (apparent molecular masses of 98 and 104 kD) are generated by alternative splicing of a 126-bp exon (6a) between the sixth and seventh exon of the smaller transcript (Steinhauer, 1992)

Transcript length - 4.1 kb for the 104 kDa isoform and 3.2 kb for the 98 kDa isoform

Bases in 5' UTR - 1332

Exons - 8

Bases in 3' UTR - 505+

PROTEIN STRUCTURE

Amino Acids - 811 and 853 for the 98 and 104 kD proteins respectively (Steinhauer, 1992)

Structural Domains

Otu is a proline-rich, hydrophilic novel protein (Steinhauer, 1989).

Otu protein C-terminal sequence contains two domains with a statistically significant homology to regions of human, rat and mouse microtubule associated proteins (MAP2s) (Tirronen, 1995).

Three potential structural motifs have been identified by analysis of the Otu sequence: an N-terminal cysteine protease domain, a central Tudor domain, and proline-rich motifs in the C-terminal region. While the amino acids that are unique to the large Otu isoform are part of an important functional domain, the nature of the domain and its boundaries have been unknown. This study has established that the region that encodes the putative Tudor domain, which was identified solely by sequence analysis, corresponds to the domain that is required for the early Otu function. Of the otu transgenes analyzed, those encoding Otu polypeptides that contain the Tudor domain homology region promote normal egg chamber differentiation through the early stages, whereas polypeptides that lack a portion or all of these sequences are defective in the early Otu function. Thus, the Tudor domain homology region is a functional domain of Otu (Glenn, 2001).

The biochemical function of Tudor domains is unknown. Ten Tudor domain repeats are found in the D. melanogaster Tudor protein, which is required for pole cell formation and normal abdominal development. One copy of this domain is found in the protein encoded by the D. melanogaster homeless gene, which appears to function in RNA transport and localization during oogenesis, and Tudor domains are also found in other proteins that interact with RNA. However, the Tudor domain located in the central region of SMN, a snRNP assembly factor required for survival of motor neurons, appears to mediate protein/protein rather than protein/RNA interactions. The Tudor domain of Otu does not appear to be required for the Otu/mRNP interaction because truncated Otu polypeptides, missing this domain, were affinity-selected with oligo-dT. However, oligo-dT selection experiments were performed in the presence of wild-type Otu, and it is possible that Otu oligomerizes via a domain in the N-terminal region. If so, the truncated proteins may have been oligo-dT selected due to an association with endogenous Otu, and full-length Otu, but not the truncated proteins, could be responsible for the observed mRNP interaction (Glenn, 2001).

As yet, the relevant functional domain(s) in the N-terminal and C-terminal regions of Otu have not been delineated. The only discernable feature of the C-terminal Otu region is that it is rich in proline residues. It is plausible that these proline-rich motifs comprise a functional domain because the deletions in otu¹⁴, otu-423, and otu-627, all of which are defective in the late Otu function, remove some or all of these motifs. The N-terminal region contains the cysteine protease domain. However, Otu may not be an active protease because an essential cysteine residue, found in the catalytic site of other members of the cysteine protease family, is replaced by a serine in the homologous region of Otu. If it lacks protease activity, the cysteine protease homology region in Otu may be a protein/protein interaction domain. The N-terminal region is also responsible for the Otu/mRNP association. Given that no known RNA binding motifs are found in Otu, the mRNP association of Otu is most likely indirect. For example, it is possible that Otu interacts with an RNA binding protein via amino acids in the N-terminal region, using the cysteine protease domain or a separate domain (Glenn, 2001).

The modification of cellular proteins by ubiquitin (Ub) is an important event that underlies protein stability and function in eukaryotes. Protein ubiquitylation is a dynamic and reversible process; attached Ub can be removed by deubiquitylating enzymes (DUBs), a heterogeneous group of cysteine proteases that cleave proteins precisely at the Ub-protein bond. Two families of DUBs have been identified previously. This study describe new, highly specific Ub iso-peptidases, that have no sequence homology to known DUBs, but which belong to the OTU (ovarian tumour) superfamily of proteins. Two novel proteins were isolated from HeLa cells by affinity purification using the DUB-specific inhibitor, Ub aldehyde (Ubal). These proteins were named otubain 1 and otubain 2, for OTU-domain Ubal-binding protein. Functional analysis of otubains shows that the OTU domain contains an active cysteine protease site (Balakirev, 2003).

date revised: 11 June 2022

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.