Two otu transcripts are selectively localised in Drosophila oogenesis by a mechanism that requires a function of the otu protein

ELSEVIER

Mechanisms

of Development

52 (1995) 65-75

Two otu transcripts are selectively localised in Drosophila oogenesis by a mechanism that requires a function of the otu protein Mika Tirronewb,

Vesa-Pekka

Lahtib, Tapio I. Heinob, Christophe

Roosa.b,*

%stitute of Biotechnology, FiN-OOOi4, University of Helsinki. Helsinki, Finland bDepartment of Biosciences, Division of Genetics, P.O. Box 56, FIN-00014, University of Helsinki, Helsinki. Finland Received

11 January

199.5; revision received 5 April 1995; accepted

7 April I 995

Abstract The ovarian tumour gene (otu) is required for several processes during Drosophita oogenesis. The locus encodes two protein isoforms that have been proposed to act during different stages of oogenesis. Here we show that the corresponding otu mRNAs display a dynamic pattern of expression during oogenesis. The 4.1 kb mRNA encoding the 104 kDa isoform is expressed throughout adult oogenesis, but is mainly restricted to nurse cells. The 3.2 kb mRNA encoding the 98 kDa protein isoform is selectively localised in the oocyte up to stage 9. Both mRNAs are expressed abundantly in nurse cells at stages 10-11. We propose that the oocyte-specific function of otu is realised by the 98 kDa isoform. We show that the export of the 3.2 kb mRNA from the nurse cell nuclei requires a functional otu protein. The otu protein is also required for the correct distribution of the pumilio and oskar mRNAs, while the &c-D, KIO and stuufen mRNAs are localised in wild type fashion in otu mutants. Furthermore, we have observed a region of homology between the carboxy-terminal part of the otu protein and the mammalian microtubule associated proteins. The more severe the mutation

in this region of homology, the more disturbed mRNA distribution is observed in otu mutants. Keywords:

Ovarian tumour;

Oogenesis; mRNA processing; Microtubule; Polytenisation; Cell cycle; MAP2

1. Introduction In Drosophila, oogenesis offers an excellent model system for studying the genetic control of cell division and differentiation. Oogenesis begins with the mitotic activation of the stem cell, which lies at the anterior tip of each ovariole. The daughter of the stem cell, a cystoblast, undergoes four mitotic divisions to produce a cyst of 16 sister cells. These cells, termed cystocytes, remain interconnected by cytoplasmic bridges that ensue from incomplete cytokinesis during the mitotic divisions. The bridges become modified into specific actin-rich structures termed ring canals (Yue and Spradling, 1992; Xue and Cooley, 1993; Robinson et al., 1994). One of the two cells containing four canals will develop into an oocyte, while the remaining 15 differentiate into nurse cells. These determination events occur in a specific compart-

* Corresponding author, P.O. Box 45, FIN-00014 University of Helsinki, Helsinki, Finland. Tel.: +358 0 434 6022; Fax: +358 0 434 6028; E-mail: [email protected].

0925-4773/95/$09.50 0 1995 Elsevier Science Ireland Ltd. All rights reserved SSDl 0925-4773(95)00390-M

ment at the anterior tip of each ovariole, the germarium. Before the 16-cell cyst leaves the germarium, it is surrounded by a layer of somatically derived follicle cells. The continuous post-germarial maturation of the egg chamber within the ovariole has been subdivided into 14 stages according to the subcellular structure and proportional size of the germ cells (Cummings and King, 1969) (for a complete description of oogenesis, see King, 1970; Mahowald and Kambysellis, 1980). The otu gene is one of several genes whose products are required to regulate the differentiation of the cystoblast into an oocyte and 15 nurse cells. Females homozygous for mutant alleles of otu are sterile and display a variety of abnormalities in cystocyte division and differentiation (Storto and King, 1988). The otu alleles have been divided into three major classes on the basis of their most common phenotypes (King et al., 1986; Storto and King, 1988). The quiescent (QUI) class contains alleles that lead to defects in oogonial proliferation, and produce chamberless ovarioles. Alleles in the oncogenic (ONC) class produce ‘ovarian tumours’, which are egg chambers

66

M. Tirronen et al. I Mechanisms of Developmeni 52 (1995) 65-75

containing hundreds of undifferentiated germ cells. Those alleles that allow the cystocytes to differentiate as nurse cells and where egg chambers might even include an oocyte belong to the differentiated (DIF) class. Some heteroallelic combinations are fertile, which suggests that the otu products have several functional domains (King et al., 1986). The otu gene has been proposed to act in several processes, for example; in germline sex determination (Bopp et al., 1993; Oliver et al., 1993; Pauli et al., 1993; Wei et al., 1994; Nagoshi et al., 1995; for a review, see Steinmann-Zwicky, 1992a,b), in germline dosage compensation (Oliver et al., 1993), and in cytoskeletal reorganisations within the syncytial cysts (Storto and King, 1988; Steinhauer and Kalfayan, 1992; Bopp et al., 1993; Lin et al., 1994; see also Theurkauf et al., 1992, 1993). Although Ott has been cloned (Mulligan et al., 1988; Chambe and Laird, 1989; Steinhauer et al., 1989), its functions are not yet understood. The gene encodes two ovarian transcripts (3.2 kb and 4.1 kb) that are produced by alternative splicing of a common pre-mRNA (Mulligan et al., 1988; Steinhauer and Kalfayan, 1992). The two protein products diverge by a stretch of 42 amino acids encoded by an extra exon present in the longer mRNA. The otu proteins are expressed in the cytoplasm of germ cells throughout oogenesis (Steinhauer and Kalfayan, 1992). Based on genetic data it has been proposed that the larger, 104 kDa isoform, is required for early differentiation and division control of germ cells (Steinhauer and Kalfayan, 1992; Bae et al., 1994). The 98 kDa isoform has been proposed to act in early stages of germ line sex determination (Nagoshi et al., 1995) and during late stages of oogenesis (Steinhauer and Kalfayan, 1992; Bae et al., 1994). However, the specific expression patterns of the otu products could not be distinguished (Comer et al., 1992; Steinhauer and Kalfayan, 1992). We now report the specific expression pattern of the two otu transcripts during oogenesis. We also show that the otu function is required for the export of nuclear mRNAs, including the otu mRNA.

Probes:

-

P3+4 P4

n

exons 1

234

5

6

6a

7

6

Genomic

DNA

3.2 kb mRNA

Fig. 1. Origin of otu clones used for in situ hybridisations. The probe ‘P4’ derives from the extra exon present only in the 4.1 kb mRNA, the probe ‘P3 + 4’ recognizes both the 3.2 kb and the 4.1 kb mRNA. Each box represents one exon in the genomic sequence of otu (EMBU GenbanWDDBJ accession number M30825; Steinhauer and Kalfayan, 1992). The dashed line at the end of the 4.1 kb mRNA demonstrates that the 3’ end of the longer mRNA is unknown (Steinhauer and Kalfayan, 1992).

2. Results 2.1. Distribution of otu transcripts during oogenesis We analysed the expression pattern of the 4.1 kb otu mRNA using a probe derived from the extra exon 6a (Fig. 1). The 4.1 kb otu mRNA is expressed throughout oogenesis (Fig. 2A,C). The signal was most prominent in nurse cells with little or no signal in the oocyte. A detectable signal could be observed in the ooplasm at stages 89 only if the staining was prolonged extensively (Fig. 2A). At stage 10, an intense signal was seen in nurse cells (Fig. 2C). The signal persisted until stage 11, when the nurse cells degenerated. By using a probe common to both otu transcripts (Fig. l), we detected, in addition to the already described pattern, a signal that was not observed with the probe specific to the 4.1 kb mRNA. We conclude that this signal is specific to the 3.2 kb mRNA. From the germarial region 2A up to stage 9 we observed a specific signal in the oocyte (Fig. 2B). At stage 9, the signal dispersed, transiently formed large aggregates, and then disappeared. At stages 10-l 1, an abundant signal was detected in nurse cells (Fig. 2D). In germarial region 2A, which contains two syncytial cysts, the signal was detected in two cells in the middle of each of the syncytial cysts (Fig. 2E). In region 2B, which contains only one cyst, a single cell expressed the signal, and in region 3 (stage 1) the otu expressing cell could be identified as an oocyte (Fig. 2F). In addition to the oocyte-specific accumulation of the signal, the probe common to both transcripts detected a small accumulation in the nucleus of each nurse cell. This signal was observed from germarial region 2 onwards. The accumulations grew in size and reached their maximal density at postgermarial stage 3, whereafter they disappeared (Fig. 2F,G). Similar mRNA dots have been observed with other probes for example within nuclei of embryonic cells, and they have been proposed to represent nascent RNAs at the site of transcription (O’Farrel et al., 1989; Shermoen and O’Farrel, 1991). The appearance of nuclear 3.2 kb otu mRNA accumulations correlated with increased chromosome polyteny. In the wild type nurse cell nuclei the chromosomes become polyploid after the cystocyte divisions are completed in germarial region 2. We observed that the mRNA accumulations grew in size as the polyploidy increased. The homologous chromosomes remain conjoined up to stage 4, whereafter the strands dissociate (Mulligan and Rasch, 1985; Storto and King, 1988). At that stage, the nuclear 3.2 kb otu mRNA accumulations disappeared. We believe that the accumulations represent nascent transcripts associated with the template chromatin. 2.2. Localisation of otu mRNAs in otu mutants To study whether alterations in the otu function have an effect on the distribution of the otu transcripts, we analysed the expression of otu in females which have a

M. Tirronen et al. I Mechanisms of Development 52 (1995) 65-75

SQ

Fig. 2. Expression pattern of the otu mRNAs in Drosophila ovaries. (A) The expression of the 4.1 kb mRNA as observed by the probe specific to the exon 6a (‘P4’). The signal is observed from the tip of the germarium onwards. (B) The signal pattern observed by a probe recognising both mRNAs (‘P3 + 4’). A specific signal accumulates in the oocyte (arrows). The alkaline phosphatase staining reaction was shorter in (B) than (A). (C) At stage 10, a heavy nurse cell signal is observed by the probe ‘P4’. (D) At the same stage, the probe ‘P3 + 4’ detects a prominent signal in the nurse cells. Either little or no signal is in the ooplasm. (E) In the germarium, the 3.2 kb rnRNA, as observed by the probe ‘P3 + 4’, appears to accumulate within two cells in the middle of each of the two adjacent cysts in region 2A (arrows). (F) In germarial region 3, the accumulated signal restricts to a single cell at the posterior pole of the cyst (big arrow). Note the small mRNA aggregates associated with the nurse cell nuclei (small arrows). The accumulated signal in region 2 is out of the focal plane. (G) The nuclear-associated mRNA aggregates (arrows) reach their maximal density at postgermarial stage 3. Anterior to the left.

truncated otu protein. In otu-‘, a base substitution leads to a premature stop codon and to a truncation of 190 amino acids from the carboxy-terminal end of the otu proteins (Steinhauer and Kalfayan, 1992). In otd4, a base substitution leads to a truncation of 156 amino acids from the carboxy-terminal end (Steinhauer and Kalfayan, 1992). We included a third allele, otu’, in these analyses, although it has not been characterised on a molecular level. The mutant phenotype of otu7 is reminiscent of that of otd and otd4. All these alleles belong to the ‘differentiated’ class, and produce mostly egg chambers with

nurse cells and an oocyte. The molecular lesions and the phenotypes of the selected otu mutants are summarised in Table 1. With the probe specific to the 4.1 kb mRNA we did not observe any alterations in the mutant egg chambers besides some rare aggregates in a few nurse cell nuclei at stage 10 (Fig. 3A). These aggregates were observed in otu5 and otu7 mutant egg chambers, but not in otuz4 mutants. The probe common to both transcripts detected a clearly disturbed pattern of mRNA distribution in all these mutants. We deduced that this pattern was specific

68

M. Tirronen et al. I Mechanisms

of Development

52 (1995) 65-75

Table 1 Summary Allele

of the different Class

properties

of the otu mutants used (see also Stotto and King, 1988; Steinhauer

Molecular

lesion

and Kalfayan,

1992)

otu expression 4.1 kb mRNA

3.2 kb mRNA

002

DIF

C to T substitution at position 4018: introduction of a stop codon, 190 amino acid truncation after residue 621

Almost normal, faint aggregates in nurse cell nuclei at stage 10

Severely disturbed. No oocyte signal, large aggregates in nurse cell nuclei, nurse cell cytoplasmic signal reduced

otu7

DIF

Not known

As in otu’

As in otd

0tlP

DIF

C to T substitution at position 4120: introduction of a stop codon, 156 amino acid truncation after residue 655

Normal

Oocyte signal weakened, aggregates in some nurse cell nuclei, nurse cell cytoplasmic signal reduced

otu’3

ONC

G to A substitution at position 2779: alteration of the splice acceptor site of exon 6a

No signal

Oocyte signal weakened, normal

to the 3.2 kb mRNA. In otu5 and otu7 mutant egg chambers, heavily stained mRNA accumulations were seen associated with the nuclei of nurse cells (Fig. 3B,D). From germarial region 2 up to stage 3 egg chambers, the accumulations were observed as in the wild type. But, while they disappeared in the wild type after stage 3, they grew in size throughout oogenesis in otd and otu7 mutant egg chambers. A drastic reduction in the cytoplasmic staining of the nurse cells was associated with the nuclear mRNA accumulations. This reduction was most significant in vitellogenic egg chambers (Fig. 3B). Moreover, we did not detect any oocyte specific accumulation of the 3.2 kb mRNA, although a morphologically distinguishable oocyte was formed. It appears that instead of being exported into the nurse cells and the oocyte, the 3.2 kb mRNA is sequestered within the nuclei of the nurse cells in otu5 and otu7 mutants. In otd4 mutant ovaries, we observed the 3.2 kb otu mRNA in the oocyte, but the signal was significantly reduced as compared to the wild type. At the same time, weak accumulations of the mRNA in nurse cell nuclei were observed (Fig. 3E). The cytoplasmic staining of the 3.2 kb mRNA in nurse cells in some stage 10 egg chambers was reduced. Taken together, these results suggest that the otu protein is required for the nuclear export of the 3.2 kb otu mRNA. In ofu5 mutants, where the carboxy-terminal

otherwise

truncation of the otu protein is longer than in otd4, mRNA export was more severely disturbed.

the

2.3. The pattern of nuclear mRNA accumulations correlates with the chromatin structure The nuclear mRNA accumulations in otu mutants were observed only in the close vicinity of the nuclear membrane and the chromosomal material within the nuclei. Interestingly, the pattern of accumulations was different in otu7 and otd mutant egg chambers. In otu7 mutants, the 3.2 kb otu mRNA was observed only as a single aggregate in each of the nurse cell nuclei (Fig. 4A). Each aggregate was attached to one of the polytene chromosomes (Fig. 4B). In otu5 mutants, the signal was sometimes observed as several aggregates within nurse cell nuclei (Fig. 3C). In such cells the polyploid chromosomes were dispersed and not polytenised (Fig. 4D) (Storto and King, 1988; Tapio Heino et al., unpublished). Although our data does not rule out other explanations, the pattern of mRNA accumulation in these mutants strongly suggests that the aggregates represent nascent otu transcripts at the site of synthesis on the chromatid template. The increased mRNA accumulation in nurse cell nuclei of otu mutants seems to be a direct continuation of the accumulation observed in the wild type nurse cells up to stage 3. In otu7 mutants the homologous chromosomes, that normally dissociate after stage 4, remain conjoined

Fig. 3. Expression of the ofu mRNAs in otu mutants. (A) The probe specific to the exon 6a (‘P4’) detects a wild type-like expression pattern in otu’ mutants. Occasionally, weak mRNA accumulations in stage 10 nurse cells can be detected (arrows). (B) In otu7 mutants, the probe ‘P3 + 4’ displays an RNA accumulation associated with each of the nurse cell nuclei, The oocyte-specific signal is not observed and the staining of nurse cell cytoplasm is reduced. especially m the eldest egg chamber (stage 10). (C) As a control, a wild type stage 10 egg chamber probed by ‘P3 + 4’ shows a prominent signal in nurse cell cytoplasm. (D) In od mutants (stage 6 egg chamber), the probe ‘P3 + 4’ detects mRNA accumulations associated with nurse cell associated with nuclei (arrows) and no signal in the oocyte. (E) In otu I4 mutants (stage 6 egg chamber), the probe ‘P3 + 4’ detects weak accumulations nurse cell nuclei (small arrow) and a weak oocyte-specific signal (big arrow). (F) In the wild type (stage 6 egg chamber), the probe ‘P3 + 4‘ detects only an oocyte-specific signal. (G) In otuJ3 mutants (stage 6 egg chamber), the probe ‘P3 + 4’ detects an oocyte-specific signal and a cytoplasmic staining in nurse cells. Anterior to the left.

M. Tirronen et ul. I Mechunisms

and become polytenised as the egg chambers mature. Accordingly, the mRNA accumulations keep growing as the polyploidy level increases. It has been shown earlier that long or abundant transcripts can be detected by in situ

of Development

52 (1995) 65-75

69

hybridisation as a band at the site of synthesis in the salivary gland chromosomes of the wild type larvae (Boyd et al., 1991; Zachar et al., 1993). However, it is important to note, that the cytoplasmic staining of the mRNA in the

M. Tirronen et al. I Mechanisms

of Development

5.2 (1995) 65-75

of the mRNA, as in otzd or otu7 mutants, was observed in otuJ3 mutant egg chambers. Moreover, in those egg chambers in which the oocyte was formed, a wild typelike signal was observed in the oocyte (Fig. 3G). These observations suggest that the 104 kDa otu protein is not essential for the mRNA localisation function of otu, and if it does participate in the localisation process, the 98 kDa isoform is able to substitute for it.

‘jFig. 4. The pattern of mRNA aggregates correlates with the chromosomal organisation (A) In otu7 mutant egg chambers, the mRNA aggregates are observed as a single band associated with each of the nurse cell nuclei. (B) Hoechst staining of the same nucleus shows that the mRNA aggregate (arrow) associates with one of the polytene chromosomes. (C) In otu* mutant egg chambers the mRNA aggregates are occasionally observed as several distinct granules. (D) In such nurse cells, the nuclear DNA is not organised as polytene chromosomes. The focal plane is different as in (C) to reveal the overall pattern of the nuclear DNA.

nurse cells is strongly reduced, and the oocyte-specific accumulation of the signal is completely absent in otu7 and otd mutants. Therefore, our data suggest that the mRNA accumulations in otu mutants are enhanced by aberrant export of the transcripts out of the nurse cell nuclei. 2.4. Requirement of the 104 kDa isoform for the otu mRNA distribution Both otu isoforms are likely to be expressed from the germarial stages onwards. However, their functions in the early stages of oogenesis seem to be distinct (Steinhauer and Kalfayan, 1992; Bae et al., 1994; Nagoshi et al., 1995). Which one of the two isoforms is required for the mRNA localisation function? To address this question, we analysed the otu expression in otd3 mutants. In otuJ3, the splice acceptor site of the extra exon 6a is altered by a base substitution. This leads to lack of the 104 kDa isoform (Steinhauer and Kalfayan, 1992). Most of the otuJ3 egg chambers develop as ‘tumorous’. Some egg chambers, however, produce nurse cells, and a minority of these egg chambers also produce an oocyte (Storto and King, 1988). With a probe specific to exon 6a, we did not observe any signal in otuJ3 mutant egg chambers (not shown). In contrast, the probe specific to both mRNAs detected a clear staining pattern (Fig. 3G). No nuclear accumulation

2.5. The otu function is required for the localisation of several mRNAs The alterations in the otu mRNA localisation in the aforementioned otu mutants could, in principle, be due to some not characterised lesions in the oru transcripts themselves (see Steinhauer and Kalfayan, 1992). If this was the case, the otu mutations would not have a similar effect on other mRNAs expressed in nurse cells. To test this, we studied the expression of several other mRNAs in otu’ mutant ovaries. Bicaudal-D (Suter et al., 1989; Suter and Steward, 1991) bicoid (Berleth et al., 1988), KZO (Haenlin and Roos et al., 1987; Cheung et al., 1992) oskar (Ephrussi et al., 1991; Kim-Ha et al., 1991) pumilio (Macdonald, 1992) and staufen (St Johnston et al., 1991) are needed either in the oocyte or in the egg, where they are transported after synthesis in nurse cells. We have shown earlier that the Bicaudal-D, bicoid and KIO mRNAs are localised in the wild type fashion in otu5 mutant egg chambers (Tirronen et al., 1992). Now, we extended these analyses to oskar, pumilio and staufen. We observed that the staufen mRNA is expressed in otd mutant egg chambers as in the wild type (data not shown). The expression of the oskar mRNA, however, was altered. The posterior pole localisation of the oskar mRNA that is normally observed in the oocyte after stage 9 (Fig. 5A,B) (Ephrussi et al., 1991; Kim-Ha et al., 1991) was reduced and diffuse (Fig. 5C,D). This alteration was even more severe in egg chambers mutant for otu7. In both of these mutants the early transport of the oskar mRNA from the nurse cells to the oocyte, which normally takes place during previtellogenic stages (Fig. 5A), was observed as in the wild type (Fig. SC). These observations suggest that the otu function is required after stage 9 for the translocation of the oskar mRNA from the anterior margin of the ooplasm to the posterior pole. Nevertheless, the otu5 mutation has different effects on oskar and otu mRNA: the former being mislocalised in the ooplasm, the latter being sequestered in nuclei of the nurse cells. Although the mislocalisation of oskar is informative, it does not let us rule out the hypothesis according to which otu mRNA mislocalisation would ensue from conformational changes in the mRNA itself. The distribution of the pumilio (pum) mRNA was also altered in otu5 and oru7 mutant egg chambers. In the wild type, the pumilio mRNA is expressed in germ cells from the germarium onwards until stages 3-4, whereafter the

M. Tirronen et al. I Mechanisms

71

of Development 52 (1995) 6.5-75

Fig. 5. Effect of mutations in otJ on the localisation of specific mRNAs. (A) the wild type egg chambers up to stage 9 (S9). The signal is observed in the oocyte. At stage 9 it localises to the posterior pole of the ooplasm (arrow). (B) In stage 10 egg chamber, a prominent signal is at the posterior pole (arrow). (C) Expression of the oskar mRNA in otu’ mutant egg chambers up to stage 9. Movement of the mRNA from the posterior-most nurse cells to the oocyte seems to be inefficient at stage 9 (small arrow). At the same stage, the signal gets dispersed in the ooplasm and it partially moves to the posterior pole (big arrow). (D) In stage 10 egg chamber of the ofus mutant, the oocyte specific signal at the posterior pole is strongly reduced (arrow). (E) Expression of the pumilio mRNA in the wild type..At stage 10, the signal is abundant in nurse cells. (F) Expression of the pumilio mRNA in oruS mutant. The mRNA accumulates in the nuclei of nurse cells (arrows), and the cytoplasmic staining is reduced. Anterior to the left.

mRNA level is reduced. At stage 9, the expression begins again, but then it is restricted to the nurse cells (Fig. 5E). Upon nurse cell degradation at the end of oogenesis, the pumilio transcript is transferred into the oocyte together with the contents of the nurse cells (Macdonald, 1992). In otu5 and otu7 egg chambers, the pumilio transcript accumulated as densely staining aggregates in nurse cell nuclei (Fig. 5F). This was observed from stage 2 onwards. The pattern of accumulations was different between otu5 and otu7 egg chambers in a similar manner as observed for the otu mRNA. In otu14 the localisation of the pumilio mRNA was similar to normal. Taken together, our results strongly suggest that the otu protein is required for the localisation of several mRNAs within the syncytial cysts. Localisation of the otu, pumilio and oskar transcripts is dependent on the function of otu, while the localisation of the BicaudalD, KIO, bicoid and staufen mRNAs is not. The fact that the requirement of the otu and pumilio mRNAs for the otu function is different from that of oskar sup-

ports the view of the multiple function.

requirements

of the otu

2.6. The otu proteins show local homologies to the mammalian microtubule associated proteins Since otu’ and otu14 mutations are known to induce deletions in the carboxy-terminal region of the protein, we wanted to test if any entry in the SwissProt or PIR data bases would display homology to the missing domains. Earlier studies have revealed a weak homology between the otu protein and the Drosophila bag-ofmarbles (barn) protein (McKearin and Spradling, 1990). The barn function is not known, but mutations in barn generate tumours in the germline (McKearin and Spradling, 1990). In this sense, the barn mutant phenotype is related to that of the ‘tumorous’ alleles of otu. However, while the barn function is required in both male and female germline, the otu alleles do not affect male germ cells (King et al., 1978; Tirronen et al., 1993). The homology between barn and otu is intriguing, but it does not

12

M. Tirroncn et al. I Mechanisms

A

of Development

52 (1995) 65-75

D Alignment 227 33 277

protein

149

(i, reference:

otu-barn)

ILENYKLCNFHSTNGNQSINARKGGRLEIKNQEERKASGSSGHEPNDLLP I. .I: .I II.. I :Il..Il.: ILHST"LRPRSKbP~G"LtTGPKaKQLQK~N"~NRKSKGSASADNI~KLP MCPNRLESCVRQ :. ::I. .. ITIEKLHMIGLH

288 : 160

Relative quality: Similarity (%): Identity (%):

B

r

(ii) ’ __d1 Alignment

-8

693

U-J

c--(iii, I,‘:

1735 743

~_ 1785

XiF2-s g_ z

632

:z-

1181

SPSSNGSQFSFYTTPSPHHHLITPPRLL I.11 :I. .: . ..I. SRSSVASPRRLSNVSSSGSINtLESPaL

I 2

400

200 5

161

I

Relative quality: Similarity (%): Identity (%):

33 T;

(iii)

682

VESTPPPSPEVANATEQSPLEKSAYAKRN 1.1:. /. .: ..,. :.

1231

IVSEPABVPSEEEEIEAGGEYDKLLFRSD

641

600

800

93 691

I3141

l..I

:

:.I:..

.I

:_:

Relative quality: Similarity (%): Identity (%):

29 T;

(iv)

r

otu protein , C

:

I

YNMGVDLHWRMSHHTPPDELGMFGYHQQNNTDQQAGRTVVIGAT~DNLTA : .I. . . . . . . :::::....::.I: . . ..:. CPPeVSSADLSTDEKGEVQMEFIQLPKEESTETPDIPA~PSCVTQPQPEA

Alignment

28 4;

(ifl

ANATEQSPLEKSAYAKRNLNSVKVRGKRPEQLQDIKDSLGPAAFLPTPTP ..I . ..I. . . ..l1: :.: : :. [..::.:I EKAQAKVGSLDNAHHVPGGGNVKIDSQKLNFREHAKARVDHGAEIITQSP

Alignment

II

:

7

18 I

1

I

143

stab codon stop codon

RMSHHTPPDELGMFGYHQQNNTDQQAGRTVVIGATEDNL~AVESTPPPSP ,: : ._::.:::: .I:.:.: . . . . . . . /.I:. RIVQVVTAEAVAVLKGEQEKEAQYKDQPAALPLAAEETAN~PPSPPPSPA EVANATEQSPLEKSA . II :..I SEQTATVEEDLLTRS

in otu14

in oh5

..I.I/..:

705 . 157

Relative quality: Similarity (%): Identity (%):

30 Tg

Fig. 6. Matrix comparison of the otu sequence to selected protein sequences and sequence alignments of statistically significant areas of homology (see Section 4). (A) Stretches of the otu protein share homologies with the bag-of-marbles protein (McKearin and Spradling, 1990). (B) The otu protein contains several stretches homologous to the mouse microtubule associated protein-2 (MAP2, database accession number P20357). The dots show the homologies found with the program Compare, while the diagonals emphasise those homologies that were classified as statistically significant with the program Bestfit. The boxed homologies are numbered with reman numbers and aligned in (D). The shaded area is used to help the reader in finding the corresponding region on the polypeptide shown in (C). (C) The exons (numbered from 2 to 8) making up the 104 kDa otu protein are shown as boxes and the stop codons introduced by two of the mutations are marked (the first exon is non-coding and the exon 6a is not numbered). (D) Amino acid alignment of the areas of homology between the otu carboxy-terminal end and the mouse MAP2 that were found to be statistically significant. The alignments have the same roman numbers as in (A) and (B). Bars display identities, colons high similarity and dots similarity. The quality of each alignment is shown as ‘relative quality ’ , ‘% identity’ and ‘% similarity’. The ‘relative quality’ takes both identities and similarities into account and also the length of the alignment, so it is a fair parameter to compare alignments inter se.

provide any explanation for the molecular function of either of the proteins. We found that the otu protein carboxy-terminal sequence contains two domains with a statistically significant homology to regions of the human, rat and mouse microtubule associated proteins (MAP2s) (Fig. 6). MAPS enhance microtubule assembly and they become bound to the newly formed microtubules (for a review, see Mandelkow and Mandelkow, 1995). The otd4 is missing one domain of homology, while the od is missing both domains. Interestingly, the phenotype and the pattern of mRNA distribution is more severely disturbed in otd than in otd4 mutants. These results suggest that the ho-

mology region is required for the mRNA localisation function of otu. However, the homology does not concern the MAP2 signature domain and it is not yet possible to assess the biological significance of the homology itself. Future analyses should reveal whether the otu protein directly associates with the microtubular structures of the germ cells. 3. Discussion Here, we report the specific expression pattern of the two otu mRNAs during Drosophila oogenesis. Previous genetic data suggested that otu encodes two products that

M. Tirronen et al. I Mechanisms oj’Development 52 (199.5) 6.5-75

are utilised at different stages of oogenesis (King et al., 1986; Storto and King, 1988). It was proposed that of the two otu protein isoforms, the 104 kDa isoform acts in early proliferation and differentiation of germ cells, while the 98 kDa isoform acts later (Steinhauer and Kalfayan, 1992; Bae et al., 1994). The 98 kDa isoform has also been shown to act in the sex determination of the female germ line (Nagoshi et al., 1995) In this report, our results show that the localisation of the two otu transcripts is selectively regulated. The 4.1 kb mRNA, encoding the 104 kDa isoform, is present predominantly in nurse cells, while the 3.2 kb mRNA encoding the 98 kDa isoform is localised in the oocyte up to stage 9, whereafter it appears abundantly in nurse cells. Both the 3.2 kb mRNA and the 4.1 kb mRNA become abundant in nurse cells at stage 10. Therefore, our in situ data suggest that the function of the 104 kDa is not restricted to early oogenesis. In otul’ mutants, where the 104 kDa otu isoform is specifically abolished (Steinhauer and Kalfayan, 1992), the egg chambers occasionally escape the ‘tumorous’ fate and form nurse cells and an oocyte. However, these egg chambers never reach maturity and arrest at pseudostage 14 (King et al., 1986; Storto and King, 1988). These observations suggest that the 104 kDa isoform is required, not only for the early division control of germ cells, but also during the late stages of oogenesis. From our in situ data we propose that the oocytespecific functions of otu are realised by the 3.2 kb mRNA and the corresponding 98 kDa otu isoform. The oocytespecific accumulation of the 3.2 kb mRNA was altered in otu5, otu’ and otu’” mutants. In otu5 and otu’ mutants, the oocyte specific accumulation of the 3.2 kb mRNA was completely blocked. It has been shown that in all these mutants the oocyte maturation is abnormal or the oocyte is unstable (Storto and King, 1988). In otu7 mutants, some oocyte nuclei are partially transformed, and become polyploid; yet they remain distinguishable from nurse cell nuclei. Our data revealed that in the vitellogenic egg chambers of otu5 mutants the oskur mRNA is mislocalised around the cortex of the ooplasm. This seems to be due to an inefficient translocalisation of the oskur mRNA from the anterior to the posterior pole. The 98 kDa otu isoform is also likely to be required for the full maturation of nurse cells. The significant reduction in the cytoplasmic distribution of the 3.2 kb mRNA in nurse cells would explain the arrest of nurse cell maturation at stage 12 in the aforementioned mutants. 3.1. Re-evaluation of a model explaining the otu phenovpes Storto and King (1988) proposed that the otu product is required in increasing amounts throughout oogenesis, and that mutant phenotypes of otu can be viewed as graded responses by germ cells to different levels of the otu gene products synthesised in the mutant cells them-

73

selves. In the wild type, the otu proteins are expressed in increasing levels as the egg chambers mature (Steinhauer and Kalfayan, 1992). In otu’ and otuJ4 mutant egg chambers the level of the otu protein is significantly reduced. It has been proposed that the carboxy-terminal deletions destabilise the mutant proteins encoded by the otu’ and otuJ4 alleles. This would then lead to a reduction of the otu protein level and, finally, to the mutant phenotypes (Steinhauer and Kalfayan, 1992). Our results provide another explanation for the reduction of the otu protein levels in the aforementioned mutants. We propose that the altered otu proteins in otu5 and otuJ4 mutants lead to an increased accumulation of the otu mRNA within the nurse cell nuclei. The accumulation of the 3.2 kb mRNA would in itself hinder the distribution of the transcript into the cytoplasm leading to a reduced level of the 98 kDa protein. Hence, the severity of otu phenotypes could depend on the level of the otu mRNA in the germ cell cytoplasm. This is supported by the fact that the mRNA level is more normal in otuJ4 mutants than in otu5 mutants. The egg chamber maturation proceeds further in otu14 than in otuS mutants. According to our view, the phenotype of otu5, otu7 and otu14 mutants would in fact result from a cumulative ef-’ feet of a reduced dose of several different ovarian mRNAs. This view is supported by the effect of the otu mutations on the localisation of the pumilio mRNA. The pumilio transcript is not required for oogenesis per se, but according to our view, it represents one transcript among others, that are not properly localised in the otu mutants. Earlier electron microscopy studies have detected large ribonucleoprotein (RNP) accumulations within the nurse cell nuclei of the otu’ mutant (Bishop and King, 1984). Our data suggest that the otu and pumilio mRNA form part of the RNP accumulations, but the large size of the deposits suggest that several other mRNAs may be involved (Tapio Heino et al., unpublished). The nature of the otu function specificity for certain mRNAs earns further attention. Nevertheless, some pleiotropic effects of otu mutations could reflect the fact that several different mRNAs are sequestered in the mutant nurse cell nuclei. 3.2. Function of the otu protein in the distribution of specific mRNAs The otu protein has so far been observed in the cytoplasm, not in the nuclei of nurse cells (Steinhauer and Kalfayan, 1992). Therefore, it seems unlikely that the otu protein would be directly involved in the processing or export of nuclear mRNAs. Chromatin structure is also one target of the otu function as indicated by the abnormal polytenisation in otu mutant nurse cells. Organisation of the template chromatin may well have an effect on the elongation or processing of the nascent transcripts. But, similarly, the regulation of the chromatin organisation requires a nuclear function.

74

h4. Tirronen et al. I Mechanisms of Development 52 (I 995) 65-75

Several recent studies suggest that mRNA processing, chromatin structure and cell cycle regulation are coupled, but do not establish the nature of the connection (Forrester et al., 1992; Amberg et al., 1993). One general failure common to all mutant classes of otu is an abnormal number of germ cells (Storto and King, 1988). In the tumorous class, the germ cells have been widely thought to have undergone erroneous sex determination, and to possess an identity of male germ cells (Bopp et al., 1993; Oliver et al., 1993; Pauli et al., 1993; Wei et al., 1993). However, the nature of the sexual transformation is obscure, since the germ cells simultaneously express both male and female marker traits (Bae et al., 1994). These observations suggest that the primary error in otu mutants is not in the sexual identity of the germ cells. Could the otu function be primarily required for the regulation of the cell cycle or endomitosis? The otu gene cannot be a general cell cycle regulator, since it is required primarily in the female germ line (Tirronen et al., 1993), but the otu function might be to modulate the properties of a general cell cycle regulation complex in a female germ line specific manner. Control of cystocyte divisions requires several functions that are not utilized in other cell types, such as centrosome-fusome association (Storto and King, 1989), centriole migration from the nurse cell to the oocyte (Mahowald and Strassheim, 1970), MTOC activation in the oocyte (Theurkauf et al., 1993), specific microtubule rearrangements (Theurkauf et al., 1992; 1993) and control of nurse cell endoreduplication (Storto and King, 1988). The exact role of the otu proteins in oogenesis remains elusive. Nevertheless, our sequence comparison data give interesting clues about one possible molecular function of otu. Several stretches of the otu protein sequence share local homologies with the human, rat and mouse microtubule associated proteins (MAP2s). This raises the possibility that the otu function is associated with the microtubular structures of the germ cells. Microtubules are a key component in centrosomes and in the mitotic spindle, and they play a crucial role in several processes of the cell cycle. A number of recent studies have provided evidence of a link between microtubules and the localisation of specific mRNAs, e.g. the oskur mRNA (Pokrywka and Stephenson, 1991; Theurkauf et al., 1993; Clark et al., 1994). Interestingly, mutations affecting protein kinase A (PKA) that act in the germ line to disrupt microtubule distribution, affect the localisation of the oskar mRNA to the posterior pole of the oocyte in a manner similar to what we observe in otd mutants (Lane and Kalderon, 1994). The value of these observations is not to connect the otu function to that of PKA, but to reveal the multiple steps where otu could participate in monitoring and regulating microtubule assembly during oogenesis. Future analyses will reveal whether otu mediates its effect on cell cycle regulation and mRNA distribution by regulating some aspect of microtubule function.

4. Experimental

procedures

4.1. Drosophila stocks The flies were reared under uncrowded conditions at 25°C in bottles containing standard Drosophila medium (malt, cornmeal, yeast, agar, propionic acid, Nipagin). The mutant strain of otuJ3 was reared at 18°C to enhance the germ cell maturation in favour of the ‘differentiated’ egg chambers (Storto and King, 1988). 4.2. mRNA hybridisations in situ In situ hybridisations to whole gonads were performed according to Tautz and Pfeifle (1989) with modifications according to Anne Ephrussi, Tom Serano and Robert Cohen (personal communication). The first fixation of dissected tissue was made by adding 1 vol. of 4% paraformaldehyde (in 100 mM Hepes, pH 6.9, 2 mM MgS04, 1 mM EGTA) and 4 ~01s. heptane to the tissue in an Eppendorf tube, which was shaken vigorously for 20 min. The fixation solution and most of the heptane was replaced with an equal amount of methanol and shaking was continued 3 min. The tissue was rinsed with methanol, then with a 1:3 solution of methanol in ethanol and finally with ethanol, then stored in absolute ethanol for at least 2 h at -20°C. The second fixation and permeabilisation were performed as follows: the ethanol was replaced by methanol and the tissue was incubated at room temperature for 3 min. Then the tissue was rinsed 5 min in a 1:2 solution of methanol in PP (PP is 4% paraformaldehyde in phosphate-buffered saline (PBS) with 0.1% Tween-20) and fixed for 20 min in a 1: 10 solution of dimethylsulfoxide (DMSO) in PP at room temperature. Following several rinses in PBT (PBT is PBS including 0.1% Tween-20), the tissue was treated with proteinase-K (50pg/ml in PBT for 30 min at room temperature). The treatment was stopped by three consecutive washes with a solution of PBT, the first one containing 2 mg/ml glycine, then fixed a third time for 20 min in PP. The samples were then rinsed for 15 min with five changes of PBT. The hybridisation and washing were performed according to Tautz and Pfeifle (1989). 4.3. Sequence comparisons In a first step, stretches of the otu polypeptide were compared to the SwissProt and PIR databases using the program Blast (Altschul et al., 1990). After the selection of sequences showing enough homology to the carboxyterminal end of the otu protein, they were in a second step compared pairwise to otu using the GCG (Genetic Computer Group software) program Compare. By selecting a large window size (40) and a relatively high stringency (21) we wanted to favour rather long areas of homology. From the result of the matrix comparisons (shown as dots in Fig. 6B), we selected the areas of homology from the otu protein carboxy-terminal end that are affected by the otu5 and otu14 mutations for further analyses (shaded on

h4. Tirronen et al. I Mechanisms of Development 52 (1995) 65-75

Fig. 6B). In a last step we used the GCG program Bestfit to find the optimal alignment in the area of homology (Fig. 6D). The comparison between bag-of-marbles and ~tu (McKearin and Spradling, 1990) is included as a reference (Fig. 6A,D). The same parameters were used for all comparisons. Acknowledgements We are indebted to Monica Steinmann-Zwicky for her critical comments on the manuscript. We thank Anne Ephrussi, Tom Serano and Robert Cohen for the valuable suggestions they made to improve our in situ hybridisation method. We gratefully thank Anne Ephrussi, Paul Macdonald and Daniel St Johnston who promptly answered our clone requests. References Altschul, S.F., Gish, W., Miller, W., Myers, E.W. and Lipman, D.J. (1990) J. Mol. Biol. 215,403-10. Amberg, D.C., Fleischmann, M., Stagljar, I., Cole, C.N. and Aebi, M. (1993) EMBO J. 12: 233-241. Bae, E., Cook, K.R., Geyer, P.K. and Nagoshi, R.N. (1994) Mech. Dcv. 47, 151-164 Berleth, T., Burri, M., Thoma, G., Bopp, D., Richstein, S., Frigerio, G., Noll, M. and Niisslein-Volhard, C. (1988) The EMBO J. 7, 17491756. Bishop, D.L. and King, R.C. (1984) J. Cell Sci. 67,87-119. Bopp, D, Horabin, J.I., Lersch, R.A., Cline, T.W. and Schedl, P. (1993) Development 118,797-g 12. Boyd, L., O’Toole, E. and Thummel, C.S. (1991) Development 112, 981-995. Chambe, M.A. and Laird, CD. (1989) Nucleic Acids Res. 17, 3304. Cheung, H.-K., Serano, T.L. and Cohen, R.S. (1992) Development 114, 653-661. Clark, I., Giniger, E., Ruohola-Baker, H., Jan, L.Y. and Jan, Y.N. (1994) Curr. Biol. 103: 189-203. Comer, A.R., Searles, L.L. and Kalfayan, L. (1992) Gene 118, 171179. Cummings, M.R. and King, R.C. (1969) J. Morphol. 128,427-442. Ephrussi, A., Dickinson, L.K. and Lehmann, R. (1991) Cell 66,37-50. Forrester, W., Stutz, F., Rosbash, M. and Wickens, M. (1992) Genes Dev. 6: 1914-1926. Haenlin, M., Roos, C., Cassab, A. and Mohier, E. (1987) EMBO J. 6, 801-807. Kim-Ha, J., Smith, J.L. and MacDonald, P.M. (1991) Cell 66,23-36. King, R.C. (1970) Ovarian Development in Drosophila melanoguster, Academic Press, New York. King, R.C., Bahns, M., Horowitz, R. and Larramendi, P. (1978) Int. J. Insect Morphol. Embryol. 7, 359-378.

75

King, R.C., Mohler, D.J., Riley, S.F., Storto, P.D. and Nicolazzo, P.S. (1986) Dev. Genet. 7, l-20. Lane, M.E. and Kalderon D. (1994) Genes Dev. 8.2986-2995. Lin, H., Yue, L. and Spradling, A.C. (1994) Development 120, 947956. Macdonald, P.M. (1992) Development 114, 221-232. Mahowald, A.P. and Kambysellis, M.P. (1980) In Ashbumer, M. and Wright, T.R.F.(eds.), The Genetics and Biology of Drosophila, Academic Press, New York. Vol. 2d, pp. 141-224. Mahowald, A.P. and Strassheim, J.M. (1970) J. Cell Biol. 45, 306-320. Mandelkow, E. and Mandelkow, E.-M. (1995) Curr. Opin. Cell Biol. 7, 72-8 1. McKearin, D.M. and Spradling, A.C. (1990) Genes Dev. 4,2242-2251. Mulligan, P.K. and Rasch, E.M. (1985) Histochemistry 82, 233-247. Mulligan, P.K., Mohler, J.D. and Kalfayan, L.J. (1988) Mol. Cell. Biol. 8, 1481-1488. Nagoshi, R.N., Patton, J.S., Bae, E. and Geyer, P.K. (1995) Development 121.579-587. O’Farrel, P.H., Edgar, B.A., Lakich, D. and Lehner, C.F (1989) Science 2469635640. Oliver, B., Kim, Y.-J. and Baker, B.S. (1993) Development 119, 897908. Pauli, D., Oliver, B. and Mahowald, A.P. (1993) Development 119, 123-134. Pokrywka, N.J. and Stephenson, E.C. (1991) Development 113.55-66. Robinson, D.N., Cant, K. and Cooley, L. (1994) Development 120, 2015-2025. Shermoen, A.W. and O’Farrel, P. (1991) Cell 67.303-310. St Johnston, D., Beuchle, D. and Niisslein-Volhard, C. (1991) Cell 66, 51-64. Steinhauer, W.R. and Kalfayan, L.J. (1992) Genes Dev. 6, 233-243. Steinhauer, W.R., Walsh, R.C. and Kalfayan, L.J. (1989) Mol. Cell. Biol. 9, 5726-5732. Steinmann-Zwicky, M. (1992a) BioEssays 14.513-518. Steinmann-Zwicky, M. (1992b) Semin. Dev. Biol. 3.341-347. Storto, P.D. and King, R.C. (1988) Dev. Genet. 9,91-120. Storto, P.D. and King, R.C. (1989) Dev. Genet. 10: 7&86. Suter, B. and Steward, R. (1991) Cell 67, 917-926. Suter, B., Romberg, L.M. and Steward, R. (1989) Genes Dev. 3, 19571968. Tautz, D. and Pfeifle, C. (1989) Chromosoma 98,81-85. Theurkauf, W.E., Smiley, S., Wong, M.L. and Alberts, B.M. (1992) Development 115.923-936. Theurkauf, WE., Alberts, B.M., Jan, Y.N. and Jongens, T.A. (1993) Development 118, 1169-l 180. Tirronen, M., Partanen, M., Heino, T.O., Heino, T.I. and Roos, C. (1992) Mech. Dev. 40, 113-126. Tirronen, M., Heino, T.I. and Roos, C. (1993) Roux’s Arch. Dev. Biol. 202,306-3 11. Wei, G., Oliver, B., Pauli, D. and Mahowald, A.P. (1994) Dev. Biol. 161,318-320. Xue, F. and Cooley, L. (1993) Cell 72.681693. Yue, L. and Spradling, A.C. (1992) Genes Dev. 6, 2443-2454. Zachar, Z., Kramer, I., Mims, P. and Bingham, P.M. (1993) J Cell Biol. 121.729-742.