Localization and association to cytoskeleton of COLL1α mRNA in Paracentrotus lividus egg requires cis- and trans-acting factors

Mechanisms of Development 99 (2000) 113±121

www.elsevier.com/locate/modo

Localization and association to cytoskeleton of COLL1a mRNA in Paracentrotus lividus egg requires cis- and trans-acting factors Daniele P. Romancino, Serena Dalmazio, Melchiorre Cervello, Giovanna Montana, Lucrezia Virruso, Angela Bonura, Roberto Gambino, Marta Di Carlo* Istituto di Biologia dello Sviluppo CNR, via Ugo La Malfa 153, 90146 Palermo, Italy Received 31 July 2000; received in revised form 15 September 2000; accepted 15 September 2000

Abstract COLL1a mRNA is asymmetrically distributed in the Paracentrotus lividus egg. Here we examine the involvement of the cytoskeleton in the localization process of collagen mRNA. The use of drugs such as colchicine and cytochalasin B reveals a perturbation of localization collagen mRNA. Moreover, the presence of speci®c cis-and trans-acting factors involved in cytoskeleton binding and the localization process was investigated. By Northwestern experiment we found that the 3 0 UTR of COLL1a mRNA is also able to bind two proteins of 54 and 40 kDa in a cellular fraction containing the cytoskeleton. Finally, we found that the protein of 54 kDa is LP54, a protein that binds the 3 0 UTRs of P. lividus maternal bep messengers and is necessary for their localization. q 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: RNA localization; 3 0 UTR; RNA-binding protein; Paracentrotus lividus

1. Introduction Localization of messenger RNAs and proteins is an intriguing mechanism utilized by the cell to obtain a gradient of proteins and establish the embryonic axis (Ding and Lipshitz, 1993). Both messenger RNAs and proteins have been found localized in oocytes and embryos, as well as in a variety of differentiated and polarized cells from yeast to mammalian (Bashirullah et al., 1996). Moreover, localized RNAs practically program the asymmetric cell fate of cells that receive that cytoplasm. Mechanisms used to transport and anchor RNAs in the cytoplasm include vectorial transport out of the nucleus, direct cytoplasmic transport in association with the cytoskeleton, and local anchoring at particular cytoplasmic sites (Wilhelm and Vale, 1993; Hesketh, 1996). The majority of localized RNAs are targeted to particular cytoplasmic regions by cis-acting RNA elements that almost always reside in the 3 0 -untranslated regions (UTRs) (Ding and Lipshitz, 1993; Bashirullah et al., 1998). Moreover, by means of deletion experiments speci®c sequences suf®cient for localization have been mapped inside the 3 0 UTR of some mRNAs such as

* Corresponding author. Tel.: 139-91-680-9538; fax: 139-91-680-9548. E-mail address: [email protected] (M. Di Carlo).

BLE1 in bicoid of Drosphila or VgLE in Vg1 of Xenopus (Macdonald et al., 1993; Macdonald and Kerr, 1998; Mowry and Melton, 1992). cis-Acting elements play their role associated with trans-acting factors, many of which are RNA-binding proteins. Moreover, some of these trans-acting factors can function in translational control during or after localization. In order to get a good knowledge of this mechanism and the factors that are involved it is useful to understand the normal development of an early embryo or differentiating cells. trans-Acting factors for some messengers have been identi®ed, mainly in the Xenopus and Drosophila systems. Some of these factors are common to several messengers. For example, Staufen protein in Drosphila is able to bind to the 3 0 UTRs of bicoid, oskar and prospero mRNAs at different stages of development or neurogenesis (Ephrussi et al., 1991; St Johnston, 1995; Kim-Ha et al., 1991; Broadus et al., 1998). Moreover, the 3 0 UTR of bicoid mRNA is able to bind two other proteins, exuperantia and swallow, while the 3 0 UTR of oskar mRNA also binds the Bruno protein that in turn is associable with gurken mRNA (Kim-Ha et al., 1995). In Xenopus, the Vg1 RNA-binding protein (RBP), which probably coincides with Vera, binds to the 3 0 UTR of both Vg1 mRNA and TGFb -5 (Schwartz et al., 1992; Elisha et al., 1995; Deshler et al., 1997). Moreover, Vg1 mRNA for its localization was also found to utilize a

0925-4773/00/$ - see front matter q 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0925-477 3(00)00486-X

114

D.P. Romancino et al. / Mechanisms of Development 99 (2000) 113±121

non-translatable RNA Xlsirt, but it is not known whether it interacts directly or via the cytoskeleton network (Kloc and Etkin, 1994). Many features of mRNA localization appear to have been conserved during evolution, suggesting that this is an old mechanism for producing cytoplasmic asymmetry. In the invertebrate sea urchin three maternal messengers have been identi®ed, called bep1, bep3 and bep4, localized at the animal pole of the Paracentrotus lividus egg and embryos (Di Carlo et al., 1990, 1994; Montana et al., 1996). These messengers are anchored to the cytoskeleton by cis-acting factors residing in their 3 0 UTRs and by a common transacting factor, an RNA-binding protein called LP54 (Montana et al., 1997; Romancino et al., 1998; Costa et al., 1999). Collagen is a protein which shows a particular amino acid repeated motif Gly±X±Y that permits formation of an alphahelical structure. This molecule is present in the extracellular matrix of all metazoan phyla and plays a fundamental role in a number of processes such as morphogenesis and differentiation. In sea urchin four cDNA coding collagens have been isolated: two ®brillar collagens, called COLL1a or Hpcol1 (D'Alessio et al., 1989; Tomita et al., 1994), and COLL2a (D'Alessio et al., 1990) and two non-®brillar collagens, called COLL3a and COLL4a (Esposito et al., 1993, 1994). COLL1a mRNA is expressed principally from the late gastrula to the pluteus stage (D'Alessio et al., 1989), but a very small amount is also present in P. lividus egg (Gambino et al., 1997). By means of in situ hybridization it has been demonstrated that COLL1a mRNA is spatially distributed in egg and embryos (Gambino et al., 1997). In this report we investigate the cellular machinery and the factors that are involved in COLL1a mRNA localization.

reaction (RT±PCR) analysis. We utilized two primers designed at about the 5 0 end and at about the 3 0 end of the 3 0 UTR. Thus these primers amplify a fragment of about 700 bp. Moreover, we utilized, as control, two primers that amplify a fragment of about 400 bp of the histone H2A mRNA (Spinelli and Albanese, 1990), an RNA, as previously demonstrated, not exclusively present in the cytoplasm (Montana et al., 1997). Fig. 1A shows that a larger amount of COLL1a mRNA is present in the supernatant fraction of unfertilized eggs and is mainly present in the pellet fraction of fertilized eggs, suggesting that COLL1a mRNA is more associated with the cytoskeleton after fertilization. When the primers for H2A mRNA were employed, an ampli®ed band was mainly found in the supernatant samples of both unfertilized and fertilized eggs (Fig.1B). We would note that although the RT±PCRs were carried out in a linear range of PCR ampli®cation, they represent the approximate difference in RNA level. These results are in agreement with the different spatial distribution observed by in situ hybridization of COLL1a mRNA, in which a gradient of distribution is detectable in unfertilized eggs, in contrast with a more restricted localization found in the fertilized egg (Gambino et al., 1997). 2.2. Effect of cytoskeleton inhibitors on localization of COLL1a mRNA Cytoskeleton is mainly composed by microtubules,

2. Results 2.1. Association of COLL1a with the cytoskeleton It is known that most localized mRNAs are associated with the cytoskeleton. COLL1a mRNA is localized in the P. lividus egg and we decided to investigate whether this messenger is also associated with the cytoskeleton. For this aim we separated a cytoskeleton fraction from unfertilized and fertilized eggs. By treatment with detergent and subsequent centrifugation we obtained a detergent-soluble fraction (supernatant) and a detergent-insoluble fraction (pellet). The pellet contains the cytoskeleton elements and all the cytoplasmatic components associated with them. The supernatant fraction contains the remaining cytoplasmatic components. From these two fractions we extracted the total RNAs. COLL1a mRNA is only expressed to a low extent in P. lividus and is dif®cult to detect by Northern blot hybridization (Gambino et al., 1997). For this reason we decided to analyze the presence or absence of COLL1a mRNA in each fraction by reverse transcriptase±polymerase chain

Fig. 1. Analysis of COLL1a (A) and H2A (B) mRNAs presence in cytoskeleton fraction. RT±PCR products of RNA extracted from total eggs (1) as control or pellet (2,4) and supernatant (3,5) fractions of unfertilized (2,3) and fertilized (4,5) eggs. M indicates the 123 bp ladder utilized. The PCR products are indicated by an arrow.

D.P. Romancino et al. / Mechanisms of Development 99 (2000) 113±121

115

Fig. 2. Effect of drugs on localization of COLL1a mRNA. Whole-mount in situ hybridization of fertilized eggs (control) (A), and fertilized eggs treated with colchicine (B) and cytochalasin B (C) incubated with antisense COLL1a mRNA. Whole-mount immunohistochemistry of fertilized eggs incubated with antia -tubulin (D).

micro®laments and intermediate ®laments. To determine which of the cytoskeletal elements interacts with COLL1a mRNA we utilized two drugs able to cause depolymerization of microtubules or micro®laments: colchicine and cytochalasin B respectively. After separate incubation with the two drugs, the fertilized eggs were ®xed to investigate the localization of COLL1a mRNA by in situ hybridization. As shown in Fig. 2, both colchicine (Fig. 2B) and cytochalasin B (Fig. 2C) provoke a release of the message from its normally ®xed distribution (Fig. 2A), indicating that COLL1a mRNA is associated both with tubulin and actin ®laments. Moreover, the presence and distribution of microtubules in the eggs utilized as control was evidenced by whole-mount immunohistochemistry of the same batch of eggs incubated with anti-a-tubulin (Fig. 2D).

mRNA, no competition was observed. This result indicates that in the proteic extract one or more proteins are present that are able to recognize a signal in the 3 0 UTR of the COLL1a mRNA and that the formed complex is speci®c. 2.4. Speci®c protein binds the 3 0 UTR of COLL1a mRNA To identify the proteins that bind the 3 0 UTR of COLL1a mRNA we carried out a Northwestern experiment. Cytoplasmic proteins extracted from unfertilized and fertilized eggs were separated by sodium dodecyl sulfate±polyacrylaimide gel electrophoresis (SDS±PAGE). The Western blot was incubated with the 3 0 UTR of COLL1a mRNA or with

2.3. Complex formation by COLL1a mRNA and egg proteic extract To investigate whether proteins present in total cytoplasmic extract of eggs could recognize speci®c sequences in the 3 0 UTR of COLL1a mRNA, we synthesized a fragment 750 nucleotides long containing all this region in vitro, utilizing DIG-UTP. This sense transcript was incubated with total egg protein extracts with or without heparin and analyzed by gel-retardation. The reaction was visualized on non-denaturing gel (Fig. 3A). The addition of protein extracts retarded the mobility of the 750-nucleotide RNA, indicating that a complex was formed. This complex is also maintained in the sample incubated with heparin, which eliminates non-speci®c protein binding. Moreover, to demonstrate the speci®city of the complex, we incubated the total egg protein extracts with the collagen fragment transcribed with the DIG-UTP together with 200-fold and 400-fold molar excess of the same fragment transcribed without DIG-UTP. After electrophoresis a competition was observed (Fig. 3B). Moreover, when the total egg extracts were incubated with the DIG-UTP transcript of the 3 0 UTR of the COLL1a mRNA together with the 400fold molar excess of a non-labeled in vitro transcript of H2A

Fig. 3. Bandshift analysis of speci®c complexes by COLL1a mRNA and egg cytoplasmic extract. (A) 3 0 UTR of COLL1a RNA alone (RNA) or incubated with the extract (TE) and the same treated with heparin (H). (B) the Total egg extract incubated with digoxigenin labeled 3 0 UTR of COLL1a mRNA (1) as control, or plus 200-fold (2) or 400-fold (3) molar excess of non-digoxigenin labeled 3 0 UTR of COLL1a mRNA, or plus 400-fold molar excess of non-digoxigenin-labeled H2A mRNA (4).

116

D.P. Romancino et al. / Mechanisms of Development 99 (2000) 113±121

the in vitro transcribed 3 0 UTR of COLL1a mRNA. Binding with both the 54 and 40 kDa proteins was detected exclusively with the pellet fraction containing the cytoskeleton elements (Fig. 5). The result obtained by the Northwestern experiment suggests that both the proteins might be involved in COLL1a mRNA association with cytoskeleton network and consequently that they may be involved in the RNA localization process. 2.6. The 3 0 UTR of COLL1a mRNA recognized the LP54

Fig. 4. Binding of 3 0 UTR COLL1a mRNA to speci®c proteins. (A) Western blot of proteins isolated from unfertilized (1) and fertilized (2) eggs incubated with labeled 3 0 UTR of COLL1a mRNA (coll) or with 3 0 UTR of bep3 mRNA (bep3). The lines indicate the migration position of the proteins utilized as marker. (B) Western blot of proteins isolated from fertilized eggs incubated with labeled 3 0 UTR of COLL1a RNA (1) or plus 200fold (2) or 400-fold (3) molar excess of non-radioactive labeled 3 0 UTR of COLL1a mRNA, or plus 400-fold molar excess of non-radioactive labeled H2A mRNA (4).

the 3 0 UTR of bep3 mRNA as a control, both synthesized in vitro utilizing radioactive UTP. Two bands were detected by binding with the 3 0 UTR COLL1a mRNA, one band of 54 kDa and another of 40 kDa in both unfertilized and fertilized egg extracts (Fig. 4A). The 54 kDa band coincides with the band recognized by the 3 0 UTR transcript of bep3 mRNA. The 40 kDa band is speci®c, instead, for the 3 0 UTR of COLL1a mRNA. Moreover, when we incubated different strips of a Western blot containing proteins isolated from fertilized eggs with the labeled fragment of COLL1a mRNA together to 200-fold or 400-fold molar excess of the same cold transcript, a decreased signal compared with the control was observed (Fig. 4B). Moreover, when a Western blot containing proteins isolated from fertilized eggs was incubated with 400-fold molar excess of cold in vitro transcript of the H2A mRNA together with the labeled transcript of the 3 0 UTR of the COLL1a mRNA, no competition was observed.

To determine, in order, whether the protein of 54 kDa is the LP54 and thus, the protein that binds bep mRNAs, we performed a binding experiment utilizing the previously isolated LP54 (Costa et al., 1999). A Western blot containing total egg extract, puri®ed LP54 and BSA as negative control was incubated with in vitro transcript of sense 3 0 UTR COLL1a mRNA. The result of the Northwestern experiment is shown in Fig. 6A. A band is detectable in the line in which the puri®ed LP54 was loaded and in the positive control sample, whereas no band was detected in the lane in which BSA was loaded. Moreover, competition with cold in vitro transcript of sense 3 0 UTR COLL1a mRNA was demonstrated by incubating different strips of a Western blot containing puri®ed LP54 with 200-fold and 400-fold excess of cold in vitro transcript 3 0 UTR COLL1a mRNA together with labeled in vitro transcript 3 0 UTR COLL1a mRNA (Fig. 6B). No decrease in the signal was observed when 400-fold molar excess of cold H2A mRNA was utilized as competitor (Fig. 6B). This result clearly indicates that the protein of 54 kDa recognized by 3 0 UTR of COLL1a mRNA is the LP54 that binds the 3 0 UTR of all bep mRNAs.

2.5. The 3 0 UTR of COLL1a mRNA is associated with the cytoskeleton To determine whether the proteins of 54 and 40 kDa associate the 3 0 UTR of COLL1a mRNA with the cytoskeleton, we treated P. lividus with detergents, and after centrifugation the proteins from soluble (supernatant) and unsoluble (pellet) fractions were isolated. These proteins were separately electrophoresed together with a total egg lysate as a control. The Western blot was incubated with

Fig. 5. Presence of both 54 and 40 kDa proteins in detergent unsoluble and soluble fractions of eggs. Western blot of proteins isolated from pellet (1,3) and supernatant (2,4) fractions of unfertilized (1,2) and fertilized (3,4) eggs incubate with labeled 3 0 UTR of COLL1a mRNA. The lines indicate the migration position of the proteins utilized as marker.

D.P. Romancino et al. / Mechanisms of Development 99 (2000) 113±121

3. Discussion In many cases messenger RNA localization is thought to be an active process that requires the cytoskeleton (Wilhelm and Vale, 1993; Bashirullah et al., 1998). cis-Acting signaling is necessary in order to direct localization, and it is mapped principally in the 3 0 UTR of transcripts and RNAbinding proteins that interact with these signals (Bashirullah et al., 1998). Preceding observations indicated that the collagen mRNA (COLL1a ) is distributed as a gradient in P. lividus unfertilized egg and is restricted in a cortical zone after fertilization, suggesting a shift through the cytoskeleton network (Gambino et al., 1997). Here we provide evidence that localization of this mRNA is due to association with the cytoskeleton elements and we begin to identify the factors involved in this association. By RT±PCR analysis utilizing RNA extracted from soluble or unsoluble (cytoskeleton) fractions of P. lividus eggs we established a more marked association of COLL1a with the cytoskeleton in the fertilized eggs than in the unfertilized eggs. Thus, the association and the more restricted localization in a pole of the egg after fertilization suggest a shift of collagen mRNA along one of the principal ®laments. This is in agreement with the fact that in sea urchin polymerization of microtubules and micro®laments and the assembly of the cytoskeleton network occur after fertilization. Rearrangement of actin and tubulin ®bers after fertilization has, indeed, been demonstrated in sea urchin (Burgess and

Fig. 6. Binding of the 3 0 UTR of COLL1a mRNA to LP54. (A) Western blot of total extract proteins (1), puri®ed LP54 (2) and BSA (3) incubated with labeled 3 0 UTR of COLL1a mRNA. The lines indicate the migration position of the proteins utilized as marker. (B) Western blot of puri®ed LP54 incubated with labeled 3 0 UTR of COLL1a RNA (1) or plus 200-fold (2) or 400-fold (3) molar excess of non-labeled 3 0 UTR of COLL1a mRNA, or plus 400-fold molar excess of non-labeled H2A mRNA (4).

117

Schroeder, 1977; Harris et al., 1980a,b). For example, in Strongilocentrotus purpuratus, different arrangements of tubulin ®bers occur after fertilization. Microtubules are found in the egg cortex soon after fertilization, and are oriented more or less perpendicularly to the egg surface. About 30 min later, the time of the pronuclear fusion, the cortical microtubules forms a characteristic spiral array that will disappear before the streak stage (Harris et al., 1980a,b). At the moment we cannot exclude that COLL1a mRNA follows the movement of microtubules. Moreover, some mRNAs are localized by microtubule-dependent mechanisms, such as Vg1 mRNA, which moves to the vegetal pole of the Xenopus oocytes (Yisraeli et al., 1990), or bicoid and oskar mRNAs, which localize at opposite poles of the Drosophila oocytes (Pokrywka and Stephenson, 1991; Ephrussi et al., 1991). By contrast, localization of b -actin mRNA in cultured ®broblast requires actin ®laments (Kislauskis et al., 1994). Treatment with drugs that depolymerize microtubules or micro®laments produces a release of the COLL1a message from its ®xed position, indicating that both these structures are involved in maintaining the position of COLL1a mRNA. The same result was obtained for P. lividus maternal bep mRNA, which is trapped in the microtubule and micro®lament network at the animal cortex of fertilized eggs (Romancino et al., 1998; Costa et al., 1997). However, these results do not give us any information about the ®laments that might be involved in possible transport. Moreover, we investigate whether the 3 0 UTR of COLL1a mRNA is able to bind one or more proteins that may play the role of trans-acting factors and whether this complex may permit association with the cytoskeleton. By bandshift and Northwestern experiments we established that the 3 0 UTR of collagen mRNA binds two speci®c proteins of 54 and 40 kDa. Previously we demonstrated that the 3 0 UTR of bep mRNAs is only able to bind a protein of 54 kDa (Montana et al., 1997). The presence of two proteins might be due to the fact that the 3 0 UTR of collagen is 750 nucleotides long whereas the 3 0 UTR of bep mRNAs ranged between 220 and 140 nucleotides (Montana et al., 1998). Thus the 40 kDa protein might be speci®c for binding with the 3 0 UTR of collagen mRNA. Moreover, both proteins seem to mediate the association with the cytoskeleton, as demonstrated by Northwestern experiments in which both the proteins were detected only in the insoluble cytoplasmic fraction. The presence of an additional protein might also be due to the localization of COLL1a mRNA in the egg and during development. This collagen is localized on one side of the egg and of the embryos until the 16-cell stage and has been suggested to be involved in oral±aboral axis determination (Gambino et al., 1997), whereas localization of bep mRNAs and proteins is along the animal±vegetal axis (Di Carlo et al., 1994; Montana et al., 1996). Thus this type of collagen may play a role in determining the diversity of the extracellular matrix and in¯uencing the morpho-

118

D.P. Romancino et al. / Mechanisms of Development 99 (2000) 113±121

genetic program and signal transduction of particular territories. Consequently, speci®c factors for speci®c territories might be involved and the 40 kDa might be the candidate. Moreover, we demonstrated that the protein of 54 kDa is the LP54, the protein that has been extensively demonstrated to be involved in bep mRNA localization (Montana et al., 1997; Costa et al., 1999). Thus this protein might be a trans-acting factor common to several messengers that are not localized in the vegetal half of the P. lividus egg. In other systems it has been demonstrated that a RNA-binding protein can recognize and interact with cis-acting signals of different mRNAs and drive their localization, also at different development stages. In Xenopus, for example, Vg1/ VERA RNA-binding protein recognizes a motif in both Vg1 and TGFb -5 mRNAs and is involved in their localization at the vegetal pole of the oocyte (Schwartz et al., 1992; Elisha et al., 1995). In Drosophila the best characterized example is the RNA-binding protein Staufen (stau) that is required for the localization of oskar (osk) mRNA at the posterior pole of the oocytes, the anchoring of bicoid (bcd) mRNA at the anterior part of the egg and for the basal localization of prospero mRNA during asymmetric divisions of embryonic neuroblasts (Ferrandon et al., 1994; Ephrussi et al., 1991; Broadus et al., 1998). Moreover, the proteins identi®ed here might have different domains to bind regions of the RNA, microtubules and/ or micro®laments and perhaps also other proteins. Different domains have been identi®ed, for example, in stau protein. Stau contains ®ve double-stranded (ds) RNA-binding domains (dsRBDs) that have been conserved throughout animal evolution and each of them plays a different role in different mRNAs. dsRBDs 1, 3 and 4 are the domains that bind to dsRNA in vitro, while 2 and 5 do not (Micklem et al., 2000). When the binding activity of dsRBD3 is abolished by mutation, localization of both osk and bcd is prevented. Moreover, dsRBD2, which has a short insertion splitting it into two halves, is required for the microtubuledependent localization of osk mRNA. DsRBD5 is instead required for the derepression of osk mRNA translation, since localized. DsRBD5 has also been demonstrated to drive the localization of prospero mRNA in an actin-dependent manner in embryonic neuroblasts (Micklem et al., 2000). We have not yet investigated the possibility of the 3 0 UTR of COLL1a mRNA folding in a secondary structure in which double strand stems and loops are present that might be speci®cally recognized by one or more domains of the 54 and 40 kDa proteins. The possibility of folding in secondary structures has been demonstrated for bep mRNAs and the LP54 recognizes a minimal sequence, and probably structure, of 80 nucleotides (Montana et al., 1998). The importance of a secondary structure has also been seen for several localized messengers and recently it has been demonstrated that stau dsRBD3 binds optimally in vitro to a non-speci®c RNA hairpin with a minimal stem of 12 bp (Ramos et al., 2000).

We do not yet know why this type of collagen is localized in a region of the P. lividus egg. Localization of mRNA is often combined with local translation control. Most of the current knowledge indicates that messenger RNAs before and during localization are translationally repressed and derepressed until the protein is required (Gunkel et al., 1998). Translation repression of nos and osk transcript requires regulatory sequences also present in the 3 0 UTRs that bind to RNA-binding proteins called, respectively, Smaug and Bruno (Smilbert et al., 1996; Kim-Ha et al., 1995) Moreover, for osk mRNA a correlation has been demonstrated between translational derepression and the presence of an element present in the 5 0 -end region that binds a protein of 50 kDa involved in this process (Gunkel et al., 1998). Thus we cannot rule out the possibility that one of the proteins that binds the 3 0 UTR of COLL1a mRNA might have a role in translational regulation, perhaps with the help of other RNA-binding proteins that permit communication between the 3 0 and the 5 0 ends.

4. Experimental procedures 4.1. Cell fractionation with detergent of P. lividus eggs A cellular fraction was obtained using the non-ionic detergent, salt concentration and temperature employed by Montana et al. (1997). Total RNAs were extracted from the supernatant and the pellet fractions using standard procedures (Maniatis et al., 1982). Proteins were isolated by dissolving the supernatant fraction in 2£ sample buffer (10£ sample buffer: 100 mM Tris (pH 6.8), 20% SDS, 1.4% b-mercaptoethanol, 20% glycerol and 0.2% bromophenol blue), and the pellet fraction in 1£ sample buffer. 4.2. RT±PCR Complementary DNA was synthesized from total egg RNA or from RNA extracted as described above, using superscript reverse transcriptase (Gibco-BRL) and oligo d(T) primer. PCR was performed in a volume of 50 ml containing, 100 mM Tris±HCl (pH 8.3), 50 mM KCl, 2 mM of the appropriate primers for collagen: for PL-56 5 0 AGGTATGTGCATTTGGTG-3 0 and PL-36 5 0 -AGCACACATGTGGTCATT-3 0 or for H2A 5 0 -TCTAGCCAAGAACCATCGCTTC-3 0 and 5 0 -CTTGGGAAGCAGCACAGCCT-3 0 (gift of Dr L. Anello), 0.2 mM dNTP, and 1.5 mM MgCl2. The thermal cycler was programmed at 948C for 3 min and at 948C for 1 min, 588C for 1 min, 728C for 1 min for 25 cycles. An aliquot of each sample was utilized for a second round of PCR for 20, 25 and 30 cycles to determine whether DNA ampli®cation was linear. An aliquot of the reactions (10 ml for collagen samples and 2 ml for H2A samples) was analyzed on 1% agarose gel.

D.P. Romancino et al. / Mechanisms of Development 99 (2000) 113±121

4.3. Preparation of total cytoplasmic egg proteins Unfertilized and fertilized eggs (2 g) were homogenized in 2 volumes of TGKEG (50 mM Tris (pH 7.5), 25% glycerol, 50 mM KCl, 0.1 M EDTA, 0.5 M dithiothreitol (DTT), 1 mM phenylmethylsulfonyl ¯uoride (PMSF)) on ice. The extracts were cleared of most of the yolk and corical debris by centrifugation 15 700 £ g for 10 min at 48C and then centrifuged in a Beckman type 60 Ti rotor for 60 min at 55 000 rpm. Aliquots of the supernatants were stored at 2808C. 4.4. Drug treatment of eggs 4.4.1. Microtubule drug treatment Eggs were incubated at 188C for 1 h in sea water collected from the Mediterranean sea and ®ltered through a 0.45-mm Millipore ®lter, containing 1 mg/ml colchicine (Sigma), and after fertilization using standard procedures (Giudice, 1973) the eggs were incubated with colchicine for 30 min. 4.4.2. Micro®lament drug treatment Fertilized eggs were resuspended in 50 ml of ®ltered sea water containing 25 mg/ml of cytochalasin B for 1 h at 208C. After these treatments the fertilized eggs were washed in sea water and utilized for in situ hybridization. 4.5. In vitro synthesis of transcripts To synthesize sense or antisense RNA a fragment of COLL1a , containing the 3 0 UTR region, was subcloned in XbaI±XhoI sites of Bluescript plasmid. The plasmid was linearized by digestion at the XhoI or XbaI site and transcription containing T3 or T7 polymerase respectively was carried out utilizing a digoxigenin (DIG-UTP) transcription kit (Roche) according to the manufacturer's instructions. To synthesize sense H2A RNA the plasmid (gift of G. Spinelli) was linearized by digestion at the XhoI site and transcription containing T3 polymerase and the transcription kit described above. To synthesize the sense collagen RNA for a Northwestern experiment the plasmid was linearized by digestion with XbaI enzymes and transcription was carried out utilizing T7 polymerase. To synthesize, instead, the H2A RNA the plasmid was digested with XhoI enzyme and transcription was carried out utilizing T3 polymerase. In both the cases [ 32P]UTP with the Riboprobe kit (Promega) according to the manufacturer's instructions was employed. 4.6. Whole-mount in situ hybridization and immunohistochemistry of P. lividus eggs treated with drugs Eggs treated with colchicine or cytochalsin B as described above or eggs collected 10±15 min after fertilization were ®xed in 2.5% glutaraldehyde and dehydrated with ethanol washes from 10 to 70%. After hydration the eggs

119

were processed according to Romancino et al. (1998). Whole-mount immunohistochemistry of the same batch of eggs, collected 10±15 min after fertilization, was carried out according to Romancino et al. (1998). Anti-a-tubulin (T5168, clone B-5-1-2, Sigma) diluted 1:500 was utilized as primary antibody. 4.7. Gel retardation analysis The egg cytoplasmic extracts (30 mg) were mixed with 10 ng sense DIG-UTP transcribed 3 0 UTR COLL1a mRNA with in the binding conditions described by Montana et al. (1997). Competition experiment was carried out by incubating the egg cytoplasmic extract 200- or 400-fold molar excess of speci®c collagen transcript, or 400-fold molar excess of non-speci®c H2A transcript for 10 min before addition of the DIG-UTP transcribed RNA. Electrophoresis of RNA protein complex was carried out in 1.5% agarose gel at 80 V for 3 h. After being transferred to nylon membrane the ®lter was processed according to the procedure of Montana et al. (1997). 4.8. Northwestern experiment Thirty micrograms of proteins obtained either from cytoplasmic extracts or by cell fractionation with detergent and 50 ng of puri®ed LP54 or BSA were electrophoresed in 10% SDS±PAGE. The Western blot was incubated in Binding Buffer (BB) (10 mM Tris (pH 7.5), 50 mM NaCl, 1 mM EDTA) for 1 h at 308C. Binding reaction was carried out in 1£ BB, 5£ Denhardt's, tRNA (20 mg/ml), and mixed with 1:5 £ 105 cpm/ml of sense [ 32P]RNA for 2 h at 308C. The ®lter was washed three times in BB for 15 min at room temperature and then autoradiographed utilizing X-AR5 ®lm (Kodak). Competition experiments were carried out by incubating the ®lter in binding buffer with 200- or 400-fold molar excess of speci®c (collagen) RNA or with 400-fold molar excess non-speci®c (H2A) RNA competitors for 10 min at 308C before addition of labeled RNA. 4.9. Gel electrophoresis and immunoblotting All the samples (30 mg) obtained from the different procedures described were resuspended in an equal volume of 2£ sample buffer and subjected to electrophoresis on 10% SDS±PAGE. Proteins were transferred onto nitrocellulose membranes using an LKB Multiphor apparatus (LKB, Sweden) and the conditions recommended in the manufacturer's instruction manual. The ®lter was processed according to standard procedures (Harlow and Lane, 1988). Acknowledgements We thank Professor Giovanni Spinelli for the gift of the H2A plasmid and Dr Letizia Anello for the H2A primers.

120

D.P. Romancino et al. / Mechanisms of Development 99 (2000) 113±121

This project was supported by NRC Bilateral Contract grant.

References Bashirullah, A., Cooperstock, R.L., Lipshitz, H.D., 1998. RNA localization in development. Annu. Rev. Biochem. 67, 335±394. Burgess, D.R., Schroeder, T.E., 1977. Polarized bundles of actin ®laments within microvilli of fertilized eggs. J. Cell Biol. 74, 1032±1037. Broadus, J., Fuerstenberg, S., Doe, C.Q., 1998. Staufen-dependent localization of prospero mRNA contributes to neuroblast daughter-cell fate. Nature 391, 792±795. Costa, C., Rinaldi, A.M., Romancino, D.P., Cavalcante, C., Vizzini, A., Di Carlo, M., 1997. Centrifugation does not alter spatial distribution of Bep4 mRNA in Paracentrotus lividus egg. FEBS Lett. 410, 499± 501. Costa, C., Romancino, D.P., Ingrassia, A., Vizzini, A., Di Carlo, M., 1999. Isolation of a trans-acting factor involved in localization of Paracentrotus lividus maternal mRNAs. RNA, 1290±1298. D'Alessio, M., Ramirez, F., Suzuki, H.R., Solursh, M., Gambino, G., 1989. Structure and development expression of a sea urchin ®brillar collagen gene. Proc. Natl. Acad. Sci. USA 86, 9303±9307. D'Alessio, M., Ramirez, F., Suzuki, H.R., Solrush, M., Gambino, R., 1990. Cloning of a fribillar collagen expressed in the mesenchyme cells of developing sea urchin embryo. J. Biol. Chem. 265, 7050±7054. Di Carlo, M., Montana, G., Bonura, A., 1990. Analysis of the sequence and expression during sea urchin development of two members of a butanoextractable proteins. Mol. Reprod. Dev. 25, 28±36. Di Carlo, M., Romancino, D.P., Montana, G., Ghersi, G., 1994. Spatial distribution of two maternal messengers in Paracentrotus lividus during oogenesis and embryogenesis. Proc. Natl. Acad. Sci. USA 91, 5622± 5626. Deshler, J.O., Highett, M.I., Schnapp, B.J., 1997. Localization of Xenopus Vg1 mRNA by Vera protein and the endoplasmic reticulum. Science 276, 1128±1131. Ding, D., Lipshitz, H.D., 1993. Localized RNAs and their functions. Bioessay 15, 651±658. Elisha, Z., Havin, L., Ringel, I., Yisraeli, J.K., 1995. Vg1 RNA binding protein mediates the association of Vg1 RNA with microtubules in Xenopus oocytes. EMBO J. 14, 5109±5114. Ephrussi, A., Dickinson, L.K., Lehmann, R., 1991. oskar organizers the germ plasma and directs localization of the posterior determinant nanos. Cell 66, 37±50. Esposito, J.K., D'Alessio, M., Di Liberto, M., Ramirez, F., 1993. Complete primary structure of a sea urchin type IV collagen a chain and analysis of the 5 0 end of its gene. J. Biol. Chem. 268, 5249±5254. Esposito, J.K., Suzuki, H., Georrijon, C., Garrone, R., Solursh, M., Ramirez, F., 1994. Identi®cation of a cell lineage-speci®c gene coding for a sea urchin a2(IV)-like collagen chain. J. Biol. Chem. 269, 13167± 13171. Ferrandon, D., Elphick, L., Nusslein-Volhard, C., St Johnston, D., 1994. Staufen protein associates with the 3 0 UTR of bicoid mRNA to form particles which move in a microtubule-dependent manner. Cell 79, 1221±1232. Gambino, R., Romancino, D.P., Cervello, M., Vizzini, A., Isola, M.G., Virruso, L., Di Carlo, M., 1997. Spatial distribution of collagen Type I mRNA in Paracentrotus lividus eggs and embryos. Biochem. Biophys. Res. Commun. 238, 334±337. Giudice, G., 1973. Developmental Biology of the Sea Urchin Embryo, Academic Press, New York/London. Gunkel, N., Yano, T., Markussen, F., Olsen, L.C., Ephrussi, A., 1998. Localization-dependent translation requires a functional interaction between the 5 0 and 3 0 oskar mRNA. Genes Dev. 12, 1652±1664.

Harlow, E., Lane, D., 1988. Antibodies: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Harris, P., Osborn, M., Weber, K., 1980a. A spiral array of microtubules in the fertilized sea urchin egg cortex examined by indirect immuno¯uorescence and electronic microscopy. Exp. Cell Res. 126, 227± 236. Harris, P., Osborn, M., Weber, K., 1980b. Distribution of tubulin-containing structures in the egg of the sea urchin Strongilocentrotus purpuratus from fertilization through ®rst cleavage. J. Cell Biol. 84, 668±679. Hesketh, J.E., 1996. Sorting of messenger RNAs in the cytoplasm: mRNA localization and the cytoskeleton. Exp. Cell Res. 225, 219±236. Kim-Ha, J., Smith, J.L., Macdonald, P.M., 1991. oskar mRNA is localized to the posterior pole of the Drosophila oocytes. Cell 66, 22±35. Kim-Ha, J., Kerr, K., Macdonald, P.M., 1995. Translational regulation of oskar messenger RNA by Bruno, an ovarian RNA binding protein, is essential. Cell 81, 403±412. Kislauskis, E.H., Zhu, X., Singer, R.H., 1994. sequences responsible for intracellular localization of b-actin messenger RNA also affect phenotype. J. Cell Biol. 127, 441±451. Kloc, M., Etkin, L.D., 1994. Delocalization of Vg1 mRNA from the vegetal cortex in Xenopus oocytes after destruction of Xlsirt RNA. Science 265, 1101±1103. Macdonald, P.M., Kerr, K., 1998. Mutational analysis of an RNA recognition element that mediates localization of bicoid mRNA. Mol. Cell Biol. 18, 3788±3795. Macdonald, P.M., Kerr, A., Leask, J.L., Smith, J.L., Leask, A., 1993. RNA regulatory element BLE1 directs the early steps of bicoid mRNA localization. Development 118, 1233±1243. Maniatis, T., Fritsch, E.F., Sambrook, J., 1982. Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Micklem, D.R., Adams, J., Grunert, S., St Johnston, D., 2000. Distinct roles of two conserved domains in oskar mRNA localization and translation. EMBO J. 19, 1366±1377. Montana, G., Romancino, D.P., Di Carlo, M., 1996. Cloning, expression and localization of a new member of a Paracentrotus lividus cell surface multigene family. Mol. Reprod. Dev. 44, 36±43. Montana, G., Bonura, A., Romancino, D.P., SbisaÁ, E., Di Carlo, M., 1997. A 54-kDa protein speci®cally associates the 3 0 untranslated region of three maternal mRNAs with the cytoskeleton of the animal part of the Paracentrotus lividus egg. Eur. J. Biochem. 247, 183±189. Montana, G., SbisaÁ, E., Romancino, D., Bonura, A., Di Carlo, M., 1998. Folding and binding activity of the 3 0 UTRs of Paracentrotus lividus bep messengers. FEBS Lett. 425, 157±160. Mowry, K.L., Melton, D.A., 1992. Vegetal messenger RNA localization direct by a 340-nt sequence element in Xenopus oocytes. Science 255, 991±994. Pokrywka, N.J., Stephenson, E.C., 1991. Microtubules mediate the localization of bicoid mRNA during Drosophila oogenesis. Development. 113, 56±66. Ramos, A., Grunert, S., Adams, J., Micklem, D.R., Proctor, M.R., Freund, S., Bycroft, M., St Johnston, D., Varani, G., 2000. RNA recognition by a Staufen double-stranded RNA-binding domain. EMBO J. 19, 997± 1009. Romancino, D.P., Montana, G., Di Carlo, M., 1998. Involvement of the cytoskeleton in localization of Paracentrotus lividus maternal Bep mRNA and proteins. Exp. Cell Res. 238, 101±109. Schwartz, S.P., Asienthal, L., Elisha, Z., Oberman, F., Yisraeli, J.K., 1992. A 69-kDa RNA-binding protein from Xenopus oocytes recognizes a common motif in two vegetally localized maternal mRNAs. Proc. Natl. Acad. Sci. USA 89, 11895±11899. Smilbert, C.A., Wilson, J.E., Kerr, K., Macdonald, P.M., 1996. Smaug protein represses translation of unlocalized nanos mRNA in the Drosophila embryo. Genes Dev. 10, 2600±2609. Spinelli, G., Albanese, I., 1990. Molecular and cellular biology of the sea

D.P. Romancino et al. / Mechanisms of Development 99 (2000) 113±121 urchin embryo. In: Adyoidi, K.G., Adyoidi, R.G. (Eds.). Reproductive Biology of Invertebrates, IBH Press, Oxford/New Delhi, pp. 283±390. St Johnston, D., 1995. The intracellular localization of messenger RNAs. Cell 81, 161±170. Tomita, M., Kinoshita, T., Izumi, S., Tomino, S., Yoshizato, K., 1994. Characterization of sea urchin ®brillar collagen and its cDNA clone. Biochim. Biophys. Acta 1217, 131±140.

121

Wilhelm, J.E., Vale, R.D., 1993. RNA on the move: the mRNA localization pathway. J. Cell Biol. 123, 269±274. Yisraeli, J.K., Sokol, S., Melton, D.A., 1990. A two step model for the localization of maternal mRNA in Xenopus oocytes: involvement of microtubules and micro®laments in the translocation and anchoring of Vg1 mRNA. Development. 108, 289±298.