Alternative Splicing of the Drosophila Dscam Pre-Mrna Is Both Temporally and Spatially Regulated

Alternative Splicing of the Drosophila Dscam Pre-mRNA Is Both Temporally and Spatially Regulated

Alicia M. Celotto and Brenton R. Graveley Department of Genetics and Developmental Biology, University of Connecticut Health Center, Farmington, Connecticut 06030 Manuscript received May 31, 2001 Accepted for publication July 6, 2001

ABSTRACT The Drosophila melanogaster Down syndrome cell adhesion molecule (Dscam) gene encodes an axon guidance receptor that can express 38,016 different mRNAs by virtue of alternative splicing. The Dscam gene contains 95 alternative exons that are organized into four clusters of 12, 48, 33, and 2 exons each. Although numerous Dscam mRNA isoforms can be synthesized, it remains to be determined whether different Dscam isoforms are synthesized at different times in development or in different tissues. We have investigated the alternative splicing of the Dscam exon 4 cluster, which contains 12 mutually exclusive alternative exons, and found that Dscam exon 4 alternative splicing is developmentally regulated. The most highly regulated exon, 4.2, is infrequently used in early embryos but is the predominant exon 4 variant used in adults. Moreover, the developmental regulation of exon 4.2 alternative splicing is conserved in D. yakuba. In addition, different adult tissues express distinct collections of Dscam mRNA isoforms. Given the role of Dscam in neural development, these results suggest that the regulation of alternative splicing plays an important role in determining the specificity of neuronal wiring. In addition, this work provides a framework to determine the mechanisms by which complex alternative splicing events are regulated. neurons the central and peripheral nervous systems (Schmucker 250,000ف ROSOPHILA melanogaster has D that are connected in a similar manner in each et al. 2000). Dscam mutants, which are early larval lethal, individual animal. Neurogenesis involves the migration have axon guidance defects in the embryonic ventral of axons, which often occurs over long distances and nerve cord and Bolwig’s nerve (Schmucker et al. 2000). through complex environments, to their targets. It is Together, these results suggest that Dscam encodes an thought that the pattern of neuronal wiring is deter- axon guidance receptor that plays an important role in mined by a molecular code that involves specific interac- specifying neuronal wiring. tions between receptor molecules on a given neuron An extraordinary feature of the Dscam gene is that it and extracellular ligands (Albright et al. 2000). The can potentially produce 38,016 different mRNAs by vir- receptors that participate in this process can be grouped tue of alternative splicing (Black 2000; Schmucker et into at least two classes, axon guidance receptors and al. 2000; Graveley 2001). The Dscam gene contains 115 synaptic cell adhesion molecules (Tessier-Lavigne and exons, 20 of which are constitutive and 95 of which are Goodman 1996). alternatively spliced (Figure 1). The alternatively spliced The D. melanogaster cell adhesion molecule (Dscam) exons are organized into four clusters. The exon 4, 6, gene (Schmucker et al. 2000), which encodes an axon 9, and 17 clusters contain 12, 48, 33, and 2 variable guidance receptor with similarity to the human Down exons, respectively. The exons within each cluster are syndrome cell adhesion molecule (Yamakawa et al. 1998), alternatively spliced in a mutually exclusive manner was recently identified by virtue of the interaction of its such that Dscam mRNAs contain only one variable exon protein product with Dock, an adaptor protein involved from each cluster. Alternative splicing of exons 4, 6, in a signal transduction pathway required for normal and 9 alters the encoded protein sequence of three axon guidance (Garrity et al. 1996). The Drosophila extracellular immunoglobulin repeats. It is possible that Dscam protein possesses an extracellular domain con- different Dscam isoforms may interact with distinct sets taining 10 immunoglobulin repeats and four fibronectin of axon guidance cues. Alternative splicing of the Dscam type III domains, which are connected to a transmem- pre-mRNA may therefore be central to the mechanisms brane domain and a cytoplasmic domain (Schmucker et specifying neuronal wiring. Thus, understanding the al. 2000). Dscam mRNA is expressed in neurons in both mechanisms of Dscam alternative splicing may provide insight into the genetic basis of neurogenesis in addition to providing an important model for investigating the Corresponding author: Brenton R. Graveley, Department of Genetics regulation of complex alternative splicing events. and Developmental Biology, University of Connecticut Health Center, 263 Farmington Ave., Farmington, CT 06030-3301. To begin to understand Dscam pre-mRNA alternative E-mail: [email protected] splicing in greater detail, we determined whether the

Genetics 159: 599–608 (October 2001) 600 A. M. Celotto and B. R. Graveley alternative splicing of this gene is regulated. For the kuba5). In each case, the exon 3 primer was end labeled with ␥ 32 purposes of this study we focused our attention on the [ - P]ATP and T4 polynucleotide kinase. PCR was carried out for 45 sec at 94Њ, 45 sec at 55Њ, and 1 min at 72Њ for 35 exon 4 cluster, which contains 12 variable exons. We cycles using Taq DNA polymerase (Promega). Initial experi- find that the alternative splicing of exon 4 is regulated ments were performed to determine whether the observed throughout development and in a tissue-specific man- frequency that each exon is utilized fluctuates with the number ner. We also cloned and sequenced the exon 4 cluster of PCR cycles. A sampling of the products every 2 cycles re- of the Dscam gene from a related Drosophila species, vealed that although the total amount of product increased with each cycle, no significant differences were observed in D. yakuba, and found that the alternative splicing of D. the percentage of each band between cycle 20, at which the yakuba Dscam exon 4 cluster is developmentally regu- products are first detectable, through cycle 40, which is beyond lated in a similar manner. Thus, the regulated alterna- the exponential phase of the reaction (data not shown). Thus, tive splicing of the Dscam exon 4 cluster is evolutionarily the extent of amplification of each of the 12 Dscam RT-PCR conserved. These results suggest that the regulation of products with respect to one another is relatively unaffected by cycle number. We believe that this is due to the fact that Dscam alternative splicing plays an important role in all of the products are the same size, amplified by the same neural development. primers, and therefore equally subject to saturation in the reaction. Single-strand conformation polymorphism gel electrophor- MATERIALS AND METHODS esis: To detect PCR products containing each of the exon 4 variants, we used single-strand conformation polymorphism Fly stocks and collection: D. melanogaster, Canton-S strain, (SSCP)/multidetection enhancement (MDE) gels. The SSCP and wild-type D. yakuba (obtained from R. Reenan) were used gels (20 cm ϫ 45 cm ϫ 0.4 mm) contained 0.6ϫ TBE and for all experiments. All Drosophila stocks were grown at room 25% MDE gel solution (BioWhittaker Molecular Applications, temperature (21Њ–23Њ) on standard cornmeal molasses media. Walkersville, MD). Ten microliters of stop solution (95% for- For all time course profiles, flies were placed in a cage and mamide, 1 m NaOH, 0.25% bromophenol blue, 0.25% xylene allowed to lay eggs on grape juice plates with supplemental cyanol) was added to 2 ␮l of each PCR reaction. This solution yeast paste for 1–3 hr. Animals were harvested at various devel- was heated to 95Њ for 2 min, placed on ice for 5 min, and 2 opmental time points, washed with PBS, and stored at Ϫ80Њ ␮l of each sample was loaded onto the gel. Each SSCP gel until needed. was run for 24 hr at a maximum of 8 W and a constant RNA isolation: Total RNA was isolated using the LiCl-urea temperature of 25Њ. The gel was dried and exposed to film method (Auffray and Rougeon 1980). Flies (0.2–0.5 g) at with an intensifying screen at Ϫ80Њ overnight. Quantitation various stages of development were homogenized in 3 ml of was performed using a Molecular Dynamics (Sunnyvale, CA) a3m LiCl, 6 m urea solution and the homogenate was stored storm phosphorimager and ImageQuant software. at Ϫ20Њ overnight. The homogenate was removed and centri- Cloning of Dscam cDNA fragments: RT-PCR products ob- fuged at 13,000 rpm for 5 min. The pellet was resuspended tained using the Dmexon3us and Dmexon5ds (D. melanogas- in 300 ␮lof1ϫ TE (10 mm Tris, pH 8.0, 1 mm EDTA)/0.1% ter) or yakuba3 and yakuba5 (D. yakuba) primers were cloned SDS. The solution was extracted twice with phenol:chloroform into the pCRII-TOPO-TA vector (Invitrogen, San Diego) as and the RNA was precipitated by adding 7.5 ␮lof4m NaCl described by the manufacturer. Clones were picked by random and 750 ␮l of 100% ethanol. and sequenced using the DSCAMex5rev primer (D. melanogas- To examine the tissue-specific profile of Dscam exon 4 ex- ter) or yakuba5 (D. yakuba) to identify one clone containing pression, tissues (legs, wings, antennae, and heads) dissected each variant of exon 4. These clones were used as templates from five animals were ground with a pestle in a microcentri- for PCR using the primers Dmexon3us and Dmexon5ds (D. fuge tube containing 300 ␮lofa3m LiCl, 6 m urea solution melanogaster) or yakuba3 and yakuba5 (D. yakuba) to generate and 50 ␮g of yeast tRNA as a carrier. After storing the solution a set of SSCP migration standards. overnight at Ϫ20Њ the tube was then centrifuged at 13,000 Cloning the exon 4 region of the D. yakuba Dscam gene: D. rpm for 5 min and the pellet was resuspended in 50 ␮lof1ϫ yakuba genomic DNA was prepared from 50 adult flies using TE/0.1% SDS. The RNA solution was extracted twice with the QIAamp DNA mini kit (QIAGEN, Valencia, CA). The D. phenol:chloroform and precipitated by adding 1 ␮lof4m yakuba Dscam exon 4 region was amplified from 300 ng D. NaCl and 250 ␮l of 100% ethanol. yakuba genomic DNA, using the Dmexon3us and Dmexon5ds Primers: The following oligonucleotides were used in this primers. The PCR reaction was carried out for 1 min at 94Њ, study: 1 min at 55Њ, and 15 min at 72Њ for 35 cycles using 2.5 units Ј Ј of LA Taq (TaKaRa, Berkeley, CA). The PCR product was Dmexon3us: 5 -TGCCGACCAAAAAGGACC-3 cloned into the pCRII-TOPO-TA vector (Invitrogen) as de- Dmexon5ds: 5Ј-ACGGATGTGCAGCTCTCCAG-3Ј Ј Ј scribed by the manufacturer. The sequence of the exon 4 yakuba3: 5 -AGGACCCGTCTTTCTCAAGG-3 region of the D. yakuba Dscam gene was determined from two yakuba5: 5Ј-CCAGAGGGCAATACCAGATAC-3Ј Ј Ј independent clones. Automated sequencing was performed DSCAMex5rev: 5 -CGGATGTGCAGCTCTCCAGAGGGC-3 at the University of Connecticut Health Center Molecular Reverse transcription-polymerase chain reaction:Reverse tran- Core Facility. The sequence of the D. yakuba Dscam exon 4 re- scription (RT) reactions contained 20 units RNase inhibitor gion has been deposited in GenBank (accession no. AF385930). (Promega, Madison, WI), 200 units of SuperScript II (GIBCO BRL, Gaithersburg, MD), 5 ␮g of total fly RNA or the entire preparation of tissue RNA, 500 ng of random hexamers, and RESULTS buffers provided by the manufacturer. The reaction was car- Analysis of Dscam exon 4 alternative splicing: The ried out at 42Њ for 1 hr. Three microliters of the RT reaction was used as a template for polymerase chain reaction (PCR) Drosophila Dscam gene contains 95 alternative exons using primers that anneal to exons 3 and 5 (D. melanogaster, arranged in four clusters (Figure 1). The exon 4, 6, 9, Dmexon3us and Dmexon5ds; D. yakuba, yakuba3 and ya- and 17 clusters contain 12, 48, 33, and 2 variable exons, Dscam Alternative Splicing 601

Figure 1.—Organization of the D. melanogaster Dscam gene. The D. melanogaster Dscam gene contains 115 exons spanning bp. Twenty exons are 60,000ف constitutively spliced (open boxes) and 95 exons are alternatively spliced (shaded boxes). The alternatively spliced exons are organized into four clusters (exons 4, 6, 9, and 17) that contain 12, 48, 33, and 2 alternative exons each. The exons within each cluster are alternatively spliced in a mutually exclusive manner. The exon 4 cluster is enlarged to depict the relative spacing of the 12 exon 4 variants in this region.

respectively. We are interested in understanding the Dscam exon 4 in D. melanogaster: To determine if Dscam alternative splicing of the Dscam transcripts in greater alternative splicing is regulated, we measured the fre- detail and in particular whether the alternative splicing quency at which each Dscam exon 4 variant is utilized is regulated. For the purposes of these studies, we fo- throughout development. Total RNA harvested from cused on the alternative splicing within the exon 4 flies at various stages of development was used as a tem- cluster. All of the exon 4 variants are very similar in size, ranging from 159 to 171 nucleotides (nt; 4.11 ϭ 159 nt; 4.1, 4.2, 4.3, 4.5, 4.6, 4.7 ϭ 162 nt; 4.9 ϭ 168 nt; and 4.4, 4.8, 4.10, 4.12 ϭ 171 nt). Likewise, the 12 possible RT-PCR products obtained using primers in the constant exons 3 and 5 differ from one another by only 12 nt. As a result, traditional methods such as aga- rose gel electrophoresis cannot be used to analyze the alternative splicing of exon 4. By separating the RT- PCR products on a SSCP gel, which separates molecules based on conformational differences (Orita et al. 1989), we were able to distinguish the majority of the 12 RT- PCR products from one another (Figure 2). The identity of each band was assigned in two ways. First, standards were generated using PCR products generated from 12 cDNA clones containing each exon 4 variant spliced to exons 3 and 5 (lanes 2–13). Second, each band from a reaction similar to that shown in lane 1 was excised from the SSCP gel, cloned, and sequenced. Although the majority of the RT-PCR products migrate at a distinct position in the gel, the RT-PCR products obtained from some mRNAs comigrate. Specifi- cally, RT-PCR products containing exons 4.3 and 4.12 migrate as a single band as do the RT-PCR products Figure 2.—Analysis of Dscam alternative splicing by SSCP synthesized from mRNAs containing exons 4.5, 4.7, and gel electrophoresis. An RT-PCR reaction was performed using 4.9. Using this method, we can determine the relative a 32P-labeled exon 3 primer and an unlabeled exon 5 primer frequency with which the majority of the exon 4 variants with total RNA isolated from adult flies (lane 1). In addition, are utilized within each RNA sample. The percentage individual Dscam cDNA clones containing each of the exon 4 inclusion of each exon is calculated by dividing the total variants (lanes 2–13) were used as templates in PCR reactions with the same primers as in lane 1. The reactions were then pixels in each individual band by the sum of the pixels resolved on an SSCP gel for 24 hr at 25Њ. The gel was dried in all bands in a single lane. and exposed to film. The identities of the bands are indicated Developmentally regulated alternative splicing of on the left. 602 A. M. Celotto and B. R. Graveley

plate for RT-PCR reactions with primers in exons 3 and 5 and the RT-PCR products were resolved on SSCP gels (Figure 3). The frequency at which most of the exon 4 variants are utilized does not change significantly throughout development. However, the splicing of two exons, 4.2 and 4.8, appears to be highly regulated (Table 1). Exon 4.2 displays the most striking developmental of the Dscam transcripts in 0- to 12- %1ف changes. Only hr embryos contain exon 4.2. However, in first instar larvae (L1), exon 4.2-containing transcripts make up of the total Dscam mRNAs. An analysis of RNA %20ف isolated hourly from embryos raised at 22Њ revealed that Dscam transcripts containing exon 4.2 first appear at hour 12, which corresponds to embryonic stage 15 (data not shown). The relative abundance of Dscam transcripts containing exon 4.2 remains high throughout the re- of the %44ف ,mainder of development and, in adults total Dscam mRNAs contain exon 4.2. This represents a 40- to 50-fold increase in exon 4.2 utilization between embryos and adults. The expression of Dscam mRNAs containing exon 4.8 is the opposite of the expression pattern of exon 4.2- containing transcripts. Approximately 20% of all Dscam mRNAs in embryos contain exon 4.8. The abundance of exon 4.8-containing transcripts decreases throughout %1ف the remainder of development and in adults only of the total Dscam mRNAs contain exon 4.8. We conclude that alternative splicing of some of the Dscam exon 4 variants is dramatically regulated throughout development. Tissue-specific alternative splicing of Dscam exon 4 in D. melanogaster: The diversity of Dscam proteins generated by alternative splicing is thought to play an important role in determining the specificity of neuronal wiring (Schmucker et al. 2000). One prediction of this model is that neurons in different tissues would express Figure 3.—The alternative splicing of Dscam exon 4 is devel- different Dscam isoforms to direct their axons to specific opmentally regulated. (A) RT-PCR reactions were performed, using as a template RNA harvested from flies at various points addresses. To begin testing this model, we examined in development. RNA was harvested from embryos collected the relative abundance of each exon 4-containing Dscam every 3 hr for 12 hr; from larvae collected at the L1, L2, mRNA isoform in different adult tissues. nonwandering L3 (L3N), and wandering L3 (L3W) stages; RNA was harvested from antennae, heads, wings, and and from prepupae, pupae and pharate adults, and newly legs dissected from adult flies. These RNA samples were eclosed adults. The RT-PCR reactions were resolved on an SSCP gel, which was dried and quantitated using a Molecular subjected to RT-PCR and the products were separated Dynamics phosphorimager. The frequency at which an exon on SSCP gels (Figure 4). These results show that the is utilized in each sample was calculated by dividing the total collection of Dscam transcripts is significantly different %45ف pixels in each band by the sum of the pixels in all bands in in each body part examined. For example, whereas of the %42ف a single lane. The data shown are representative of at least of the total Dscam transcripts in legs and ف -three independent experiments. (B) The data in A were quan titated and the behavior of representative exons is graphically transcripts in wings contain exon 4.2, only 16% of represented. Some exons, such as exon 4.8, decrease through- Dscam transcripts isolated from heads contain exon 4.2 out development, whereas other exons, such as exon 4.2, in- (lanes 2–4). Similar differences are observed for many crease throughout development. Most exons, however, do not of the other exon 4 variants (see Table 2). We conclude change significantly throughout development, represented that the alternative splicing of the Dscam exon 4 variants here by exon 4.1. The quantitated data are summarized in Table 1. is regulated in a tissue-specific manner. Cloning and analysis of the exon 4 region of the D. yakuba Dscam gene: The most striking change we observed is the developmental regulation of exon 4.2, which is not utilized in early embryos. To determine Dscam Alternative Splicing 603 2.4 1.8 2.4 1.5 0.9 5.4 6.6 3.3 5.1 Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ 1.3 12.9 1.0 5.8 1.2 10.9 0.9 6.9 1.8 2.5 0.6 4.4 3.7 44.2 1.4 11.1 4.4 1.3 Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ least three independent Pharate 1.0 4.4 0.6 9.3 1.3 7.1 1.7 8.3 1.5 6.6 1.3 9.2 1.2 31.7 0.9 12.7 2.0 10.7 Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ 1.1 9.2 1.4 8.5 1.8 16.1 1.6 14.3 0.9 11.5 1.4 6.8 2.4 8.1 2.5 8.6 1.9 17.1 Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ D. melanogaster 1.1 6.3 0.9 5.8 2.4 16.0 0.8 15.0 0.7 11.0 2.0 11.0 2.3 12.3 2.0 11.1 1.7 11.5 Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ 3.2 12.8 0.7 10.5 0.9 11.0 1.2 5.0 7.9 18.8 1.9 12.6 7.0 11.4 5.3 12.7 0.5 5.3 Figure 4.—The alternative splicing of Dscam exon 4 is regu- Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ lated in a tissue-specific manner. RNA harvested from tissues (antenna, head, wings, and legs) dissected from adult flies was used as a template in RT-PCR reactions using primers in

exon 4 alternative splicing in exons 3 and 5. The reactions were resolved on an SSCP gel, TABLE 1 1.5 7.7 0.3 6.9 0.5 6.9 0.9 11.1 0.6 7.6 1.4 18.7 1.9 14.6 1.8 14.3 1.3 12.2 which was dried and quantitated with a Molecular Dynamics Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ phosphorimager (see Table 2). The data shown are represen-

Dscam tative of three independent experiments.

whether the alternative splicing of exon 4.2 has been 0.4 22.8 1.1 5.5 0.6 7.7 2.4 7.0 3.7 10.0 3.8 12.3 1.4 13.0 1.4 9.5 4.0 12.2 conserved in other Drosophila species, we cloned the Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ genomic DNA encompassing the exon 3–5 region of the Dscam gene from D. yakuba, which is estimated to have diverged from D. melanogaster 7–15 million years ago (Powell 1997; Li et al. 1999; Robin et al. 2000). 0.7 0.5 0.9 6.3 0.9 10.7 0.6 8.7 2.5 12.7 2.1 16.2 3.3 14.2 5.8 21.8 2.8 9.1 The sequence of the exon 4 region of the D. yakuba Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ

Developmental regulation of Dscam gene is similar to the D. melanogaster gene throughout its length (Figure 5, Table 3). Like D. melanogaster, the D. yakuba Dscam gene contains 12 variants of exon

Embryo4. Larval stage The exon 4 variants are on average 95% identical 0.5 0.9 1.2 7.2 0.7 8.4 1.2 6.8 1.6 11.2 3.5 19.3 1.1 14.4 5.5 18.3 1.0 13.3 Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ between D. melanogaster and D. yakuba at the nucleotide level. Most of the exonic nucleotide changes are silent third-position changes. As a result, the protein sequences encoded by these exons are nearly identical 1.4 0.7 0.4 7.6 0.9 7.1 1.5 7.4 1.0 11.5 5.1 20.2 2.0 13.6 4.6 19.7 2.8 12.2 between the two species. Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ As expected, the sequences of the intron are more divergent than the exons. The nucleotide sequence of the introns separating the exon 4 variants is an average of 82% identical between D. melanogaster and D. yakuba. The introns between exons 3 and 4.1 and between exons

All data represent the percentage inclusion of each exon calculated as described in the Figure 3 legend. The standard deviation was4.12 calculated from at and 5 are 78 and 77% identical between the two 4.1 11.6 experiments. L3 non, L3 nonwandering; L3 wand, L3 wandering. 4.2 1.6 Exon 0–3 hr 3–6 hr 6–9 hr 9–12 hr L1 L2 L3 non L3 wand Pupae adult Virgins 4.3/12 16.6 4.4 6.4 4.5/7/9 8.3 4.6 16.3 4.8 21.8 4.10 6.7 4.11 10.8 species. 604 A. M. Celotto and B. R. Graveley

TABLE 2 Tissue-speciﬁc alternative splicing of Dscam exon 4 in D. melanogaster

Exon Antenna Head Wings Legs 4.1 4.7 Ϯ 0.4 12.3 Ϯ 3.7 13.6 Ϯ 4.3 11.5 Ϯ 4.3 4.2 26.9 Ϯ 6.8 15.8 Ϯ 2.5 42.4 Ϯ 11.5 44.8 Ϯ 14.6 4.3/12 17.4 Ϯ 11.1 22.8 Ϯ 5.6 11.6 Ϯ 5.9 9.2 Ϯ 2.5 4.4 6.4 Ϯ 2.8 3.6 Ϯ 0.1 3.2 Ϯ 1.6 2.7 Ϯ 0.2 4.5/7/9 14.1 Ϯ 5.8 12.4 Ϯ 1.0 10.8 Ϯ 10.0 7.6 Ϯ 2.4 4.6 7.7 Ϯ 3.8 10.3 Ϯ 1.0 5.7 Ϯ 1.0 8.7 Ϯ 2.8 4.8 10.1 Ϯ 4.7 9.9 Ϯ 2.1 6.9 Ϯ 1.9 8.4 Ϯ 1.1 4.10 3.2 Ϯ 1.6 5.9 Ϯ 1.5 2.0 Ϯ 0.8 3.4 Ϯ 1.8 4.11 9.5 Ϯ 3.7 6.9 Ϯ 1.3 3.8 Ϯ 0.4 3.7 Ϯ 0.3 All data represent the percentage inclusion of each exon calculated as described in the Figure 3 legend. The standard deviation was calculated from three independent experiments.

A comparison of the splice site sequences flanking throughout D. yakuba development as in D. melanogaster each of the exon 4 variants revealed that the 5Ј splice (Figure 6C and Table 4). However, in both cases, the sites are more conserved between the two species than magnitude of the changes is lower in D. yakuba than in the 3Ј splice sites (Table 3). For example, between the D. melanogaster (Figure 6, B and C). We conclude that two species there are only 2 nucleotide changes among a similar developmental pattern of Dscam exon 4-regu- the 12 exon 4 5Ј splice sites, whereas there are 25 nu- lated alternative splicing occurs in both D. melanogaster cleotide changes among the 12 exon 4 3Ј splice sites. and D. yakuba. The regulation of Dscam alternative splicing is conserved between D. melanogaster and D. yakuba: We next DISCUSSION tested whether the developmental pattern of exon 4.2 alternative splicing observed in D. melanogaster is con- The Drosophila Dscam gene can potentially generate served in D. yakuba. The alternative splicing pattern of more alternatively spliced mRNAs than any other gene exon 4 throughout D. yakuba development was deter- known in nature (Black 2000; Graveley 2001). Here, mined by performing RT-PCR, using D. yakuba-specific we examined whether the alternative splicing of this primers in exons 3 and 5 on total RNA harvested from gene is regulated. In particular, we have focused on the embryos, larvae, and adults. The reactions were resolved exon 4 region of the Dscam gene, where the splicing on an SSCP gel to visualize the various exon 4-containing machinery must select 1 of 12 possible alternative exons. PCR products (Figure 6A). The migration pattern of These experiments demonstrated that the alternative the D. yakuba exon 4-containing PCR products differs splicing of this region is controlled throughout develop- from that of D. melanogaster. We therefore determined ment and in a tissue-specific manner. In addition, we the identity of each band with standards derived from found that at least some of this regulated alternative cloned D. yakuba Dscam exon 4-containing cDNAs (data splicing is evolutionarily conserved. This work lays the not shown). In addition, each band was excised from foundation to begin examining the biochemical mecha- the gel, PCR amplified, cloned, and sequenced. As with nisms involved in controlling the extraordinarily com- D. melanogaster, we find that D. yakuba Dscam transcripts plex alternative splicing of this gene, which plays an containing exon 4.2 are not expressed in embryos but important role in determining the specificity of neu- are expressed in both larvae and adults (Figure 6, ronal wiring. A and B, Table 4). We also find that the relative abun- Our most striking observation is the developmental dance of the exon 4.8-containing transcripts decreases regulation of exon 4.2 alternative splicing. This exon is

Figure 5.—Comparison of the exon 4 region of the Dscam gene from D. melanogaster and D. yakuba. The exon 4 region of the Dscam gene was cloned from D. yakuba by PCR using primers in exons 3 and 5. The entire nucleotide sequence was determined and aligned to the sequence of the D. melanogaster Dscam gene using the program Advanced PipMaker (Schwartz et al. 2000). The data are shown as a percentage identity plot. The locations of the exons are shown. The height of the horizontal bars within the graph indicates the degree of homology between the two sequences. Dscam Alternative Splicing 605

TABLE 3 Comparison of the exon 4 region of the D. melanogaster and D. yakuba Dscam genes

Exon Exon Intron Intron Exon 3Ј splice site 5Ј splice site size identity (%) size identity (%) 3NAD.m. CG/GTAAGT 345 96a 1421 78 D.y. CG/GTAAGT 308a 1471 4.1 D.m. CGGCCTTTTCCCAG/TGG D.m. CG/GTGCGA 162 94 177 87 D.y. GGCCTTTTTCCCAG/TGG D.y. TG/GTGCGA 162 174 4.2 D.m. TCCTACCTGTTTAG/TCG D.m. TG/GTGTGT 162 97 361 84 D.y. TCCTTACTGTTTAG/TCG D.y. TG/GTGTGT 162 373 4.3 D.m. CATTGCTGTTTTAG/TCG D.m. CG/GTTTGT 162 98 460 90 D.y. CATTGCTGTTTTAG/TCG D.y. CG/GTTTGT 162 464 4.4 D.m. GAACTCACCTTCAG/TTG D.m. GG/GTACAC 171 94 132 88 D.y. GAACTTACATCCAG/TTG D.y. GG/GTACAC 171 137 4.5 D.m. CTCTTGCTTTACAG/TCG D.m. TG/GTACAG 162 93 124 83 D.y. CTCTTGCTTTACAG/TCG D.y. TG/GTACAG 162 132 4.6 D.m. ATTTTAAATCGCAG/TTG D.m. TG/GTATTA 162 93 114 74 D.y. TTTTCGAATCGCAG/TTG D.y. TG/GTATAA 162 122 4.7 D.m. GCACACCTTTGCAG/TTG D.m. CG/GTGCGA 162 96 107 87 D.y. GAACACCTTTGCAG/TTG D.y. CG/GTGCGA 162 115 4.8 D.m. TATTCGATTCAAAG/TCG D.m. TG/GTACAG 171 95 129 73 D.y. GTTTCGATTCAAAG/TCG D.y. TG/GTACAG 171 131 4.9 D.m. TTCTATCGACTCAG/TGG D.m. GG/GTACTT 168 96 115 84 D.y. TCATTTCGGCTTAG/TGG D.y. GG/GTACTT 168 117 4.10 D.m. CTGATTTCCTTCAG/TTG D.m. TG/GTAATC 171 97 104 74 D.y. CTGATTTTCTTCAG/CTG D.y. TG/GTAACC 171 120 4.11 D.m. CTCCCGTCTTGCAG/TGG D.m. TG/GTAGGA 159 98 137 81 D.y. GCACCGTCTTGCAG/TGG D.y. TG/GTAGGA 159 126 4.12 D.m. CGTACACTTTGCAG/TTG D.m. CG/GTAAAT 171 94 1156 77 D.y. CGTAAACTTTGCAG/TTG D.y. CG/GTAAAT 171 1099 5 D.m. TATCAAAATATCAG/ATG N/A 150 92a D.y. CATCAACATATCAG/ATG 122a D.m., D. melanogaster; D.y., D. yakuba; NA, not available. a Due to our cloning method, the entire portions of exons 3 and 5 were not isolated from D. yakuba. The percentage identity for exons 3 and 5 was therefore calculated only for those portions for which D. yakuba sequence was available. not utilized in early embryos but is present in the major- profiles are examined at a more detailed level. This is ity of Dscam mRNAs in adults. Our finding that exon due to the fact that the tissues we examined undoubtedly 4.2 is similarly regulated in D. yakuba suggests that this contain thousands to tens of thousands of neurons. It splicing event has an important function. A more careful will be critical to determine how many, and which, Dscam analysis revealed that in D. melanogaster, exon 4.2 is first mRNA isoforms are expressed in individual neurons. utilized between 11 and 12 hr of embryogenesis at 22Њ Our comparison of Dscam exon 4 alternative splicing (data not shown). To determine the function of the in D. melanogaster and D. yakuba demonstrates that both exon 4.2-containing Dscam isoforms it will be necessary exon 4.2 and exon 4.8 are regulated in a similar manner to identify which neurons express Dscam mRNAs con- in the two species. However, we also note that some taining exon 4.2. exons are utilized at different frequencies in the two The extraordinary diversity of Dscam isoforms is likely species. For example, in D. melanogaster exons 4.5, 4.7, to be critical to its function as an axon guidance recep- and 4.9 (which migrate as a single band in the SSCP tor. Presumably, individual neurons will be found to gels) together make up between 8 and 14.3% of the express different Dscam isoforms, the collection of total Dscam transcripts throughout development (Table which will determine the path that a neuron will take 1). In contrast, in D. yakuba, 12–23% of the total Dscam in finding its synaptic target. Consistent with this idea, transcripts contain exon 4.5 and 13–16% contain exon we found that the alternative splicing of Dscam exon 4 4.7 (Table 4). Thus, while there are similarities in how is regulated in a tissue-specific manner. Although our the alternative splicing of some of the exon 4 variants experiments demonstrate that there are dramatic differ- is regulated in these two species, there are also signifi- ences in the collection of Dscam isoforms expressed in cant differences in the relative abundance of several the tissues we examined, we imagine that even greater Dscam mRNA isoforms. It is possible that such differ- differences will be observed when the Dscam expression ences could underlie behavioral variations between the 606 A. M. Celotto and B. R. Graveley

Figure 6.—The developmentally regulated alternative splicing of Dscam exon 4 is conserved in D. yakuba. (A) RNA harvested from D. yakuba embryos, larvae, and adults was used as a template in RT-PCR reactions using exon 3 and 5 primers. The reactions were resolved on an SSCP gel, which was dried and quantitated with a Molecular Dynamics phosphorimager. The identity of each band was determined by excising each band, cloning and sequencing it, and using a set of standards generated by using as a template D. yakuba Dscam cDNA clones containing each exon 4 variant. We were unable to identify a band corresponding to exon 4.3, even after sequencing Ͼ100 cDNA clones. The graphs quantitatively compare the relative frequency at which exon 4.2 (B) and exon 4.8 (C) are used throughout development in D. melanogaster and D. yakuba.

two species. In addition, these differences could have exons? We can envision at least two models. The first arisen if the morphological differences of these two is a positive regulatory model. In this case, each exon species require axons to travel along alternate routes to 4 variant would be skipped by default. The binding of find a similar target. a regulatory factor or complex at or near a specific Regulatory mechanisms of Dscam alternative splicing: exon would then activate the splicing of that exon. The Our results demonstrate that the alternative splicing of second model involves negative regulation. In this case, Dscam exon 4 is regulated throughout development and each exon 4 variant would be included by default and in a tissue-specific manner. What types of regulatory the selection of a single exon would require the active mechanisms may act in controlling the splicing of these repression of the other 11 exon 4 variants. Future work will be aimed at defining the biochemical mechanisms TABLE 4 responsible for regulating exon 4 alternative splicing. Comparative sequence analysis can be used to gain Developmentally regulated alternative splicing of insight into important aspects of gene structure and Dscam exon 4 in D. yakuba may highlight potential cis-acting regulatory regions (Kent and Zahler 2000). We cloned and sequenced Exon Embryo Larvae Adult the exon 4 region of the Dscam gene from D. yakuba, 4.1 7.4 Ϯ 0.7 7.3 Ϯ 0.4 3.3 Ϯ 0.2 which is estimated to have diverged from D. melanogaster ;million years ago (Powell 1997; Li et al. 1999 15–7ف Ϯ 0.3 10.7 Ϯ 2.3 14.2 Ϯ 1.7 0.4 4.2 Ϯ Ϯ Ϯ 4.4 8.7 0.05 6.2 0.8 6.8 0.2 Robin et al. 2000). The sequence of the D. yakuba exon 4.5 23.0 Ϯ 2.4 21.5 Ϯ 0.7 21.2 Ϯ 0.9 4 region is quite similar to that of D. melanogaster 4.6/10 9.8 Ϯ 0.4 6.4 Ϯ 0.7 4.8 Ϯ 0.7 4.7 13.1 Ϯ 0.4 15.3 Ϯ 1.9 16.7 Ϯ 0.3 throughout its length (Figure 5). The genes from both 4.8 14.7 Ϯ 1.2 11.1 Ϯ 0.01 9.8 Ϯ 0.4 species contain 12 exon 4 variants and the spatial organi- 4.9/12 10.6 Ϯ 1.0 10.9 Ϯ 1.1 12.1 Ϯ 2.6 zation of the exons and introns is remarkably conserved 4.11 12.6 Ϯ 0.08 10.7 Ϯ 1.0 10.9 Ϯ 1.0 (Table 3). The nucleotide and protein sequences of the Ͼ All data represent the percentage inclusion of each exon alternative exons are highly conserved ( 90% identity), calculated as described in the Figure 3 legend. The standard while the intron sequences flanking the alternative ex- identity). Unfortunately, this %80ف) deviation was calculated from two independent experiments. ons are less similar Dscam Alternative Splicing 607 level of conservation is too high to clearly identify impor- of an organism (Black 2000; Graveley 2001). One tant cis-acting regulatory elements. Determining the se- solution to this paradox is that processes such as alterna- quence of the Dscam exon 4 region from more distantly tive splicing function to generate a tremendously diverse related flies will be necessary to identify potential splic- proteome from a relatively small number of genes. It is ing regulatory elements. estimated that at least 55% of all human genes (Kan et al. To determine whether the splicing of the exon 4 2001), and perhaps as many Drosophila genes, encode variants is regulated by controlling 5Ј or 3Ј splice site alternatively spliced transcripts and some of these are utilization, we compared the sequences of the splice alternatively spliced to generate an amazingly large sites flanking each alternative exon. This analysis re- number of distinct mRNA isoforms (Black 2000; Gra- vealed that the exon 4 5Ј splice sites are more conserved veley 2001). However, the amount of protein diversity between D. melanogaster and D. yakuba than the exon 4 generated by alternative splicing of the D. melanogaster 3Ј splice sites (Table 3). While the Drosophila 5Ј splice Dscam pre-mRNAs is unparalleled (Schmucker et al. site consensus sequence is rather constrained, the 3Ј 2000). Understanding the mechanisms by which Dscam splice site consensus sequence is larger and more re- alternative splicing is regulated will provide insight into laxed (Mount et al. 1992). Thus, the nucleotide differ- not only the basic mechanisms of regulated alternative ences we observe between the two species may simply splicing but also into how the specificity of neuronal reflect neutral evolutionary change. Alternatively, these wiring is genetically encoded. changes may indicate that the regulation of exon 4 Ј We thank D. Black, J. Clemons, A. Das, E. Fleming, K. Hertel, K. alternative splicing may be achieved through 5 splice Lynch, W. Mattox, M. Palladino, R. Reenan, L. Zipursky, and members site selection instead of 3Ј splice site selection. Regard- of our laboratory for discussions and/or comments on the manuscript. less, the analysis of sequences from more distantly re- This work was supported by startup funds and a Research Initiation lated species, coupled with biochemical experiments, and Support Enhancement (RISE) grant (to B.R.G.) from the Univer- will be required to elucidate the mechanisms control- sity of Connecticut Health Center. ling the alternative splicing of these exons. A method for analyzing complex alternative splicing events: The unusual exon-intron structure and alterna- LITERATURE CITED tive splicing patterns of the Dcsam gene required the Albright, T. D., T. M. Jessell, E. R. Kandel and M. I. Posner, 2000 development of a new method for analyzing alternative Neural science: a century of progress and the mysteries that splicing. Given that the 12 exon 4 variants are nearly remain. Neuron 25(Suppl): S1–55. Auffray, C., and F. Rougeon, 1980 Purification of mouse immuno- identical in size, methods such as agaorse gel electro- globulin heavy-chain messenger RNAs from total myeloma tumor phoresis or denaturing polyacrylamide gels were not RNA. Eur. J. Biochem. 107: 303–314. suitable for separating the RT-PCR products derived Black, D. L., 2000 Protein diversity from alternative splicing: a chal- from the Dscam mRNAs. However, we found that SSCP lenge for bioinformatics and post-genome biology. Cell 103: 367– 370. gels were well suited for this analysis. These gels separate Garrity, P. A., Y. Rao, I. Salecker, J. McGlade, T. Pawson et al., DNA molecules on the basis of their conformation, 1996 Drosophila photoreceptor axon guidance and targeting which is determined by their nucleotide sequence requires the dreadlocks SH2/SH3 adapter protein. Cell 85: 639– 650. rather than their size (Orita et al. 1989). Although Graveley, B. R., 2001 Alternative splicing: increasing diversity in the Dscam gene appears to have the most complicated the proteomic world. Trends Genet. 17: 100–107. organization, there are numerous examples of eukaryo- Kan, Z., E. C. Rouchka, W. R. Gish and D. J. States, 2001 Gene structure prediction and alternative splicing analysis using geno- tic genes that contain mutually exclusive alternatively mically aligned ESTs. Genome Res. 11: 889–900. spliced exons of the same size. The vast majority of the Kent, W. J., and A. M. Zahler, 2000 Conservation, regulation, syn- known examples contain only two exons (Stamm et al. teny, and introns in a large-scale C. briggsae-C. elegans genomic alignment. Genome Res. 10: 1115–1125. 2000), but at least one, the Drosophila myosin heavy Li, Y. J., Y. Satta and N. Takahata, 1999 Paleo-demography of the chain (Standiford et al. 1997, 2001), contains multiple Drosophila melanogaster subgroup: application of the maximum tandemly arranged exons. In fact, like Dscam, the Dro- likelihood method. Genes Genet. Syst. 74: 117–127. Mount, S. M., C. Burks, G. Hertz, G. D. Stormo, O. White et al., sophila myosin heavy chain gene contains three clusters, 1992 Splicing signals in Drosophila: intron size, information each containing multiple alternatively spliced exons. In content, and consensus sequences. Nucleic Acids Res. 20: 4255– this case, the exon 7 cluster contains 4 variants, the exon 4262. Orita, M., Y. Suzuki, T. Sekiya and K. Hayashi, 1989 Rapid and 9 cluster contains 3 variants, and the exon 11 cluster sensitive detection of point mutations and DNA polymorphisms contains 5 variants (Standiford et al. 1997, 2001). The using the polymerase chain reaction. Genomics 5: 874–879. relatively simple method we developed could be effec- Powell, J. R., 1997 Progress and Prospects in Evolutionary Biology: The tively used to analyze the alternative splicing of tran- Drosophila Model. Oxford University Press, New York. Robin, G. C., R. J. Russell, D. J. Cutler and J. G. Oakeshott, 2000 scripts from any gene containing mutually exclusive The evolution of an alpha-esterase pseudogene inactivated in the exons. Drosophila melanogaster lineage. Mol. Biol. Evol. 17: 563–575. The complexity of alternative splicing: One of the Schmucker, D., J. C. Clemens, H. Shu, C. A. Worby, J. Xiao et al., 2000 Drosophila Dscam is an axon guidance receptor exhib- most unexpected discoveries of the genomic era is that iting extraordinary molecular diversity. Cell 101: 671–684. gene number does not correlate with the complexity Schwartz, S., Z. Zhang, K. A. Frazer, A. Smit, C. Riemer et al., 608 A. M. Celotto and B. R. Graveley

2000 PipMaker—a web server for aligning two genomic DNA muscle-speciﬁc alternative exon splicing of Drosophila myosin sequences. Genome Res. 10: 577–586. heavy chain transcripts. Genetics 157: 259–271. Stamm, S., J. Zhu, K. Nakai, P. Stoilov, O. Stoss et al., 2000 An Tessier-Lavigne, M., and C. S. Goodman, 1996 The molecular biol- alternative-exon database and its statistical analysis. DNA Cell ogy of axon guidance. Science 274: 1123–1133. Biol. 19: 739–756. Yamakawa, K., Y. K. Huot, M. A. Haendelt, R. Hubert, X. N. Chen Standiford, D. M., M. B. Davis, W. Sun and C. P. Emerson, Jr., et al., 1998 DSCAM: a novel member of the immunoglobulin 1997 Splice-junction elements and intronic sequences regulate superfamily maps in a Down syndrome region and is involved in alternative splicing of the Drosophila myosin heavy chain gene the development of the nervous system. Hum. Mol. Genet. 7: transcript. Genetics 147: 725–741. 227–237. Standiford, D. M., W. T. Sun, M. B. Davis and C. P. Emerson, Jr., 2001 Positive and negative intronic regulatory elements control Communicating editor: A. J. Lopez