|Home | About | Journals | Submit | Contact Us | Français|
Development via regeneration or budding shares some known genetic pathways with embryogenesis, but no concerted effort has been made to identify genes unique to asexual development. We have identified a novel gene that plays a role in cyclical bud formation and asexual organogenesis in the colonial ascidian Botryllus schlosseri. Athena mRNA is transcribed at high levels during the 24–36 hour interval of programmed cell death and new bud initiation at the conclusion of the budding cycle (takeover). Knockdown of Athena by RNAi and antisense morpholinos induced defects in the development of new buds ranging from retardation in growth and abnormal organogenesis to hollow buds lacking organs. As genetic intervention in this organism has not been possible, this study establishes the use of RNAi and morpholinos in Botryllus as well as describing the knockdown phenotype of a new gene.
Developmental programs have arisen several times during metazoan evolution. Embryogenesis begins from a single cell produced by the fusion of gametes. By contrast, asexual modes of development, which include regeneration and budding, proceed from multiple cells in the adult body. Both agametic and gametic modes of development involve the establishment of body axes, patterning and cell differentiation. Given these common requirements, it would be expected that many genetic pathways would be shared between embryogenesis, budding and regeneration. Indeed homologs of many known morphogens, signaling components and transcription factors including BMPs (Reinhardt et al., 2004), homeobox genes (Bely and Wray, 2001), TGF-β and WNT pathways (Hobmayer et al., 2000; Hobmayer et al., 2001), and T-box genes (Technau and Bode, 1999) have been shown to play similar roles in regeneration of cnidarians and annelids. However, many of the specialized processes that occur in gametes even before fertilization that establish polarity or anticipate protein requirements by stockpiling mRNAs have no parallel in agametic reproduction. Development via budding or regeneration does not occur in isolation, but involves integration of signals from the pool of founding cells as well as nearby differentiated tissues. Genes unique to budding and regeneration certainly exist, although their discovery will require forward approaches in nonstandard model organisms.
Colonial ascidians are ancestral marine chordates endowed with a remarkable capacity for propagation via asexual reproduction. At least nine separate modes of asexual propagation have been recognized across three taxa (aplousobranchs, phlebobranchs, and stolidobranchs) that comprise this group of animals (Nakauchi and Kawamura, 1986). The stolidobranch Botryllus schlosseri undergoes weekly cycles of regeneration from the adult body wall, or palleal budding. A Botryllus colony is comprised of a gelatinous tunic encasing bodies, termed zooids, which grow in rosette-shaped systems that share a common circulatory system and excurrent siphon (Berrill, 1941a; Sabbadin, 1969). Individual zooids possess a heart, digestive tract, oral siphon, and asynchronously arising testes and ovaries (Milkman, 1967). Asexual development in botryllid ascidians is unique as it occurs as a highly coordinated weekly cycle during which three successive generations of asexually derived zooids develop simultaneously and the oldest generation succumbs to apoptotic death and resorption (Berrill, 1941b; Lauzon et al., 1992). Botryllus is an ideal species for the study of palleal budding because developing buds are easily observed though the transparent tunic from both dorsal and ventral sides when the organism is cultured on glass slides.
The morphology of budding in ascidians has been described in remarkable detail (Pizon, 1893; Berrill, 1941a; Berrill, 1941b; Izzard, 1973). Nascent buds progress through the stages of hollow blastula-like sphere to a structure with three tissue layers (Brien, 1968). A few investigators have made forays into the problem of patterning in asexual bud development. Pioneering work by Kawamura in another colonial ascidian, Polyandrocarpa misakiensis demonstrated a role for retinoids in bud axis formation (Kawamura et al., 1993; Katsuyama and Saiga, 1998) and established that bud polarity is determined by position on the parental zooid (Kawamura and Watanabe, 1983; Kawamura, 1984). In Botryllus, work by Izzard (1973) and Sabbadin et al. (1975) has begun to elucidate the role of parental positioning and vascularization in determining bud polarity. However, almost nothing is known about the molecular program of budding in ascidians. We have identified a gene unique to botryllid ascidians that is differentially regulated during the blastogenic cycle and demonstrated by two separate knockdown techniques that it participates in asexual bud development and organogenesis.
The blastogenic cycle in Botryllus is organized into stages A through D (Watanabe, 1953; Fig 1). A single cycle takes place over 7 days at 18°C, but because 3 generations of bodies coexist, development from birth of a bud to death of a zooid requires 3 cycles or 21 days. The appearance of a bud rudiment as a thickening in the atrial wall of the primary bud is coordinated with the onset of a new blastogenic cycle (stage A-1 or day 1) at which point parent zooids open their siphons and begin to feed. The bud primordium grows into a hemisphere that narrows at the base and pinches off to create a closed vesicle. A gut rudiment is formed from a fold of the innermost layers and elaboration of three inner chambers is complete by day 7, which concludes the cycle with the “grandparental” zooid death and its resorption. Following takeover, in which the primary buds replace the former zooids, a new cycle begins as a new bud rudiment grows from the wall of the now primary bud. Organogenesis in the primary buds is completed with the initiation of heartbeat on developmental day 10 (7+3) and the opening of siphons for filter-feeding on day 15 (7+7+1) as the bud becomes a zooid. Death of the zooids occurs at the end of the cycle on day 21 (Berrill, 1941a; Brien, 1968).
Athena was identified by differential display in a screen for developmentally modulated Botryllus transcripts. Athena mRNA is dramatically upregulated during takeover, the 24–36 hour period of apoptotic death and resorption of zooids at the conclusion of the blastogenic cycle, termed stages D-1 through D-4, and nearly undetectable during the rest of the cycle (Lauzon et al., 1996). The full length cDNA sequence corresponding to the differentially expressed fragment, obtained by subcloning, 5′ primer extension, and 5′ RACE, comprises 632 nucleotides with a single open reading frame encoding 142 amino acids (Fig 2A–B; Genbank accession number DQ092483). The developmental upregulation of Athena transcript during takeover was confirmed in several different colonies by Northern blot (Fig. 2C), although rare message could be detected at all stages by nested PCR (data not shown). Repeated efforts to localize Athena transcript by RNA in situ hybridization were not successful. Tiozzo et al. (2005) recently described expression patterns of the homeobox gene Pit-X during embryogenesis and asexual budding in Botryllus schlosseri using in situ hybridization. We too have successfully utilized this technique to localize several genes including cytoplasmic actin in Botryllus (data not shown). Consequently, our lack of success in localizing Athena is likely due to low abundance of expression of the transcript. However, we could detect Athena in surgically excised buds and dying zooids from colonies in takeover by quantitative RT-PCR (Fig 2D). Presence of transcript in both buds and zooids at this stage is surprising, since they appear to be undergoing opposing processes of growth and programmed cell death (Lauzon et al., 1992; 1996).
Potential function of Athena could not be inferred from sequence comparison, as significant homology could not be found to genes of model organisms, including the solitary ascidian Ciona intestinalis. The predicted protein lacks known conserved domains and is largely hydrophilic except for a 20 residue stretch at the N-terminus which contains several hydrophobic residues but does not appear to be a transmembrane domain by available search algorithms. A low-stringency Southern blot with DNA from a related colonial ascidian, Botrylloides simodensis, yielded a single band, suggesting a potential homologue (data not shown).
Genetic knockdowns of Athena were induced with morpholinos (MO), double-stranded (ds) RNA and short-interfering (si) RNA, delivered by injection or soaking. Previous studies demonstrated the specific knockdown of translation by MO in Ciona intestinalis embryos (Satou et al., 2001). However, we wanted to establish the efficacy of knockdown technologies in mature Botryllus since budding begins only after metamorphosis. Experiments were carried out in groups of stage-synchronized clonal replicates in order to control for natural variation in the blastogenic cycle between genotypes. RNAs and morpholinos were co-microinjected with dye into terminal vascular structures called ampullae. Distribution throughout the colony, including zooids and developing buds, occurred via the common blood circulation. In addition, juvenile colonies (1–3 zooids) were subjected to repeated rounds of soaking in RNAs; distribution of RNAs throughout the body likely occurs via the neural gland, which regulates osmolarity (Ruppert et al., 2003).
Blastogenic defects were observed 3 to 12 days following injection with Athena dsRNA and short-interfering RNA (siRNA) and 2 to 7 days following injection with MOs. Regardless of the substance or route of delivery, effects were not observed within the same cycle if treatment occurred after stage A-2. This result implies that translation of Athena during (or at the conclusion of) takeover is required for correct bud development in the following cycle. Defects were rarely observed in zooids or during takeover; in very few cases, resorption of old zooids was delayed into stage A-2 (day 2), whereas it is normally completed in A-1, or day 1. Affected structures included mainly secondary buds and in some cases primary buds as well.
Knockdown phenotypes were categorized as mild, definitive and severe (Table 1). Mild defects consisted of slight delays in the development of secondary buds relative to control colonies and normal progression of the blastogenic cycle. Mildly affected buds were often smaller than control counterparts at some points during development, but appeared normal 1–2 days later. These kinds of aberrations were observed at similar frequencies in all groups, regardless of what treatment, if any, was given (Table 1, column 1), demonstrating that temporary desynchronizations in the blastogenic cycle between generations of buds is within the range of normal development.
Buds that were deficient in organogenesis, retarded in their growth or development more than 36 hours were categorized as definitive phenotypes. We observed failure of secondary bud evagination closure at stage B-2 (Fig. 3A) as compared to stage B-2 control buds (Fig 3B). In addition, developmentally delayed stage A-2 primary buds (Fig. 3C) and delayed stage D-2 secondary buds (Fig. 3E) were observed. Abnormally small buds were often characterized by incomplete organogenesis. Whereas the initiation of three atrial folds delineating atrial from branchial cavities can be observed microscopically in normal C-1 secondary buds, the contents of Athena knockdown buds advanced beyond stage C-1 frequently appeared as a disorganized mass of cells (stage A-2 primary bud in Fig. 3C vs. 3D). In some cases, affected buds appeared to be hollow vesicles (stage D-2 secondary bud in Fig. 3E vs. 3F) or lacked extracorporeal vascular connection (Fig. 3E). This definitive phenotype was observed in knockdown animals produced by injection with dsRNA, siRNA and MOs and could be reproduced by soaking juvenile colonies in dsRNA and siRNA (Table 1).
More potent effects were observed following injection with Athena MOs than either dsRNA or siRNA. In two instances, both primary and secondary buds entirely failed to develop on the zooids on one side of a colony (Fig. 3G); the region of absent buds coincided with the site of MO injection in both cases. Defective developmental timing was noted in two colonies by the appearance of bud rudiments 2 to 3 days early on the wall of secondary buds at stage D-3 (Fig 3G, compared to normal D-3 colony in Fig 3 H). Defects in timing of organogenesis appeared as secondary bud heartbeat initiated three days early in stage D-3 (Fig. 3I) or in a single case as an ectopic beating heart arose in the extracorporeal vasculature of a colony (data not shown).
A significant effect (p=0.0035, chi square=8.53) was found in all classes of knockdown colonies relative to those receiving control RNA or control MOs using an ordinal logistic model. The effects produced by the method of delivery (soak versus injection) or the agent given (dsRNA, siRNA or MO) were statistically indistinguishable. Given the precedence of nonspecific defects induced by dsRNAs in vertebrates(Tuschl et al., 1999; Caplen et al., 2000; Oates et al., 2000; Zhao et al., 2001), each group was compared to its control by chi-square test; we found significant differences between Athena RNAi versus control RNAi groups (p < 0.001) and between Athena MO versus control MO group (p=0.0097). The latter test was repeated with a combined definitive/severe category to produce a probability of 0.017 by Fisher’s Exact test. Finally, injection of Athena MO was found to be more effective than injection of RNA (p=0.01); however, differences in the effectiveness of these two knockdown methods were not significant if data from RNA soaking was included. These analyses served to verify the efficacy of Athena knockdown by three separate methods as well as establishing the use of knockdown techniques in Botryllus.
Inhibition of Athena transcription was verified in several colonies treated with dsRNA or siRNA by real time quantitative RT-PCR. Abundance of Athena mRNA relative to tubulin was established in a normalized cDNA library from all stages and in several normal and control stage D animals (Fig. 4, expressed as the difference in the cycle threshold, ΔCt). Athena cDNA could not be detected in 4 of 4 colonies treated with siRNA by injection or soaking. As expected, the lower ΔCt value obtained for the stage D control animal reflects a greater abundance of Athena transcript present during takeover relative to a normalized, pooled cDNA library.
Histological analysis of colonies with definitive bud defects confirmed the in vivo observations of failed organogenesis and revealed abnormalities in cellular morphology. In sections through primary buds of a colony on day 4 following MO injection, we observed an absence of organ structures (Fig. 5A) relative to stage B-1 buds of the synchronized control colony on day 4 (Fig. 5B). At the time of injection, the stage C secondary buds of this colony were initiating organogenesis with inward folds in the atrial epithelium; two days later, at the onset of the new blastogenic cycle, the former secondary buds (now primary buds) were smaller than those of the dye-injected control clone (Fig. 5D) and failed to grow or develop in the following days. The inner (atrial) epithelial cells of the MO-treated primary bud appear squamous (Fig 5A), whereas secondary bud epithelia normally become cuboidal at the initiation of organogenesis in stage C-1 (Fig 5C; Berrill, 1941a; Kawamura, 1984). With respect to the known expression of Athena late on day 6 as takeover begins, the absence of organ structures and cuboidal morphology of the epithelium implies that this bud underwent developmental regression following MO injection. Interestingly, protrusions on either side of the bud in Fig. 5A may represent attempts at formation of the atrial folds that give rise to the neural complex and branchial sac.
Taken together, these observations suggest that Athena provides a signal during the death of the old generation and takeover by the next generation of zooids that is critical to development of a new set of buds. As with much work in a nonstandard model organism, the questions raised outnumber the answers; however, with the establishment of techniques for genetic knockdown by MO or RNAi in adult Botryllus colonies, perhaps the most important result here is the introduction of new tools for future work.
Mariculture of Botryllus schlosseri colonies was carried out at 18°C as described elsewhere (Boyd and Weissman, 1986). Genotypically identical cycle-synchronized single systems were used as genetic knockdown and control subjects in each set of experiments. Oozooids used for soaking inhibition were cultured on chambered slides. For histology, colonies were fixed in 1% paraformaldehyde/filtered seawater and processed in JB-4 plastic (Polysciences, Warrington, PA), as described (Lauzon et al., 1992). Sections 2 μm thick were generated with a Histoknife (Diatome Inc., Switzerland), stained with toluidine blue and observed under bright field microscopy, as previously described (Lauzon et al., 1992).
Differential display was performed in duplicate on Botryllus total RNA from various blastogenic stages (Lauzon et al., 1996) using the RNAmap kit (GenHunter Corp, Brookline, MA; see also Liang et al., 1993). Bands of interest were re-amplified with the original PCR primers and used as probes for RNA slot blot analysis (Lauzon et al., 1996) in order to confirm the expression pattern. cDNA clones were obtained by probing a Botryllus stage D-2 cDNA library (Unizap cDNA synthesis system, Stratagene) and sequenced with Sequenase (US Biochemical). The 5′ upstream region of Athena was amplified using RACE as described elsewhere (Maruyama et al., 1995).
A high stringency Northern blot was carried out using 32P-labeled probe generated from Athena cDNA by random priming method (Feinberg and Vogelstein, 1983). Quantitation of gene expression was normalized against Botryllus cytoactin in each blot.
Complementary 32P end-labeled oligonucleotides (5′-TTACCATTTAGTGAAACGCC-3′; Operon Technologies) were incubated with total RNA isolated from stage D-2 colonies. cDNAs were extended at 37° C for one hour in a 50 μl reaction mixture containing 50mM Tris-HCl, pH 8.3, 75mM KCl, 10 mM DTT, 3mM MgCl2, 40 units of RNase block (Stratagene), 1mM dNTPs and 50 U MMLV reverse transcriptase (Stratagene) followed by RNase treatment and phenol chloroform extraction. Yeast tRNA (Sigma) was used as a negative control in a separate reaction. cDNA extension products were denatured and fractionated through a 7 M urea/6% polyacrylamide sequencing gel with sequencing ladders generated from M13mp18 using a 32P end-labeled −40 primer (US Biochemicals).
Double stranded RNA (dsRNA) corresponding to the full coding region of Athena was synthesized as described (Kennerdell and Carthew, 1998; Carthew and Kennerdell, 1999). Botryllus cDNA was amplified with PCR primers comprised of a 5′ T7 promoter sequence and 20–24 nucleotides complementary to the 5′ and 3′ untranslated regions of the Athena gene ( 5′-TAATACGACTCACTATAGGGAGACCACAGTCACACGCACGACTATACGA-3′ and 5′-TAATACGACTCACTATAGGGAGACCACACACAGGTGACATGGTCTTCAT-3′; T7 promoter sequences underlined), which generated a 540 bp product. A control dsRNA of the same size was created by amplification of vector sequence from pGEM-T (Promega, Madison, WI) with T7 primer and an adaptor primer of the sequence 5′TAATACGACTCACTATAGGGAGCTCTCCCATATGGTCGACC-3′. Thermocycling parameters used were 10 cycles of 94°-40°-72°C for 30s, 1 min and 1 min, followed by 35 cycles of 94°-55°-72°C for 30s, 1 min and 1 min. Product was extracted, quantified and reverse transcribed. using the RiboProbe In Vitro Transcription System (Promega). Short-interfering (si) RNA oligonucleotides, designed as 21-mers with 2 overhanging 3′ thymidines, were obtained from Dharmacon (Boulder, CO). Athena siRNA comprised the sequences 5′-UAAAACCUGAAUGGGUCGCdTdT-3′ and 5′-GCGACCCAUUCAGGUUUUAdTdT-3′, corresponding to nucleotides 120 to 139. Control siRNA oligos 5′-UUAGCGGAGUGGCAGGUCUdTdT-3′ and 5′-AGACCUGCCACUCCGCUAAdTdT-3′ corresponded to nucleotides 167–186 of pGEM, a sequence that includes a stop codon. siRNAs were purified, annealed, and quantified according to the manufacturer’s instructions, with a final concentration of 50 μM. Integrity of siRNA complexes was verified by a modified RNase protection assay. Briefly, 20 pmols of each siRNA complex was radiolabeled in 10 μl containing 1X polynucleotide kinase buffer, 10 U polynucleotide kinase (New England Biolabs, Beverly, MA), and 10 μCi 32P-γ-ATP (New England Nuclear) for 1 hour at 37°C, then purified on a G-25 sephadex column. Half of the labeled siRNA was treated with 10 units S1 nuclease (Promega) in an 8 μl reaction with 1x nuclease buffer for 1 hour at 37°C. Both S1 nuclease-treated and untreated samples were electrophoresed through 10% acrylamide and autoradiographed. Antisense Athena MO (5′-TCGTTCGATTGAATCGGTAGTCTTC-3′) and a control MO (5′-CCTCTTACCTCAGTTACAATTTATA-3′) were obtained from Gene Tools LLC (Philomath, OR).
Oligonucleotides and RNAs were diluted in Botryllus buffer (25 mM HEPES, 10 mM cysteine, 50 mM EDTA in seawater) and 0.1% phenol red or isosulfan blue prior to microinjection. Final concentrations were 1 mM for MO, 20–200 μM for siRNAs, and 1–2 μM for dsRNAs. 0.1–1.0 μl of RNA or MO was microinjected into the blood vasculature of a single Botryllus system via a proximal ampulla. Oozooids were soaked in a total volume of 1 mL containing 20 pmols RNA for 12 hours at a time every other day. Experimental design included genotype-matched, blastogenic cycle-matched controls for each MO or RNA injection. This was necessitated by the natural variation in blastogenesis, growth rates, and health between genotypes. Observations of each colony in the study were performed under stereomicroscopy every 24 hours independently by D.J.L and R.J.L. Bud diameter measurements were made with an ocular micrometer.
Tissues for RT-PCR analysis were prepared by surgical excision of intact zooids and buds using a Wheeler dissecting knife (Ernest Fullam, Inc., Latham, NY) as previously described (Lauzon et al., 2002). Tissues were flash frozen in liquid nitrogen and maintained at −80°. RNA was extracted with Nucleospin RNA columns (BD Biosciences Clontech, Palo Alto, CA) or by lysis in CHAPS buffer for 30 minutes on ice, followed by centrifugation for 20 minutes at 14000 X G to pellet debris, and isopropanol coprecipitation with 2 μl RNase-free glycogen (Roche, Indianapolis, IN). Samples were treated with 5 U RNase-free DNaseI (Roche) and reverse transcribed with 200 U Superscript (Invitrogen). Real time PCR was carried out in a Perkin Elmer ABI Prism 7000 using SYBR Green detection (Applied Biosystems) with Athena primers 5′GAGCTGCTTCCAGACAATATCACA-3′ and 5′-CGGTTGATGCGCTACATGTT-3′ which amplified nucleotides 269-349. Athena expression was normalized to that of β-tubulin, which was amplified as a 70 base pair product using the primers 5′-GGTACTCGCTCACCAAGTCGTT-3′ and 5′-CTGGCGAAGGTATGGATGAGA-3′. Nested Athena PCR reactions employed diluted product from primers described above. Thermocycling was carried out for10 minutes at 95°C followed by 40 cycles of 95°C for 30s, 56°C for 1 min, and 72°C for 1 min. Analysis of real time PCR was performed with ABI Prism 7000 SDS software (Applied Biosystems).
Data analysis was performed using the JMP statistics package.
The authors wish the acknowledge the efforts of Karla Palmeri and ViKi Takas toward animal care, Katherine Ishizuka for expertise in histological sectioning, the statistical savvy and patience of Chris Harley, the bioinformatics expertise of Eugene Koonin, and the input of Anthony DeTomaso and David Zapol. Support for this work includes NIH grant DK54762-02 to I.L.W. and an NSF Predoctoral fellowship to D.J.L, and an NIH-AREA grant (1R15GM067746-01) and a Union College Faculty research Grant to R.J.L.
Support: NIH grant DK54762-02 to I.L.W., NSF Predoctoral fellowship to D.J.L., NIH-AREA grant (1R15GM067746-01) to R.J.L., Union College Faculty research Grant to R.J.L.