Sp2 Is a Maternally Inherited Transcription Factor Required for Embryonic Development*

The Sp family of transcription factors is required for the expression of cell cycle- and developmentally regulated genes, and the deregulated expression of a handful of family members is associated with human tumorigenesis. Sp2 is a relatively poorly characterized member of the Sp family that, although widely expressed, exhibits little or no DNA binding or transcriptional activity in human and mouse cell lines. To begin to address the role(s) played by Sp2 in early metazoan development we have cloned and characterized Sp2 from zebrafish (Danio rerio). We report that 1) the intron/exon organization and amino acid sequence of zebrafish Sp2 is closely conserved with its mammalian orthologues, 2) zebrafish Sp2 weakly stimulates an Sp-dependent promoter in vitro and associates with the nuclear matrix in a DNA-independent fashion, 3) zebrafish Sp2 is inherited as a maternal transcript, is transcribed in zebrafish embryos and adult tissues, and is required for completion of gastrulation, and 4) zebrafish lines carrying transgenes regulated by the Sp2 promoter recapitulate patterns of endogenous Sp2 expression.

The Sp family of transcription factors is required for the expression of cell cycle-and developmentally regulated genes, and the deregulated expression of a handful of family members is associated with human tumorigenesis. Sp2 is a relatively poorly characterized member of the Sp family that, although widely expressed, exhibits little or no DNA binding or transcriptional activity in human and mouse cell lines. To begin to address the role(s) played by Sp2 in early metazoan development we have cloned and characterized Sp2 from zebrafish (Danio rerio). We report that 1) the intron/exon organization and amino acid sequence of zebrafish Sp2 is closely conserved with its mammalian orthologues, 2) zebrafish Sp2 weakly stimulates an Sp-dependent promoter in vitro and associates with the nuclear matrix in a DNA-independent fashion, 3) zebrafish Sp2 is inherited as a maternal transcript, is transcribed in zebrafish embryos and adult tissues, and is required for completion of gastrulation, and 4) zebrafish lines carrying transgenes regulated by the Sp2 promoter recapitulate patterns of endogenous Sp2 expression.
Sp/XKLF proteins are sequence-specific DNA-binding proteins that share a highly conserved, carboxyl-terminal DNA binding domain featuring three "zinc fingers" of the Cys 2 -His 2 class (1-3). Four conserved nucleotides within each zinc finger specify the interaction of the DNA binding domain with a nonameric GC-rich sequence that is often located proximal to sites of transcriptional initiation. Sp proteins comprise one subfamily of Sp/XKLF proteins for which nine members (Sp1-Sp9) have been described in mammals. Each family member carries an amino-terminal trans-activation domain that is comprised of three subdomains, termed A, B, and C (4). The A and B subdomains feature alternating regions enriched in glutamine or serine/threonine residues, whereas the C subdomain carries an abundance of charged amino acids (5,6). Each subdomain is sufficient to drive transcription when tethered individually to a DNA binding domain, albeit to levels that are less significant than elicited when multiple subdomains are examined together. trans-Activation of target genes by Sp proteins is de-pendent on post-translational modifications as well as their physical interaction with each other, additional sequence-specific DNA-binding proteins, and components of the basal transcription complex.
The biochemical and functional properties of Sp1, Sp3, and Sp4 have been characterized most thoroughly, and each has been shown to regulate the expression of a constellation of genes involved in fundamental biological processes, including cell cycle control, differentiation, development, and oncogenesis (1,(7)(8)(9). Sp1, Sp2, and Sp3 are expressed in many, if not all, mammalian tissues, whereas the expression of Sp4 -9 is considerably more restricted. Sp-related proteins have been identified in a wide variety of species, including zebrafish and invertebrates, and mechanisms governing Sp-dependent transcription are evolutionarily conserved. For example, mammalian family members can at least partially supplant the functions of invertebrate orthologues (3,10,11).
The creation of animals that lack individual Sp family members has yielded valuable insights into their requirement for metazoan development. Sp1 nullizygous mice perish in mid-gestation and exhibit a broad range of severe developmental abnormalities (12). When analyzed in chimeric animals, loss of Sp1 function leads to a cell autonomous defect that prevents the contribution of Sp1 nullizygous cells to tissues (12). In contrast to these findings, Sp1 appears to be superfluous for the growth and differentiation of mouse embryonic stem cells in vitro, and thus, at least some Sp1 functions can be provided by other transcription factors (12). Sp3 nullizygotes exhibit post-natal lethality due to respiratory failure as well as profound deficiencies in tooth, bone, and hematopoietic development (13)(14)(15). A majority of Sp4-deficient mice perish within a few days of birth. Surviving nullizygous pups are smaller than their wildtype or heterozygous littermates and, although fertile, males are non-productive (16,17). Sp5-deficient mice do not exhibit developmental defects; however, loss of Sp5 function exaggerates the vertebrae loss that is characteristic of mice carrying the Brachyury mutation (18,19). In apparent contradiction to these findings, zebrafish Sp5 has been shown to be a direct target of the Wnt/␤-catenin pathway and is required for mesoderm induction and the posteriorization of neuroectoderm (20 -24). Sp6 nullizygous mice are smaller than wild-type littermates, and 20% perish before 2 months of age due to unknown causes (25,26). Surviving animals initially exhibit delayed tooth development but by one year of age present with an excess number of teeth with complete loss of enamel. Mice lacking Sp7 function have difficulty breathing and die immediately after birth (27). Although possessing an embryonic skeleton that is patterned in a fashion indistinguishable from wild-type animals, Sp7 is required for osteoblast differentiation, and consequently, nullizygotes lack bone entirely. Sp8-deficient mice die during or immediately after birth and present with a variety of defects, including exencephaly of the forebrain and spina bifida as well as loss of nasal, palate, and tail structures (28,29). Thus, although the functions of Sp proteins may partially overlap, it is clear that each performs duties that cannot be provided by other Sp family members or unrelated transcription factors.
Sp2 is an enigmatic member of the Sp family that has yielded few insights into its role(s) in cell or organismal physiology (30 -34). Studies examining Sp2 function(s) in vivo have not as yet been reported, nor has Sp2 been determined to be required for vertebrate development. The Sp2 DNA-binding domain is the least conserved (75%) among Sp family members. Utilizing a PCR-based protocol and recombinant human Sp2 protein, we identified a consensus Sp2 DNA-binding site (5Ј-GGGCGG-GAC-3Ј) that is bound with high affinity (K d ϭ 225 pM) in vitro and is distinct from a consensus site developed for Sp1 (31,35,36). Yet, in transient transfection assays, Sp2 only weakly transactivates promoters carrying consensus Sp2-binding sites or well characterized Sp-dependent promoters that are readily induced by Sp1 and Sp3 (31). In contrast to Sp1, Sp3, or Sp4, little or no soluble Sp2 DNA binding activity has been detected in many human and mouse cell lines despite its widespread expression (31). To identify domains of Sp2 that limit its capacity to stimulate transcription and/or bind DNA, we prepared and characterized the activities of chimeric molecules in which portions of Sp1 and Sp2 were exchanged. Studies using Sp1/Sp2 chimeras revealed that the Sp2 trans-activation and DNA binding domains are each negatively regulated in vitro, and further analyzes have shown that each of these domains carries amino acid sequences that independently target Sp2 to the nuclear matrix (32). It remains uncertain whether the tethering of Sp2 to the nuclear matrix accounts for the limited capacities of Sp2 to bind DNA and stimulate transcription.
As a first step toward identifying Sp2 functions in vivo, we wished to determine whether Sp2 is an essential gene. One adopted strategy was to isolate the zebrafish Sp2 orthologue, determine its pattern of expression, and then eliminate Sp2 function via microinjection of gene-specific morpholinos. As a consequence of these studies we report that 1) the structures and functions of zebrafish and mammalian Sp2 are closely conserved, 2) zebrafish Sp2 is a maternally inherited transcript, is transcribed in embryos as well as adult tissues, and is required for the completion of gastrulation, and 3) a transgene governed by the zebrafish Sp2 promoter faithfully recapitulates patterns of endogenous Sp2 expression.

EXPERIMENTAL PROCEDURES
Cells-Zebrafish Zf4 (CRL-2050) and COS-1 cells were obtained from the ATCC (Manassas, VA) and cultured as suggested by the supplier.
Animals-Zebrafish (Danio rerio) were purchased from EkkWill Waterlife Resources, Ruskin, FL. Cultivation of adult zebrafish was performed as described (37). Adult zebrafish were maintained at 28°C on a 14-h light and 10-h dark cycle, and embryos were collected by natural mating. All procedures were approved by the Institutional Animal Care and Use Committee of North Carolina State University.
Bioinformatic Analysis-Genes that flank zebrafish Sp2 were identified using the NCBI map viewer, and a reference sequence for each gene was identified via the NCBI Entrez Gene data base (38). To verify the identity of hypothetical zebrafish genes, a BLAST analysis was executed using the zebrafish reference sequences as queries, and the hypothetical zebrafish genes were assigned the same symbol as the human gene. The location of these genes in the human and mouse genomes was determined using the NCBI Entrez Gene data base (38).
Whole Mount in Situ Hybridization-In situ hybridization of zebrafish embryos was performed essentially as described (45,46). A 1026-bp fragment of the zebrafish Sp2 3Ј-untranslated region was subcloned into pCR4-TOPO (Invitrogen) using primers Sp2-F7 and Sp2-R9 and the PCR. Sense and antisense digoxigenin-labeled probes were synthesized in vitro using T7 and T3 polymerases, respectively, and a proprietary kit (MEGAscript; Ambion, Inc., Austin, TX) supplemented with DIG RNA labeling mix (Roche Applied Science) and RNAguard TM (GE Healthcare). Labeled RNA probes were purified using a proprietary kit (MEGAclear TM ; Ambion) and hybridized to fixed and permeabilized embryos in 50% formamide. After extensive washes, hybridized embryos were incubated with anti-digoxigenin alkaline phosphatase antibodies (Roche Applied Science), and antigen/antibody complexes were detected with BM Purple (Roche Applied Science). Hybridized embryos were imaged using a Nikon AZ100 Macro/Micro Zoom microscope.
Cloning and Characterization of Zebrafish Sp2 Promoter-An 1175-bp genomic fragment proximal to the zebrafish Sp2 transcriptional start site (as determined by 5Ј-RACE) was iso-lated via the PCR and subcloned at the KpnI and HindIII sites of pGL3-Basic (Promega Corp., Madison, WI), generating pGL3-Sp2. A derivative construct that carries 315 bp of genomic sequences immediately proximal to the Sp2 transcriptional start site was isolated by the PCR and subcloned in pGL3-Basic, generating pGL3-Sp2⌬860. The relative transcriptional activities of pGL3-Basic, pGL3-Sp2, and pGL3-Sp2⌬860 were determined via transient transfection of Zf4 cells in conjunction with phRL-⌬53MLP as indicated above. The sequence of the 1175-bp genomic fragment isolated from the zebrafish Sp2 locus has been deposited in GenBank TM (GU128957).
Generation of Sp2-mCherry and CDH17-EGFP Transgenic Zebrafish-The 1175-bp genomic fragment carried by pGL3-Sp2 was subcloned upstream of mCherry coding sequences and an SV40-derived poly-A signal in plasmid pDB739 (pminiTol2/ MCS (47)), generating pDB739-ZfSp2-mCherry. A TOL2 transposase cDNA in plasmid pDB600 (47) was transcribed in vitro using T3 polymerase and a proprietary kit (mMESSAGE mMACHINE; Ambion), and the resulting capped messages were purified using MEGAclear TM (Ambion). One-to two-cell stage fertilized embryos were microinjected with 20 pg of pDB739-ZfSp2-mCherry and 25 pg of purified TOL2 transposase message. Surviving microinjected embryos were raised to adulthood, and founders were identified after matings with wild-type animals and detection of mCherry in developing embryos via fluorescence microscopy. Embryos that scored positive for mCherry expression were raised to adulthood and interbred with wild-type animals to generate F2 progeny. F2 and F3 progeny were analyzed for mCherry expression during embryogenesis, adolescence, and adulthood using a Nikon AZ100 Macro/Micro Zoom microscope. The CDH17 promoter was isolated from genomic DNA via the PCR using genespecific primers (5Ј-AGATCTCTTATGTATGAGGCA-3Ј and 5Ј-GAGCTCCTCTCCTCAAAAGAC-3Ј) and subcloned upstream of EGFP and an SV40-derived poly-A signal in plasmid pDB739 (pminiTol2/MCS (47)), generating pDB739-CDH17-EGFP. Transgenic founders were obtained and characterized as outlined for Sp2-mCherry transgenic lines and inter-crossed with Sp2-mCherry animals to generate doubletransgenic zebrafish lines. Expression of mCherry and EGFP in double-transgenic animals was localized using a Nikon AZ100-C1 Confocal Macro/Micro Zoom microscope.

Identification of a Candidate Zebrafish
Orthologue of Mammalian Sp2-To identify a candidate Sp2 orthologue, a computer-assisted search of the zebrafish genome (assembly Zv6) was performed using nucleotides encoding the aminoterminal 514 amino acids of human Sp2. This amino-terminal region excluded the Sp2 DNA binding domain, and thus, we reasoned it should identify a candidate orthologue as opposed to irrelevant zinc finger-containing genes. This search identified a closely related hypothetical gene, LOC556835, located on zebrafish chromosome 11. To determine whether this hypothetical gene is indeed a likely orthologue of mammalian Sp2, we sought evidence for syntenic markers in sequences flanking LOC556835. Interrogation of the zebrafish genome identified eight syntenic markers flanking LOC556835 that are shared with sequences adjoining the human and mouse Sp2 genes on chromosomes 17 and 11, respectively (Fig. 1A).
Should LOC556835 be a zebrafish Sp2 orthologue, we reasoned that its amino acid sequence should exhibit greater conservation with human Sp2 than with other Sp family members. To address this issue we performed a phylogenetic analysis using the predicted amino acid sequence of LOC556835 and sequences of nine Sp family members (Sp1-9) from the human and mouse genomes. Consistent with the conclusion that LOC556835 is indeed an Sp2 orthologue, its amino acid sequence was more similar to human Sp2 than paralogues that compose the mammalian Sp family (Fig. 1B). This phylogenetic analysis was extended by comparing the LOC556835 amino acid sequence with Sp2 orthologues from nine vertebrate species. As would be expected based on their evolutionary distance, the amino acid sequence of LOC556835 is more closely related to Xenopus Sp2 than their mammalian orthologues (Fig.  1C). We conclude from these analyses that LOC556835 is indeed an orthologue of human Sp2, and we will refer to this locus as zebrafish Sp2.
Zebrafish Sp2 Is a Maternally Inherited Message and Is Expressed in Developing Embryos and Adult Tissues-To determine whether zebrafish Sp2 is transcriptionally active, a series of RT-PCR assays was performed with gene-specific primers and RNAs harvested from unfertilized eggs, embryos at various stages post-fertilization, and five adult tissues (ovary, liver, kidney, spleen, and intestine). Each RNA sample was examined with oligonucleotide primers (Sp2-F3 and Sp2-R3) spanning predicted exons 3-6 of the zebrafish Sp2 locus. A parallel set of primers designed to amplify zebrafish elongation factor 1 were employed as an internal amplification control. Consistent with the notion that zebrafish Sp2 is expressed in unfertilized oocytes, early development, and in adult tissues, a single amplification product of the predicted size (771 bp) resulted from each of these reactions (Fig. 2). In particular, copious amounts of this RT-PCR product were detected using RNAs prepared from zebrafish ovary and unfertilized oocytes. Sequencing of this amplified product from embryonic liver and kidney RNAs confirmed that each was colinear with predicted exons 3-6 of zebrafish Sp2 (data not shown). Although equivalent amounts of RNA were analyzed in each reaction, the amounts of RT-PCR products appeared to decrease after fertilization, eventually stabilizing after 24 h and then increasing in concert with development. Whether these apparent fluctuations in gene expression reflect differences in Sp2 mRNA abundance will require additional analyses with more highly quantitative assays, such as real-time PCR. We conclude from these results that zebrafish Sp2 is a maternally inherited message and is expressed in developing embryos as well as adult tissues.
The Intron/Exon Structure and Amino Acid Sequence of Zebrafish Sp2 Are Evolutionarily Conserved-Alignments of zebrafish and mammalian Sp2 sequences suggested that irrelevant genomic sequences had been appended erroneously to the 5Ј end of LOC556835. To identify the 5Ј end of zebrafish Sp2 mRNA, a 5Ј-RACE strategy was employed in which randomprimed first strand cDNAs prepared from zebrafish embryo-, kidney-, and spleen-derived RNAs were amplified with an antisense zebrafish Sp2 oligonucleotide. Single amplified PCR products resulted from each of these reactions, and DNA sequencing revealed that each carried identical 5Ј ends that were inconsistent with the first hypothetical exon assigned to LOC556835. cDNAs carrying the entirety of the zebrafish Sp2 coding sequence were isolated subsequently from embryo and kidney first-strand cDNAs via RT-PCR using primers (Sp2-F8 and Sp2-R4) derived from 5Ј and 3Ј Sp2-untranslated sequences. Single amplified products of 1960 bp were recovered, and each was sequenced in its entirety. Two additional oligonucleotides (Sp2-F7 and Sp2-R9) and RT-PCR were employed to amplify additional 3Ј-untranslated sequences, these fragments were sequenced in their entirety and combined with sequences from coding regions, and a 2962-bp zebrafish Sp2 cDNA that includes coding and non-coding sequences was deposited in GenBank TM (EF613286).
To compare the genomic organization of zebrafish and human Sp2, we first identified intron/exon boundaries by aligning the 2962-bp zebrafish Sp2 cDNA sequence with zebrafish genomic DNA. This alignment resulted in the identification of eight coding exons that span 24,378 bp of genomic DNA. In comparison, the human Sp2 locus features seven coding exons that span 31,855 bp. In keeping with the conclusion that zebrafish and human Sp2 are orthologues, their intron/exon boundaries are virtually identical (Table 1). For example, each gene begins with two small exons (zebrafish Sp2, 26 and 77 bp; human Sp2, 58 and 81 bp), the first of which is predicted to encode only two amino acids, and each gene encodes a penultimate exon that carries the first two of three zinc fingers. The sole organizational difference between the orthologues is that exons 3 and 4 of zebrafish Sp2 are joined in human Sp2, forming exon 3.
To determine the extent to which the zebrafish Sp2 amino acid sequence is conserved relative to human Sp2, the two sequences were aligned, and amino acid identities and similarities were calculated for each functional domain. The greatest degrees of homology were noted between the DNA binding domains of the orthologues. Amino acids comprising the zinc fingers of human and zebrafish Sp2 are 90% identical, whereas their trans-activation domains are considerably less conserved (49 -67% identity in subdomains A-C) with the degree of amino acid conservation greatest within the C sub-domain. Inspection of the zebrafish Sp2 amino acid sequence reveals that it carries a bipartite nuclear localization sequence (NH 2 -KRKHICHIAGCEKTFRKT-CO 2 H) identified within human Sp2 as well as eight other species examined (data not shown and Ref. (32)). We conclude from these comparative analyses that the genomic organization and amino acid sequence of zebrafish Sp2 is conserved relative to its mammalian counterparts.
Zebrafish Sp2 Localizes to the Nuclear Matrix in a DNAindependent Fashion and trans-Activates an Sp-dependent Promoter Weakly-As a first step toward the functional characterization of zebrafish Sp2, we prepared an expression vector carrying a cDNA that had been tagged at its carboxyl termi-FIGURE 2. Sp2 expression in unfertilized zebrafish eggs, developing embryos, and adult tissues. Total RNAs were prepared from unfertilized zebrafish eggs (0 hpf), developing embryos (6 -72 hpf), adolescents (6 dpf), and adult (ovary, liver, kidney, spleen, intestine) tissues, reverse-transcribed, and subjected to thermal cycling with gene-specific primers (Sp2-F3 and Sp2-R3). Zebrafish ef1a transcripts were amplified in parallel as an internal control. Amplification reactions in which the template was not included (no template) were performed in parallel as a negative control. Reaction products were resolved on an agarose gel and stained with ethidium bromide. nus with a 10-amino acid epitope derived from influenza HA. COS-1 cells were transiently transfected with this expression vector as well as expression vectors encoding human and mouse Sp2 cDNAs that had been epitope-tagged similarly (31). Denatured cell extracts were prepared 48 h post-transfection and examined by Western blotting with an anti-HA antibody. As shown in Fig. 3, expression of zebrafish Sp2 resulted in the synthesis of an 85-kDa protein that approximates in size proteins produced from human and mouse Sp2 cDNAs. An additional HA-tagged protein of ϳ64 kDa was also noted in zebrafish Sp2-transfected cell extracts that is likely a degradation product or results from internal translational initiation. We have reported that human Sp2 is a constituent of the nuclear matrix and associates with this nuclear compartment in a DNA-independent fashion (32). To determine whether zebrafish Sp2 partitions similarly, we prepared an EGFP-Sp2 expression vector (pEGFP-ZfSp2/flu) and transiently expressed this fusion protein in zebrafish Zf4 cells.
Transfected cells and nuclear matrices prepared from transfected cells were examined subsequently by direct fluorescence microscopy. Akin to our previously reported results for human Sp2, expression of EGFP-Sp2 in Zf4 cells resulted in its accumulation within non-nucleolar portions of cell nuclei in a "speckled" pattern (Fig. 3B). Also consistent with our previous results, EGFP-Sp2 associated with the nuclear matrix as indicated by its retention after treatment with micrococcal nuclease and ammonium sulfate (Fig. 3B). Thus, we conclude that zebrafish Sp2 is a karyophilic protein that associates with the nuclear matrix of zebrafish cells in a DNA-independent fashion.
We have reported that Sp2 DNA binding activity is negatively regulated in human and mouse cells and that Sp2 transactivates several Sp-dependent promoters weakly in vitro (31). To determine whether zebrafish Sp2 regulates transcription akin to its mammalian orthologues, a series of transient cotransfection experiments was performed in COS-1 cells with a well characterized, Sp-dependent luciferase reporter gene gov- erned by the hamster DHFR promoter (DHFR-Lux (31)). In keeping with our previous results and consistent with the contention that zebrafish Sp2 is functionally analogous to mammalian Sp2, zebrafish Sp2 stimulated DHFR transcription modestly (Fig. 3C). In keeping with previous results, this DHFRdependent reporter gene was strongly activated in cells transfected with full-length mammalian Sp1 and Sp3 (Ref. 31 and data not shown).
Sp2 Expression Is Distributed throughout the Early Embryo-RT-PCR results presented in Fig. 2 indicate that Sp2 is a maternally inherited message and is expressed during embryogenesis. To determine the distribution of Sp2 message in cells and tissues of developing embryos, sense and antisense digoxigenin-labeled probes derived from a portion of the zebrafish Sp2 3Ј-untranslated region were synthesized in vitro and employed in a series of in situ hybridization experiments. Hybridizations utilizing an antisense probe detected Sp2 message in the developing embryo as early as 3 h post-fertilization (Fig. 4A), throughout epiboly (Fig. 4, B  and C), and segmentation as well as the pharyngula period ( Fig. 4D and data not shown). Sp2 message was distributed widely during the first 48 h post-fertilization, achieving greatest abundance in the head region by 24 h (Fig. 4D). In contrast, a sense probe derived from an identical portion of the Sp2 3Ј-untranslated region did not detect a message at any stage of early development (Fig. 4). We conclude from these results that Sp2 message is expressed early and widely in the developing zebrafish embryo.

An 1175 Base Pair Genomic Fragment Isolated from the Zebrafish Sp2 Locus Regulates Transcription
in Vitro-To identify genomic sequences that specify Sp2 transcription, 1175 bp of genomic DNA proximal to the zebrafish Sp2 transcriptional start site (as determined by 5Ј-RACE) was isolated using the PCR and subcloned upstream of the firefly luciferase gene in pGL3-Basic (creating pGL3-Sp2). Computer-assisted inspection of this sequence revealed a number of putative transcription factor-binding sites, including two candidate CAAT boxes (rectangular boxes, Fig. 5A). An obvious TATA box was not identified; however, a predicted TBP-dependent core element for TATA-less promoters was noted (ellipse, Fig. 5A). To determine whether this genomic fragment carries sequences in common with analogous portions of the human and mouse Sp2 genes, we determined the major transcriptional start sites of human and mouse Sp2 using 5Ј-RACE and aligned zebrafish genomic sequences with genomic sequences immediately proximal to the transcriptional start sites of human and mouse Sp2. This alignment revealed close conservation between human and mouse genomic sequences but only limited homology with zebrafish sequences within a 315-bp region immediately adjacent to the zebrafish Sp2 transcriptional start site (Fig.  5B). Interestingly, the putative CAAT boxes and TBP-dependent core element upstream of the zebrafish Sp2 transcriptional start site overlap with nucleotides that exhibit the greatest sequence conservation with human and mouse genomic sequences (shaded nucleotides, Fig. 5A).
To determine whether the 1175-bp genomic fragment proximal to the zebrafish Sp2 transcriptional start site is capable of directing transcription, zebrafish Zf4 cells were transiently transfected with pGL3-Basic or pGL3-Sp2. Resulting luciferase activities were determined relative to a Renilla luciferase expression construct used as an internal control. As shown in Fig. 5C, mean pGL3-Sp2-directed luciferase activity was ϳ10fold greater than pGL3-Basic. To determine whether zebrafish sequences exhibiting limited homology with human and mouse genomic sequences are functionally significant, we created a second luciferase construct that carries the proximal 315 bp of zebrafish genomic DNA (creating pGL3-Sp2⌬860). Consistent with the notion that the majority of zebrafish Sp2 transcriptional activity is carried by this proximal segment, mean pGL3-Sp2⌬860-directed luciferase activity was equivalent to that of pGL3-Sp2. We conclude from these results that a 315-bp genomic DNA fragment immediately proximal to the Sp2 tran-  FEBRUARY 5, 2010 • VOLUME 285 • NUMBER 6 JOURNAL OF BIOLOGICAL CHEMISTRY 4159 scriptional start site is sufficient to direct transcription in Zf4 cells.

Transgenic Zebrafish Carrying an Sp2-mCherry Reporter Gene Faithfully Recapitulate Patterns of Sp2 Expression in Early
Development and Adulthood-To determine whether the 1175-bp genomic fragment shown to have in vitro promoter activity is sufficient to recapitulate patterns of Sp2 transcription in vivo, we prepared a series of transgenic zebrafish lines in which expression of a fluorescent reporter gene, mCherry, is governed by the Sp2 promoter. This reporter gene was inserted into the zebrafish genome using the TOL2 transposase, four female founders were identified by virtue of the expression of mCherry in their descendants, and F3 progeny were derived via matings with wild-type animals. Dissemination of the mCherry transgene to each generation was monitored by fluorescent microscopy and varied from 36 -43% (F1) to 100% (F3). Consistent with evidence from in situ hybridization experiments (Fig. 4), mCherry expression was detected uniformly in early embryos derived from crosses between founders and wild-type males (Fig. 6, A and  B). Expression of mCherry was detected throughout developing embryos during the first 48 hpf, accumulating within the head region by 24 h (Fig. 6, C and D). To identify the developmental stage at which the Sp2-mCherry transgene begins to be transcribed, we prepared total RNAs at various developmental stages (one-cell stage to 30% epiboly) from embryos derived from crosses between transgenic males and wild-type females. Endogenous Sp2 and mCherry mRNAs were detected subsequently using gene-specific primers and RT-PCR. As shown in Fig. 6E, mCherry expression was detected initially in blastula stage embryos (256 -1000 cells; top row, lane 4), and the abundance of this message increased FIGURE 5. Molecular and functional characterization of the zebrafish Sp2 promoter. A, shown is the genomic sequence upstream of the zebrafish Sp2 transcriptional start site as defined by 5ЈRACE. A 1175-bp genomic fragment immediately adjacent to the zebrafish Sp2 transcriptional start site was cloned and sequenced. A cytosine at Ϫ1 relative to the transcriptional start site is underlined. Shaded nucleotides exhibit the greatest homology with sequences upstream of the major human and mouse Sp2 transcriptional start sites (as identified in B). Putative binding sites for trans-acting factors (rectangles indicate predicted CAAT boxes, and an ellipse indicates the predicted TBP-dependent core element for TATA-less promoters) are indicated. Also indicated is a 315-bp proximal genomic fragment (double-spaced sequence) that exhibits intermittent homology with genomic sequences upstream of the human and mouse Sp2 genes, aligned in B, and analyzed for transcriptional activity in C. B, alignment of genomic sequences upstream of the major transcriptional start sites of the human, mouse, and zebrafish Sp2 genes. Sequences were aligned using ClustalW, and identical (black boxes) and similar (gray boxes) nucleotides were annotated using BoxShade 3.21. C, shown is a transient transcription assay using genomic fragments upstream of the zebrafish Sp2 transcriptional start site. Genomic fragments (1175 and 315 bp) upstream of the zebrafish Sp2 transcriptional start site were cloned upstream of the firefly luciferase gene in pGL3-Basic (pGL3) generating pGL3-Sp2 (Sp2) and pGL3-Sp2⌬860 (Sp2-⌬860), respectively. Zebrafish Zf4 cells were transiently transfected with 90 ng of pGL3-Basic, pGL3-Sp2, or pGL3-Sp2⌬860 as well as a Renilla luciferase reporter gene as a control for transfection efficiency. Levels of mean -fold relative trans-activation (ϮS.E.) are derived from three independent plates of transfected cells. Each plate of transfected cells received a total of 100 ng of DNA.
in subsequent developmental stages (top row, lanes 5 and 6). Consistent with our previous results, endogenous Sp2 message was detected at all developmental stages including unfertilized embryos (Fig. 6E; bottom row, lane 1).
The intensity of mCherry expression in transgenic animals declined beginning 3 days post-fertilization, diminishing to a relatively low expression level in most tissues by 4 days of age (Fig. 7, A and B, and data not shown). As overall levels of transgene expression declined, robust mCherry expression was revealed within the glomerulus, pronephric tubules, and ducts of the developing kidney (Fig. 7, A and B). Indeed, expression within these tissues was coincident with that of CDH17, a well characterized marker of kidney development (Fig. 7B). Low, uniform levels of transgene expression were noted in adults with the exception of the ovary, where mCherry expression was detected in developing oocytes (Fig. 7C). We conclude that the 1175-bp genomic fragment carried by these transgenic animals carries sequences that faithfully recapitulate the developmental and tissue-specific expression of zebrafish Sp2.
Sp2 Is Required for Completion of Gastrulation-Given that Sp2 is a maternally inherited message and is expressed during blastogenesis, we speculated that Sp2 function may be required for one or more early developmental steps. To address this possibility, we synthesized three antisense morpholinos designed to abrogate Sp2 protein expression upon microinjection into fertilized embryos. One synthetic morpholino, Morph4, was designed to prevent Sp2 protein synthesis by blocking translational initiation and elongation. Two additional morpholinos, Morph2 and Morph3, were designed to prevent the splicing of Sp2 mRNA. Morph2 spans splice donor sequences flanking the second predicted coding exon (exon 2), whereas Morph3 spans splice donor sequences flanking the third coding exon (exon 3). Zebrafish embryos were microinjected with a 1.3-nl mixture of Sp2 or a control morpholino (0.5-4 ng) at the one-to two-cell stage, and the effects of each on embryonic development were monitored. Consistent with the conclusion that zebrafish Sp2 is required for early development, microinjection of each Sp2 morpholino induced identical, profound developmental defects in a dose-dependent fashion (Fig. 8, B-E, and data not shown). The vast majority (greater than 95%) of embryos microinjected with 4 ng of Morph2-4 arrested development between 50 and 75% epiboly and induced lethality within 24 h. Microinjection of lower concentrations (1-2 ng) of Morph2-4 produced defective embryos that developed grossly abnormal dorsalized features including a decreased anterior/posterior axis, deformed head and tail structures, and partial development of somites (data not shown). To confirm these results, one-to two-cell stage embryos were injected with a control morpholino (Morph7) composed of Morph2 nucleotides that had been synthesized in a randomized order. As expected, microinjection of Morph7 did not result in developmental arrest or aberrations (Fig. 8F). To determine whether microinjection of Sp2-specific morpholinos resulted in the de-regulated maturation of Sp2 message, total RNAs were prepared from Morph2-or Morph 3-injected embryos 9 -10 hpf, and RT-PCR was employed to detect mature Sp2 message. Consistent with expectations, treatment with Morph2 or Morph 3 abrogated or greatly diminished splicing 3Ј of the exon 2 and 3 donor sequences spanned by these synthetic oligonucleotides (Fig. 8G). Given these findings, we conclude that zebrafish Sp2 is required for the completion of gastrulation.

DISCUSSION
The Sp family of transcription factors has been shown to regulate the expression of a diverse set of genes required for cell-cycle progression, differentiation, apoptosis, and general "housekeeping" functions. A subset of Sp family members has been shown to be essential for development, whereas others have been shown to be required for the specification and maturation of specific tissues. Before the work reported here, little was known about the expression of Sp2 in vivo or its requirement for metazoan development. As part of an ongoing effort to study Sp2, we identified a zebrafish Sp2 orthologue and analyzed its functional properties, patterns of expression, and requirement for zebrafish development. These efforts have led to a number of interesting findings. First, consistent with its conserved amino acid sequence, zebrafish Sp2 was shown to be functionally analogous to its mammalian orthologues with respect to subnuclear localization and trans-activation. Second, zebrafish Sp2 is a maternally inherited message, and Sp2 transcription can be detected within blastula stage embryos as well as subsequent stages of development and in adult tissues. Third, a genomic fragment with Sp2 promoter activity was identified, and transgenic animals carrying an Sp2 promoter-dependent  FEBRUARY 5, 2010 • VOLUME 285 • NUMBER 6

Requirement for Sp2 in Early Development
reporter gene recapitulated patterns of endogenous Sp2 expression. Finally, one-to two-cell stage zebrafish embryos microinjected with Sp2-specific morpholinos led to the arrest of embryonic developmental in gastrulation (50 -75% epiboly).
Despite intensive investigations into the biochemical and functional properties of Sp2, its role in cell or organismal physiology has remained uncertain. Widely expressed in human and mouse cell lines, little or no soluble Sp2 DNA binding activity has been detected in these settings (31). The vast majority of human Sp2 is associated with the nuclear matrix in foci (nuclear speckles) that are distinct from promyelocytic oncogenic domains and are largely invariant in size and distribution as a function of cell-cycle progression (32). A high affinity, consensus Sp2 DNA-binding site (5Ј-GGGCGGGAC-3Ј) has been identified for recombinant human Sp2 protein using a PCRassisted cloning strategy, yet Sp2 has shown little or no capacity to trans-activate promoters carrying consensus DNA-binding sites in transient-transfection experiments (31). Careful analyses of Sp1/Sp2 chimeras have shown that this is due to the negative regulation of the Sp2 trans-activation and DNA binding domains in vivo. Moreover, each of these domains carries sequences (nuclear matrix targeting sequences) that function to tether Sp2 to the nuclear matrix (32). Each of these biochemical and functional properties distinguishes Sp2 from other Sp family members, and as a consequence we have argued that Sp2 is likely to play a unique role in cell and/or organismal physiol-ogy. Given its structural and sequence homology with mammalian Sp2, it is perhaps not surprising that zebrafish Sp2 shares these biochemical and functional hallmarks. Yet, the work reported herein extends these findings significantly by revealing that Sp2 is a maternally inherited message that is required for the completion of gastrulation.
RT-PCR results indicate that unfertilized zebrafish oocytes are enriched for Sp2 mRNA. Sp2 message levels appear to diminish after fertilization, eventually stabilizing and remaining at constant levels throughout development and in adult tissues. To begin to define sequences required for Sp2 expression, a genomic fragment adjacent to the zebrafish Sp2 transcriptional start site was isolated and analyzed for promoter activity in transient transfection assays. This 1175-bp fragment stimulated transcription in zebrafish Zf4 cells, and in vitro promoter activity was delimited to a 315-bp fragment proximal to the Sp2 transcriptional start site. This proximal region carries putative CAAT-boxes, lacks an obvious TATA box, and possessed limited sequence homology with analogous genomic sequences upstream of the major transcriptional start sites for human and mouse Sp2. To determine whether sequences necessary to recapitulate patterns of endogenous Sp2 expression were carried by the 1175-bp genomic fragment, transgenic zebrafish lines were prepared in which the expression of an mCherry reporter gene was directed by the Sp2 promoter. Consistent with patterns of endogenous Sp2 expression, mCherry was detected uniformly throughout early embryos and became concentrated in the head region by 24 h post-fertilization. Overall mCherry expression diminished to a low level by 4 days postfertilization, except within the developing kidney, where robust Sp2 expression was detected in the glomerulus, pronephric tubules, and ducts. Indeed, expression of Sp2 in these regions of the kidney was indistinguishable from CDH17, a well characterized zebrafish kidney marker (48). Sp2-directed mCherry expression in adults was quite low with the exception of the ovary, where mCherry was expressed strongly in developing oocytes. This concentration of mCherry expression in the adult ovary is consistent with the observation that Sp2 message is maternally inherited. Using mCherry-specific primers and RNAs prepared from embryos fertilized by transgenic males, transcription directed by the Sp2 promoter can be detected as early as the blastula (256 -1000 cells) stage. Thus, we conclude that the Sp2 genomic fragment employed to generate these transgenic animals carries all sequences necessary to promote embryonic as well as tissue-specific expression in adult animals.
Given their ease of observation and manipulation, zebrafish embryos have provided valuable insights into expression patterns and functions of a variety of transcription factors, including Sp family members. Two orthologues of Sp5, sp5 (also known as bts1) and sp5l (also known as spr2), have been studied in zebrafish. Each is expressed in late blastula stage embryos and is responsible for mediating the induction of mesoderm and neuroectoderm by the Wnt/␤-catenin and fibroblast growth factor signaling pathways (20 -24). Elimination of sp5 causes modest patterning defects of the mesoderm, whereas loss of sp5l does not result in embryos carrying obvious developmental defects. Consistent with the conclusion that the functions of sp5 and sp5l may partially overlap, their simultaneous elimination leads to mesodermal defects that are significantly more severe. Sp8 and Sp9 are expressed during zebrafish somitogenesis within the developing central nervous system and fin buds. Elimination of Sp9 expression in early embryos results in major defects in neurogenesis and brain segmentation, whereas loss of Sp8 and Sp9 expression in 2-day-old embryos is correlated with defects in fin outgrowth (49,50). As for sp5 and sp5l, Sp8 and Sp9 are regulated by Wnt/ ␤-catenin and fibroblast growth factor signals released by the mesenchyme to induce fin development (49). In contrast to these Sp family members, the requirement for Sp2 function appears earlier in development and more absolute. More than 95% of embryos microinjected with three independent Sp2 morpholinos arrested development before the completion of gastrulation (50 -75% epiboly). These results indicate in dramatic fashion that Sp2 is required for one or more functions in early embryonic development that cannot be supplanted by other Sp family members or unrelated transcription factors. These critical functions remain to be discovered; however, they are likely to be evolutionarily conserved. In parallel with studies reported here, we have generated a mouse line that is conditionally nullizygous for Sp2 function and have shown that Sp2 is also required for early mammalian development. 3 Taken together, results from distinct species indicate that Sp2 may be unique among Sp family members in its absolute requirement at the earliest stages of metazoan development. It will be of interest to identify key embryonic targets of Sp2 function as well as determine whether Sp2 is required later in life for the development and/or maintenance of adult tissues. G, total RNAs prepared from control and morpholino-injected embryos were amplified using gene-specific primers and RT-PCR. Left, RT-PCR reactions were prepared with RNAs from control or Morph2-injected embryos using primers complementary to sequences within the first (Sp2-F8) and third (Sp2-R2) coding exons. These primers produce a 348-bp amplification product. Lane 1, no DNA template. Lane 2, amplification using RNA from uninjected control embryos. Lane 3, amplification using RNA from Morph2-injected embryos. Right, RT-PCR reactions were prepared with RNAs from control or Morph3-injected embryos using primers complementary to sequences within the third (Sp2-F11 and Sp2-12) and fourth (Sp2-R6) coding exons. Lanes 1-3 employed primers Sp2-F11 and Sp2-R6 producing a 617-bp amplification product, whereas lanes 4 -6 employed primers Sp2-F12 and Sp2-R6 producing a 379-bp amplification product. Lanes 1 and 4, no DNA template. Lanes 2 and 5, amplification using RNA from uninjected control embryos. Lanes 3 and 6, amplification using RNA from Morph3-injected embryos.