Blood.
Extra large 
|
|
|
Figure 1. Scl expression is knocked down by MO injection. (A) Genomic structure of the scl gene. Exons (ex) depicted as boxes, intron sizes not to scale. Atg and spliceMO binding sites and primers for RT-PCR analysis in panel C are shown (red lines and black arrows). The major splice variant obtained after spliceMO injection is indicated by the red line. (B) AtgMO injection blocked translation of a coinjected scl:GFP fusion mRNA that included the MO target sequence. GFP fluorescence was examined at germ ring stage (5.7 hpf) and was found to be strongly reduced in the presence of the atgMO (11 x magnification; views are animal [a] and lateral [l] as indicated). (C) RT-PCR analysis revealed formation of an alternative splice product following spliceMO injection (426 bp, upper red arrow), in addition to residual wt product (401 bp, lower red arrow). (D) DNA sequencing of the alternative splice product revealed the use of a cryptic splice donor 25 bp into the intron (A). These 25 bases (shown in red) caused a frameshift which, when translated, would give rise to a truncated Scl lacking the bHLH. (E) Live 24-hpf embryos (15 x original magnification); lateral views; anterior, left. While spliceMO-injected embryos exhibit normal overall morphology, atgMO-injected embryos displayed severe nonspecific defects by 24 hpf.
- The first MO was designed to target the scl atg and inhibit translation (Figure 1A).
- Coinjection of the atgMO severely reduced the fluorescence detected in scl:GFP mRNA–injected embryos (Figure 1B), showing that the atgMO was active.
- Morphant embryos displayed specific effects on blood gene expression at early stages of development (discussed later) but as development proceeded they exhibited increasingly severe nonspecific defects, including widespread necrosis and shortening of the axis, which prevented analysis of Scl-depletion in older embryos (Figure 1E).
- We therefore designed a second MO, using a splice-blocking approach,36 spanning the splice donor sequence of exon 3, upstream of the essential bHLH domain (for a review, see Lecuyer and Hoang37) (Figure 1A).
- To show that the spliceMO interfered with normal scl pre-mRNA splicing, RT-PCR analysis of spliceMO-injected embryos was carried out, revealing a severe reduction in the amount of normal splice product (Figure 1C, lower red arrow) and instead the formation of an aberrantly spliced RNA (Figure 1C, upper red arrow).
- Sequencing showed inclusion of 25 nucleotides of intron 3, due to use of a downstream cryptic splice donor, which caused a frameshift resulting in translation of a truncated protein lacking the bHLH domain (Figure 1D).
- Importantly though, spliceMO-injected embryos developed without the extensive nonspecific defects observed following injection of the atgMO (Figure 1E), and survived for several days.
- First, the MO substantially reduced scl expression without negative effects on the overall morphology of the embryos (Figure 1E).
- Although the atgMO generated extensive nonspecific defects on injection (Figure 1E), it nevertheless revealed the same defects in hematopoietic gene expression at early stages of development (eg, loss of gata1 and pu.1 expression at 10 somites [14 hpf]; data not shown).
Blood.
Extra large 
|
|
|
Figure 2. Hematopoietic gene expression is lost or severely reduced in Scl-depleted embryos. (A-H) Whole-mount embryos (15 x original magnification); lateral views; anterior, left; numbers of embryos represented in gray. (A, B) Erythroid gata1 and {{beta}}E1 expression was lost in all morphants at 22 hpf. (C-E) In 26-hpf morphants, expression of primitive myeloid genes, pu.1, l-plastin, and mpo, was severely reduced. Remaining cells were restricted to the heart region (arrows) and posterior ICM (arrowheads). (F-H) HSCs were lost in Scl-depleted embryos. Runx1 and c-myb expression at 26 hpf associated with the DA was lost in scl morphants (F, G, arrows). In both cases, nonhematopoietic expression was unaffected (arrowheads).30,35 Ikaros expression in primitive blood and HSCs was lost in 29-hpf morphants (C, arrow). In each case, any remaining expressing cells were restricted to the posterior ICM.
- Expression of gata1, beta embryonic globin 1 (E1) and aminolevulinate synthase 2 (alas2), which, at this stage, is found in the primitive erythroid lineage, was lost entirely (Figure 2A-B, data not shown).38-40 Similarly, expression of pu.1, l-plastin, and mpo, which are expressed in the primitive myeloid lineage (predominantly macrophages), was significantly down-regulated in 26-hpf embryos (Figure 2C-E; see Crowhurst et al27 and references therein).
- Some expressing cells remained under the heart and in the posterior ICM (Figure 2C–E, arrows and arrowheads).
- Expression of runx1 and c-myb at 26 hpf in HSCs was almost entirely lost in morphant embryos (Figure 2F-G, arrows), while expression of ikaros at 29 hpf was lost in both circulating primitive erythrocytes and HSCs (Figure 2H, arrow).
- Scl-negative tissues such as the runx1 and c-myb–expressing nonhematopoietic lineages, the nkx2.5+ cardiac mesoderm, and the pax2.1+ pronephric duct progenitors were completely unaltered (Figure 2F-G and Figure 3A-B).
- Expression of terminal differentiation genes, E1 and alas2, and regulators, gata1, pu.1, c-myb and ikaros, was never observed in morphants at early stages of PLM development (Figure 2A-B, Figure 4A,I, and data not shown).
Blood.
Extra large 
|
|
|
Figure 3. SpliceMO-mediated Scl knockdown is highly specific. (A, B) Wholemount 10-somite (14 hpf) embryos (10.5 x magnification); posterior/anterior views as indicated; dorsal, top; numbers of embryos represented in gray. (A) Coinjection with correctly spliced scl mRNA but not myoD mRNA, together with the scl spliceMO, rescued early gata1 and pu.1 expression in the PLM and the PLM and ALM, respectively. Nkx2.5 in the heart (A) and Pax2.1 in the pronephric duct mesoderm (B) were unaffected by these injections. Ectopic anterior expression of des at 10 somites (14 hpf) confirms that myoD mRNA was active.
- Scl-negative tissues such as the runx1 and c-myb–expressing nonhematopoietic lineages, the nkx2.5+ cardiac mesoderm, and the pax2.1+ pronephric duct progenitors were completely unaltered (Figure 2F-G and Figure 3A-B).
- Third, MO-induced hematopoietic defects were rescued by coinjection of a correctly spliced scl mRNA: expression of gata1 in the PLM, and of pu.1 in the ALM and the PLM, was restored in 10-somite (14 hpf) embryos (Figure 3A).
- Coinjection of the scl spliceMO with an alternative member of the bHLH family, the murine myogenic transcription factor, myoD, did not restore gata1 or pu.1 expression, confirming specificity of the rescue (Figure 3A).
- The myoD RNA was able to ectopically induce the expression of one of its downstream targets, desmin (des), in the head at 10 somites (14 hpf), as previously described,15 thereby proving its activity (Figure 3A).
- As observed with blood markers (Figure 3A), endothelial gene expression could be restored to more normal levels in the lateral plate mesoderm of morphant embryos by coinjection with scl mRNA, although as expected, scl overexpression also caused ectopic endothelial gene expression in the paraxial mesoderm14 (data not shown).
Blood.
Extra large 
|
|
|
Figure 4. Scl is a critical regulator of PLM development. (A-H) Whole-mount embryos (12 x magnification); posterior views; dorsal, top; numbers of embryos represented in gray. Gata1 expression was never observed in the PLM of morphant embryos (A). Runx1 was at first dependent on Scl, initiated weakly in the PLM of morphants at 6 somites (12 hpf) (B, arrow), and by the 10- to 12-somite stage (14 hpf-15 hpf) appeared almost as robust as wt (B, arrowheads). dra was initially expressed normally in morphants, but appeared reduced from 7 somites (12.5 hpf) (C, arrows). The posterior-most region of dra expression was unaffected (C, arrowhead). biKLF expression was lost in morphants by 10 somites (14 hpf) (D). Hhex expression initiated as normal, appeared reduced at 7 somites (12.5 hpf) (E, arrows) and was lost by 10 somites (14 hpf) (E). Gata2, fli1, and lmo2 expression in the PLM was unaffected in 10-somite (14 hpf) morphants (F-H). (I) Data from in situ analyses are summarized schematically. Bars represent the changing expression of genes over time. Bars marked by asterisks represent gene expression in morphants. Purple, Scl-independence; red, Scl-dependence; green, scl expression.
- An advantage of the fish system is that the cells giving rise to blood and endothelium can be observed at approximately 20- to 30-minute intervals as individual somites are added (Figure 4I).
- Scl expression in the PLM initiates between 1 to 2 somites (10.5 hpf), shortly after gastrulation, and is an early marker for these cells14 (Figure 4I).
- Expression of terminal differentiation genes, E1 and alas2, and regulators, gata1, pu.1, c-myb and ikaros, was never observed in morphants at early stages of PLM development (Figure 2A-B, Figure 4A,I, and data not shown).
- Runx1 expression in morphants started at 6 somites (12 hpf) instead of 4 somites (11.3 hpf), but increased until at 10 to 12 somites (14 hpf-15 hpf) it appeared similar to wt (Figure 4B,I).
- Genes in this category include regulators, such as biKLF and draculin (dra), initially expressed more broadly in lateral plate mesoderm before the onset of scl expression, and later restricted to blood precursors.41,42 Control of early expression of these genes must be Scl-independent, but expression of both genes was lost in most of the PLM by 10 somites (14 hpf) in MO-injected embryos (Figure 4C-D,I).
- Reduction in dra expression (ie, Scl dependence) was evident from the 7-somites (12.5 hpf) stage (Figure 4C, arrows; Figure 4I), but residual expression remained in the PLM around the tailbud (Figure 4C, arrowhead).
- Gain-of-function experiments have shown that scl can induce ectopic hhex expression.43 Nevertheless, in scl morphants, hhex expression initiated normally and persisted until 6 somites (12 hpf; Figure 4E,I).
- Genes in this category include gata2 and fli1, which appear in the PLM before scl and are thought to be its activators, and lmo2, which appears in the PLM soon after scl and acts in a complex with it.15,30,38,44,45 In scl morphants, the expression of these genes was unaffected at 10 somites (14 hpf) or earlier (Figure 4F-I, data not shown).
- Fli1 and lmo2 expression in the ALM at 10 somites (14 hpf) was unaffected by Scl-depletion (D, E). (F) Data from such in situ analyses are summarized schematically, as described for Figure 4.
- This loss of expression did not reflect loss of cells as fli1 expression was normal at this stage of development (Figure 4G).
Blood.
Extra large 
|
|
|
Figure 5. Scl is a critical regulator of early myeloid development. (A-E) Whole-mount embryos (13 x magnification); anterior views; dorsal, top; numbers of embryos represented in gray. Pu.1 expression in the ALM initiated weakly in morphants and 1 somite later than wt (A, arrows), remaining weak until 10 somites (14 hpf) (A). dra was expressed normally in Scl-depleted embryos until 6 somites (12 hpf), then began to appear reduced (B, arrows). Similarly, hhex expression was reduced at 7 somites (12.5 hpf), and was lost by 10 somites (14 hpf) (C, arrows). Fli1 and lmo2 expression in the ALM at 10 somites (14 hpf) was unaffected by Scl-depletion (D, E). (F) Data from such in situ analyses are summarized schematically, as described for Figure 4.
- As in the PLM, scl is expressed in the ALM from 1 to 2 somites (10.5 hpf)14 (Figure 5F).
- As seen in the PLM, pu.1, c-myb, and ikaros expression was substantially reduced or lost in the ALM of scl morphants (Figure 5A, arrows, Figure 5F, data not shown).
- Initiation of runx1 expression in the ALM, as in the PLM, was delayed in MO-injected embryos, in this instance from 8 somites (13 hpf) to 10 somites (14 hpf) (Figure 5F, data not shown).
- Expression persisted, albeit weakly, until at least 15 somites (16.5 hpf), although the few runx1+ cells remaining did not migrate (Figure 5F, data not shown).
- However, by 7 somites (12.5 hpf) when dra expression is becoming restricted to the ALM, its expression was substantially reduced in morphants, and was lost by 10 somites (14 hpf) (Figure 5B, arrows, Figure 5F).
- The same profile of hhex expression seen in the PLM (namely switched on normally at 4 somites [11.3 hpf] but starting to decay at 7 somites [12.5 hpf] and gone by 10 somites [14 hpf]), was seen in the ALM (Figure 5C,F).
- Also as seen in the PLM, expression of fli1 and lmo2 was unaffected in the ALM of MO-injected embryos at 10 somites (14 hpf) and earlier (Figure 5D-F, data not shown).
- However, its expression in the ALM at 5 somites (11.6 hpf) was unaffected by Scl-depletion (Figure 5F, data not shown).
Blood.
Extra large 
|
|
|
Figure 6. Endothelial development is severely disrupted in Scl-depleted embryos. (A-C) Flat-mount embryos; anterior, left (A-B, 18 x magnification; C, 21 x magnification). (D-H) Whole-mount embryos; lateral view; close-up of trunk/tail region; anterior, left. (I-M) Wholemount embryos (45 x magnification); lateral view; close-up of trunk region; anterior, left. Numbers of embryos represented in gray. At 10 somites (14 hpf), flk1 and flt4 expression in the ALM and PLM was severely reduced in morphant embryos (A, B). At 15 somites (16.5 hpf), fli1 expression was significantly reduced in morphants (C), and the line of DA precursors forming at the midline was not as distinct as in wt (C, arrows). In 26-hpf morphants, fli1 and tie1 expression was severely reduced in the trunk region and lost in ISVs (D, E, arrows). Flk1 expression in the DA was lost, with remaining expression ventrally and laterally (F). At 26 hpf, expression of flt4 and tie2, mainly restricted to the PCV, was not significantly down-regulated in morphant embryos (G, H). (I-M) DA-but not PCV-specific gene expression was lost in Scl-depleted embryos. Expression of artery-specific genes hrT (I), deltaC (J), notch5 (K), and ephrinB2a (L) was entirely lost in morphants. EphB4 expression in the PCV, however, was not depleted (M).
- To study vasculogenesis in Scl-depleted embryos, we analyzed expression of the VEGF receptors, flk1 and flt4.30,46 Both were substantially reduced in scl morphants (both spliceMO- and atgMO-injected embryos) at 10 somites (14 hpf) (Figure 6A-B, data not shown).
- Around this time, a subset of these VEGF receptor–expressing angioblasts in the PLM begin to migrate toward the midline producing a line of cells that will form the DA, here shown expressing fli1 at 15 somites (16.5 hpf) (Figure 6C, wt, arrow).46 Posterior cardinal vein (PCV) progenitors remain lateral at this time, and eventually coalesce at the midline, ventral to the DA, as a result of more general morphologic movements.
- In MO-injected embryos, the distinct line of DA precursor cells at the midline was barely evident (Figure 6C, arrows).
- In addition, fli1 expression in PCV precursors was severely reduced (Figure 6C, arrowheads).
- In 26-hpf scl morphant embryos, expression of fli1 and the later endothelial marker tie147 in the midline where the 2 axial vessels develop was significantly down-regulated (Figure 6D-E).
- In support of this interpretation, the strong expression of flk1 in the DA and ISVs was absent (Figure 6F).
- In contrast, expression of flt4 and tie2 (tek), which are more restricted to the PCV,47 were not as significantly down-regulated in morphants (Figure 6G-H).
- HrT, notch5, deltaC, and ephrinB2a have artery-specific vascular expression patterns and are restricted to the DA at this early stage of zebrafish development.48 All 4 were entirely lost in morphants, whereas expression of the vein-specific marker, ephB4, was intact (Figure 6I-M).48
Blood.
Extra large 
|
|
|
Figure 7. Genetic regulatory networks controlling early PLM and ALM development. Direct relationships, illustrated by continuous lines, are defined by 3 criteria: (1) target gene expression is affected by perturbation of activator; (2) target gene and activator are coexpressed; (3) target gene promoter/enhancer sequences contain binding sites for activators, or length of time between perturbation of activator and effect on target gene is probably insufficient to allow for synthesis of intermediates. Where criteria 1 and 2 are met but 3 is unknown, the relationship is described as indirect and depicted by a dashed line. Lmo2 does not bind DNA, but is known to be an obligate member of a multiprotein complex containing Scl (represented here by the "and" function).15,37 PLM (Erythroid): A gata1 enhancer contains binding sites for the multiprotein complex containing Scl and Lmo2.49 Our data show that initiation of gata1 expression is Scl-dependent. Initially, runx1 expression is dependent on Scl, whereas hhex and dra are Scl-independent. After 7 somites (12.5 hpf), hhex and dra become dependent on Scl, whereas runx1 expression gradually becomes Scl-independent. ALM (Myeloid): Pu.1 expression is initially dependent on Scl, whereas hhex and dra are independent. After 7 somites (12.5 hpf), hhex and dra become Scl-dependent. Unknown activators of Scl-independent genes are depicted as genes X, Y, and Z. It is possible that X and Y represent the same gene due to similarities in timing of involvement.
- This period of time is unlikely to be long enough for the synthesis of intermediates; therefore, our data suggest that Scl activates gata1 directly in embryonic precursors to erythroid cells (Figure 7).
- Thus, pu.1 looks like a direct target for Scl, acting with Lmo2 and Gata2, in myeloid precursors (Figure 7).
- By 7 somites, however, hhex, dra, and all other hematopoietic gene expression measured in this study had become dependent on Scl, either directly or indirectly (Figure 7).
- Expression of runx1 in the PLM follows that of scl by less than 1 hour, making it another possible direct target (Figure 7).
- Over time, an Scl-independent pathway is able to compensate for the absence of Scl, and runx1 expression levels in both the ALM and PLM increase (Figure 7).