Amphioxus Gli knockout disrupts the development of left–right asymmetry but has limited impact on neural patterning

Introduction

Hedgehog (Hh) signaling is a well-characterized intercellular signaling pathway that functions in many capacities and contexts in the embryonic development of most animals. Canonical examples include parasegment boundary formation in Drosophila, and limb and spinal cord patterning in vertebrates (reviewed by Jiang and Hui 2008). Hh protein is synthesized, processed and exported by Hh gene expressing cells, with the signal received by other cells via the receptor Patched (Ptch). Ptch interacts with another membrane protein, Smoothened (Smo), to relay the signal to the nucleus. The strength of signal received coupled with length of signal exposure can lead to different transcriptional outcomes, meaning Hh can in some contexts work as a morphogen and deliver robust patterning of cells according to their distance from a source of the signal (Briscoe and Small 2015).

Relay of the Hh signal from the Ptch-Smo complex to the nucleus is a complex process. In vertebrates this may involve translocation of proteins to the primary cilium, though the exact mechanism may differ between lineages (Elliott and Brugmann 2019; Huangfu and Anderson 2006). However, a consistent feature of Hh signaling is that it feeds through the Gli transcription factors. Gli proteins have five C2H2 zinc finger domains that mediate DNA binding, as well as many other regions and sites that are functionally significant (Hui and Angers 2011). Importantly, the way that Hh signaling regulates the transcriptional regulatory activity of Gli and hence target gene transcription seems to be deeply conserved in animals (von Mering and Basler 1999). Essentially, in the absence of Hh signaling full length Gli protein is modified from the C-terminus, leaving a truncated N-terminal form that still includes the zinc fingers but which acts as a transcriptional repressor (reviewed by Hui and Angers 2011). When Hh signaling is relayed into a cell, C-terminal modification is inhibited leaving a more full-length Gli protein that acts as a transcriptional activator. This balance between activator and repressor forms of Gli under Hh signal control is one of the ways in which quantitative effects of signaling are realized as different transcriptional outcomes (Briscoe and Small 2015; Sagner and Briscoe 2019).

Most invertebrates have a single Gli gene. However the genome duplications that occurred in early vertebrate evolution have resulted in most vertebrate lineages having three Gli paralogues: Gli1, Gli2 and Gli3 (Abbasi et al. 2009). Studies of mammalian Gli genes show these paralogues have diverged in function. C-terminal modification of both Gli2 and Gli3 proteins is regulated by Hh signaling. However Gli3 tends to function more as a repressor and Gli2 as an activator (Hui and Angers 2011). These characteristics help generate a dynamic network downstream of Hh that may explain some of the morphogenetic properties of the pathway. Mammalian Gli1 functions differently: the protein is shorter than Gli2 and Gli3 and lacks some N-terminal domains, and it tends to act as an activator in a positive feedback circuit downstream of Hh (Hui and Angers 2011; Park et al. 2000). Gli gene function in other vertebrate lineages has not been as well studied as it has in mammals, though it has been noted zebrafish has two gli2 paralogues and that gli1 and gli2 genes may not function the same way as they do in mammals (Karlstrom et al. 2003; Ke et al. 2008).

The cephalochordate amphioxus predates the vertebrate genome duplications but shares many basic features of vertebrate development (Holland et al. 2008; Putnam et al. 2008), including a conserved pathway that regulates organismal asymmetry via the Nodal and Pitx genes (Li et al. 2017) and a dorsal neural tube with organized patterns of cell differentiation (Leung and Shimeld 2019). Amphioxus has a Hh pathway that functions in both these processes, as established by knockout of the amphioxus Hh gene (Wang et al. 2015). Hh knockout compromises the development of left–right asymmetry, leading to animals with symmetrical organization (Hu et al. 2017). Hh knockout also affects the differentiation of cell types in the central nervous system (Ren et al. 2020). Amphioxus has a single Gli gene like most other invertebrates (Shimeld 2008), consistent with amphioxus not sharing the vertebrate genome duplications. Amphioxus Gli, however, is alternately spliced to yield transcripts encoding different isoforms: a full-length protein called GliL and a truncated form missing part of the C terminus called GliS (Shimeld et al. 2007). It was previously considered that these resembled the full-length and C-terminus truncated forms of Gli in other species that are formed by Hh signaling regulating protein processing; this was in part tested in heterologous systems, in which amphioxus Gli isoforms were expressed in vertebrate and Drosophila embryos and the impact on target gene expression monitored (Shimeld et al. 2007). However, Gli function in vivo in amphioxus has not been studied.

Here, we use TALEN gene knockout to examine Gli function in amphioxus. Two knockout lines were developed, one near the 5' end of the gene and a second that just affects the GliL transcript. Gli knockout animals showed compromised development of left–right asymmetry at both morphological and molecular levels and resemble (but are not identical to) Hh knockouts in this regard (Hu et al. 2017). However, neural patterning is only weakly affected compared to the impact of Hh gene knockout. Possible reasons for these differences are discussed.

Results

Generating Gli knockouts

Positioning of TALEN knockout sites is shown in Fig. 1A. One TALEN was designed to the 5' region of the Gli gene and should result in a knockout of the whole gene. A second TALEN was designed near the 5' end of an exon that is specific to the GliL transcript. This should leave GliS functional but eliminate functions specific to GliL, as the mutated GliL transcript closely resembles GliS. Hereafter, these two mutants are referred to as Gli and GliL, respectively (e.g., as Gli^−/− and GliL^−/− when in the homozygous state). It is not possible to design a mutant specific for GliS as the entire coding sequence of GliS is included in GliL. Molecular characterization of knockout lines showed both genetic manipulations resulted in missense mutations (Fig. 1B). In the following analysis, we obtained GliS^−/− and GliL^−/− embryos by crossing GliS^+/− or GliL^+/− animals and distinguished the homozygous embryos from the heterozygous and wild type ones based on their morphological or gene expression differences.

Figure  1.  Generation of amphioxus Gli mutants. A Schematic maps of the GliL and GliS transcripts showing the positions of the Gli TALEN (a) and the GliL TALEN (b). B Sequence comparison of wild type and mutant loci for the Gli TALEN (top) and GliL TALEN (bottom)

DownLoad: Full-Size Img PowerPoint

The role of Gli in left–right asymmetry

Hh knockout amphioxus show characteristic morphological defects in the pharynx related to disruption of left–right asymmetry patterning (Hu et al. 2017). In wild-type animals the mouth forms on the left side, the first gill slit expands on the right side, and pharyngeal organs (preoral pit, endostyle, club-shaped gland) are asymmetrically organized. Hh^−/− animals appear symmetrical: they do not form a mouth opening, gill slits lie on the ventral midline, left-sided pharyngeal structures are duplicated on the right side, and right-sided pharyngeal structures are lost (Hu et al. 2017).

Gli^+/+ and Gli^+/− larvae showed normal asymmetric pharynx morphology as described above in wild-type animals (Fig. 2A). However, like Hh^−/− larvae, Gli^−/− larvae showed symmetrical pharynx morphology (Fig. 2A; Hu et al. 2017): the preoral pit was duplicated on both sides of the body, and the first gill slit was positioned on the ventral midline and did not expand up the right side as in wild-type animals. In addition, we also noticed a pronounced feature that differed between the two phenotypes: in Gli^−/− larvae the mouth was duplicated on both left and right sides, while in Hh^−/− larvae the mouth was absent (Fig. 2A). Similar results were found for GliL^−/− larvae, compared to GliL^+/+ and GliL^+/− larvae, which appeared normal (Fig. 2A). The expression of Ptch was weaker in Gli^−/− and GliL^−/− embryos as compared to wild type and heterozygous embryos (Fig. 2B).

Figure  2.  Phenotyping of amphioxus Gli mutants. A Morphological comparison of Gli^−/− and GliL^−/− larvae with wild type and heterozygous larvae, with lateral and ventral views showing symmetric pharynx morphology in knockout larvae. gs gill slit, m mouth, pp preoral pit. B Expression of Ptch in Gli^−/− and GliL^−/− embryos compared to wild type and heterozygous embryos. For the pairs of images shown in B, the numbers indicate the number of times the expression pattern was observed out of the total number of embryos analyzed for that cross. For example, for the images in the GliL analysis 51 embryos in total were analyzed, with 38 showing the pattern above and 13 the pattern below. All scale bars are 50 μm

DownLoad: Full-Size Img PowerPoint

In wild-type larvae, asymmetrical pharyngeal morphology is accompanied by asymmetric expression of genes that mark pharyngeal territories (Hu et al. 2017). We studied the expression of such marker genes in wild type and Gli knockout embryos. In wild type and heterozygote embryos Krox expression marks the right-side glandular region of the club-shaped gland (Knight et al. 2000), Nkx2.1 the endostyle (Venkatesh et al. 1999), Pou4 the mouth (Candiani et al. 2006), and Pit the preoral pit (Candiani et al. 2008). All four genes showed normal expression patterns in wild type and heterozygote embryos of both mutants (Fig. 3A–H). However, expression of all four genes was abnormal in Gli^−/− and GliL^−/− embryos (Fig. 3A'–H'). The right-side expression domains of Krox and Nkx2.1 were lost, while the left-side expression domain of Nkx2.1 was duplicated and presented on both sides (Fig. 3A', B', E', F', compare to arrows on A, B, E, F, respectively). Asymmetric expression of Pou4 and Pit became symmetric with the appearance of another site of expression on the opposite side of the pharynx (Fig. 3C', D', G', H', compare arrows on C, D, G, H, respectively). These data show that the development of asymmetric pharynx organs is compromised in both Gli knockout mutants.

Figure  3.  Expression of pharynx marker genes in amphioxus Gli knockout embryos. Each pair of panels (e.g., A, A') shows expression of the gene indicated above in the wild type and heterozygous embryos (above) compared to expression in knockout embryos (below), with genotype indicated on the far left. White lines mark the separation between the left and right sides of embryos. For A and E, the arrows mark asymmetric expression lost in the embryos in the panels below. For B and F, blue and red arrows mark a symmetrical expression domain and right sided asymmetric expression domain, respectively. In B', F' only the asymmetric expression domain (blue arrows) is visible. On C, D, G, and H the arrows mark left-sided expression domains, which are bilateral in the embryos shown in C', D', G', and H'. For the pairs of images (A, A' etc.), the numbers indicate the number of times the expression pattern was observed out of the total number of embryos analyzed for that cross. For example, for the images shown in A, A', 26 embryos in total were analyzed, with 20 showing the pattern above and 6 the pattern below. The scale bar on A is 50 μm and applies to all panels

DownLoad: Full-Size Img PowerPoint

Early development of asymmetry has been studied in amphioxus, revealing that Hh acts upstream of a conserved molecular pathway involving Nodal, its secreted inhibitors Cerberus (Cer) and Lefty, and its target the transcription factor Pitx (Hu et al. 2017; Li et al. 2017). To test if the same molecular pathway was being disrupted in Gli knockout embryos as it was in Hh knockout embryos, we studied the expression of Nodal, Cer, Lefty and Pitx in embryos of both Gli mutants (Fig. 4). Expression of all four genes was normal in wild type and heterozygous embryos derived from crosses of both Gli knockout lines, with Cer predominantly expressed on the right side and the other three genes on the left side (Fig. 4A–H). However, both Gli^−/− and GliL^−/− embryos showed abnormal expression of all four genes: asymmetric expression of Cer was lost (Fig. 4A', E'), while asymmetric expression of Nodal, Lefty and Pitx became symmetric with the appearance of ectopic expression of all three genes on the right side (Fig. 4B'–D', F'–H'). Note that embryos showing Cer expression occur at an earlier developmental stage than those for the other three genes; this reflects the epistatic relationships between these genes, with Cer acting upstream of Nodal, Lefty and Pitx as previously described (Li et al. 2017).

Figure  4.  Expression of early left–right asymmetry genes in amphioxus Gli knockouts embryos. Each pair of panels (e.g., A, A') shows expression of the gene indicated above in the wild type and heterozygous embryos (above) compared to expression in knockout embryos (below), with genotype indicated on the far left. For the pairs of images (A, A', etc.), the numbers indicate the number of times the expression pattern was observed out of the total number of embryos analyzed for that cross. For example, for the images shown in A, A', 40 embryos in total were analyzed, with 29 showing the pattern above and 11 the pattern below. The scale bar on A is 50 μm and applies to all panels. Embryos in A, A', E, E' are at the early neurula stage, other embryos at the late neurula stage

DownLoad: Full-Size Img PowerPoint

The role of Gli in neural tube development

Hh signaling also functions in vertebrate neural tube patterning, and analysis of amphioxus Hh mutant embryos supports the developmental role in patterning the amphioxus neural tube (Ren et al. 2020). Amphioxus Gli^−/− and GliL^−/− larval neural tubes were not overtly different from those of wild-type embryos in terms of gross morphology (Fig. 2A), although this is also the case in Hh knockout embryos. Hh knockout embryos do show disrupted expression of genes expressed in the ventral region of the neural tube, including Vacht, Isl, Lhx3, Mnxa, Nkx6 and OligA-C genes (Ren et al. 2020). We examined the expression of these genes in Gli and GliL knockout embryos compared to wild type and heterozygote embryos. No differences were observed in the expression of the Vacht, Isl, Lhx3, Mnxa and Nkx6 genes between wild type and mutant embryos (Supplementary Fig. S1). OligA, OligB and OligC, amphioxus specific paralogues of the Olig bHLH transcription factor family, were affected in Gli^−/− embryos (Fig. 5A'–C') compared to Gli^+/− and Gli^+/+ embryos (Fig. 5A–C). Specifically, expression of these genes was lost from a number of cells in the neural tube in a pattern very similar to that observed in Hh knockout embryos (Ren et al. 2020). However, in GliL^−/− embryos the expression of these three genes was indistinguishable between all three genotypes (Fig. 5A"–C"), indicating that GliS can cover the necessary functions of Gli in this context.

Figure  5.  Expression of Olig family genes in amphioxus Gli knockouts embryos. The expression of OligA (A–A"), OligB (B–B") and OligC (C–C") is shown in wild type, heterozygous and mutant embryos. For Gli knockouts, Gli^−/− embryos (A'–C') are shown compared to wild type and heterozygous embryos (A–C). For GliL knockouts, all three genotypes are shown combined (A"–C") as the offspring of the Gli^± x Gli^± cross used to generate them, as there were no observed differences in gene expression. For the pairs of images (A, A', etc.) the numbers indicate the number of times the expression pattern was observed out of the total number of embryos analyzed for that cross. For example, for the images sown in A, A', 19 embryos in total were analyzed, with 15 showing the pattern above and 4 the pattern below. For A", B", and C" only one pattern was observed. The scale bar on A is 50 μm and applies to all panels. Note that our experimental design means that embryos shown in A and A' were analyzed at the same time in the same tube, controlling for the between-tube variation in staining intensity sometimes encountered in in situ hybridization. The embryos represented by A" were analyzed in the same tube as each other but in a different experiment to A and A'. Thus, while A and A' are directly comparable, neither should be compared to A" for subtle differences in staining intensity. The same applies for B–B" and C–C"

DownLoad: Full-Size Img PowerPoint

Discussion

By designing TALENs to the 5' end of amphioxus Gli, and to the region specific to the GliL transcript, we were able to generate mutants that are likely to be a full loss of function of Gli, and loss of function of the GliL isoform only, respectively. Both mutants showed lower levels of Ptch expression. Ptch is a conserved downstream target of Hh signaling and thus we would expect disruption of Hh signaling to lead to a drop in its expression. Hh signaling is required for the correct development of asymmetry in amphioxus (Hu et al. 2017; Li et al. 2017). Gli^−/− embryos mostly phenocopy Hh^−/− embryos in this respect, showing similar disruption to pharynx development that is underlain by similar changes to genes that are normally asymmetrically expressed (Hu et al. 2017; Li et al. 2017). This is consistent with Gli playing the pivotal role as the transcription factor immediately downstream of the Hh signaling. There was one clear difference between Hh^−/− and Gli^−/− larvae, with the mouth lost in the former and duplicated in the later. The mouth is also lost in amphioxus Smo^−/− embryos, although initial asymmetry expression of the mouth marker Pou4 is retained (Hu et al. 2021).

GliL^−/− embryos display the same phenotype as Gli^−/− embryos. This means that full length Gli protein is required for left–right patterning of the amphioxus pharynx by Hh signaling, while the truncated form, GliS, is not sufficient on its own to cover this role. Previous work in heterologous expression systems has shown that in these contexts GliL works as an activator, while GliS functions more as a repressor (Shimeld et al. 2007). Our analysis of Gli^−/− and GliL^−/− embryos hence indicates that the activator function of amphioxus Gli is essential for correct left–right asymmetric development.

We found clear differences between the phenotypes of Hh and Gli knockouts in neural tube patterning. Hh knockout embryos show significant differences in cell type specification in the neural tube (Ren et al. 2020). However, Gli knockout embryos only showed a difference in the expression of the Olig gene family, with other markers including ones disrupted in Hh knockout embryos showing normal expression. Overall, the impact of Gli knockout on neural cells was less pervasive than would be expected from either the canonical view of Gli as the sole mediator of Hh signaling, or from the major phenotypes identified in the pharynx in both Gli and GliL mutant embryos. Furthermore, the neural tubes of GliL knockout embryos were not different from wild-type embryos (as analyzed by the genes selected in this study), implying that perhaps only repressive functions of Gli are needed at this developmental stage.

The reasons for the difference in the strength of Gli knockout phenotype between the pharynx and neural tube are unknown and require further investigation, as do the differences in mouth opening between Gli and Hh knockouts. However, we know that the regulation and function of Gli are complex, involving feedback loops and with multiple phenotypic impacts of modified Gli proteins arising through regulated protein processing or through mutation (Falkenstein and Vokes 2014; Hui and Angers 2011). Furthermore, in mice both Shh and Gli3 mutants have severe spinal cord phenotypes, while the phenotype of Shh/Gli3 compound homozygotes is less severe than either single mutant (Litingtung and Chiang 2000). This indicates that how different mutants feed into the gene regulatory network that turns Hh signaling into patterning via the Gli proteins can have an impact on the phenotypic consequences. Resolving this in amphioxus will be a challenge. However, the interbreeding of genetic knockouts to produce compound mutants has the potential to further develop our insight into this pathway.

Materials and methods

Animal and embryo cultivation

Amphioxus (Branchiostoma floridae) were acquired from Dr. Jr-Kai Yu's lab in the Institute of Cellular and Organismic Biology at Taiwan's Academia Sinica. They were cultured and induced to spawn following the protocols used for B. belcheri (Li et al. 2013); for a brief description see Yuan et al. (2020). Fertilization and subsequent culturing of the embryos were carried out at 27–28 ℃ as described by Liu et al. (2013).

Mutant generation

Two TALEN pairs (BfGli-TALEN2 and BfGliL-TALEN2) targeting amphioxus Gli coding regions were designed and assembled according to Li et al. (2014). Their positions and targeting sequences are shown in Fig. 1A, B. Mutation detection was conducted as described by Li et al. (2017). Primers used for amplification of BfGli-TALEN2 flanking regions were BfGli-TALEN2-PCR-F1: AACGGAAAGTGGCGCTGAGT and BfGli-TALEN2-PCR-R1: CTTAGTTTCAGTGTGGGTAG, and those for BfGliL-TALEN2 were BfGliL-TALEN1-PCR-F1: GCACCGAAGCTCCCGAATA (or Bf-GliL-TALEN2-F3: CAGATGCAGTATCAGGCCAA), and BfGliL-TALEN1-PCR-R1: AAAGTTCGCGCTTTGGTCCT. As described by Li et al. (2017) and Ren et al. (2020) mutant lines were maintained as heterozygotes since homozygous mutants were not viable. To generate homozygous mutant embryos for gene expression analysis, heterozygotes were crossed, yielding homozygous mutant embryos at an expected level of 25% reflecting the mendelian ratio of 3:1. Gene expression patterns were scored as wild type (reflecting the known expression pattern for a gene) or non-wild type. Non-wild type expression patterns were consistent between individual embryos. In each analysis, the proportion of embryos in each category did not significantly deviate from the expected 3:1 ratio (chi squared test, p-values range from 0.69 to 0.93).

Probes, in situ hybridization and imaging

All probes used in the study are described by Li et al. (2017) and Ren et al. (2020). Embryos and larvae at required developmental stages were fixed with 4% paraformaldehyde (PFA) in MOPS buffer (pH 7.4) at 4 ℃ for around 12 h. In situ hybridization was then performed essentially according to a previous description (Yu and Holland 2009). After staining, embryos were washed with phosphate-buffered saline, mounted in 80% glycerol and photographed using an inverted microscope (Olympus, IX71).

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s42995-023-00195-w.

Acknowledgements

We thank the two anonymous reviewers for their critical and helpful comments. This work was supported by grants from the National Natural Science Foundation of China (Nos. 32070815, 32070458, 31872186 and 32061160471) and from the Youth Innovation Fund Project of Xiamen (3502Z20206032).

Author contributions

GL, SS and YW designed the experiments. QR generated the mutants, and XH performed all other experiments. QR and XH prepared the figures. All authors interpreted the data. GL and SS wrote the manuscript.

Data availability

All data generated during this study are included in this published article.

Declarations

Conflict of interest

The authors declare no competing interests.

Animal and human rights statement

This article does not contain any studies with human participants or animals performed by any of the authors.