Characterization of gli2 2b genes in zebrafish hindbrain development 3

In contrast, both gli2 and gli2b show similar late expression in the hindbrain and its deficiency in morphants is in line with expression of these genes in this part of brain.. Both Gli

Fig 3-14 gli2b MO blocks zebrafish gli2b mRNA splicing (A) Partial gli2b

genomic sequence used for designing splicing morpholino Sequence

complementary to morpholino is underlined in red Exons (gli2b cDNA:1125-1314,

1315-1425, 1426-1560 nt) are in cyan and introns are in black Arrows indicate primers showed in B (B) Diagram shows the strategy of using splice MO (Red bar) for blocking the normal splice site and organization of predicted mRNA The blue and green arrows represent the primers for PCR used to test whether the splice MO

affects formation of gli2b mRNA The sequences for primers F1, F2, R2 are

indicated in A; R1 is indicated in Fig 3-2 (C) RT-PCR results Control RNA (lane

C) was extracted from gli2bMMS morphants; morphant RNA (lane M) was extracted from gli2bSPL Primers used are indicated in B EF1α primers were used

for internal control of mRNA quality

3.4.2 Classification of phenotypes of gli2b morphants

Both the gli2bATG and gli2bSPL morphants showed similar morphological changes Compared with wild type or control MO injected embryos, gli2b morphants

demonstrated dosage-dependent phenotypes (Table 3-1) The mild phenotype (Type I)

of gli2b morphants consisted of the enlarged hindbrain ventricle and smaller brain and eyes by 48 hpf (Fig 3-15C, D) The severe phenotype (Type II) of gli2b morphants

demonstrated also abnormality of the trunk and in ~50% cases conversed otoliths at

48 hpf (Fig 3-15E, F) Interestingly, after injection of a high dose (>2 pmol) of gli2b

MO, a defect of the convergence-extension movement could be observed By the end

of gastrulation these morphants showed a phenotype similar to that of mutants with affected BMP signaling (Fig 3-15 H, I), indicating that the early function of Gli2b might be related to dorso-ventral (D-V) patterning This early function may be due to maternal expression of this gene and not Hh related Therefore, to circumvent the

early defect and investigate the function of gli2b in neural development, we typically

injected embryos with the almost 10 times lower dose of morpholino (0.3 pmol)

Table 3-2 Phenotypes obtained after injection of gli2b morpholino

oligonucleotides in zebrafish embryos

gli2bAtg MO

injected(pmol)

Injected number

Mortality

>24hpf(Num/percentage)

Looks normal

Mild phenotype

Strong phenotype

Mortality

>24hpf(Num/percentage)

Looks normal

Mild phenotype

Strong phenotype

Fig 3-15 Classification of Gli2b morphants (A-F) Gli2b morphant with mild

phenotype (Type I) and severe phenotype (Type II) at 30 hpf and 48 hpf The mild phenotype (C, D) of Gli2b morphant is characterized by an enlarged hindbrain ventricle but relatively normal trunk The severe phenotype (E, F) of Gli2b morphant

shows not only enlarged hindbrain ventricle, but also curled trunk (H, I) gli2b

morphants in high doses (dose I, 1 pmol; dose II, 1.5 pmol) showed convergence-extension defects (arrows) in different levels compared with control (G)

3.4.3 Comparison of the phenotypes of gli2 and gli2b morphants

Both the gli2b and gli2 morphants by 48 hpf were characterized by the similarly

enlarged 4th ventricle and smaller brain and eyes (Fig.3-16B, C) Further analysis of

the Gli2b morphant under the differential interference contrast (DIC) microscope

showed the intact horizontal myoseptum in contrast to Gli2 morphants, which lacked this structure (Fig.3-16E,F; 45/52 embryos) Even when injected with the much higher

doses of MO (1 - 2 pmol), the gli2b morphants acquired the curl-down tail phenotype

with much more severe abnormality of the brain, but preserved the horizontal

myoseptum gli2b is not expressed in the adaxial slow muscle cells, thus in muscle development gli2b probably plays a minor role if any In contrast, both gli2 and gli2b

show similar late expression in the hindbrain and its deficiency in morphants is in line with expression of these genes in this part of brain

ctrl/brain gli2b MO/brain gli2 MO/brain

ctrl/trunk gli2b MO/trunk gli2 MO/trunk

D

Fig 3-16 Comparative analysis of Gli2b and Gli2 morphants Both Gli2b and

Gli2 morphants show similar abnormality in the hindbrain but the trunk is affected differently Gli2b (B) and Gli2 morphants (C) have enlarged fourth ventricle

(arrowhead) In the trunk of Gli2b morphant (E) the horizontal myoseptum (arrow) is

normal (D), while in Gli2 morphant the horizontal myoseptum is abnormal and

somites U-shaped (F)

3.4.4 Gli2b morphant showed cell proliferation defects in the hindbrain

In the 30 hpf hindbrain, the radial astrocytes (RA) stained with zrf-1 antibody

appeared as seven double segmental clusters that form curtains along rhombomere boundaries (Trevarrow et al, 1990) (Fig 3-17A) (for discussion of characteristics of cell types, see Morest and Silver, 2003) These clusters were reduced in Gli2b

morphants (Fig 3-17B; 10/10 embryos) Further increase of gli2b MO dose led to decrease of zrf-1 staining at this stage, but a gross morphology of morphants remains

relatively normal (not shown) Perhaps, elimination of RAs could be due to apoptosis

in the hindbrain triggered by the knockdown of Gli2b However, TUNEL staining of Gli2b morphants failed to detect increase of apoptosis (data not shown) Another possibility is that cell proliferation slowed down Consistent with this explanation, the number of mitotic cells detected by the anti-phospho-histone H3 (P-H3) antibody was reduced in the hindbrain both dorsally in the ventricular zone (VZ) and ventrally in association with RGCs (Fig 3-17C, D) Our analysis of the number and distribution

of proliferating cells in the consecutive sections of the hindbrain of Gli2b morphants and controls (n=3, sections/embryo=15) demonstrated moderate reduction of cell proliferation in the ventricular zone and more significant reduction in the ventral hindbrain (Fig 3-17E) Since the RAs are the neural progenitors, these results indicated that Gli2b could be necessary in the progenitor cells, including the neural progenitors in the hindbrain

Fig 3-17 gli2b and cell proliferation (A, B) Whole mount immunostaining of

radial astrocytes in 30 hpf control (A) and Gli2b morphant (B) by zrf-1 antibody followed by POD staining (C, D) Cross sections of 30 hpf control (C) and Gli2b morphant (D) at the hindbrain level stained by anti-P-H3 (purple) and zrf-1 (green) antibody (E) Chart shows the average number of proliferating cells in the ventricular zone of 4th ventricle and the ventral region associated with processes of RA cells in continuous sections of hindbrain of Gli2b morphant and wild type embryos

Ventricle RA

3.4.5 zath3 expression was affected in Gli2/Gli2b morphants

Since expression of gli2b takes place in the CNS (Ke et al, 2005) and Gli2b

knockdown caused abnormality of cell proliferation and differentiation, we decided to

check markers of early neurodifferentiation zath3 (neurod4) is involved in early neurogenesis downstream of ngn1 (Bae et al, 2003; Park et al, 2003; Wang et al, 2003) zath3 detects two main groups of cells First, the looped clusters in the

hindbrain consist of cell groups extended along the D-V axis; these are associated with curtains of RAs on both sides of rhombomeric boundaries (DV clusters) (Fig 3-18A) Second, the ventral (V) clusters in a central part of rhombomeres 2, 3, 4 and 6

(Fig 3-18C) In the single Gli2 and Gli2b morphants, the dorsal extent of zath3

expression in the DV clusters is shifted more ventrally (Fig 3-18B, E), while the V clusters diminished (Fig 3-18D, E; 10/10 embryos each) In the double Gli2/Gli2b

morphants, zath3 expression in the DV clusters is strongly reduced; however, some zath3 transcripts were still detected ventrally in rhombomeres 3 and 4 (Fig 3-18F;

10/10 embryos) Thus it seems that in both positions Gli2b acts in parallel with Gli2,

but at different D-V levels zath3-positive cells demonstrate different dependence on

Gli2 proteins

To get better understanding of this phenomenon, we checked expression of zath3

in two mutants affecting Hh pathway In yot -/-, where in result of mutation the dominant repressor form of Gli2 (Gli2DR) appeared (Karlstrom et al, 1999), the

expression of zath3 in the DV clusters was similar to that in control In contrast, the V clusters in rhombomeres 2 and 3 disappeared and the ectopic domain of zath3 expression in rhombomere 5 appeared (Fig 3-18G) In smu-/- mutant deficient in the receptor of Hh signaling, Smoothened (Chen et al, 2001; Varga et al, 2001), the

zath3-positive DV clusters were intact while all V clusters disappeared (Fig 3-18I)

Similar result was obtained after Gli2b knockdown of Gli2b on a yot -/- background (Fig 3-18H; 10/10 embryos) Hence, it seems that Hh-dependent Gli2 activating role

is required for expression of zath3 only in the ventral hindbrain probably containing

more differentiated neurons Here some other factors may also be involved in addition

to the two Gli2s In the dorsal hindbrain, the maintenance of neuroblasts could be due

to the repressive role of Gli2 (Dai et al, 1999; Ruiz i Altaba, 1999; Sasaki et al, 1999) Alternatively, Gli2s may act within a context of dorsal signaling other then Hh In both models Gli2b seems acts in parallel with Gli2

Fig 3-18 Gli2/Gli2b regulate expression of zath3 zath3 expression in

hindbrain in 30 hpf in controls (A,C,E,G,I) or Gli2b MO injected embryos (B, D,

F, H,J) White broken lines indicate the r4 (A,C) control; (B,D) Gli2b morphant;

(E) Gli2 morphant; (F) Gli2/Gli2b morphant; (G) yot -/- mutant; (H) yot mutant/Gli2b morphant; (I) smu -/- mutant; (J) smu -/- mutant/Gli2b morphant Black lines indicate the zath3-positive clusters in the ventral hindbrain Abbreviation: e,

-/-ear; r4, rhombomere4

3.4.6 Gli2/2b and regulation of expression of bmp2

The bone morphogenetic proteins (BMPs) play roles in specifying the dorsal neural tube (Dickinson et al, 1995; Liem et al, 1995; Kodjabachian et al, 1999), while ventrally their activity is antagonized by Hh signaling (Goulding et al, 1993; Roelink

et al, 1994) In the dorsal aspect of neural tube Gli2 and 3 were shown to act as repressors of Hh targets (Aza-Blanc et al, 2000; Meyer and Roelink, 2003; Ruiz i Altaba, 1999; Sasaki et al, 1999; von Mering and Basler, 1999) The double knockdown of Gli2/Gli2b caused the reduction of ventralizing Hh signaling and

affected the dorsal expression of zath3 This could be reflected in changes of transcription of dorsalizing factors We tested the relative expression level of bmp2a, bmp2b and bmp4 mRNA in 28 hpf gli2/2b morphants by the quantitative Real-Time

PCR (Fig 3-19) Total RNA extraction from 50 embryos per sample was carried out

in two different batches of experiments and Real-Time PCR was carried out in

triplicates for each sample It showed that in Gli2/Gli2b morphants all bmp genes were expressed at much higher level The increase of bmp2b expression was higher than that of bmp2a or bmp4 These results illustrated a role of the two zebrafish Gli2

in antagonizing BMP signaling at a distance from the source of Hh signaling

Therefore, bmp’s could be amongst targets repressed by Gli2s in the dorsal brain

Relative amount of mRNA in Gli2/2b morphants (MO) compared to that in

Wild type (WT) embryos

Fig 3-19 Gli2b and expression of bmps Total RNA was extracted from 50

Gli2/2b morphants or wide type embryos RT-PCR was carried out in triplicate for each sample The experiment was repeated once more

3.4.7 Inhibition of gli2b caused the disruption of notch1 and gfap expression

Notch signaling is known to maintain the neural progenitor fate and facilitate the glial differentiation (reviewed by Wang and Barres, 2000) The activity of Notch1 leads to strong activation of Her4 resulting in suppression of expression of

neurogenin1 transcription and reduction of the number of primary neurons (Takke et

al, 1999) Since the knockdown of gli2b affects the neural progenitors in the hindbrain,

we would like to examine whether this event was due to the disruption of Notch signaling

notch1a was expressed in the hindbrain forming lateral loops and the intermediate

medio-lateral clusters where expression was higher then laterally (Fig.3-20A) In

gli2b morphants, notch1a expression was reduced in the whole hindbrain (Fig.3-20B)

The lateral loops disappeared; the intermediate medio-lateral clusters were

substantially reduced (Fig 3-20B; 10/10 embryos) notch1b was expressed similar to that of notch1a but it did not formed the intermediate medio-lateral clusters, whereas its expression was enhanced at the midline (Fig 3-20C) In gli2b morphants, the lateral loops were also absent, but the midline notch1b expression remained albeit at

much more reduced level (Fig 3-20D; 10/10 embryos) Similarly expression of one

more marker linked to gliogenesis, gfap was substantially reduced at the midline

while strongly reduced in lateral loops (Fig 3-20G, H; 10/10 embryos) In summary,

the expression of two notch genes was dramatically reduced in the lateral hindbrain,

but less so medially This illustrated differential requirement in Gli2b along the M-L axis of the neural tube

The neural differentiation marker, neurogenin1 (ngn1), is expressed in position that corresponds to that of the intermediate medio-lateral clusters of notch1a Similar

to this marker expression of ngn1 was slightly affected in gli2b morphants (Fig 3-20E,

F; 10/10 embryos) Since Ngn1 acts as a proneural gene, the intermediate mediolateral clusters of cells expressing Ngn1 could represent neuronal precursors In contrast, the

midline expressing gfap and notch1b and the lateral loops, expressing notch1a and 1b,

could represent the two different populations of non-neuronal precursors The severe disruption of these and decrease in expression of zrf-1 antigen indicate much more severe decrease in early gliogenesis comparing to that of neurogenesis Thus, Gli2b in the hindbrain may act preferentially in regulating early differentiation of non-neuronal cell lineages

Fig 3-20 Expression of early neurodifferentiation markers in the hindbrain of

Gli2b morphant at 30 hpf (A, B) Expression of notch1a in Gli2b morphant (A) and

control (B) (C, D) Expression of notch1b in Gli2b morphant (C) and control (D) (E, F) Expression of ngn1 in Gli2b morphant (E) and control (F) (G, H) Expression of gfap in Gli2b morphant (G) and control (H) The relative positions of expression

domains are indicated in left side of the panel: midline, blue bar; intermediate

3.4.8 Inhibition of gli2b did not affect the segmentation of hindbrain

Since loss-of-function of Gli2b caused the disruption of segmented

notch1a/notch1b expression, it is interesting to examine whether this is due to the

disruption of the rhombomere segmentation in hindbrain In the zebrafish, several

genes, including krox20, hox genes and valentino (val), are used as markers to

ascertain a proper course of hindbrain segmentation (Moens et al, 1996; Oxtoby et al,

1993; Prince et al, 1998) krox20 and val may regulate hox gene expression in specific segment patterns in hindbrain, and hox genes establish identity of individual segments

Therefore, expression of these genes in both uninjected controls and Gli2b morphants was compared As shown in Figure 3-21A-H (10/10 embryos each), there was no difference between the expression of these markers in control embryos and that in Gli2b morphants, indicating that Gli2b is not involved in either the early onset of segmentation or the later maintenance of specific identity of rhombomeres

Fig 3-21 Expression of segmentation markers in Gli2b morphants (A,B)

Expression of val in r5 and r6 in control (A) and Gli2b morphants (C, D) Expression

of krox20 in r3 and r5 in control (C) and Gli2b morphants (D) (E, F) Expression of hoxa2 in r2-r5 in control (E) and Gli2b morphants (F) (G, H) Expression of hoxb3 in

r5-7 in control (G) and Gli2b morphants (H)

3.5 Characterization of the gli2/gli2b function within Hh pathway

The activities of Gli transcription factors are regulated by Hh signaling Gli proteins are responsible for regulating the expression of Hh-downstream genes

(reviewed by Ingham and McMahon, 2001) In Drosophila, the single Gli homolog, Cubitus interruptus (Ci) has potential activator or repressor function in Hh signaling,

depending on the posttranslational modifications in the presence or absence of Hh In vertebrates, a sum of activities of all Gli proteins is supposed to regulate all Hh-related activities in vertebrates For example, while in absence of Hh Gli2 acts as

repressor of Hh targets, in presence of Hh, Gli2 is supposed to be the Hh-dependent activator; this is supported by down-regulation of Hh target genes in mouse Gli2

mutants (Ding et al, 1999; Ruiz i Altaba, 1999)

However, a functional analysis of Gli2 functions in different vertebrate model animal provided evidence that during evolution functions of Gli2 could vary For example, knockdown of Gli2 in the zebrafish resulted in a relatively minor defect of

Hh signaling (Karlstrom et al, 2003), which could be due to existence of a second

gli2-related gene (gli2b) identified in our study (Ke et al, 2005) Therefore, it is of

interest to find specific developmental roles of the two Gli2 The aim of the following

sections is to understand how exactly gli2b and gli2 are involved in regulating events

of Hh signaling and which events they regulate For this purpose, several zebrafish

mutants affecting Hh signaling (smu and yot) in combination with gli2b

loss-of-function experiments by morpholino knockdown were used To evaluate

effects of experimental interference several known markers regulated by Hh (nkx2.2, islet1 and netrin1) were used for WISH or immunohistochemistry analysis

3.5.1 gli2b and nkx2.2 expression

Nkx2.2 is regulated by Hh signaling pathway and used as marker to define the p3

neural precursors in ventral neural tube, which are located in the most ventral region just above floor plate P3 cells give rise to V3 interneurons in mammals (Briscoe et al,

1999) In cyclops (cyc -/-) mutants, which initially lack neuroectodermal expression of

shh, the expression of nkx2.2 is absent And overexpression of shh mRNA results in the ectopic expression of nkx2.2 (Barth and Wilson, 1995) Thus, it could be informative to check whether Gli2b regulates nkx2.2 expression

In Gli2b morphant, nkx2.2 expression only slightly expanded (Fig.3-22A, B; 5/5 embryos) In Gli2 morphant, nkx2.2 expression was similar to that in Gli2b morphant (data now shown) These results indicated a minor role of both gli2 and gli2b in regulating expression of nkx2.2 in the ventral hindbrain Due to duplication of gli2 in

zebrafish, the function of Gli2b might be redundant with Gli2 Thus, loss of a single

Gli2 might not be enough to affect the Hh signaling in ventral hindbrain To address whether deficiency of both Gli2 and Gli2b would affect nkx2.2 expression, we

checked its expression after knockdown of both Gli2b and Gli2, but no obvious difference has been found compared with single Gli2 or Gli2b morphants (data not

shown) Possibly Gli2s of zebrafish does not regulate the nkx2.2-positive ventral cells,

or other Gli members function redundantly with Gli2s in regulating these cells

In yot mutant, mutation causes truncation of Gli2 and formation of the

dominant-negative (DN) Gli2 lacking C-terminal activator domain This protein acts

as a constitutively active (CA) inhibitor of Hh signaling (Karlstrom et al, 1999)

nkx2.2 expression in yot -/- was maintained only in r4 (Fig.3-22C) Thus, while

CA-DN-Gli2 blocked Hh signaling in almost all ventral hindbrain the residual activity

of some other Gli could be responsible for maintaining nkx2.2 expression in r4

It was proposed that r4 plays an important role of the signaling center within a hindbrain (Maves et al, 2002) Hence it could be of interest to study a role of Gli2b in

this rhombomere The expression of nkx2.2 was completely lost in the hindbrain after injection of Gli2b MO into yot -/- (Fig.3-22D; 5/5 embryos) This phenotype is similar

to that of smu-/- mutants deficient in Hh signaling (Fig.3-22E) These results indicate that Gli2b in parallel with some other Gli plays a role in regulation of expression of

nkx2.2 in r4 As such Gli2b could be one of the factors necessary to establish r4 as a

regional signaling center within a hindbrain In this regard, the function of Gli2b is clearly different from that of Gli2 Importantly, this function of Gli2b was not predicted based on expression pattern of this gene It seems that Glis other than Gli2s

are necessary for induction of nkx2.2 expression in other rhombomeres

mutant (E)

3.5.2 Inhibition of Gli2b caused abnormal development of oligodendrocytes

The oligodendrocytes derive from the ventral neural progenitors, which express

olig2 (Lee et al, 2005; Park et al, 2002) The Hh signaling is both necessary and

sufficient for induction of oligodendrocytes in the spinal cord of vertebrates (Pringle

et al, 1996; Poncet et al, 1996; Orentas et al, 1999; Alberta et al, 2001; Soula et al, 2001) Given a role of Gli2b within a context of Hh signaling in the ventral neural tube, it may be necessary for development of oligodendrocytes After knockdown of

Gli2b, the morphants showed a decrease or even complete loss of olig2 expression in the ventral hindbrain (Fig 3-23A, B; 9/10 embryos) In contrast, expression of olig2

in the spinal cord of Gli2b morphant only slightly decreased compared with controls (data not shown)

The embryos of ET3 transgenic GFP line (Parinov et al, 2004) express GFP in

position of olig2 expression in the ventral neural tube suggesting that oligodendrocytes in these embryos could be GFP-tagged Since mbp transcripts are expressed in oligodendrocytes (Brosamle and Halpern, 2002), we used the anti-mbp probe for double staining of the ET3 embryos against mbp mRNA and anti-GFP antibody to detect GFP protein expression The mbp probe stains the perinuclear space

(blue), while GFP was distributed evenly in the cytoplasm (brown) There are many positive cells in the ventral part of the neural tube and detailed analysis of colocalization of these two markers in this part of neural tube needs to be done on

cross sections This was not necessary in this study, since the colocalization of mbp

and GFP is more obvious in the dorsal neural tube, where only a few cells expressing these markers were detected at this stage (Fig 3-23C, D) Thus, a population of GFP-positive cells in the ET3 represents oligodendrocytes

By 4 dpf, ET3 larvae expressed GFP in the hindbrain (Fig 3-23F) and ventral spinal

cord (Fig 3-23G) The Gli2b morphants demonstrated substantial reduction of GFP expression in the hindbrain (Fig 3-23I; 19/21 embryos) and posterior trunk (Fig.3-23J; 16/21 embryos) while expression in the anterior trunk was much less affected (not shown) Why such differential sensitivity in respect of Gli2b exists along the A-P axis

is currently unclear One possibility is that at different A-P level cells require specific levels of Gli2b, second, the population of GFP positive cells cells contains oligodendrocytes and some other as yet unknown cell type, third, at different A-P levels distribution of functions between proteins of Gli family may vary or these proteins may have different origin (maternal v zygotic) For further details to be understood this analysis in future needs to be continued

Fig 3-23 gli2b is required for differentiation of oligodendrocytes olig2

expression in hindbrain in 30 hpf control (A) and gli2b morphant (B) In the ET3 line

GFP maps oligodendrocytes as shown by double staining of anti-GFP antibody and

antisense mbp probes [C, representing the white frame in (E)] Enlarged (D) from the black frame in (C) reveals the mbp mRNA (blue, arrows) in GFP expressing cells

(brown) (E) Control embryos of ET3 line and ET3 Gli2b morphants (H) 4 dpf ET3 controls, hindbrain (F) and posterior trunk (G) GFP expression in the ET3 Gli2b morphant, hindbrain (I) and posterior trunk (J)

3.5.3 Gli2b and neuronal differentiation

The early changes in organization of neural precursors prompted us to explore a

role of Gli2b in neurodevelopment Islet1 is an early marker of neurodifferentiation downstream of nkx2.2a (Higashijima et al, 2000; Inoue et al, 1994; Korzh et al, 1993,

1998) In the mammalian hindbrain, the branchiomotor neurons (BMNs) innervate the branchial arch-derived muscles that control jaw movements, facial expression, larynx and pharynx (reviewed by Chandrasekhar, 2004) In the 48hpf control hindbrain, anti-Isl1 antibody detects the trigeminal (V), facial (VII), vagal (X) BMNs and sensory Rohon-Beard (RB) cells (Fig 3-24A, B) Importantly, during late neurodevelopment Gli2b is expressed in r4 and 6 (Ke et al, 2005) Given the fact that the facial BMNs migrate into r6-7 from r4 (Chandrasekhar, 2004), it is of interest to explore behavior of these cells in Gli2b morphants A number of Isl1-positive BMNs

in Gli2b morphant is similar to that in controls, but they spread much more along the M-L axis (Fig 3-24C; 9/10 embryos) In contrast, in the yot -/- mutant, the trigeminal (V) and vagal (VII) BMNs were lost; the facial (VII) BMNs were still present, but instead of rhombomeres 6 and 7 they occupied rhombomeres 4-6 (Fig 3-24E; 10/10 embryos) This could be due to deficient migration of these cells from the

rhombomere 4 While injection of gli2b MO into yot -/- mutant resulted in even

stronger reduction of Isl-1 staining of facial BMNs, some cells of this cluster remained Isl-1-positive in demonstration that they still could reach rhombomeres 6 and 7 while migrating from rhombomere 4 (Fig 3-24G; 8/10 embryos) Therefore, migration and differentiation of facial BMN could be dependent on factors other than Gli2b

The secondary motor neurons (SMN) of the abducens (VI) nerve in the r6 are recognized by zn-5 antibody (Fashena and Westerfield, 1999; Fig 3-24I) Importantly,

expression of Gli2b is maintained in the ventral rhombomere 6 (Ke et al, 2005) The SMNs of the abducens nerve were not detected in Gli2b morphants (Fig 3-24J) Thus, this cell cluster seems is dependent on Gli2b

The loss-of-function of Gli2b showed no effect on the PMNs (not shown) However, the late-born secondary motor neurons (SMNs), which express GFP in the spinal cord of the Islet1-GFP transgenics, were affected in Gli2b morphants (Fig 3-24L,N; 36/52 embryos) The disruption of SMNs was most obvious in the mid-trunk region (Fig 3-24N) These results also demonstrated that not only the oligodendrocytes, but also the SMNs have differential sensitivity in respect of a level

of Gli2b in the ventral neural tube These cells populations react differently in respect

of changes in the Gli2b level While oligodendrocytes are more sensitive at the level

of the hindbrain and posterior neural tube, the SMNs are more sensitive in the intermediate spinal cord So, it seems that a level of Gli2b in the neural tube is a subject of tight developmental control

Another Islet-1-positive cell lineage found in the anterior dorsal spinal cord - RB

sensory neurons - is not affected in yot -/- (Fig 3-24F) In contrast, their number was reduced in Gli2b morphants (Fig 3-24D; 6/10 embryos) and even more so after

injection of Gli2b MO into yot -/- mutant (Fig 3-24H; 8/10 embryos) demonstrating,

first, a redundant role of Gli2b in this cell lineage and, second, incomplete

suppression of Gli2b function in the dorsal neural tube of yot -/- mutants

Fig 3-24 Gli2/Gli2b and differentiation of motor neurons and Rohon-Beard sensory neurons (RB) (A) Islet1-positive branchial motor neurons (BMNs) and

(B) RBs in 48 hpf controls The rhombomere numbers are in Arabic The motor nuclei of cranial nerves numbers are in Roman as follows: V, trigeminal; VI, abducens; VII, facial; X, vagal (C) In Gli2b morphants the motor nuclei (in particular, the facial one, two-headed arrow) were further away from the ventral

midline (D) In gli2b morphants the RBs were slightly reduced (E) BMNs and (F) RBs in yot -/- mutant The trigeminal and vagal motor nuclei are absent in yot -/- The migration of cells of the facial nucleus from r 4 to r 6, 7 is affected (E) No

changes in RBs was found in yot -/- (G) The residual cells of facial nucleus of yot

-/-were depleted even further after injection of Gli2b MO while some remaining cells (arrows) migrated to r 6, 7, while RBs (arrowhead) were almost gone (H) (I) cells

of the abducent nucleus (VI) were detected by zn5 antibody in 48 hpf controls These cells were lost in Gli2b morphants (J) (K-L) Position related changes of later developed motor neurons in controls (K) and Gli2b morphants (L) Enlarged mid-trunk regions were showed respectively in (M) and (N)

Abbreviations: e, ear

3.5.5 Gli2b and axonogenesis

Netrin is one of the axon guidance molecules produced by glial cells of the floor plate (Kennedy et al, 1994; Serafini et al, 1996) It was shown that in a context of axonal guidance Netrin interacts with Shh (Charron et al, 2003) We decided to

investigate whether gli2b is involved in regulation of netrin1a expression in the hindbrain In 30 hpf control embryos, netrin1a is strongly expressed in the floor plate

and boundaries of rhombomeres with peak of expression in the rhombomere 4 (Fig

3-25A, C) While the floor plate-associated netrin1a expression was expanded in Gli2b morphant, its dorsal expression was reduced (Fig 3-25B, D; 10/10 embryos)

In yot -/- mutants, netrin1a expression disappeared in the floor plate except the

rhombomere 4 (Fig 3-25E), but it remained along dorsal boundaries of rhombomeres

Thus, expression of netrin1a is regulated differently ventrally and dorsally Interestingly, the injection of gli2b MO into yot -/- mutant resulted in a severe reduction of netrin1a transcripts in the hindbrain (Fig 3-25F; 10/10 embryos) These results indicated that, first, the netrin1a patterning is regulated by the Hh signaling involving Gli2b, and, second, knockdown of Gli2b in yot -/- blocks multiple events of the Hh-dependent gliogenesis in the hindbrain

Fig 3-25 Gli2/Gli2b regulate expression of netrin1a (A) Lateral and (C)

dorsal view of netrin1a expression in the hindbrain of 30 hpf controls and Gli2b morphants (B, D) In Gli2b morphants netrin1a expression in the ventral hindbrain was up-regulated and expanded laterally (E) In yot -/- , netrin1a expression was less

affected in r4 However, its expression in the dorsal hindbrain was less affected

After injection of Gli2b MO into yot -/- , a faint netrin1a expression remained only

in the ventral r4 (black arrow, F) Abbreviations: e, ear

3.5.6 Gli2b and formation of axonal scaffold

Since gli2b is expressed throughout the hindbrain and optic tectum, it could be of

interest to examine whether loss of Gli2b function affects later events of neurodevelopment, such as, for example, axonogenesis The mouse monoclonal antibody against acetylated tubulin labels processes of all neurons and has been used

to examine the earliest neuronal tracts in the brain of the zebrafish (Chitnis et al, 1990) At 30 hpf, the axonal scaffold in Gli2b morphants is already affected The axonal tracts seem to be more defasciculated The axonal tract deriving from the trigeminal sensory cluster is more heavily stained, while the organized pattern of axons in commissures along the M-L axis, which are cleanly separated in controls, is perturbed in morphants (Fig 3-26A, B; 9/10 embryos) By 48 hpf other abnormalities also became obvious The Gli2b morphants contain less of acetylated-tubulin-positive fibers in the dorsal hindbrain and midbrain (Fig 3-26D; 8/10 embryos) In parallel, there is further (comparing to 30 hpf) increase in staining in the ventral neural tube Also, some of the axons (4/10 embryos) change their orientation from the D-V axis to

A-P axis (Fig 3-26F, enlarged frame of D) Importantly, gli2b is expressed in the

dorsal hindbrain and perhaps, these expression domains are required for extension of axons from the ventral neural tube This expression may play a role in providing positional cues in the substrate through which the growth cone/process navigate Gli2b is expressed in the optic tectum and dorsal hindbrain and later on in the dorsal clusters (Fig 3-8, 3-9) In the midbrain of controls, the acetylated tubulin-positive axons appear in the optic tectum and cerebellum (Fig 3-26G) But in Gli2b morphants, no axonal tracts could be found in the optic tectum (arrow) and cerebellum (dashed line) (Fig 3-26H; 9/10 embryos) Here, a mechanism similar to that in the more posterior hindbrain could be involved also

Fig 3-26 Axonal scaffold and formation of neuronal clusters in the 48 hpf Gli2b morphant The anti-acetylated tubulin antibodies detect the axonal scaffold in

(A) 30 hpf wild type control and (B) Gli2b morphant Only the right side of the hindbrain is shown The axonal scaffold of the 48 hpf wild type control (C) and morphant hindbrain (D) (E, F) Blow-up of regions in red frames in (C, D) shows a change of axonal growth in the morphant from the D-V to the A-P direction (G), control midbrain contains staining associated with axons (H) the midbrain of Gli2b morphant contains no axon-associated staining in optic tectum (arrow) and dorsal midbrain-hindbrain boundary (dashed line) present in wild type controls (I) formation of the motor neuron cluster corresponding to the facial (VII) nerve was

Chapter IV

DISCUSSION

IV Discussion

The developmental comparative studies demonstrated divergent requirements for Gli2 in zebrafish and mouse In particular, in presence of Hh in mice, Gli2 acts as the Hh-dependent activator and in absence of Hh signaling, Gli2 may perform a role of repressor of Hh targets in the neural tube (Ding et al, 1998; Matise et al, 1998; Ruiz i Altaba, 1999; Sasaki et al, 1999), while the zebrafish Gli2 was postulated to play minor roles in CNS development (Karlstrom et al, 2003) This led to attempts to explain apparent difference in Gli2 function in mice and zebrafish by redistribution of roles within the zebrafish Gli family in favor of other proteins- Gli1 and Gli3 (Chandrasekhar et al, 1999; Chen et al, 2001; Tyurina et al, 2005; Vanderlaan et al, 2005; Varga et al, 2001)

Due to genome duplication in teleost lineage a function of some of the zebrafish genes is divided between two genes (Postlethwait et al, 1999; Chong et al, 2001) Up to date, four Gli family members have been identified compared with only

three Gli family members in mammals The additional gli which is likely a paralog of gli2, or gli2b, is a most recent addition to zebrafish zebrafish gli family Among three

zebrafish Gli proteins, which function has been studied, Gli1 plays a sole role as transcription activator, which is similar to Gli1 in other vertebrate species Gli3 has bi-potential roles similar with mammalian Gli3, with the activator function in the ventral neural tube and repressor function in the dorsal neural tube (Tyurina et al, 2005) However, although structurally, Gli2 also has both activator and repressor domains, the function of Gli2 is far from being fully understood in zebrafish and

mammals, in particular in the CNS Therefore, in view of recent identification of a

second gli2 gene in zebrafish, the present study was initiated to study the function of

Gli2 in zebrafish neural development in more detail In result we will present a different model of redistribution of functions of mammalian Gli2 between Gli2 and Gli2b

Within a frame work of this thesis, the full length gli2b cDNA was isolated and cloned The 4730 bp gene of gli2b was analyzed for its unique features and

mapped to a conserved region on zebrafish linkage group 11 Its expression and function in early zebrafish development were analyzed by various experimental tools Its expression in the hindbrain was further analyzed with several mutants affecting

different signaling pathways Our data suggested that induction of gli2b positive

neural precursors in the hindbrain is dependent on Shh and Notch signaling These results have been published (Ke et al, 2005) Further, our unpublished result (Ke et al,

in preparation) demonstrated that loss of Gli2b resulted in disruption of different neural precursors as well as neuron migration and axon projection Besides, Gli2b is involved in position-related patterning of motor neuron and oligodendrocytes

The study also elucidated the role of gli2/gli2b in development of the zebrafish

brain, especially the hindbrain Our results indicated that Gli2b plays a more important role in regulating Hh signaling in zebrafish hindbrain than Gli2 Also, our results indicated that Gli2 dominant repressor was not sufficient to interfere with all

possible functions of Gli2b in yot -/-; in particular, Gli2b might have an essential activator role in the rhombomere4 These observations also indicated the involvement

of Gli2b as an activator in Hh signaling during ventral neural patterning and as a repressor in dorsal neural patterning in absence of Hh signaling Thus a combined activity of two zebrafish Gli2s performs functions similar to those of mammalian Gli2

4.1 Zebrafish Gli2b belongs to Ci/Gli zinc finger transcription factor family

The Cubitus interruptus (Ci)/Gli family of zinc-finger transcription factors is

widely conserved both in structure and in general function between invertebrates and vertebrates and is involved in the numerous Hh-dependent developmental processes

In Drosophila, this family is represented by one protein, Cubitus interruptus; in

vertebrates, the family is represented by at least 3 members, Gli1, 2 and 3, which are thought to have arisen as a result of two successive rounds of gene duplication, much

in the same way as the developmentally important Hox gene clusters (Ruddle et al, 1994; Garcia-Fernandez et al, 1994)

In zebrafish, there are at least four gli genes (gli1, gli2, gli2b and gli3) The additional member of gli family, gli2b might be produced from the additional recent

round of genome duplication in teolests followed by reduction in a number of genes resulting in maintenance of about 20-30% of teleost genes as duplicated in comparison to mammals (Amores et al, 1998; Ke et al, 2005; Venkatesh, 2003) Gli2b

is highly conserved with zebrafish Gli2 in the zinc-finger domain (96.9% identity), but shares only 62.5% identity for overall amino acid sequence The deduced PKA binding domains are also conserved between Gli2b and Gli2 In addition, both N-terminal repressor domain and C-terminal activator domain containing VP16-like

TAF binding domain and CBP binding site are found in Gli2b, indicating a possibility

of bi-potential regulation of downstream gene expression depending on the presence

of Hh signaling The presence of slightly more divergent CBP binding domain compared with Gli2 suggests some difference in activating abilities of Gli2b, but this hypothesis so far is not supported by experimental evidence

4.2 Genome mapping of gli2b reveals a conserved synteny

Based on genetic mapping, the region containing gli2b might be syntenic to the chromosomal regions containing Gli2 in mouse chromosome 1 and zebrafish linkage group 9 Using systeny analysis, we found that zebrafish myog, ndpk6, gli2b are located close to each other similar to their homolog genes Myog, Nme7 and Gli2

in mouse chromosome 1 By comparing the region including gli2 region in zebrafish

linkage group 9 and mouse chromosome 1, we found another set of closed located

genes (en1, ihha, dermo1 and gli2) Interestingly, this set of genes is located on the other side of gli2 in mouse chromosome 1 However, no other genes show homology when comparing linkage groups where zebrafish gli2b and gli2 are located One

possible explanation could be the incomplete information on distribution of gene loci

Another explanation is that some chromosomal rearrangement happened just near gli2/2b gene in zebrafish during the evolutionary process of elimination of most duplicated genes, and in result the region where gli2b is located was much more affected This is less likely scenario, since the two eng1 genes are still found in the

zebrafish genome Thus more likely the fragment of chromosome 11 containing eng1

and gene homologous to ihhb (Ehh) has been translocated elsewhere (chr.1 and chr.6)

4.3 Comparative analysis of gli2/2b expression and function

Many mammalian genes are represented in teleost genome by duplicated copies (Postlethwait et al, 1998; Amores et al, 1998; Wittbrodt et al, 1998; Meyer and Schartl 1999; Volff and Schartl 2003) The increased gene copy number could be the consequence of a whole-genome duplication (Amores et al, 1998) or the result of frequently occurring local duplication events (Robinson-Rechavi et al, 2001), which is less likely No matter what is a mechanism by which they arise, the functional gene duplicates can be maintained principally by neofunctionalization of a duplicate, characterized by protein activities acquired de novo and/or by subfunctionalization (Force et al, 1999) The cause of subfunctionalization could be the differential decay

of specific regulatory or coding sequences in each gene copy and the subsequent need for the presence of both duplicates with complementary activities to achieve the functions of the ancestor gene (Force et al, 1999)

Our analysis of expression pattern of gli2/gli2b demonstrates that the

maintenance of these two genes has been mainly due to subfunctionalization The protein structure of Gli2/Gli2b is relatively well conserved, indicating that they may play similar developmental role Whether this is a case is difficult to conclude in absence of data illustrating interaction of these proteins with their targets After all some evolutionary modification took place in functional domains of these two proteins And yet in absence of this information, I prefer to leave this matter and discuss other possibilities

The maintenance of both gli2 duplicates in zebrafish might be largely due to the

change of regulatory sequences The diagram in Fig.4-1 illustrates possible

evolutionary modification of developmental regulation of gli2 and gli2b expression based on the expression patterns in Fig.3-8 gli2 retains the regulatory elements

controlling the expression at the telencephalon-diencephalon boundary, midbrain, neurohypophysis and somites, but it loses some efficiency of regulatory element

controlling the expression in hindbrain; gli2b retains the regulatory elements

controlling the expression in telencephalon boundary, posterior hypothalamus and hindbrain, but it loses some efficiency of regulatory element controlling the expression in midbrain and totally loses the regulatory element on the expression in diencephalons boundary and somites

The differential expression of gli2/gli2b might reflect their function in different regions Consistent with their expression, knockdown of gli2/gli2b results in the

different phenotypes While knockdown of Gli2 causes the loss of horizontal myoseptum, knockdown of Gli2b does not affect horizontal myoseptum These results supports the hypothesis that in teleosts subfunctionalization indeed is consistent with the maintenance of most duplicated genes (Chiang et al, 2001b,c;Lister et al, 2001; Serluca et al, 2001; Altschmied et al, 2002)

Fig 4-1 Diagram of the regulatory regions of gli2 and gli2b genes During

evolution, both of them might maintain some regulatory elements T, telencephalon boundary; D, diencephalons boundary; M, midbrain; H, hindbrain; NHp, neurohypophysis; PHp, posterior hypothalamus; S, somites Dark grey frames indicate the change of the regulatory efficiency on gene expression

4.4 Mutant analysis demonstrated that gli2b is regulated by an integration of

different signaling pathways

Although the role of Gli proteins in Hh signaling has been extensively studied,

the regulation of gli expression is not yet fully understood It is clear that gli1 is a direct target of Hh signaling, positively mediated by Gli2 In contrast, gli3 could be

negatively regulated by Hh signaling However, how Gli2 is regulated remains largely unknown

Nevertheless, emerging evidence from experiments in Xenopus indicated that gli2 expression could be regulated by different signaling pathways In Xenopus, Gli2 can

be induced by Shh and mediate some of Hh related events (Ruiz i Altaba et al, 1998) Gli2 is also sufficient to induce ventroposterior development, functioning in the FGF-Brachyury regulatory loop (Brewster et al, 2000) Gli2 directly induces

brachyury, a gene required and sufficient for mesodermal development, and gli2 is in

turn induced by FGF signaling These results suggest a molecular basis for an integration of FGF and Hh inputs in Gli-expressing cells that respond to these signals (Brewster et al, 2000) Besides Hh and Fgf, other signaling pathways may also

regulate gli expression Wnt signaling has been recently suggested to affect Gli2 and Gli3 expression in chick somites (Borycki et al, 2000) As expression of gli2b is not present in the posterior mesoderm, perhaps, it is gli2 but not gli2b, which regulatory

regions could be responsive to the posteriorizing effects of FGF-Brachyury and Wnt, mentioned above

Our results also support the hypothesis that gli2b could be regulated by different