Mutations affecting morphogenesis during gastrulation and tail formation in the zebrafish, Danio rerio

Rapid progress has been made in the understanding of the cellular behavior underlying the morphogenetic movements during vertebrate gastrulation and tail formation. In gastrulating zebrafish embryos, three morphogenetic movements can be distinguished: epiboly, involution and dorsal convergent extension, which are driven by radial and medial cell intercalation and active cell migration (Warga and Kimmel, 1990).

However, very little is known about the genes controlling these movements. Treatment of pregastrula embryos with α-amanitin blocks all gastrulation movement, indicating that they depend on zygotic gene expression (Kane et al., 1996). In zebrafish, two mutants with defects in gastrulation movements have been described: spadetail mutant embryos lack the trunk somites and have a greatly enlarged tail (Kimmel et al., 1989). These defects are caused by a reduction of dorsal convergence movements (Ho and Kane, 1990). cyclops mutant embryos (Hatta et al., 1991) exhibit a fusion of the eyes and a lack of ventral specification within the neuroectoderm. The anterior defects might be caused by defects in cell migration of the anterior dorsal mesendodermal cells (Thisse et al., 1994).

Among the cloned genes that are expressed during gastrulation, only a few have been shown to be involved in gastrulation movements. Studies in Xenopus animal cap explants have shown that maternal factors that induce mesoderm simultaneously initiate convergent extension movements, suggesting that the early processes of pattern formation and gastrulation may be coupled and share elements of the same genetic control system (Smith and Howard, 1992). This control system might also include genes that are zygotically expressed during gastrulation: goosecoid, a homeobox gene expressed in the anterior dorsal mesoderm, seems to play a role in the migration of anterior dorsal mesodermal cells to their final destination in the head region (Niehrs et al., 1993).

According to recent studies in Xenopus, tail formation is a continuation of gastrulation and shares components of the same control system that is used in the trunk (Gont et al., 1993; Tucker and Slack, 1995). Cell-labeling studies indicate that cells in the tailbud, the organizer of the tail, derive from cells of the blastopore lip, the organizer of the trunk. Moreover, the cellular movements in the tail resemble those in the blastopore. The tight linkage of gastrulation and tail formation is also revealed by the phenotypes of ventralized and dorsalized zebrafish mutants: in all mutants isolated, the dorsoventral patterning of both trunk and tail are affected (Hammerschmidt et al., 1996; Mullins et al., 1996). However, additional morphogenetic processes are initiated after gastrulation that appear to be uniquely required for the posterior extension of the body axis and the outgrowth of the tail. In the zebrafish, these are the ventral movement of the tailbud on the yolk sac, the constriction of the posterior part of the yolk sac to form the yolk tube, and the detachment of the tip of the tail from the yolk tube. Therefore, there could be genes that are uniquely required for specific aspects of zebrafish tail development.

Here, we describe the isolation of zebrafish mutants that show specific defects during gastrulation and tail formation. one-eyed pinhead mutants display defects in the specification of the prechordal plate, which lead to an impaired rostral outgrowth of the body axis, while in spadetail and trilobite mutants, convergent and extension movements are affected. pipetail and kugelig mutants display specific defects during tail formation.

Mutations were induced and zebrafish maintained, mated and raised as previously described (Mullins et al., 1994). Embryos were kept at 28°C and staged according to Kimmel et al. (1995). For photography, embryos were dechorionated and mounted in 5% methylcellulose in E3 medium (Westerfield, 1994). Older embryos were anesthetized with 1/10 volume of 0.2% MESAB (Sigma).

Immunostainings were carried out as described elsewhere (Schulte-Merker et al., 1992) with slight modifications: the blocking, but not the incubation solution, was supplemented with 100 μM biotin; the incubation solution of the secondary antibody was supplemented with 1% normal goat serum (Vector); and the secondary antibody was preincubated for 2 hours with zebrafish embryos.

In situ hybridization was carried out as described elsewhere (Schulte-Merker et al., 1992) with the following modifications: hybridization and all washes were carried out at 65°C. Washes were as follows: 1× 20 minutes in 5× SSCT, 50% formamide; 2× 20 minutes in 2× SSCT, 50% formamide; 1× 10 minutes in 2× SSCT, 25% formamide, 1× 10 minutes in 2× SSCT, 2× 30 minutes in 0.2× SSCT, 1× 10 minutes in PBST. Stained embryos were fixed for 2-6 hours at room temperature in 4% paraformaldehyde/PBS, transferred to methanol and stored at 4°C in benzylbenzoate:benzylalcohol (2:1, v:v).

In our screen for mutants affected in morphogenesis and pattern formation during zebrafish embryogenesis, a number of mutants were isolated that show abnormalities in gastrulation and tail formation. In complementation tests, at least 18 genes were identified. At least seven of them are required for the establishment of the basic body plan, and mutant embryos display partial dorsalization (five genes, Mullins et al., 1996) or ventralization (two genes, Hammerschmidt et al., 1996). A number of genes with unique phenotypes affect specific aspects of early processes, such as epiboly (Kane et al., 1996). Here, we describe genes required for other morphogenetic processes during gastrulation and tail formation (one eyed pinhead, dirty nose, trilobite, spadetail, pipetail, kugelig and others; see Table 1). Additional mutants affecting gastrulation are silberblick (Heisenberg et al., 1996) and cyclops (Brand et al., 1996).

One allele of one-eyed pinhead, oep^tz257, was isolated, and two more alleles were found in other laboratories (Schier et al., 1996; U. Strähle, S. Jesuthasan, P. Blader, P. Garcia-Villalba, K. Hatta and P. W. Ingham, personal communication). Morphologically, the phenotype of oep mutants is recognizable at the tailbud stage (Fig. 1A): the body axis is shorter, the head region is thicker and no prechordal plate can be found. Despite the lack of the anterior structures, no increase in the rate of cell death could be detected in the anterior region at gastrula and somite stages by acridine orange stainings (not shown). The notochord, the somites, and other structures in the posterior trunk develop normally during early somite stages (Fig. 1B), although the tailbud is more prominent in oep mutants (Fig. 1B). A dorsal view on the head region reveals that the anterior part of the forebrain is reduced, while the eye vesicles are fused (Fig. 1C); brain structures posterior to the forebrain are normally formed (Fig. 1D). In the trunk (Figs 1E, 2K), the notochord is slightly undulated, and the yolk tube is smaller and thicker, suggesting that the extension of the body axis is impaired. The body axis is curled down, and no floorplate cells can be morphologically detected (Fig. 1F). The hatching gland, a derivative of the prechordal plate, is not formed.

With molecular markers, alterations are first visible in the presumptive dorsal mesoderm, where the expression of goosecoid (gsc; Stachel et al., 1993; Schulte-Merker et al., 1994; Thisse et al., 1994) is progressively reduced, beginning before gastrulation, as shown here at 40% epiboly (Fig. 2A). At 70% epiboly, no gsc transcripts have been detected in most oep mutant embryos, while a few show some residual gsc staining. In those embryos, the gsc expression domain lacks the arrowhead-like anterior protrusion found in wild-type siblings (not shown), suggesting that in the mutants, the cells with the reduced gsc expression fail to move normally towards and beyond the animal pole. The lack of the prechordal plate is also revealed by the absence of its derivatives, the hatching gland precursors, which are marked by the expression of the forkhead domain gene fkd2 (R.M.W. and J.O., unpublished). At the 10-somite stage, no or only very few fkd2-positive cells (up to 10 versus about 200 in wild type) are present anterior of the head region (Fig. 2D).

In contrast to the prechordal plate, the axial mesoderm of the trunk, the notochord anlage, appears normal during gastrulation, as indicated by the expression of no tail (ntl; Schulte-Merker et al., 1992; Fig. 2C). At the end of gastrulation, the notochord is slightly shorter, but not broader, and the density of ntl-positive cells is increased. No significant difference has been found in the number of ntl-positive nuclei in the axis of oep and wild-type embryos.

In the neuroectoderm, the earliest differences between oep mutants and wild-type siblings have been found at midgastrula stages, about 3 hours after the alterations of the gsc expression in the mesendoderm had become apparent. As an early neural marker, we used the forkhead domain gene fkd3 (J. O., unpublished), which is expressed in presumptive neuroectodermal cells beginning shortly before the onset of gastrulation. While the expression of fkd3 appears normal at shield stage (not shown) and in the posterior region of the presumptive neuroectoderm at 70% epiboly, the domain anterior of the presumptive diencephalic-mesencephalic boundary is completely absent (Fig. 2E), revealing anterior deficiencies in the neuroectoderm of oep mutants. The lack of anterior forebrain structures is also reflected by the expression of other genes at later stages of development: the expression of pax6 (formerly called pax[a]; Krauss et al., 1991) at the 3-somite stage in the presumptive diencephalon reaches the anterior boundary of the head rather than fading out as in wild-type embryos (Fig. 2F), suggesting that brain regions rostral to the diencephalic pax6 expression domain are missing in oep mutants. In addition, the expression of dlx3 (Akimenko et al., 1994) in the region of the olfactory placodes, which normally overly the telencephalon, is absent in oep mutants while it is normally expressed in the more posteriorly derived otic placodes (Fig. 2H).

Other markers reveal defects in the ventral specification of the neuroectoderm throughout the whole length of the axis: the expression of sonic hedgehog (shh, Krauss et al., 1993) is missing in ventral cells of the brain and the floorplate of the trunk in oep mutants (Fig. 2J,K). An additional neuroectodermal defect of oep mutants, the fusion of the eyes, is anticipated by the expression of pax2 (formerly called pax[b]; Krauss et al., 1991), which is normally expressed in the optic stalk in proximal regions of the eye primordia. In oep mutants, this expression is missing, consistent with the subsequent lack of a separation of the eye field (Fig. 2G).

Defects in derivatives of the prechordal plate have also been found in dirty nose (dns) mutants. Similar to oep mutants, dns mutant embryos lack the hatching gland and fail to hatch. However, dns embryos show no obvious defects during gastrulation. At day 1 of development, brownish looking apoptotic cells, possibly the hatching gland precursors, are found in the anterior region of the head underneath the eyes (Fig. 1G,H). When manually dechorionated, the mutants survive to adulthood.

With molecular markers, alterations in cell specification can already be detected at early gastrula stages, when the expression of gsc in dorsal mesodermal cells is slightly reduced in dns mutants (Fig. 2B). At the 10-somite stage, before cell death can be detected in this population of cells by acridine orange staining, the staining for Fkd2 protein (R. M. W. and J. O., unpublished) is much weaker in the hatching gland precursors of the prechordal plate, and the stained nuclei appear much smaller than in wild-type siblings (Fig. 2D). The fkd2-expressing cells of dns mutants stay at the anterior border of the head region rather than spreading onto the yolk sac to form the hatching glands.

While the main defects in oep and possibly dns mutants are observed in the anterior region of the developing body axis, spadetail (spt) and trilobite (tri) mutants display alterations over the entire length of the axis. Two new alleles of the pre-viously identified gene spt (Kimmel et al., 1989) have been isolated, spt^tm41 and spt^tq5, both of which are of similar strength to the previously known allele spt^b104. In spt mutant embryos, no somites are formed in the trunk, while the notochord appears rather normal (Fig. 3A, see below). The tailbud is greatly enlarged (Fig. 3B).

Two alleles of tri have been isolated, tri^tk50 and tri^tc240, with tri^tk50 causing a clearly stronger phenotype than tri^tc240 (see also Solnica-Krezel et al., 1996). In contrast to spt mutants, tri mutant embryos have trunk somites. Compared to wild-type siblings, the somites are more compressed in the anteroposterior axis and broader in the mediolateral axis, as already apparent at the 2-somite stage. The notochord is also shorter and slightly broader (Fig. 3A). Although the length of the body axis is reduced, the broader notochord shows little indication of undulation compared to other short mutants like pipetail (see below).

The morphological alterations in the axis of spt and tri mutant embryos suggest that convergent extension movements at the ventrolateral and the dorsal side, respectively, are affected. This is supported by the altered expression pattern of marker genes. At the 80% epiboly stage, the lack of dorsal convergent movements of the ventrolaterally derived mesoderm in spt mutant embryos is revealed by the expression pattern of sna1 (Hammerschmidt and Nüsslein-Volhard, 1993; Thisse et al., 1993), which stays restricted to ventrolateral marginal regions rather than forming two paraxial stripes (Fig. 4A). At the 4-somite stage, the expression of myoD (Weinberg et al., 1996) in two adaxial lines adjacent to the notochord is strongly reduced (Fig. 4C), and the expression in more lateral regions, corresponding to the positions where the somites are normally formed in the wild-type siblings, is completely missing. These paraxial deficiencies have been shown to be caused by a failure of muscle precursor cells to converge to the dorsal axis (Ho and Kane, 1990). Instead, these cells remain in ventrolateral positions and eventually end up in the tailbud, which is much thicker and contains many more cells in the posterior region of spt mutants at early somite stages, as revealed by the expression of eve1 (Fig. 4D; Joly et al., 1993).

The axial mesoderm is only slightly affected in spt mutants. The ntl expression domain is slightly broader than in wild-type siblings at early somite stages (Fig. 4B, see also Fig. 4C for the increased distance between the myoD-positive adaxial stripes). These observations suggest that the defects in gastrulating spt mutant embryos are largely restricted to the dorsal convergence movements of ventrolateral cells towards the dorsal axis.

In contrast, embryos homozygous for the strong tri^tk50 allele exhibit only a subtle difference in the expression pattern of sna1 in ventrolateral positions at 70% epiboly (Fig. 4E). In the dorsal midline, however, alterations can be detected at the end of gastrulation. As revealed by the expression pattern of ntl, the presumptive notochord is considerably broader and shorter, although no significant difference in the number of ntl-positive cells has been found (Fig. 4F). As revealed by the expression of myoD, the paraxial mesoderm of tri^tk50/tri^tk50 embryos has about 60% of its normal length (Fig. 4G). The eve1 expression domain in the tailbud of tri mutant embryos, although about twice as broad as in wild type, appears to contain the normal number of cells (Fig. 4H). These observations suggest that in tri mutants, in contrast to spt mutants, ventrolateral cells converge normally towards the axis, but convergent extension movements in the axis itself are reduced.

Seven short mutants defining two complementation groups were isolated which show an impaired outgrowth of the tail. Five alleles of pipetail (ppt) were isolated, ppt^ta98, ppt^te1, ppt^tc271, ppt^ti265 and ppt^th278, which all are of similar strength. ppt mutant embryos can be morphologically identified at the 2-somite stage, when their body axis is shorter, spanning an angle of about 270° instead of about 300° as in wild-type siblings (see Fig. 5A for 8-somite stage). Normally, the tailbud moves ventrally after the end of gastrulation. This movement seems to be blocked in ppt mutants, leaving the tailbud closer to the point of germ ring closure. The whole body axis becomes compressed: the notochord is undulated (Fig. 5B) and the somites are thinner along the anteroposterior axis (Fig. 5A,B). The formation of the yolk tube starts normally, although the yolk tube is about twice as thick and half as long as in wildtype embryos at the 24-somite stage. The region posterior to the yolk tube is several times shorter than in wild-type siblings and remains attached to the yolk tube much longer than in wildtype siblings (Fig. 5C). At day 5 of development, the ppt mutant larvae have only about 80% of the normal length. Although the notochord and the neurocoel are bent at the tip of the tail (Fig. 5E), the gap of the ventral melanophore stripe is present, indicating that despite the compression posterior structures are properly patterned.

In addition to the defects in the tail, ppt mutants display an impaired growth of the head skeleton. In wild-type larvae, cartilaginous elements of the neurocranium and the pharyngeal skeleton extend anteriorly beyond the eyes at days 4 and 5 of development. In ppt mutants, however, these elements show a reduction in length and abnormal cartilage differentiation, features shared by the ‘hammerhead’ class of mutants (Fig. 5D, Piotrowski et al., 1996).

Two alleles of kugelig (kgg) were isolated. The kgg phenotype becomes obvious at the 14-somite stage. In contrast to ppt mutants, the ventral migration of the tailbud on the yolk sac occurs normally in kgg mutants. However, the constriction of the posterior part of the yolk sac and the subsequent formation of the yolk tube are blocked. The yolk tube of kgg mutants has at most 10% of the normal length, leading to a rather spherical (‘kugelig’ in German) body shape (Fig. 5C). The region posterior to the yolk tube is only slightly shorter than in wild-type embryos. During subsequent development, the posterior part of the tail often curls up and folds back in a rather sharp angle. In contrast to ppt mutants, notochord and neurocoel end normally in the tail tip of kgg mutant embryos. In the trunk, however, the neurocoel seems slightly enlarged and disrupted at variable positions. Although not fully understood yet, these neurocoel defects might be secondary consequences of the curled shape of the trunk.

In situ analysis of ppt mutants was carried out with markers specific for different regions of the tailbud and the tail. In early somite stages, eve1 (Fig. 5L), ntl (Fig. 5H) and sna1 (not shown), all of which are expressed in presumptive mesodermal cells of the tailbud, show a normal expression pattern, suggesting that the formation of the posterior mesoderm is normal. The same applies to genes expressed in posterior neuroectoderm like pou2 (Takeda et al., 1994) and cad1 (Joly et al., 1992) (not shown). The expression of wnt[a], a zebrafish wnt gene with highest structural similarity to the mouse genes wnt3 and wnt3a (Krauss et al., 1992), although of normal intensity, reflects the impaired ventral growth of the tailbud described above. While in a wildtype embryo of the 5-somite stage the expression domain acquires the shape of a full semicircle, the expression domain in the mutant is still round with a small circular gap in the middle (Fig. 5G), showing that the tailbud has maintained the same conformation as at the end of gastrulation. Only one of all tested marker genes shows a significant reduction in the level of expression from early somite stages on: wnt[b], a zebrafish wnt gene with highest structural similarity to the mouse genes wnt5a and wnt5b (Krauss et al., 1992). Its expression is strongly reduced in the whole posterior region, including the posteriormost somites (Fig. 5F). At the 28-somite stage, the impaired outgrowth of the region posterior to the yolk tube is reflected by the eve1 expression. In addition to the presumptive mesodermal cells at the tip of the tail, eve1 is expressed in a line of distal cells that spans, ventrally and dorsally, from the anus to the tail tip (Fig. 5L). In the mutant, this distal expression domain does not extend. Rather, it is broader than in wild-type siblings and remains at its original position close to the posterior end of the yolk tube, while inner regions of the tail attempt to extend normally (Fig. 5L, see also Fig. 5C).

The trunk phenotype is reflected by the expression pattern of ntl and myoD. The ntl expression domain in the notochord is undulated (Fig. 5H), as are the myoD-positive adaxial lines (Fig. 5J). The somitic myoD expression pattern is more compressed in the anteroposterior axis, and the expression in the posterior somites is broader, extending to more lateral positions (Fig. 5J).

In contrast to ppt mutants, posterior somites in kgg mutants are compressed without showing a lateral expansion (Fig. 5K). Since kgg mutants show an impaired formation of the yolk tube, we especially investigated the expression pattern of genes expressed in the region of the presumptive anus and just anterior to it, such as dlx3 (Akimenko et al., 1994) and cad1 (Joly et al., 1992). However, none of these genes shows an altered expression before the phenotype becomes morphologically visible (not shown).

We have identified several genes required for the various morphogenetic movements that occur during zebrafish gastrulation and tail formation. Mutants in one-eyed-pinhead (oep) and dirty nose (dns) display defects in prechordal plate formation. Mutants in spadetail (spt) and trilobite (tri) display alterations in different aspects of dorsal convergent extension movements, while in pipetail (ppt) and kugelig (kgg) mutants morpho-genetic processes specific to tail formation are affected. Mutations in additional genes required for a normal extension of the body axis are described in Table 1. No mutants lacking involution were isolated, while four mutants affecting epiboly movements have been found (Kane et al., 1996).

one-eyed-pinhead (oep) mutants show four main defects: a lack of the anteriormost structures in both the mesoderm and the neuroectoderm, a fusion of the eyes, and a lack of ventral specification in the central nervous system over the entire length of the body axis. With molecular markers, alterations first become visible in the presumptive prechordal plate of the mesoderm, starting before the onset of gastrulation. The defects in the neuroectoderm appear later during gastrulation and somitogenesis. The temporal development of the different phenotypic traits suggests that the alterations in the neuroectoderm might be a consequence of the defects in the prechordal plate. This notion is supported by findings that in normal vertebrate development, the separation of the eye field depends on vertical signals emanating from the underlying axial head mesendoderm (Brun, 1981). Interestingly, the ventralization of the neural tube is blocked over the entire length of the body axis, although in the posterior regions of the axis, no structural defects have been found in the underlying mesoderm, the notochord. This observation might suggest that the ventral specification of the neuroectoderm takes place at early stages of gastrulation (60% epiboly), when the goosecoid-positive presumptive prechordal plate cells underlie the not-yet-extended neuroectoderm over its entire anteroposterior length. In wild-type zebrafish embryos, the prechordal cells are known to be very motile (R. M. W. and D. A. K., unpublished results) and are supposed to display active cell migration, as it has been demonstrated for prechordal cells in Xenopus (Winkelbaur, 1990; Winkelbaur and Selchow, 1992). The altered shape of the residual gsc expression domain in some oep mutants suggests that oep is required for a proper rostral movement of cells at the anterior edge of the hypoblast. The coincidence of the lack of gsc expression and the lack of these anterior movements in oep mutants is consistent with the finding that ectopic gsc can initiate cell migration movements in Xenopus embryos (Niehrs et al., 1993).

These anterior movements appear to be necessary for the normal extension of the body axis, as the body axis is compressed in oep mutants at late gastrula stages and the notochord is slightly undulated in older embryos. A similar compression of the body axis is observed in other mutants with defects in the prechordal plate (cyclops, Brand et al., 1996; silberblick, Heisenberg et al., 1996), and in mutants with defects in the axis itself (trilobite, see below) or at its posterior end (pipetail, see below). Altogether, the phenotypes of the various mutants suggest that the extension of the body axis is driven by different forces: an anterior movement of prechordal cells at the anterior edge of the hypoblast, a posterior movement of the tailbud at the posterior end of the axis, and extension within the axis driven by medial intercalation.

While oep, cyc, slb and ppt appear to be involved in movements of anterior and posterior cells at the leading edges of the axis, spadetail (spt) and trilobite (tri) are required for convergent extension movements of cells between these edges. As previously described, spt mutant embryos display defects in the movement of ventrolateral cells to the dorsal midline. The extension movements in the midline, however, appear rather normal at gastrula and early somite stages. In contrast to spt, tri mutant embryos exhibit normal convergent movements of ventrolateral cells to the dorsal side, while the convergent extension in the axis is strongly impaired. Convergent extension movements have been shown to be driven by mediolateral cell intercalations (Warga and Kimmel, 1990). The phenotypes of spt and tri mutants, however, indicate that convergent extension movements in ventrolateral regions towards the axis and on the dorsal side of the embryo along the length of the axis can be uncoupled from one another.

Previous studies have revealed that the formation of the tail is a continuation of gastrulation and depends on genes already required during gastrulation (Gont et al., 1993; Tucker and Slack, 1995). Much of the material of the tail is formed in the tailbud and later in the tip of the tail, where a ‘mini-gastrulation’ takes place. pipetail (ppt) and kugelig (kgg) mutants, however, show specific defects during tail formation, although gastrulation and formation of tail material in the tailbud and the tip of the tail appear largely normal. This suggests that the genes regulate processes that occur in addition to the continued gastrulation processes in order to allow a normal outgrowth of the tail. Alternatively, the effects of the mutations during gastrulation might be very subtle or compensated by maternally supplied gene products. In ppt mutant embryos, the tailbud fails to move more ventrally on the yolk sac after closure of the germ ring. Later, the detachment of the tip of the tail from the yolk is impaired. Of all posterior marker genes tested, only wnt5 shows a significantly reduced expression in the posterior region of ppt mutants from early developmental stages on, suggesting that ppt might be zebrafish wnt5 or a gene involved in the activation of the wnt5 expression. We are currently testing whether the mutation in ppt is physically linked to the wnt5 gene.

In contrast to ppt mutant embryos, kgg mutants exhibit a normal ventral movement of the tailbud after closure of the germ ring and a rather normal detachment of the forming tail from the yolk. However, the formation of the yolk tube is strongly impaired, which leads to a compression of the posterior part of the body axis.

The mutants described above display specific and unique effects on different morphogenetic processes of early zebrafish development. We expect them to be valuable tools for the investigation of the various morphogenetic processes that occur during gastrulation and tail formation.

M. H. and F. P. contributed equally to this work. We are very grateful to Dr Andrew McMahon, in whose laboratory much of the mutant analysis has been carried out. Additionally, we would like to thank Ed Sullivan for his help and advice during the setting-up of a fish facility in the McMahon laboratory. Dr Eric Weinberg generously supplied us with the myoD cDNA prior to publication. Published reagents were obtained from Drs Marie-Andrée Akimenko, Jean-Stéphane Joly, Stefan Krauss and Stefan Schulte-Merker.