|Home | About | Journals | Submit | Contact Us | Français|
The complicated trajectory of facial motor neuron migration requires coordination of intrinsic signals and cues from the surrounding environment. Migration begins in rhombomere (r) 4 where facial motor neurons are born and proceeds in a caudal direction. Once facial motor neurons reach their target rhombomeres, they migrate laterally and radially from the ventral neural tube. In zebrafish, as facial motor neurons migrate through r5/r6, they pass near cells that express olig2, which encodes a bHLH transcription factor. In this study we found that olig2 function is required for facial motor neurons to complete their caudal migration into r6 and r7 and form stereotypical clusters. Additionally, embryos that lack mafba function, in which facial motor neurons also fail to complete caudal migration, lack olig2 expression in r5 and r6. Our data raise the possibility that cells expressing olig2 are intermediate targets that help guide facial motor neuron migration.
Forming a nervous system requires that cells not only be born at the right place and time, but often that they also migrate. Migratory, post-mitotic neurons arise from neuroepithelial precursors, extend leading processes and move, sometimes in a multidirectional fashion, toward their target destinations (Rakic, 1990; Hatten, 1999; Rakic, 1999; Marin and Rubenstein, 2003). Neuronal migration disorders, many of which are associated with intractable epilepsy, highlight the significance of proper neuronal migration (Copp and Harding, 1999). Among known populations of migrating neurons are facial motor neurons, which are specified in rhombomere (r) 4 and then migrate caudally to occupy other regions of the hindbrain (Altman and Bayer, 1982; Ashwell and Watson, 1983; Auclair et al., 1996; Garel et al., 2000; Studer, 2001). Whereas this caudal migration is conserved across species, the target rhombomere can vary. The facial motor neurons of many species, such as mouse, shark, lizard and salamander, terminate in r5 and r6 (Barbas-Henry, 1982; Roth et al., 1988; Gilland and Baker, 1993; Guthrie, 2007), whereas in other species, such as zebrafish, facial motor neurons migrate intro r6 and r7 (Chandrasekhar et al., 1997). In birds, only a small subset of the facial motor neurons migrate caudally into r5 (Jacob and Guthrie, 2000). Following caudal migration to the target rhombomeres, facial motor neurons move both laterally and radially (Chandrasekhar, 2004; Guthrie, 2007; Song, 2007).
Various molecules that are necessary for facial motor neuron migration have been described. For example, facial motor neurons express the transcription factors Nkx6.1, Hoxb1, Phox2b and Pbx4 (Goddard et al., 1996; Cooper et al., 2003; Muller et al., 2003; Pattyn et al., 2003; Guthrie, 2007). Facial motor neurons fail to migrate caudally in mice that either lack Nkx6.1 or Hoxb1 functions or in which Phox2a is substituted for Phox2b, indicating that each of these factors is necessary for the initial phase of facial motor neuron migration (Goddard et al., 1996; Pattyn et al., 2000; Studer, 2001; Muller et al., 2003; Pattyn et al., 2003; Coppola et al., 2005; Song et al., 2006). Zebrafish embryos that lack function of lzr/pbx4, which encodes a Hox binding partner, similarly exhibit a facial motor neuron migration initiation defect (Cooper et al., 2003). The non-canonical Wnt/planar cell polarity (PCP) genes trilobite/strabismus (Bingham et al., 2002; Jessen et al., 2002), off-limits/frizzled3a (Wada et al., 2005; Wada et al., 2006), off-road/celsr2 (Wada et al., 2006), landlocked/scribble1 (Wada et al., 2005), colgate/hdac1 (Nambiar et al., 2007) and prickle1b (Rohrschneider et al., 2007) are also required for the caudal migration of facial motor neurons. For example, frizzled3a and celsr2 are expressed in the surrounding neuroepithelium and act as chemorepellents, prohibiting the facial motor neurons from initiating their lateral and radial migration before entering r6 and r7 (Wada et al., 2006). The roles many of these genes play in facial motor neuron migration appear to be independent of their roles in the Wnt/PCP pathway regulating convergent and extension cell movements (Wada et al., 2005; Wada et al., 2006; Nambiar et al., 2007). Additionally, loss of tbx20 function, which results in a down-regulation of PCP components, disrupts facial motor neuron migration at several points (Song et al., 2006). During specific phases of migration facial motor neurons regulate the expression of Ret, a GDNF-receptor subunit (Pachnis et al., 1993) and the cell surface proteins TAG-1 (Furley et al., 1990) and Cadherin8 (Redies and Takeichi, 1996) raising the possibility that these genes help establish the necessary pathways for migration in each rhombomere (Garel et al., 2000).
In addition, tissue transplantation experiments provided evidence that r5 and r6 are the source of factors that influence facial motor neuron migration (Studer, 2001). Consistent with those observations, mouse and zebrafish mutations that perturb development of r5 and r6 cause aberrant facial motor migration (Moens et al., 1996; Chandrasekhar et al., 1997; Schneider-Maunoury et al., 1997; Garel et al., 2000; Studer, 2001). For example, r5 and r6 are mispatterned in mafba−/− embryos, creating a single rhombomere called rX (Moens et al., 1996), and facial motor neuron migration becomes disorganized as they move caudally (Chandrasekhar et al., 1997; Bingham et al., 2010). We recently showed that neuroepithelial cells of r5 and r6 in zebrafish express olig2, which encodes a bHLH transcription factor (Zannino and Appel, 2009). Here we show that, in the absence of olig2 function, facial motor neurons failed to complete caudal migration. Instead, many of the facial motor neurons remained in r4 and r5 and appeared disorganized. Our data raise the possibility that olig2 contributes to facial motor neuron migration by regulating the environment through which they migrate. Consistent with this interpretation, olig2 expression was reduced specifically in rX of mafba−/− embryos while expression in other regions remained unchanged, thus indicating that the facial motor neuron migration defect exhibited in mafba−/− embryos is due, at least in part, to the lack of olig2 function in r5 and r6. Additionally, mafba−/− embryos expressed sdf1a, a chemokine necessary for facial motor neuron migration (Sapede et al., 2005), at reduced level in r6. However, the expression of sdf1a was unaffected in olig2 morphants suggesting that olig2 is not necessary for sdf1a expression. This indicates that both olig2 and sdf1a are downstream of mafba and required for facial motor neuron migration, but may function independent of each other.
Previously, we demonstrated that neuroepithelial precursor cells in r5 and r6, which lie along the path of migrating facial motor neurons, express olig2 (Zannino and Appel, 2009). However, the facial motor neurons themselves do not express olig2. Olig2 function can be efficiently blocked by antisense morpholino oligonucleotides (MOs) (Park et al., 2002; Bae et al., 2005; Filippi et al., 2005; Lewis et al., 2005; Yeo and Chitnis, 2007; Borodovsky et al., 2009; Schebesta and Serluca, 2009; Zannino and Appel, 2009; Yang et al., 2010; Lim et al., 2011). To investigate the possibility that olig2 influences facial motor neuron migration, we therefore used MOs designed to block translation of olig2 (Zannino and Appel, 2009) and examined facial motor neuron distribution using Isl immunocytochemistry and Tg(isl1:GFP)rw0 transgenic reporter expression. The Isl antibody labels all motor neuron cell bodies (Korzh et al., 1993) whereas the Tg(isl1:GFP)rw0 transgene labels all motor neurons and their axonal projections except for the abducens and glossopharyngeal motor neurons (Higashijima et al., 2000).
At 28 hours post fertilization (hpf) in wild-type and control MO-injected embryos two branchiomotor populations were revealed by Tg(isl1:GFP)rw0 reporter expression, nV clustered in r2 and r3, and nVII positioned from r4 to r6 (Fig. 1A,G). The axons of nV extended toward their exit point in r2. The facial motor neurons (nVII) formed bilateral, ventromedial columns extending from r4 into r6 and extended axons toward their r4 exit point (Fig. 1A,G). By 33 hpf a third branchiomotor population, nX, was specified in the caudal hindbrain (Fig. 1B,H) and the facial motor neurons resided primarily in r6 and r7, forming clusters in each rhombomere. The more posteriorly located facial motor neurons also began a lateral migration, moving away from the midline (Fig. 1B,H). While the positions of nV and nX remained largely unchanged, the facial motor neurons completed their migration by 50 hpf and formed a large cluster in r6 and a smaller, looser cluster in anterior r7 (Fig. 1C,I). These clusters were bilaterally located and occupied positions slightly lateral to the midline. At 3 days post fertilization (dpf) the distribution of branchiomotor neurons marked by Tg(isl1:GFP)rw0 reporter expression remained largely unchanged in wild-type and control embryos (Supp. Fig. S1A and data not shown).
We detected a defect in facial motor neuron migration as early as 28 hpf in olig2 deficient embryos. The facial motor neurons in olig2 MO-injected embryos clustered predominately in r4, with only a few cell bodies located in r5 or more posteriorly. However, their axons extended correctly towards the r4 exit point (Fig. 1D). While some of the facial motor neurons in olig2 MO-injected embryos migrated into r6 by 33 hpf, the majority continued to reside in r4 and r5, failing to form the appropriate clusters (Fig. 1E). The few facial motor neurons that reached r6 appeared to undergo normal lateral movement (Fig. 1E). Mislocalization of facial motor neuron cell bodes in olig2 MO-injected embryos persisted at 50 hpf and 3 dpf with the majority of cells located too anteriorly, failing to make the appropriate clusters in r6 and r7 (Fig. 1F, Supp. Fig. S1B). However, the facial motor neuron axons in olig2 MO-injected embryos appeared fasciculated and exited r4 similar to wild-type embryos. Distributions of trigeminal, glossopharyngeal and vagal motor neurons revealed by Isl immunocytochemistry and isl1 RNA expression in wild-type and olig2 MO-injected embryos were similar to those described here using transgenic reporter expression (Supp. Fig. S2). Additionally, co-injection of the olig2 MO along with a p53 MO, which inhibits apoptosis that could result from non-specific effects of the olig2 MO through the p53 pathway, had no effect on the facial motor neuron migration phenotype at 24, 33, and 50 hpf (Supp. Fig S3A–I). We conclude that loss of olig2 function impedes facial motor neuron movement through r5 and r6, consistent with the possibility that r5 and r6 olig2+ cells contribute to a guidance mechanism for facial motor neuron migration.
One possible explanation for the above results is that loss of olig2 function disrupts hindbrain patterning thereby disrupting positional cues necessary for facial motor neuron migration. We therefore used gene expression to investigate hindbrain patterning in the absence of olig2 function and to more precisely compare facial motor neuron migration in control and olig2 deficient embryos.
In wild-type 19 hpf embryos, hoxb1a expression marked the origin of facial motor neurons in r4 (Fig. 2A). As reported (Zannino and Appel, 2009), olig2 expression was restricted to r5/r6 and did not extend into r4 (Supp. Fig. S4A). Accordingly, the hoxb1a expression domain was unaffected in olig2 MO-injected embryos and included isl1+ cells that failed to migrate caudally (Fig. 2B). Quantitative PCR analysis further confirmed that hoxb1a expression was not decreased in embryos lacking olig2 function (Supp. Fig. S4B). Additionally, the expression level of prickle1b, a downstream target of hoxb1a expressed in facial motor neurons (Rohrschneider et al., 2007; Bingham et al., 2010), was unchanged at 24 hpf in olig2 MO-injected embryos as compared to wild-type and control MO-injected embryos (Supp. Fig. S4C). At 22 and 26 hpf, egr2b expression was similar in wild-type and olig2 MO-injected embryos, indicating that olig2 function is not necessary for patterning r3 and r5, but in comparison to wild type, many more facial motor neurons occupied r4 and r5 in olig2 deficient embryos (Fig. 2C–F). We next used hoxb4a expression to mark r7 at several developmental time points. Again, expression was similar in wild-type and olig2 knockdown embryos but revealed a deficit of isl1+ facial motor neurons in r7 of MO-injected embryos (Fig. 2G–J, data not shown). Therefore, the facial motor neuron migration defect in olig2 MO-injected embryos does not appear to result from disruption of anteroposterior hindbrain patterning.
We then quantified the facial motor neuron migration defect in olig2 MO-injected and control MO-injected embryos compared to wild-type embryos at three different time points. We divided the olig2 loss of function phenotype into two classes, a mild phenotype in which some facial motor neurons were located in the correct positions, and a severe phenotype in which the vast majority of facial motor neurons were located more anteriorly than wild-type facial motor neurons and failed to form the stereotypical clusters. At 28 hpf, approximately 10 hours into their migration, 100% (45/45) of wild-type embryos had facial motor neurons migrating in a line extending from r4 into r5 and r6 and 90.9% (10/11) of embryos injected with a control MO were comparable to the wild type. 89.9% (62/69) of olig2 MO-injected embryos exhibited a facial motor neuron deficit at 28 hpf, with 50.7% exhibiting a mild phenotype and 39.1% a severe phenotype. By 33 hpf, midway through migration, 95.5% (21/22) of wild-type embryos and 84.2% (16/19) of control MO-injected embryos had facial motor neurons beginning to form their stereotypical clusters in r6 and r7 with few remaining in r4. In contrast, only 8.3% (3/36) of olig2 MO-injected embryos were similar to wild type, with 69.4% demonstrating a mild phenotype and 22.2% showing a severe phenotype. Finally, at 50 hpf, at least two hours after migration should be complete, 96.7% (30/31) of wild-type embryos and 80.0% (16/20) of control MO-injected embryos had facial motor neurons located in r6 and r7 and in their stereotypical clusters. In contrast 97.7% of olig2 MO-injected embryos exhibited a facial motor neuron migration defect, 56.7% with a mild phenotype and 40.0% with a severe phenotype.
One possible explanation for the delayed and incomplete migration of facial motor neurons in our loss of function experiments is that the olig2 MO slowed development, which can result from off-target effects (Corey and Abrams, 2001; Heasman, 2002; Eisen and Smith, 2008). To investigate this, we monitored the movements of cells outside the CNS, which should be independent of olig2 function. In zebrafish, the cells of the posterior lateral line ganglia migrate posteriorly at a consistent rate and their position along the body axis can be used as a developmental staging tool (Kimmel et al., 1995). At 23, 33 and 50 hpf the posterior lateral line ganglia, revealed by Elavl immunocytochemistry, were in similar positions in wild-type, control injected and olig2 MO-injected embryos (Fig. 3) indicating that the facial motor neuron migration defect did not result from generalized developmental delay.
To investigate directly the migratory behaviors of facial motor neurons in embryos lacking olig2 function, we performed in vivo time-lapse imaging. As noted above, in wild-type embryos the Tg(isl1:GFP)rw0 reporter first revealed facial motor neurons by 19 hpf (Fig. 4). Over the next several hours more facial motor neurons expressed the reporter and the cells began to migrate posteriorly along the midline of the neural tube. At approximately 23 hpf the facial motor neuron axon bundles extended anteriorly from the migrating cells to exit the neural tube from r4. By 22 hpf the facial motor neurons began to enter r6, and by 26 hpf they turned laterally. By 25 hpf reporter gene expression began to mark vagus motor neurons in the caudal hindbrain and at 28 hpf a small population of facial motor neurons formed a cluster in r7. Over the next 10 hours the facial motor neurons continued to migrate caudally into r6 and r7, where they turned laterally, while their axons extended rostrally toward their r4 exit point.
Similar to wild-type embryos, the facial motor neurons in olig2 MO-injected embryos were first visible by transgenic reporter expression by 19 hpf, on either side of the midline (Fig. 5). Over the next few hours some of these cells began to migrate posteriorly, however, they did not move as far posteriorly as in the wild-type embryos. At 23 hpf the facial motor neuron axon bundles started to extend anteriorly, as in wild-type embryos. While a small number of cells continued their caudal migration, the majority of facial motor neuron cell bodies remained clustered in r4 and r5. Around 29 hpf the first vagal motor neurons (nX) became apparent and subsequently increased in number, as in wild-type embryos. The apparent reduction of nX motor neurons in the olig2 MO-injected embryos is due to the focal plane of the time-lapse images. Other flat mount images showed similar numbers of nX motor neurons in wild-type and olig2 morphant embryos (Fig. 1C, F, I, Fig. 3G–I, Supp. Fig. S2). At the end of this time-lapse sequence, at 36 hpf, the majority of facial motor neuron cell bodies remained in r4 and r5 with only a few occupying r6 and r7. However, they correctly extended their axons, which properly exited from r4.
There are many conditions that perturb facial motor neuron migration, for example, mafba−/−embryos in which r5 and r6 fail to be specified and a precursor rX remains (Moens et al., 1996). The facial motor neurons in these embryos do not form organized clusters and instead are scattered throughout rX. Whereas the facial motor neuron axons still correctly exit from r4 they are defasciculated. In addition, the glossopharyngeal motor neurons (nIX), which are born in r6 and migrate to r7, are absent in mafba−/− embryos (Chandrasekhar et al., 1997; Chandrasekhar, 2004). Because the facial motor neuron migration phenotype in mafba−/− embryos resembles the defect we observe with functional loss of olig2, which is initially expressed in r5 and r6, we reasoned that olig2 may be downstream of mafba and explain the facial motor neuron migration phenotype. To begin, we tested whether olig2 expression requires mafba function. At 24 hpf, wild-type embryos expressed olig2 in r5 and r6 (Fig. 6A, arrow), as well as more anteriorly and laterally along the edge of the hindbrain and in the diencephalon. By contrast, rX expression of olig2 was almost entirely absent from mafba−/− embryos (Fig. 6B, arrow) whereas other regions of olig2 expression remained unchanged. This is consistent with our observation that facial motor neuron migration is disrupted in embryos lacking olig2 function and raises the possibility that the deficit of r5/r6 olig2+ cells in mafba−/− embryos contributes to their facial motor neuron defect. However, loss of olig2 function in mafba−/− embryos cannot account for all the mafba−/−phenotypes because the glossopharyngeal motor neurons, labeled by anti-Alcama (formerly zn8) antibody, which are absent in mafba−/− embryos, were not lost in embryos injected with olig2 MO (Supp. Fig. S5, arrows). We previously demonstrated that loss of olig2 function significantly reduced the number of abducens motor neurons and kept cells that expressed olig2+ in a more precursor-like state (Zannino and Appel, 2009). However, at both 48 and 56 hpf, the glossopharyngeal motor neurons were unaffected in olig2 MO-injected embryos (Supp. Fig. S5B,D, arrows).
Because olig2 is a transcription factor not expressed by facial motor neurons, it is likely that it regulates other r5/r6 genes required for facial motor neuron migration. One candidate is sdf1a, which encodes a chemokine necessary for facial motor neuron migration (Sapede et al., 2005). Therefore, we examined the expression of sdf1a in wild-type and mafba−/− embryos. At 18 hpf in wild-type embryos sdf1a was expressed from r4-r6, with expression strongest in r6 (Fig. 6C, also see see6G).6G). Additionally, we observed a stripe of sdf1a expression in the anterior hindbrain as well as expression adjacent to the hindbrain along r2 and r3 (Fig. 6C). At 24 hpf the expression of sdf1a was restricted to r4 and r6, whereas the expression lateral to r2 and r3 was maintained as was the expression in the head (Fig. 6E). Although the r4 sdf1a expression, and other anterior populations, were unchanged in mafba−/− embryos, the rX expression of sdf1a appeared greatly reduced at 18 hpf (Fig. 6D). Concurrently, at 24 hpf in mafba−/− embryos the r6 expression of sdf1a was lost whereas other regions of expression were unaltered (Fig. 6F). These data suggest that sdf1a acts downstream of mafba in r5/r6.
To more precisely determine the relationship between olig2, sdf1a and facial motor neuron migration we examined the expression of sdf1a in olig2 MO-injected embryos. At 18 hpf the expression of sdf1a in olig2 MO-injected embryos appeared weaker, but in a pattern similar to that of wild-type and control MO-injected embryos (Fig. 6C,G,H). Consistently, at 24 hpf, there was no difference in the sdf1a expression between control MO-injected and olig2 MO-injected embryos (Fig. 6I–J). Thus, sdf1a expression does not appear to require olig2 function. These results indicate that both sdf1a and olig2 act downstream of mafba in regulation of facial motor neuron migration but then act independently of each other.
Facial motor neurons are one of many populations of neurons that migrate from where they are born to occupy other regions of the CNS. The trajectory of facial motor neurons is unique in that they move in a caudal direction until the cell bodies reach a specific rhombomere at which point they turn and migrate laterally and radially to occupy the dorsal neural tube (Chandrasekhar, 2004; Guthrie, 2007). Specifically, many of the cues necessary for initiating caudal migration, such as hoxb1, pbx4, nkx6.1, scribble1, hdac1, strabismus and prickle1b (Goddard et al., 1996; Studer, 2001; Bingham et al., 2002; Cooper et al., 2003; Muller et al., 2003; Pattyn et al., 2003; Cooper et al., 2005; Wada et al., 2005; Nambiar et al., 2007; Rohrschneider et al., 2007), as well as some of the cues required for the final radial migration, reelin, celsr2 and frizzled3a (Rossel et al., 2005; Wada et al., 2006), are known. In addition, the facial motor neurons are in close proximity to the early-formed medial longitudinal fascicle (MLF) and may interact with and require this tract for their caudal migration (Sittaramane et al., 2009). However, we still lack a detailed understanding of the mechanisms that guide their directional movement.
The work described here demonstrates that olig2, which encodes a basic helix-loop-helix (bHLH) transcription factor (Lu et al., 2000; Zhou et al., 2000), is required for the caudal migration of facial motor neurons. The facial motor neurons in embryos injected with antisense morpholino oligonucleotides designed to interfere with olig2 function initiated their caudal migration, but failed to completely migrate into r6 and r7 to form their stereotypical clusters. However, olig2 is not expressed by facial motor neurons, but rather in neuroepithelial precursor cells located in r5 and r6, along the route of migrating facial motor neurons (Zannino and Appel, 2009)(Fig. S4A). olig2 must, therefore, have an indirect role on facial motor neuron migration and understanding its function will provide insight into the intermediate steps between the initiation of caudal migration and radial migration.
One way that loss of olig2 function could disrupt facial motor neuron migration is by altering hindbrain patterning. In the spinal cord, Olig2 is necessary for dorsoventral patterning by inhibiting expression of genes that encode other transcription factors (Mizuguchi et al., 2001; Novitch et al., 2001; Zhou et al., 2001; Lu et al., 2002; Takebayashi et al., 2002). Consequently, alteration of positional cues that depend on mechanisms that establish polarity in the neural tube could interfere with cell migration. However, unlike the spinal cord, the hindbrain does not express olig2 in a continuous longitudinal domain suggesting that Olig2 does not function in dorsoventral patterning of the hindbrain as it does for spinal cord. Indeed, loss of olig2 function does not appear to alter dorsoventral patterning of r5 and r6 in zebrafish (Zannino and Appel, 2009). Furthermore, data presented in this manuscript indicates that loss of olig2 function also does not interfere with anterioposterior patterning of the hindbrain. Additionally, we found that migration of posterior lateral line ganglia was unaffected in olig2 MO-injected embryos, indicating that the effects of olig2 MO on facial motor neuron migration are specific to that population. We also examined embryos at 3 dpf, 24 hours after the completion of wild-type migration, and found that the facial motor neuron migration phenotype in olig2 MO-injected embryos persisted. We note that olig2-expressing cells are also found in more anterior brain regions, some near areas giving rise to the MLF, thus raising the hypothesis that the facial motor neuron migration defect seen with loss of olig2 function could be an indirect result of impaired MLF development. However, closer examination of the olig2 expression pattern reveals that the olig2+ cells are expressed in neurons contributing to the dopaminergic system in the diencephalon (Borodovsky et al., 2009) whereas the neurons contributing to the MLF are located in ventral mesencephalon (Ahsan et al., 2007; Wolman et al., 2008; Sittaramane et al., 2009). We therefore think it is unlikely that facial motor neuron migration defects in olig2 deficient embryos result indirectly from changes in hindbrain pattern, MLF formation or developmental delay.
A second possibility for the indirect role of olig2 on facial motor neuron migration is that the r5/r6 olig2-expressing cells function as a guidepost for the facial motor neurons. The timing and location of olig2 expressing cells in r5 and r6 situates them in the right place and time to directly regulate the facial motor neurons as they migrate through those rhombomeres. Facial motor neurons are born in r4 at 18 hpf (Chandrasekhar, 2004), before olig2 can be detected in r5 and r6 by RNA in situ hybridization. By 22 hpf, when the facial motor neurons are entering r5 and r6, olig2 expression was detected. Additionally, the earliest deficit in facial motor neuron migration was seen at 22 hpf. Four hours later, at 26 hpf, the facial motor neurons in r6 began to migrate away from the midline and by 28 hpf the facial motor neurons reach r7. Abducens motor neurons, as detected by olig2:EGFP and anti-Alcama antibody labeling, were not observed until 33 hpf and Sox10, the first maker expressed by oligodendrocyte precursor cells (OPC), was not seen until 40 hpf, indicating that the effects of loss of olig2 function seen in facial motor neuron migration is due to the olig2+ precursor cells, not specified abducens motor neurons or OPCs. The migration of two other neuronal cell types, pontine neurons in the hindbrain and telencephalic interneurons, is regulated as they progress past a population of cells expressing or secreting necessary cues. Pontine neurons change from a ventral to a rostral migration as they pass facial motor neurons secreting Slit2/3 (Geisen et al., 2008) and interneurons derived from the medial ganglion eminence destined for the cortex circumvent the striatum expressing semaphorins (Nobrega-Pereira and Marin, 2009). We previously showed that in the absence of olig2 function, the r5/r6 olig2+ cells remained in a more precursor-like state and failed to properly exit the cell cycle and differentiate into abducens motor neurons and oligodendrocytes (Zannino and Appel, 2009). Therefore, these immature cells might also lose the ability to guide facial motor neurons as they enter r5 and r6. Similarly, in mafba−/− embryos, rhombomeres 5 and 6 fail to be specified and a precursor rX remains. Facial motor neurons in these embryos initiate migration from r4 but fail to form clusters and appear disorganized throughout the rX region, a phenotype similar though more severe to olig2 MO-injected embryos. In this work we demonstrate that olig2 expression is specifically lost in rX of mafba−/− embryos, thus placing olig2 downstream of mafba function in regulating facial motor neuron migration and indicating that the mafba facial motor neuron phenotype is at least partly due to the loss of olig2 in r5 and r6. Whereas loss of olig2 function can explain some of the mafba−/− phenotypes, such as loss of abducens motor neurons (Zannino and Appel, 2009) and the facial motor neuron migration defect, other mafba−/− phenotypes, such as the loss of glossopharyngeal motor neurons, are unrelated to olig2 function.
The most probable explanation for the role of olig2 is that it regulates the transcription of another gene necessary for facial motor neuron migration. One candidate gene is sdf1a, a chemokine necessary for facial motor neuron migration, as well as lateral line migration (David et al., 2002; Sapede et al., 2005). There are multiple sources of Sdf1a throughout the developing embryo, consistent with different populations requiring Sdf1a for their migration (Perlin and Talbot, 2007). Previous work demonstrated that sdf1a is expressed before facial motor neuron migration in the hindbrain posterior to r4 (Knaut et al., 2005) and at 24 hpf in r4 (David et al., 2002). Here we further define this expression pattern, demonstrating that sdf1a is expressed in r4–r6 at 18 hpf, becoming restricted to r4 and r6 by 24 hpf, and downregulated in these rhombomeres by 33 hpf (data not shown). Additionally, the olig2 MO phenotype is similar to the phenotype seen in embryos injected with a MO designed to sdf1a, in which facial motor neurons remain largely in r4 and r5 (Sapede et al., 2005). However, on closer examination, the expression of sdf1a was not changed in olig2 MO-injected embryos, indicating that although the two genes are both required for facial motor neuron migration, sdf1a does not require olig2 function. Notably, the loss of both olig2 and sdf1a in mafba−/− embryos can explain why the mafba−/− facial motor neuron phenotype is more severe than loss of olig2 function alone. Other genes that might be regulated by olig2 include those of the PCP pathway, which are known to play crucial roles in facial motor neuron migration (Wada and Okamoto, 2009). Given the very localized expression of olig2 in the hindbrain, changes in expression of PCP genes in the absence of olig2 function might be very subtle and will require detailed investigation.
Wild-type and transgenic fish were raised either in the Vanderbilt University Zebrafish Facility or the University of Colorado Denver Anschutz Medical Campus Zebrafish Facility and embryos collected from pair matings. The embryos, raised at 28.5°C, were staged according to morphological criteria (Kimmel et al., 1995) and hours post-fertilization (hpf). We used the following transgenic lines: Tg(olig2:EGFP)vu12 (Shin et al., 2003) and Tg(isl1:GFP)rw0 (Higashijima et al., 2000). The valB337 (also known as mafba) mutant zebrafish were raised in the University of Massachusetts Zebrafish Facility.
The following previously described RNA probes were used: egr2b (also known as krox20) (Oxtoby and Jowett, 1993), olig2 (Park et al., 2002), isl1 (Inoue et al., 1994), hoxb1a (McClintock et al., 2002), hoxb3a (Hadrys et al., 2004), hoxd4a (Moens and Prince, 2002) and sdf1a (David et al., 2002). Embryos were fixed in 4% paraformaldehyde (pfa) and then stored in 100% methanol at −20°C. The in situ RNA hybridization was performed as previously described (Hauptmann and Gerster, 2000) followed by a color reaction with BM purple (Roche Diagnostics). For double RNA labeling probes were labeled with either digoxygenin or fluorescein. The first probe was detected with the appropriate antibody conjugated to alkaline phosphatase and followed by a color reaction with BM purple. Washing the embryos with 0.1M glycine, pH 2.2, followed by a 20 min incubation with 4% pfa inactivated alkaline phosphatase of the first antibody and the appropriate second antibody was then applied and developed with a solution of INT/BCIP (Roche Diagnostics). Once developed, the embryos were dissected from the yolk and mounted in 70% glycerol for whole mount imaging on bridged cover-slips. All images were captured using a Zeiss Axio Observer compound microscope equipped with DIC and epifluorescence optics and a Retiga Exi digital color camera. Once captured, images were imported into Adobe Photoshop and adjustments were limited to contrast, levels, color matching settings and cropping.
For immunocytochemistry we used the following primary antibodies: mouse anti-Isl (39.4D5, 1:100; Developmental Studies Hybridoma Bank (DSHB)), rabbit anti-GFP (A11122; 1:500, Invitrogen), mouse anti-Elavl (clone 16a11; 1;100; Molecular Probes), and mouse anti-Alcama (also known as zn8; 1:1000; DSHB). For fluorescence detection, the following Alexa Fluro secondary antibodies were used: 568 goat anti-mouse and 488 goat anti-rabbit (1:200; Invitrogen).
Embryos for whole mount antibody labeling were fixed in 4% AB fix (4% pfa, 8% sucrose, 1x PBS) overnight at 4°C and pre-blocked with 10% sheep serum/BSA-1x PBS for 1 hr at RT. The embryos were incubated in primary antibody for 24 hr at 4°C, washed semi-continuously with 1x PBS with 0.2x Trition (PBSTx) for 2 hr at RT, and then incubated with the secondary antibody for 12 hr at 4°C, followed by 3 hr of semi-continuous washes with 1x PBSTx. These embryos were then dissected from the yolk and mounted on bridged cover-slips in 70% glycerol for imaging. Images were collected using a PerkinElmer UltraVIEW VoX confocal imaging system mounted on a Zeiss Axio Observer compound microscope. Images were imported into Adobe Photoshop and adjustments were made to contrast and levels settings. Embryos for sectioning were embedded in 1.5% agar/5% sucrose, frozen with 2-methyl-butane chilled by immersion in liquid nitrogen, and sectioned using a cryostat microtome (10 μm). Sections were re-hydrated with 1x PBS and pre-blocked for 30 min in 2% sheep serum/BSA-1x PBS. The sections were incubated with primary antibody overnight at 4°C, washed extensively with 1x PBS and incubated with the appropriate fluorescent secondary antibody for 2 hr at RT. Once the secondary antibody was washed off sections were covered with Vectashield (Vector Laboratories).
Embryos for whole mount in situ RNA hybridization followed by antibody labeling were fixed as above in 4% AB fix for 2 hrs at RT. The in situ RNA hybridization was performed as described in the previous section, however, the embryos were not treated with 100% methanol. Once the color reaction with BM purple concluded the embryos were washed with 1x PBS with % Tween (PBSTw) for two 5 min washes. The embryos were then fixed in 4% pfa for 20 min to deactivate the alkaline phosphatase. The fix was washed off with 1x PBSTx followed by 10% block solution for 1 hr. Whole mount immunocytochemistry was completed as described above. Following in situ RNA labeling and antibody labeling, embryos were dissected from the yolk and mounted on bridged cover-slips in 70% glycerol for imaging. DIC images of in situ RNA hybridizations were overlapped with the corresponding fluorescent images using Volocity software. Images were exported and analyzed as described above.
An antisense morpholino (MO) oligonucleotide with the sequence 5′-ACACTCGGCTCGTGTCAGAGTCCAT-3′ (Gene Tools, LLC) was designed to the olig2 translation start site (Park et al. 2002; Zannino and Appel 2009). We also used a Standard Control MO (Gene Tools, LLC). Both morpholinos were re-suspended in distilled water for a stock solution of 3 mM. The stock solution was further diluted with water and phenyl red and 4–5 ng was injected into the yolk of one- to two-cell stage embryos.
The embryos were manually dechorionated and transferred to embryo medium containing PTU. Embryos for time-lapse imaging were anesthetized using Tricaine and immersed in 0.8% low-melting temperature agarose. They were then mounted in dorsal orientations in glass-bottom 35mm Petri dishes. Images were captured using a 20x objective (NA=0.8) mounted on the confocal microscope described above. A heated stage and chamber kept the embryos at 28.5°C and Z image stacks were collected every 5–15 min. The data sets were analyzed using Volocity software and exported as QuickTime files to create movies.
Total RNA was isolated from whole embryos using TRIzol reagent (Invitrogen). Then second stand cDNA was synthesized using Superscript Reverse Transcriptase (Invitrogen) and utilized for quantitative PCR using QuantiTect SYBR Green PCR mix (Qiagen) with gene specific primers. We used a DNA Engine Opticon (MJ Research) instrument to quantify the PCR amplification and the relative amount of mRNA was compared to a control gene (βactin) run at the same time. For hoxb1a CTAGTGACAGCTATAACGCTGATGGACGAC foward primer and CTTGTTGTCCCAGTTCCACCATAGGTAAGG reverse primer were used, and for prickle1b CCCTTCTGTTGTGGATGCTTTGAGTCATTG forward primer and AGAGGTCCTGTAATCTGTTGCTGAG reverse primer were used. A Student’s t-test was used to analyze qPCR results from three separate trials.
Distribution of branchiomotor neurons marked by Tg(isl1:GFP)rw0reporter expression at 3 dpf. Dorsal whole mount with anterior to the l ft. A: Wild-type embryo. B: olig2 MO-injected embryo. Cranial motor neurons labeled “n”.
Expression of isl RNA and protein in wild-type, control MO-injected and olig2 MO-injected embryos. Dorsal whole mount with anterior to the left. A–C: isl1 RNA expression in 24 hpf embryos. D–F: Isl protein expression detected by immunocytochemistry in 24 hpf embryos. G–I: isl1 RNA expression in 33 hpf embryos. J–L: Isl protein expression detected by immunocytochemistry in 33 hpf embryos. M–O: isl1 RNA expression in 50 hpf embryos. P–R: Isl protein expression detected by immunocytochemistry in 50 hpf embryos. Otic vesicles outlined.
Tg(isl1:gfp)rw0 reporter expression reveals that the facial motor neuron migration defect in olig2 MO-injected embryos is not affected by co-injection of p53 MO. Panels are images showing dorsal views of whole embryos with anterior to the left. A–C: Wild-type embryos. D–F: olig2 MO-injected embryos. G–I: olig2 MO-injected embryos co-injected with p53 MO. Arrows indicate posteriorly migrated facial motor neurons. Bar indicates facial motor neurons along their caudal migration. Cranial motor neurons labeled “n”.
olig2 is not expressed in r4 and olig2 loss of function has no effect on hoxb1a or prickle1b expression. A: Wild-type expression at 22 hpf of olig2 in r5 and r6 in blue and egr2b in r3 and r5 in red. B: qPCR of hoxb1a expression in wild-type, control MO-injected, and olig2 MO-injected embryos at 24 and 33 hpf. C: qPCR of prickel1b expression in wild type, control MO-injected, and olig2 MO-injected embryos at 24 hpf. Error bars represent standard deviation. p value calculated by Student’s t-test is listed above each bar and compared to wild-type embryos.
Glossopharyngeal motor neuron formation does not require olig2 function. Tg(olig2:EGFP) embryos labeled with anti-Alcama antibody (red) in wild-type and olig2 MO-injected embryos. Lateral sections, dorsal to the top, anterior to the left. A,B: Wild-type embryos. C,D: olig2 MO-injected embryos. Glossopharyngeal motor neurons indicated by arrows. Rhombomeres labeled “r”.
Grant sponsor, NIH; Grant number: NS046668; Grant number: MH064913; Grant number: NS038183
We thank members of the Appel lab, Sagerström lab and Genia Olesnicky for discussions. The anti-Isl antibody, developed by T. M. Jessell, was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biological Sciences, Iowa City, IA 52242. This work was supported by NIH grant NS046668, Training in Fundamental Neuroscience Grant T32 MH064913 (D.A.Z), and NIH Grant NS038183.