Transgenic NBT-Katushka γ-cry-Venus [Xla.Tg(tubb2b:Katu shka;cryga:Venus)^EXRC; Love et al., 2011] and custom-bred NBT-Katushka pax6-GFP [Xla.Tg(tubb2b:Katushka;cryga:Ven us;pax6:GFP;CMV:DsRED)^Manzini] larval Xenopus laevis were raised in the animal husbandry facility at the Institute of Animal Physiology of the Justus-Liebig-University of Gießen, Germany. Transgenic NBT-Katushka pax6-GFP frogs were naturally bred from transgenic NBT-Katushka γ-cry-Venus and pax6-GFP CMV-DsRED [Xla.Tg(pax6:GFP;CMV:DsRED)^Papal; Hartley et al., 2001] Xenopus laevis. Animals were kept in water tanks of 1.8–7.5 l at a water temperature of 19–22°C and fed with a mix of spirulina and chlorella algae (MS-Tierbedarf). For all experiments, tadpoles of stages 48–52 were used as a representative of the premetamorphic developmental phase (Nieuwkoop and Faber, 1994). All animal procedures were performed following the guidelines of Laboratory Animal Research of the Institutional Care and Use Committee of the Justus-Liebig-University of Gießen (649_M).

All animals were anesthetized in 0.02% MS-222 solution (ethyl 3-aminobenzoate methanesulfonate; TCI Germany), dissolved in tap water, and killed by severing the spinal cord at the level of the brainstem. Tissue blocks containing the whole brain and spinal cord were dissected. Palatial tissue covering the brain was removed, and the blocks were collected in frog Ringer solution (98 mM NaCl, 2 mM KCl, 1 mM CaCl₂, 2 mM MgCl, 5 mM Na-pyruvate, 5 mM glucose, 10 mM HEPES, pH 7.8, osmolarity of 230 mOsmol/l). For fixation, the samples were immersed in 4% formaldehyde for 1 h and washed with PBS (137 mM NaCl, 2.7 mM KCl, 8 mM Na₂HPO₄, 1.4 mM KH₂PO₄, dissolved in purified water, pH 7.4). Several blocks were then embedded in 5% of low-melting point agarose (Sigma-Aldrich) and cut into 300-μm-thick coronal slices using a half-automatic vibratome (Leica VT1200S; Leica Biosystems).

To help localize neurons with tubb2b-dependent fluorescence in the larval forebrain, we performed immunohistochemical stainings against Pax7, calretinin (CR), and tyrosine hydroxylase (TH). All tissue slices were permeabilized using PBST (PBS containing 0.2% Triton-X100; Carl Roth). Unspecific binding sites were blocked with PBST with 2% normal goat serum (NGS; MP Biomedicals) for 1 h at room temperature. The slices were incubated with primary antibodies (1:100) in PBST and 2% NGS at 4°C for 3 days. Katushka was enhanced with anti-tRFP (AB233, polyclonal, derived from rabbit; BioCat) in combination with anti-HuC/D (A-21271, monoclonal, derived from mouse; Thermo Fisher Scientific), anti-calretinin (6B3, monoclonal, derived from mouse; Swant), anti-Pax7 (PAX7, monoclonal, derived from mouse), or anti-tyrosine hydroxylase (22941, monoclonal, derived from mouse; Immunostar). PAX7 was deposited to the DSHB by Kawakami, A. (DSHB Hybridoma Product PAX7). GFP was enhanced with anti-GFP (ab1218, monoclonal, derived from mouse; Abcam). The primary antibodies were washed off with PBS, and the samples were incubated with the secondary antibodies Alexa Fluor 594 goat anti-rabbit (Invitrogen, Thermo Fisher Scientific; 1:100) and Alexa Fluor 488 goat anti-mouse (Invitrogen, Thermo Fisher Scientific; 1:100) in PBS with 2% NGS at 4°C for 3 days. After several washing steps with PBS, the samples were transferred to a recording chamber for multiphoton microscopy (A1R MP; Nikon).

All images were recorded at a z-resolution of 1 μm at an excitation wavelength of 780 nm. Image processing was performed using open-source processing software ImageJ (Schindelin et al., 2012).¹ The raw image data were acquired as multidimensional image stacks. Images obtained from whole mount preparations are shown as maximum projections of all planes of the recorded z-stacks. Images obtained from the coronal slices are shown as individual z-planes. Due to the size of the samples, multiple image stacks were stitched together to show larger brain areas in one picture (Preibisch et al., 2009). Some images were cropped to show details. Brightness and contrast were linearly adjusted in all images. Labeled cells were manually counted using ImageJ ROI Manager in the OB as an example brain region. Mean counts are presented with standard deviation.

In anurans, the OB is the most anterior telencephalic region (Figure 1L) and is structured in six distinct layers (Byrd and Burd, 1991; Manzini and Schild, 2010; Manzini et al., 2022). The OB, which is connected to the olfactory epithelium via the olfactory nerve (ON), is the first relay station of the olfactory system (Manzini and Schild, 2010). We found tubb2b-dependent fluorescence in cell bodies of olfactory receptor neurons (ORNs) in the MOE (Figures 1A–C) and associated axons within the ON (Figures 1I–K). Terminals of these axons in the glomerular layer of the main and accessory OB also show tubb2b-dependent fluorescence (projection fields 1–9, α, β, γ, and δ; Figures 1D–G, 2A–D). Within the OB of larval Xenopus laevis, axon terminals of ORNs form nine large and four small projection fields, which are highly conserved among anurans (Gaudin and Gascuel, 2005; Weiss et al., 2020, 2021). We detected tubb2b-dependent fluorescence in axon terminals innervating all of these regions (projection fields 1–9, α, β, γ, and δ; Figures 1D–G, 2A–D) and in a bundle of fibers that project through the OB toward more caudal brain regions (Figure 1I; white arrow).

To visualize all mature neurons of the OB and compare the pattern of tubb2b-expressing cells, antibody stainings against HuC/D were done in NBT-Katushka γ-cry-Venus transgenic tadpoles (Figures 3A–C). HuC/D expression was detected in all cell layers of the whole OB (Figures 3A,B; green). We found that tubb2b is not expressed in all neurons and surprisingly just in a limited number of cells (Figures 3A,C; magenta). Unlike HuC/D, tubb2b-positive cells were found only in the mitral cell layer and close to projection fields 1–9 (Figures 3A,C). For the whole OB, we counted an average number of 6,517 ± 210 HuC/D-positive cells (Figure 3D; green; n = 3). In comparison, we calculated an average number of 1,588 ± 191 tubb2b-positive cells (Figure 3D; magenta; n = 10). As the overall density of tubb2b-positive cells compared to HuC/D-positive cells, we calculated an average of 24.4% (Figure 3D). To assign tubb2b-dependent fluorescence to the specific bulbar cell types, we performed antibody stainings against CR and TH in transgenic NBT-Katushka γ-cry-Venus tadpoles and used transgenic NBT-Katushka pax6-GFP tadpoles. CR-positive cells were predominantly located in the granule cell layer of the OB (Figure 1I; green) and, to a lesser extent, in the mitral cell layer. All CR-positive cells did not show tubb2b-dependent fluorescence (Figure 1I; magenta). TH-positive cells were located in both the mitral cell layer and the periglomerular layer of the OB (Figure 1J; green), but not in the granule cell layer. Again, no cells exhibiting both TH-staining and tubb2b-dependent fluorescence were identified (Figure 1J; magenta). tubb2b-dependent fluorescence was exclusively located in some, but not all, cells of the mitral cell layer of the main and the accessory OB (Figures 1E–H, 2C,D, 3A,C, 4A; magenta). In this layer, we also found pax6-positive cells (Figures 1F–H,K; green), some of them also having tubb2b-dependent fluorescence (Figures 1G,H; asterisk). Additionally, a group of cells that show tubb2b-dependent fluorescence was located in the nucleus accumbens (Acc; Figures 1K, 2D). This group of cells is also visible in Figure 4A. Laterally, adjacent to Acc, we found pax6-positive cells (green) and cells with tubb2b-dependent fluorescence (magenta) in the striatum (Str; Figure 1K).

Figure 4A shows a coronal slice at a posterior position of the anterior telencephalon. At this level in the anterior–posterior axis, we found tubb2b-dependent fluorescence (magenta) in cells of the Str, Acc, and accessory OB (projection field 8). First, cells show tubb2b-dependent fluorescence in the pallium close to the lateral ventricle (Figures 2D, 4A). A little farther back in the telencephalon, tubb2b-dependent fluorescence is located in cells of the medial pallium (Mp; Figures 2E,F, 4B,E), lateral pallium (Lp; Figures 2E, 4B,D,E), and ventral pallium (Vp; Figures 2E,F, 4D,E). tubb2b-dependent fluorescence is also present in cells of the medial amygdala (MeA; Figures 2F, 4C,E,G), the central amygdala (CeA; Figures 4G,H), and the septum (S; Figure 4C). Several regions with tubb2b-dependent fluorescence also show pax6-dependent fluorescence (Figures 4B–D,G,H; green). Thereby, the most prominent group of cells that shows tubb2b-dependent fluorescence was present at the level of the CeA (Figures 4G,H). An additional group of cells with tubb2b-dependent fluorescence was found close to the dorsal pallidum (DP) and at the position of the lateral forebrain bundle (lfb; Figures 2F, 4E). tubb2b-dependent fluorescence-positive fibers were additionally observed in the optic nerve (OpN; Figures 2F, 4E,F,H) and cell bodies in the preoptic area (POA) close to the lateral ventricle (V; Figures 2F, 4E,F). At the level of the anterior commissure (ac), we found tubb2b-dependent fluorescence in fibers that project to the contralateral brain hemisphere (Figures 2F, 4H). Moreover, we found that cells of the hypothalamus show strong tubb2b-dependent fluorescence (Hyp; Figures 2G, 5B–D).

In the diencephalon (Figure 5E), we found tubb2b-dependent fluorescence in the subventricular zone of prosomeres 1–3 (p1–3; Figures 2G, 5A–D). In p2, some cells were found in the Th (Figures 5B,C). Similar to the observations made in the telencephalon, tubb2b-dependent fluorescence was not found in all neurons of the diencephalon. For instance, at the level of the prethalamus and Th, we found CR-positive fibers that clearly did not show tubb2b-dependent fluorescence (Figures 5B,C; green). Additionally, no tubb2b expression was seen in CR-positive cell bodies of the Th (Figures 5B,C; green). Moreover, a group of Pax7-positive cells (green) in prosomere 3 (p3) were tubb2b negative (Figure 5D).

β-Tubulin isotypes can occur in different tissues and different subtypes of cells from the same tissue (Eddé et al., 1981; Moura Neto et al., 1983; Sullivan and Cleveland, 1986). In the African clawed frog (Xenopus laevis), class II β-tubulin protein is exclusively and widely expressed in the nervous system (Dworkin-Rastl et al., 1986; Oschwald et al., 1991; Bieker and Yazdani-Buicky, 1992) and found to be specially expressed in neurons (Moody et al., 1996). In this study, we monitored the activity of class II β-tubulin tubb2b promoter in the forebrain of premetamorphic Xenopus. We confirm that it is active in neurons in most main forebrain regions, but we unambiguously found that it is not active in all neurons of these regions. In the OB, tubb2b-dependent fluorescence is absent from TH- and CR-immunoreactive interneurons, and only located in the mitral cell layer and periglomerular layer. Cell counts of tubb2b-positive cells in comparison to postmitotic HuC/D-positive cells of the OB show that tubb2b is only active in 24.4% of all mature neurons in this region. In more posterior areas of the telencephalon, the expression of tubb2b is lacking in pax6-positive neurons of the CeA, as well as the Mp and Lp. Within the diencephalon, we did not find tubb2b-dependent fluorescence in CR-positive neurons and fibers of the Th. Pax7-positive neurons in prosomere 3 also lacked tubb2b-dependent fluorescence. The results of the present study show that class II β-tubulin promoter is not active in all forebrain neurons. According to our observations made in the transgenic lines we used, β-tubulin class II cannot be considered a pan-neuronal marker in Xenopus laevis.

Polyploidy is a common condition in frogs, and Xenopus laevis is one of the few allotetraploid species with four sets of chromosomes (Bisbee et al., 1977; Kobel and Du Pasquier, 1986; Evans et al., 2004; Session et al., 2016). Thus, we cannot fully exclude a mosaic expression of the genome-inserted transgenes in the Xenopus lines we investigated. Also, expression patterns of genes and proteins can vary during the development of the animal. In zebrafish (Danio rerio), the expression of another tubulin isotype, class I β-tubulin (zβ1), is restricted to the nervous system (Oehlmann et al., 2004). In larvae, zβ1 is expressed in defined peripheral and central nervous system zones that comprise early differentiating neurons. During development, zβ1 decreases, and in adults, its expression is limited to some proliferative brain zones (including the OB) and the olfactory epithelium (Oehlmann et al., 2004). Similarly, tubb2b-dependent fluorescence in Xenopus larvae is also present in brain regions with proliferative activity, such as the Acc, Lp, Mp, Dp, the ventral POA, and the ventral Hyp (D’Amico et al., 2011). In contrast to the zβ1 expression in zebrafish, in larval Xenopus, tubb2b-dependent fluorescence was also present in brain regions not described to have proliferative activity, for example, the DP and Th (D’Amico et al., 2011). This characteristic expression pattern suggests that in premetamorphic Xenopus, β-tubulin class II could play a role in neurogenesis and have other functions in mature neurons. Because in this study, we focus on premetamorphic Xenopus larvae, it must be considered that the tubb2b expression in other developmental tadpole stages and in the adult frog could be different. Not only the expression patterns within the same tissue change but also the activity within individual neurons might be different. It has been shown that the expression of β-tubulin isotypes in single mammalian cells can change during development (Guo et al., 2011). In larval Xenopus laevis, like in mice, class II β-tubulin is involved in axonal growth and elongation (Moody et al., 1996). In mice, β-tubulins I, II, and III are expressed in neurons of the adult brain (Guo et al., 2011). In cultured neurons of newborn mice, β-tubulin II is no longer expressed in some parts of the cell bodies after differentiation and starts being expressed in neurites, where it might contribute to the regulation of neurite outgrowth (Guo et al., 2011). In the present study, we found that in premetamorphic Xenopus laevis, tubb2b-dependent fluorescence is mainly located in cell bodies. We observed a very strong tubb2b-dependent fluorescence in axons of the ON and OpN. Additionally, we found fibers of the ON that project toward the midbrain and the contralateral brain hemisphere without terminating in the OB. All three structures have their axons cover a relatively long distance, and their cell bodies are located in peripheral organs (Cima and Grant, 1982; Burd, 1991; Pinelli et al., 2004). Because we did not find tubb2b-positive neurites of intracerebral neurons, we assume that the expression of tubb2b in fibers from the periphery could play a special role and might be important to enable efficient intracellular transport over long distances.

In vertebrates, the olfactory system has an exceptional regenerative ability (Graziadei and Dehan, 1973; Graziadei and Monti Graziadei, 1983; Steuer et al., 2014; Yu and Wu, 2017; Calvo-Ochoa et al., 2021). Within the olfactory epithelium, olfactory sensory neurons undergo a constant natural turnover (Graziadei and Monti Graziadei, 1983; Higgs and Burd, 2001; Brann and Firestein, 2014) and have the ability to regrow after a lesion (Schwob, 2002). In premetamorphic Xenopus laevis, 7 weeks after ON injury, newborn olfactory sensory neuron axons fully reinnervate the OB (Hawkins et al., 2017). It has been shown that newly formed glomerular clusters and second-order neurons respond to olfactory stimuli during functional calcium imaging (Hawkins et al., 2017). In mammals, class II β-tubulin is associated with axonal growth and elongation (Joshi and Cleveland, 1989). Thus, it might also be involved in the regeneration of ON axons. In Xenopus embryos, class II β-tubulin has already been found during the differentiation of pioneering neurons (Moody et al., 1996); moreover, our results show that tubb2b-dependent fluorescence is present in the vast majority of axonal projections from the olfactory organs to the OB in premetamorphic Xenopus.

Therefore, we observed that some of these axonal projections bypass the OB and terminate in higher brain regions. Most likely, these fibers belong to the EBOS, which has been described in many aquatic vertebrates (Demski and Northcutt, 1983; Schmidt et al., 1988; Von Bartheld and Meyer, 1988; Hofmann and Meyer, 1989; Pinelli et al., 2004). Studies conducted in goldfish (Carassius auratus) and salamander (Triturus alpestris and Salamandra salamandra) show that EBOS fibers target regions like the hypothalamus and are potentially involved in the perception of pheromones (Demski and Northcutt, 1983; Schmidt et al., 1988). In larval Xenopus laevis, it has been shown that a particular glomerulus, the so-called γ-glomerulus, is innervated by ipsilateral and contralateral axonal projections originating in the olfactory epithelium (Kludt et al., 2015). These fibers bypass the ipsilateral OB, cross the brain midline via the anterior commissure, and project toward the contralateral OB (Kludt et al., 2015). The γ-glomerulus is particular in the sense that it is not innervated by axons of odorant-sensitive receptor neurons, but instead by temperature-sensitive fibers (Kludt et al., 2015). In the present study, we also observed tubb2b-dependent fluorescence in fibers projecting through the anterior commissure. Consequently, some of the tubb2b-dependent fibers detected in the OB possibly belong to the previously described temperature-sensitive system.

A large number of cells in the mitral cell layer of the OB show tubb2b-dependent fluorescence. In the vertebrate OB, the mitral cell layer and the external plexiform layer are populated with cell bodies of projection neurons, the so-called mitral and tufted cells (Nagayama et al., 2014; Kosaka and Kosaka, 2016). The somata of the various interneurons of the OB, for example, juxtaglomerular cells and granule cells, are, respectively, located in the glomerular layer and the granule cell layer (Nagayama et al., 2014; Manzini et al., 2022). In larval Xenopus laevis, like in other amphibians (Hoffman, 1963), the mitral cell layer and the external plexiform layer are not clearly disjunct, and thus, an unambiguous assignment of the different bulbar neuron types to the various layers of the OB is not easy (Scalia et al., 1991; Nezlin and Schild, 2000). To overcome this limitation, in the present study, we employed TH and CR immunostainings in the OB of larval NBT-Katushka γ-cry-Venus Xenopus to get information about the cell type of cells featuring tubb2b-dependent fluorescence. In the OB of larval Xenopus, TH antibodies have been shown to stain dopaminergic interneurons in the glomerular and external plexiform layers (González et al., 1994; Boyd and Delaney, 2002), and antibodies against CR have been shown to stain a subset of mitral cells in the mitral cell layer and interneurons (granule cells) in the granule cell layer (Porteros et al., 1997; Castro et al., 2008; Morona and González, 2013). In the present study, we did not find a single TH-positive cell that also shows tubb2b-dependent fluorescence. This shows that cells exhibiting tubb2b-dependent fluorescence are not dopaminergic TH-immunoreactive interneurons. Similarly, we did not detect a single CR-positive cell featuring tubb2b-dependent fluorescence. This rules out that granule cells in larval Xenopus express tubb2b-dependent fluorescence.

Another study conducted in Xenopus identified that Pax6 is expressed in cells of the periglomerular and internal granule cell layers (Bandín et al., 2014). Within the periglomerular layer, Pax6 is present in a subset of dopaminergic interneurons that also express in the TH (Bandín et al., 2014). In this study, we also found that pax6 promoter-driven fluorescence occurs in cells throughout the glomerular layer, which suggests that these cells might also be dopaminergic neurons. More posteriorly, we found pax6-dependent fluorescence in the mitral cell layer of the OB and the accessory OB of larval Xenopus, but not in the granule cell layer. In the mitral cell layer, we also found tubb2b-dependent fluorescence and a subset of cells featuring both pax6- and tubb2b-dependent fluorescence. Due to the location of the tubb2b-positive cell bodies and the observation that TH- and CR-immunoreactive interneurons are tubb2b-negative, we assume tubb2b is exclusively expressed in projection neurons. Additionally, the presence of double-positive pax6 and tubb2b cells could be an indication that pax6 is located in a subset of projection neurons as well. If this is the case, it remains the question of what distinguishes the tubb2b from the pax6 projection neuron population and what role double-positive tubb2b and pax6 cells play in the OB.

In Xenopus laevis, very little is known about the terminal processing centers of the olfactory system. However, potential target regions of OB neurons, like the Acc and MeA, have been reported in adult Xenopus laevis (Marín et al., 1997a,b; Moreno and González, 2003; Moreno et al., 2005). Our study shows that cells expressing tubb2b are present in all potential target regions of OB neurons. In the ventral telencephalon, we found tubb2b-dependent fluorescence in cell bodies of the Acc, bed nucleus of the stria terminalis (BST), the MeA, and axons of the OpN. The Acc is a main component of the basal ganglia (Marín et al., 1997a,b). Retrograde tracing studies in adult Xenopus have shown that the Acc receives input from neurons of the OB, and thus, it might be involved in olfactory information processing (Marín et al., 1997a,b). The BST, which is part of the extended amygdala (Alheid et al., 1995; Moreno et al., 2012), and the MeA are also considered part of the olfactory cortex (Swanson and Petrovich, 1998; Moreno and González, 2003; Moreno et al., 2005). Both might have a similar function as the teleost habenula, which is involved in the control of fear responses (Agetsuma et al., 2010; Lee et al., 2010; Miyasaka et al., 2014). Additionally, we also found tubb2b-dependent fluorescence in the CeA, which is the main component of the amygdaloid complex for integration and control of cardiovascular, respiratory, and gastrointestinal functions (Jolkkonen and Pitkänen, 1998; Moreno and González, 2005). Posteriorly in the telencephalon, we have detected tubb2b-dependent fluorescence in cells of the pallium, including the LA, which is part of the Vp (Moreno and González, 2004). In teleost fish, the dorsal part of the pallium is homolog to the olfactory cortex of mammals (Wullimann and Mueller, 2004; Miyasaka et al., 2014), and it was shown that in adult Xenopus laevis, the pallium is a target of neurons originating from the OB (Moreno et al., 2005). Similar to the MeA, the LA is also a terminal processing center of the olfactory system (Moreno and González, 2004; Moreno et al., 2005). Due to the wiring pattern of the LA, it is discussed to be a homolog of the mammalian basolateral complex (Moreno et al., 2005).

In addition to higher olfactory processing centers, we found that other main forebrain regions also show tubb2b-dependent fluorescence. In the telencephalon, tubb2b-positive cells are located in the Str, DP, POA, and Hyp. The DP and posterior tuberculum, together with the Str, are considered a homolog of the mammalian nigrostriatal pathway and to regulate different motor functions of the frog (Hoke et al., 2007). The Hyp serves as a control unit for endocrine mechanisms and regulates general functions of the body and behavioral patterns (Bruce, 2009; Domínguez et al., 2013, 2014). In the diencephalon, we observed tubb2b-dependent fluorescence in cells of p1–p3. A small number of tubb2b-positive cells were visible in the Th of p2. Although the thalamus serves as a relay station for sensory information, it does not process olfactory input (Jones, 2007; Bandín et al., 2015). Although tubb2b-dependent fluorescence has been detected in several regions that might be involved in olfactory information processing, we observed that it is clearly not restricted to it. We also found expression in fibers of the OpN, cell bodies in the area of the BST, the S, the POA, and Hyp.

In the past years, the promoter of class II β-tubulin (tubb2b) has been used as a tool to create transgenic Xenopus lines, like the tubb2b-GCaMP6s Rno.elas-GFP (NXR; Cat# 0.0107; Horb et al., 2019), to drive the expression of the genetically encoded calcium indicator GCaMP6s in the nervous system (Offner et al., 2020). Measurements of neuronal activity have been already possible in the OB and the olfactory amygdala (Offner et al., 2020) and could be performed in all regions (cells) featuring tubb2b-dependent fluorescence shown in the present study. Not only the tubb2b GCaMP6s transgenic line is a great advantage for functional studies but also our custom-bred NBT-Katushka pax6-GFP line is a great tool to mark and characterize a broad field of brain regions in larval Xenopus.