Eur J Neurosci
May 1, 2015;
Microtubule-associated protein tau promotes neuronal class II β-tubulin microtubule formation and axon elongation in embryonic Xenopus laevis.
Compared with its roles in neurodegeneration, much less is known about microtubule
-associated protein tau''s normal functions in vivo, especially during development. The external development and ease of manipulating gene expression of Xenopus laevis embryos make them especially useful for studying gene function during early development. To study tau''s functions in axon
outgrowth, we characterized the most prominent tau
isoforms of Xenopus embryos and manipulated their expression. None of these four isoforms were strictly analogous to those commonly studied in mammals, as all constitutively contained exon 10, which is preferentially removed from mammalian fetal tau
isoforms, as well as exon 8, which in mammals is rare. Nonetheless, like mammalian tau
, Xenopus tau
exhibited alternative splicing of exon 4a, which in mammals distinguishes ''big'' tau
of peripheral neurons, and exon 6. Strongly suppressing tau
expression with antisense morpholino oligonucleotides only modestly compromised peripheral nerve
outgrowth of intact tadpoles, but severely disrupted neuronal microtubules containing class II β-tubulins while leaving other microtubules largely unperturbed. Thus, the relatively mild dependence of axon
development on tau
likely resulted from having only a single class of microtubules disrupted by its loss. Also, consistent with its greater expression in long peripheral axons, boosting expression of ''big'' tau
increased neurite outgrowth significantly and enhanced tubulin acetylation more so than did the smaller isoform. These data demonstrate the utility of Xenopus as a tool to gain new insights into tau''s functions in vivo.
Eur J Neurosci
Tubulin I+II Ab1
[+] show captions
Fig. 1. Xenopus expressed tau isoforms during development. Total RNAs
from stage (st.) 20, 29/30, 37/38, juvenile frog brain, spinal cord or st. 56
tadpole tail were reverse transcribed and amplified by PCR. PCR products
were then separated by agarose gel electrophoresis and stained with ethidium
bromide. The contrast has been reversed to improve visibility of the bands.
Arrowheads point to the four major bands that mark the earliest forms of tau
expressed at the beginning of axon development (st. 20; 1.2, 1.4, 1.9 and
2 kb). These bands were excised and cloned for sequencing. During development,
large isoforms (1.9–2 kb) became the most prominent ones in the tail
of metamorphosing tadpoles, which is enriched with peripheral neurons. This
expression and close resemblance to human peripheral nervous system tau
(see text) are consistent with their identification as Xenopus big taus. The
smaller isoform, at 1.2 kb, became most prominent in adult CNS.
Fig. 2. Comparison of Xenopus laevis and human mapt genes. (A) To identify the exons, the nucleotide and the amino acid sequences of the four tau PCR
products from Fig. 1, as well as those of two previously characterized X. laevis mapt cDNAs (XTP-1 and XTP-2) cloned from late-stage tadpole tail (Olesen
et al., 2002), were aligned against nucleotides of the X. tropicalis genomic sequence and against the amino acid sequences of nine human tau isoforms (see
Supporting Information), respectively. The six Xenopus isoforms were generated by alternative splicing of four exons (exons 4, 4a, 6 and 12a) as indicated.
Exons that were constitutively present in all four isoforms are indicated in black. As a result of the insertion of exon 12a (124 nt), which has no human parallel,
into XTP-2, a frameshift is introduced and a premature stop codon (*) is encountered after the first amino acid of exon 13. (B) The degree of amino acid identity
[Identities (%)] as well as the number of amino acids (aa#) encoded for each exon of X. laevis and human tau.
Fig. 3. Tau protein expression began at early stages of axon outgrowth and increased with neuronal development through adulthood. (A–C) Parasagittal views
of st. 22, 29/30 and 37/38 embryos processed for tau immunofluorescence in whole-mount and imaged by confocal laser-scanning microscopy. Images were
taken with the 109 (NA 0.5) objective and represent maximal intensity projections created from stacks of 26, 17 and 20 optical sections, respectively, with
voxel sizes of 1.8 9 1.8 9 8.0 lm (A and B) and 2.5 9 2.5 9 7.8 lm (C). At st. 22 (A), representing early axonal outgrowth, tau immunoreactivity was most
intense in axon tracts within the hindbrain (arrow) and ventral spinal cord (S.C.), and was also visible in peripheral motor axons projecting to anterior somites.
At st. 29/30 (B), representing a stage of maximal axonal outgrowth, tau immunoreactivity became more intense throughout the brain and spinal cord, and could
be visualized in axons of the trigeminal nerve (Vmd, mandibular branch of the trigeminal nerve). By st. 37/38 (hatchling tadpole; C), tau immunoreactivity was
abundant in axon tracts throughout the brain and spinal cord, in the retina (eye), and cranial nerves (X, vagus; IX, glossopharyngeal; VII, facial; Vmd). The
image in (C) was cropped when rotated. (D) Tau immunoreactivity in neuronal perikarya and axons within a transverse section through the ventral diencephalon
of a juvenile frog. The bracket indicates the optic tract. The image is a single optical confocal laser-scanning microscopy section taken with the 209 objective
(NA 0.75), with a pixel size of 0.45 9 0.45 lm. (E and F) Tau immunoreactivity within transverse sections through the retina of a juvenile frog at low (E) and
high (F) magnifications. Images are single confocal laser-scanning microscopy optical sections taken with the 209 objective (NA 0.75) with the scan zoom at
0.7 and 2.0, and pixel sizes of 0.60 9 0.60 lm and 0.45 9 0.45 lm for (E) and (F), respectively. Abbreviations: GCL, ganglion cell layer; INL, inner nuclear
layer; IPL, inner plexiform layer; ONL, outer nuclear layer; OPL, outer plexiform layer; PL, photoreceptor layer. Scale bars: 100 lm (A–E); 20 lm (F).
Fig. 4. In intact embryos, tau knockdown reduced peripheral nerve development. Embryos were injected at the two-cell stage with tau antisense morpholino
oligonucleotide (MO) and immunostained at st. 37/38 in whole-mount for tau, to visualize effects of the MO on tau expression (A), and for peripherin (B1–2,
C1–2), to visualize effects on peripheral nerve development. (A) View in the horizontal plane illustrating normal tau immunostaining within the brain (arrowhead
1), cranial peripheral nerves (e.g. arrowhead 2) and spinal cord (arrowhead 3) on the uninjected side of the embryo (left of the dotted line) and reduced
tau immunostaining on the side descended from the blastomere injected with MO (right of the dotted line). Rostral is toward the top. The image represents a
maximal intensity projection stack of 33 confocal laser-scanning microscopy sections (objective: 109; NA 0.5) with a voxel size of 2.5 9 2.5 9 6.4 lm. (B1
and 2) Parasagittal views through trunk level somites illustrating diminished robustness of peripheral motor nerve development (arrowheads) on the injected
(B2) vs. uninjected (B1) side. Images represent a single optical section (109; NA 0.5) taken at comparable depths through the somites; voxel size is
1.8 9 1.8 9 5.1 lm. (C1 and 2) Parasagittal views through the head illustrating diminished robustness of cranial nerve development on the injected (C2) vs.
uninjected (C1) side. Rostral is toward the right. Images represent a maximal intensity projection stack of 15 confocal laser-scanning microscopy sections (109;
NA 0.5) with a voxel size of 1.6 9 1.6 9 3.1 lm. Scale bars: 100 lm for all panels; (B1 and 2) are at the same magnification, as are (C1 and 2). (D) Average
lengths ( SEM) of cranial and spinal peripheral nerves of tau and control MO-injected animals immunostained for peripherin at st. 37/38 (N = 3). **Significantly
different from controls, as determined by t-test, P < 0.01. Abbreviations: II, optic nerve; IX, glossopharyngeal nerve; Vmd, mandibular branch of the
trigeminal nerve; Vop, ophthalmic branch of the trigeminal nerve; VII, facial nerve; SN, spinal motor nerves; X, vagus nerve.
Fig. 5. Tau knockdown selectively disrupted the structural integrity of microtubules containing a neuronal class II b-tubulin. Newly differentiating, embryonic
cultured spinal cord neurons containing tau antisense morpholino oligonucleotide (tau MO; B1–3, D2, E2, F2, G2) or control MO (A1–2, D1, E1, F1, G1) were
stained for the neuron-specific b-tubulin II isotype (Nb-tubulin; Xenopus gene symbol, tubb2b), ubiquitous a- and b-tubulins, tyrosinated (Tyr-), and acetylated
(Ac-) a-tubulins, respectively. (A1–B3) In contrast to the generally filamentous structures seen in the axons and growth cones (A2) of control MO neurons
stained for Nb-tubulin, significantly more short, discontinuous fragments (arrowheads in B1 and 2) were seen in the axons and growth cones (B3) of tau MO
neurons. (C) Quantitation of the number ( SEM) of discontinuous, fragmented microtubules (puncta) stained by the Nb-tubulin antibody along the neuritic
shaft (left; /100 lm) and within growth cones (right; /100 lm2) of control and tau MO cultured neurons (three cultures each). **Significantly different from
control, as determined by t-test; P < 0.01. (D1–G2) As determined by immunostaining, neither the levels of expression nor distributions of ubiquitous b-tubulins
(D1 and 2), ubiquitous a-tubulins, Tyr-tubulin (F1 and 2) and Ac-tubulin (G1 and 2) differed between tau MO and control MO cultured neurons. Scale bars
(20 lm in all panels) in (A2), (B1), (E1), (F1) and (G1) apply to (B3), (B2), (E2), (F2) and (G2), respectively. Neurons were imaged by conventional epifluorescence
with a 1009 objective (NA 1.4). (H) Quantitation of the relative fluorescence intensity (RFI; see Statistical analyses in Materials and methods for
details) within neurites for each of the tubulin antibodies in (D1–G2) revealed no significant differences in tau MO neurons relative to that of control MO neurons
(RFI = 1.0; t-test, three cultures each).
Fig. 6. Big tau overexpression selectively enhanced neurite outgrowth and expression of acetylated a-tubulin in cultured neurons. Cultured spinal cord neurons
from embryos that were injected with 5 ng green fluorescent protein (GFP) RNA (left column; A1–F1), tau-402 RNA (middle column; A2–F2) or tau-688
RNA (right column; A3–F3) were immunostained for multiple tubulin isoforms (abbreviations as in Fig. 5) and for C-terminal tail domain-phosphorylated NFM
(pNF-M), as indicated. Whereas the intensities and distributions of immunostaining for the neuronal b-tubulin II isotype (A1–3), pan-specific a- and b-tubulins
(C1–3 and B1–3, respectively), tyrosinated a-tubulin (D1–3) and phosphorylated NF-M (F1–3) were indistinguishable among the three groups, acetylated
a-tubulin (E1–3) immunostaining was more intense in the axon shaft of both tau-688 and tau-402 overexpressing neurons, compared with those expressing
GFP. Arrowheads in (F1–3) point to immunofluorescence in neuronal perikarya representing cross-reactivity with a nuclear antigen unrelated to neurofilaments
(Wood et al., 1985; Szaro & Gainer, 1988). Scale bars (20 lm in all panels) in (A1), (A2), (A3) and (F2) apply to (B1–E1), (B2–E2), (B3–E3) and (F1 and
F3); respectively. (G–I) Morphometric analyses for average ( SEM) total length of the neuritic arbor (G), length of the longest primary neurite (H), and the
number of branches (I) per neuron as a function of the dose of RNA injected into the embryonic blastomere. Each bar represents the mean value over three separate
cultures ( SEM) of the average across the neurons in each culture. Total numbers of neurons in the three cultures were: 5 ng GFP RNA (132); 2 and
5 ng tau-688 RNA (123 and 126, respectively); 2 and 5 ng tau-402 RNA (138 and 141, respectively). The bar across the top of each graph indicates the level
of statistical significance of the one-way ANOVA conducted across all groups. Brackets indicate the level of significance for differences between individual
groups, as determined by a Tukey post hoc test: * and **, P < 0.05 and P < 0.01, respectively. (J–M) Quantitation of the relative fluorescence intensity (RFI)
within the neurite ( SEM; three cultures each) for cells injected with 5 ng GFP (RFI = 1.0), tau-402 or tau-688, and then immunostained for ubiquitous btubulin
(J), ubiquitous a-tubulin (K), Tyr-tubulin (L) and Ac-tubulin (M). Whereas no significant differences were observed for the first three antibodies (J–L;
ANOVA), staining for Ac-tubulin (M) was significantly greater in tau-688 neurons than for both tau-402 and GFP neurons (**P < 0.01; Bonferroni post hoc
t-test), and it was also greater for tau-402 neurons than for GFP neurons (*P < 0.05).