Evolution of the spatial distribution of MAP1B phosphorylation sites in vertebrate neurons

Sequence alignments

Two databases were searched to obtain MAP1B sequences: The National Centre for Biotechnology Information (http://www.ncbi.nlm.nih.gov/) and Ensembl (http://www.ensembl.org/index.html). Rat MAP1B was used as a query sequence and databases were searched using Basic Local Alignment Search Tool (blast). Sequences were aligned using the ClustalW alignment programme (Thompson et al. 1994).

Tissue preparation for immunoblotting

Embryos and hatchlings from the following species were used: stage 31 dogfish (Scyliorhinus cuniculus), 5-day-old zebrafish (Danio rerio), stage 42 frog (Xenopus laevis), 17-day-old snake (Elaphe guttata) and embryonic day 4 chick (Gallus gallus). Animals were anaesthetized by immersion in an aqueous solution (0.1% w/v) of tricaine methanesulphonate and their brains and spinal cords removed. Early postnatal rat (Rattus norvegicus) whole brain homogenate was used as a positive control. Neural tissue was homogenized using a tight-fitting glass–glass homogenizer in buffer (100 mm 2-(N-morpholino)ethanesulfonic acid, pH 6.5, 1 mm EGTA, 0.5 mm MgCl₂, 1 mm 1,4-dithiothreitol, 5 mm NaF) at 4 °C. An equal volume of × 2 sodium dodecyl sulphate gel electrophoresis buffer was then added, the sample boiled and stored at −20 °C.

Immunoblotting

Proteins were separated by sodium dodecyl sulphate-polyacrylamide gel electrophoresis using 8% gels. Proteins were then transferred to nitrocellulose membranes by wet transfer. Non-specific binding was blocked by the addition of blocking buffer (BB) [4% non-fat milk solids (Marvell) and 0.1% Tween in phosphate-buffered saline (PBS-T)] for 1 h at room temperature (25 °C). Blots were then incubated with a primary antibody in BB at 4 °C overnight. The primary antibodies were: mouse monoclonal antibody (mAb) 125 (1 : 1000; Ulloa et al. 1993), rabbit polyclonal antibody (pAb) pS824 (1 : 1000; Lin et al. manuscript in preparation), pAb BUGS (1 : 1000; Trivedi et al. 2005), pAb SuperBUGS (1 : 1000; Trivedi et al. 2005), pAb pS1388 (1 : 1000; Scales et al. 2009) and pAb pS1392 (1 : 1000; Scales et al. 2009). A mAb against acetylated α-tubulin (mAb 6-11B-1) was used as a loading control (1 : 1000; Sigma). The next day, the primary antibodies were removed and blots were washed twice in 0.1% PBS-T. Peroxidase-conjugated secondary antibodies (Sigma; 1 : 1000) were added in BB and left to incubate for 1 h at room temperature. Blots were then washed several times with PBS-T and developed with Pierce supersignal West Pico Chemiluminescent substrate and exposed to Clear Blue X-ray films (Fujifilm).

DNA sequencing

RNA was extracted from the same snake, zebrafish and dogfish material that was used for immunoblotting. The brain and spinal cord were dissected out and placed into Trizol (Invitrogen) and RNA was isolated following the manufacturer’s instructions. The cDNA was synthesized from RNA using the Superscript III 1st strand RT-PCR kit (Invitrogen). MAP1B was sequenced from each species using the polymerase chain reaction (PCR). Primers were designed to produce products that covered the majority of the MAP1B sequence.

Immunocytochemistry

Embryos and hatchlings were anaesthetized by immersion in an aqueous solution (0.1% w/v) of tricaine methanesulphonate and fixed by immersion in 3% formaldehyde and 0.2% glutaraldehyde in phosphate-buffered saline (PBS) for 3 h at room temperature. Tissues were embedded in a gelatine (5 mg mL⁻¹) and albumin (0.3 g mL⁻¹) solution polymerized with glutaraldehyde (2.5% v/v). Transverse sections of embryos and hatchlings were cut at 50 μm with a vibroslice (Camden Instruments) and placed into PBS. Endogenous peroxidase was neutralized by adding a 4 : 1 methanol and 0.3% hydrogen peroxide solution and incubating at room temperature on a rocker for 20 min. The sections were then washed with PBS and left overnight at 4 °C in BB [5% (v/v) normal horse serum, 5% (v/v) normal goat serum, 50 mm l-lysine and 0.2% Triton X-100 in PBS]. Immunolabelling was performed by incubating sections overnight with pAb SuperBUGS (1 : 1000–1 : 1500) or pAb pS1388 (1 : 400–1 : 700) diluted in BB. The next day, primary antibodies were removed and sections washed in PBS. Sections were incubated with peroxidase-conjugated goat anti-rabbit (DAKO) secondary antibody diluted in BB and left for 2 h. Sections were washed twice in PBS followed by Tris-buffered saline. Immunolabelling was visualized by adding diaminobenzidine at 0.5 mg mL⁻¹ in Tris-buffered saline containing 0.01% hydrogen peroxide. Sections were washed in PBS and allowed to air dry on 1% gelatine-coated microscope slides. They were then dehydrated in 70% ethanol in water and then in 100% ethanol for 15 min, cleared in xylene and mounted with distyrene, plasticizer, xylene. Sections were viewed with an Olympus BH2 microscope equipped with Nomarski optics and images captured with a DP70 colour CCD camera (Olympus). Images, as Tiff files, were manipulated in Adobe Photoshop.

Culture and immunofluorescence microscopy of cerebral cortical neurons

Cerebral cortices were dissected from day 17–18 rat embryos in Hank’s balanced salt solution and dissociated using papain (Worthington) followed by trituration with a pipette. For biochemical analysis, cells (8 × 10⁵) were plated in six-well plates (Nunc) coated with poly-d-lysine (100 μg mL⁻¹; Sigma) and laminin (20 μg mL⁻¹; Sigma), whereas for immunofluorescence, cells (1 × 10⁴) were plated on 13-mm glass coverslips that were similarly coated. Cultures were incubated at 37 °C in 5% CO₂ in humidified air in Neurobasal medium supplemented with 2% (v/v) B27 supplement, 2 mm glutamine/100 IU mL⁻¹ penicillin/100 IU mL⁻¹ streptomycin and 0.45%d-(+) glucose.

Cortical cultures to be used for immunofluorescence were fixed either with 3% formaldehyde and 0.2% glutaraldehyde in PBS at 37 °C for 10 min or with the same fixative containing 0.2% Triton X-100 in PIPES, HEPES, EGTA, MgSO₄ buffer (Schliwa & van Blerkom, 1981) and 0.33 μm phalloidin (Molecular Probes) at 37 °C. Fixed cultures were treated for 1 h with BB [5% (v/v) normal horse serum, 5% (v/v) normal goat serum, 50 mm L-lysine and 0.2% (v/v) Triton X-100 in PBS]. Following blocking, cells were incubated for 1 h with the following antibodies diluted in BB: pAb N-19 (Santa Cruz, 1 : 50), a phosphorylation-independent antibody that recognizes all forms of MAP1B, and mAb YL 1/2 (Sera Lab; 1 : 50), which recognizes tyrosinated α-tubulin. The appropriate Alexa-conjugated secondary antibodies or Alexa-conjugated phalloidin (Molecular Probes) were diluted in BB and applied for 1 h. Finally, cultures were washed three times with PBS, mounted onto microscope slides in FluorSave (Calbiochem) and viewed under a Leica TCS confocal microscope equipped with Argon, Krypton and HeNe lasers. Cells were imaged with × 63 or × 100 PLANAPO oil-immersion objectives and Tiff images captured (1024 × 1024) using Leica TCS software. Images, as Tiff files, were manipulated in Adobe Photoshop.

Culture and immunofluorescence microscopy of frog nodose ganglion neurons

Frog (X. laevis) nodose ganglia from hatchlings were incubated in type III collagenase (0.125%, Worthington) for 3 h at 28 °C, followed by trituration in Rosewell Park Memorial Institute (RPMI) medium (Gibco), diluted to 80% with water and containing 10% bovine serum. Dissociated neurons were plated onto 13-mm glass coverslips coated with poly-d-lysine (0.1%; Sigma) and 10 μg mL⁻¹ laminin–nidogen complex (a gift from T. Sasaki, Max-Planck-Institut für Biochemie, Martinsried, Germany). Cultures were incubated overnight at 28 °C in 5% CO₂ in humidified air in 80% RPMI medium.

Cultures were fixed after 1 day in vitro with 3% formaldehyde and 0.2% glutaraldehyde in PBS at 37 °C for 10 min and were labelled for immunofluorescence as for cortical cultures (see above) except that mAb AA6 (Sigma; 1 : 50) was used to label all forms of MAP1B.

Quantitative analysis of MAP1B axonal gradients

To quantify the relative changes in the amount of phosphorylated MAP1B along the axon, cortical cultures were double-labelled with pAb N-19 (1 : 50) and pAb SuperBUGS (1 : 100), and fluorescence images were obtained by confocal microscopy. The axon was divided into three contiguous segments of equal length from the cell body to the growth cone and the mean fluorescence intensity in each channel determined using ImageJ software (http://rsb.info.nih.gov/ij/index.html). Ten axons were measured for each condition (fixed conventionally or with fixatives containing detergents) from three independent experiments. An unpaired Student’s t-test was used for statistical analysis.

Conservation of glycogen synthase kinase 3β phosphorylation sites on MAP1B in vertebrates

In previous work, we identified 29 phosphorylation sites on juvenile rat (R. norvegicus) MAP1B by a combination of mass spectrometry, site-directed mutagenesis, in-vitro kinase assays with recombinant MAP1B proteins and phospho-specific antibodies (Trivedi et al. 2005; Scales et al. 2009). Two of these sites [serine 1260 (S1260) and threonine 1265 (T1265)] were shown to be non-primed GSK-3β sites (Trivedi et al. 2005) and a third site (S1388) was found to be a GSK-3β site primed by phosphorylation of S1392 by dual-specificity tyrosine phosphorylation-regulated kinase 1A (Scales et al. 2009). We compiled a list of MAP1B sequences across the vertebrates, including representatives from mammals, birds, reptiles, amphibians and fish, by searching sequence databases and by sequencing MAP1B DNA produced by reverse transciption-PCR from dogfish (S. caniculus), snake (E. guttata) and zebrafish (D. rerio) using degenerate primers based on zebrafish, Takifugu rubripes, Tetraodon nigroviridis and Gasterosteus aculeatus MAP1B for dogfish and degenerate primers based on lizard (Anolis carolinensis) MAP1B for snake. We aligned these MAP1B sequences in the region of the identified primed and non-primed GSK-3β sites (Table 1). This revealed that both non-primed GSK-3β sites (S1256 and T1261 in rat MAP1B) are highly conserved. The S1256 site is present in all classes of vertebrates including teleost and elasmobranch fish but is absent from lamprey (P. marinus) and, curiously, from zebrafish MAP1B (Table 1A). The T1261 site is even more highly conserved, although it is a serine in teleost fish (Table 1A). Furthermore, the sequence between these two non-primed sites and the flanking sequence upstream and downstream of the two sites is also highly conserved, particularly the proline immediately downstream of the phospho-acceptor amino acid residue, which is a feature of GSK-3β and other so-called proline-directed serine/threonine kinases (Doble & Woodgett, 2003). The conservation of the flanking sequence begins to break down in the fish, with the notable exception of dogfish. In contrast, although the primed GSK-3β site (S1388 in rat MAP1B) is also highly conserved, being absent only in the lamprey (not shown), the presence of the priming phosphorylation site (S1392 in rat MAP1B) is patchy in fish MAP1B, being absent from the puffer fish (T. rubripes), and several other species of teleost fish, and from lamprey (not shown) (Table 1B). Thus, in the fish, although the primed GSK-3β site is present, its priming site is frequently not present. Again, the notable exception is the dogfish.

Table 1.

High sequence conservation of MAP1B phosphorylation sites and flanking regions among vertebrate classes

Open in a new tab

Phosphorylation of MAP1B sites in vivo

Sequence analysis does not reveal whether phosphorylation of these sites occurs in vivo. To examine this question we utilized our panel of affinity-purified phospho-specific antibodies to phosphorylation sites on MAP1B (Trivedi et al. 2005; Scales et al. 2009). We collected samples of developing neural tissue from the central nervous system of dogfish, zebrafish, frog (X. laevis), snake and chick (G. gallus), and probed protein immunoblots with our panel of phospho-specific antibodies. We used rat postnatal brain homogenate as an internal marker for MAP1B. We found that phospho-specific pAbs that recognize non-primed GSK-3β sites (pS1256 and pT1261 on rat MAP1B, pAb BUGS and pAb SuperBUGS, respectively) (Trivedi et al. 2005) cross-react with proteins in developing dogfish, frog, snake and chick neural tissue that migrate to a similar position to rat MAP1B in immunoblots (Fig. 1A). Similarly, we found that a phospho-specific pAb (pAb pS1388) that recognizes a primed GSK-3β site (pS1388 on rat MAP1B) and a phospho-specific pAb (pAb pS1392) that recognizes the priming site (pS1392 on rat MAP1B) (Scales et al. 2009) cross-react with proteins in developing dogfish, frog, snake and chick neural tissue that migrate to a similar position to rat MAP1B in immunoblots (Fig. 1B). None of these antibodies cross-react with zebrafish neural tissue (Fig. 1). The failure of pAb BUGS and pAb pS1392 to recognize MAP1B from zebrafish is expected because a phosphorylatable amino acid is absent at these sites, whereas the failure of pAb SuperBUGS to recognize MAP1B from zebrafish, despite the presence of phospho-acceptor residues at the appropriate sites, is probably due to the divergent flanking sequence at these sites (Fig. 1A,B). Although zebrafish MAP1B contains the primed pS1388 site it lacks the priming site and therefore would not be expected to react with pAb pS1388.

To test the idea that the flanking sequence influences phospho-specific antibody binding, we looked for a MAP1B phosphorylation site that has a conserved flanking sequence in zebrafish. One such site is S824 in rat MAP1B (Fig. 1C). This site is the most highly conserved vertebrate MAP1B phosphorylation site, being present in all vertebrate sequences that we have assembled and, furthermore, the upstream and, particularly, the downstream flanking sequences are also highly conserved (Fig. 1C). We have two phospho-specific antibodies that recognize pS824 in rat MAP1B: mAb 125 (Ulloa et al. 1993) and pAb pS824 (Lin et al. manuscript in preparation). Both of these antibodies recognize proteins of the appropriate molecular weight in neural tissue samples from dogfish, frog, snake and chick by immunoblotting (Fig. 1C). Furthermore, mAb 125 cross-reacted with a protein in zebrafish neural tissue that migrated to a slightly lower position than rat MAP1B in immunoblots (Fig. 1C). Surprisingly, however, given the reactivity with mAb 125, pAb pS824 did not react with zebrafish MAP1B. Tomasiewicz & Wood (1999) detected MAP1B in immunoblots of zebrafish and reported that the molecular weight was 82 kDa. However, this is probably not MAP1B. We calculated the molecular weight of this protein from immunoblots as 250 kDa, which is similar to the molecular weight estimated from zebrafish MAP1B DNA (222 kDa).

Immunolocalization of phosphorylated MAP1B in developing spinal cord

Previous work by us has shown that MAP1B phosphorylated at GSK-3β non-primed sites is uniquely localized in growing axons of differentiating neurons where it is often present in a concentration gradient that is highest distally, i.e. towards the growth cone (Trivedi et al. 2005; Scales et al. 2009). In contrast, MAP1B phosphorylated at GSK-3β primed sites is distributed throughout the length of growing axons and is also found in the somato-dendritic compartment of differentiating neurons (Scales et al. 2009). To investigate whether this difference in localization of the two forms of phosphorylated MAP1B is conserved throughout the vertebrates we immunolabelled transverse sections of developing spinal cord from representative vertebrate groups with pAb SuperBUGS, an antibody against pT1261 in rat (Trivedi et al. 2005), a non-primed GSK-3β site, and pAb pS1388 (Scales et al. 2009), an antibody against pS1388, a GSK-3β primed site (Fig. 2). We chose to investigate the expression of MAP1B phosphorylated isoforms in developing spinal cord because neuronal development is well documented in this tissue and the trajectories of growing axons from a small number of identified neurons are readily distinguishable. In embryonic rat spinal cord, pAb SuperBUGS stained growing axons in a gradient that was highest towards the growth cone as previously described (Fig. 2A) (Trivedi et al. 2005; Scales et al. 2009). For example, the axons of dorsal root ganglion cells were only stained at the poles of the dorsal root ganglion and, similarly, the axons of primary motor neurons were only stained after their exit from the spinal cord in the ventral roots and not within the ventral horn (Fig. 2A). Neuronal cell bodies were unstained. This distribution pattern of non-primed GSK-3β phosphorylated MAP1B seen in rat was strikingly conserved in the chick embryonic spinal cord (Fig. 2A). In the snake, frog (not shown) and dogfish spinal cord the staining was much weaker and, although it was clear that only axons were stained with pAb SuperBUGS, it was not clear whether the axons were stained in a gradient (Fig. 2A).

In contrast to the distribution of non-primed GSK-3β phosphorylated MAP1B, MAP1B phosphorylated at primed GSK-3β sites, as revealed by labelling with pAb pS1388, was uniformly distributed throughout the length of growing axons and also present in neuronal cell bodies in embryonic rat spinal cord, as previously described (Fig. 2B) (Scales et al. 2009). For example, axons in the dorsal root ganglia were labelled throughout their course at both the ganglion poles and the equator and primary motor neuron axons were labelled both within the ventral horns and throughout the ventral roots. This distribution pattern of primed GSK-3β sites was conserved in the chick, frog and dogfish developing spinal cord with labelling of axons along their entire length and neuronal cell body labelling, for example of motor neurons in the ventral horn (Fig. 2B). In the snake (not shown), frog and dogfish spinal cord, the staining of axons with pAb pS1388 was stronger than that seen with pAb SuperBUGS, perhaps because of a closer homology of the flanking sequences with the rat (Table 1B), and there was no evidence of an axonal labelling gradient (Fig. 2B). It was clear that neuronal cell bodies were also labelled by pAb pS1388, particularly in frog and dogfish (Fig. 2B).

The sub-cellular distribution of glycogen synthase kinase 3β phosphorylated MAP1B primed and non-primed sites is preserved in cultured embryonic neurons

To determine whether the striking distribution patterns of primed and non-primed GSK-3β phosphorylated MAP1B sites seen in spinal neurons in vivo was preserved in cultured neurons, we labelled cultures of embryonic cortical neurons with pAb pS1388 and pAb SuperBUGS. Embryonic cortical neurons were cultured for 2 days on laminin, by which time the majority of neurons had extended one long process, destined to become an axon, and several minor processes that will become dendrites. pAb SuperBUGS only labelled long processes in these cultures, whereas pAb pS1388 labelled the entire neuron including the nascent axon and the minor processes and cell bodies (somato-dendritic compartment; Fig. 3). There was a clear gradient in the intensity of axonal labelling by pAb SuperBUGS with the highest labelling towards the growth cone (Fig. 3A). In contrast, pAb pS1388 axonal labelling was uniform along the axon length (Fig. 3B).

The sub-cellular distribution of glycogen synthase kinase 3β phosphorylated MAP1B primed and non-primed sites in cultured neurons is evolutionarily preserved

Our analysis of GSK-3β phosphorylated MAP1B primed and non-primed sites in spinal cord sections from representative vertebrates (Fig. 2) showed that the distribution of GSK-3β phosphorylated MAP1B primed and non-primed sites in lower vertebrates (snake, frog and dogfish) was broadly similar to that seen in higher vertebrates (rat and chick). However, we could not determine whether GSK-3β phosphorylated MAP1B non-primed sites were distributed in a gradient in axons, as they are in higher vertebrates (Figs 2 and 3). To investigate this we cultured frog (X. laevis) neurons and labelled them with pAb SuperBUGS and pAb pS1388 (Fig. 4). This revealed a distribution pattern of the two sites that was similar to that seen in rat cortical neurons; GSK-3β phosphorylated MAP1B non-primed sites were present only in long neurites and in a gradient that was highest towards the growth cone (Fig. 4A), whereas GSK-3β phosphorylated MAP1B primed sites were present throughout the neuron including the somato-dendritic and axonal compartments (Fig. 4B).

The axonal gradient of polyclonal antibody SuperBUGS labelling is not due to an underlying gradient of MAP1B or to binding of a masking protein

MAP1B is more concentrated in the distal parts of growing axons and axons tend to have larger diameters distally (Black et al. 1994), so we determined whether the axonal gradient labelled by pAb SuperBUGS was independent of the increase in total MAP1B along the axon. Cortical cultures were double labelled with a phosphorylation-independent antibody that recognizes all forms of MAP1B (pAb N-19) and with pAb SuperBUGS (Fig. 5A). We quantified axonal labelling by ratioing the fluorescence intensity for total MAP1B and pAb SuperBUGS immunofluorescence using ImageJ software in three contiguous segments of the axon from the cell body to the growth cone. This showed that, although there was a proximo-distal increase in total MAP1B along the axon, it was far less than the proximo-distal increase in pAb SuperBUGS labelling as reflected by an increase in the fluorescence intensity ratio between the first and second axon segment (Fig. 5B). There was also an increase between the second and third segment but this did not reach statistical significance. Therefore, the axonal gradient of phosphorylated MAP1B is independent of the concentration of MAP1B (Fig. 5B).

Previously, we have shown that mAb 150, which recognizes an uncharacterized phospho-epitope on MAP1B, labels growing axons of dorsal root ganglion neurons in a gradient that is highest distally (Bush et al. 1996). However, after simultaneous detergent extraction and fixation of these neurons, a procedure that preserves the cytoskeleton and removes soluble proteins, mAb 150 labelled axons uniformly along their length, suggesting that a masking protein, removed by detergent extraction, blocked access of the antibody to MAP1B in the proximal region of axons (Bush et al. 1996). To test whether the axonal gradient seen with pAb SuperBUGS is due to a reverse gradient of a masking protein that blocks access of pAb SuperBUGS to MAP1B we simultaneously detergent extracted and fixed cortical cultures. This had no effect on the labelling of axons by pAb SuperBUGS; axonal gradients were still present and had a similar quantitative distribution along the axon in relation to total MAP1B as in conventionally fixed cultures (Fig. 5C). This suggests that the axonal gradient is due to a genuine gradient of phosphorylated MAP1B (Fig. 5B,C).

In growth cones, MAP1B binds to dynamic microtubules and is phosphorylated at both glycogen synthase kinase 3β primed and non-primed sites

It has been suggested that MAP1B is involved in maintaining a population of dynamic microtubules in growth cones that are important for axon growth and growth cone pathfinding (Gordon-Weeks & Fischer, 2000). Stable and dynamic microtubules in growth cones can be distinguished using specific antibodies to post-translationally modified forms of tubulin that are markers for stability. To examine the distribution of GSK-3β primed and non-primed phosphorylated MAP1B in growth cones in relation to the cytoskeleton, we fixed cortical cultures with fixatives containing detergent to preserve the cytoskeleton and remove soluble proteins (Bush et al. 1996) and then labelled them with either pAb SuperBUGS or pAb pS1388, phalloidin to reveal F-actin and with mAb YL 1/2, which labels dynamic microtubules. This showed that the two forms of phosphorylated MAP1B in growth cones, recognized by pAbs SuperBUGS and pS1388, exclusively co-localize with dynamic microtubules (Fig. 6A,B).

Conservation of glycogen synthase kinase 3β phosphorylation sites on MAP1B in the vertebrates

To gain insight into the function of GSK-3β phosphorylated MAP1B sites, we have taken an evolutionary approach arguing that highly conserved sites are indicative of general importance for MAP1B function and therefore for neuronal development. We assembled MAP1B amino acid sequences at identified GSK-3β primed and non-primed sites from data in sequence databases and by sequencing reverse transciption-PCR products from MAP1B RNA. Inspection of the amino acid sequence of MAP1B from representatives of all of the vertebrate classes at GSK-3β primed and non-primed sites revealed a remarkable conservation of phospho-acceptor amino acid residues and flanking sequences. Non-primed GSK-3β MAP1B sites were more highly conserved than primed GSK-3β MAP1B sites, which were largely absent in teleost fish, suggesting that the former evolved earlier than the latter. It will now be important to extend this analysis to the invertebrates. There is a MAP1B homologue in fruit flies (Drosophila melanogaster) called futsch that is expressed in some neurons in fly embryos (Hummel et al. 2000; Roos et al. 2000) and a GSK-3 homologue called shaggy/Zeste-white 3 that is present in axons (Bobinnec et al. 2006). Furthermore, futsch is a phosphorylation substrate of shaggy at a site recognized by pAb BUGS (Gögel et al. 2006), an antibody raised against a non-primed GSK-3β phosphorylation site on MAP1B (Trivedi et al. 2005), suggesting that non-primed GSK-3β MAP1B sites are present in invertebrates.

Conservation of the intracellular distribution of primed and non-primed glycogen synthase kinase 3β phosphorylated MAP1B sites

MAP1B is present in both the somato-dendritic compartment and the axon of differentiating neurons. We have previously shown that phosphorylated non-primed GSK-3β sites on MAP1B are only present in the axonal compartment of developing rodent neurons and that they are often distributed along the axon in a gradient that is highest distally, towards the growth cone (Trivedi et al. 2005; Scales et al. 2009). In contrast, phosphorylated primed GSK-3β MAP1B sites are present throughout the neuron in both the somato-dendritic compartment and in the axon, where they are uniformly distributed (Trivedi et al. 2005; Scales et al. 2009). Here, we confirmed these findings and extended them to other vertebrates. In the chick in particular, these two distribution patterns for primed and non-primed GSK-3β MAP1B sites are strongly conserved.

We also showed that the distribution pattern of primed and non-primed GSK-3β phosphorylation sites on MAP1B is preserved in cultured embryonic rat cortical neurons and in frog nodose ganglion neurons. This finding opens up the possibility of studying primed and non-primed GSK-3β phosphorylation sites on MAP1B in vitro. We used these cultures to investigate the properties of the axonal gradient of a non-primed GSK-3β phosphorylation site on MAP1B at T1261 recognized by pAb SuperBUGS (Trivedi et al. 2005). Previous work has shown that another phospho-epitope on MAP1B, recognized by mAb 150, is also distributed in growing axons in a gradient that is highest distally (Bush et al. 1996). However, after simultaneous detergent extraction and fixation, mAb 150 labelled axons uniformly along their length, suggesting that a masking protein, removed by detergent extraction, blocked access of the antibody to MAP1B in the proximal region of axons (Bush et al. 1996). Here, we found that the axonal gradient of a non-primed GSK-3β phosphorylation site on MAP1B at T1261 in rat recognized by pAb SuperBUGS remained unchanged after detergent extraction and fixation, suggesting that the gradient is real and not apparent. How the gradient is established is unknown. GSK-3β is uniformly distributed in neurons but the phosphatase, which has yet to be identified, may not be.

Distinct functions of primed and non-primed glycogen synthase kinase 3β phosphorylated MAP1B sites

The differential distribution of MAP1B phosphorylated at primed and non-primed sites in differentiating neurons suggests that there is a differential regulation of primed and non-primed sites in neurons and that these sites have different functions in neuronal differentiation. Indeed, there is evidence that primed GSK-3β sites are involved in axon branching, whereas non-primed GSK-3β sites are important for axon growth (Kim et al. 2006). Zebrafish MAP1B does not have a primed GSK-3β site, corresponding to S1392 in rat MAP1B. This predicts that zebrafish axons might be less branched than in other vertebrates; however, there are no comparisons of axon branching between vertebrates. In future experiments it will be important to test this hypothesis, for example by separately mutating primed and non-primed GSK-3β sites on MAP1B.

Our understanding of the functional role of GSK-3β phosphorylation sites on MAP1B is poor. In general, protein phosphorylation has two consequences: (i) the modification of intrinsic function, such as catalytic activity, and (ii) the generation or modification of a binding site either for another protein or for an intramolecular domain. One intrinsic function of MAP1B is to bind to microtubules and therefore this function might be modulated by phosphorylation. MAP1B has been reported to have a range of binding partners, including kinases (e.g. death-associated protein kinase; Harrison et al. 2008), neurotransmitter receptors (e.g. GABAc; Hanley et al. 1999) and cytoplasmic proteins (e.g. glyceraldehyde-3-phosphate dehydrogenase; Cueille et al. 2007). This has led to the suggestion that MAP1B is a scaffolding protein enabling the interaction between other proteins by binding to them (Riederer, 2007). How phosphorylation might regulate binding is unknown.