|Home | About | Journals | Submit | Contact Us | Français|
Reversible tyrosine phosphorylation, catalysed by receptor tyrosine kinases and receptor tyrosine phosphatases, plays an essential part in cell signaling during axonal development. The receptor protein tyrosine phosphatase PTPσ has been implicated in the growth, guidance and repair of axons from retinal axons. This phosphatase has also been implicated in motor axon growth and innervation. Insect orthologues of PTPσ are also implicated in the recognition of muscle target cells. A potential extracellular ligand for vertebrate PTPσ has been previously localised in developing skeletal muscle. The identity of this muscle ligand is currently unknown, but is appears unrelated to the heparan sulphate ligands of PTPσ. In this study, we have used affinity chromatography and tandem mass spectrometry to identify nucleolin as a binding partner for PTPσ in skeletal muscle tissue. Nucleolin, both from tissue lysates and in purified form, binds to PTPσ ectodomains. Its expression pattern overlaps with that of the PTPσ-binding partner previously localised in muscle and we demonstrate that a significant amount of muscle-associated nucleolin is present on the cell surface of developing myotubes. Furthermore, two nucleolin-binding components, lactoferrin and the HB-19 peptide, can block the interaction of PTPσ probes with muscle in tissue sections. These data suggest that cell surface-associated nucleolin is a potential component of the muscle binding sites for PTPσ, and would be accessible on the cell surface to axonal PTPσ.
Vertebrate nervous system development relies on a multitude of guidance cues to stimulate axonal extension and guidance as well as to establish stable synaptic connections with targets such as muscles. Interpretation of these environmental signals by growth cones involves multiple receptor classes such as cell adhesion molecules (CAMs) , DCC and Neuropilins [2,3], and enzymes involved in phosphotyrosine signaling, such as the receptor protein tyrosine kinases (RPTK’s)  and receptor protein tyrosine phosphatases (RPTPs) [5-7].
Evidence for the role of phosphotyrosine signaling during axon growth and guidance has come from studies that show fibroblast growth factor (FGF) receptor signalling, a RPTK acting alongside neural CAMs, promotes neurite growth [8,9]. Also, the Eph family of RPTKs are a large group of enzymes whose members regulate retinal axon guidance in direct response to graded patterns of their ligands, the ephrins [10-12]. It is therefore unsurprising that the counterbalancing enzymes, the RPTPs are also implicated in many of these processes. There are 21 recognized human RPTPs , with homologues and orthologues across the vertebrate phylum and in invertebrates. RPTPs show strong developmental expression in central and peripheral nervous systems, coinciding with significant events such as axonogenesis, target contact, synaptogenesis and plasticity [14,15]. The type IIa subfamily is representative of these neural RPTPs. These enzymes have two cytoplasmic phosphatase domains and an extracellular region consisting of immunoglobulin like domains and fibronectin type III repeats, similar to the NCAM family of cell adhesion molecules . Members of the type IIa RPTPs include vertebrate LAR, PTPδ and PTPσ, leech hmLAR and Drosophila DLAR and DPTP69D. Studies have implicated PTPδ and PTPσ in retinal axon development in both chick and Xenopus [16-18]. In Drosophila, DLAR and DPTP69D have roles in photoreceptor, and commissural axon guidance [19-21]. Mouse LAR deficiency leads to reduced size of basal forebrain cholinergic neurons and diminished hippocampal innervation [22,23], whereas PTPδ deficiency causes impaired learning and enhanced hippocampal long-term potentiation . PTPσ deficiency leads to the most extreme defects with hypomyelination of peripheral nerves, abnormalities in development of the hypothalamus and pituitary, and ataxias [25,26].
Type IIa RPTPs have also been implicated in a further area of neural development, that of the neuromuscular system. In Drosophila, DLAR and DPTP69D are required for axon guidance of motor neurons [27,28], and DLAR has also been implicated in neuromuscular synaptic plasticity [29-32]. LAR-related RPTPs also influence synaptogenesis in muscles in other species [33,34]. A recent study of mice with a double gene deficiency in PTPσ and PTPδ, demonstrated that these two RPTPs are critical for the generation of a branched innervation pattern in the diaphragm, and subsequent motor neuron survival . In the chick embryo, there is strong expression of PTPσ in vertebrate spinal and cranial motor neurons [36,37] and evidence from affinity probe assays of a potential PTPσ ligand in developing muscle . Recent studies using gene knockdown in the chick spinal cord have also shown that type IIa RPTPs play a role in motor axon growth [Stepanek, 2005 #2]. These collective data show that vertebrate type IIa RPTPs, like DLAR in the fly, are involved in neuromuscular development.
Despite the functional data implicating RPTPs in neuromuscular development, little is known about the signaling role of these RPTPs and how such signaling is regulated. One way to understand these events might be to identify the extracellular ligands of the type IIa RPTPs in the neuromuscular system. For example, it is known that PTPδ and an isoforms of LAR can interact homophilically and that LAR can also bind heterophilically to the laminin-nidogen complex [O′Grady, 1998 #62]. Heparin sulphate proteoglycans (HSPGs) have also been identified as potential ligands FOR PTPσ . and recently syndecan and dally-like, both HSPGs, have been reported as functional ligands for neuronal LAR in Drosophila motor axons and neuromuscular synaptogenesis [29,31,32]. Nevertheless, the potential ligand of PTPσ within developing muscle of the chick embryo appears not be HSPG-related, and only interacts with the short protein isoform of PTPσ expressed in motor neurons . Given the interest in PTPσ in neuromuscular development, we have undertaken the identification of this potential muscle ligand using an affinity chromatography approach. We report that chick nucleolin, expressed on the cell surface of developing muscle cells, is a PTPσ binding protein. Nucleolin expression correlates with the location of PTPσ binding sites on developing muscle and nucleolin-binding proteins and peptides can perturb this PTPσ binding. These data demonstrate that cell surface nucleolin is a candidate ligand for PTPσ and is likely to be part of the PTPσ-binding site in developing skeletal muscles.
We have shown previously that PTPσ binds to an unidentified ligand(s) in the developing muscle of the chick . To identify PTPσ binding proteins and potential ligands, an immobilised fusion protein consisting of the first six subdomains of the PTPσ ectodomain fused to alkaline phosphatase (termed FN3d-AP, figure 1A, 1B) was used to perform affinity chromatography on solubilised muscle tissue from 10 day old (E10) chick embryos. To identify specifically-retained proteins that interact with PTPσ, we performed (as a negative control), chromatography on alkaline phosphatase-conjugated sepharose. Detergent extracts of chick muscle tissue were loaded onto these two columns and proteins were eluted using a high salt buffer. Eluted proteins were compared after SDS gel electrophoresis and this revealed a complex pattern of protein bands. The only reproducible difference observed was a 95 kDa band identified as being present in the eluate from the PTPσ column, but absent in the control eluate (Figure 1B). For protein identification, multiple affinity runs were performed, eluates concentrated, separated by SDS gel electrophoresis and stained with coomassie. The band of interest was excised from the gel, digested with trypsin, and analysed by tandem mass spectrometry. As shown in Figure 1A, 21 peptides were sequenced and found to correspond to chicken nucleolin (SwissProt accession number P15771). All 21 peptides could be identified within the C-terminal region of the nucleolin sequence (Figure 1B) with no peptide sequence tags being obtained from the N-terminal part of nucleolin. This is likely due to the clustering of glutamic acid residues within the N-terminal region, preventing the formation of reasonably sized peptides for MS/MS analysis. The calculated mass of nucleolin based on its sequence is 76 kDa, however it migrates at approximately 100 kDa in SDS gel electrophoresis due to post-translational modifications and a high content of negatively charged amino acids . The identity of the 95 kDa protein band as Nucleolin was confirmed by immunoblotting eluates using anti-nucleolin antibody (Figure 1D). This revealed a band at approximately 95 kDa present in the PTPσ eluate only. These data confirm that nucleolin is a binding partner for PTPσ under these conditions.
To address whether nucleolin can bind directly to PTPσ, we carried out solid phase binding assays using recombinant, myc-tagged chick nucleolin purified from transfected 293T cells. Purification of nucleolin is very difficult since it is accompanied by a high degree of protein degradation (D. Alete, A. Hovanession, unpublished work). Although our purified chick nucleolin was similarly only around 20% intact (data not shown), it was considered of sufficient quality for initial solid phase overlay assays. The purified nucleolin was immobilised on charged microtiter plates and incubated with varying concentrations of conditioned media containing FN3d-AP. The data revealed significant binding of PTPσ to nucleolin above that of the BSA control (Figure 3), demonstrating that PTPσ can bind to nucleolin directly. Nevertheless, this assay was near the limit of detection for this interaction, since successive dilution of the probe soon led to loss of signal. The low yields of the AP probe and of nucleolin, together with the inevitable partial degradation of nucleolin, meant that calculations of binding affinity are unrealistic at this stage.
To assess if the nucleolin identified by affinity chromatography is potentially a muscle ligand for PTPσ, we compared the expression pattern of nucleolin with the distribution of the PTPσ muscle ligand seen when using the receptor affinity probe (RAP) assay [41,42] (figure 4). E10 cranial tissue sections were cut and stained using FN3d-AP fusion protein (figure 4A and C) or anti-nucleolin antibody (figure 4B and D). The strongest binding of the FN3d-AP fusion protein is localised to the muscle tissue (figures 4A and C), with binding also observed in motor nerves and other scattered cells and matrix . The most intense staining is observed on the myotubes of the developing muscle (arrows, figure 4C) . Immunofluorescence staining using nucleolin antibody displayed a closely overlapping staining pattern to that seen in the RAP assay, with most of the staining localised to developing muscle (figure 4B). Increased magnification showed the most intense staining was, as with the RAP stain, within the myotubes and as patches on myotube surfaces (arrow figure 4D). Technical limitations mean that we cannot directly demonstrate how much of the RAP stain and nucleolin fluorescence directly overlap. Nevertheless, these data demonstrate that the developmental expression of nucleolin within developing muscle is largely consistent with the location of the PTPσ muscle ligand(s).
We have shown that nucleolin is a PTPσ binding protein and that its expression corresponds to the location of PTPσ muscle ligands. In order to ascertain whether FN3d-AP binds to nucleolin in chick muscle tissue, we tested whether the pentameric pseudopeptide 5(Kψ(CH2N)PR)-TASP (referred to as HB-19 ) could perturb the binding of FN3d-AP. HB-19 is a potent inhibitor of HIV entry into the cell, acting by specifically binding to, and forming complexes with, cell surface nucleolin [44-46]. It has been shown to exert this effect independently of heparin sulphate proteoglycans (HSPGs), by binding the C-terminal tail of nucleolin containing the RGG domain, consisting of residues 656-707 . HB-19 (10 μM) or BSA (0.5 mg/mL) were pre-bound to chick sections, washed, and RAP analysis then performed as described earlier (figure 5). Pre-binding of HB-19 prevented subsequent FN3d-AP binding to both muscle and basement membranes (figure 5 A, B and data not shown). We were surprised that HB-19 blocked the basement membrane sites, since these were thought to be only HSPG-dependent. Nevertheless, immunostaining using a biotinylated form of HB-19 revealed binding to muscle tissue (data not shown) as well as basement membranes in the retina (figure 5 G), possibly explaining why HB-19 could block FN3d-AP binding to its HSPG ligands. No non-specific disruptive effects were observed by HB-19 on the binding of antibodies to the antigens laminin and myosin (data not shown).
Further perturbation experiments were carried out using the protein lactoferrin. Lactoferrin, an iron binding protein of the transferrin family, is present in external secretions and the secondary granules of polymorphonuclear leukocytes . Lactoferrin is a highly basic protein  that bind to, and is internalised by, cell surface nucleolin, with the binding site located within the C-terminal RGG domain of nucleolin . We pre-incubated sections of E10 chick embryos with 0.5 mg/mL lactoferrin. This preincubation with lactoferrin effectively blocked Fn3d-AP binding to muscle tissue (figure 5, C and D). Like HB-19, lactoferrin also perturbed PTPσ binding to the basement membrane of the retina (figure 5, E and F). Lactoferrin is also a HSPG binding protein [48,51], and may therefore be directly interfering competitively with PTPσ binding to basement membrane-associated HSPGs. It is unlikely, however, that lactoferrin blocks PTPσ binding in muscles through an effect on HSPGs , since our previous work has shown that the muscle binding site is not HSPG-related .
To confirm that lactoferrin can bind to nucleolin from developing muscle, we carried out affinity chromatography of muscle lysates using immobilised lactoferrin. Nucleolin was specifically retained on the lactoferrin column (figure 1E). Immunoblots using antibodies against actin and myosin, as a control for non-specific binding, were negative (data not shown). Pre-binding of lactoferrin to muscle sections also showed no non-specific, disruptive effect on the binding of antibodies to antigens such as laminin and myosin (data not shown).
Both HB-19 and lactoferrin bind to the C-terminal RGG domain of nucleolin. Therefore if PTPσ also binds to this domain, we would predict that an antibody raised to a sequence outside this domain might have little effect on the RAP assay signal. This was tested with an anti-nucleolin antibody (Santa Cruz, USA) raised against amino acids 271-520. No effect on PTPσ binding to muscle was observed (data not shown). There is currently no RGG-specific antibody to test directly whether it can block PTPσ binding.
These data demonstrate first that two known nucleolin binding components (HB-19 and Lactoferrin) can specifically inhibit PTPσ binding to its muscle ligand. Second, binding of PTPσ to nucleolin is likely to be mediated through the RGG domain of nucleolin.
In order to function as a potential ligand for PTPσ, nucleolin must be present on the surface of the target tissue. Originally nucleolin was reported to be exclusively present within the nucleus , however more recent studies have shown that nucleolin is present on the surface of a variety of cell lines [40,46,50,53,54] and on the surface of endothelial cells in angiogenic blood vessels [55, 56]. To address whether nucleolin is also present on the surface of developing muscle cells, we carried out immunofluorescence analysis of primary chick muscle cells isolated from E10 embryos (figure 6). These were grown in moderately high serum, as this has been reported to promote the cell surface localisation of nucleolin in other cell types (?). Live, non-permeabilised cells and fixed, semi-permeabilised cells were co-stained with anti-nucleolin and anti-myosin antibodies. Anti-nucleolin staining of the live non-permeabilised cells showed intense punctate patches on the outside of the cells (figure 6B). No myosin staining was observed in non-permeabilised cells, which confirms the integrity of the membrane. By contrast, semi-permeabilised cells showed high levels of nucleolin staining within the cytoplasm, with myosin staining also observed within these cells (figure 6A). Since paraformaldehyde was used for partial permeabilisation, we did not expect to see nucleoloar localisation of nucleolin in these experiments.
To determine more precisely if the staining observed on the non-permeabilised cells was present on myotubes, live cells were first incubated with anti-nucleolin antisera, then fixed and permeabilised and treated with myosin antibody (figure 6 C and D). We observed punctate staining of nucleolin on the surface of myotubes (arrow figure 7D), as well as on the surface of non-myosin expressing cells. Some of the punctate localisation also occurred at the cell-cell interface between these cells and myotubes (arrowheads figure 6D). It is of interest that a punctate pattern of nucleolin is also seen on the surface of Hela cells after treatment of live cells with a nucleolin ligands, midkine . It is possible therefore that some of the punctate pattern in muscles cells is caused by clustering of nucleolin by the antibodies in the live cells.
To address further the question of which cells express surface nucleolin in muscle, we performed confocal microscopy on chick muscle sections co-stained with nucleolin and myosin antibodies (figure 6E and F). In a three-dimensional reconstruction (figure 6F), intracellular overlap is observed as yellow staining, but in addition we observed punctate regions of nucleolin on myotube surface (arrows in figure 6E). Furthermore, three-dimensional rendering of the myotubes using the green channel (myosin) to differentiate between intracellular nucleolin signal (yellow, arrowhead figure 6F) and surface nucleolin (red), shows nucleolin present on the surface of developing myotubes (arrows, figure 6F). This surface nucleolin should therefore be accessible to cell surface PTPσ.
Studies carried out previously have identified two distinct ligand locations for PTPσ, within the basement membrane of the retina  and in developing muscle . The basement membrane-associated ligands have been identified as HSPG’s. Here, using affinity chromatography, tandem mass spectrometry and RAP affinity assays, we report the identification of the multifunctional protein nucleolin as a potential new ligand present in developing muscle. It was confirmed that nucleolin and the PTPσ ectodomain can directly interact. Furthermore, we demonstrated for the first time that nucleolin is expressed on the surface of developing myotubes, and that its localisation in muscle overlaps that of the previously characterised PTPσ interactor.
Nucleolin was first described as a major nuclear protein consisting of a negatively charged N-terminal domain, an RNA-binding domain and a C-terminal domain rich in RGG motifs . Nucleolin has been reported to be involved in a diverse array of cellular processes, including cell proliferation and growth, cytokinesis, replication, embryogenesis and nucleogenesis . More recently, numerous studies have reported nucleolin as being present on the cell surface [55,56,60,61] and to function as a ligand/receptor for a number of different proteins including lactoferrin , pleiotrophin , achran sulphate , HIV , L-selectin  and midkine . Nucleolin does not have a classic secretion signal and it is therefore not known how it reaches the cell surface. Nucleolin can, however, reach cell surfaces without endogenous HSPG production  and Nucleolin has even been reported to function as a shuttle between the cell surface and the nucleus .
Our study has now shown that nucleolin is also found on the surface of developing myotubes. Indeed, the overall expression of nucleolin appears generally high in muscle at the embryonic stage examined, although its non-nucleolar localisation was also observed in several other tissue sites. It is possible that there is a much more restricted pattern of expression of the cell surface form of nucleolin, but current antibodies cannot specifically distinguish it from intracellular forms. The patch-like or punctate pattern of nucleolin on myotubes in culture and also in muscle sections, may also indicate that nucleolin has a role at localised areas of the developing muscle membrane. It is noteworthy that several binding partners of nucleolin are also found in a punctate pattern on cell surfaces and in the case of pleiotrophin can co-patch nucleolin . Our previous data  suggest that the PTPσ ligand does not colocalise, at least not exclusively, with developing neuromuscular junctions. Our recent examination of the localization of nucleolin and acetylcholine receptor indicates that nucleolin may overlap with, but is not notably enriched in, neuromuscular junctions. Therefore, although we do not yet understand the function of this nucleolin, it may form part of a muscle surface complex that recognises axonal molecules such as PTPσ before neuromuscular junction formation.
In the blocking experiments with lactoferrin and HB19, our data also showed a blockade of PTPσ interactions with other known ligand sites, in particular those of HSPGs in retinal basement membranes. It is not clear what this means at present since the basement membrane interactions of PTPσ are absolutely dependent on HSPGs. If they are also dependent on nucleolin, then this might invoke a receptor complex containing both HSPGs and nucleolin, both of which might be necessary for a functional interaction with PTPσ in the retinal inner basement membrane. There is no further evidence for this currently, but it is interesting to note that of the molecules known to bind to surface nucleolin, pleiotrophin, midkine, lactoferrin and PTPσ also bind to HSPGs [46,51]. For example, both midkine and PTPσ have both been reported to bind to the HSPG agrin [39,64]. Although this may be coincidental, it suggests an underlying theme that these molecules share some binding properties and may therefore interact with nucleolin in a similar fashion. Having said this, the situation in muscle is still distinct, since PTPσ binding occurs independently of heparan sulphate . Furthermore, nucleolin does not require HSPGs to reach the cell surface .
The mechanism of molecular interaction between PTPσ and the muscle-associated nucleolin remains to be determined. However, in light of the fact that HB-19 and lactoferrin bind to the RGG domain at the C-terminal tail of nucleolin, and both these components perturb the interaction between PTPσ and its muscle ligand, it is plausible to suggest that PTPσ interacts with the RGG domain of surface nucleolin.
The biological significance of the interaction between PTPσ and nucleolin in muscle has also yet to be elucidated. Although PTPσ, along with PTPδ, influence motor axon growth and branching within the target field [65,66](new ref.), nucleolin itself has not so far been implicated in muscle or neuromuscular development. Furthermore, although nucleolin may act as a co-receptor for HIV for example, the normal molecular function of cell surface nucleolin in any type of cell is still relatively unclear. From the present study, we could hypothesise that nucleolin might serve as part of a receptor complex on the surface of developing muscle, recognising adhesive molecules such as PTPσ present on incoming growth cones of motor neurons. Recent studies have indeed shown that cell surface nucleolin can function as a cell adhesion molecule .
To address the function of nucleolin in muscles, methods must first be developed for isolating large amounts of non-degraded nucleolin. If nucleolin is involved in a recognition complex, it will then be possible directly to test what the cellular and biochemical consequences are of PTPσ-nucleolin interactions.
The Fnd3-AP protein represents a truncated ectodomain region of cPTPσ1 (amino acids 1 to 597), fused at its carboxy terminus to the placental alkaline phosphatase (AP) gene in vector pBG as described previously . The Fnd3-AP expression vector was transfected into 293T cells (grown in Dulbecco’s modified Eagles medium, 10% fetal calf serum, 1% penicillin-streptomycin mixture; sigma) using calcium phosphate. Conditioned medium containing the secreted FN3d-AP fusion protein was collected after 6 to 7 days, sterile filtered, buffered to pH 7.4 with 20 mM HEPES, and stored at 4°C. Receptor affinity probe (RAP) assays were carried out on unfixed tissue cryosections as described previously .
1 mL of anti-PLAP (placental alkaline phosphatase, Sigma) was packed into a Fast Protein Liquid Chromatography (FPLC) column (Amersham Biosciences). Purification of FN3d-AP was carried out using an AKTA FPLC system (Amersham Biosciences). The column was equilibrated using 5 column volumes of 0.05M Tris, 0.5M NaCl (pH 8.0) at a flow rate of 0.5 ml min-1 and flow through absorbances measured at 280 nm. Conditioned media was centrifuged at 1000 rpm for 5 mins, the supernatant recovered and loaded onto the column. The column was then washed with 5 column volumes of the equilibration buffer to remove unbound components. Bound components were eluted using 0.05M glycine, 0.5M NaCl (pH 2.8). Fractions (500 μL) were collected directly into tubes containing 50 μL of a 1.0M Tris-HCL (pH 9.0) solution. Purified fusion constructs (determined by the absorbance at 280 nm) were pooled, desalted using a PD10 desalting column (Amersham Biosciences) and their purity determined by gel electrophoresis.
Limb and chest muscle tissue (1 g) was dissected from E10 chick embryos and homogenized with a glass homogenizer in 10 ml 4% CHAPS, 100 mM KH2PO4, pH 7.5, 5% glycerol, protease inhibitor cocktail (Roche). The lysate was vortexed for 1 h at 4°C, centrifuged (4000 rpm), the supernatant recovered and diluted 1:2 in PBS and incubated with affinity matrix (1mL of CnBr sepharose covalently coupled to 2 mg FN3d-AP, AP or lactoferrin) overnight at 4°C. Chromatography was carried out on an AKTA FPLC system (Amersham Biosciences) under the following conditions; the column was washed with 5 column volumes of PBS containing 0.1% Tween, 30 mM EDTA. Bound components were eluted using PBS containing 0.5 M NaCl (2 ml) into 200 μl fractions, separated by polyacrylamide gel electrophoresis (PAGE; 6% gel) under reducing conditions and visualized by silver stain .
Bands from a coomassie-stained gel were cut out and subjected to digestion with trypsin as follows; Gel pieces were washed three times in 30 μL 50% CH3CN with agitation. The gel pieces were dried in a vacuum centrifuge for 10 min, reduced with 10 mM DTT, 10 mM NH4HCO3, pH 8.0 (15 μl) for 45 min at 50°C followed by alkylation with 50 mM idoacetamide, 10 mM NH4HCO3 for 1 h at room temperature in the dark. Gel pieces were washed three times in 30 μl 50% CH3CN andvacuum-dried before being re-swollen with 50 ng of modified trypsin (Promega, Southampton, UK) in 5 μl 10 mM NH4HCO3. The pieces were then overlaid with 10 mM NH4HCO 3(10 μl) and incubated for 16 h at 37°C. The samples were centrifuged and supernatant recovered. Peptides were further extracted twice with 10 μl 5% trifluoroacetic acid in 50% CH3CN and the supernatants pooled. Peptide extracts were vacuum-dried and resuspended in 5 μl of ddH2O containing 20 mM NH4HPO4. Digested peptide mixtures were separated by nanoHPLC (Ultimate; LC Packings, Amsterdam, Holland) equipped with a PepMap column (75μm_15cm; LC Packings) at a flow rate of 300nl/min. Eluting peptides were analysed by electrospray ionization (ESI)-MS/MS in a quadrupole/orthogonal acceleration time-of-flight (Q-Tof) mass spectrometer (Micromass, Wythenshaw, Manchester, U.K.) using a nanoelectrospray ion source and ESI emitters with a 15μm tapered end (New Objective, Woburn, MA, U.S.A.). Proteins were identified using the SwissProt database with Mascot search engine (www.matrix-science.com). A parent ion tolerance of + 3 m/z; peptide ion tolerance of + 3 m/z; one missed cleavage; fixed carbamidomethylation of cysteines and variable oxidation of methionines were specified.
A myc-tagged chick nucleolin expression vector (gift from E.A. Nigg) was transfected into 293T cells as described earlier. Cells were cultured for two days, washed in cold PBS and solubilised in 4% CHAPS, 100 mM KH2PO4, pH 7.5, 5% glycerol, protease inhibitor cocktail (Roche). The lysate was spun at 3000 rpm and the supernatant was recovered. Myc-tagged nucleolin was purified using anti-myc agarose (sigma) as described for fusion protein purification.
15 μg nucleolin and bovine serum albumin (BSA) were immobilised on a 96 well microtiter plate for 2 hrs at RT. Remaining binding sites were saturated by overnight incubation in PBS containing 2% Goat serum (Dako). Wells were incubated for 3 h at room temperature with conditioned media containing AP fusin proteins (FN3d). After 4 washes in PBS and one in SEAP buffer (0.5 mM MgCl2, 1M diethanolamine pH 9.8), the bound AP activity was determined by adding 200 μL SEAP buffer containing 10 mM p-nitrophenyl phosphate. Progress curves were recorded for 1 h at room temperature, at 405 nm, using a Dynex MRX microplate reader.
Affinity purified samples were separated on a 6% Tris-glycine gel by SDS-PAGE and transferred onto PVDF membrane for 40 mins at 120 V. Membranes were blocked with PBS containing 5% milk powder for 1 h and incubated with rabbit polyclonal anti-nucleolin antibody (1: 5000, Santa Cruz) in PBS, 5% milk, 0.05% Tween-20 overnight at 4°C. After extensive washing, the membranes were incubated with peroxidase coupled goat anti-rabbit antibody (1:10,000, Dako), and bound antibody detected using ECL (Amersham Biosciences).
E10 chick embryo heads were frozen in OCT compound (Tissue Tek), cryosectioned (10 to 12 μm) and mounted on superfrost plus slides (VWR). For immunohistochemistry, sections were fixed for 5 mins with -20°C methanol then blocked with 1% Goat serum, 0.05% Tween in PBS for 30 mins at room temperature. Primary antibodies used were rabbit anti-nucleolin (Santa Cruz) at 1:100 dilution, , mouse anti-laminin 1 (mab31) at 1:100. Antibodies were diluted in PBS containing 1% goat serum. 0.05% tween 20, and incubated on sections for 1 h at room temperature. Sections were then washed three times in PBS, 0.05% Tween 20 and secondary antibodies, (goat anti-rabbit biotin conjugated, 1:100, Dako; goat anti-mouse FITC conjugated 1:100, Dako) were incubated for 1 h at room temperature. After three final washes FITC labeled sections were mounted with Vectashield Hardset™ mounting medium with DAPI (Vector labs). For biotin labeled sections, slides were incubated for a further 30 mins in PBS, 0.05% Tween, containing streptavidin conjugated Cy3 (1:400, Amersham biosciences), washed and mounted as described. For HB-19 peptide staining, biotinylated HB-19 peptide  was diluted in PBS containing 1% goat serum, 0.05% tween 20, to a final concentration of 10 μM and incubated for 1 h at room temperature. Sections were washed and incubated with streptavidin conjugated Cy3 as described above. Sections were analyzed using an axiophot fluorescence microscope (Ziess).
For immunofluorescence, embryonic myotubes cultures were established from E10 chick muscle tissue. E10 trunk tissue was dissociated enzymatically and plated at 106 cells per 60mm plate on fibronectin-coated glass coverslips in Dulbecco’s modified Eagles medium, 2% chick serum, 4% fetal calf serum, 1% penicillin-streptomycin mixture (sigma) for 72 hrs. For cell surface staining, primary antibodies (rabbit anti-nucleolin, 1:100; mouse anti-myosin (F59), 1:100) were diluted in growth media and incubated with the cells for 30 mins at 4°C. Cells were then washed three times with cold PBS and fixed with 4% Para formaldehyde for 15 mins at 4°C. Secondary antibodies (anti-rabbit biotin conjugated, anti-mouse FITC conjugated) were diluted 1:100 in PBS, 1% goat serum, 0.05% Tween 20 and incubated with the cells for 1 h at room temperature. Cells were washed, mounted and analyzed as described above. To semi-permeabilise cells for intracellular staining, cells were incubated in 4% paraformaldehyde for 30 mins at 4°C in the first instance and then stained as described above.
E10 chick head sections were stained as described earlier with antibodies against myosin and nucleolin with FITC and Cy3 secondary’s respectively. The sections were examined using a Leica TCS 4D laser scanning confocal microscope (leica, Milton Keyes, UK). The 488 nm line of the laser was used to visualize the FITC-myosin and the 568 nm line used for the Cy3-nucleolin. Using these wavelengths, separation of the fluorescent signals from the two fluorophores was almost complete. A series of optical sections 1 μm apart were taken through a depth of 20 μm. All images were stored digitally, and 3D reconstruction and visualization carried out using the Volocity software (Improvision Ltd, Coventry, UK). The parameters were set as follows: green channel 100% density, red channel 1% density, medium noise filter.
We thank Clare Faux and Juan Pedro Martinez-Barbera for critical reading of the manuscript. The research was funded by the Wellcome Trust (071418).