[1] 14w Semaphorin3A regulates axon growth independently of growth cone repulsion via modulation of TrkA signaling

[1] 253w Regulation of axon growth is a critical event in neuronal development. Nerve growth factor (NGF) is a strong inducer of axon growth and survival in the dorsal root ganglia (DRG). Paradoxically, high concentrations of NGF are present in the target region where axon growth must slow down for axons to accurately identify their correct targets. Semaphorin3A (Sema3A), a powerful axonal repellent molecule for DRG neurons, is also situated in their target regions. NGF is a modulator of Sema3A-induced repulsion and death. We show that Sema3A is a regulator of NGFinduced neurite outgrowth via the TrkA receptor, independent of its growth cone repulsion activity. First, neurite outgrowth of DRG neurons is more sensitive to Sema3A than repulsion. Second, at concentrations sufficient to significantly inhibit Sema3A-induced repulsion, NGF has no effect on Sema3A-induced axon growth inhibition. Third, Sema3A-induced outgrowth inhibition, but not repulsion activity, is dependent on NGF stimulation. Fourth, Sema3A attenuates TrkA-mediated growth signaling, but not survival signaling, and over-expression of constitutively active TrkA blocks Sema3A-induced axon growth inhibition, suggesting that Sema3A activity is mediated via regulation of NGF/TrkA-induced growth. Finally, quantitative analysis of axon growth in vivo supports the possibility that Sema3A affects axon growth, in addition to its well-documented role in axon guidance. We suggest a model whereby NGF at high concentrations in the target region is important for survival, attraction and inhibition of Sema3A-induced repulsion, while Sema3A inhibits its growth-promoting activity. The combined and cross-modulatory effects of these two signaling molecules ensure the accuracy of the final stages in axon targeting.

[1] 142w Nervous system development is critically reliant on highly stereotypic and precise patterns of neuronal connectivity. This developmental process involves axon growth and guidance, as well as neuronal selection. A growing number of target-derived survival molecules has been identified. Interestingly, many of these molecules also act as axon growth promoters for the same neuronal populations. For example, in developing DRG neurons stem cell factor, hepatocyte growth factor and NGF all function as survival molecules, in addition to their roles as robust axon growth promoters [1][2][3][4][5]. The growth-promoting effects of these molecules increase with concentration, thus axons close to their targets should grow more rapidly. Paradoxically, once axons reach their target region, axon growth must slow down to enable appropriate selection of the target. There must thus be a differential regulation of target-derived molecules that enables survival and attraction, but concurrently reduces growth activity.

[2] 195w One plausible mechanism may involve growth inhibitory molecules. Here we utilized the NGF-induced growth of DRG neurons to test whether growth inhibitory molecules can modulate the growth properties of NGF without affecting its survival or guidance activities [6]. We tested Sema3A, a potent repellent guidance molecule, as a possible growth inhibitory molecule [7]. The role of Sema3A in DRG neuronal guidance is well characterized. Close examination of many experiments in co-culture assays indicated that Sema3A could affect axon growth in addition to its directional effects. However, there was a tendency to interpret these effects on growth as secondary, resulting from its guidance role. The effects of Sema3A on growth were recently measured directly [8]. Interestingly, this study found that inhibition of RanBPM reduced the axon growth effects of Sema3A without affecting its repulsion activity, suggesting that these two activities may represent different signaling cascades [8]. However, since Sema3A is a potent repellent and death-inducing molecule in addition to its axon growth inhibition property, it is difficult to see how this molecule can reduce growth without inducing cell death or repulsion [9]. One explanation for this differential activity may be a balance between NGF and Sema3A.

[3] 162w NGF is a strong modulator of axon growth in DRG neurons. This molecule is also a potent modulator of Sema3A repellent and death-inducing activity [9][10][11]. However, the balance between the ability of Sema3A to reduce growth and that of NGF to induce growth has not been tested to date. We therefore studied whether Sema3A could regulate NGF-induced axon growth. Using a collapse assay to measure repulsion, and an explant assay to measure neurite growth, we show that Sema3A fully inhibits axon growth at concentrations sufficient to induce only weak repulsion, suggesting that Sema3A inhibition of growth is not the result of growth cone collapse, but rather a separate activity. We further demonstrate that Sema3A-induced growth inhibition is mediated via the modulation of NGF-induced TrkA activation. Finally, analysis of the total axon coverage/surface area in Sema3A null embryos suggests that the detected abnormalities in sensory axon patterning are likely to reflect a lack of axon growth modulation, in addition to defects in directionality.

[1] 171w Sema3A is best known as a growth cone repellent molecule. The effects of this molecule as an axon growth inhibitor have been noted previously [8], but were generally considered to be a side effect of axon repulsion. To begin testing the relationship between Sema3A axon growth inhibition versus growth cone repulsion, we began by testing the sensitivity of neurons in separate assays to these two activities. The effect of Sema3A on axon growth was assayed by measuring the difference between the lengths of axons at the time Sema3A was added (defined as t = 0) and 6 h later (see Experimental/materials and methods for detailed description). The measured average axon growth in the presence of 0, 6, and 7.5 pM Sema3A (Fig. 1A, B) shows that Sema3A completely inhibits axon growth at the 7.5 pM concentration. Parallel experiments using a growth cone collapse assay in DRG explants from the same embryos showed that these neurons are less sensitive to Sema3A-induced growth cone collapse than to axon outgrowth inhibition (Fig. 1C, D).

[2] 95w NGF is a known modulator of Sema3A-induced cell death and repulsion [9][10][11]. We therefore tested its effect on Sema3Ainduced inhibition of axon growth, by comparing the effects of increased concentrations of NGF on growth cone collapse versus axon growth. DRG neurons were grown in 2.5 ng/ml NGF for 20 h, and NGF was added to a concentration of 40 ng/ml 30 min before the addition of Sema3A. Consistent with previous studies [10,11], and as shown in Fig. 2A, growth cone collapse was significantly blocked (from 70% in low NGF to about 44% in high NGF).

[3] 147w To test the effects of elevated concentrations of NGF on growth we carried out a neurite outgrowth assay. In this experiment explants were grown in the presence of 2.5 ng/ml NGF. After 20 h, fresh medium with 2.5 or 40 ng/ml NGF was added for 30 min. At this time (t = 0) explants were photographed and Sema3A was added (0, 6, 7.5 or 10 pM). The explants were photographed once more following 6 h of incubation. The difference between t = 0 and t = 6 h was calculated. Surprisingly, although NGF is a strong modulator of repulsion (as shown in the collapse assay), it is unable to modulate Sema3A-induced neurite growth inhibition (Fig. 2B). These results demonstrate that neurons grown in the presence of high NGF concentrations, in conjunction with low concentrations of Sema3A, will slow their growth rate without changing their directionality of growth.

[4] 141w Axon growth is regulated by soluble growth factors, such as neurotrophins, which affect both neurite outgrowth and neuronal survival. In order to examine growth effects independently of survival roles we used DRG neurons from BAX null mice. Neurons obtained from these embryos are no longer dependent on neurotrophins for their survival, but are responsive to their growth-promoting activity [18]. In this experiment neurons from BAX null mice were tested in a dissociated growth assay with or without NGF in the culture (Fig. 3A, B). In the absence of NGF, Sema3A slightly inhibited growth of DRG axons, although this inhibition was not statistically significant (p =0.97, two-tailed t test). Neurons stimulated with NGF grew significantly more, but Sema3A inhibited their growth by 60%. These results indicate that growth as a result of NGF stimulation is much more sensitive to inhibition by Sema3A.

[5] 134w Previous studies have shown that the expression of NP1, the Sema3A binding unit, is affected by NGF, so that low NGF levels result in lower levels of NP1 protein expression [11,19]. We therefore tested the sensitivity of DRG neurons to Sema3A-induced collapse in the absence of NGF. In these experiments neurons from BAX null mice were grown for 20 h in the presence or absence of NGF. At this time neurons were tested in a growth cone collapse assay using 30 pM Sema3A (Fig. 3C). Both in the presence and absence of NGF, neurons are similarly sensitive to Sema3A-induced growth cone collapse. Taken together, these results show that Sema3A inhibits NGF-induced axon growth. In contrast NGF activity is not required for Sema3A guidance activity, suggesting that the mechanisms mediating these two activities are different.

[6] 345w Sema3A-induced growth inhibition is dependent on NGF activity, and substantial evidence indicates that NGF-induced growth is at least in part mediated by TrkA, a cell surface tyrosine kinase NGF receptor. TrkA contains ten evolutionarily conserved tyrosines in its cytoplasmic domain, of which three, Y670, Y674 and Y675 (human TrkA sequence nomenclature), are present in the autoregulatory loop of the kinase domain that controls tyrosine kinase activity [20]. Phosphorylation of these residues further activates the receptor. Phosphorylation of the other tyrosine residues promotes signaling by creating docking sites for adapter proteins (reviewed in [21]). Research has focused on interactions mediated by phosphorylated tyrosines Y490 (SHC site) and Y785 (PLCγ site), which are the major phosphorylated tyrosine residues that are not in the kinase activation domain [22]. To test the possibility that Sema3Ainduced axon growth is mediated by inhibition of NGF-induced activation of TrkA, we monitored the phosphorylation level of this receptor following Sema3A stimulation. Fig. 2. NGF modulates Sema3A-induced growth cone collapse, but not Sema3A-induced axon-growth inhibition. DRG explants were cultured in the presence of 2.5 ng/ml NGF for 20 h (A, B). At this time, neurons were re-fed with fresh medium containing 2.5 (Low) or 40 (High) ng/ml NGF. A) Growth cone collapse. After refeeding with low NGF (gray bars) or high NGF (empty bars) the explants were incubated for another 40 min with or without 30 pM Sema3A. The explants were then fixed and stained with rhodamine phalloidin. Growth cone collapse results represent the mean +/-s.e.m. of three independent experiments (n N 1000 growth cones for each treatment). B) Neurite outgrowth assay. DRG explants from the same embryos as in A were cultured with medium containing low NGF (gray bars) or high NGF (empty bars) concentrations as in A. For each NGF concentration, the explants were treated with 0, 6, 7.5, or 10 pM Sema3A. Each explant was monitored at time 0 and t = 6 h (as described in Fig. 1). Neurite outgrowth assay results represent the mean +/-s.e.m. of each treatment across three independent experiments (n N 240 axons for each treatment).

[7] 92w We first monitored the phosphorylation levels of phospho-Tyr 490 by Western blots. Neurons were grown for 20 h in the presence of 10 ng/ml NGF. The cells were then re-fed with fresh medium and NGF, and incubated for 60 min before addition of Sema3A. Neurons were then harvested at 0, 5, 30 and 60 min following Sema3A stimulation (t = 0 is 60 min following re-feeding of neurons with fresh NGF) (Fig. 4A). The results show that the addition of Sema3A significantly accelerates the kinetics of de-phosphorylation of the TrkA Tyr490 site.

[8] 248w To further test the activation levels of TrkA we also monitored the phosphorylation levels of Tyr785 in immunofluorescence experiments (Fig. 4B). In this set of experiments explants were grown for 20 h in the presence of 10 ng/ml NGF. At this time, control cultures were fixed (low NGF). Other explants were re-fed with fresh NGF (40 ng/ml), and incubated for an additional period of 30 min before adding 6 pM Sema3A. Control cultures had no Sema3A added. Neurons were fixed after 1 or 5 min and phosphorylation levels were then monitored in the growth cones. To estimate the levels of phosphorylation we stained the neurons using phospho-TrkA785 antibody. To correct for non-specific changes in protein levels at the growth cones we labeled all proteins non-specifically using 5-(4,6-dichlorotriazin-2-yl) aminofluorescein (DTAF), as described previously [23]. To normalize the fluorescence intensity of phospho-TrkA we divided it by the DTAF measurement of the same area. As a control, some explants were stained with anti-TrkA and analyzed for fluorescence intensity by the same method (Fig. 4C). Quantification of normalized fluorescence intensity for the different antibodies and different conditions is shown in Fig. 4D (only growth cones with no obvious morphological changes were analyzed). As with the western blot results it is evident that Sema3A stimulation reduces the phosphorylation levels of TrkA in these two positions, known to play an important role in axon growth stimulation by NGF. Taken together, our results show that Sema3A-induced growth inhibition is mediated by inhibition of TrkA activity.

[9] 236w Our results show that Sema3A induced a rapid dephosphorylation of two adaptor sites of TrkA (phospho-Tyr490 and phospho-Tyr785). It is not likely, however, that Sema3A is able to globally suppress TrkA signaling since NGF is able to inhibit Sema3A-induced collapse and death. We therefore assume that elements of the TrkA-activated signaling pathway are not affected by Sema3A stimulation. To test this idea we monitored the levels of phospho-Akt, a downstream signaling cascade of TrkA (Tyr751, the PI3K binding site), and phospho-ERK42/44, another downstream signaling cascade of TrkA (regulated by Tyr490 and Tyr785). We monitored the activation levels of Akt (phosphorylations at Thr308 and at Ser473) and activation levels of ERK42/44 (phosphorylations at Thr202 and Tyr204 of Erk1 or Thr185 and Tyr187 of Erk2) by Western blots (Fig. 5). Neurons were grown for 20 h in the presence of 10 ng/ml NGF. The cells were then re-fed with fresh medium and NGF, and incubated for 60 min before addition of Sema3A. Neurons were then harvested at 0, 0.5, 2, 4 and 8 h following Sema3A stimulation (t = 0 is 60 min following re-feeding of neurons with fresh NGF). Sema3A is unable to suppress Akt activation suggesting that the TrkA-induced PI3K pathway is not affected by Sema3A (Fig. 5A-C). Since Akt is a major modulator of neuronal survival, this result is consistent with our previous findings, which showed that NGF is able to suppress Sema3A-induced death [9].

[10] 59w To further test the idea of differential inhibition of TrkA by Sema3A we also tested the levels of phospho-ERK42/44 following Sema3A stimulation (Fig. 5D-F). In contrast to the phospho-Akt levels, phospho-ERK44/42 is suppressed by Sema3A stimulation. This result is consistent with the ability of Sema3A to suppress the TrkA phosphorylation sites Tyr490 and Tyr785, both involved in ERK activation.

[11] 140w Our results indicate that TrkA undergoes rapid dephosphorylation following Sema3A stimulation. We wanted to test whether this TrkA inactivation is necessary for Sema3A-induced inhibition of axon growth. To this end, we generated constitutively active TrkA by replacing Y683 and Y684 with E683 and D684 (rat nomenclature) in the activation loop [24]. Constitutively active HA-TrkA YY683-4ED , tyrosine dead HA-TrkA K538A (a mutated form that lacks tyrosine kinase activity) and the unrelated receptor HA-Caspar2 (data not shown) were nucleofected into dissociated primary DRG neurons (Fig. 6). The neurons were then grown for 20 h before stimulation with or without Sema3A for 6 h. The results show that the expression of constitutively active TrkA completely blocked the effects of Sema3A on axon growth inhibition. In contrast, inactive TrkA or Caspar2 had no effect on the ability of Sema3A to inhibit neurite growth.

[12] 284w It has been shown that sensory axons in Sema3A null mice enter tissues they normally avoid in wild-type animals [25]. Since we show in vitro that Sema3A is also a growth modulator, we assumed that axons in Sema3A null mice would exhibit overgrowth in addition to the effects of its guidance defects. To Fig. 3. Sema3A inhibits neurite growth in an NGF-dependent manner. DRG neurons from E12.5 BAX mutant embryos were cultured in the absence of NGF for 20 h. At this time, neurons were re-fed with fresh medium either without NGF or with 10 ng/ml NGF (A, B, C). A, B) Neurite outgrowth assay. Neurons were cultured for an additional incubation period of 6 h in the presence of 30 pM Sema3A (empty bars), or without Sema3A (grey bar). The neurons were then fixed and stained with anti-TrkA antibody (to monitor similar neuronal populations as been used in Figs. 1 and 2). A) Representative examples of neurons for each condition are shown. B) To calculate neurite growth the average length at t = 0 was subtracted from the average length at t = 6. The average growth of neurites in the presence of NGF and the absence of Sema3A was defined as 100% growth. Percent growth results represent the mean +/-s.e.m. of three independent experiments (n N 70 neurons for each treatment, see Experimental/materials and methods for details). C) Growth cone collapse assay. Explants were treated with (empty bars) or without (gray bars) 30 pM Sema3A. After an additional incubation period of 40 min the explants were fixed and stained with rhodamine phalloidin. Growth cone collapse results represent the mean +/-s.e.m. of three independent experiments (n N 1000 growth cones for each treatment).

[13] 132w test this we used whole-mount staining to visualize the total coverage of axons in wild-type and mutant mice. We then measured the total coverage of axons in particular areas such as the limbs and around the eye. We used stereological principles to calculate the total axon converge per embryo area (Fig. 7). We assumed that if Sema3A functions only as a guidance molecule, the total axon coverage should not be significantly different, although some axons will enter inappropriate regions of the embryo. If, however, growth rate in the absence of Sema3A is also significantly accelerated, the total coverage should increase significantly. Indeed, we found the total axon length in the absence of Sema3A to be dramatically higher than in wild-type mice, a finding consistent with a role for Sema3A in growth regulation.

[14] 151w NGF has been shown to act as a modulator of Sema3A. For instance, Sema3A-mediated repulsion is attenuated by increasing NGF concentrations before the addition of Sema3A [10,11]. Sema3A is also a potent cell death inducer for embryonic DRG neurons [9]. Again, similar to repulsion, this effect is blocked at higher concentrations of NGF [9]. In contrast to these modulatory effects of NGF on repulsion and death, elevated concentrations of NGF are not able to block Sema3A-mediated axon growth inhibition. In fact, contrary to repulsion and cell death, it seems that Sema3A requires stimulation by NGF to act as an efficient growth inhibitor. This conclusion is based on the experiment in which Sema3A caused weak neurite outgrowth inhibition in cultures without NGF (Fig. 3A,B). From these results it is apparent that Sema3A is a modulator of NGF-induced neurite growth activity, rather than NGF acting as a modulator of Sema3A-induced axon growth inhibition.

[15] 100w Axon growth is dependent on different factors, including growth factors and extracellular matrix proteins (for instance, [27]). Our results indicate that Sema3A-induced axon growth inhibition is dependent on the identity of the growth stimulator, and may indicate that there is a specific relationship between Sema3A and NGF. NGF is a growth inducer, and neurite outgrowth inhibition by Sema3A therefore indicates its ability to modulate NGF signaling. There appears to be a complex interregulation of NGF and Sema3A, in which at high concentrations of NGF and low concentrations of Sema3A, Sema3A attenuates NGF-induced growth, while NGF attenuates repulsion induced by Sema3A.

[16] 113w Sema3Awas shown to be a potent modulator of axon guidance in vitro. The detection of inappropriate and premature invasion of axons into some tissues in Sema3A null mice is consistent with the possibility that Sema3A acts as a guidance molecule in vivo. However, guidance errors alone cannot explain the massive increase in total axon coverage detected in Sema3A null mice (Fig. 6). Our in vitro analysis, which demonstrates axon growth regulation independent of Sema3A-guidance activity, together with our in vivo demonstration of a massive enhancement in total axon coverage, suggests that it is most likely that Sema3A functions as a growth regulator in vivo in addition to its function as a guidance molecule.

[17] 93w In conclusion, this work demonstrates that axon growth regulation is an important part of Sema3A activity in vitro, and contributes to the formation of accurate patterns of axonal connectivity in vivo. In contrast to guidance activity, axon-growth inhibition is mediated by regulating the activation levels of a subset of the TrkA signaling cascade, causing a rapid dephosphorylation of PLCγ and SHC docking sites, which mediates axon growth. In contrast, Sema3A has no apparent affect on the PI3K site of TrkA (as determined indirectly by the levels of activated Akt) that regulates neuronal survival.

[1] 87w Sema3A is best known as a guidance molecule, in most cases acting as a growth cone repellent signal [26]. The inhibitory effect of Sema3A on axon growth has also been noted previously [8]. However, whether axon growth regulation is a side effect of repulsion or an independent property of this molecule was not determined. In this study we tested the role of Sema3A as a growth regulator, and whether this function is dependent on repulsion. The molecular mechanism by which Sema3A modulates axon growth was also investigated.

[1] 126w NGF was obtained from Sigma (St. Louis, MO, USA). The anti-HA, antiphospho-TrkA490, anti-phospho-Akt (Ser473), anti-phospho-Akt (Thr308) and anti-phospho-Erk42/44 (Thr202/Tyr204) antibodies were purchased from Cell Signaling Technology (Danvers, MA, USA). Anti-neurofilament 2H3 (developed by Thomas M. Jesell and Jane Dodd) was obtained from the Developmental Studies Hybridoma Bank, developed under the auspices of the NICHD and maintained by the University of Iowa, Department of Biological Sciences, Iowa City, IA, USA 52242. Anti-TrkA (RTA) was generously provided by Dr. Reichardt, University of California, USA). Anti-phospho-TrkA794 (the PLCγ site in Rat, equivalent to Tyr785 in human) was generously provided by Dr. Moses Chao, Skirball Institute of Biomolecular Medicine, New York University School of Medicine, New York, NY. Secondary antibodies were obtained from Jackson Immunoresearch Laboratories, Inc. (Jackson, PA, USA).

[2] 65w Total protein dye 5-(4,6-dichlorotriazin-2-yl) aminofluorescein (DTAF) was obtained from Sigma (St. Louis, MO, USA). Protease inhibitor cocktail was obtained from Roche Diagnostics (Mannheim, Germany), Matrigel from BD Biosciences (Franklin Lakes, NJ, USA), and rhodamine phalloidin and Opti-MEM from Invitrogen (Carlsbad, CA, USA). Tissue culture reagents were purchased from Biological Industries (Kibbutz Beit Haemek, Israel). All other reagents were purchased from Sigma (St. Louis, MO, USA).

[3] 20w The following plasmids were generous gifts from other researchers: pCDNA-HA-TrkA from Dr. Moses Chao; and pCDNA3-HA-Caspar2 from Dr. Elior Peles.

[4] 101w ICR mice were obtained from Harlan Laboratories. Pregnant mice were obtained following overnight mating (day of vaginal plug is defined as embryonic day 0.5). We used E12.5 and E13.5 embryos. The Sema3A-null mice have been described previously [12]. Sema3A embryos were genotyped using the following PCR primers: 5′ TGATGGCGAAAAGACTGTGT, 5′ CACACGCACAGAG-GAATC and 5′ ACCAAATTAAGGGCCAGCTC. BAX null mice were obtained from the Jackson laboratory [13]. BAX null embryos were genotyped using the following PCR primers: 5′-GTT GACCAGAGTGGCGTAGG, 5′-CCG-CTTCCATTGCTCAGCGG, and 5′-GAGCTGATCAGAACCATCATG. Animal handling adhered strictly to national and institutional guidelines for animal research, and was approved by the Ethics Committee of our institution.

[5] 46w DRG explants were tested for growth cone collapse as previously described [9,14]. Each experiment included 3-4 wells (100-200 growth cones per well) and was repeated using 3-4 embryos for each treatment. An observer who was blind to the identities of the treatments counted the growth cones.

[6] 39w Mouse E13 DRG neurons were electroporated using mouse Nucleofector kit (Amaxa). Transfection of the neurons was carried out according to the manufacturer's protocol, using program G13. We used the following plasmids: pCDN3-HA-TrkA K547A , pCDN3-HA-TrkA YY683,684ED , and pCDN3-HA-Caspar2.

[7] 9w Western blot analysis was performed as previously described [9].

[8] 9w Partially purified Sema3A was prepared as previously described [9].

[9] 140w Whole-mount immunostaining with 2H3 was performed essentially as described [15]. In brief, embryos were fixed for 2 h in 4% paraformaldehyde and 10 mM PBS (pH 7.0), washed three times in PBS + 0.2% Triton X-100 and dehydrated through a methanol series. Endogenous peroxidase activity was quenched overnight at 4 °C with 80% MeOH and 3% H 2 O 2 , followed by rehydration through a methanol series. After washing for 3 h in a Tris-buffered saline solution containing 0.2% Triton X-100 (PBST), the embryos were incubated with the 2H3 antibody (1:100 dilution of 2H3 hybridoma culture supernatant with PBST-containing 2% skim milk) for 3-4 days at 4 °C. The embryos were then incubated for one day with HRP-coupled anti-mouse Ig antibody (Jackson; 1:200 dilution in PBST containing 2% skim milk, 2% goat serum). HRP activity was detected with diaminobenzidine.

[10] 243w DRG neurons were fixed in 4% paraformaldehyde and 10% sucrose. Cells were incubated overnight at 4 °C with primary antibodies diluted in PBS containing 5% BSA, 0.1% Triton X-100. The cells were washed on the following day and then incubated with Cy5-(1:100) antibody (Jackson). Neurons were counterstained with the non-specific total protein dye DTAF. Staining was analyzed using a laser-scanning confocal microscope (Olympus IX70). Images were captured using a 60x oil objective (NA 1.35) at room temperature and confocal acquisition software (Fluoview, Olympus, Hamburg, Germany). The interval between imaged optical sections was 1 μm. Sequential scanning for Cy5 (Em: 565-660 nm) and DTAF (Em: 510-530 nm) was performed. Laser intensity and acquisition parameters were identical for each treatment. For quantifications, fluorescent images of the same growth cones were captured in parallel for both TrkA antibody and DTAF treatment. The amount of fluorescence within the area of the growth cone was calculated digitally for each channel using ImageJ software, giving the mean fluorescence intensity per unit area. The mean levels of background fluorescence of adjacent areas were similarly calculated and subtracted from the growth cone value to yield the corrected intensity measurements. We then divided the values obtained from each TrkA staining by the values obtained from the same growth cone in the DTAF channel to calculate the normalized intensity for each growth cone. The fluorescence intensities were calculated for 20 growth cones under each condition. Each experiment was repeated at least three times.

[1] 99w DRG explants were cultured for 20 h. At this time (t = 0), medium was changed and the explants were grown for an additional period of 6 h with or without Sema3A. Explants were photographed at t = 0 and after 6 h. In order to monitor a large number of neurites, we divided each explant into four quarters and photographed these same quarters at t = 0 and t = 6 h. The lengths of the neurites were measured using Image-Pro software, and the differences between the lengths at t = 0 and t = 6 were calculated.

[2] 142w We used dissociated neuron cultures to test neurons from BAX null mice and TrkA-transfected cells. In these experiments, cells were cultured for 20 h. At this time (t = 0), some cultures were fixed, while in the rest of the cultures the medium was changed and the neurons were cultured for an additional period of 6 h with or without Sema3A. At this time cultures were fixed and immunostained (with anti-HA in transfection experiments, or anti-TrkA in the BAX experiments), and thirty fields were then photographed (systematically recording one field after the other). An observer who was blind to the identities of the cultures photographed the neurons. Twenty-five to thirty neurons were measured for each treatment, and each experiment was repeated three times. Neurite length was measured using NeuronJ, an ImageJ plug-in (a neurite tracing and quantification program written by Erik Meijering).

[3] 98w Images of embryos were obtained using an SZX12 stereomicroscope (Olympus, Hamburg, Germany) equipped with a C-4040 digital camera (Olympus, Hamburg, Germany). To estimate the degree of growth we measured the total axon length using stereological principles, essentially as described [16,17]. In brief, the projection of the neuronal network in each limb was superimposed with a square grid (a set of parallel lines) using Adobe Photoshop software. The total axon length was estimated using Buffon's needle problem equation: L =(πd/M2) I (L-axon length, d-distance between the grid lines, I-number of intersections between the axons and the grid lines, M-Magnification).

[4] 172w We show that axon growth inhibition by Sema3A is not a simple consequence of growth cone repulsion. First, at low concentrations of Sema3Awhich are sufficient to generate collapse of only a minority of growth cones (about 40%), maximal axon growth inhibition is already observed (growth arrest of neurites is almost 100%). Therefore, about 60% of neurites stop growing without their growth cones having collapsed. Second, elevated concentrations of NGF are able to reduce Sema3A-induced repulsion by about 37%, while axon growth inhibition is not reduced. From this experiment it is clear that NGF can negatively modulate Sema3A-induced repulsion, but cannot affect its axon growth inhibition. Finally, when growing neurons from BAX null mice in the absence of NGF (circumventing the association between growth factors and survival dependence), Sema3A functions as a strong repellent molecule, but its effects on growth are very weak. Taken together, it is apparent that Sema3A-induced repulsion and axon growth inhibition are two separate functions, and are likely mediated by different signal transduction cascades which remain to be investigated.

[5] 554w Our results indicate that induction of axon growth by NGF is necessary for Sema3A to effectively block axon growth. This indicates that Sema3A is somehow able to specifically block the signaling of NGF-induced growth. This effect may be mediated by directly modulating the NGF receptor or by affecting the signaling cascade downstream to the NGF receptor. To begin Fig. 4. Sema3A inhibits TrkA activation. A) DRG neurons were cultured in the presence of 10 ng/ml NGF for 20 h. At this time neurons were re-fed with fresh medium containing 10 ng/ml NGF, and left to incubate for an additional 1 h. At the end of this incubation period (defined as t = 0), medium with or without 30 pM Sema3A was added. Protein was extracted at the indicated time points (min) (see diagram in upper panel). Relative changes in the phosphorylation levels of p-Trk Y490 were measured by western blot analysis using a phospho-specific antibody. To determine protein levels each membrane was re-blotted for actin. Each experiment was repeated three times and a representative result is shown. B) Quantification of bend intensity using scanning densitometry (Quantity One, BioRad) of three blots representing three different experiments. Results were normalized to actin. Control at time 5′ was defined as 100%. C) DRG explants were cultured in the presence of 10 ng/ml NGF for 20 h. At this time, neurons were either fixed (Low NGF) or re-fed with fresh medium containing 40 ng/ml NGF (High NGF) and incubated for an additional period of 30 min. At this point, 6 pM Sema3A was added for 1 or 5 min (see diagram in upper panel). Neurons were fixed, stained with phospho-TrkATyr785 or anti-TrkA, and counterstained with DTAF for total protein labeling (data not shown). exploring these possibilities we began by testing the activation levels of the NGF receptor. The receptor for NGF comprises TrkA and p75NTR. Since many studies have shown the importance of TrkA in mediating axon growth, we focused on this receptor. Following NGF binding, TrkA activation involves dimerization and auto-phosphorylation of ten evolutionarily conserved tyrosines, of which three, Y670, Y674 and Y675 (human TrkA sequence nomenclature), are present in the autoregulatory loop of the kinase domain that controls tyrosine kinase activity. Phosphorylation of the other tyrosine residues promotes 5. Sema3A is able to suppress only part of the TrkA-activated-signaling cascade. DRG neurons were cultured in the presence of 10 ng/ml NGF for 20 h. At this time neurons were re-fed with fresh medium containing 10 ng/ml NGF, and left to incubate for an additional 1 h. At the end of this incubation period (defined as t = 0), medium with or without 30 pM Sema3A was added. Protein was extracted at the indicated time points (h). Relative changes in the phosphorylation levels of p-Akt Ser473, p-Akt Thr308 (A) or p-ERK42/44 (D) were monitored by western blot analysis using phospho-specific antibodies. To determine protein levels, each membrane was re-blotted for actin and total Akt (in the case of Akt) or total ERK. Each experiment was repeated three times and a representative result is shown. Quantification of band intensity for p-Akt Ser473 (B), p-Akt Thr308 (C), p-ERK42 (E) and p-ERK44 (F) using scanning densitometry (Quantity One, BioRad) of three blots representing three different experiments (mean +/-s.e.m.) are shown. Results were normalized to actin (B,C) or total ERK (E,F).

[6] 214w signaling by creating docking sites for adapter proteins containing phosphotyrosine-binding (PTB) or src-homology-2 (SH-2) motifs (reviewed in [21]). Our results show that Sema3A is able to accelerate the dephosphorylation of TrkA at two major sites: Y490 and Y785. Another important regulatory site of TrkA is Tyr751, a PI3K activation site, an important regulator of the pro-survival cascade mediated by Akt. The inability of Sema3A to inhibit active Akt levels indicates that Sema3A is able to suppress only a specific subset of TrkA signaling pathways. These finding in conjunction with the results showing that constitutively active TrkA is able to block the inhibitory activity of Sema3A on axon growth, it is likely that Sema3A acts upstream to the TrkA Neurite length was measured at t = 0 and t = 6 for each construct (see Experimental/materials and methods for details). The average length at t = 0 was subtracted from the average length at t = 6 in order to calculate growth. Cultures transfected with pCDN3-HA-TrkA TrkA K538A and incubated without Sema3A were defined as 100%. All other conditions were normalized accordingly. Neurons treated with (empty bar) or without Sema3A (gray bars). No statistical difference was found between control and Sema3A treatment in neurons transfected with pCDN3-HA-TrkA YY683/684ED , (p = 0.8, two-tailed t test).

[7] 51w receptor. In principle Sema3A may exert its inhibitory effect on TrkA signaling through modulation of NGF binding, TrkA dimerizationor activation/recruitment of a TrkA targeted phosphatase. Our findings showing that Sema3A inhibits only specific elements of NGF-induced TrkA signaling indicate that the recruitment of a specific phosphatase is the most likely mechanism.

[8] 117w Interestingly, our results are very different from those reported previously with respect to Sema3F-induced effects in sympathetic neurons. Here, it was shown that the addition of Sema3F has no effect on phospho-TrkA, though it is able to suppress the activation of ERK and Akt [28]. These differences are intriguing, although it is important to note the significant differences between the two systems, which include different ligands (Sema3A versus Sema3F) and receptors (NP1/PlexinA4 versus NP2/PlexinA3) [29][30][31]. Moreover, additional differences in signaling between the two semaphorins have been reported, including the involvement of cGMP and Rho-kinase in Sema3A-induced DRG collapse, in contrast to the lack of cGMP and Rho-kinase involvement in the case of Sema3Finduced sympathetic neuron collapse [28,32].