Abstract
Learning and memory depend on morphological and functional changes to neural spines. Non-muscle myosin 2b regulates actin dynamics downstream of long-term potentiation induction. However, the mechanism by which myosin 2b is regulated in the spine has not been fully elucidated. Here, we show that filamin A-interacting protein (FILIP) is involved in the control of neural spine morphology and is limitedly expressed in the brain. FILIP bound near the ATPase domain of non-muscle myosin heavy chain IIb, an essential component of myosin 2b and modified the function of myosin 2b by interfering with its actin-binding activity. In addition, FILIP altered the subcellular distribution of myosin 2b in spines. Moreover, subunits of the NMDA receptor were differently distributed in FILIP-expressing neurons and excitation propagation was altered in FILIP-knockout mice. These results indicate that FILIP is a novel, region-specific modulator of myosin 2b.
Similar content being viewed by others
Introduction
Filamentous actins (F-actins) are important structural components that, together with a variety of actin-binding proteins, underlie a broad range of cellular responses. For example, F-actins and their binding partners play pivotal roles in cell motility and directionality at the leading edge of migrating cells1. In addition, actin-binding proteins dynamically regulate the structure of neural spines, which are actin-rich protrusions on neurons and major sites for excitatory synaptic transmission involved in learning and memory2,3. To date, more than 100 actin-binding proteins have been identified; muscle-type myosin and filamin were among the first such proteins to be characterised, whereas non-muscle myosins, including myosin 2b, are emerging members of this group and have attracted much attention because of their essential involvement in diverse fundamental cellular functions, spine structure dynamics and learning and memory4,5,6,7. Although there is much interest in the mechanisms of actin-binding protein regulation, which is critical for essential cellular functions, these mechanisms have not yet been fully elucidated. We previously identified a novel filamin A binding protein, filamin A-interacting protein (FILIP or FILIP-1) and demonstrated that FILIP participates in actin dynamics by accelerating the calpain-dependent degradation of filamin A8, which, when mutated, causes human migration disorder9,10. We have demonstrated that FILIP, whose mRNA is localised in the ventricular zone of the cortex during the perinatal stage, is potentially involved in radial migration in the cortex by introducing exogenous FILIP into the ventricular zone8. However, the in vivo role of FILIP, especially in adults, has not yet been clarified. Here, we generated a FILIP-knockout mouse to address these issues. Unexpectedly, we found that FILIP is likely to regulate spine structure by modulating the activity of myosin 2b.
Learning and memory are dependent on the activity and morphology of the neural spines11. Morphological changes to a spine depend on actin dynamics12. In the hippocampus, myosin 2b is a key regulator of the changes in spine morphology related to learning and memory4,6. However, how and why neural spine morphology varies across the cortex remains unknown, especially in response to learning. During the learning response, spine enlargement is observed during long-term potentiation (LTP) in the hippocampus11, whereas learning induces a reduction in the volume of the spine head in the piriform cortex13. Although the mechanisms underlying such differences have not been fully elucidated, our data suggest that FILIP, which is expressed in the piriform cortex but not in the hippocampus, is one of the molecules responsible for these differences.
Results
Generation of a FILIP-knockout mouse
To investigate the function of FILIP, FILIP was disrupted through homologous recombination. In embryonic stem cells, a 2.8-kb genomic fragment containing a portion of exon 5 (the largest exon of FILIP) was replaced with in-frame β-galactosidase and neomycin resistance genes (see Supplemental Fig. S1a online). The chimeras were backcrossed with C57BL/6 mice to produce FILIP-heterozygous mice. The appearance of and histological samples from the FILIP-heterozygous (FILIP+/−) mice were indistinguishable from those of their wild-type (FILIP+/+) littermates. The homozygous (FILIP−/−) mice were indistinguishable from their normal littermates in terms of appearance. The disruption of the FILIP gene and absence of FILIP protein were confirmed by northern blot and western blot analyses (see Supplemental Fig. S1b, c online). Although two alternatively spliced forms of FILIP, a long form (L-FILIP) and a short form (S-FILIP), were observed in the rat, only one form, corresponding to the long form of FILIP, was detected in mice. Hereafter, we refer to this form as FILIP instead of L-FILIP.
Targeting of FILIP revealed limited localisation of FILIP in the brain
Because the mutant allele conferred β-galactosidase expression under the control of the FILIP promoter (see Supplemental Fig. S1d online), we examined the distribution of FILIP-expressing cells in the adult mouse brain by visualising β-galactosidase activity. With the exception of modestly accumulated neurons in the upper cortical layer of the FILIP+/− mice, no obvious difference was observed in terms of the distribution between FILIP+/− and FILIP−/− mice (see Supplemental Fig. S2a–f online).
Many β-galactosidase-positive cells were detected in the forebrain, particularly in the ventral portion and in the deep nuclei (Fig. 1 and Supplemental Fig. S2a–f online), including the anterior olfactory nucleus, the piriform cortex and the olfactory tubercle as well as in the nucleus accumbens, the globus pallidus and the amygdaloid complex. In addition, positive cells were observed in the neocortex, especially in the visual and the motor cortices (Fig. 1). In contrast, β-galactosidase-positive cells were rarely observed in the hippocampus. In the diencephalon, β-galactosidase activity was faint, except in cells in the arcuate nucleus (Fig. 1f and Supplemental Fig. S2c online).
FILIP is expressed in glutamatergic neurons
β-galactosidase-positive cells were positive for the neuronal markers NeuN and MAP-2 (see Supplemental Fig. S2g online)14,15 and for Brn1, which is expressed principally in neurons of layers II–V16, the visual cortex and the piriform cortex (see Supplemental Fig. S2h online). We further investigated the types of neurons that express FILIP. FILIP containing neurons were not positive for GAD67 in the piriform cortex (see Supplemental Fig. S2g online), indicating that FILIP is expressed in glutamatergic neurons but not in GABAergic neurons.
FILIP regulates spine length
We next investigated whether FILIP controls neuronal morphology. Neurons were visualised using the Golgi-Cox staining method and the lengths of the spines (the distance between the spine neck close to the dendrite and the tip; Fig. 2a) on the apical dendrites of the layer II pyramidal neurons (superficial pyramidal neurons) in the piriform cortex were measured. The mean length of the spines was shorter in FILIP−/− mice (mean ± s.d., 1.00 ± 0.16 μm; n = 15) than in the control FILIP+/− and FILIP+/+ littermates (1.17 ± 0.16 μm; n = 19; Fig. 2b–c; Student's t-test, two-tailed, p = 0.00358). The proportion of the stubby, mushroom and thin types of spines is indicative of spine maturation17,18. The measured differences of spine types did not achieve significance (Fig. 2d). Then, we constructed an inducible knockdown vector for FILIP and transfected the piriform neurons with this vector (Fig. 2e). We found that the mean spine length of FILIP-knockdown neurons was shorter (0.95 ± 0.09 μm; n = 20) than that of control neurons (1.06 ± 0.06 μm; n = 12; Fig. 2f–g; Student's t-test, two-tailed, p = 0.00059). We then studied the influence of FILIP knockdown on the spine morphology with primary cultured piriform neurons (Fig. 2h). We observed that the mean spine length of FILIP-knockdown neurons (1.29 ± 0.84 μm; 398 spines in 5 neurons) was significantly shorter than that of control neurons (1.53 ± 0.91 μm; 432 spines in 5 neurons; Fig. 2i; Student's t-test, two-tailed, p = 0.00011). We also found that the estimated spine volume of FILIP-knockdown neurons (0.33 ± 0.41 μm3; 241 spines in 5 neurons) was significantly larger than that of control neurons (0.22 ± 0.24 μm3; 326 spines in 5 neurons; Fig. 2j; Wilcoxon rank sum test, p = 0.00015).
FILIP was not expressed in pyramidal neurons in the hippocampus (Fig. 1f–h and Supplemental Fig. S2c–e online). Therefore, these hippocampal neurons can be used as an ideal system to study the role of FILIP through ectopic expression studies. Hippocampal neurons were taken from E17.5 mice, cultured for 20 days in vitro (DIV20) and transfected with an expression vector for Asp-Tyr-Lys-Asp-Asp-Asp-Asp-Lys (FLAG)-tagged FILIP (FILIP-FLAG). Two days after transfection, we examined the localisation of exogenous FILIP based on FLAG staining. FILIP was identified in the spines as well as in dendrites (Fig. 3a). We further investigated whether the ectopic expression of FILIP influences spine morphology. We transfected the FILIP expression vector into hippocampal neurons at DIV17; 3 days later, we measured the spine length and volume (Fig. 3b–e). The mean spine length in the neurons with FILIP was longer (0.97 ± 0.40 μm; 912 spines in 13 neurons) than in those without exogenous FILIP (0.87 ± 0.38 μm; 843 spines in 12 neurons; Fig. 3d; Student's t-test, two-tailed, p < 0.00001). The spine head volume was reduced compared with spines without FILIP (0.19 ± 0.20 μm3; 433 spines from 11 FILIP-expressing neurons and 0.27 ± 0.27 μm3; 397 spines from 11 control neurons; Fig. 3e; Wilcoxon rank sum test, p < 0.00001).
FILIP binds to non-muscle myosin heavy chain IIb
Although FILIP binds to filamin A and is involved in controlling filamin A degradation8, filamin A expression decreased in the postnatal telencephalon despite its high expression during the embryonic period (Fig. 4a). Recently, it has been reported that small amounts of filamin A are expressed in the dendrites and cell soma but not spines of adult neurons19. Therefore, we examined how FILIP exerts its function in the absence of filamin A in the spine. We determined that non-muscle myosin heavy chain IIb (NMHC IIb) is a novel FILIP-binding partner. In general, non-muscle myosin is composed of 2 heavy chains and 4 light chains20. NMHC IIb is a heavy chain of myosin 2b and is essential for the actin-binding and ATPase activity of myosin 2b5. FILIP-FLAG and NMHC IIb in COS-7 cells were successfully co-immunoprecipitated using an anti-NMHC IIb antibody (Fig. 4b). The NMHC IIb binding sites on FILIP were investigated using several fragments of FILIP (Fig. 4c, d). The fragment containing amino acids (aa) 1–652 of FILIP did not co-immunoprecipitate with NMHC IIb; however, the 687–1212 aa fragment of FILIP did co-immunoprecipitate with NMHC IIb (Fig. 4d). Therefore, the NMHC IIb binding site was assigned to the C-terminal region between 687–1212 aa of FILIP. Because the 960–1212 aa fragment of FILIP did not co-immunoprecipitate with NMHC IIb (Fig. 4d), the 687–960 aa region of FILIP is likely to be important for binding to NMHC IIb. Conversely, FILIP-FLAG co-immunoprecipitated with the 1–782 aa and 1–331 aa regions of NMHC IIb but not with the 90–331 aa region (Fig. 4e–g). The myosin heavy chain has a globular head domain that is critical for the contractive activity of myosin 2b. Biochemically, the globular head domain of NMHC IIb consists of the myosin N-terminal, the upper 50 kDa, the lower 50 kDa and the converter subdomain; crystal structure analyses indicate that ATP binds in a pocket between the upper 50 kDa and the N-terminal subdomain located within the 1–90 aa region of NMHC IIb20. Our data indicate that FILIP binds to the globular head domain near the ATP binding site that controls the conformational change of NMHC IIb and is needed for the contractile cycle, which consists of dissociation and binding to F-actins.
FILIP influences the subcellular distribution of myosin 2b
Because FILIP binds to the head domain of NMHC IIb, which is essential for the binding of NMHC IIb to actin fibres, it is likely that FILIP modifies the ability of myosin 2b to bind to F-actins. First, we determined how FILIP is involved in the interaction of myosin 2b with F-actins using COS-7 cells. Because F-actin is capable of forming stress fibres in COS-7 cells, the use of COS-7 cells allowed us to investigate the influence of FILIP on the interaction of myosin 2b with F-actin. In terms of myosin 2b distribution, we classified cells into the following two categories: cells with stress fibre-like distribution and cells with granular distribution. “Stress fibre-like distribution” indicates that the cells have thick and long fibre-like structures of NMHC IIb and “granular distribution” indicates that NMHC IIb exhibits a particle-like localisation without thick fibre-like structures. Whereas the stress fibre-like distribution of myosin 2b was dominant in COS-7 cells without endogenous FILIP, the number of cells with granular myosin 2b distribution surpassed that of cells with stress fibre-like distribution in the presence of FILIP (see Supplemental Fig. S3a, b online). We next investigated whether FILIP influenced the formation of actin stress fibres, as myosin 2 is a component of actin stress fibres21,22. The number of cells with actin stress fibres decreased in the presence of FILIP (see Supplemental Fig. S3c, d online). Because FILIP binds to the N-terminal domain of myosin 2b, which is essential for binding to actin fibres and the binding of myosin 2b to actin fibres is essential for the function of myosin 2b, we hypothesised that FILIP interferes with the activity of myosin 2b. We then compared the appearance of NMHC IIb in COS-7 cells in the presence of the myosin 2b inhibitor blebbistatin. Blebbistatin binds to the myosin globular head domain and blocks the rigid binding of myosin 2b to actin23. A similar “granular distribution” of NMHC IIb was observed in the presence of blebbistatin (see Supplemental Fig. S3e, f online).
FILIP interferes with the binding of myosin 2b to F-actins
If our hypothesis that FILIP interferes with myosin 2b binding to F-actins is correct, the amount of myosin 2b associated with F-actins should decrease in the presence of FILIP. F-actins exist in equilibrium with free globular (G)-actins. Because actins in the cytosol can be fractionated into F-actin-rich (Triton-soluble (TS) fraction in Supplemental Fig. S4a online) and G-actin-rich fraction (cytosolic (CS) fractions in Supplemental Fig. S4a online), we examined the amounts of NMHC IIb in these two fractions. We found that the amount of NMHC IIb in the G-actin-rich fraction (free NMHC IIb) increased in the presence of FILIP, whereas the amount of NMHC IIb in the F-actin-rich fraction decreased (see Supplemental Fig. S4b online), suggesting that less NMHC IIb is associated with F-actins in the presence of FILIP. These results indicated that FILIP controlled the binding of myosin 2b to F-actin. As FILIP controls the degradation of filamin A, we investigated whether the amount of myosin 2b was also influenced by FILIP expression in our culture conditions. We found that the total amounts of myosin 2b and filamin A did not change much after FILIP expression in some culture cells (see Supplemental Fig. S5a online). However, we found that the amount of NMHC IIb increased by approximately 40% in the hearts of FILIP−/− mice (see Supplemental Fig. S5b online), suggesting that deletion of FILIP increased the amount of NMHC IIb. We also found that the mean intensity of immunostaining of myosin 2b was significantly increased in the primary cultured piriform neurons of the FILIP−/− mice (187.5 ± 23.6; 44 neurons) compared with that of the control mice (149.8 ± 32.0; 44 neurons; Welch's t-test, two-tailed, p < 0.00001; see Supplemental Fig. S5c online). This change in the intensity of NMHC IIb in FILIP-expressing neurons was also observed in the FILIP+/− piriform cortex (see Supplemental Fig. S5d online).
FILIP influences the subcellular distribution of myosin 2b in spines
It has been demonstrated that myosin 2b localises to the lower part of the spine head and the spine neck7,24 and that the inhibition of myosin 2b leads to spine elongation and reduced spine head volume4,7. We performed myosin 2b knockdown and obtained similar results (see Supplemental Fig. S6 online). Therefore, it is likely that FILIP modulates spine morphology through myosin 2b. We then investigated whether the subcellular localisation of NMHC IIb was altered in spines in the presence or absence of FILIP. We transfected morphologically pyramidal neurons that had been isolated and cultured from the piriform cortices of the FILIP−/− mice and those of the control littermates with tdTomato as a volume marker and an EGFP-tagged NMHC IIb and analysed the subcellular distribution of exogenous NMHC IIb. The subcellular distribution of NMHC IIb correlated well with the tdTomato signals in the spine of control mice but not in that of the FILIP−/− mice (Fig. 5a–c). We also investigated the distribution of internal NMHC IIb in the piriform cortex neurons taken from FILIP−/− mice and found that the number of spines with accumulated signals of NMHC IIb increased in the neurons from the FILIP−/− mice (Fig. 5d, e). These results indicate that FILIP altered the subcellular distribution of myosin 2b from its proximally accumulated pattern to an ubiquitous localisation.
Overexpression of NMHC IIb inhibits the spine elongation induced by exogenous FILIP expression
We observed that exogenous FILIP influenced endogenous NMHC IIb distribution in the spine of hippocampal neurons (Fig. 6a, b). If we assume that exogenous FILIP elongated the spine through binding to myosin 2b and interfering with its function, then exogenous NMHC IIb should be able to rescue the influence of FILIP on spine morphology. We transfected the FILIP expression vector and NMHC IIb expression vector into primary cultured hippocampal neurons and examined spine length. We found that the exogenous expression of NMHC IIb inhibited the spine elongation caused by exogenous FILIP (Fig. 6c, d).
FILIP intermingles with NMDA receptor-mediated signalling
Experimentally, glycine is applied to neurons to enhance the activation of NMDA receptors25,26. We therefore investigated whether FILIP influences spine morphology in the presence of glycine. In these experiments, we performed morphological studies on hippocampal neurons that had been cultured for 2 weeks (DIV15), which is when spines are in the maturation process in vitro27 (Fig. 7a). Without glycine, the spine length and spine head volume were not significantly different regardless of FILIP (Fig. 7b–d). In contrast, the spine length decreased in the presence of FILIP when glycine was applied (1.15 ± 0.82 μm, 1513 spines from 11 control cells/glycine (−); 1.14 ± 0.82 μm, 1438 spines from 11 control cells/glycine (+); 1.20 ± 0.73 μm, 1421 spines from 11 FILIP-expressing cells/glycine (−); 1.05 ± 0.71 μm, 1736 spines from 11 FILIP-expressing cells/glycine (+); Fig. 7c). Interestingly, the application of glycine resulted in an increase in the spine head volume and the degree of this enlargement was enhanced by FILIP expression (0.144 ± 0.202 μm3, 1245 spines from 11 control cells/glycine -; 0.195 ± 0.276 μm3, 1241 spines from 11 control cells/glycine +; 0.161 ± 0.193 μm3, 1207 spines from 11 FILIP-expressing cells/glycine + and 0.233 ± 0.300 μm3, 1451 spines from 11 FILIP-expressing cells/glycine +; Fig. 7d). In addition, we studied the effects of blebbistatin on spine length in the presence of glycine. A similar decrease in spine length was observed by the blebbistatin application as by FILIP (1.19 ± 0.90 μm, 396 spines with blebbistatin in the absence of glycine; 0.97 ± 0.87 μm, 305 spines with blebbistatin in the presence of glycine; 1.07 ± 0.81 μm, 555 spines with vehicle only in the absence of glycine; 1.15 ± 0.91 μm, 543 spines with vehicle only in the presence of glycine; Fig. 7e).
We further investigated how FILIP influences the intracellular distribution of NMDA receptors, especially the NR1 (NR1) and NR2A (NR2A) subunits. We counted the numbers of NR1- and NR2A-positive deposits in spines and divided them by the number of observed spines. We defined this value as the ‘density of subunits in spines’. When FILIP was expressed, the densities of NR1- and NR2A-positive deposits were significantly lower (Fig. 7f, g; Fisher's exact test, two-tailed, NR1 p = 0.0211, NR2A p = 0.00009), suggesting that FILIP moderates NMDA receptor activity in the spine. We also investigated the proportion of NR2A-positive spines to all observed spines in the primary cultured piriform neurons in which the expression of FILIP was suppressed by inducible RNAi vectors. We observed a significant increase in NR2A-positive spines in FILIP-depleted piriform cortex neurons (control, 2.7% of 377 spines; FILIP knockdown, 5.6% of 444 spines; Fisher's exact test, two-tailed, p = 0.03807).
The response to NMDA is reduced in the piriform cortex of FILIP−/− mice
Because FILIP modified the distribution of NMDA receptors in the piriform cortex, we investigated whether responses to NMDA differed in the FILIP−/− mice compared with wild-type mice using calcium imaging techniques and a bath application of NMDA (see Supplemental Fig. S7 online). Neurons were less responsive to NMDA in the FILIP−/− mice compared with the wild-type littermates (Fig. 7h). Supplemental Fig. S8 depicts the FILIP effects on spines (see Supplemental Fig. S8 online).
The intracortical excitation propagation is abnormal in FILIP−/− mice
Because FILIP deletion resulted in an abnormal response to NMDA in the piriform cortex, we investigated whether FILIP influenced neuronal activity. Unlike other areas in the cortex, the visual cortex, where FILIP is expressed in layers II/III (Fig. 1g–i and Fig. 8a), receives strong and sole input from the thalamus, which allows us to investigate how neuronal activity propagates in the cortex by stimulating the thalamic input fibres. We stimulated the white matter of explants taken from the visual cortex so that axons from the lateral geniculate nucleus were activated and studied how neuronal activity was propagated using a voltage-sensitive dye28. The first excitation was observed in layer VI approximately 4.2 ms after stimulation (Fig. 8b). In the control, the excitation was subsequently propagated to the upper layers and spread laterally (horizontally) in layers II/III and layer VI (Fig. 8b). In the FILIP−/− mice, the degree of excitation in the upper cortical layers was lower than in the controls. In particular, the lateral propagation was reduced in layers II/III 8.4 ms after the stimulation (Fig. 8b). Whereas the peak amplitude of excitation in layer VI was not significantly different between the control and the FILIP−/− mice, it was significantly different in layers II/III, where FILIP is principally expressed (Fig. 8a–c). The reduction in the horizontal propagation of excitation in layers II/III of the FILIP−/− mice persisted even when GABAergic transmission was inhibited by treatment with the GABAA-receptor blocker bicuculline (Fig. 8d). We next investigated whether this phenomenon depended on the modification of myosin function by FILIP. We used blebbistatin to block myosin 2b activity and found that the excitation propagation was reduced in layers II/III of the visual cortex in wild-type mice, whereas that of FILIP−/− mice was not changed (Fig. 8e, f).
Discussion
We showed that FILIP bound to and influenced the subcellular distribution of, myosin 2b. FILIP is likely to facilitate the degradation of NMHC IIb by modulating enzyme accessibility, because we observed that exogenous FILIP prevented NMHC IIb from binding to actin fibres and it has been reported that the binding of myosin to actin fibres results in the protection of myosin from enzymatic digestion29. Although it is difficult to exclude the possibility that any compensatory activity of the FILIP knockout is responsible for an increase of myosin 2b, we observed an increase in the amount of NMHC IIb in FILIP−/− mice. We previously reported that FILIP enhances filamin A degradation8. Therefore, FILIP is likely to be a meta-regulator that orchestrates the activities of the major actin-binding proteins filamin A and myosin 2b.
We showed that FILIP deletion leads to a shorter spine length in the piriform cortex and that exogenous FILIP results in an elongated spine length in well-developed hippocampal neurons. These results indicate that FILIP, which is expressed in a region-specific manner, is a key molecule that confers unique regional spine characteristics.
The NR1, NR2A and NR2B subunits of the NMDA receptor bind to the myosin regulatory light chain, which is one component of myosin 2b. It has been demonstrated that such binding influences the function of NMDA receptors and the intracellular trafficking of these subunits30,31. As FILIP bound to the heavy chain of myosin 2b and altered its distribution, it is possible that the distribution of NMDA receptor is also altered from the synaptic area to other regions, for example dendritic shafts. Furthermore, as FILIP accelerates the degradation of filamin A through calpain activity8, the cleavage of NR2A receptor by the activated calpain may have resulted in the alteration of its distribution from the synaptic to extrasynaptic area32. We presumed that this alteration of localisation is the cause of the decrease of NR1 and NR2A signals in the spines of FILIP expressing neurons. While it is not completely clear how extrasynaptically localised NMDA receptors influence the neurons32, it is possible that they are responsible for the high responses to glycine in FILIP expressing hippocampal neurons. Interestingly, at DIV15, when spine lengths are becoming shorter and the spine head wider27, FILIP did not exert any significant influence on the spine morphology of hippocampal neurons without glycine. Because the maturation of cultured hippocampal neurons progresses greatly in the 2–3 weeks after the initiation of culture27, our data suggest that FILIP did not influence the developmental changes of spine morphology in any apparent way.
It has been demonstrated that myosin 2 activity is required for NMDA receptor-dependent synaptic plasticity and for LTP-related dendritic spine actin polymerisation6. It is possible that FILIP contributes to regional spine characteristics in learning11,13,33 through the modification of NMDA receptor and myosin 2b function, while our glycine treatment on the FILIP expressing hippocampal neurons resulted in the enlargement of spine head volume. While we administered our glycine treatment without AP5 preconditioning26,34 to avoid the alteration of expression of NMDA receptors by AP5 on the spines35, it may be that the glycine treatment did not fully mimic the in vivo condition, especially in the case of the our primary culture neurons. As we have previously discussed, FILIP showed no significant effects on the spine morphology of primary cultured hippocampal neurons at DIV15 without glycine treatment. This result indicated that the conditions of the primary culture neurons in our glycine treatment were somewhat different from those of piriform neurons in vivo at the adult stage. It is possible that the morphological changes of FILIP expressing hippocampal neurons could not mimic that of the piriform neurons in their response to learning due to the difference in the in vivo and in vitro conditions.
The results of our blebbistatin treatment on the cortex indicated that myosin 2b function represents the function of FILIP on the horizontal propagation of excitation in the upper layer. As myosin 2b in the growth cone of axons is involved in axon elongation36 and the myosin 2b activity regulates axon branching37, there is a possibility that alterations in the subcellular distribution of myosin 2b due to FILIP altered the intra-cortical circuit related to the horizontal propagation of excitation. Furthermore, the decrease of FILIP expressing cells in the upper layer of cortex is probably one of the causes of the decrease of horizontal propagation of excitation because β-galactosidase positive cells decreased in the cortex of FILIP−/− mice compared to FILIP+/− mice.
Methods
Method details are given in the Supplemental information online.
Animals
The mice were maintained in the animal room at the Division of Laboratory Animal Resources, University of Fukui and Hyogo College of Medicine. The day of birth was designated P0. All experiments were conducted in accordance with the Regulations for Animal Research at University of Fukui and the Regulations for Animal Experimentation in Hyogo College of Medicine. The Animal Research Committee, University of Fukui and the President of Hyogo College of Medicine under the review of the Hyogo College of Medicine Animal Experiment Committee approved the experiments.
Generation of FILIP-knockout mice
Conventional methods were used for the generation of FILIP-knockout mice.
Vector construction
The full-length rat l-FILIP cDNA was amplified using PCR and inserted into the pCAGGS vector38, which contains 3 × FLAG sequences and IRES GFP (pCAGGS FILIP IRES GFP). The empty-vector control was the pCAGGS vector expressing IRES GFP (pCAGGS IRES GFP). Vectors that express a truncated form of FILIP were constructed using the KOD-plus-mutagenesis kit (TOYOBO CO., LTD, Tokyo, Japan). The full-length and various fragments of NMHC IIb were amplified using PCR and inserted into the pCMV VSV vector to produce fragments tagged with VSV-G under the CMV promoter. The shRNA vector for NMHC IIb and NMHC IIb-resistant vector were constructed. The conditional knockdown vector for FILIP was constructed using the Tol2 transposon-mediated technique39. The target nucleotide was bp 1433–1453 of the mouse FILIP cDNA (GenBank accession number: BC131965.1). The primers for the vector construction are shown in the Supplemental information online.
Northern blot analyses
Conventional protocols were used for northern blot analyses40.
Histochemical detection of β-galactosidase and immunostaining
After fixation with 4% paraformaldehyde (PFA), the brains were cut into 30-μm sections with a cryostat. The sections were stained with X-gal staining solution (1 mg/ml X-gal, 2 mM MgCl2, 5 mM K3Fe(CN)6, 5 mM K4Fe(CN)6, 0.01% NP-40) at 37°C for 40–48 hr. For immunohistochemical analyses, the sections were incubated with antibody dilution solution for 30 min and incubated overnight at 4°C with the antibody dilution buffer containing appropriate concentration of antisera. The signals were visualised with Alexa Fluor 568-conjugated anti mouse IgG or Alexa Fluor 488-conjugated anti mouse IgG (Life Technologies Corporation, Grand Island, NY).
Golgi staining method
The FD Rapid GolgiStain kit (FD Neurotechnologies, Inc., Columbia, MD) was used for Golgi-Cox staining according to the manufacturer's protocol. The spine morphology in piriform layer II neurons was classified according to Harris's report41.
The primary culture of neurons from the hippocampus and the piriform cortex
The hippocampus and piriform neurons at E17.5 were cultured on polyethyleneimine-coated coverslips with growth medium (neurobasal medium; Life Technologies Corporation) containing MACS Supplement B27 PLUS (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany) and l-glutamine). For the analysis of spine morphology, the vectors were transfected using Lipofectamine 2000 (Life Technologies Corporation) at DIV17 or 18. To calculate the spine head volume and length, Z-stacked images were captured using a confocal microscope (LSM 5 Pascal, Carl Zeiss MicroImaging, GmbH, Jena, Germany) at DIV21. The spine head width and spine head length of the mushroom- or thin-type spines were measured using the ImageJ image analysis program (NIH, Bethesda, MD)42. Spine head volume was calculated according to the method of Knafo13. To treat cells with glycine, the vectors were transfected as described above at DIV12. At DIV15, the cells were incubated with 200 μM glycine-containing HEPES-buffered balanced salt solution (HEPES-BSS) supplemented with 0.5 μM tetrodotoxin, 1 μM strychnine and 20 μM bicuculline for 10 min following incubation with normal HEPES-BSS for 20 min. To be treated with 10 μM blebbistatin during the glycine treatment, the cells were preincubated with growth medium containing 10 μM blebbistatin for 30 min; then, the cells were treated with glycine as described above, with the exception of using HEPES-BSS supplemented with 10 μM blebbistatin. NMDA receptors were visualised using a polyclonal anti-NR1 antibody (Sigma-Aldrich Co. LLC, St. Louis, MO) or a polyclonal anti-NMDAR2A antibody (EMD Millipore Corporation, Billerica, MA) followed by Alexa Fluor 633-conjugated anti-rabbit IgG (Life Technologies Corporation). pCAGGS-tdTomato, pT2K-CAGGS-rtTAM2, pT2K-TBI-shRNAmir and pCAGGS-T2TP (Tol2 transposase) were transfected at a ratio of 1:4:2:4 using Lipofectamine 2000 at DIV16. The culture media were changed to doxycycline-containing media (1 μg/ml) at DIV17; the cells were observed at DIV21.
Immunoprecipitation
COS-7 cells that had been cultured in a 6-cm dish were lysed in 400 μl of ice-cold lysis buffer 24 hr after transfection with the plasmid vectors. Pre-cleared lysates were incubated with antibody-bound Protein G Dynabeads (Life Technologies Corporation) at 4°C for 3 hr. After the Dynabeads were rinsed three times in lysis buffer, the immunoprecipitated proteins were eluted in SDS sample buffer. Details are provided in the Supplemental information online.
Western blot analyses and anti-FILIP antisera
The cortices of the ICR mouse brains and hearts were homogenised in lysis buffer. The insoluble materials were removed through centrifugation. The protein concentration was measured using protein assay CBB solution (Nacalai Tesque, Inc., Kyoto, Japan). Protein lysates (5 μg) or immunoprecipitation products were separated through SDS-PAGE and transferred onto polyvinylidene difluoride membranes (EMD Millipore Corporation). After the membranes were blocked, they were incubated with primary antibodies followed by secondary antibodies coupled to HRP (1:2,000; BD Biosciences, Franklin Lakes, NJ). The peroxidase activity was detected using enhanced chemiluminescence. The blot densities were quantified using the ImageJ program. Anti-FILIP antisera were raised against a recombinant peptide of rat FILIP (ESQEMPMGRTILK) in rabbit.
In utero electroporation gene transfer of the inducible FILIP knockdown vector
We performed in utero electroporation-mediated gene transfer as previously reported39. pCAGGS-tdTomato, pT2K-CAGGS-rtTAM2, pT2K-TBI-shRNAmir and pCAGGS-T2TP (Tol2 transposase) were transfected at a ratio of 1:10:5:10 into the embryonic brains of ICR mice at E14.5. At P21 to P28, the delivered pups were administered doxycycline via the drinking water (the final concentration was 2 mg/ml + 5% sucrose). The brains were dissected out at P28 and cut coronally into 100-μm slices with a Vibratome (DOSAKA EM Co., LTD., Kyoto, Japan).
Preparation for optical recording
We performed optical recording with some modifications from the previous reports43,44. For optical imaging of gross neuronal excitation, the slices (400-μm thick) were prepared from the visual cortices of FILIP−/− and their FILIP+/+ or FILIP+/− littermates at one month of age. The slices were stained in a bath filled with RH-482 (0.1 mg/ml; 20 min). After completing the optical recordings under perfusion with Ringer's solution, the slice was perfused with Ringer's solution containing 2 μM bicuculline for 30 min. The optical recording in the presence of bicuculline was then performed.
Optical recording
The light absorption change at 700 ± 32 nm was recorded using an imaging system (Deltaron 1700; Fujifilm Corporation, Tokyo, Japan) with 128 × 128 pixel photo sensors at a frame rate of 0.6 ms. Starting at 10 ms before each stimulus, the image sensor took 128 consecutive frames at a sampling interval of 0.6 ms. A reference frame, which was taken immediately before each series of 128 frames, was subtracted from the subsequent 128 frames. An electric pulse was given to the white matter. The ratio image was then calculated by dividing the image data by the reference frame.
Ca2+ imaging
We performed Ca2+ imaging with some modifications from the previous report44. The brains were removed and transferred to an ice-cold aerated solution (95% O2 and 5% CO2) containing (in mM) 120 choline chloride, 2.4 KCl, 26 NaHCO3, 1.2 NaH2PO4, 0.5 CaCl2, 7 MgCl2, 1.2 ascorbic acid and 15 glucose. Coronal slices (300-μm thick) containing the piriform cortex were prepared using a vibrating microtome. Each slice was loaded for 45 min at room temperature with 10 μM Fluo-4/AM (Life Technologies Corporation) in the presence of 0.01% Pluronic F-127 (Life Technologies Corporation). The slices were then washed thoroughly with Ringer's solution and set in a chamber (0.2 mL) on an upright microscope (BX51WI; Olympus, Tokyo, Japan). Each slice was perfused with Ringer's solution containing (in mM) 127 NaCl, 2 KCl, 1.2 KH2PO4, 1.3 MgSO4, 2.4 CaCl2, 26 NaHCO3 and 10 glucose (oxygenated with 95% O2 and 5% CO2) at room temperature (25 ± 2°C). The confocal images of fluo-4 fluorescence were captured in a 0.4-mm2 area of the piriform cortex at a sampling interval of 500 ms using a CSU10 Nipkow spinning-disk confocal microscope (Yokogawa Electric, Tokyo, Japan) equipped with an EM CCD camera (iXON EM; Andor Technology Ltd., Belfast, Northern Ireland, UK). Fluo-4 fluorescence at 518 nm was excited by light at 488 nm from a semiconductor laser. NMDA (50 μM) or a high-K+ solution were puff-applied for 10 s beginning at 10 s after the start of recording. The high-K+ solution was identical to the Ringer's solution except for (in mM) in addition of 76 NaCl and 50 KCl.
Statistical analyses
To analyse the statistical significance of means, we used an unpaired two-tailed Student t-test when two samples had equal variances and Welch's t-test when two samples had unequal variances. To analyse statistical significance of the spine head volume, we used the Wilcoxon rank sum test or Welch's t-test. To analyse statistical significance of ratio, we used Fisher's exact test.
References
Le Clainche, C. & Carlier, M. F. Regulation of actin assembly associated with protrusion and adhesion in cell migration. Physiol. Rev. 88, 489–513 (2008).
Honkura, N., Matsuzaki, M., Noguchi, J., Ellis-Davies, G. C. & Kasai, H. The subspine organization of actin fibers regulates the structure and plasticity of dendritic spines. Neuron 57, 719–729 (2008).
Hotulainen, P. & Hoogenraad, C. C. Actin in dendritic spines: connecting dynamics to function. J. Cell Biol. 189, 619–629 (2010).
Ryu, J., Liu, L., Wong, T. P., Wu, D. C. et al. A critical role for myosin IIb in dendritic spine morphology and synaptic function. Neuron 49, 175–182 (2006).
Vicente-Manzanares, M., Ma, X., Adelstein, R. S. & Horwitz, A. R. Non-muscle myosin II takes centre stage in cell adhesion and migration. Nat. Rev. Mol. Cell Biol. 10, 778–790 (2009).
Rex, C. S., Gavin, C. F., Rubio, M. D., Kramar, E. A. et al. Myosin IIb regulates actin dynamics during synaptic plasticity and memory formation. Neuron 67, 603–617 (2010).
Hodges, J. L., Newell-Litwa, K., Asmussen, H., Vicente-Manzanares, M. & Horwitz, A. R. Myosin IIb activity and phosphorylation status determines dendritic spine and post-synaptic density morphology. PLoS One 6, e24149 (2011).
Nagano, T., Yoneda, T., Hatanaka, Y., Kubota, C. et al. Filamin A-interacting protein (FILIP) regulates cortical cell migration out of the ventricular zone. Nat. Cell Biol. 4, 495–501 (2002).
Fox, J. W., Lamperti, E. D., Ekşioğlu, Y. Z., Hong, S. E. et al. Mutations in filamin 1 prevent migration of cerebral cortical neurons in human periventricular heterotopia. Neuron 21, 1315–1325 (1998).
Feng, Y. & Walsh, C. A. The many faces of filamin: a versatile molecular scaffold for cell motility and signalling. Nat. Cell Biol. 6, 1034–1038 (2004).
Matsuzaki, M., Honkura, N., Ellis-Davies, G. C. & Kasai, H. Structural basis of long-term potentiation in single dendritic spines. Nature 429, 761–766 (2004).
Matus, A. Actin-based plasticity in dendritic spines. Science 290, 754–758 (2000).
Knafo, S., Libersat, F. & Barkai, E. Olfactory learning-induced morphological modifications in single dendritic spines of young rats. Eur. J. Neurosci. 21, 2217–2226 (2005).
Mullen, R. J., Buck, C. R. & Smith, A. M. NeuN, a neuronal specific nuclear protein in vertebrates. Development 116, 201–211 (1992).
Whittaker, S. G., Wroble, J. T., Silbernagel, S. M. & Faustman, E. M. Characterization of cytoskeletal and neuronal markers in micromass cultures of rat embryonic midbrain cells. Cell Biol. Toxicol. 9, 359–375 (1993).
Molyneaux, B. J., Arlotta, P., Menezes, J. R. & Macklis, J. D. Neuronal subtype specification in the cerebral cortex. Nat. Rev. Neurosci. 8, 427–437 (2007).
Fiala, J. C., Feinberg, M., Popov, V. & Harris, K. M. Synaptogenesis via dendritic filopodia in developing hippocampal area CA1. J. Neurosci. 18, 8900–8911 (1998).
Holtmaat, A. J., Trachtenberg, J. T., Wilbrecht, L., Shepherd, G. M. et al. Transient and persistent dendritic spines in the neocortex in vivo. Neuron 45, 279–291 (2005).
Noam, Y., Phan, L., McClelland, S., Manders, E. M. et al. Distinct regional and subcellular localization of the actin-binding protein filamin A in the mature rat brain. J. Comp. Neurol. 520, 3013–3034 (2012).
Sellers, J. R. Myosins: a diverse superfamily. Biochim. Biophys. Acta 1496, 3–22 (2000).
Pellegrin, S. & Mellor, H. Actin stress fibres. J. Cell Sci. 120, 3491–3499 (2007).
Goeckeler, Z. M., Bridgman, P. C. & Wysolmerski, R. B. Nonmuscle myosin II is responsible for maintaining endothelial cell basal tone and stress fiber integrity. Am. J. Physiol. Cell Physiol. 295, C994–1006 (2008).
Kovács, M., Tóth, J., Hetényi, C., Málnási-Csizmadia, A. & Sellers, J. R. Mechanism of blebbistatin inhibition of myosin II. J. Biol. Chem. 279, 35557–35563 (2004).
Korobova, F. & Svitkina, T. Molecular architecture of synaptic actin cytoskeleton in hippocampal neurons reveals a mechanism of dendritic spine morphogenesis. Mol. Biol. Cell 21, 165–176 (2010).
Thomson, A. M., Walker, V. E. & Flynn, D. M. Glycine enhances NMDA-receptor mediated synaptic potentials in neocortical slices. Nature 338, 422–424 (1989).
Lu, W., Man, H., Ju, W., Trimble, W. S. et al. Activation of synaptic NMDA receptors induces membrane insertion of new AMPA receptors and LTP in cultured hippocampal neurons. Neuron 29, 243–254 (2001).
Papa, M., Bundman, M. C., Greenberger, V. & Segal, M. Morphological analysis of dendritic spine development in primary cultures of hippocampal neurons. J. Neurosci. 15, 1–11 (1995).
Tanifuji, M., Sugiyama, T. & Murase, K. Horizontal Propagation of Excitation in Rat Visual Cortical Slices Revealed by Optical Imaging. Science 266, 1057–1059 (1994).
Mornet, D., Bertrand, R. U., Pantel, P., Audemard, E. & Kassab, R. Proteolytic approach to structure and function of actin recognition site in myosin heads. Biochemistry 20, 2110–2120 (1981).
Amparan, D., Avram, D., Thomas, C. G., Lindahl, M. G. et al. Direct interaction of myosin regulatory light chain with the NMDA receptor. J. Neurochem. 92, 349–361 (2005).
Bajaj, G., Zhang, Y., Schimerlik, M. I., Hau, A. M. et al. N-methyl-D-aspartate receptor subunits are non-myosin targets of myosin regulatory light chain. J. Biol. Chem. 284, 1252–1266 (2009).
Gladding, C. M. & Raymond, L. A. Mechanisms underlying NMDA receptor synaptic/extrasynaptic distribution and function. Mol. Cell. Neurosci. 48, 308–320 (2011).
Van Harreveld, A. & Fifkova, E. Swelling of dendritic spines in the fascia dentata after stimulation of the perforant fibers as a mechanism of post-tetanic potentiation. Exp. Neurol. 49, 736–749 (1975).
Fortin, D. A., Davare, M. A., Srivastava, T., Brady, J. D. et al. Long-term potentiation-dependent spine enlargement requires synaptic Ca2+-permeable AMPA receptors recruited by CaM-kinase I. J. Neurosci. 30, 11565–11575 (2010).
Molnár, E. Long-term potentiation in cultured hippocampal neurons. Semin. Cell Dev. Biol. 22, 506–513 (2011).
Yu, P., Santiago, L. Y., Katagiri, Y. & Geller, H. M. Myosin II activity regulates neurite outgrowth and guidance in response to chondroitin sulfate proteoglycans. J. Neurochem. 120, 1117–1128 (2012).
Ohnami, S., Endo, M., Hirai, S., Uesaka, N. et al. Role of RhoA in activity-dependent cortical axon branching. J. Neurosci. 28, 9117–9121 (2008).
Niwa, H., Yamamura, K. & Miyazaki, J. Efficient selection for high-expression transfectants with a novel eukaryotic vector. Gene 108, 193–199 (1991).
Iguchi, T., Yagi, H., Wang, C. C. & Sato, M. A tightly controlled conditional knockdown system using the tol2 transposon-mediated technique. PLoS One 7, e33380 (2012).
Sambrook, J. & Russell, D. Molecular Cloning: A Laboratory Manual (CSHL Press, NYUSA, 2001).
Harris, K. M., Jensen, F. E. & Tsao, B. Three-dimensional structure of dendritic spines and synapses in rat hippocampus (CA1) at postnatal day 15 and adult ages: implications for the maturation of synaptic physiology and long-term potentiation. J. Neurosci. 12, 2685–2705 (1992).
Collins, T. J. ImageJ for microscopy. Biotechniques 43, 25–30 (2007).
Kiritoshi, T., Ikeda, H. & Murase, K. Long-term potentiation of neuronal excitation in the central nucleus of the rat amygdala revealed by imaging with a voltage-sensitive dye. Brain Res. 1349, 32–40 (2010).
Ikeda, H., Kiritoshi, T. & Murase, K. Contribution of microglia and astrocytes to the central sensitization, inflammatory and neuropathic pain in the juvenile rat. Mol. pain 8, 43 (2012).
Acknowledgements
We are grateful to Dr. M. Takahashi (Hokkaido Univ) for the myosin 2b cDNA, to Dr. Y. Yoshimura (NIPS), Dr. Y. Ishikawa (NAIST) and Dr. Y. Oka (Osaka Univ) for helpful discussions, to M. Murota, H. Yoshikawa, S. Kanae, C.C. Wang and H. Miyagoshi for technical assistance and to T. Taniguchi for secretarial assistance. This work was supported in part by the Competitive Allocation Fund and the Multidisciplinary Program for Elucidating the Brain Development from Molecules to Social behaviour (Fukui Brain Project) of University of Fukui, Grant-in-Aid for Researchers (Hyogo College of Medicine, 2013) and a Grant-in-Aid for Scientific Research and the Strategic Research Program for Brain Sciences “Integrated research on neuropsychiatric disorders” from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan.
Author information
Authors and Affiliations
Contributions
H.Y. carried out most of the experiments and analysed the data. M.S. initially conceived the project and directed the research. T.N., S.M. and M.O. generated the FILIP knockout mice. M.K. prepared the samples for histological analyses. R.M.T. and T.I. prepared plasmid vectors. M.-J.X. and K.K. prepared and manipulated the primary culture of hippocampal neurons. H.I and K.M carried out the optical recording and the Ca2+ imaging. M.S. wrote the manuscript together with H.Y. and all authors contributed to the final version.
Ethics declarations
Competing interests
The authors declare no competing financial interests.
Electronic supplementary material
Supplementary Information
Supplemental figures and methods
Rights and permissions
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder in order to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-sa/4.0/
About this article
Cite this article
Yagi, H., Nagano, T., Xie, MJ. et al. Filamin A-interacting protein (FILIP) is a region-specific modulator of myosin 2b and controls spine morphology and NMDA receptor accumulation. Sci Rep 4, 6353 (2014). https://doi.org/10.1038/srep06353
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/srep06353
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.