Activity-dependent local translation of matrix metalloproteinase-9

Abstract

Local, synaptic synthesis of new proteins in response to neuronal stimulation plays a key role in the regulation of synaptic morphogenesis. Recent studies indicate that matrix metalloproteinase-9 (MMP-9), an endopeptidase that regulates the pericellular environment through cleavage of its protein components, plays a critical role in regulation of spine morphology and synaptic plasticity. Here, we sought to determine whether MMP-9 mRNA is transported to dendrites for local translation and protein release. First, dendritic transport of MMP-9 mRNA was seen in primary hippocampal neuronal cultures treated with glutamate and in dentate gyrus granule cells in adult anesthetized rats after induction of long-term potentiation. Second, rapid, activity-dependent polyadenylation of MMP-9 mRNA; association of the mRNA with actively translating polysomes; and de novo MMP-9 protein synthesis were obtained in synaptoneurosomes isolated from rat hippocampus. Third, glutamate stimulation of cultured hippocampal neurons evoked a rapid (in minutes) increase in MMP-9 activity, as measured by cleavage of its native substrate, β-dystroglycan. This activity was reduced by the polyadenylation inhibitor, thus linking MMP-9 translation with protein function. In aggregate, our findings show that MMP-9 mRNA is transported to dendrites and locally translated and that the protein is released in an activity-dependent manner. Acting in concert with other dendritically synthesized proteins, locally secreted MMP-9 may contribute to the structural and functional plasticity of the activated synapses.

Figure 1.

Localization of MMP-9 and Arc mRNA by fluorescent in situ hybridization in the granule and molecular layers of dentate gyrus. A, B, E, F, Representative staining of dentate gyrus coronal sections from control brain (A, E) or 2 h after high-frequency stimulation leading to LTP (B, F). C, G, The sense probes showed uniform background levels on the section. D, H, Higher magnification of the border region between dentate gyrus molecular and granule cell layers to visualize granule cells extending dendrites into the molecular layer. I, J, Fluorescence intensity was measured with Fiji software in molecular and granule cell layers on three slices from three HFS-stimulated animals and three controls. The sense probe was quantified for molecular and granule cell layers and subtracted as a background from the intensity of fluorescence measured for molecular and granule cell layers antisense probes. MMP-9 and Arc mRNA was upregulated in the molecular and granule cell layers 2 h after HFS compared with unstimulated controls. *p < 0.05; **p < 0.01; ***p < 0.005, unpaired t test. n = 3 animals per group. The graph represents mean values ± SEM. GL, Granule cell layer; ML, molecular layer.

Figure 2.

Dendritic transport of MMP-9 mRNA is regulated by synaptic stimulation in hippocampal neurons. Fluorescent in situ hybridization combined with immunofluorescence shows MMP-9 mRNA granules in dendrites stained with anti-FMRP antibody. A, For the control and before bicuculline application the cells were incubated for 3 h with inhibitors. B, C, To activate synaptic release of glutamate, cultured hippocampal neurons were stimulated by application of bicuculline for 10 min (B) or 20 min (C). D, In situ hybridization with the sense probe. E, The number of MMP-9-positive mRNA granules was counted per 50 μm of dendrite, 10 neurons for each experimental condition (n = 10). Similar results were obtained in three independent experiments. ***p < 0.001, Mann–Whitney test. The graph represents mean values ± SEM. ctr, Control.

Figure 3.

MMP-9 mRNA-containing granules show direct movement along dendrites of rat hippocampal neurons. Hippocampal neurons were cotransfected with two vectors, pSyn-MS2-MMP9UTR vector containing the MMP-9 coding sequence with the 3′UTR and 24 copies of MS2-binding site RNA hairpin, under the control of synapsin-1 promoter; and pMS2-GFP vector expressing MS2-binding protein as a fusion with GFP and nuclear localization sequence. After expression of both constructs, the GFP fusion should bind as a dimer to each MS2-binding site of the RNA construct (top). GFP-labeled MMP-9 mRNA observed in the living cell by time-lapse microscopy shows many small granules in the dendrites. Sequential images were taken every 20 s; the arrow shows one granule moving bidirectionally along the dendrite.

Figure 4.

MMP-9 protein is enriched in the synaptoneurosomal fraction. A, Characterization of protein content of the synaptoneurosomal fraction. Crude homogenate (H), filtrate (F), SNs, and supernatant [cytosolic (C)] fractions were analyzed by Western blot to determine the distribution pattern of synaptic and nonsynaptic proteins. Synaptic protein PSD-95 and MMP-9 were enriched in the SN. The glial protein GFAP and the nuclear protein c-Jun were either diminished or absent in the SN fraction. B, The activity of MMP-9 released by NS or enhanced in stimulated (10 μm glutamate, 50 μm NMDA for 5 or 15 min) synaptoneurosomes, as measured by the DQ gelatin assay. Error bars represent SD; n = 3.

Figure 5.

MMP-9 mRNA is rapidly polyadenylated in synaptoneurosomes after glutamate stimulation. A, Schematic representation of the rat MMP-9 3′UTR structure with putative poly(A) sites and CPEB binding sites. B, Alignment between the MMP-9 3′UTR regions from rat and mouse containing CPEB binding sites. Consensus sequences are bolded. C, E, MMP-9 and αCaMKII mRNA polyadenylation measured by PAT assay. Synaptoneurosomes were either NS or stimulated with glutamate for 3, 6, or 12 min. M, Size marker. Oligoadenylated and polyadenylated forms of MMP-9 and αCaMKII mRNAs are indicated by black arrows. D, F, Dispersion graph representing the distribution of the MMP-9 and αCaMKII polyadenylated transcripts in NS (black line) and glutamate-stimulated (3, 6, and 12 min; gray lines) SN. The signal intensity along the lane has been plotted against the poly(A) tail length, estimated from the molecular markers loaded on the same gel.

Figure 6.

Polyribosome association and local translation of MMP-9 mRNAs in synaptoneurosomes. A, The synaptoneurosomes were lysed; separated on a sucrose density gradient; and, according to the absorbance profile, divided into five fractions: free mRNP, monosomal fraction, light polysomes, heavy polysomes, and RNA granules. B, A254 absorbance profiles from sucrose density gradients. C, Radioactive RT-PCR with primers specific for MMP-9, Arc, and control (LSM) mRNAs show polyribosome association of MMP-9 and Arc mRNAs in control (NS) and stimulated (10 μm glutamate, 50 μm NMDA, 15 min) synaptoneurosomes. D, Polysomal incorporation of MMP-9 and Arc mRNAs is expressed as a percentage of mRNA in fractions 3 and 4 (polysomes). Synaptoneurosomal stimulation leads to increased association of MMP-9 and Arc mRNAs with polyribosomes. *p < 0.05, unpaired t test; n = 5. The graph represents mean values ± SEM. E, Local synthesis of MMP-9 protein detected with click chemistry. The proteins newly synthesized in synaptoneurosomes incorporate HPG that, in click reaction, is tagged with biotin and enriched on streptavidin dynabeads. Western blot on precipitated proteins show induction of MMP-9 and Arc protein synthesis in synaptoneurosomes after the stimulation. Application of 20 μm anisomycin inhibits local translation of MMP-9 and Arc proteins. F, Quantification of MMP-9 and Arc bands intensity shows an increase of protein synthesis in SN after stimulation and its inhibition by anisomycin. Error bars represent SD; n = 3. The results were not statistically significant. The bands were quantified by measuring their intensity in a Fiji program; the background subtraction was done separately for each lane. S, stimulated; NS, not stimulated.

Figure 7.

Glutamate-dependent increase in MMP-9 proteolytic activity is regulated by polyadenylation in cultured hippocampal neurons. A, The effect of MMP-9 inhibitor I on activity-induced β-DG cleavage. Application of MMP-9 inhibitor I to the hippocampal neurons before glutamate stimulation results in reduction of β-DG cleavage. B, D, Representative Western blots of β-DG and GAPDH. Hippocampal neurons were preincubated with 20 μm cordycepin 10 min before either bicuculline (synaptic protocol) or glutamate stimulation. GAPDH was a protein loading control. C, E, Quantification of β-DG 30 kDa product of MMP-9 cleavage intensity from three independent experiments (n = 3). *p < 0.05; **p < 0.01;***p < 0.001, Mann–Whitney test. The graph represents mean values ± SEM. ctr, Control.