green fluorescent protein
lysosome-associated membrane protein 2
mechanistic target of rapamycin
ribosomal S6 kinase 1
Ras homologue enriched in brain
tuberous sclerosis complex-2
- • Mechanical stimuli play a major role in the regulation of skeletal muscle mass.
- • Signalling through a protein kinase called the mechanistic target of rapamycin (mTOR) is essential for mechanically induced changes in muscle mass; however, the mechanism(s) via which mechanical stimuli regulate mTOR signalling have not been defined.
- • In this study, mouse skeletal muscles were stimulated with eccentric contractions (ECs) to determine if the mechanical activation of mTOR signalling is associated with changes in the phosphorylation of the tuberous sclerosis complex-2 (TSC2) and the targeting of both mTOR and TSC2 to the lysosome.
- • Our results demonstrate that ECs induce hyper-phosphorylation of TSC2, enhanced lysosomal targeting of mTOR and nearly abolish the lysosomal targeting of TSC2.
- • These novel observations suggest that alterations in the lysosomal targeting of mTOR/TSC2 could play a fundamental role in the mechanism via which mechanical stimuli regulate mTOR signalling and ultimately skeletal muscle mass.
Abstract The goal of this study was to determine whether the mechanical activation of mechanistic target of rapamycin (mTOR) signalling is associated with changes in phosphorylation of tuberous sclerosis complex-2 (TSC2) and targeting of mTOR and TSC2 to the lysosome. As a source of mechanical stimulation, mouse skeletal muscles were subjected to eccentric contractions (ECs). The results demonstrated that ECs induced hyper-phosphorylation of TSC2 and at least part of this increase occurred on residue(s) that fall within RxRxxS/T consensus motif(s). Furthermore, in control muscles, we found that both mTOR and TSC2 are highly enriched at the lysosome. Intriguingly, ECs enhanced the lysosomal association of mTOR and almost completely abolished the lysosomal association of TSC2. Based on these results, we developed a new model that could potentially explain how mechanical stimuli activate mTOR signalling. Furthermore, this is the first study to reveal that the activation of mTOR is associated with the translocation of TSC2 away from the lysosome. Since a large number of signalling pathways rely on TSC2 to control mTOR signalling, our results have potentially revealed a fundamental mechanism via which not only mechanical, but also various other types of stimuli, control mTOR signalling.
It is well recognized that mechanical stimuli play a central role in the regulation of skeletal muscle mass and the maintenance of muscle mass is essential for mobility, disease prevention and quality of life (Goldberg et al. 1975; Seguin & Nelson, 2003). Over the last decade it has become apparent that the activation of a protein kinase called the mechanistic target of rapamycin (mTOR) is essential for mechanically induced changes in muscle mass (Bodine et al. 2001; Goodman et al. 2011). However, the mechanism(s) via which mechanical stimuli activate mTOR signalling remain vaguely defined. Nevertheless, advancements in our understanding of the mechanisms that control mTOR signalling are being made. For example, recent studies have demonstrated that both phosphatidic acid (PA) and GTP-, but not GDP-, bound Ras homologue enriched in brain (Rheb) can directly stimulate mTOR kinase activity in vitro (Sancak et al. 2007; You et al. 2012). Furthermore, it has been reported that both of these molecules are enriched at the lysosome and a growing body of evidence indicates that mTOR signalling is regulated, in part, by controlling its association with the lysosome (Sengupta et al. 2010; Zhao et al. 2012). For instance, it has been shown that nutrients such as amino acids cause mTOR to translocate to the lysosome, and it is thought that this event promotes mTOR signalling by enhancing the association of mTOR with PA and GTP-Rheb (Sancak et al. 2010; Yoon et al. 2011). On the other hand, stimulants such as growth factors are believed to primarily control mTOR signalling by regulating the GTP-loading state of Rheb (Huang & Manning, 2008).
The GTP-loading state of Rheb is regulated by the tuberous sclerosis complex-2 (TSC2) which functions as a GTPase-activating protein (GAP) that converts active GTP-Rheb into inactive GDP-Rheb (Huang & Manning, 2008). Based on recent studies, it has been proposed that growth factors induce changes in the phosphorylation of TSC2 and this inhibits the ability of TSC2 to function as a GAP (Huang & Manning, 2008). As a result, growth factors induce an accumulation of active GTP-Rheb and, in turn, promote mTOR signalling. This model is supported by several studies which have shown that growth factor-induced phosphorylation of TSC2 inhibits its GAP activity towards Rheb in vivo; however, the molecular mechanisms that are responsible for this inhibition remain a subject of intense debate (Huang & Manning, 2008). For instance, stimulation with growth factors does not alter TSC2's GAP activity when measured in vitro (Cai et al. 2006). Based on this point, it has been argued that rather than inducing changes in the intrinsic GAP activity, phosphorylation of TSC2 induces changes in its subcellular localization so that it can no longer function as a GAP for Rheb. In support of this possibility, previous studies have demonstrated that growth factor-induced phosphorylation of TSC2 causes it to translocate from a crude membrane to cytosolic fraction (Cai et al. 2006; Miyazaki et al. 2010). Furthermore, just like mTOR and Rheb, TSC2 was recently shown to colocalize with a well-established lysosomal marker called lysosome-associated membrane protein 2 (LAMP2) (Eskelinen, 2006; Dibble et al. 2012). Hence, it seems very plausible that growth factor-induced phosphorylation of TSC2 causes it to translocate away from the lysosome. Consequently, TSC2 would lose its ability to function as a GAP for Rheb at the lysosome and, in turn, promote the activation of the lysosome-associated mTOR. Although intriguing, the existence of such a mechanism has yet to be experimentally verified.
As illustrated above, a growing body of evidence suggests that changes in the targeting of mTOR to the lysosome, and phosphorylation-induced changes in the localization of TSC2, play fundamental roles in the regulation of mTOR signalling. However, the potential role of these mechanisms in the mechanical activation of mTOR signalling remains largely unexplored. Therefore, the goal of this study was to determine whether mechanically induced changes in mTOR signalling are associated with changes in the phosphorylation of TSC2 and the targeting of both mTOR and TSC2 to the lysosome.
Male C57BL6 mice (Jackson Laboratories, Bar Harbor, MA, USA), at 8–12 weeks of age, were randomly assigned to different experimental groups. All animals were fed ad libitum, and immediately prior to surgical procedures the mice were anaesthetized with either inhaled isoflurane (1–5%) in O2 or an intraperitoneal injection of ketamine (100 mg kg−1) plus xylazine (10 mg kg−1). Upon the completion of experimental treatments, mice were killed by cervical dislocation while under anaesthesia. All methods were approved by the Institutional Animal Care and Use Committee of the University of Wisconsin-Madison and conform to the principles of UK regulations, as described in Drummond (2009).
Contractions of the lower leg muscle were induced by stimulating the sciatic nerve with an SD9E Grass Stimulator (Warwick, RI, USA) as previously described (O’Neilet al. 2009). The right (eccentrically contracted, EC) and left (control) tibialis anterior (TA) muscles were collected 1 h post stimulation.
Immunoprecipitation of TSC2
Upon collection, TA muscles were immediately frozen in liquid N2 and homogenized with a Polytron in ice-cold NP-40 lysis buffer (10 mm Tris-HCl (pH 7.5), 100 mm NaCl, 2 mm EDTA, 1% NP-40, 1 mm DTT, 1 mm phenylmethylsulfonyl fluoride (PMSF), 20 μg ml−1 of leupeptin, pepstatin, aprotinin and soybean trypsin inhibitor, 25 mm NaF, 25 mmβ-glycerophosphate and 1 mm Na3VO4). Whole muscle homogenates were centrifuged at 2500 g for 5 min and 1000 μg of protein from the supernatant was diluted to a volume of 0.5 ml with NP-40 buffer. Samples were then incubated with either anti-TSC2 (Cell Signaling, Danvers, MA, USA) or pre-immune anti-rabbit IgG (Santa Cruz Biotechnology, Dallas, TX, USA) antibodies at 4°C for 2 h. During this incubation, 40 μl of protein A agarose beads (Santa Cruz) were blocked in ice-cold 1% bovine serum albumin (BSA)-PBS for 1 h and then washed 3 times with PBS. The antibody-containing samples were incubated with the blocked beads at 4°C for 2 h. The beads were then pelleted by centrifugation at 500 g for 30 s and washed 4 times with NP-40 buffer, and 3 times with high salt wash buffer (40 mm Hepes (pH 7.4), 400 mm NaCl, 2 mm EDTA, 0.3% CHAPS). After the washes, the pellets were dissolved in Laemmli buffer, heated to 100°C for 5 min, pelleted at 500 g for 30 s and then the supernatant was subjected to Western blot analysis.
λ phosphatase treatment
TSC2 was immunoprecipitated as described above and then subjected to two additional washes with TBS (50 mm Tris-HCl (pH 7.5) and 150 mm NaCl). The beads were then resuspended in 50 μl of phosphatase buffer (50 mm Tris-HCl (pH 7.5), 100 mm NaCl, 2 mm DTT, 0.01% Brij-35, 2 mm MnCl2 and 400 units of λ phosphatase (New England Biolabs, Ipswich, MA, USA)) and incubated at 30°C for 1 h. The reaction was terminated by the addition of Laemmli buffer and then the samples were prepared for Western blot analysis as described above.
Upon collection, TA muscles were immediately frozen in liquid N2 and homogenized with a Polytron in ice-cold fractionation buffer (20 mm Tris-HCl (pH 7.5), 250 mm sucrose, 1 mm EDTA, 1 mm EGTA, 1 mm DTT, 1 mm PMSF, 25 mm NaF, 25 mmβ-glycerophosphate, 1 mm Na3VO4, and 20 μg ml−1 of leupeptin, pepstatin, aprotinin and soybean trypsin inhibitor). The homogenates were pre-cleared with centrifugation at 1000 g (4°C) for 10 min. The pre-cleared supernatant was centrifuged at 100,000 g (4°C) for 1 h. The resulting supernatant (cytosolic fraction) was retained and the pellet (crude membrane fraction) was washed twice in fractionation buffer and centrifuged at 100,000 g (4°C) for 10 min. The membrane pellet was then resuspended in 1X Laemmli buffer with a volume that was twice that of the original pre-cleared supernatant. The cytosolic fraction was dissolved with an equal volume of 2X Laemmli buffer. The same volumes of membrane and cytosolic samples (1:1 ratio) were then heated at 100°C for 5 min and subjected to Western blot analysis.
Western blot analysis
Western blot analyses were performed as previously described (Goodman et al. 2011). Images of the blots were either captured on film or with a Chemi410 camera mounted to a UVP Autochemi system (UVP, Upland, CA, USA). Densitometric measurements were carried out on ImageJ (NIH; http://rsb.info.nih.gov/nih-image/). Anti-phospho ribosomal S6 kinase 1 (p70S6k(389); sc-11759-R) was purchased from Santa Cruz. Anti-Na+,K+-ATPase (A6Fs) was purchased from Developmental Studies Hybridoma Bank (Iowa City, IA, USA). Peroxidase-conjugated anti-rabbit (PI-1000) and anti-mouse antibodies (PI-2000) were purchased from Vector Laboratories (Burlingame, CA, USA). Anti-Lactate dehydrogenase (LDH)(C28H7) (no. 3558), anti-mTOR (no. 2972), anti-phos pho RxRxxS*/T*(23C8D2) (no. 10001S), anti-p70S6k(49D7) (no. 2708S), anti-TSC2(D93F17) (no. 4308S) and anti-tubulin (no. 2148) were purchased from Cell Signaling.
Cell culture transfection and immunohistochemistry
Culturing and transfection of C2C12 myoblasts with 2 μg of RNA interference (RNAi) vector (pcDNA 6.2-GW/EmGFP-miR; Invitrogen, Grand Island, New York, USA) was performed as previously described (Goodman et al. 2010). The specific RNAi target sequences used in the vectors include the following. TSC2: CCGCTGGACTACAAGTGCAACCTAT; mTOR: TAAAGAGCTCGTCCACCCACT; and scramble: GTCTCCACGCAGTACATTT. The transfected myoblasts were plated on chamber slides (Nunc-Lab Tek), grown for 24 h in 10% fetal bovine serum (FBS)–Dulbecco's modified Eagle's medium (DMEM) antibiotic/antimycotic-free growth media and then switched to 10% FBS–DMEM with antibiotics and antimycotics for an additional 48 h. Immunohistochemistry was performed as previously described in Sancak et al. (2010) using mTOR(7C10) (no. 2983; Cell Signaling) or TSC2(D93F12) (no. 4308S; Cell Signaling) as primary antibodies, and Dylight 594-conjugated anti-rabbit (no. 111-515-144; Jackson Immuno-Research, West Grove, PA, USA) as the secondary antibody.
Skeletal muscle immunohistochemistry and image acquisition
Upon collection, TA muscles were immediately submerged in OCT (Tissue-Tek; Sakura, Torrance, CA, USA) and frozen in liquid N2-chilled isopentane. Semi ultra-thin sections (2 μm) were taken at the mid-belly and fixed in −30°C acetone for 10 min. After incubating in PBS for 15 min, the sections were then incubated in solution A (PBS with 5% normal goat serum (Jackson Immuno-Research) and 0.3% CHAPS) for 1 h. After three 5 min washes with PBS, samples were incubated for 1 h with solution B (PBS with 0.5% BSA and 0.3% CHAPS) containing primary antibodies specific for either TSC2 or mTOR, LAMP2 (ab13524, Abcam Cambridge, MA, USA) and laminin (ab14055, Abcam). After three 5 min PBS washes, samples were incubated for 1 h with solution A containing secondary antibodies (DyLight 594-conjugated goat anti-rabbit IgG (no. 111-515-144, for TSC2 or mTOR), Dylight 488-conjugated goat anti-rat IgG (no. 112-485-167, for LAMP2) and Dylight 405-conjugated goat anti-chicken (no. 103-475-155, for laminin); Jackson Immuno-Research). Samples were then washed 3 times for 10 min with PBS. Images of the different fluorophores were captured with a Nikon DS-QiMc camera that was mounted to a Nikon 80i epifluorescence microscope equipped with a ×60 water immersion objective (Nikon, Tokyo, Japan). The exposure times used for each image were determined by the auto-exposure feature in Nikon's NIS-Elements D software and all images were captured and analysed by an individual who was blinded to the sample identification.
Frequency scatterplots that compared the intensity of the signal for LAMP2 versus the intensity of the signal for either mTOR or TSC2 within every pixel of the images were generated with the WCIF ImageJ plug-in (http://www.uhnresearch.ca/facilities/wcif/imagej/). The ImageJ JACoP plugin (http://rsbweb.nih.gov/ij/plugins/track/jacop.html) was then used to quantify the number of pixels in each image that were considered to be as intensely positive for both LAMP2 and mTOR or TSC2. Specifically, cytofluorograms of each image were generated with JACoP and then the number of pixels exceeding an intensity threshold of 60 relative light units for both mTOR or TSC2 and LAMP2 were counted as ‘intense colocalized pixels’.
All values reported in the manuscript and in the figures are expressed as means + SEM. Statistical significance was determined using Student's t test and differences between groups were considered significant when P≤ 0.05. All statistical analyses were performed on SigmaStat software (San Jose, CA, USA).
Eccentric contractions induce an increase in TSC2 phosphorylation
To determine whether the mechanical activation of mTOR signalling is associated with alterations in the phosphorylation of TSC2, mouse TA muscles were subjected to mechanical stimulation by means of a bout of eccentric contractions (ECs). Consistent with previous studies, we found that ECs induced a robust increase in mTOR signalling as assessed by changes in the phosphorylation of p70S6k on the threonine 389 residue (O’Neil et al. 2009). It was also determined that ECs induced an upward mobility shift in TSC2, while the total amount of TSC2 was not significantly altered (Fig. 1A).
Upward mobility shifts are often observed in proteins that become hyper-phosphorylated (e.g. total p70 in Fig. 1). Therefore, to determine whether the upward mobility shift in TSC2 resulted from increased phosphorylation, we immunoprecipitated TSC2 and then incubated the immunoprecipitates with, or without, λ phosphatase. As shown in Fig. 1B, λ phosphatase eliminated the EC-induced upward mobility shift indicating that ECs promote an increase in the phosphorylation of TSC2.
Previous reports have shown that TSC2 can be phosphorylated on a variety of different sites and many of these sites lie within a consensus motif (RxRxxS*/T*, where * represents the phosphorylated residue and x is any amino acid) that is utilized by several members of the AGC family of kinases (Huang & Manning, 2008; Pearce et al. 2010). Thus, in order to further confirm that ECs induce an increase in TSC2 phosphorylation, we probed immunoprecipitates of TSC2 with an antibody that recognizes the phosphorylated RxRxxS*/T* consensus motif. Our findings revealed that ECs induced a >6-fold increase in the RxRxxS*/T* phosphorylation of TSC2 (Fig. 1C).
Eccentric contractions induce translocation of TSC2
Phosphorylation of TSC2 on residues that lie within RxRxxS*/T* consensus motifs have been implicated in the regulation of TSC2 localization. For example, subcellular fractionation experiments have demonstrated that growth factors cause TSC2 to translocate from a membranous to cytosolic fraction, and this effect can be prevented with non-phosphorylatable mutations of RxRxxS*/T* consensus motif residues (Cai et al. 2006; Miyazaki et al. 2010). Based on these findings, we set out to determine whether ECs also cause TSC2 to translocate away from membranous structures. As shown in Fig. 2, our results indicated that the relative proportion of TSC2 found in membranes was reduced by 20% following a bout of ECs (P≤ 0.05). On the other hand, ECs did not significantly alter the amount of mTOR in the membranes.
Eccentric contractions induce dissociation of TSC2 from the lysosome
Next, we used immunohistochemistry (IHC) to further define how ECs alter the subcellular localization of TSC2. Specifically, we first validated the specificity of our antibody for the detection of TSC2 under IHC conditions by transfecting C2C12 myoblasts with a vector that expresses the green fluorescent protein (GFP) and an RNAi against TSC2, or GFP and a scramble (control) RNAi. As shown in Fig. 3, the TSC2(D93F12) antibody produced a strong punctate signal in both non-transfected cells and in cells that had been transfected with control RNAi vector (Fig. 3A and B). More importantly, the immunoreactivity of this antibody was effectively eliminated in cells transfected with the TSC2 RNAi vector (Fig. 3C and D).
Having validated our TSC2 antibody, we then performed IHC analyses on both control and EC muscles. In control muscles, TSC2 was widely distributed throughout the fibres in weakly stained fine punctate structures, as well as in more intensely stained large punctate structures (Fig. 3G). The distribution pattern of the large intense puncta was markedly similar to a pattern that we had observed with the lysosomal marker LAMP2. Therefore, to determine whether the large intense punctate structures of TSC2 were colocalized with the lysosome, we stained control muscles with TSC2, LAMP2 and laminin as a marker of the sarcolemma. The results indicated that the large intense TSC2 puncta were highly colocalized with the signal for LAMP2 (Fig. 3E–G). Remarkably, when the same IHC analyses were performed on muscles that had been subjected to ECs, the large intense TSC2 puncta were essentially undetectable (Fig. 3H–J). To graphically illustrate this effect, we generated frequency scatterplots that compared the staining intensity of LAMP2 and TSC2 within every pixel from 23 randomly selected images per group (Fig. 3K and L). We also generated frequency scatterplots for each individual sample and then counted the number of pixels that exceeded a threshold value that represented the typical strength of the signal observed in the intense colocalized TSC2/LAMP2 puncta. The results from these analyses revealed that ECs induced a 90% decrease in the number of intense colocalized pixels (Fig. 3M). In other words, the results indicated that ECs dramatically reduce the amount of TSC2 that colocalizes with the lysosome.
Eccentric contractions enhance the association of mTOR with the lysosome
Presumably, the EC-induced dissociation of TSC2 from the lysosome would inhibit the ability of TSC2 to function as a GAP for lysosome-associated Rheb. As a result, the lysosome-associated Rheb would accumulate in its active GTP-bound state and, in turn, promote the activation of lysosome-associated mTOR. We reasoned that such events might be part of the mechanism via which ECs activate mTOR, but in order for this mechanism to exist, mTOR would have to be associated with the lysosomes in muscles that had been subjected to ECs. However, to date, the precise subcellular location of mTOR in skeletal muscle has not been established. Therefore, we used IHC to determine whether mTOR was associated with the lysosomes in skeletal muscle. In these experiments, we first validated the specificity of our mTOR antibody by transfecting C2C12 myoblasts with an RNAi against mTOR. As shown in Fig. 4, the mTOR(7C10) antibody produced a strong punctate signal in both non-transfected cells and in cells that had been transfected with a scrambled (control) RNAi vector (Fig. 4A and B). Furthermore, we found that the immunoreactivity of this antibody was abolished in cells transfected with the mTOR RNAi vector (Fig. 4C and D). Next, we stained both control and EC muscles for mTOR, LAMP2 and laminin. The results indicated that, in control muscles, mTOR was distributed throughout the fibres in intense punctate structures and these structures were typically, although not always, colocalized with the signal for LAMP2 (Fig. 4E–G). The same general observations were also made in muscles subjected to ECs. However, a subtle, but potentially very important increase in the degree of colocalization between mTOR and LAMP2 was observed (Fig. 4H–J). To visualize this effect, we generated frequency scatterplots (Fig. 4K and L), and we quantified this effect with the same approach that was used to assess TSC2/LAMP2 colocalization (Fig. 4M). The results from these analyses revealed that ECs induce a 67% increase in the number of intense colocalized pixels (Fig. 4M). Hence, it can be concluded that ECs promote an enhanced association of mTOR with the lysosome.
Based on our data, and those of others, we have developed a conceptual model of how mechanical stimuli might activate mTOR signalling (Fig. 5). Specifically, we propose that in the resting state, skeletal muscle lysosomes are enriched with PA, mTOR, Rheb and TSC2. In this state, the presence of TSC2 keeps the lysosome-associated Rheb in its inactive GDP-bound state, and thus, signalling by mTOR is relatively inactive. In response to mechanical stimulation, TSC2 becomes hyper-phosphorylated, and it is proposed that this change in phosphorylation causes TSC2 to dissociate from the lysosomes. As a result, the lysosome-associated Rheb is now able to obtain its active GTP-bound state and, in turn, promote the activation of the lysosome-associated mTOR. Furthermore, mechanical stimulation increases the association of mTOR with the lysosomes, and thus, the potential for mTOR to interact with one of its direct activators (PA and GTP-Rheb) is enhanced. Hence, when combined, the events described in this model might explain how mechanical stimuli can induce a robust activation of mTOR signalling.
In addition to potentially explaining how mechanical stimuli activate mTOR, our conceptual model also highlights several gaps in our current knowledge. For example, previous cell culture studies have shown that lysosomes are particularly enriched with Rheb and PA (Saito et al. 2005; Zhao et al. 2012). Yet, the subcellular localization of these molecules in skeletal muscle has not been defined, and it is not known if mechanical stimuli alter the GTP-bound status of Rheb. Furthermore, we have proposed that the mechanically induced increase in the phosphorylation of TSC2 causes it to dissociate from the lysosomes. This idea is supported by previous studies which have shown that the phosphorylation of TSC2 on residues that lie within RxRxxS*/T* consensus motifs (e.g. S939 and S981) can regulate its association with the membrane (Cai et al. 2006). However, it has also been suggested that the phosphorylation of the S939 and S981 residues is controlled by a pathway that involves the phosphatidylinositol 3-kinase (PI3K)-dependent activation of protein kinase B (PKB) (Dan et al. 2002; Cai et al. 2006). This is a potentially important point because we have previously demonstrated that the mechanical activation of mTOR signalling occurs through a PI3K/PKB-independent mechanism (Hornberger et al. 2004; O’Neil et al. 2009). Hence, if mechanical stimuli utilize the S939 and S981 residues to control the localization of TSC2 and mTOR signalling, then the mechanically induced phosphorylation of these residues would be expected to occur through a PI3K/PKB-independent pathway. Alternatively, mechanical stimuli might utilize different phosphorylation site(s) to control the localization of TSC2. With this point in mind, it is important to consider that the exact residues that become phosphorylated in response to mechanical stimulation are currently not known, and thus, the functional role of these phosphorylation events remains to be determined.
Lastly, it has been well established that a variety of different types of stimuli utilize the Rheb-GAP activity of TSC2 to control mTOR signalling (Huang & Manning, 2008; Tomasoni & Mondino, 2011). Yet, the molecular mechanism(s) that control the ability of TSC2 to function as a GAP for Rheb have not been resolved. Our results have potentially shed light on this mechanism by demonstrating that the mechanical activation of mTOR signalling is associated with the translocation of TSC2 away from the lysosomes. To the best of our knowledge, changes in the association of TSC2 with the lysosome have never been described, and a mounting body of evidence suggests that this type of event could serve as a fundamental regulatory mechanism that is used to control mTOR signalling.
All of the work in this study was conducted in the laboratory of T. A. Hornberger. B.L.J., J.-S.Y., J.W.F., C.A.G. and T.A.H. contributed to the conception/design of experiments; B.L.J., J.-S.Y., J.W.F., C.A.G., D.M.G. and T.A.H. contributed to the collection/analysis of data; B.L.J. and T.A.H. wrote the paper. All authors approved the final submitted version.
This work was supported by the National Institutes of Health grant AR057347.