Intracellular expression profiling by laser capture microdissection: three novel components of the neuromuscular junction

Javad Nazarian1,2, Khaled Bouri2 and Eric P. Hoffman1,2

1 The Institute for Biomedical Sciences, George Washington University
2 Center for Genetic Medicine, Children’s National Medical Center, Washington, District of Columbia


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
The neuromuscular junction (NMJ) is a regionally specialized area of myofibers defined, in part, by specific gene expression from underlying myonuclei. We sought to obtain a more complete picture of the mRNA transcripts and proteins playing a role in NMJ formation and maintenance using laser capture microdissection (LCM) and to define expression profiles of the nuclear domain at the NMJ. NMJs (800) were isolated from normal mouse tibialis anterior muscle by LCM, with an equal amount of adjacent non-NMJ regions isolated. Many known components of the NMJ were found significantly differentially expressed. Three differentially expressed potential novel components of the NMJ were chosen for further study, and each was validated by immunostaining with and without blocking peptides (3/3), quantitative RT-PCR (3/3), and in situ hybridization (1/3). The three genes validated were dual-specificity phosphatase-6 (DUSP6), ribosomal receptor-binding protein-1 (RRBP1), and vacuolar protein sorting-26 (VPS26). Query of each of these novel components in a 27-time point in vivo muscle regeneration series showed expression commensurate with previously known NMJ markers (nestin, {alpha}-ACh receptor). Understanding and discovering elements responsible for the integrity and function of NMJs is relevant to understanding neuromuscular diseases such as spinal muscular atrophy. Our LCM-based mRNA expression profiling provided us with new means of identification of specific genes potentially responsible for NMJ stability and function and new candidates for involvement in disease pathogenesis.

dual-specificity phosphatase-6; ribosomal receptor-binding protein-1; ES130; vacuolar protein sorting-26; nuclear domain


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
A MATURE MYOFIBER typically contains >100 myonuclei, with three regional "myonuclear domains" identified: neuromuscular junction (NMJ), myotendinous junction (MTJ), and extrajunctional regions (XNMJ) (12). The mRNA expression patterns of these three myonuclear domains are thought to be relatively distinct, although they change as a function of development and can show interdomain communication (1, 3, 31).

The NMJ has been considered a classical model for the establishment and maintenance of nuclear domains and the extent to which gene expression is restricted at these regions. A classical example of such restricted gene transcription is the almost exclusive expression of acetylcholine (ACh) receptor (AChR) subunits at the NMJ (13). Although NMJs consist of <0.1% of the muscle fiber surface, they contain >95% of the AChRs (43).

There are three developmental phases involved in the formation of the NMJ. During myofiber development or regeneration, before innervation, certain NMJ components are expressed throughout the muscle fiber (e.g., nicotinic AChRs). Once a motor neuron contacts the myofiber, expression of these components is suppressed at noncontact regions, leading to highly localized, restricted myonuclear domains for these components. Synapse-specific transcription program by the subsynaptic nuclei is initiated along with the transcriptional arrest of NMJ-specific genes (such as AChRs) in non-NMJ myonuclei (3234).

A key event in the localization and stabilization of the NMJ is the release of agrin by the motor neuron terminal. Agrin, a proteoglycan released by the motorneuron, induces AChR clustering (25). Agrin-deficient knockout (KO) mice lack differentiated NMJs and die soon after birth due to respiratory failure (14). Muscle-specific kinase (MuSK), a receptor tyrosine kinase that undergoes autophosphorylation (48), is the main component of the agrin-binding complex (20). Signal-transducting molecules are then thought to bind to these sites and propagate the necessary signaling cascade, although these downstream pathways are poorly defined. As with many receptor tyrosine kinases, MuSK does not bind agrin directly but through a poorly defined "myotube-associated specificity complex" (MASC) (15). Other neurotrophic factors known as ARIA [AChR-inducing activity, or neuregulin (NRG)] also activate transcription of AChR genes at the NMJ myonuclei (34, 35). Rapsyn is also an NMJ-specific protein that interacts with AChRs and directly binds to the cytoplasmic domain of ß-dystroglycan (4, 8). Further studies have indicated that rapsyn is not required for MuSK phosphorylation (2). Therefore, the nature of protein kinases and phosphatases responsible for MuSK activation and regulation remains unknown. Apel et al. (2) suggested that a putative rapsyn-associated transmembrane linker protein (RATL) may facilitate rapsyn-MuSK linkage.

Upon formation and stabilization of the NMJ, the concentration of AChRs at the postsynaptic membrane reaches up to 10,000 molecules/µm2, whereas it is reduced to as low as 10/µm2 molecules at the XNMJ (33). Thus the formation and stabilization of the NMJ myonuclear domain is critical for appropriate contact and signaling between the motor neuron and myofiber. Considerable progress has been achieved in defining many of the protein-protein signaling events during this process (22), but less is known regarding the protein components of the NMJ and the transcriptional events at this region.

The specific and comprehensive transcription program of NMJ-associated nuclei (of motorneuron, myofiber, and Schwann cells) has not been previously established because of many obstacles in dealing with NMJs. The main obstacles in doing so have been the relatively small size of the NMJ (35–90 µm) and the low abundance of these in muscle tissue (1 per myofiber, where the myofiber itself stretches from tendon to tendon). We hypothesized that we could define the specific transcription program of the NMJ by combining emerging technologies, namely laser capture microdissection (LCM) coupled with RNA amplification Affymetrix GeneChip arrays. Here we report technique development [isolation of NMJs by LCM, and mRNA profiling of small RNA amounts (>0.5 ng)] and definition of candidates of mRNAs corresponding to the NMJ-associated nuclei. We also present protein validation of three of these potential NMJ-localized transcripts.


    MATERIALS AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Mice
Eight (2-mo-old C57Bl/10) mice were euthanized by cervical dislocation, and tibialis anterior (TA) muscles were collected. The tissue was immediately frozen in isopentane cooled in liquid nitrogen and stored at –80°C. Muscle samples were used for laser microdissection (2 mice), immunostaining, and cDNA synthesis (6 mice). All animal procedures and experiments complied fully with the principles set forth in the "Guide for the Care and Use of Laboratory Animals" prepared by the Committee on Care and Use of Laboratory Animals of the Institute of Laboratory Resources, National Research Council, and were approved by the Children’s National Medical Center’s Institutional Animal Use and Care Committee.

Tissue Dehydration and NMJ Collection by LCM
Mouse TA muscle was sectioned (8-µm thick) on RNase Away (Ambion, Austin, TX)-treated Superfrost Plus (Fisher Scientific, Houston, TX) microscope glass slides. NMJs were then stained for acetylcholinesterase (see below). Slides were then washed with 70% ethanol for 30 s and four rounds of 100% ethanol for 1 min each, placed in xylene for 5 min, and finally placed in hood to air dry for 30 min.

Dehydrated tissues were used immediately for NMJ and XNMJ microdissection. Arcturus HS CapSure caps (Arcturus, Mountain View, CA) were used to collect cells. LCM was operated as described by the manufacturer.

NMJ Identification by Acetylcholinesterase Staining
Karnowsky and Root’s method was used for NMJ localization. Briefly, acetylcholinesterase was stained during an enzymatic process, using acetylthiocholine iodide. Five milligrams of acetylthiocholine iodide were dissolved in 0.1 M sodium acetate buffer (pH 5.2). To this solution, 0.5 ml of 0.2 M sodium citrate, 1 ml of 30 mM cupric sulfate, 1 ml of H2O, and 1 ml of 5 mM potassium ferricyanide were added and mixed. Three hundred microliters of this solution were placed on sectioned tissues for 25 min until sufficient red/brown color developed.

RNA Amplification and Expression Profiling
Roughly 1,000 NMJs were microdissected per experiment. Stratagene (La Jolla, CA) Nanoprep RNA isolation kit was used to isolate RNA from the caps. Isolated total RNA was then subjected to two rounds of T7-based cRNA amplification to yield biotin-labeled cRNA.

Briefly, 50–100 ng of total RNA were converted to double-stranded cDNA, using a T7-promoter primer. This cDNA was then used as a template to derive the production of cRNA using the Megascript T7 in vitro transcription kit (Ambion, Austin, TX). For the second-round amplification, 200 ng of the first round-amplified cRNA were used to synthesize double-stranded cDNA, first using a random primer to drive the first-strand synthesis and then using the T7 promoter/primer to derive the second-strand synthesis. This amplified cDNA was used to synthesize biotinylated cRNA from the T7 promoter using the Enzo BioArray High Yield RNA transcription labeling kit (Enzo, New York, NY). Biotin-labeled cRNA was then cleaned up using RNeasy minicolumns. Twelve to sixteen micrograms of labeled cRNA were then fragmented and hybridized to MG-U74A v.2 Affymetrix GeneChip probe arrays. Hybridization and detection of hybridized probes were performed as described in the GeneChip Expression Analysis Technical Manual. Present call for the four U74A v.2 chips were 23, 29, 25.9, and 26.9%. The GAPDH 3'/5' ratios were 5.3, 1.4, 5.52, and 7.4, respectively. All profiles of this experiment are accessible via a public website (http://www.ncbi.nlm.nih.gov/geo/; accession nos. GSM-38542, GSM-38543, GSM-38545, GSM-38544).

Quantitative Multiplex Fluorescence-RT-PCR
Total RNA was isolated from both NMJ and XNMJ of three mice, using LCM as described before, and amplified one time to obtain cRNA. About 200 ng of cRNA were then used to produce single-stranded cDNA, and then 10 ng of cDNA were input into each reaction for quantitative RT-PCR.

The RT-PCR method involves the coamplification of both a control (bone morphogenesis protein-4; BMP4) and an experimental mRNA in the same reaction (53). PCR reactions were done and the products resolved using automated sequencers. Infrared fluorescent-labeled (LiCor, Lincoln, NE) and nonfluorescent-labeled primers (Invitrogen, Carlsbad, CA) were designed and synthesized according to Supplemental Table S1 (available at the Physiological Genomics web site).1 The RT-PCR mixtures included 1.0 µl of cDNA, 10x PCR buffer, dNTPs, each primer (1 pmol), and Taq Gold DNA polymerase (Perkin Elmer, Foster City, CA). Cycling conditions were 94°C for 12 min and 24–27 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 1 min. Reactions were then mixed with denaturing solution and denatured at 95°C for 5 min. Denatured solutions were then run on 5% polyacrylamide gels, using LiCor automated sequencers. The LiCor sequencer scanner allowed detection of the intensity of the fluorescent signal for each product. The Scanalytics (Fairfax, VA) software provided a nanogram intensity value for each band corresponding to the quantitative measure of the relative levels of mRNA. The quantity of the expression level for each experimental gene vs. each control gene was calculated, based on the ratio between the intensity value of the experimental gene and the control sample (BMP4), for a total of three comparisons for the NMJ group and three for the XNMJ group (see Fig. 5). Student’s t-test was performed to generate P values between NMJ and XNMJ for each experimental gene.

Immunohistochemistry
Polyclonal antibodies have been used successfully by others (7) for detecting NMJ-associated molecules such as nestin. To increase specificity and reduce unspecific staining, we chose to use affinity-purified polyclonal antibodies raised against synthetic peptides. Polyclonal rabbit anti-mouse antibodies were generated by Bethyl Laboratories (Montgomery, TX). The following peptides were used to produce antibodies: dual-specificity phosphatase-6 (DUSP6), CDIESDLDRDPNSATDSDGSPLS and CDNRVPTPQLYFTTPSNQNVYQVDSLQST; ribosomal receptor-binding protein-1 (RRBP1), CKTKKKEEKPNKIPEHD and CQEAPKQDAPAKKKSGSRKKG; vacuolar protein sorting-26 (VPS26), CGETRKMAEMKTEDGKVEKHY and CKLRKQRTNFHQRFESPDSQASAEQ. Briefly, peptides were synthesized and conjugated to keyhole limpet hemocyanin (KLH) before being injected into rabbits. Hyperimmune serum from rabbits was processed over the respective immunosorbent assay to capture antibodies specific for each peptide. {alpha}-Bungarotoxin antibody was purchased from Calbiochem (San Diego, CA) and cy3-conjugated donkey anti-rabbit secondary antibody was purchased from Jackson Laboratory (Bar Harbor, ME).

Serial 5-µm-thick frozen mouse TA muscle sections were cut with cryostat, mounted on Superfrost Plus slides (Fisher Scientific), and fixed in cold anhydrous acetone. Sections were then blocked for 15 min in 5% horse serum and 1x PBS and incubated with primary antibody (1:10,000) for 1 h at room temperature. Washes (3 for 5 min each) were done with 5% horse serum and 1x PBS, and sections were then incubated with secondary antibody for 1 h.

In Situ Hybridization
The design of oligonucleotide probes for DUSP6 was done with the Helios BioSciences (Paris, France) software OligoETC. Four distinct, nonoverlapping anti-sense probes and the corresponding sense control probes were synthesized based on the DUSP6 sequence; all were confirmed for uniqueness via comparison to the murine genome. Frozen sections of mouse TA muscle were first postfixed in paraformaldehyde, histochemically stained for acetylcholinesterase (50), and then hybridized for in situ detection of DUSP6 mRNA. The four anti-sense oligonucleotides were end-labeled with 35S-nucleotides, using terminal deoxynucleotide transferase (TdT), mixed, and cohybridized to frozen sections. Parallel sections were hybridized to the control sense mix of four labeled oligonucleotides.

Hybridized and washed slides were then coated with emulsion, exposed for 2 wk, and developed. Slides were counterstained with hematoxylin and eosin.


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
NMJ Isolation and Expression Profiling
To isolate NMJs, we employed LCM (Arcturus). Briefly, longitudinal frozen sections of TA muscle were mounted on microscope slides and dehydrated. A histochemical stain was done for ACh-esterase (AChE), with the complete staining procedure reduced to 10 min to minimize mRNA degradation (Fig. 1A). A "cap" containing a thin thermoplastic (ethylene vinyl acetate) film was then placed on the stained sections, a low-power laser beam was directed to 7.5-µm areas stained positively with AChE, and three laser pulses were used to collect each NMJ (Fig. 1). We found two technical hurdles to requiring optimization. One was appropriate dehydration (see MATERIAL AND METHODS). The second was nonspecific adherence of pieces of tissue to the cap. Arcturus (Mountain View, CA) provides adherence pads (CapSure cleanup pads) to aid in cleaning undesired material from the caps. However, we found these pads insufficient for NMJs. First, these highly adhesive pads removed NMJs along with the unwanted tissues. Second, the pads are opaque, making it impossible to observe the specific regions eliminated from the cap using the microscope. We overcame this obstacle by using transparent Post-it adhesive notes (3M, St. Paul, MN). The relatively low degree of adhesiveness of Post-it notes permitted the removal of loosely adhering nonspecific tissue. At the same time, the transparency of the material allowed the visualization of the cap surface so the operator can choose the area to be cleaned (Fig. 2, A–D). We also found that mRNA integrity was best when the tissue processing through to RNA isolation was done in 4 h or less.



View larger version (121K):
[in this window]
[in a new window]
 
Fig. 1. Isolation of neuromuscular junctions (NMJs) stained for acetylcholinesterase by laser capture microdissection (LCM). A: histochemical stain for acetylcholinesterase. B: laser pulses were used to adhere NMJs to an overlying plastic film. C and D: lifting of the plastic cap removes the NMJ regions from the myofibers (C), with the NMJs adhering to the cap (D). Material is then used for mRNA isolation and expression profiling. Scale bar = 30 µm.

 


View larger version (106K):
[in this window]
[in a new window]
 
Fig. 2. Removal of unwanted tissue from LCM caps. A: a plastic cap after LCM, with both desired NMJ material (blue arrows) and nondesired adherent tissue (black arrows). B–D: a transparent Post-it note is lightly adhered to the cap (B) and then lifted such that the unwanted material is transferred to the transparent Post-it (C), leaving only desired NMJs on the LCM cap (D).

 
A quality control metric for Affymetrix arrays is the percentage of "present" calls (44). This metric measures the percentage of probe sets on the microarray where the signal intensity difference between the 11 perfect-match and 11 paired-mismatch probes is above a specific threshold. This is a method of approximating background or nonspecific hybridization to probe sets, and values of 30–40% present calls are considered adequate signal-to-noise ratios when one round of amplification is used. In the case of two rounds of amplification, as presented here, the percentage of present calls is typically not as high as with one round. All of the microarrays presented here showed a percentage present call >23%. A second metric is an approximate measure of cDNA integrity, assessed by relative expression of probe sets at the 3'- and 5'-ends of the same transcript. Typical measures of GAPDH 3'/5' ratios for one round of amplification are <3.0. Again, these quality control metrics must be relaxed for LCM and two rounds of amplification. All of our samples showed GAPDH 3'/5' ratios of <7.4, with an average of 4.9 (see MATERIALS AND METHODS).

We hybridized four Affymetrix U74A v.2 GeneChips (2 NMJs and 2 non-NMJs) and queried 12,000 genes. Signal intensities for each probe set were generated using Microarray Suite (MAS)5. Signal intensity values were compared by pairwise analyses of the four microarray profiles, and only those probe sets showing at least one "present call" were retained for further analyses (4,633/12,424 probe sets; 37% of total probe sets). NMJ-enriched transcripts were defined as those that showed higher expression (>2-fold) in the NMJ profiles relative to non-NMJ. This analysis generated a list of 143 genes that were then correlated with fold-change approximations (Table 1).


View this table:
[in this window]
[in a new window]
 
Table 1. Transcripts enriched at the NMJ, using expression profiling

 
We first interrogated the 143 gene list for documented NMJ-associated transcripts and/or proteins. {epsilon}-AChR, {alpha}-AChR, and nestin were each known NMJ-enriched transcripts that were detected as significantly enriched in the NMJ expression profiles (Table 1, top). We also used a less stringent approach, where MAS4 probe set signals were compared with the use of pairwise analyses of the four microarray profiles (Supplemental Table S2). This nonstatistical analysis showed more extensive gene lists showing strong fold changes between NMJs and non-NMJs and included additional known NMJ-associated transcripts (acetylcholinesterase, nestin, and {alpha}-dystrobrevin). Nestin is a member of intermediate filament (IF) protein family (24). Initial studies had indicated that nestin expression is high during early developmental stages, whereas it decreases in adult myofiber (39). However, Vaittinen et al. (45) showed that nestin localizes exclusively to MTJs and NMJs.

We prioritized differentially expressed transcripts by focusing on those that showed strong enrichment at the NMJ, as suggested by the strongest P values and highest fold changes. We further prioritized this subset by querying for muscle specificity in GenBank [e.g., expressed sequence tags (ESTs) originally identified in muscle, heart, or both] and by literature inquiries regarding known or possible function and possible relevance to nerve-muscle interaction. Three transcripts were chosen for further validation: DUSP6, RRBP1, and VPS26.

Quantitative Multiplex Fluorescence-RT-PCR and Immunostaining Assays
To validate our expression profile data, we performed quantitative multiplex fluorescence (QMF)-RT-PCR using infrared primers (11, 53). Initially, we selected nestin, a known NMJ-associated gene, as a positive control for QMF-RT-PCR analysis. Nestin is known to be highly enriched at the NMJ (45). To select a control gene that is equally expressed in NMJ and XNMJ, we analyzed our data with GeneSpring software (Silicon Genetics, Redwood City, CA). On the basis of signal intensities and present/absent calls, we identified transcripts that were of similar signal to our test transcripts, yet were predicted to be equally expressed at both NMJ and XNMJ from the microarray data. From this list, we selected BMP4 as the control gene. We used LCM to isolate roughly 400 NMJs and an equal area of the non-NMJs. RNA was isolated, and single-strand cDNA was synthesized and used as templates for PCR. QMF-RT-PCR analysis confirmed the high expression of nestin at the NMJs (Fig. 3), whereas BMP4 (the control gene) was shown to be expressed equally between NMJ and non-NMJ. We then performed QMF-RT-PCR on the selected transcripts and validated the high expression of DUSP6, RRBP1, and VPS26 (Fig. 3).



View larger version (47K):
[in this window]
[in a new window]
 
Fig. 3. Quantitative multiplex fluorescence (QMF)-RT-PCR validates enrichment of dual-specificity phosphatase (DUSP)6, ribosomal receptor-binding protein (RRBP)1, and vacuolar protein sorting (VPS)26 at the NMJ. A: electrophoretogram of QMF-RT-PCR for each experimental transcript relative to an internal control [bone morphogenesis protein (BMP)4]. A positive control, nestin, shows enrichment to the NMJ similar to that seen with the 3 novel NMJ transcripts (DUSP6, RRBP1, and VPS26). B: bar charts of 3 replicate QMF-RT-PCR quantitations. Each of the 3 experimental transcripts showed significant enrichment at the NMJ vs. adjacent extrajunctional material (XNMJ). *P > 0.05. **P > 0.01.

 
To further validate the expression profiling data, we produced affinity-purified polyclonal antibodies to the three selected genes (DUSP6, VPS26, and RRBP1). For each transcript, two nonoverlapping peptides were designed against hydrophilic regions that showed little or no homology to other proteins in the National Center for Biotechnology Information GenBank. The peptides were conjugated to KLH and coimmunized into rabbits, with resulting sera affinity purified using each immobilized peptide separately. The affinity-purified peptide antibodies were then tested for specificity and sensitivity, using primarily immunostaining of mouse TA muscle, although DUSP6 antibodies were also tested by immunoblot using Torpedo electroplaques and mass spectrometer identification (data not shown). DUSP6 was also validated by in situ hybridization, as described below.

DUSP6.
DUSP6 is a 42-kDa cytoplasmic protein that dephosphorylates the "classical" growth factor-activated MAPKs ERK1 and ERK2 but not stress-activated MAPKs (17, 27). Dual-specificity protein phosphatases (DSPs) have been shown to catalyze the removal of phosphate groups form threonine and tyrosine residues, within a signature sequence of TXY, that are essential for the enzymatic activity (6).

With the use of expression profiling, DUSP6 was found to be highly expressed at the NMJ (7.2-fold) compared with the non-NMJ (Table 1). QMF-RT-PCR confirmed the high expression levels of this protein at the NMJ (Fig. 3). We also employed in situ hybridization assays and showed the localization of DUSP6 mRNA to the NMJ (Fig. 4, Aa and Ab), confirming our QMF-RT-PCR results, whereas the sense strand probes (negative control) failed to produce detectable signals (Fig. 4B). Immunostaining assays, using either of the two anti-peptide antibodies, showed that DUSP6 localized to the NMJ (Fig. 5Aa) in serial sections of mouse TA muscle compared with {alpha}-bungarotoxin ({alpha}-BTX) immunostaining (Fig. 5Ab).



View larger version (72K):
[in this window]
[in a new window]
 
Fig. 4. In situ hybridization shows localization of DUSP6 mRNA to the NMJ. Shown is histochemical staining for acetylcholinesterase (blue arrows), with exposed silver grains in the overlying emulsion (black arrows). Aa and Ab: hybridization of 4 DUSP6 anti-sense oligonucleotide probes labeled with 35S. B: hybridization of the sense (control) probes for the same regions of the DUSP6 sequence. Hybridization to the DUSP6 mRNA is seen overlying the NMJs only in the anti-sense probes (Aa and Ab). Scale bar = 50 µm.

 


View larger version (10K):
[in this window]
[in a new window]
 
Fig. 5. Immunostaining assays of serial sections show localization of 3 proteins to the NMJ. Immunostainings: DUSP6 (Aa), RRBP1 (Ba), and VPS26 (Ca). Ab, Bb, and Cb: {alpha}-bungarotoxin. Scale bar = 30 µm.

 
Our laboratory has previously identified transcriptional regulatory pathways specific to mouse muscle regeneration by generating temporal expression profiles (51, 52). Query of DUSP6 in the 27-time point in vivo muscle regeneration series showed transient increases in expression, as one would expect for a component of the NMJ (Fig. 6). Specifically, DUSP6 expression showed highest expression at 2–3 days during regeneration (5-fold increase relative to time 0 at 2 days postinjection). This preceded the upregulation of {alpha}-AChR (10-fold at 3 days) and nestin (4.5-fold at 3.5 days postinjection).



View larger version (16K):
[in this window]
[in a new window]
 
Fig. 6. The 27-time point in vivo muscle regeneration series shows a transient increase in the expression of known [nestin and {alpha}-acetylcholine receptor ({alpha}-AChR)] and novel (DUSP6, RRBP1, and VPS26) NMJ-associated genes. Query of DUSP6 showed highest expression at 2–3 days during regeneration (5-fold increase relative to time 0 at 2 days postinjection), which preceded the upregulation of {alpha}-AChR (10-fold at 3 days) and nestin (4.5-fold at 3.5 days postinjection). RRBP1 showed transient induction, with a peak expression at 2 days (>2-fold), whereas VPS26 was expressed at relatively low levels.

 
RRBP1.
RRBP1 (also called ES/130, p180) was originally reported as a heart morphogenesis inducer in chicken embryonic heart (30). RRBP1 is a membrane protein with a short luminal tail, with most of the protein projected into the cytoplasmic region (23). By expression profiling, we found high expression levels of RRBP1 at the NMJ compared with the non-NMJ (Table 1). This was confirmed by quantitative RT-PCR (Fig. 3). Of the two affinity-purified peptide antisera produced against RRBP1, one showed colocalization with BTX to the NMJ (Fig. 5, Ba and Bb), whereas the second showed high background staining in muscle and was not studied further. Query of RRBP1 in the 27-time point muscle regeneration series showed transient induction, with a peak expression at 2 days (>2-fold; Fig. 6). Similar to DUSP6, the RRBP1 transcriptional induction correlated roughly with that of nestin and {alpha}-AChR, although increased expression was sustained for a longer time period during regeneration.

VPS26.
Organelles in eukaryotic cells are maintained by their specific organelle sorting proteins such as VPS26. VPS26, a 38-kDa protein, is the mammalian homolog of yeast vacuolar protein sorting-26 (Vps26p). Microarray showed VPS26 to be highly enriched at the NMJ (Table 1), a finding confirmed by QMF-RT-PCR (Fig. 3). Immunostaining assays using both peptide antibodies showed that VPS26 localized to the NMJ in serial sections of mouse TA muscle compared with {alpha}-BTX immunostaining (Fig. 5, Ca and Cb). Secondary antibodies, primary antibodies mixed with competing peptides, and preimmune sera were also used as negative controls (data not shown). The use of preimmune sera did not stain the NMJs.


    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Identification of Potential Components of the NMJ by LCM and Expression Profiling
The NMJ has been studied as a prototypical synapse for decades, and considerable knowledge exists concerning molecular structure (34) and molecular function (22). Nevertheless, it is likely that many if not most components of the NMJ remain undiscovered. For example, two proteins/complexes have been hypothesized to exist, based on observed signaling cascades (MASC, RATL), yet the identity of these proteins or corresponding genes remains unknown (2, 15). Studies of the NMJs are complicated by the small size of the region, difficulties in applying standard fractionation techniques for NMJ isolations, and the nature of nuclear domains within a syntium that defines the NMJ transcriptome. Here we report use of LCM methods to isolate NMJs and compare the mRNA profiles of NMJ-enriched fractions with non-NMJ captured material.

Validation of DUSP6, RRBP1, and VPS26 as Novel Components of the NMJ
We focused on three potential novel components of the NMJ that were detected by LCM microarray analyses: DUSP6, RRBP1, and VPS26. To validate these three proteins as novel components of the NMJ, we first used quantitative RT-PCR from distinct LCM preparations (Fig. 3). Each of the three transcripts showed high enrichment within the LCM-collected NMJ fractions.

We then produced affinity-purified polyclonal peptide antibodies against the three proteins. Two nonoverlapping 18-mer peptides were designed against each protein, and affinity-purified antibodies were developed. Because of the low concentration of NMJ protein in total muscle homogenates, we focused on characterization of immunoreactive protein by immunostaining of frozen sections. All three proteins showed costaining with BTX, indicating preferential expression and localization to the NMJ (Fig. 5). In situ hybridization of DUSP6 mRNA also validated enrichment of this message to the NMJ (Fig. 4). We note that the NMJ material contains terminal glial cells as well as the motor neuron terminal and possible contaminating connective tissue and neighboring non-NMJ myonuclei. However, we feel that these three transcripts are not expressed in glia or motor neurons, as immunoblotting of sciatic nerve tested negative in all three (data not shown). Thus we conclude that these three proteins are novel components of the NMJ.

DUSP6.
Muda et al. (26), using Northern analysis, showed that DUSP6 expression level was high in lungs; low in heart, brain, spleen, liver, and kidney; and undetectable in skeletal muscle. This is not surprising, since the small size of NMJs could contribute to the lack of protein detection. Similarly, nestin, a component of the NMJ, was originally reported as absent in adult muscle (39).

DUSP6 is a member of a phosphatase family [dual-specificity (threonine/tyrosine) phosphatases; DSPs] known to bind and inactivate MAPKs through dephosphorylation of key sites within the MAPK proteins. MAPKs play an important role in mediating intracellular signaling events, mitogenic signal transduction, survival, stress response, and programmed cell death. In response to extracellular stimuli, MAPKs are activated by phosphorylation by threonine/tyrosine dual-specificity MAPK kinases (MAPKKs). Once activated, MAPKs activate a series of substrates that lead to changes in the transcription of specific genes.

DUSP6 (also called MKP3, PYST1, rVH6) is a 42-kDa cytoplasmic protein that dephosphorylates the "classical" growth factor-activated MAPKs, ERK1 and ERK2, but not stress-activated MAPKs (17, 27). ERK is downstream of both of the key NMJ components, agrin and neuregulin-1 (see above), and our data suggest that DUSP6 may be the relevant phosphatase that regulated the phosphorylation status of ERK1/2 at the NMJ. Interestingly, ERK2 has been shown to directly interact with the NH2-terminal noncatalytic domain of DUSP6 (17, 27).

Our data set up the hypothesis that DUSP6 may negatively regulate agrin- and neuregulin-activated kinases at the NMJ. ERK1 transmits downstream signals, in part, through Src homology (SH)2/SH3 domain-containing proteins, including one known as v-Crk. Transgenic mice overexpressing v-Crk show abnormal NMJs, consistent with the importance of ERK1 signaling for appropriate NMJ structure and function (29, 41, 42, 49). The kinases discussed here likely work at the level of signaling via protein phosphorylation, and such signaling processes are not measured by the mRNA profiling reported here. Further functional studies of these proteins are required to assess the role of DUSP6 at the NMJ, including its protein targets.

RRBP1.
The primary structure of RRBP1 consists of a hydrophobic NH2-terminus that includes a membrane-binding domain, a highly conserved basic tandem repeat (ribosome-binding domain), and an acidic coiled-coil COOH-terminal domain with an unknown function (47). There is recent evidence that RRBP1 may have a role in neuronal synapses, as it has been shown to bind KIF5B, and KIF5B in turn is known to bind synaptic proteins [synaptosomal-associated protein (SNAP)-24] at the NMJ (10, 46).

RRBP1 has been studied in yeast, where it is a member of the endoplasmic reticulum (ER) stress response and associated unfolded-protein response (UPR). In yeast, RRBP1 (p180) stabilizes mRNAs (5, 40, 47). Hyde et al. (18) showed that expression of exogenous p180 in yeast stabilizes mRNAs that are targeted to the secretory pathway, and p180-mediated stabilization was shown to be UPR independent. Expression of p180 was shown to regulate RNA stability at the level of mRNA turnover by targeting it to the ER membrane, rather than at the transcription level. This is achieved possibly by localizing mRNAs to the ER membrane and saving them from cell cycle-dependent degradation. Our data suggest the model where RRBP1 may be responsible for stabilizing the highly active gene transcriptional machinery of subjunctional myonuclei.

VPS26.
Organelles in eukaryotic cells are maintained by their specific organelle sorting proteins. Mannose-6-phosphate receptors (MPRs) mediate the vesicular transport to mammalian lysosomes by cycling between the trans-Golgi network (TGN) and a prelysosomal endosome (21). Cation-independent (CI) MPRs and cation-dependent (CD) MPRs are the two distinct MPRs that play a central role in the cycling process. In yeast, a complex of five proteins (Vps35p, 29p, 26p, 17p, and 5p) named "retromer" plays a role in the retrieval of some receptors (3638). Vps26p, which directly binds to Vps35p, plays an important role in stabilizing this retromer complex (28). Little is known about the molecular mechanisms regulating the retrieval of MPRs. Our observation that VPS26 is NMJ associated may suggest the role of this protein in NMJ stability by maintaining/retrieving NMJ-bound receptors. Therefore, VPS26 may play its role once the assembly of different NMJ complexes is accomplished. This may explain the relatively low expression of VPS27 in the muscle regeneration time series (Fig. 6).

Using an LCM-based expression profiling approach, we have identified three novel NMJ-associated proteins and provided many more potential candidates. A more complete understanding of the structure and function of the NMJ may lead to a better understanding of neurodegenerative diseases such as amyotrophic lateral sclerosis (ALS) (9). For example, familial ALS is often caused by a gain of function mutation of superoxide dismutase-1 (SOD1). However, it has become clear that overexpression and aggregation of mutant SOD1 in motorneurons and/or glia are not sufficient to cause the disease (16, 19). A recent report suggests that accumulation of mutant SOD1 at the NMJ may play an important role in disease pathophysiology (44a). Thus investigating the structure and function of the NMJ in SOD1 mutant mice could lead to an understanding of the specific perturbation of nerve-muscle interaction.


    GRANTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
This work was supported by the Erin Godla Fund for ALS Research (http://www.juvenileals.org) and the W. M. Keck Foundation.


    FOOTNOTES
 
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).

Address for reprint requests and other correspondence: E. P. Hoffman, Center for Genetic Medicine, Children’s National Medical Center, 111 Michigan Ave. NW, Washington, DC 20010 (E-mail: EHoffman{at}cnmcresearch.org).

10.1152/physiolgenomics.00227.2004.

1 The Supplemental Material for this article (Supplemental Tables S1 and S2) is available online at http://physiolgenomics.physiology.org/cgi/content/full/00227.2004/DC1. Back


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 

  1. Allen DL, Roy RR, and Edgerton VR. Myonuclear domains in muscle adaptation and disease. Muscle Nerve 22: 1350–1360, 1999.[CrossRef][ISI][Medline]
  2. Apel ED, Glass DJ, Moscoso LM, Yancopoulos GD, and Sanes JR. Rapsyn is required for MuSK signaling and recruits synaptic components to a MuSK-containing scaffold. Neuron 18: 623–635, 1997.[CrossRef][ISI][Medline]
  3. Apel ED, Lewis RM, Grady RM, and Sanes JR. Syne-1, a dystrophin- and Klarsicht-related protein associated with synaptic nuclei at the neuromuscular junction. J Biol Chem 275: 31986–31995, 2000.[Abstract/Free Full Text]
  4. Bartoli M, Ramarao MK, and Cohen JB. Interactions of the rapsyn RING-H2 domain with dystroglycan. J Biol Chem 276: 24911–24917, 2001.[Abstract/Free Full Text]
  5. Becker F, Block-Alper L, Nakamura G, Harada J, Wittrup KD, and Meyer DI. Expression of the 180-kD ribosome receptor induces membrane proliferation and increased secretory activity in yeast. J Cell Biol 146: 273–284, 1999.[Abstract/Free Full Text]
  6. Camps M, Nichols A, and Arkinstall S. Dual specificity phosphatases: a gene family for control of MAP kinase function. FASEB J 14: 6–16, 2000.[Abstract/Free Full Text]
  7. Carlsson L, Li Z, Paulin D, and Thornell LE. Nestin is expressed during development and in myotendinous and neuromuscular junctions in wild type and desmin knock-out mice. Exp Cell Res 251: 213–223, 1999.[CrossRef][ISI][Medline]
  8. Cartaud A, Coutant S, Petrucci TC, and Cartaud J. Evidence for in situ and in vitro association between ß-dystroglycan and the subsynaptic 43K rapsyn protein. Consequence for acetylcholine receptor clustering at the synapse. J Biol Chem 273: 11321–11326, 1998.[Abstract/Free Full Text]
  9. Dawbarn D and Allen SJ. Neurotrophins and neurodegeneration. Neuropathol Appl Neurobiol 29: 211–230, 2003.[CrossRef][ISI][Medline]
  10. Diefenbach RJ, Diefenbach E, Douglas MW, and Cunningham AL. The ribosome receptor, p180, interacts with kinesin heavy chain, KIF5B. Biochem Biophys Res Commun 319: 987–992, 2004.[CrossRef][ISI][Medline]
  11. Di Giovanni S, Knoblach SM, Brandoli C, Aden SA, Hoffman EP, and Faden AI. Gene profiling in spinal cord injury shows role of cell cycle in neuronal death. Ann Neurol 53: 454–468, 2003.[CrossRef][ISI][Medline]
  12. Engel AG and Franzini-Armstrong C. Myology. New York: McGraw-Hill, 1994.
  13. Fontaine B, Sassoon D, Buckingham M, and Changeux JP. Detection of the nicotinic acetylcholine receptor {alpha}-subunit mRNA by in situ hybridization at neuromuscular junctions of 15-day-old chick striated muscles. EMBO J 7: 603–609, 1988.[Abstract]
  14. Gautam M, Noakes PG, Moscoso L, Rupp F, Scheller RH, Merlie JP, and Sanes JR. Defective neuromuscular synaptogenesis in agrin-deficient mutant mice. Cell 85: 525–535, 1996.[CrossRef][ISI][Medline]
  15. Glass DJ, Bowen DC, Stitt TN, Radziejewski C, Bruno J, Ryan TE, Gies DR, Shah S, Mattsson K, Burden SJ, DiStefano PS, Valenzuela DM, DeChiara TM, and Yancopoulos GD. Agrin acts via a MuSK receptor complex. Cell 85: 513–523, 1996.[CrossRef][ISI][Medline]
  16. Gong YH, Parsadanian AS, Andreeva A, Snider WD, and Elliott JL. Restricted expression of G86R Cu/Zn superoxide dismutase in astrocytes results in astrocytosis but does not cause motoneuron degeneration. J Neurosci 20: 660–665, 2000.[Abstract/Free Full Text]
  17. Groom LA, Sneddon AA, Alessi DR, Dowd S, and Keyse SM. Differential regulation of the MAP, SAP and RK/p38 kinases by Pyst1, a novel cytosolic dual-specificity phosphatase. EMBO J 15: 3621–3632, 1996.[Abstract]
  18. Hyde M, Block-Alper L, Felix J, Webster P, and Meyer DI. Induction of secretory pathway components in yeast is associated with increased stability of their mRNA. J Cell Biol 156: 993–1001, 2002.[Abstract/Free Full Text]
  19. Jaarsma D, Haasdijk ED, Grashorn JA, Hawkins R, van Duijn W, Verspaget HW, London J, and Holstege JC. Human Cu/Zn superoxide dismutase (SOD1) overexpression in mice causes mitochondrial vacuolization, axonal degeneration, and premature motoneuron death and accelerates motoneuron disease in mice expressing a familial amyotrophic lateral sclerosis mutant SOD1. Neurobiol Dis 7: 623–643, 2000.[CrossRef][ISI][Medline]
  20. Jennings CG, Dyer SM, and Burden SJ. Muscle-specific trk-related receptor with a kringle domain defines a distinct class of receptor tyrosine kinases. Proc Natl Acad Sci USA 90: 2895–2899, 1993.[Abstract/Free Full Text]
  21. Kornfeld S. Structure and function of the mannose 6-phosphate/insulin-like growth factor II receptors. Annu Rev Biochem 61: 307–330, 1992.[CrossRef][ISI][Medline]
  22. Lai KO and Ip NY. Postsynaptic signaling of new players at the neuromuscular junction. J Neurocytol 32: 727–741, 2003.[CrossRef][ISI][Medline]
  23. Langley R, Leung E, Morris C, Berg R, McDonald M, Weaver A, Parry DA, Ni J, Su J, Gentz R, Spurr N, and Krissansen GW. Identification of multiple forms of 180-kDa ribosome receptor in human cells. DNA Cell Biol 17: 449–460, 1998.[ISI][Medline]
  24. Lendahl U, Zimmerman LB, and McKay RD. CNS stem cells express a new class of intermediate filament protein. Cell 60: 585–595, 1990.[CrossRef][ISI][Medline]
  25. McMahan UJ. The agrin hypothesis. Cold Spring Harb Symp Quant Biol 55: 407–418, 1990.[Medline]
  26. Muda M, Theodosiou A, Gillieron C, Smith A, Chabert C, Camps M, Boschert U, Rodrigues N, Davies K, Ashworth A, and Arkinstall S. The mitogen-activated protein kinase phosphatase-3 N-terminal noncatalytic region is responsible for tight substrate binding and enzymatic specificity. J Biol Chem 273: 9323–9329, 1998.[Abstract/Free Full Text]
  27. Muda M, Theodosiou A, Rodrigues N, Boschert U, Camps M, Gillieron C, Davies K, Ashworth A, and Arkinstall S. The dual specificity phosphatases M3/6 and MKP-3 are highly selective for inactivation of distinct mitogen-activated protein kinases. J Biol Chem 271: 27205–27208, 1996.[Abstract/Free Full Text]
  28. Reddy JV and Seaman MN. Vps26p, a component of retromer, directs the interactions of Vps35p in endosome-to-Golgi retrieval. Mol Biol Cell 12: 3242–3256, 2001.[Abstract/Free Full Text]
  29. Reichman CT, Mayer BJ, Keshav S, and Hanafusa H. The product of the cellular crk gene consists primarily of SH2 and SH3 regions. Cell Growth Differ 3: 451–460, 1992.[Abstract]
  30. Rezaee M, Isokawa K, Halligan N, Markwald RR, and Krug EL. Identification of an extracellular 130-kDa protein involved in early cardiac morphogenesis. J Biol Chem 268: 14404–14411, 1993.[Abstract/Free Full Text]
  31. Rossi SG, Vazquez AE, and Rotundo RL. Local control of acetylcholinesterase gene expression in multinucleated skeletal muscle fibers: individual nuclei respond to signals from the overlying plasma membrane. J Neurosci 20: 919–928, 2000.[Abstract/Free Full Text]
  32. Ruegg MA and Bixby JL. Agrin orchestrates synaptic differentiation at the vertebrate neuromuscular junction. Trends Neurosci 21: 22–27, 1998.[CrossRef][ISI][Medline]
  33. Sanes JR and Lichtman JW. Development of the vertebrate neuromuscular junction. Annu Rev Neurosci 22: 389–442, 1999.[CrossRef][ISI][Medline]
  34. Sanes JR and Lichtman JW. Induction, assembly, maturation and maintenance of a postsynaptic apparatus. Nat Rev Neurosci 2: 791–805, 2001.[CrossRef][ISI][Medline]
  35. Schaeffer L, de Kerchove d’Exaerde A, and Changeux JP. Targeting transcription to the neuromuscular synapse. Neuron 31: 15–22, 2001.[CrossRef][ISI][Medline]
  36. Seaman MN. Cargo-selective endosomal sorting for retrieval to the Golgi requires retromer. J Cell Biol 165: 111–122, 2004.[Abstract/Free Full Text]
  37. Seaman MN, Marcusson EG, Cereghino JL, and Emr SD. Endosome to Golgi retrieval of the vacuolar protein sorting receptor, Vps10p, requires the function of the VPS29, VPS30, and VPS35 gene products. J Cell Biol 137: 79–92, 1997.[Abstract/Free Full Text]
  38. Seaman MN, McCaffery JM, and Emr SD. A membrane coat complex essential for endosome-to-Golgi retrograde transport in yeast. J Cell Biol 142: 665–681, 1998.[Abstract/Free Full Text]
  39. Sejersen T and Lendahl U. Transient expression of the intermediate filament nestin during skeletal muscle development. J Cell Sci 106: 1291–1300, 1993.[Abstract/Free Full Text]
  40. Spear E and Ng DT. The unfolded protein response: no longer just a special teams player. Traffic 2: 515–523, 2001.[CrossRef][ISI][Medline]
  41. Tanaka S and Hanafusa H. Guanine-nucleotide exchange protein C3G activates JNK1 by a ras-independent mechanism. JNK1 activation inhibited by kinase negative forms of MLK3 and DLK mixed lineage kinases. J Biol Chem 273: 1281–1284, 1998.[Abstract/Free Full Text]
  42. Teng KK, Lander H, Fajardo JE, Hanafusa H, Hempstead BL, and Birge RB. v-Crk modulation of growth factor-induced PC12 cell differentiation involves the Src homology 2 domain of v-Crk and sustained activation of the Ras/mitogen-activated protein kinase pathway. J Biol Chem 270: 20677–20685, 1995.[Abstract/Free Full Text]
  43. Tsim KW and Barnard EA. The signaling pathways mediated by P2Y nucleotide receptors in the formation and maintenance of the skeletal neuromuscular junction. Neurosignals 11: 58–64, 2002.[CrossRef][ISI][Medline]
  44. Tumor Analysis Best Practices Working Group. Expression profiling—best practices for data generation and interpretation in clinical trials. Nat Rev Genet 5: 229–237, 2004.[CrossRef][ISI][Medline]
  45. Turner BJ, Lopes EC, and Cheema SS. Neuromuscular accumulation of mutant superoxide dismutase 1 aggregates in a transgenic mouse model of familial amyotrophic lateral sclerosis. Neurosci Lett 350: 132–136, 2003.[CrossRef][ISI][Medline]
  46. Vaittinen S, Lukka R, Sahlgren C, Rantanen J, Hurme T, Lendahl U, Eriksson JE, and Kalimo H. Specific and innervation-regulated expression of the intermediate filament protein nestin at neuromuscular and myotendinous junctions in skeletal muscle. Am J Pathol 154: 591–600, 1999.[Abstract/Free Full Text]
  47. Vilinsky I, Stewart BA, Drummond J, Robinson I, and Deitcher DL. A Drosophila SNAP-25 null mutant reveals context-dependent redundancy with SNAP-24 in neurotransmission. Genetics 162: 259–271, 2002.[Abstract/Free Full Text]
  48. Wanker EE, Sun Y, Savitz AJ, and Meyer DI. Functional characterization of the 180-kD ribosome receptor in vivo. J Cell Biol 130: 29–39, 1995.[Abstract]
  49. Watty A, Neubauer G, Dreger M, Zimmer M, Wilm M, and Burden SJ. The in vitro and in vivo phosphotyrosine map of activated MuSK. Proc Natl Acad Sci USA 97: 4585–4590, 2000.[Abstract/Free Full Text]
  50. Weinstein DE, Dobrenis K, and Birge RB. Targeted expression of an oncogenic adaptor protein v-Crk potentiates axonal growth in dorsal root ganglia and motor neurons in vivo. Brain Res Dev Brain Res 116: 29–39, 1999.[ISI][Medline]
  51. Young C, Lindsay S, Vater R, and Slater CR. An improved method for the simultaneous demonstration of mRNA and esterase activity at the human neuromuscular junction. Histochem J 30: 7–11, 1998.[CrossRef][ISI][Medline]
  52. Zhao P and Hoffman EP. Embryonic myogenesis pathways in muscle regeneration. Dev Dyn 229: 380–392, 2004.[CrossRef][ISI][Medline]
  53. Zhao P, Iezzi S, Carver E, Dressman D, Gridley T, Sartorelli V, and Hoffman EP. Slug is a novel downstream target of MyoD. Temporal profiling in muscle regeneration. J Biol Chem 277: 30091–30101, 2002.[Abstract/Free Full Text]
  54. Zhou J and Hoffman EP. Pathophysiology of sodium channelopathies. Studies of sodium channel expression by quantitative multiplex fluorescence polymerase chain reaction. J Biol Chem 269: 18563–18571, 1994.[Abstract/Free Full Text]




This Article
Free upon publication
Abstract
Full Text (PDF)
Supplemental Tables
All Versions of this Article:
21/1/70    most recent
00227.2004v1
Citation Map
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Download to citation manager
Search for citing articles in:
ISI Web of Science (1)
Google Scholar
Articles by Nazarian, J.
Articles by Hoffman, E. P.
Articles citing this Article
PubMed
PubMed Citation
Articles by Nazarian, J.
Articles by Hoffman, E. P.


HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Visit Other APS Journals Online
Copyright © 2005 by the American Physiological Society.