1 Division of Cardiovascular Medicine, University of California, Davis, California 95616-8636
2 Division of Endocrinology, Clinical Nutrition, and Vascular Medicine, University of California, Davis, California 95616-8636
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
vascular; differential display; estrogen; hormones
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Although the mechanisms by which ovarian steroids affect vascular function are not clear, ovarian hormones are becoming increasingly recognized as complex and powerful cardiovascular mediators (4, 9, 35, 11). Possible mechanisms by which ovarian hormones, principally estrogen, act on vascular tissue include enhanced vasodilatation via increased production of nitric oxide, a potent vasodilator (7, 36), changes in plasma concentrations of lipids and reductions in vascular cholesterol accumulation (7, 12), and prevention of apoptotic cell death (2, 30).
In addition, as recently reported (13), changes in vascular gene expression also occur in response to estrogen. Vasoactive genes known to be estrogen-regulated include inflammatory mediators such as interleukins IL-4 and IL-6 (3, 8) and adhesion molecules VCAM-1 and ICAM-1 (5, 2628). In addition, we have demonstrated that the substance P receptor is upregulated by estrogen (32). More recently, using differential display mRNA gene expression profiling of human aortic endothelial cells, our laboratory has reported previously uncharacterized estradiol-sensitive dose- and time-dependent upregulation of three additional genes with potential importance to vascular function: aldose reductase, a caspase homolog, and plasminogen activator inhibitor-1 (33).
In this report, we extend our recent observations in cultured aortic endothelial cells to examine the effect of loss of ovarian hormones on the genetic profile of the entire vascular wall. Data from a variety of experimental systems strongly indicate that the loss of ovarian estrogens, as occurs in postmenopausal women and following ovariectomy, is a potent signal for regulation of gene expression (4, 9, 12, 35). It is therefore highly probable that ovariectomy is an essential trigger for significant genetic reprogramming of the vessel wall. However, this remains a relatively uninvestigated area. The current studies were undertaken to establish an initial characterization of changes in gene expression patterns in vessels in response to loss of ovarian sex hormones and to identify specific vascular genetic targets for ovarian hormones.
Based on our previous work, we reasoned that the loss of ovarian hormones would be coupled to a significant genetic reprogramming in the vascular wall of aortas. Experiments were therefore designed using the technique of mRNA gene display (14) to identify differentially expressed genes in response to ovariectomy in normal mouse aortas. This strategy enabled us to identify genes that were selectively induced and/or downregulated by ovariectomy and to establish relative changes in gene expression levels compared with intact mice. This knowledge is important to advance our current understanding of the profile of genes under hormonal regulation in the vasculature.
![]() |
MATERIALS AND METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
For reamplification, cloning and sequencing, materials, and reagents were obtained from the following sources: AmpliTaq DNA polymerase (Perkin-Elmer, Foster City, CA); T7 promoter and M13 reverse primers (Midland Certified Reagent, Midland, TX); Invitrogen TOPO TA cloning kit, containing 10x ligation buffer, T4 DNA ligase, SOC medium, INVF' E. coli competent cells, and pCR2.1-TOPO vector (Invitrogen, Carlsbad, CA); Luria-Bertani (LB) medium and agar (Media Room, UC Davis); ampicillin (Apothecon, Princeton, NJ); 5-bromo-4-chloro-3-indolyl-ß-D-galactopyranoside (X-Gal; Boehringer Mannheim, Indianapolis, IN); KH2PO4 and K2HPO4 (Fisher Scientific, Fairlawn, NJ); and Terrific Broth containing tryptone, yeast extract, and glycerol (Difco, Detroit, MI).
Strategy, overview, and advantages of the experimental approach used to identify and verify differential expression.
To identify genes whose expression was differentially regulated by ovariectomy in mouse aortas, we bilaterally ovariectomized normal postpubescent C57BL/6J mice, removed their aortas, and compared gene expression to aortas from C57BL/6J mice with intact ovaries. Total RNA was extracted from the aortas. Three 1-base anchored oligo-dT primers with 5' HindIII sites, used in combination with a series of 16 additional arbitrary 13-mer oligonucleotide upstream primers (also containing 5' HindIII sites), were used to reverse transcribe and amplify the mRNAs from the control and ovariectomized aorta samples. Following amplification, the mRNA subpopulations were resolved on DNA sequencing gels. Hundreds of bands representing expressed gene fragments were screened for differential expression in aortas from intact and ovariectomized mice. DNA from excised bands was eluted, directly sequenced, and queried against National Center for Biotechnology Information (NCBI) databases using the Basic Local Alignment Search Tool (BLAST) algorithm. A match of interest was defined as having >70% homology to a known gene. Following database verification, genes with potential physiological relevance to vascular function were selected for further verification. To confirm differential expression of apparently differentially displayed cDNAs, gene-specific primers were designed and used to amplify the reverse transcription products, which were subsequently analyzed by both semi-quantitative and real-time RT-PCR.
The technique of delta RT-PCR used in these studies is based on reverse transcription and amplification of 3'-terminal RNA sequences (3, 14). The method relies on a combination of thermophilic DNA polymerases with 3'-5' exonuclease (proofreading) activity to produce larger, easier to detect, higher fidelity, and not overcycled PCR products than is possible with conventional PCR (20). These studies were also designed to minimize false-positive results and optimize sensitivity. This technology ensures that differentially expressed genes are more readily identified and cloned than with prior generations of differential display. Specifically, the use of three 1-base anchored oligo-dT primers to subdivide the mRNA population reduces the redundancy and under-representation of the subpopulation of mRNAs (14). In addition, with built-in restriction sites at the 5' ends of both anchored and arbitrary 13-mers, the longer primer pairs produce highly selective and reproducible cDNA patterns (34).
Ovariectomy and isolation of mouse aorta.
Eight female C57BL/6J mice were obtained from Charles River (San Diego, CA). Four mice were bilaterally ovariectomized at 67 wk of age and the other four mice left intact. Both sets of mice were kept in as near identical housing conditions as possible controlling for the same environment, food, light, and temperature conditions. The ovariectomized mice were allowed to recover for a week, and both groups were killed 56 wk later using pentobarbital anesthesia (50 mg/kg). Under RNase-free conditions, the aortas were immediately removed by micro-excision of the entire aorta from the heart to the iliac bifurcation. The aortas were then trimmed of adherent fat, rinsed clean of blood in diethyl pyrocarbonate (DEPC)-treated PBS, and stored at -80°C for subsequent extraction of total RNA.
RNA isolation.
The aortas of ovariectomized and intact mice were homogenized in a tissue homogenizer, and total RNA was isolated using TRI-Reagent (Molecular Research, Cincinnati, OH), per the manufacturers instructions. Each aorta yielded 1416 µg of total RNA, the predicted yield per the manufacturers estimates. Total RNA was then treated with DNase I for removal of contaminating genomic DNA. All lab ware and solutions used to isolate RNA were rendered RNase-free prior to use. RNA samples were stored in DEPC-treated water at -80°C until ready for use.
Serum estradiol assay.
To confirm that following ovariectomy levels of circulating estradiol were low as expected, serum estradiol levels were determined in ovariectomized mice and compared with those of intact control mice using an estradiol enzyme immunoassay (EIA) kit (Cayman Chemical, Ann Arbor, MI) following the manufacturers instructions. Using this method, we found plasma estradiol levels were 168 pg/ml and 23 pg/ml, in intact and ovariectomized mice, respectively.
Differential mRNA display analysis of mouse aortas.
Differential display was performed using an RNAimage mRNA differential display fingerprinting system from GenHunter. Eight samples of total RNA (0.2 µg each), prepared individually from each of the aortas of the four intact and the four ovariectomized mice, were used as template for reverse transcription to cDNA prior to performing differential display. RNA samples were not pooled for any of the analyses. Fibroblast RNA provided with the GenHunter kit, and H2O samples minus the reverse transcriptase, were used for positive and negative RNA controls, respectively. To subdivide the mRNA population for each RNA sample into three subpopulations, total RNA was first converted into first-strand cDNA with MMLV reverse transcriptase, and each of three different 1-base anchored H-T11M oligo-dT primers [where H = the 5' end restriction site (HindIII), T11= an 11-bp long poly-T tail, and M = G, C, or A].
For fingerprinting, gels were hybridized with 2.0 µCi -33P-labeled cDNA (3.5 ng/sample) prepared by reverse transcription of total tissue RNA. PCR amplification was performed with DNA polymerase and the same H-T11M oligo-dT anchor primers and "H-AP" arbitrary primers used for the reverse transcription reactions; see Table 1 for primer sequences. A total of 32 primer pairs were used to amplify the 8 aortic tissue samples. Each primer pair would be expected to yield a different fingerprint and identify a different set of differentially expressed RNAs, typically 12 differential bands per pair. Based on previously published axioms (15), the 16 arbitrary primers used, combined with each of the three 1-base anchored primers, would be expected to detected between 56% and 64% of the total aortic mRNAs.
|
Identification, recovery, and reamplification of apparently differentially displayed cDNAs.
Bands demonstrating apparent differential expression by analysis of gel radiographs were identified by two independent observers. Differential display was suggested by the presence of bands in both samples from control aortas and absence in both samples from ovariectomized aortas, or vice versa. An example is provided in Fig. 1. Thus only transcripts that reproducibly differentially displayed were selected for further analysis. Bands that appeared to be differentially expressed were excised from the gels, then cDNA was eluted and reamplified by PCR using the fingerprinting H-T11M and H-AP primers. Bands were directly sequenced and cloned, and the clones were resequenced.
|
Sequence data was matched against human and mouse genome databases, including GenBank (http://www.ncbi.nlm.nih.gov/Education/index.html) and the Jackson Laboratory site (http://www.jax.org), for nucleotide sequences, and against protein and expressed sequence tag (EST) databases using FASTA (FastA) and NetBLAST based on heuristic algorithms (1, 22). In general, bands that produced direct sequence information and had high homology (>70%) to known human mRNAs were chosen and pursued for subsequent verification of differential expression.
Semi-quantitative RT-PCR-based verification analysis of differential gene expression.
For the initial verification analysis we used semi-quantitative RT-PCR, as total RNA from mouse aorta (14.5 µg total RNA/aorta) was insufficient for poly(A)+ selection and detection by conventional Northern analysis. Verification was performed with gene-specific primers (29, 33), as low levels of contaminating cDNAs in probes prepared directly from differentially displayed bands can obscure accurate verification of bands (18). The online program Primer3 (24) http://www.genome.wi.mit.edu/cgi-bin/primer/primer3.cgi was used to design specific primers for each of the genes of interests and D-glyceraldehyde-3-phosphate dehydrogenase (GAPDH), a "housekeeping" gene. Primer pairs yielded PCR products with the exact length predicted by the known sequences of the five genes (3b, 5a, 18f, 22a, and 34c) selected for verification analysis (Table 2).
|
A titration analysis curve was then constructed for the five genes of interest and GAPDH by making a dilution series of cDNA template (1:10, 1:20, 1:40, and 1:80, corresponding to approximately 3.5, 1.8, 0.9, and 0.4 ng of cDNA, respectively). cDNAs were then PCR amplified, and the band intensity signals quantified by densitometry to verify that relative differences in gene expression remained constant over a wide template concentration range. The optimal conditions for amplification (the relative proportion of upstream to downstream primers, temperatures, and cycle number) were experimentally determined. To reduce spurious priming during gene amplification, PCR was performed for 10 "touchdown" cycles (6) as follows: The first "touchdown" cycle was for 30 s at 94°C, 30 s at 61°C for annealing, and 2 min at 72°C. The next "touchdown" cycle was for 30 s at 94°C, 30 s at 60.9°C for annealing, and 2 min at 72°C. For the remaining 8 cycles the annealing temperature dropped by 0.1°C after each cycle. These touchdown cycles were followed by 1015 cycles of 30 s at 94°C, 30 s at 60°C, and 2 min at 72°C. Samples were held for a final 7-min extension cycle at 72°C, then cooled to 4°C. Correct PCR product was verified by electrophoresis through a 1% agarose gel stained with 1:5,000 SYBR Gold nucleic acid prestain.
DNA bands were visualized with a STORM Fluorimager (Molecular Dynamics, Sunnyvale, CA) and quantified by spot densitometry using STORM Fluorimager ImageQuant software. The mean densitometry signal (calculated as an average of the signal at each template dilution) was expressed as a ratio (intact: ovariectomized) for each of the five genes of interest. The 1:80 template dilution data was not used for this final analysis given difficulty in obtaining accurate readings at this very low dilution. Because there was equalization of target cDNA by GAPDH, it was not necessary to further normalize signals to GAPDH prior to directly comparing intact and ovariectomized sample signals.
Real-time RT-PCR-based confirmation analysis of differential gene expression.
Verification of differential gene expression by semi-quantitative PCR with gene-specific primers is an established approach for analysis of differential expression (18). However, the semi-quantitative RT-PCR verification experiments used the same template RNA as was used for differential display analysis. Therefore, to further confirm differential mRNA expression, a more quantitative approach, real-time PCR, was employed. For these studies, two additional mouse aortas (one ovariectomized and one intact) were analyzed. These aortas were not used in the previous differential expression and semi-quantitative PCR verification experiments. However, to mitigate against the possible influence of environmental factors on gene expression, care was taken to ensure that the additional aortas were from the same experimental animal cohort as that described earlier in MATERIALS AND METHODS.
Briefly, total RNA from the ovariectomized and intact aortas was treated with DNase (MessageClean DNase I; GenHunter) and cDNA generated by RT-PCR. A mixture of oligo-dT oligodeoxynucleotide primers (T1218) and random hexamer primers were used for these studies following the protocol for SuperScript RNase H- reverse transcriptase (GIBCO-BRL, Rockville, MD). Real-time detection of PCR was performed using the GeneAmp 5700 sequence detection system (Applied Biosystems, Foster City, CA) according to the manufacturers instructions. The same gene-specific primers used for semi-quantitative PCR analysis were used for real-time PCR detection of the five genes of interest (3b, 5a, 18f, 22a, and 34c). Equal amounts of sample cDNA, used in duplicate or triplicate, were amplified with the SYBR Green I Master Mix (Applied Biosystems). The thermal cycling parameters were: thermal activation for 10 min at 95°C followed by 40 cycles of PCR (melting for 15 s at 95°C and annealing extension for 1 min at 60°C). A standard curve was constructed using a template dilution series (1:10, 1:20, 1:40, 1:80, and 1:160) of total RNA from mouse aorta. A "no template control" was included with each PCR. Amplification efficiency for each of the five genes of interest was then validated and normalized against GAPDH. To compare the relative expression of the various genes, data was expressed as the ratio of intact:ovariectomized aortas.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Following database verification in GenBank, 20 of the apparently differentially displayed bands were found to have significant (greater than or equal to 70%) homology to known genes. These were considered for further analysis (Table 3). The remaining 12 genes had lesser degrees of homology to known genes and were therefore not considered for further study. As anticipated from GenBank searches, none of the apparently differentially expressed clones represented novel genes. However, several had short and poor matches to known genes or matches to ESTs of unknown function, potentially representing novel genes.
|
Verification of the effect of ovariectomy on aortic differential gene expression by semi-quantitative RT-PCR analysis.
Semi-quantitative RT-PCR identified significant (>2-fold greater than control) differential expression of mRNAs for five of the eight genes of interest. Coincidentally, all of the five genes were downregulated by ovariectomy (Table 3). They had cDNA sequences encoding for: Tat binding protein 1 (TBP-1; clone 3b), a heparin-binding angiogenic growth factor; skeletal muscle -tropomyosin (clone 5a), an actin-binding protein central to the control of calcium-regulated striated muscle contraction; elongation factor 1
(EF-1
; clone 18f), a polypeptide involved in vascular endothelial cell sprouting and elongation; and ganglioside-induced differentiation associated protein (clone 22a), a gene involved in signal transduction to induce cellular differentiation. Moreover, the gene for mitochondrial nicotinamide adenine dinucleotide (NADH)-ubiquinone oxidoreductase chain 6 (clone 34c), an important mitochondrial electron transport enzyme, was also downregulated by ovariectomy. One of the downregulated genes of interest (clone 18d) and two of the upregulated genes of interest (clones 21da and 32ab) demonstrated less than twofold greater expression than GAPDH and therefore were not significantly differential following ovariectomy.
To assess relative changes in differential gene expression among the five genes verified to be differentially expressed by semi-quantitative RT-PCR, direct comparisons were made between densitometry signals for intact and ovariectomized aortas. This was possible, as the cDNA PCR template between aortic samples of intact and ovariectomized mice, at four serial template dilutions, had been equalized between GAPDH and each of the five genes of interest; see Fig. 2A for an example. The semi-quantitative RT-PCR analysis of the mRNA densitometry ratios (intact:ovariectomized aorta) for GAPDH and the five genes of interest is shown in Fig. 2, B and C. As can be noted, the five differentially expressed genes demonstrated significantly greater mRNA expression in intact compared with ovariectomized mouse aortas, consistent with downregulation of gene expression by ovariectomy. The relative differences in differential expression levels between intact and ovariectomized aortas were as follows: 3.3-fold for TBP-1 (clone 3b), 3.2-fold for skeletal muscle -tropomyosin (clone 5a), 3.5-fold for EF-1
(clone 18f), 8.2-fold for ganglioside-induced differentiation associated protein (clone 22a), and 3.8-fold for NADH:ubiquinone oxidoreductase (clone 34c).
|
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The EF-1 family of polypeptides, one of the differentially displayed genes identified to be downregulated in our study, promote polypeptide chain elongation during mRNA translation. EF-1
is a ubiquitously expressed important component of the protein synthesis apparatus in eukaryotic tissues (10). EF-1
is also involved in vascular endothelial cell sprouting and elongation in vitro, as well as the formation of networks of elongated endothelial cells (17). Therefore, downregulation of EF-1
by ovariectomy could be important in retarding angiogenesis.
The second gene identified to be downregulated by ovariectomy in this study was the gene for ganglioside-induced differentiation associated protein. This gene was originally identified in a search of genes involved in ganglioside-induced differentiation of a mouse neuroblastoma cell line, where it causes cell differentiation with neurite sprouting (16). The gene is expressed in most tissues and might be involved in signal transduction to induce cellular differentiation. However, it is unknown whether it could potentially have a role in the cellular differentiation that accompanies endothelial cell sprouting in angiogenesis.
The third gene downregulated by ovariectomy was mitochondrial NADH:ubiquinone oxidoreductase chain 6, an important mitochondrial electron transport enzyme. This gene regulates the redox state of the carriers of the electron-transport chain of mitochondria that correlate with changes in a number of physiological parameters. It is proposed that NADH:ubiquinone oxidoreductase inhibitors block multiple and possibly reactive oxygen species-modulated pathways (23). In addition, cytokines, such as tumor necrosis factor- (TNF
) and IL-1
, inhibit both pyruvate dehydrogenase activity and mitochondrial function in tumor and vascular smooth muscle cells (37). If the same occurs with ovariectomy, then ovariectomy may have the potential for primary impairment of mitochondrial function and the cellular redox state.
In nonvascular animal systems ovariectomy has been previously associated with changes in the expression of genes with primary relevance to hormonal actions. For example, in rats ovariectomy is associated with a progressive increase in mRNA levels for -2,3-sialyltransferas (
-2,3-STase), one of the enzymes that incorporates sialic acid residues into follicle stimulating hormone (FSH) and an important regulator of FSH isoforms and estrogen output (4). However, ovariectomy has also been demonstrated to be associated with changes in expression of genes pertinent to vascular physiology. For example, mRNA expression for steroid hormone receptors was demonstrated to decrease after ovariectomy in aortic and venous rat vessels (11). Furthermore, in ovariectomized female pigs, expression of mRNA for both the endothelial isoform of nitric oxide synthase (eNOS) and endothelin-1, two important vasoactive peptides, is increased compared with their expression in intact female pigs (31). Interestingly, in humans, expression of the P450c5 gene, the key enzyme involved in the conversion of progestins to androgens, demonstrates significant variation in levels of expression in postmenopausal ovaries compared with normal cycling ovaries, without significant differences in the overall gene expression levels between the two states (9). Therefore, the assumption that loss of ovarian hormones inevitably results in changes in gene expression in all systems needs to be challenged and specifically tested.
The most plausible mechanism by which ovariectomy could regulate gene expression in the blood vessel wall is by ovarian steroid-stimulated regulation of mRNA transcription. However, in other systems, changes in gene expression may be a consequence of mRNA breakdown (19). Therefore, it is possible that ovariectomy could affect both the transcriptional rate of the genes identified and/or the stability of their mRNAs. Although our study was not designed to tease out these processes, the data presented provide significant new information about vascular targets for hormone action and lay the foundation for additional future mechanistic work in this area.
In this study, only genes whose sequences are found in GenBank were investigated, yielding information on regulation of known genes by loss of ovarian steroids. It would be interesting as a next step to embark on the discovery of unknown genes regulated by ovariectomy. Novel genes may be represented by EST fragments that are in GenBank, yet have no known function or characterization. In addition, poor matches to GenBank nucleotide sequences also could be novel genes. In fact, some of the ovariectomy-regulated sequences found by this study corresponded to ESTs in the human genome database with unknown function or poor matches to known genes; perhaps some of these ESTs are to novel genes. Although those ESTs were not further characterized, cloning them would be of future interest.
The bipartite differential display technique utilized for these studies, consisting of 1) RT-PCR and electrophoresis followed by 2) gene identification and verification, takes advantage of new innovations designed to improve reliability and reduce false positives in differential display. For example, the methodology improves on previous approaches to differential display by subdividing the mRNA population in three, built in restriction sites at the 5' ends of both anchored and arbitrary primers, longer PCR primers, loading of samples in duplicate, and the use of two approaches to verify and confirm differentially expressed cDNAs. However, our study has certain limitations. The techniques, although reliable in identifying differentially expressed gene products, do not provide a functional assessment of the differentially expressed genes identified. It is also possible that some of the lower sequence homology genes (<80%) may represent mismatches. Furthermore, ovariectomy could have varying effects on vascular gene expression at differing times following ovariectomy and at different phases of the reproductive cycle. This would not be revealed by the present studies as we studied mice at one point in time. An additional limitation is that loss of ovarian steroids, hormones, and peptides following ovariectomy does not permit inferences to be made about the role of specific ovarian hormones (e.g., estrogen) or ovarian vasoactive peptides (e.g., vascular endothelial growth factor and relaxin) on regulation of gene expression. Rather, our studies are an important first step toward this analysis. Lastly, our studies do not localize gene expression changes in the vessel wall. Given that the aorta is a complex tissue consisting of several cell types (endothelial cells, smooth muscle cells, and adventitia), gene expression could clearly vary by cell type relative to cell function. Nonetheless, our studies provide novel data in this evolving area of science, lay new groundwork for further research in this area, and add to our understanding of the molecular basis for female sex steroid hormone action in the vasculature.
In summary, the results of our studies serve to extend the findings of prior studies in three ways. First, we have demonstrated that ovariectomy is associated with significant changes in vascular gene expression in the vessel wall. Second, we have identified three vascular genes previously unreported to be regulated by ovarian hormones that may act as genetic targets for hormone action. And third, in our analysis we have broadened the spectrum of gene functions potentially regulated by ovarian steroids in the vasculature.
In conclusion, the genes differentially regulated by ovariectomy identified in the present study have not previously been identified to be under the control of ovarian hormones. These differentially regulated genes encode for a wide array of cellular regulatory functions. Based on the function of the differentially expressed genes identified, loss of ovarian hormones, principally but not exclusively estrogens, may have relevance to a wide spectrum of regulatory factors in the vasculature related to mitochondrial energy metabolism, cellular differentiation, and the cellular growth and vascular sprouting that characterizes new vessel formation. Thus, understanding the genetic vascular changes accompanying loss of ovarian steroids may enhance our understanding of hormonal regulation of vascular function in health and disease.
![]() |
ACKNOWLEDGMENTS |
---|
![]() |
FOOTNOTES |
---|
Address for reprint requests and other correspondence: A. C. Villablanca, Division of Cardiovascular Medicine, Univ. of California, Davis, One Shields Ave., TB 172, Davis, CA 95616-8636 (E-mail: avillablanca{at}ucdavis.edu).
10.1152/physiolgenomics.00040.2002.
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Visit Other APS Journals Online |