Figures
Abstract
Multiple quantitative trait loci (QTLs) for blood pressure (BP) have been detected in rat models of human polygenic hypertension. Great challenges confronting us include molecular identifications of individual QTLs. We first defined the chromosome region harboring C1QTL1 to a segment of 1.9 megabases that carries 9 genes. Among them, we identified the gene encoding the fibronectin type III domain containing 1 protein (Fndc1)/activator of G protein signaling 8 (Ags8) to be the strongest candidate for C1QTL1, since numerous non-synonymous mutations are found. Moreover, the 5’ Fndc1/Ags8 putative promoter contains numerous mutations that can account for its differential expression in kidneys and the heart, prominent organs in modulating BP, although the Fndc1/Ags8 protein was not detectable in these organs under our experimental conditions. This work has provided the premier evidence that Fndc1/Ags8 is a novel and strongest candidate gene for C1QTL1 without completely excluding other 8 genes in the C1QTL1-residing interval. If proven true by future in vivo function studies such as single-gene Fndc1/Ags8 congenics, transgenesis or targeted-gene modifications, it might represent a part of the BP genetic architecture that operates in the upstream position distant from the end-phase physiology of BP control, since it activates a Gbetagamma component in a signaling pathway. Its functional role could validate the concept that a QTL in itself can influence BP ‘indirectly’ by regulating other genes downstream in a pathway. The elucidation of the mechanisms initiated by Fndc/Ags8 variations will reveal novel insights into the BP modulation via a regulatory hierarchy.
Citation: Deng AY, Chauvet C, Ménard A (2016) Alterations in Fibronectin Type III Domain Containing 1 Protein Gene Are Associated with Hypertension. PLoS ONE 11(4): e0151399. https://doi.org/10.1371/journal.pone.0151399
Editor: Christian Schönbach, Nazarbayev University, KAZAKHSTAN
Received: August 4, 2015; Accepted: February 26, 2016; Published: April 11, 2016
Copyright: © 2016 Deng et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Data are available through the European Nucleotide Archive (LT158646, LT158647, LT158648, and LT158649).
Funding: AYD received funding from Canadian Institutes of Health Research for the conduct of the experiments. CC is supported by a doctoral fellowship from the Heart and Stroke Foundation of Canada. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Localization of quantitative trait loci (QTLs) has disclosed numerous chromosome regions as well as gene markers that are functionally associated with blood pressure (BP) in both humans [1] and animal models [2,3]. The genetic architecture of BP is composed of multiple QTLs as building blocks [4]. Their cumulative impact on the overall BP homeostasis has been revealed by experimental studies, and consists of limited modularized functional ‘blocks’ that fall into epistatic hierarchies [5]. In order to understand molecular mechanisms underlying the functionality of epistatic modules, we embarked on identifying gene variants causal to BP variations.
We previously defined several BP QTLs closely-linked on Chromosome (Chr) 1 in congenic strains cross-bred from hypertensive Dahl salt-sensitive (DSS) and Lewis rats [5]. They belong to one of the 2 epistatic modules from which QTLs take part in molding the overall BP [5]. In the present work, we aimed at identifying a likely causal gene candidate for such a QTL combining congenic resolution [6], a total genome sequencing and molecular analyses. We have identified potentially function-altering genetic variants in fibronectin type III domain containing 1 (Fndc1) that qualifies it to be a candidate, despite mouse knockout strains do not exist beyond embryonic cell lines [7].
Methods
Animals
Protocols for handling, keeping and treating animals in our current study have been approved by COMITÉ INSTITUTIONNEL DE PROTECTION DES ANIMAUX DU CRCHUM (CIPA). The congenic approach simulates the ‘knock- in’ principle where a Chr 1 segment of the hypertensive DSS rat has been replaced by that of normotensive Lewis rat, while maintaining the remaining genome as that of DSS [5]. All the congenic strains used in the current study are depicted in Fig 1 and came from our previous work [5]. Their designations are abbreviated to facilitate cross-referencing and presentations in the text and with our previous work.
A solid bar under congenic strains represents the DSS fragment (a white bar) that has been replaced by that of Lewis (S.L). Striped bars on ends of the solid bars denote the ambiguity of crossover breakpoints between markers. The map is not drawn to scale. Mean arterial pressures (MAPs) for all the strains are averages for the duration of the measurement. Systolic and diastolic arterial pressures are consistent with their MAPs of all the strains (data not shown). Fibronectin type III domain containing 1 (Fndc1)/ activator of G protein signaling 8 (Ags8).
Whole genome sequencing of DSS and Lewis rats
We first extracted genomic DNAs from kidneys of DSS and Lewis rats using a QIAamp DNA Mini Kit following the protocol provided by the manufacturer. 5 micrograms of extract DNAs for each rat strain were then sequenced by the Illumina technology at the Mcgill Genome Center (http://gqinnovationcenter.com/index.aspx). S1 Table outlines the methods, procedure, interpretation, sequence calling and data mining along with relative references. The genome sequences became our data base for identifying single-nucleotide polymorphisms (SNPs) differentiating DSS from Lewis (S2 Table). Our results are consistent with those published by other investigators [8].
Coding regions and intron-exon junctions of all the genes residing in the C1QTL1-containing interval (Fig 2) were curated from our genome data base and in consultation with those in the public domain [8]. When a non-synonymous mutation was found, the DNA segment was independently PCR-amplified and re-sequenced including DSS and Lewis rats by Sanger sequencing. Thus, all missense mutations were verified and are homozygous for the Lewis alleles, (i.e. LL) in the C1S.L4 congenic strain that defines C1QTL1.
The position of a mutation enumerates from the ATG start codon of that gene. The amino acid position begins from the first methionine. Dynlt1, dynein light chain Tctex-type 1; Ezr, ezrin; Fndc1/Ags8, fibronectin type III domain containing 1/ activator of G protein signaling 8 (S3 Table); LOC100912128, parkin coregulated gene protein-like; RGD1560015, similar to glycoprotein, synaptic 2; Rsph3, radial spoke 3 homolog (Chlamydomonas); Sytl3, synaptotagmin-like 3; Tagap, T-cell activation RhoGTPase activating protein (S4 Table); Tmem181, transmembrane protein 181. Pseudogenes are not included. In: intron, Ex: exon, (+) and (-): nucleotide after before a given exon respectively. Raw genome sequence data have been deposited at the European Nucleotide Archive with accesion #s: Accession#s: LT158646, LT158647, LT158648, and LT158649.
All the genes that carry missense mutations were assessed for expressions by reverse transcriptase (RT) polymerase chain reaction (PCR) in a panel of mRNAs extracted from various organs of DSS and congenic (or Lewis) rats. Copy number variations can also be found.
Quantitative real-time RT PCR (qPCR) was utilized to determine and compare expressions of a gene as reported previously [9]. Data was analyzed by applework5.
Results
Overall strategy in identifying candidate genes for a BP QTL
A summary of genome SNPs is presented in S2 Table and illustrates a global distribution of SNPs in known genes between DSS and Lewis genomes. Combining the genome information with a congenic strain defining a BP QTL, C1QTL1, we proceeded to identify all the SNPs located in the coding regions and intron-exon junctions of every gene residing in the chromosome interval harboring the QTL.
Restricting a chromosome interval harboring C1QTL1
Congenic strains are powerful in that, first, they isolate the genes functionally impacting on BP based on a cause-effect relationship from a myriad of genes existing in the entire genome [6]. Second, they are capable of separating multiple closely-linked BP QTLs. Genomic SNPs enabled us to curtail the size of chromosome segments harboring BP QTLs on Chr 1 by diminishing the ambiguous regions on both sides of congenic strains as we reported previously [5]. Among the 3 BP QTLs delineated on Chr. 1 (Fig 1), the region harboring C1QTL1 is narrowed to a segment carrying 9 genes. The number of genes in the other QTL-residing fragments is over 1400 for limiting candidate genes for a detailed molecular study. Because of it, further genetic studies were not pursued for other QTLs.
Fndc1 is a novel candidate gene for C1QTL1
(a), the congenic strain, C1S.L4, is different from DSS in the segment of 1.9 megabases (Fig 1). This segment-‘knock in’ strain lowers mean arterial pressure (MAP) by 43 mmHg (p<0.001) from that of DSS (Fig 1), indicating that a gene(s) entrapped within and among the 9 genes is responsible for the functionality of C1QTL1. This level of narrowing candidate genes is one of a few studies so far reported in the experimental genetics of polygenic hypertension.
All the genes, which exist in the C1QTL1-containing interval and carry SNPs, are homozygous in genotype for Lewis rats (i.e. LL) in the congenic strain C1S.L4, and were curated from our DSS and Lewis genome data bases. The genes carrying SNPs in the coding regions and intron-exon variations are summarized in Fig 2.
(b), two genes carry non-synonymous mutations, i.e. Fndc1 (S3 Table) and T-cell activation RhoGTPase activating protein (Tagap) (S4 Table). In contrast, all the remaining genes located in the C1QTL1-residing interval encode identical proteins.
No DSS/C1S.L4 alternative splicing was detected for the genes carrying intron-exon boundary variants (Fig 2). When a SNP is found near a splice donor or recipient site, the status of alternative splicing, or a lack of it, is assessed by RT-PCRs of the exons flanking it as follows. While taking into considerations that a gene might be differentially spliced in an organ-specific style, PCR primers located in 1 or 2 exon before and after the implicated exon are used to amplify a panel of cDNA segments from the organs where the gene is expressed (e.g. Fig 3). The resultant products were then analyzed by gel electrophoresis for size polymorphisms. So far, among 9 genes in the C1QTL1-residing interval (Fig 2), only Fndc1/Ags8 carries possible splice-site mutations, but they did not influence splicing.
The organs are from Dahl salt-sensitive (DSS) rats. Numbers to the left indicate the size of the fragment in base pairs. The full gene names are given in the legend in Fig 1 and contain missense mutations. Primers for RT-PCRs of genes are forward 5’ GGGAGAACATTCCGACACAT 3’ and reverse 5’ GCTACCATGCTGACACTGGA 3’ for Tagap; forward 5’ TGAAGCTGGCAACCTCATAA 3’ and reverse 5’ TGTGGTCTCAAGTCCATAATCG 3’ for Fndc1/Ags8; and forward 5’ ACTGCCGCATCCTCTTCCTC 3’ and reverse 5’ CCGCTCGTTGCCAATAGTGA 3’ for β-actin. Two primers for each gene are located in 2 different exons to avoid amplifying genomic DNAs contaminated in RNA preparations, since no products were seen when genomic DNAs were amplified (data not shown). Before RT-PCR, all the mRNA samples were treated with RNase-free DNAs to remove possible traces of genomic DNAs. Results shown are from 1 rat of each strain and they have been replicated with multiple rats of the same strain and Lewis rats (data not shown). All rats were males, 11 weeks of age and fed a high salt diet for 6 weeks starting from 5 weeks of age.
Thus, this mutation screening narrowed the list of probable candidates down to the 2 genes, Fndc1 and Tagap. No copy number variations, short insertions and deletions, structural variants were detected among the genes from the total genome comparisons between DSS and Lewis rats (data not shown).
(c), we analyzed gene expressions on these 2 genes. Fig 3 shows that they are expressed in several organs and are functional genes, not pseudogenes.
(d), we reasoned that for a gene to remain a meaningful candidate for C1QTL1, it should carry a missense mutation(s) that would functionally affect the protein product it encodes. We examined the potential functional impact of each missense mutation in the 2 genes (Table 1). Only the Pro1374Leu substitution of Fndc1 caused by the C4121T mutation is predicted to be damaging (Table 1), and thus may have a functional consequence. In contrast, the mutations in Tagap are predicted to be benign or neutral. Thus, Fndc1 is the best supported candidate gene for C1QTL1.
(e), Fndc1 is also known as the activator of G protein signaling 8 (Ags8)/the ischemia-inducible regulator of Gβγ subunit [12,13]. The expression of Ags8/Fndc1 is induced specifically in cardiomyocytes in response to ischemia/hypoxia, but not under other cardiac dysfunctions including prolonged tachycardia, Ang II-infused hypertrophy, and heart failure [12]. By implication, the heart seems an important organ where Ags8/Fndc1 acts. Since the pumping action of the heart is a main mechanical force in driving BP [14], we compared the cardiac Ags8/Fndc1 expressions between DSS and congenic strain C1S.L4 (Fig 1). As shown in Table 2, Ags8/Fndc1 is expressed greatly more in C1S.L4 than in DSS, indicating that it is differentially expressed due to the congenic ‘knock-in’.
(f), we noted from the organ expression pattern that Ags8/Fndc1 appears vastly expressed in kidneys (Fig 2), which is consistent with a quantitative organ assessment by other investigators [12]. Since kidneys play a vital role in long-term BP modulations [15], we compared the renal Ags8/Fndc1 expressions (Fig 1). C1S.L4 showed a higher level of Ags8/Fndc1 expression than DSS (Table 2).
Western blotting using antibodies against Ags8/Fndc1 (Y-12, Santa Cruz) could not detect the protein in the heart and kidneys of DSS and C1S.L4 rats (data not shown), presumably because the protein quantity is below the detectable level in the organs, although a differential expression is seen by a real time RT-PCR after their mRNAs were greatly amplified.
(g), finally, we reasoned that if the Ags8/Fndc1 differential expression is an inherent property of the gene itself, rather than a responder to the action of another gene in trans outside the C1QTL1-residing segment in the genome, there should be a mutation(s) present in its promoter. Indeed, several mutations are found in its putative promoter (Fig 4, S5 Table) and can potentially account for its differential expression capability.
Gene names are given in the legend for Fig 2. Nucleotides were trolled from our sequence data base of DSS and Lewis genomes (S2 Table). RGD1560015 curated in this region shares a high conserved homology with Tecr trans-2,3-enoyl-CoA reductase (Tecr) on Chr 19, thus cannot be differentiated from it. UTR, untranslated region.
In contrast, no mutations are found in the 5’ genomic regions ahead of ATG start codons for the remaining genes (Fig 4). This lack of nucleotide variations lessens the chance that any of these genes can be differentially expressed owing to its direct and innate promoter activity, and thus does not support them as possible candidate genes for C1QTL1. However, a mutation can not be excluded in other regulatory regions in addition to the promoter that may still have a function impact.
In contrast to the C1QTL1-residing segment carrying 9 genes, the chromosome regions containing C1QTL2 and C1QTL3 are too large for identifying a single strong candidate gene that either bears structural mutations or exhibits differential expressions. Thus, no additional studies were performed due to this limitation.
Discussion
The primary finding from the current work is that fibronectin type III domain containing 1 (Fndc1)/activator of G protein signaling 8 (Ags8) is a prohibited candidate gene for C1QTL1. The fibronectin type III domain refers to internal repeats found in 2% of animal proteins including extra-cellular matrices, cell-surface receptors, enzymes and muscle proteins [16]. Ags8/Fndc1 was previously shown to be up-regulated in cardiomyocytes when induced by ischemia/hypoxia [12], and highly expressed in kidneys without induction.
From our current work, Ags8/Fndc1 is supported as C1QTL1 for the following reasons. First, protein-altering mutations, particularly C4121T (Fig 2, Table 1), in Ags8/Fndc1 are associated with a change in blood pressure (Fig 1). Second, Ags8/Fndc1 is differentially expressed greatly in the kidneys and heart (Table 2) supported by numerous promoter mutations (Fig 4, S5 Table).
Gβγ signaling pathways in which Ags8/Fndc1 participate are quite diverse, and some have been linked to hypertension [17]. A Gβ3 splice variant was associated with increased hypertension in humans [18]. Ags8/Fndc1, a proapoptotic factor, has been shown to mediate the Gβγ-dependent phosphorylation of connexin 43 (Cx43) affiliated with cell permeability and a Cx43 internalization [13]. Interestingly, an endothelium-specific Cx43 mouse knockout caused hypotension and increased plasma nitric oxide content and, paradoxically an elevated Ang II level [19]. Thus, regular endothelial Cx43 gap junctions may play a role in maintaining the vascular function and a hypertensive state [20]. A Cx43 knock-in mouse strain by Cx32 remains normotensive in a 2-kidney and 1-clip model associated with a lower level of renin [21], indicating that Cx43 may mediate hypertension via renin in that model.
In contrast, an inhibition of nitric oxide synthase (Nos), a promoter of hypotension, caused a decrease in aortic Cx43 accompanied by hypertension [22] and correlated with a lower level of phosphorylation of Cx43 in the Nos model. The phosphorylation of Cx43 is a prominent mechanism regulating its function and its cellular trafficking [23].
The Ags8/Fndc1 involvement in Cx43 phosphorylation may be through activations of kinases by the Ags8/Fndc1-Gβγ pathway [23]. Our genetic work suggests that alterations and a differential production in a positive regulator, Ags8/Fndc1, upstream in a Gβγ signaling pathway may trigger BP changes. As such, Ags8/Fndc1 as C1QTL1 seems to act ‘indirectly’ on BP far away from the ‘direct’ genes closer to the end-stage BP biology (Fig 5). Interestingly, the gene encoding a new kinase, alpha kinase 2 (Alpk2), was shown to be a BP QTL candidate [9]. Alpk2 may operate in the same epistatic module/pathway as Ags8/Fndc1/C1QTL1 [5]. As such, it is tantalizing to speculate that Alpk2 might phosphorylate Cx43 in connection to the signaling by the Ags8/Fndc1-Gβγ complex.
Arrows only indicate a general direction of a pathway and do not enumerate exact steps involved.
Since Ags8/Fndc1 is a unique gene that may be required for survival, knocking it out [24] may not be a viable approach in proving its effect on BP. Evidently, no viable mice are available despite the existence of cell lines (http://www.informatics.jax.org/allele/MGI:5000326).
In contrast to the substantial support of Ags8/Fndc1 to be C1QTL1, no protein-altering mutations exist in the rat homologues of the human FURIN/FES, ADM, and PLEKHA7 (data not shown). Thus, these genes are likely to be markers for the real C1QTL3 located in their vicinity. Our current work is unique, and yet complements those of other investigators on genetic studies of rat Chr 1, which is rich in BP QTLs [25–27].
Caveats
Limitations of our current genetic studies are. First, the mechanisms by which the coding mutations in Ags8/Fndc1 change its function are unknown. Second, since Ags8/Fndc1 is pleiotropic even playing a role in response to ischemia/hypoxia, the mechanisms by which Ags8/Fndc1 would regulate the phosphorylation of Cx43 via the Gβγ pathway and lead to the BP modulation are yet to be elucidated (Fig 5). Third, the predicted functional consequences, or their lack of, for missense mutations serve as promising guidelines and need to be verified experimentally.
Fourth, since no gene expression studies were performed for remaining 8 genes, their expressions can still be relevant in a functionally-appropriate organ. Finally, epigenetic variations were not assessed and may exist among the 9 genes.
In conclusion, congenic coupled with molecular studies have identified Ags8/Fndc1 to be the most plausible gene candidate for C1QTL1. The combined impact from structural mutations as well as renal and cardiac differential expressions supports a possible functional relevance of Ags8/Fndc1 as C1QTL1. Fndc1/Ags8 appears the strongest, not necessarily the only possible, candidate. In other words, Fndc1/Ags8 definitely can not be excluded, whereas the remaining 8 genes may or may not be excluded. They are currently not supported as C1QTL1 for a lack of genetic evidence except for Fndc1/Ags8.
Ags8/Fndc1 is known to act distantly upstream in a cascade of signaling mechanisms in modifying CX43 and thus implicates the principle that a QTL might take part in a regulatory hierarchy in a pathway that ‘indirectly’, but eventually results in a BP control [5,28]. Since Ags8/Fndc1 may involve in a pathway with other QTLs [5], revealing its role may pave the way for unfolding the other QTLs in the (Gβγ signaling-CX43 phophorylation-BP) pathway. This conceptual awareness may help unravel the functional roles of certain human QTLs whose rat homologues operate in the same pathway, although no human homologue of Ags8/Fndc1/C1QTL1 per se has been detected [1].
Supporting Information
S1 Table. Total genome sequencing and identification of single nucleotide polymorphisms (SNPs): DNA Sequencing Pipeline Genome Quebec and McGill Innovation Center October 2013.
https://doi.org/10.1371/journal.pone.0151399.s001
(PDF)
S2 Table. Selective global evaluation of genome single nucleotide polymorphisms (SNPs) comparing DSS and Lewis rats.
https://doi.org/10.1371/journal.pone.0151399.s002
(PDF)
S3 Table. Coding sequence alignment of fibronectin type III domain containing 1 (Fndc1)/activator of G protein signaling 8 (Ags8) between Dahl salt-sensitive (DSS) and Lewis rats.
https://doi.org/10.1371/journal.pone.0151399.s003
(PDF)
S4 Table. Coding sequence alignment of T-cell activation RhoGTPase activating protein (Tagap) between Dahl salt-sensitive (DSS) and Lewis rats.
https://doi.org/10.1371/journal.pone.0151399.s004
(PDF)
S5 Table. Sequence alignment of putative Fndc1/Ags8 promoter between the mouse and Rat.
https://doi.org/10.1371/journal.pone.0151399.s005
(PDF)
Author Contributions
Conceived and designed the experiments: AYD. Performed the experiments: CC AM. Analyzed the data: AYD AM. Contributed reagents/materials/analysis tools: CC AM. Wrote the paper: AYD.
References
- 1. The International Consortium for blood pressure. Genetic variants in novel pathways influence blood pressure and cardiovascular disease risk. Nature 2011; 478: 103–109. pmid:21909115
- 2. Cowley AW Jr. The genetic dissection of essential hypertension. Nat Rev Genet 2006; 7: 829–840. pmid:17033627
- 3. Deng AY. Genetic basis of polygenic hypertension. Hum Mol Genet 2007; 16: R195–R202. pmid:17911162
- 4. Deng AY. Genetic mechanisms of polygenic hypertension: fundamental insights from experimental models. Journal of Hypertension 2015; 33: 669–680.
- 5. Chauvet C, Crespo K, Menard A, Roy J, Deng AY. Modularization and epistatic hierarchy determine homeostatic actions of multiple blood pressure quantitative trait loci. Hum Mol Genet 2013; 22: 4451–4459. pmid:23814039
- 6. Deng AY. Positional Cloning of Quantitative Trait Loci for Blood Pressure: How Close Are We?: A Critical Perspective. Hypertension 2007; 49: 740–747. pmid:17296871
- 7. White J, Gerdin AK, Karp N, Ryder E, Buljan M, Bussell J, et al. Genome-wide Generation and Systematic Phenotyping of Knockout Mice Reveals New Roles for Many Genes. Cell 2013; 154: 452–464. pmid:23870131
- 8. Atanur S, Diaz A, Maratou K, Sarkis A, Rotival M, Game L, et al. Genome Sequencing Reveals Loci under Artificial Selection that Underlie Disease Phenotypes in the Laboratory Rat. Cell 2013; 154: 691–703. pmid:23890820
- 9. Chauvet C, Crespo K, Menard A, Wu Y, Xiao C, Blain M, et al. alpha-Kinase 2 is a novel candidate gene for inherited hypertension in Dahl rats. J Hypertens 2011; 29: 1320–1326.
- 10. Adzhubei IA, Schmidt S, Peshkin L, Ramensky VE, Gerasimova A, Bork P, et al. A method and server for predicting damaging missense mutations. Nat Methods 2010; 7: 248–249. pmid:20354512
- 11. Kumar P, Henikoff S, Ng PC. Predicting the effects of coding non-synonymous variants on protein function using the SIFT algorithm. Nat Protoc 2009; 4: 1073–1081.
- 12. Sato M, Cismowski MJ, Toyota E, Smrcka AV, Lucchesi PA, Chilian WM, Lanier SM. Identification of a receptor-independent activator of G protein signaling (AGS8) in ischemic heart and its interaction with Gbetagamma. Proc Natl Acad Sci U S A 2006; 103: 797–802.
- 13. Sato M, Jiao Q, Honda T, Kurotani R, Toyota E, Okumura S, et al. Activator of G protein signaling 8 (AGS8) is required for hypoxia-induced apoptosis of cardiomyocytes: role of G betagamma and connexin 43 (CX43). J Biol Chem 2009; 284: 31431–31440. pmid:19723622
- 14.
Caro CG. The Mechanics of The Circulation. 1978. Oxford University Press.
- 15. Cowley AW Jr. Renal medullary oxidative stress, pressure-natriuresis, and hypertension. Hypertension 2008; 52: 777–786.
- 16. Bork P, Doolittle RF. Proposed acquisition of an animal protein domain by bacteria. Proc Natl Acad Sci U S A 1992; 89: 8990–8994. pmid:1409594
- 17. Jones MB, Siderovski DP, Hooks SB. The G betagamma dimer as a novel source of selectivity in G-protein signaling: GGL-ing at convention. Mol Interv 2004; 4: 200–214. pmid:15304556
- 18. Siffert W, Rosskopf D, Siffert G, Busch S, Moritz A, Erbel R, et al. Association of a human G-protein beta3 subunit variant with hypertension. Nat Genet 1998; 18: 45–48. pmid:9425898
- 19. Liao Y, Day KH, Damon DN, Duling BR. Endothelial cell-specific knockout of connexin 43 causes hypotension and bradycardia in mice. Proc Natl Acad Sci U S A 2001; 98: 9989–9994. pmid:11481448
- 20. SAEZ JC, BERTHOUD VM, BRANES MC, MARTINEZ AD, BEYER EC. Plasma Membrane Channels Formed by Connexins: Their Regulation and Functions. Physiol Rev 2003; 83: 1359–1400. pmid:14506308
- 21. Haefliger JA, Krattinger N, Martin D, Pedrazzini T, Capponi A, Doring B, et al. Connexin43-dependent mechanism modulates renin secretion and hypertension. J Clin Invest 2006; 116: 405–413. pmid:16440062
- 22. Haefliger JA, Meda P, Formenton A, Wiesel P, Zanchi A, Brunner HR, et al. Aortic connexin43 is decreased during hypertension induced by inhibition of nitric oxide synthase. Arterioscler Thromb Vasc Biol 1999; 19: 1615–1622. pmid:10397678
- 23. Solan JL, Lampe PD. Connexin43 phosphorylation: structural changes and biological effects. Biochem J 2009; 419: 261–272. pmid:19309313
- 24. Geurts AM, Cost GJ, Freyvert Y, Zeitler B, Miller JC, Choi VM, et al. Knockout rats via embryo microinjection of zinc-finger nucleases. Science 2009; 325: 433. pmid:19628861
- 25. Clemitson JR, Dixon RJ, Haines S, Bingham AJ, Patel BR, Hall L, et al. Genetic dissection of a blood pressure quantitative trait locus on rat chromosome 1 and gene expression analysis identifies SPON1 as a novel candidate hypertension gene. Circ Res 2007; 100: 992–999. pmid:17332427
- 26. Joe B, Garrett MR, Dene H, Rapp JP. Substitution mapping of a blood pressure quantitative trait locus to a 2.73 Mb region on rat chromosome 1. J Hypertens 2003; 21: 2077–2084. pmid:14597851
- 27. Yagil C, Hubner N, Monti J, Schulz H, Sapojnikov M, Luft FC, et al. Identification of hypertension-related genes through an integrated genomic-transcriptomic approach. Circ Res 2005; 96: 617–625. pmid:15731461
- 28. Deng AY. Response to the Editorial Do epistatic modules exist in the genetic control of blood pressure in Dahl rats? A critical perspective. Physiol Genomics 2013; 45: 1196–1198.