INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE ROLE of endothelium-derived nitric oxide (NO) in the regulation of vascular tone and organ blood flow is well established. Although three isoforms of the human nitric oxide synthase (NOS) gene family are known to exist (11, 34, 73), it is the endothelial nitric oxide synthase (eNOS) gene that accounts for the synthesis and release of bioactive endothelium-derived relaxing factor (EDRF) (34, 42, 45, 72). eNOS is a peripheral membrane protein that, after N-myristoylation and palmitoylation (5), is localized to plasmalemmal caveolae of vascular endothelium (62). Increases in cytosolic free calcium concentration facilitate NADPH-dependent electron flux through eNOS dimers, resulting in the five-electron oxidation of L-arginine to L-citrulline and synthesis of NO (33, 37).

Earlier studies from this laboratory and others have described the isolation and characterization of complementary and genomic clones for eNOS. The eNOS gene is present as a single copy in the haploid human genome and is localized to 7q35-36 (38, 39, 47, 54) and chromosome 5 in the mouse (17). The nucleotide and amino acid sequence of the murine eNOS cDNA corresponds to an open reading frame (ORF) of 1,202 amino acids, and sequence comparison has revealed 94% and 93% amino acid identity with human and bovine sequences, respectively (16). The human gene contains 26 exons and encodes a 4,052-nucleotide (nt) mRNA (38, 39). Although a canonical TATAA motif is not present, a single major transcription initiation site was identified 22 nt upstream of the translational start site using primer extension, S1 nuclease protection, and 5'-rapid amplification of cDNA ends (38). Two tightly clustered regulatory domains have been identified in the proximal promoter of the human eNOS gene. In endothelial cells, nucleoprotein complexes containing Ets family members, Sp1, variants of Sp3, MAZ, and YY1 form on these regions (30, 81). The murine eNOS promoter has recently been cloned, and cross-species comparisons have revealed a high degree of homology (70), suggesting that mechanisms active in the regulation of eNOS expression may also be conserved. The markedly restricted expression of eNOS contrasts with the broad distribution of the neuronal (nNOS) and inducible (iNOS) isoforms (11, 34, 44, 72, 73). However, as is the case with nearly all cell-restricted transcripts, some exceptions regarding eNOS transcription have been noted. Constitutive expression of the mRNA and protein for eNOS have been documented in the syncytiotrophoblast of human placental villi (8), the neuronal cells of the CA1 region (12, 29), human platelets (6, 41, 56), and, arguably, cardiac myocytes (3, 4, 25, 60, 74, 76).

The preponderance of what is known regarding expression of the eNOS gene has been derived from in vitro studies. Some murine models have been developed that have augmented our understanding of eNOS gene regulation in the in vivo setting. Murine models of targeted homologous replacement and germline inactivation of the various NOS isoforms have been described (26-28, 36, 63, 66, 69). Homozygous (/) null mutants of eNOS have exhibited elevated systemic vascular resistance and displayed abnormalities of vascular remodeling and coagulation (27, 55, 63). Mice also have exhibited pulmonary hypertension, perhaps resulting from an increase in total pulmonary resistance (69). nNOS(/) mice are also viable, are capable of reproducing, and evidence neuroprotection in models of ischemia-reperfusion of the central nervous system. Hypertrophy of the gastric pyloric sphincter is a prominent anatomic feature (26). Null mutants for the iNOS isoform are less able to combat systemic bacterial, viral, or fungal infections. iNOS(/) mice are protected, in part, from lipopolysaccharide-induced hemodynamic collapse (36). It has been suggested that single NOS knockouts may be partially compensated by changes in functional activity of other isoforms. Interbreeding among the varied NOS-deficient mice will provide interesting information regarding the biological roles of NO, especially during development. In contrast to loss-of-function models, recent studies have sought to define the consequences of overexpression of eNOS. For example, an insertional murine transgene in which the murine preproendothelin-1 promoter directed bovine eNOS in a murine model has been reported. Overexpression of eNOS by this promoter, which is known to induce transcription in the endothelium of both large vessels and the microvasculature as well as ectopically in vascular smooth muscle cells, evoked hypotension and attenuation of NO-elicited vasorelaxation in vivo (49). Recently Guillot et al. (20) described a line of transgenic mice containing 1.6 kb of the 5'-flanking region of the human eNOS promoter. Transgene expression occurred in micro- and macrovascular endothelial cells of the heart and in the blood vessels of the brain and skeletal muscle. In the coronary arteries the transgene was expressed only in a subpopulation of endothelial cells. For larger blood vessels, such as the aorta, the pattern of expression was not uniform in that the X-Gal reaction product was present in clusters of endothelial cells along the longitudinal axis of blood flow and in the regions around the ostia of aortic tributaries. Expression was absent in the vascular beds of the liver, kidney, and spleen, and only one of the founder mice showed detectable -galactosidase (LacZ) activity in a larger blood vessel of the lung. The authors concluded that this region of the promoter conferred eNOS expression by responding to vascular bed-specific pathways (20).

To examine the in vivo relevance of in vitro studies of the eNOS promoter, we have developed lines of transgenic mice carrying an insertional promoter-reporter transgene. Here we report the expression profile of transgenic murine lines (eNOSprom_nlsLacZ) hemizygous for a transgene (5,200/+28 Mu eNOS_nlsLacZ) where the -galactosidase (LacZ) reporter gene is under the transcriptional control of substantive portions of the native murine eNOS promoter. On the basis of previous in vitro studies (30, 39) and the high degree of cross-species homology (70), we postulated that this portion of the 5'-flanking region of the eNOS gene conferred the normal, restricted pattern of eNOS expression observed in vivo. In this model the expression of endogenous eNOS transcripts, and hence NO production, has not been disrupted. Findings reported here in adult mice suggest that the expression profile of the eNOSprom_nlsLacZ transgenic lines mirrors the constitutive expression pattern of eNOS in vivo.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Plasmid construction. A murine genomic eNOS clone was isolated from a Sau 3A partial-digest sublibrary prepared from a P1 bacteriophage clone isolated from a diploid Mus musculus cv129 genomic library (70). Eighteen hundred base pairs of the proximal promoter have been previously characterized (GenBank accession no. AF091262) (70). An antisense oligonucleotide (5'-GAG TCC CGG GTT GCC CAA GCC AGC TGA C-3') was used to generate a NgoM I-Sma I amplicon for the proximal promoter in which a 3'-Sma I site was created within exon 1 at +28 nt, numbered with respect to transcription, and the eNOS ATG site was mutated to prevent translation initiation. A 5,200-bp region of the murine eNOS promoter, which included 5,200 bp of the 5'-flanking region and 28 bp of the 5'-untranslated region (UTR), was isolated as a Sac I and Sma I fragment and subcloned into the 5'-Sma I site of the vector pLacC upstream of the LacZ ORF (gift of R. Palmiter, Howard Hughes Medical Institute, University of Washington, Seattle, WA). A 2,035-bp fragment containing the SV40 large T nuclear localization signal (5'-GGG CCC AAG AAG AAA CGC AAA GTG GGG AG-3') was excised with Sac I and then subcloned into the Sac I site directly upstream of the 5'-end of the LacZ ORF. The LacZ ORF included a eukaryotic translation initiation signal. A 516-bp Hind III fragment representing the 3'-UTR and 3'-flanking genomic regions of the human ₁-globin gene was incorporated at the 3'-end of the LacZ ORF to ensure proper RNA processing (43, 46). This 5,200/+28 Mu eNOS_nlsLacZ construct has been previously demonstrated to have functional promoter activity in transient transfection analyses of cultured endothelial cells (70).

Generation of transgenic mice. The construct 5,200/+28 Mu eNOS_nlsLacZ was linearized by Not I-Hind III digestion (8.7 kb) to remove vector sequences and purified by preparative gel electrophoresis (0.7% Sea Plaque, FMC Products, Rockland, ME). Recovery and purification of DNA was achieved through -agarase digestion (New England Biolabs). DNA was resuspended in microinjection buffer (10 mM Tris · HCl, pH 7.5, 0.1 mM EDTA) and purified by CsCl₂ preparative gradient ultracentrifugation in the absence of ethidium bromide (Beckman L8-70M, 70.1 Ti rotor, 49 K, 22°C). Aliquots were sequentially extracted from the cushion and subjected to analytic 0.7% agarose gel electrophoresis. Aliquots containing DNA were subsequently dialyzed extensively in microinjection buffer at 4°C. Founder mice were generated by pronuclear injection of the 5,200/+28 Mu eNOS_nlsLacZ DNA construct into C57BL × SJL F2 hybrid mouse eggs (National Institute of Child Health and Human Development Transgenic Mouse Development Facility, University of Alabama at Birmingham, Birmingham, GA) using standard approaches (59). Founder transgenic 5,200/+28 Mu eNOS_nlsLacZ mice were bred, housed, and monitored in accordance with standards set by the Canadian Animal Care Committee at the Ontario Cancer Institute Transgenic Facility (Toronto, ON, Canada). Hemizygous matings were accomplished by breeding founders with negative littermates. F1 progeny were screened for the presence of the transgene at the time of weaning (3-4 wk). Tail genomic DNA was extracted by proteinase K digestion (Boehringer Mannheim, Mannheim, Germany), followed by phenol-chloroform extraction and isopropanol precipitation. Genomic DNA samples were subjected to Southern blot analysis using an [-³²P]dCTP-labeled, nick-translated (Amersham Life Sciences, Amersham, UK) fragment (1.1 kb) from the LacZ ORF. Results were quantitated using a Molecular Dynamics PhosphorImager (Sunnyvale, CA) and ImageQuant software version 1.1 for Macintosh (Molecular Dynamics), as reported previously (15).

-Galactosidase staining. Organs (heart, lung, kidney, liver, spleen, and brain) from sexually mature hemizygous female F1 mice (age 7-11 wk) were excised, briefly rinsed in PBS, and then immersed in fixative solution (0.2% glutaraldehyde, 2% formaldehyde, 5 mM EGTA, 2 mM MgCl₂, and 100 mM sodium phosphate, pH 7.3). Microdissected organs were cut at 3.0-mm intervals to optimize fixation (4-6 h, 22°C ). Thereafter, fixed organs were briefly rinsed in PBS, blotted, and immersed in an aqueous X-Gal solution (Boehringer Mannheim; 1 mg/ml, 16 h, 22°C, pH 7.3, under light-protected conditions). Tissues were subsequently cryosectioned (5-7 µm) and counterstained with neutral red (Sigma-Aldrich) as previously described (32, 58). Expression of -galactosidase was assessed by multiple independent blinded observers.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Insertional transgene: 5,200/+28 Mu eNOS_nlsLacZ. Functional analyses of the human eNOS gene have defined important regulatory regions in the in vitro setting (30, 81). The promoter-reporter transgene 5,200/+28 Mu eNOS_nlsLacZ, numbered with respect to the transcription start site, contained 5,200 bp of the murine eNOS promoter directing expression of the LacZ gene. An SV40 nuclear localization signal was incorporated into the -galactosidase ORF to optimize histological staining and to demarcate any false positive background LacZ staining. To abet efficient processing of transcripts and produce a stable mRNA species, 516 nt of the human ₁-globin 3'-UTR and 3'-flanking genomic regions (43) were added downstream of the LacZ ORF (Fig. 1). For these studies it was elected not to use native eNOS 3'-UTR and 3'-flanking genomic sequences given the evolving understanding that eNOS mRNA species are rapidly degraded in such models of endothelial activation as cytokine treatment, hypoxia, and entry into the cell cycle (13, 40, 80). It appears that cis-RNA elements within the 3'-UTR of the eNOS mRNA are operative in this regulated expression (38). Murine eNOS 5'-flanking regions (5,200/+28) have previously been demonstrated to have functional promoter activity in transient transfection analyses of cultured endothelial cells. A 12.3-fold increase in activity was observed compared with activity of a promoterless construct, thereby confirming the competence of 5,200/+28 Mu eNOS_nlsLacZ to yield enzymatically active -galactosidase (70).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Schematic representation of 5,200/+28 Mu eNOSnlsLacZ construct used to generate insertional eNOSpromnls LacZ murine transgenic lines. -Gal, -galactosidase; eNOS, endothelial nitric oxide synthase gene; NLS, nuclear localization signal; ORF, open reading frame; UTR, untranslated region.

Histology. To examine the tissue distribution of 5,200/+28 Mu eNOS_nlsLacZ expression, organs from multiple hemizygous F1 progeny of founders (eNOSprom_nlsLacZ[1], eNOSprom_nlsLacZ[10], and eNOSprom_nlsLacZ[5]) were stained in toto with X-Gal. Organs were cryosectioned and counterstained with neutral red. Given the well-known sexual dimorphism observed in the expression of endothelium-derived NO, this body of work focused on expression patterns observed in sexually mature females. LacZ staining was absent in corresponding organs of transgene negative littermates (data not shown, n = 3). Representative tissues from transgenic mice are depicted in Fig. 2. The nuclear localization signal allowed LacZ staining of endothelial cells to be clearly distinguishable. We and others have previously reported that eNOS mRNA expression in humans appeared to be confined to larger arteries, with little or no expression detected in small arterioles, as assessed with in situ hybridization (75). Consistent with these prior studies, the following tissues derived from all the transgenic lines showed consistent staining of the endothelium: the aorta (Fig. 2a) and the coronary (Fig. 2, b and c), pulmonary (Fig. 2e), hepatic (Fig. 2f), renal (Fig. 2h), cerebral (Fig. 2i), splenic (not shown), and meningeal arteries (not shown). In all tissues small veins and venules were negative. Staining for the presence of -galactosidase was also notably absent in the microvasculature of the heart (Fig. 2c), lung (Fig. 2d), liver (Fig. 2f), spleen (data not shown), and brain (Fig. 2, j-l).

View larger version (167K): [in this window] [in a new window]   Fig. 2.   Representative cryosections of tissues derived from hemizygous eNOSpromnlsLacZ F1 mice stained for presence of -galactosidase (blue) and counterstained with neutral red; approximate magnifications are indicated. Shown are aorta (×250) (a), epicardial coronary artery (×650) (b), perforating intramuscular medium-sized coronary artery and adjacent myocardial tissue (×200) (c), pulmonary parenchyma (×100) (d), pulmonary artery and trachea (×200) (e), major intrahepatic branch of hepatic artery (×200) (f), renal papilla (×150) (g), bifurcation of main renal artery (×400) (h), cerebral artery (×700) (i), Purkinje neuronal cells of cerebellar cortex (×250) (j), hippocampus (×150) (k), and neurons of cerebral cortex (×250) (l).

View larger version (13930K): [in this window] [in a new window]   Fig. 3.   Expression of 5,200/+28 Mu eNOSnlsLacZ eNOS promoter-reporter in kidney microcirculation. A: absence of -galactosidase staining in cortical microcirculation of glomerulus and pars recta of eNOS transgenic mice (×300). B: -galactosidase staining in vasa recta of medullary circulation in eNOS transgenic mice (×200). This magnified view demonstrates positive staining in endothelial cell nuclei, but epithelial cells are uniformly negative. At arrows, where 2 cell types are in very close proximity, this distinction is clear. Red blood cells stain yellow in this micrograph. C: schematic representation of a juxtamedullary nephron illustrating the unique portal system known as the vasa recta.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Studies of constitutively expressed endothelial genes have examined the in vivo expression patterns of insertional promoter-reporter transgenic models to define the molecular basis of endothelium-restricted expression. Important issues have arisen from such studies: the relatedness of the reporter expression pattern to that of the native transcript, the reproducibility of findings across different founder lines, and the influence of the local tissue microenvironment on transgene expression (9, 22, 24, 71). For example, transgenic LacZ reporter mice containing ~2,200 bp of human von Willebrand factor (vWF) 5'-flanking regions and portions of the first exon and intron demonstrated that vWF transgene expression may be regulated by organ-specific transcriptional pathways, given that this expression pattern did not mirror the known broad endothelial distribution of the native vWF mRNA transcript (1, 2). Studies of human (5,000 bp) and murine (735 bp) Tie 1 promoter-LacZ reporter expression documented endothelial cell-restricted expression in large vessels and in the microvasculature of some, but not all, organs of adult mice. Expression of the reporter did not correspond to the expected adult expression pattern of the native Tie 1 mRNA (31). Early studies with the murine Tie 2 gene revealed that 7,200 bp of 5'-flanking regions were insufficient to emulate endogenous Tie 2 expression beyond stages of early vasculogenesis in transgenic mice (59). Subsequent Tie 2 promoter-reporter construct transgenic mice defined an endothelium-specific enhancer in the first intron of the murine gene that was critically important for both the endothelial cell specificity of Tie 2 expression in vivo (58) and the reproducibility of findings across independent founders. Nonetheless, the strong expression observed in the adult vascular endothelium of these mice was somewhat unanticipated because Tie 2 expression is downregulated in quiescent endothelium (57). It appears that the validity of utilizing the Tie 1 and Tie 2 genes to direct transgene expression in the adult setting is confounded by transcriptional pathways that are highly regulated during development and invoked in "response-to-injury" disease processes.

The important conclusions from the current work can therefore be compared and contrasted with those from prior studies that addressed the in vivo expression profile of endothelial-restricted promoter-reporters. All progeny of the eNOSprom_nlsLacZ lines consistently showed expression of the transgene in the large and medium-sized blood vessels of the heart, lung, kidney, liver, spleen, and brain. Moreover, eNOSprom_nlsLacZ transgenic mice recapitulated the constitutive expression pattern of eNOS in endothelial cells uniformly across founders. Small vessel expression was not detected with the methods used, except in the vasa recta of the kidney. Taken together, the 5'-flanking genomic region of murine eNOS, placed upstream of _nlsLacZ, functioned as a promoter in vivo and recapitulated the reported expression profiles of eNOS mRNA in vivo (33, 38, 75), even at the single-copy level.

Non-endothelial cell expression was observed in regions of the brain that were known to express eNOS in vivo: select neurons of the cerebral cortex, the CA1 neurons of the hippocampus, and the Purkinje neuronal cell layer of the cerebellum. There is evidence that neurons of the cerebellum utilize NO to communicate (14), and eNOS mRNA transcripts have been noted in the cerebellum and cerebrum (60). The distinct expression pattern of eNOS throughout the CA1 neuronal population of the hippocampus and its role in synaptic plasticity and/or long-term potentiation have previously been reviewed (12, 29). Our transgenic model also documents transcription of the eNOS promoter in the Purkinje cell layer of the cerebellum. Perhaps, as occurs in the hippocampus, the Purkinje cells of the cerebellum generate NO from eNOS and this is biologically relevant.

Prior work has sought to define a role for endogenous (3, 61) or eNOS-derived (21) NO in cardiac myocyte functioning. Northern blot, RT-PCR analyses, and in situ hybridization have shown the presence of eNOS transcripts in cardiac myocytes in vitro and in vivo (4, 60). In contrast, others have found no or very low levels of eNOS mRNA (25) and immunoreactive protein (49). The 5,200/+28 Mu eNOS_nlsLacZ expression pattern showed complete absence of LacZ staining in cardiac myocytes under baseline conditions in adult mice. Whether this insertional promoter-reporter fails to be expressed in cardiac myocytes because important transcription control regions are lacking on the genomic fragment utilized (5,200/+28) is not clear. An alternative explanation would be that transcription of the eNOS gene in cardiac myocytes is a response-to-injury phenomenon of diseased cardiac tissue. Also, we did not detect -galactosidase activity in bone marrow platelet precursors. Future studies are necessary to dissect these issues further.

Human eNOS promoter-reporter transgenic mice that incorporate 1,600 bp of 5'-flanking regions of the human eNOS gene coupled to LacZ have very recently been described (20). The results of that study are revealing in comparison with findings in the current work. Expression of LacZ was observed in endothelial cells corresponding to the aorta and major blood vessels of the heart and brain. However, large vessels were not positive in the liver, kidney, or spleen. Also notable is the lack of congruency across founders. For example, some of the founders exhibited expression in major vessels of the lung, whereas others did not. Non-endothelial cell expression was also noted in the brain of one founder. This suggested to these authors that integration site-specific expression was being observed and that local microenvironment effects might be relevant to the observed expression pattern of the transgene. Curiously, epicardial arteries showed expression of the transgene in only a subpopulation of endothelial cells, and staining in the aorta was described as patchy. It is also noteworthy that the microvasculature of the heart, brain, and skeletal muscle was positive in this model but not in the current work. The authors concluded that the eNOS promoter fragment used to develop their transgenic mice included DNA elements conferring vascular bed-specific expression and that elements outside this region may be responsible for expression in other tissues (20). The commonalities and differences of these two models will allow the details of cell-specific expression to be addressed in future studies. In this respect it should be noted that the current study used larger fragments of the promoter (5,200 vs. 1,600 bp) and that murine genomic sequences were used in the transgenic model. With these two studies as background, it is now possible to dissect out specific cis-regulatory elements involved in 1) transcriptional control differences between one endothelial cell type and another and 2) chromatin integration site-independent expression of endothelium-restricted genes. It can, however, be fairly argued that the expression profile of the murine eNOSprom_nlsLacZ transgenic mice more faithfully reproduces the predicted expression pattern of the native eNOS gene.

Vascular resistance determines the overall blood flow and, for a given perfusion pressure, is dependent on vessel number, size, and arrangement as well as the passive and active factors that alter their diameters (52). Control of vascular resistance is complex and heterogeneous, because it is altered by physiological and pharmacological factors. Indeed, the region of the murine blood vessel tree that imparts the greatest overall effects on vascular resistance is not firmly established. Clearly, NO (EDRF) is a significant regulator of blood flow in resistance arteries (19), and eNOS(/) mice are hypertensive with increases observed in total peripheral resistance (27, 63). However, robust endogenous expression of eNOS mRNA and protein in humans appears to be predominantly restricted to large blood vessels (75). Similarly, the vascular profile of 5,200/+28 Mu eNOS_nlsLacZ transgene expression was most prominent in large and medium-sized blood vessels. In comparison, expression of the reporter was notably diminished in small blood vessels. We posit that eNOS-derived NO from large or medium-sized vessels of an organ is especially noteworthy with respect to vascular resistance within that organ. Clearly, the microvasculature may, under certain biological scenarios such as development or during disease states, express appreciable levels of eNOS mRNA transcripts. It is less clear whether functional NO effects on resistance vessels can be solely ascribed to NO that is produced locally. It has been postulated that NO can be produced by the endothelium of larger blood vessels and then subsequently transported to the resistance vessels in the microvasculature complexed on circulatory proteins or within red blood cells, perhaps as S-nitrosothiols (65, 67). Clearly, this cannot be the sole mechanism of microvasculature NO action because paracrine EDRF pathways have been demonstrated in both isolated microvascular preparations and perfused organs in which blood was not the perfusate. Also, this work does not argue that physiologically important amounts of NO are not derived from the microvascular endothelium in organs, such as the heart. The biological activity of endothelium-derived NO is critically dependent on the local environment. For instance, amounts of NO necessary for physiological activation of soluble guanylate cyclase in neighboring vascular smooth muscle cells or pericytes may be much lower in vessels with a small radius or a limited number of contractile vascular smooth muscle cell layers. Future studies are necessary to examine the hypothesis that under normal conditions the biological activity of NO in the microvasculature may also prominently reflect synthesis and release of NO by proximal large and medium-sized vessels within solid organs, rather than synthesis solely at the level of the microvasculature per se.

Our failure to document microvasculature -galactosidase expression is not a technical limitation of this approach. A notable example of small vessel expression is the portal system of the kidney, specifically the vasa recta (Fig. 3). Expression of 5,200/+28 Mu eNOS_nlsLacZ in portal vessels was unique to the kidney. In comparison, the sinusoids of the hepatic portal system did not exhibit any LacZ staining. Mice from all founders demonstrated strong expression of the transgene deep within the medulla. Unique biological properties are now ascribed to this portal vascular bed. In contrast to the cortical circulation of the kidney, the medullary circulation is not autoregulated under normal conditions. Because a major contribution of this vascular bed is the reabsorption of marked amounts of water and solute produced via ultrafiltration within the glomerulus, the blood flow out of this vascular network is greater than the blood flow entering. Moreover, under conditions of antidiuresis the environment of the renal papilla is a very harsh environment for cells because it is characterized by marked gradients in the concentrations of O₂ and urea. Indeed, the appropriate functioning of the countercurrent multiplication system presents a unique environment in the body, especially for trafficking red blood cells. The vasa recta of the medullary circulation are an integral component of this tissue. For instance, aquaporin-1 (AQP1) is a member of a family of water-transporting proteins that is highly enriched in endothelial cells of the vasa recta (50), and a role for vasa recta expression of AQP1 in creation of a hypertonic medullary interstitium by countercurrent multiplication has been argued (35). UT3 (UT II) is a urea transport protein that is also enriched in endothelial cells of the descending vasa recta (53, 78) and is perhaps also important for the formation of a high urea concentration in the medullary interstitium (79). Why is eNOS enriched in the vasa recta? Diffusive loss of O₂ in the vasa recta is a feature of the countercurrent distribution of blood flow in the descending and ascending vasa recta (64), and hypoxia has been shown to suppress eNOS-derived NO production through transcriptional and posttranscriptional mechanisms (40). Therefore, the transcriptional and posttranscriptional processes that regulate eNOS in this tissue may of necessity be unique. Perhaps NO is also lost from the descending vasa recta by the countercurrent exchange process, as has been postulated for vasoactive peptides in this unique microvascular network (51). When oxygen binds to the heme irons of Hb, it promotes binding of NO to cysteine-93, thereby forming S-nitrosohemoglobin (SNO) in the R (oxygenated) structural conformation. Deoxygenation of SNO is followed by an allosteric transition to the T (deoxygenated) structural conformation state, which releases NO. Through an evolutionary conserved process, Hb is able to sense physiological oxygen gradients in tissues and, via SNO, to elicit increases in blood flow to address oxygen requirements (23, 68). This modeling of Hb and NO interactions was suggested to be important to the transit of NO from central to peripheral vascular networks. It could be argued that allosteric modifications of Hb that induce release of NO in peripheral tissues from SNO may not be present in the medullary environment, hence the need for local synthesis of NO. Our finding of strong expression of the eNOS reporter in vasa recta seems, on one hand, paradoxical given the prevailing view that medullary blood flow exhibits limited autoregulatory control, restricted primarily to the influence of the extracellular fluid status of the animal. On the other hand, robust expression of the eNOS reporter in vasa recta provides new insights into the biological control of renal interstitial fluid pressure, a major regulator of the pressure-natriuresis response (10).

In summary, the findings in the current work demonstrate that murine eNOS genomic regions spanning 5,200 to +28 nt direct expression of a reporter construct in a fashion that recapitulates the known expression profile of eNOS mRNA and protein. Future studies incorporating the eNOSprom_nlsLacZ strain of mice into other in vivo models of cardiovascular gene regulation will undoubtedly contribute information pertinent to our understanding of the interplay of gene regulation and NO production among the various NOS isoforms and other endothelium-specific genes. This expression cassette represents a novel and powerful tool with which exogenous genes may be reproducibly expressed in large and medium-sized blood vessels (e.g., green fluorescent protein, Cre recombinase, and Tet activator/repressor). It will also be of interest to utilize these reporter transgenic lines to examine expression of eNOS in models of injury and disease relevant to cardiovascular biology, particularly those characterized by perturbations of NO production and vascular reactivity [i.e., hypertension, atherosclerosis, and septic shock (7, 45, 48)].