INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE RENIN-ANGIOTENSIN SYSTEM (RAS) has been implicated in a variety of cardiovascular diseases (3, 5, 30). Activity of the RAS is thought to be regulated primarily at the level of release of renin from the juxtaglomerular cells through a negative feedback mechanism involving ANG I receptor inhibition (26). However, because extrarenal synthesis of the octapeptide ANG II appears to play an important role in activity of the RAS (5), the regulation of ANG II activity might be more complex. Less attention has been paid to the role that degradation plays in regulating ANG II levels. Aminopeptidase-A (APA; glutamyl aminopeptidase, EC 3.4.11.7) catalyzes the conversion of ANG II to [des-Asp¹]ANG II or ANG III (11, 17, 19). Although ANG III has significant biological activity, the half-life of ANG III is shorter than ANG II so that it probably represents the rate-determining step in ANG II degradation (1). Recently we reported that ANG II regulates the expression of APA in rat kidney in a cell-specific manner (9, 22). Because APA is widely distributed throughout the body, this would suggest that APA might play a physiologically significant role in regulating activity of the RAS in both renal and extrarenal tissues.

APA appears to be identical to BP-1/6C3, the pre-B cell differentiation antigen (31, 32). Expression of BP-1/6C3 in vitro has enzymatic activity that is indistinguishable from APA, and, conversely, knockouts of murine BP-1/6C3 reduce APA-like activity in tissues to background levels (14). To determine how APA is regulated in the rat, a more convenient-sized animal for conducting cardiovascular studies, we used a previously characterized partial cDNA for rat kidney APA (25) to isolate and clone partial cDNAs for APA from a kidney cDNA library. A full-length cDNA was constructed from overlapping clones and placed in a vector for expression in mammalian cells. Tissue distribution of APA was characterized by enzymatic assays, Northern blots, in situ hybridization, and immunohistochemistry.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

cDNA library screening. A rat kidney cDNA phage library constructed by priming rat kidney mRNA with oligo(dT) (Uni-Zap, Stratagene, La Jolla, CA) was screened with a previously characterized partial cDNA obtained by RT-PCR of rat kidney RNA (25). Initial screens (~4 × 10⁴ plaque-forming units per large petri dish) were phage transferred in duplicate to nitrocellulose membranes (Schleicher & Schuell, Keene, NH) and prehybridized at 59°C for 1 h in hybridization buffer minus probe. Membranes were hybridized overnight at 59°C in buffer [6× NET, 5× Denhardt's, and 0.1% SDS (20× NET = 3 M NaCl, 20 mM EDTA, 0.3 M Tris · HCl, pH 8.0)] containing a radiolabeled partial cDNA for rat kidney APA as previously described (25). The filters were then washed for 1 h at room temperature with four changes of 2× sodium chloride-sodium citrate (SSC) and 0.1% SDS followed by a 2-h wash at 59°C in 2× SSC and 0.1% SDS. The filters were then dried and exposed to X-ray film. Positive plaques were picked up and placed in SM buffer (100 mM NaCl, 10 mM MgSO₄, 50 mM Tris, pH 7.5, and 0.01% gelatin) and titrated. Additional rounds of screening positive phage plaques at greater dilution were conducted until the pure positive phage was obtained. The cDNA inserts were then rescued from the Uni-Zap phage according to the manufacturer's protocol using the R408 helper phage. The sizes and sequence of the inserts were determined by digestion of the individual plasmids with EcoR I and Xho I followed by gel electrophoresis and sequence analysis, respectively.

Initial screening resulted in isolation of several large clones (Fig. 1A, clones K1 and K2) whose sequence terminated in an open reading frame. A second rat kidney cDNA library in lambda gt10 that had been primed with random primers was kindly provided by Dr. Kevin Lynch (Charlottesville, VA) and screened as described above. cDNA inserts were digested directly from the lambda gt10 phage and subcloned into the EcoR I site of pBluescript.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Schematic drawing of rat kidney aminopeptidase-A (APA) cDNA and construction of APA expression vector. A: at top is a schematic of APA cDNA with coding region shown as a thick line and 5'- and 3'-untranslated regions shown as thin lines. Also shown are locations of major restriction sites. Bottom shows partial cDNAs that were isolated by library screening and rapid amplification of cDNA ends (5'-RACE) product. B: strategy for construction of APA expression vector. E, EcoR I; K, Kpn I; N, Nar I; S, Sac I; X, Xho I.

5'-RACE. Rapid amplification of cDNA ends (5'-RACE) was performed according to the manufacturer's recommendations (GIBCO-BRL, Grand Island, NY) to obtain sequence information on the 5'portion of the APA cDNA. First-strand cDNA synthesis of total rat kidney RNA was performed using an antisense primer (RACE-2, 5'-TGGTCAGCCGATAGACACTGTCCCC-3') based on sequence from the partial cDNA clone K5 (Fig. 1A) and Superscript II RT. An anchor sequence was then added to the 3'end of the single-stranded cDNA using TdT and dCTP. PCR was then done using an anchor primer (GIBCO-BRL) and a nested antisense primer (RACE-3, 5'-AGCTCCGGCAGCTTGGTGATC-3') and Taq polymerase (Perkin-Elmer, Norwalk, CT). The PCR product consisted of a single major band of ~600 bp (data not shown). The PCR product was subcloned into a TA cloning vector (Invitrogen) and used to transform bacteria. Individual colonies were selected, and the largest inserts were analyzed.

APA expression vector. A full-length cDNA expression vector for APA was constructed from partially overlapping cDNA clones (Fig. 1B). Plasmids (pBluescript, Stratagene) containing two clones isolated from the rat kidney cDNA library (pBlue-K1 and pBlue-K5) and the 5'-RACE product in the TA cloning vector (pCR-5') were digested with appropriate restriction enzymes (Fig. 1B) and ligated into the EcoR I and Xho I sites of the pcDNA3 (Invitrogen, Carlsbad, CA) expression vector. The full-length clone in pcDNA3 (pcDNA-APA) was fully sequenced to confirm that proper ligation had been obtained.

Transfection. Transient transfection of pcDNA-APA into mammalian cells was performed essentially as reported previously (8). Briefly, HEK-293 or CHO cells were grown to 50% confluency and transfected with the pcDNA3-APA expression vector using the Lipofectamine reagent (GIBCO-BRL) method. Control transfections consisted of the pcDNA-3 vector alone. Transfections were done in serum-free and antibiotic-free medium according to the manufacturer's protocol. One day before transfection, confluent cells grown in 15-cm petri dishes were split 1:3 and plated 4-6 × 10⁶ cells per 15-cm dish and grown overnight. For each transfection, 20 µg of DNA diluted in 1.8 ml of DMEM and 180 µl of Lipofectamine diluted in 1.8 ml of DMEM were combined, mixed gently, and incubated in RT for 10-15 min. After the incubation, 14.5 ml of DMEM was added and the mixture was overlayed onto the cells prewashed with DMEM. Cells were incubated overnight with DNA-Lipofectamine mixture. The next day the DNA-containing medium was replaced with normal growth medium containing FBS and antibiotics. The cells were either harvested 2 days later for immunoblotting or APA activity measured in situ within the wells.

In vitro translation. In vitro translation of the APA was performed using a coupled transcription/translation system (TnT Lysate System, Promega, Madison, WI). Briefly, the pcDNA3-APA vector (1 µg) was added to rabbit reticulocyte lysate containing T7 RNA polymerase, an amino acid mixture (1 mM) minus methionine, [³⁵S]methionine (sp act 10 mCi/ml, NEN), and RNasin (RNase inhibitor, 40 U/µl, Boehringer Mannheim, Indianapolis, IN) and incubated for 2 h at 30°C. An aliquot of reaction mix was diluted with sample buffer (2 ml glycerol, 2 ml 10% SDS, 0.25 mg bromophenol blue, 2.5 ml stacking gel buffer, 0.5 ml -mercaptoethanol) and separated on a 7% SDS-PAGE gel at 180 V for 2.5 h. The gel was then dried and exposed to X-ray film.

Immunoblots. Immunoblots of HEK cells transfected with the APA expression vector (pcDNA-APA) were conducted essentially as previously reported for immunoblots of rat tissues using the enhanced chemiluminescence (ECL) method (Amersham Pharmacia Biotech, Piscataway, NJ) (22). Briefly, after transient transfection, cells were harvested with lysis buffer, samples were mixed with loading buffer, and 10-25 µg protein were separated on a 10% SDS-PAGE gel. The protein was then transferred to nylon membranes at 70 V for 50 min. The membranes were washed with transfer buffer and incubated with primary antibody at a dilution of 1:1,000 in PBS buffer containing 2% BSA. The next day, the membranes were washed 3 × 5 min with PBS buffer at room temperature and then incubated with secondary antibody (1:3,000) in PBS buffer containing 3% BSA for 60 min. The membranes were then washed 3 × 5 min in PBS buffer, incubated with ECL solution, and exposed to film for 5 min.

Enzyme activity. Because APA is an ectoenzyme, APA activity can be measured in cultured cells within the culture dish without disruption of the cells. Similarly, for cells either transiently or stably transfected with APA, APA expression can be measured within the well. APA was measured in situ as described previously for membrane preparations (25) using -glutamyl-2-naphthylamide (Bachem Bioscience, Philadelphia, PA) as substrate. Enzyme specific activity was expressed as nanomoles per milligram per hour per well.

Northern blots. Northern blot analysis of total RNA from various tissues with a partial APA cDNA probe was conducted as previously described for rat kidney RNA (25). Briefly, tissue total RNA was isolated using the method of Chomczynski and Sacchi (4). Ten micrograms of total RNA was run on a denaturing 1% formaldehyde-agarose gel and transferred to nitrocellulose membranes (Nitro-pure, MSI Transfer Membrane, Fisher Scientific, Pittsburgh, PA). The membranes were then baked for 2 h at 80°C. The blots were hybridized with the APA partial cDNA labeled with -³²P-dCTP and primer extension with DNA polymerase I (Random Priming System, New England Biolabs). Prehybridization and hybridization were conducted at 44°C, and membranes were incubated overnight in hybridization buffer (25 mM KPO₄, pH 7.4, 5× SSC, 5× Denhardt's solution, 50 µg/ml salmon sperm DNA, 50% formamide, and 10% dextran sulfate). The next day, the membranes were washed two times for 15 min in 1× SSC plus 0.1% SDS at room temperature followed by two 15-min washes in 0.25× SSC plus 0.1% SDS at room temperature. The membranes were then wrapped in plastic wrap and exposed to X-ray film.

In situ hybridization. In situ hybridization of rat APA was performed identically as reported previously (25). Briefly, male Sprague-Dawley rats (175-200 g) were decapitated, and tissues were removed and frozen immediately on dry ice. Cryostat sections (10 µm) were placed on microscope slides and fixed with 3% paraformaldehyde in phosphate buffer for 5 min, dehydrated, and vacuum dried. An ³⁵S-labeled riboprobe, in either the antisense or sense orientation (control) to the APA mRNA, was transcribed in vitro using a partial APA cDNA as template. Sections were prehybridized for 2 h at 50°C in a 1:1 mixture of formamide and prehybridization buffer containing 0.6 M NaCl, 20 mM PIPES buffer (pH 6.7), 1× Denhardt's solution, 1 mM EDTA, 500 µg/ml yeast total RNA, 50 µg/ml yeast tRNA, 1 mg/ml salmon sperm DNA, and 500 µg/ml sodium pyrophosphate. Hybridization buffer was identical to prehybridization buffer, except that it contained 2.5 M -thiol nucleotide triphosphates and 200 mM dithiothreitol (Sigma). Slides were incubated overnight at 50°C. The next day, the sections were washed in 10 mM Tris · HCl (pH 8), 0.3 M NaCl, 0.1% sodium pyrophosphate, and 10 mM dithiothreitol and then treated with RNase (30 µg/ml) for 30 min at room temperature in 0.3 M NaCl, 10 mM Tris · HCl, and 1 mM EDTA. The sections were then washed for 10 min at room temperature in 0.3 M NaCl, 2 mM Tris · HCl, 1 mM EDTA, 0.05% sodium pyrophosphate, and 1 mM dithiothreitol followed by a 3-h wash at 50°C in 0.075 M NaCl, 2 mM Tris · HCl, 1 mM EDTA, and 1 mM dithiothreitol and then overnight with fresh buffer at room temperature. The next day, the sections were vacuum dried and exposed to Kodak XAR film for 1-20 days.

Immunocytochemistry. Immunocytochemical localization of APA was conducted in a similar manner as reported previously (25). Male Sprague-Dawley rats (175-200 g) were anesthetized with pentobarbital sodium (50 mg/kg ip) and perfused transcardially with ~100 ml of PBS (pH 7.4) followed by 250 ml of Zamboni's fixative (2% paraformaldehyde, 15% picric acid, 83% 0.12 M phosphate buffer, pH 7.4). Tissues were removed and postfixed for 1 wk in fixative. Slide-mounted cryostat sections were incubated overnight at 4°C with diluted immune sera (1:5,000 to 1:10,000) in PBS containing 0.1% Triton X-100 and 0.1% BSA. The sections were then rinsed with PBS and incubated 45 min with biotinylated goat anti-rabbit IgG (Vector Laboratories) diluted 1:222 with PBS-Triton X-100-BSA and then 45 min with streptavidin-horseradish peroxidase conjugate. The peroxidase reaction was developed by treating the sections with a 0.05% solution of 3,3'-diaminobenzidine (Sigma Chemical) containing 0.003% hydrogen peroxide in 50 mM phosphate buffer, pH 7.4, for 5-10 min. Some sections were lightly counterstained with 1% cresyl violet.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

A partial cDNA for rat kidney APA (25) was used to screen rat kidney cDNA libraries. A series of overlapping clones (Fig. 1) was isolated, but because the 5' terminus of each clone was still within an open reading frame it was apparent that the clones did not extend to the translation start site. Antisense primers (RACE-2 and RACE-3; Fig. 2) were constructed from the 5'-most clone and 5'-RACE performed with rat kidney poly(A)⁺ RNA to obtain the 5'end of the cDNA. A product of ~600 bp was characterized (5'-RACE; Fig. 1A). Sequence analysis indicated that the product extended 200 bp upstream of the translation start site. The 5'-RACE product was then used to rescreen the lambda gt10 rat kidney cDNA library. An additional partial cDNA clone was isolated and found to extend 38 bp beyond the RACE product (5'-clone; Fig. 1A).

View larger version (5249K): [in this window] [in a new window]   Fig. 2.   Complete cDNA sequence of rat kidney APA. A: nucleotide sequence +1 to 1768. B: nucleotide sequence 1769 to 3639. Shown is cDNA sequence with first nucleotide numbered +1. Below the nucleotide sequence is the predicted amino acid sequence of the protein. Boxed amino acid sequence from 19-40 represents single hydrophobic transmembrane-spanning domain. Shaded box represents Zn2+ binding domain with core sequence HELVHX18E. Irregular clear box represents sequence of previously characterized partial cDNA (25). Areas of sequence from which antisense primers were made for 5'-RACE are noted (RACE-2 and RACE-3) as is 5'end of 5'-RACE product (position relative to the isolated cDNA clone beginning at +1). The 5'-untranslated region start codon at position 209 is shown by thick underlining (A), and 3'-untranslated region RNA instability element (AUUUA) is underlined with a dashed line (B).

When overlapping cDNAs were combined, the full-length cDNA sequence was ~3.6 kb in length and contained an open reading frame that would encode a protein of 945 amino acids with a predicted molecular mass of 108 kDa. Human APA (i.e., BP-1/6C3) has been shown to be 957 amino acids in length and mouse APA 945 amino acids (12, 20, 31). Sequence comparisons indicated that rat APA was 85% identical with human APA and 93% identical with mouse APA (Fig. 3). Hydropathy analysis of rat APA indicated that there was a single hydrophobic domain from position +19 to +40. This is consistent with the wild-type protein known to be a type II membrane-associated protein with the amino terminus inside the cell and the catalytic domain contained within the large extracellular portion of the protein. The 5'-untranslated region (UTR) contains an ATG at position 209 followed by an in-frame TGA stop codon at position 221. The ATG at position 239 initiated the longest open reading frame of 945 amino acids and is the presumptive start codon. APA has been characterized as being a metalloenzyme with Zn²⁺ in the catalytic site (29). The signature Zn-binding motif HEVLHX₁₈E was shown previously to be essential for catalytic activity of human APA (28), and the amino acid sequence surrounding this region was conserved in all three species (Fig. 3). There were 12 possible N-linked glycosylation sites within the extracellular domain, of which six were conserved in rat, mouse, and human sequences. There were nine Cys residues within the rat APA sequence and seven within the extracellular domain, and, of these, all nine were conserved across the three species. APA exists as a homodimer, and its enzyme activity is inhibited by reducing agents, presumably reflecting an importance of intramolecular or intermolecular disulfide bonds. It is not clear which, if any, of these Cys residues are essential for APA activity. Finally, there are two conserved putative phosphorylation sites within the cytoplasmic domain. The Ser-9 residue is a protein kinase C consensus site, and Tyr-12 is a possible tyrosine kinase substrate site (Fig. 3). Although the BP-1/6C3 differentiation antigen is known to be phosphorylated (31), it is not known which residue(s) is phosphorylated.

View larger version (54K): [in this window] [in a new window]   Fig. 3.   Amino acid comparison between rat, mouse, and human APA. Amino acid sequences of the 3 proteins were aligned using GCG programs. Amino acids that are not conserved compared with rat are shown. For reference, hydrophobic domain is boxed in rat sequence and conserved Zn2+ binding domain outlined by shaded box. Possible phosphorylation sites within cytoplasmic domain are noted by * and N-linked glycosylation sites by #.

The 3'-UTR was ~563 bp in length. There was one RNA destabilization element (ATTTA) at position 3256 but no detectable polyadenylation signals (AATAAA) were seen, although the terminal region was AT rich (63%).

Three clones were spliced together and inserted into a mammalian expression vector to form pcDNA-APA (Fig. 1B). The size of the protein encoded by the APA cDNA was determined experimentally by in vitro translation using a rabbit reticulocyte lysate-coupled transcription/translation system. A major product slightly <110 kDa was obtained (Fig. 4A). This size is in agreement with the predicted size based on sequence analysis and represents the nonglycosylated form of the enzyme. There was some indication that the in vitro translation product was a doublet (Fig. 4A, inset). To determine the size of the protein expressed in mammalian cells, the expression vector was transiently transfected into HEK-293 cells and immunoblots were performed. There was a single major immunoreactive band of ~140 kDa (Fig. 4B), in close agreement with the size of the APA protein seen in immunoblots of rat kidney (25). Because HEK-293 cells have low levels of APA activity (data not shown), excess protein from nontransfected HEK-293 cells was loaded as a control and indicated that no prominent band could be seen in the 140-160 kDa region, suggesting that the band was specific for the transfected vector. Transient transfection of the pcDNA-APA expression vector in CHO cells resulted in an approximate sixfold increase in APA activity compared with CHO cells transfected with vector alone (Fig. 5).

View larger version (47K): [in this window] [in a new window]   Fig. 4.   In vitro expression of rat APA. A: in vitro translation. Expression vector pcDNA-APA was added to rabbit reticulocyte lysate with T7 polymerase and reacted at 30°C for 2 h. Lane 1, lysate without addition of DNA. Lane 2, lysate plus pcDNA-APA. Lane 3, same as lane 2 but shorter exposure. Protein size markers are shown on left (in kDa). Inset: magnification of predominantly labeled band from lane 3. Note that major product is slightly smaller than 110-kDa markers and with less exposure appears as a doublet. B: immunoblot analysis of APA after transient transfection of pcDNA-APA into HEK-293 cells. Lane 1, 10 g of HEK-293 cell lysate 48 h after transfection of pcDNA-APA. Lane 2, 20 µg of control HEK-293 cell lysate. Note that single major positive band of ~140 kDa (arrow) cannot be seen in nontransfected cells, even when 2-fold more HEK-293 cell protein was loaded.

View larger version (10K): [in this window] [in a new window]   Fig. 5.   APA activity in transiently transfected CHO cells. Cell surface APA activity was measured in cultured CHO cells transfected with vector alone or with pcDNA-APA expression vector. Enzyme activity was expressed as nmol · mg protein1 · h1 · well1. Data are expressed as means ± SE, n = 6 experiments.

The tissue distribution of APA was assessed by several different approaches, including enzymatic assays, Northern blots, in situ hybridization, and immunohistochemistry. APA activity was widely distributed (Table 1). APA activity was highest in the kidney and ileum. With the exception of the lung, there was general agreement between the intensity of the Northern blot analysis with the level of enzyme activity among the various tissues tested (Fig. 6). Hybridization with a partial cDNA for APA against total RNA from a variety of tissues resulted in labeling of an ~4.1-kb band. APA mRNA levels were highest in the kidney and ileum. Significant levels relative to the kidney were also seen in the liver and pituitary. Lower levels were seen in the heart, adrenal, and brain. We could not detect APA mRNA in the aorta, lung, or spleen. Interestingly, dissection of the heart into atrial and ventricular portions indicated that levels of APA mRNA were higher within the ventricle than the atrium. Also, RNA from LLC-PK₁ cells, a proximal tubule-like established cell line derived from pig kidney, yielded a similarly sized band as from rat tissues, consistent with LLC-PK₁ cells having detectable levels of APA activity.

View larger version (41K): [in this window] [in a new window]   Fig. 6.   Northern blot analysis of APA mRNA from rat tissues. Ten micrograms of total RNA was separated on a formaldehyde-agarose gel and transferred to nitrocellulose membranes and probed with a partial APA cDNA. Single major hybridized band was ~4.1 kb in size based on markers run in parallel (not shown). Kidney RNA was included in most blots for comparative purposes.

APA mRNA was also localized by in situ hybridization. In the kidney, APA mRNA was localized to the outer stripe of the outer medulla that appeared to follow the medullary rays and extend into the cortex (Fig. 7A). In the cortex, labeling appeared punctate. Previous studies using higher resolution autoradiography indicated that this pattern was due to labeling of proximal tubules and glomeruli (25). In the brain (Fig. 7B), APA mRNA was concentrated within the choroid plexus and the ependymal lining. Although less intense, a clearly recognizable pattern of labeling was seen adjacent to the dorsal third ventricle, consistent with labeling of the hypothalamic paraventricular nucleus. The labeling was not seen with tissue hybridized with sense strand probes (Fig. 7B'). The adrenal gland exhibited positive hybridization within the capsular region (Fig. 7C), corresponding to the zona glomerulosa. In the ileum, APA mRNA was very high in the basal portion of villi (Fig. 7D). In the pituitary, APA mRNA was primarily localized to the anterior lobe (Fig. 7E).

View larger version (129K): [in this window] [in a new window]   Fig. 7.   In situ hybridization of APA mRNA. Sections were hybridized with 35S-labeled antisense riboprobes synthesized in vitro from a partial APA cDNA. Sections were exposed to X-ray film for 2-4 wk, and reverse film images were examined. A: transverse section from rat kidney hybridized with an antisense APA riboprobe. Heavy labeling was seen in outer stripe (os) of outer medulla, and punctate labeling was seen in cortex (c). Inset: lower magnification of an adjacent section incubated with a radiolabeled sense strand APA riboprobe. B: coronal section from rat brain hybridized with an antisense APA riboprobe. Labeling was seen within ependymal (e) lining of brain, choroid plexus (cp), and hypothalamic paraventricular nucleus (pvn). B': same as B except incubated with a radiolabeled sense strand APA riboprobe. Ependyma, choroid plexus, and paraventricular nucleus are not labeled (arrows). C: section of rat adrenal hybridized with an antisense APA riboporobe. Detectable labeling was restricted to capsular region (arrows). C': same as section in C except incubated with a sense strand APA riboprobe. Capsular region is not labeled (arrows). D: section of rat ileum hybridized with an antisense APA riboprobe. Inset: lower magnification of an adjacent section incubated with a sense strand APA riboprobe. E: section of rat pituitary hybridized with an antisense APA riboprobe. Heavy labeling was seen within anterior lobe (a) but not intermediate (i) or posterior (p) lobes. Inset: lower magnification of an adjacent section incubated with a sense strand APA riboprobe.

The in situ hybridization pattern was then compared with the immunohistochemical staining pattern using antiserum against rat kidney APA. We previously reported that APA in the brain was primarily associated with microvessels and was not associated with neuronal or glial elements (10). This again appeared to be the pattern, but the Zamboni's fixative used here gave better localization to the vascular elements than the previous fixative that was used. As reported previously, heavy staining of vascular elements was seen within the hypothalamic paraventricular nucleus (Fig. 8A). This pattern exactly matches the pattern seen in the in situ hybridized sections (Fig. 7B). High resolution staining of the pattern of microvessel staining that was seen throughout the brain indicated clearly that the cellular elements that were stained were adventitial pericytes (Fig. 8, B and C). At high power, the pericyte cell body and thin processes were positively stained, whereas the endothelial cells lining the vessel wall were devoid of staining. In the heart, the in situ hybridization was nondescript (data not shown), and immunohistochemical staining indicated a fairly uniform distribution of staining within the ventricle, with staining restricted to the thin vascular elements, presumably capillaries, intermingled within the microfibrils (Fig. 8D). The liver, another organ with only a diffuse signal by in situ hybridization (data not shown), had positive staining within the central vein and the liver sinusoids (Fig. 8E). In the ileum, APA immunoreactivity was intensely rich within the enterocytes lining the lumen, with some impression that immunoreactivity was higher within the proximal portions of the microvilli (Fig. 8F). Within the adrenal gland, heavy APA immunohistochemical staining could be seen within the adrenal capsule (Fig. 8, G and H) and less intense staining within the vascular elements in the adrenal cortex and medulla. Higher resolution indicated that the cortical staining was localized to the lining of the sinusoids. The staining pattern within the pituitary (not shown) was very similar to that seen in the brain, with again, the staining being restricted to small vessels. Whereas the intermediate and posterior lobes were not enriched in RNA and thus appeared negative relative to the anterior pituitary in sections treated for in situ hybridization (Fig. 7E), immunocytochemical staining indicated that vessels were positive for APA immunoreactivity in a nearly identical pattern as seen in the brain. In the anterior lobe, the staining was even more intense but again only seen within the small vascular elements, presumably reflecting staining within capillaries.

View larger version (120K): [in this window] [in a new window]   Fig. 8.   Immunohistochemical staining of APA within rat tissues. A: low power view of paraventricular nucleus, showing heavily stained and densely distributed microvessels. Third ventricle is marked by *. Section was not counterstained. Bar = 200 µm. B: higher power view from rat cortex showing positively stained microvessels. Bar = 60 µm. C: high power magnification of a single positively stained microvessel. A pericyte cell body (thick arrow) is seen on outside of microvessel, whereas endothelial cell nuclei (thin arrows) are within vessel lumen and can be seen following inner contour of vessel wall. Immunoreactive staining can be seen extending from pericyte cell body along surface of vessel (arrowheads), whereas luminal surface of vessel (*) is devoid of staining. Bar = 10 µm. D: staining of microvessels (arrow) within cardiac ventricular muscle. Myofibrils (arrowheads) were not positively stained. Bar = 25 µm. E: liver section showing positive staining along lining of sinusoids and central vein (*). Bar = 60 µm. F: low power view of ileal microvilli. Note positive staining along entire surface of microvilli, although somewhat stronger staining nearer intestinal wall than at microvilli tips. Bar = 200 µm. G: low power view of a section of adrenal that has not been counterstained. Note heavy labeling in capsular region (arrow) and microvessels in adrenal cortex. Bar = 200 µm. H: higher power view of an adrenal section that has been counterstained. Staining was seen within capsule (arrow) and along lining of sinusoids. Bar = 50 µm.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Rat APA is highly similar to both mouse (93%) and human APA (85%). Areas functionally important for APA enzymatic activity were found to be highly conserved in all three species. In particular, APA is a zinc metalloenzyme (29) and the zinc binding domain and the regions immediately surrounding it are completely conserved. Interestingly, all nine Cys residues are conserved across all three species. Although APA is a homodimer and loses activity under reducing conditions, it is not clear whether the disulfide bonds are intramolecular or intermolecular. Mild treatment of the membrane-bound enzyme with papain can solubilize the enzyme activity, presumably by clipping the enzyme near the membrane surface (23). Under nonreducing conditions, the solubilized enzyme complex is smaller but retains activity, indicating that the dimerization does not require the transmembrane domains but not revealing whether the disulfide bonds are intramolecular or intermolecular.

The 5'-UTR contains an ATG at position 209 followed by an in-frame TGA stop codon at position 221 and, if translated, would encode a 4-amino acid peptide. Interestingly, the upstream ATG is contained within a more conserved Kozak consensus site than is the downstream start codon for the major open reading frame; compare ²⁰³T²¹² vs. ²³³AGAAAATGA²⁴². In the ribosomal scanning model of translation, the presence of the relatively well conserved Kozak site at 30 might indicate that translation arises from the upstream start codon. Furthermore, in this model, translation will not reinitiate at the downstream cistronic start codon, and downstream cistronic polypeptides are only translated if the upstream ATG is skipped. Another possibility is that the stop codon for the upstream codon is leaky and translation continues. This is an attractive possibility given that the UGA followed by C stop codon is not high fidelity (15). Furthermore, if the stop codon were skipped, the downstream initiator Met would be in frame. In this case, it would encode a 954-amino acid protein instead of 944 amino acids. Indeed, in vitro translation by rabbit reticulocyte lysates yielded a major product of ~108 kDa, the predicted size of the unglycosylated protein. Close inspection indicates that the major product appeared to be a doublet. In transiently transfected mammalian cells, however, there was no evidence of a doublet and, because APA is glycosylated, there was no indication from the size as to whether the protein contained the additional NH₂-terminal amino acids.

Ectoenzymes such as APA are generally considered to be housekeeping enzymes. Lending support to this notion are early studies from renal hypertensive patients (18). It was argued that because ANG II is a substrate for APA, renin-dependent renal hypertensive patients might be expected to have altered levels of APA. However, plasma levels of APA were unchanged, arguing that APA expression was insensitive to changes in ANG II. We reported recently, however, that plasma levels of APA do not accurately reflect what is going on within tissues (9, 22) and because ANG II is synthesized in extrarenal tissues (5), regulation of the half-life of ANG II might play a more important role in regulating ANG II levels than previously appreciated. Specifically, we found that within two-kidney, one-clip hypertensive rats, glomerular APA was increased in both clipped and nonclipped kidneys, whereas tubular APA was diminished (22). Likewise, ANG II infusion increases glomerular APA expression and angiotensin-converting enzyme inhibition decreases APA activity in spontaneously hypertensive rats (9). Thus it seems as if APA can indeed be regulated by ANG II in a cell-specific manner.

As our previous studies on the localization of APA have been conducted primarily within the kidney (25) and brain (10, 23), we extended the study here to localize APA to additional tissues. First of all, Northern blot analysis and direct measurement of APA activity were generally in agreement, with the exception of the lung, where APA activity was measurable but mRNA was not detected. APA mRNA and APA activity were highest within the kidney and ileum. Previously, we reported that APA within the kidney is heavily localized to the brush border of proximal tubule cells, with lesser staining detected within glomeruli. Both the immunohistochemical staining pattern and the in situ hybridization labeling pattern within glomeruli were consistent with expression within mesangial cells (25), and, indeed, cultured mesangial cells have APA enzyme activity and detectable APA mRNA (27). In the ileum, APA immunohistochemical staining was restricted to the brush border of enterocytes lining the intestinal microvilli. The localization pattern by low resolution in situ hybridization was consistent with this and suggested that the mRNA levels were higher in the proximal portion of the villi than in the distal tips. In general, the high levels of expression of APA in the brush border of both proximal tubule cells and ileal enterocytes are consistent with the role of these tissues in the degradation of proteins and peptides and the reabsorption of essential amino acids for nutritive purposes. These areas are enriched with a battery of proteolytic enzymes, and APA is thought to participate in the degradation of peptides generated from such degradation. Interestingly, APA would not appear to be critical in this function because BP-1/6C3 (i.e., APA)-deficient mice develop normally (14).

We previously reported that in the brain APA was localized primarily to small vascular elements, the choroid plexus, and the ependymal lining of the brain (10). The in situ hybridization was consistent with this pattern of expression. Because APA is so widely distributed, the in situ hybridization pattern was diffuse and only concentrated in structures expressing higher levels, such as the choroid plexus, or in structures more heavily vascularized, such as the paraventricular nucleus of the hypothalamus (compare Figs. 7B and 8A). Whereas in previous studies based on the pattern of staining we argued that the distribution of APA within microvessels within the brain was consistent with localization to adventitial pericytes, here, using an alternative fixation approach, the staining evidence is more convincing in this regard. Quite clearly, the positive-stained cells are on the outer surface of the vessels and can be distinguished from the endothelial cells lining the lumen (see Fig. 8C). Interestingly, pericytes and renal mesangial cells are thought to have similar origins (21) and it is thus consistent that both cell types express APA. Moreover, as we demonstrated that pericytes readily take up large molecules such as horseradish peroxidase (10), the localization of APA to the adventitial lining of cerebral capillaries would argue that APA plays a protective role as part of an enzymatic component of the blood-brain barrier. Interestingly, Alliot et al. (2) recently reported using electron microscopy that immunoreactive APA is localized on the abluminal surface (i.e., facing the interstitial space) of brain pericytes. Thus APA may be preferentially positioned to metabolize peptides arising from the brain parenchyma rather than from the circulation. This may explain why immunostaining of microvessels for APA is enriched in brain areas that also contain ANG II immunoreactive neurons, areas such as the hypothalamic paraventricular nucleus (10). We have also shown that an APA inhibitor or APA antiserum with anticatalytic activity administered within the cerebroventricular space inhibits the actions of intracerebroventricularly administered ANG II (24). Together these results suggest that expression of APA on the abluminal surface of brain microvessel pericytes may play a functional role in modulating activity of the brain ANG system.

We further localized APA to other tissues. In the heart, ANG II is known to have positive inotropic effects on cardiac function (5). Northern blot analysis indicated cardiac APA mRNA was higher within the ventricular muscle than within atrial muscle. Immunohistochemical localization indicates that this simply might reflect differences in microvessel density between these two areas, because in both atrium (not shown) and ventricle (Fig. 8D), APA was restricted to capillaries or very small caliber arterioles or veins. APA might thus influence the levels of circulating or locally produced ANG II within the heart. In the adrenal gland and liver, APA was associated with the lining of sinusoids. Again, one could speculate that this localization is consistent with either processing or degradation. In addition to sinusoids within the adrenal cortex, APA was very high within the capsule directly adjacent to (or overlapping) the zona glomerulosa. Because blood flow within the adrenal gland is from the superficial cortex to the inner medulla, ANG II reaching the adrenal gland via the circulation would be accessible to high levels of APA. Indeed, this localization pattern is consistent with physiological studies that indicate ANG III is equieffective as ANG II at stimulating aldosterone from the adrenal cortex (7) and may reflect the metabolism of ANG II to ANG III. Finally, the anterior pituitary had high levels of APA mRNA and APA immunoreactivity. ANG II is known to stimulate release of ACTH and prolactin from the anterior pituitary (6). Thus APA may play an important role in regulating ANG II levels within the anterior pituitary, either locally synthesized or from the general circulation or, possibly, the portal circulation, because ANG II immunoreactivity is present in the externa lamina of the median eminence (10).

Although our focus has been primarily on the role of APA in metabolism of ANG II, APA is also the principal enzyme involved in the degradation of CCK-8 (16), the other known biologically active peptide that is a substrate for APA. CCK-8 is a gastrointestinal hormone that stimulates pancreatic secretion, gallbladder contraction, gastric emptying, and induction of satiety (13). Although APA activity or mRNA levels are not high in the pancreas, APA levels are very high in the small intestine, suggesting that APA might be important in regulating CCK-8 levels locally.

Perspectives