INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

STUDIES OF RAT AORTIC SEGMENTS in vitro (1, 40, 41) have demonstrated a characteristic regional pattern of functional differentiation of the endothelium for the generation of three classes of signals capable of regulating the cGMP content of the underlying vascular smooth muscle. Endothelial formation of nitric oxide (NO) from L-arginine (1) and of glutamate from L-leucine (41) increase cGMP, which is reported to mediate relaxation of vascular smooth muscle (21, 35, 38). Endothelial formation of prostaglandin H₂ (PGH₂) via the cyclooxygenase arm of the eicosanoid pathway decreases cGMP (40). All three pathways exhibit a similar regional pattern of activity: maximal values just distal to the aortic arch, i.e., in the "windkessel region," and progressively decreasing values distally. On the basis of these observations and the functional importance of the windkessel elastic reservoir, whose compliance and other viscoelastic properties influence hemodynamic parameters, such as the arterial systolic and pulse pressures, work of the left ventricle, and pulse-wave velocity (15), we have hypothesized (41) that an interplay of the NO, leucine-to-glutamate (LeuGlu), and PGH₂ signal pathways mediates the dynamic regulation of smooth muscle contractility and aortic compliance in the windkessel region.

The LeuGlu pathway, the most recently described of these mechanisms, is composed of the endothelial uptake of L-[U-¹⁴C]leucine, the relatively efficient utilization of the leucine carbons for synthesis of [¹⁴C]glutamate, and the transfer of the glutamate to the underlying smooth muscle via a compartmentalized mechanism (41). This study characterizes the metabolic route of the LeuGlu pathway and examines its role in the regulation of aortic contractile responses. It is well established that leucine, an essential, "ketogenic" amino acid, is metabolized via 3-hydroxy-3-methylglutaryl (HMG) coenzyme A (CoA) to acetyl CoA and acetoacetate by a reaction sequence localized largely in the mitochondria (12, 47). Acetyl CoA so derived can enter the tricarboxylic acid cycle by the condensation reaction with oxaloacetate and result in the formation of -ketoglutarate, which can be converted to glutamate by the glutamate dehydrogenase reaction or by transamination. The studies below show that the aortic LeuGlu conversion utilizes this route. Aortic segments incubated in vitro with radioactive leucine labeled at different positions or with other radioactive precursors yield radioactive glutamate and CO₂ as predicted qualitatively and quantitatively by the catabolic sequence. Thus [U-¹⁴C]leucine, which is ketogenic, yields [¹⁴C]glutamate efficiently; [U-¹⁴C]isoleucine, only partially ketogenic, yields considerably less labeled glutamate; and [U-¹⁴C]valine, not ketogenic, yields insignificant labeled glutamate. Furthermore, the experimental results indicate that the tissue specificity and overall activity of the LeuGlu pathway is determined by three initial reactions of the catabolic sequence: cellular uptake of leucine, transamination to -ketoisocaproate, and conversion of the latter to isovaleryl CoA via oxidative decarboxylation by the branched-chain dehydrogenase complex (BCDC). The BCDC reaction is the controlling reaction whose activity accounts for the regional LeuGlu pattern of the rat aortic endothelium.

To examine the role of the LeuGlu pathway in the regulation of aortic smooth muscle contractility, we studied the effects of blocking the pathway on the contractile responses of aortic segments in vitro. Given prior evidence that LeuGlu blockade owing to treatment with leucine analogs decreases the smooth muscle cGMP content (41), we predicted increased contractile responses. In the present study, three different classes of compounds capable of interrupting the LeuGlu pathway were examined to ensure the specificity of the inhibition: 1) the leucine analogues L-leucinol and L-leucinamide, previously shown to block the conversion of [U-¹⁴C]leucine to [¹⁴C]glutamate and to decrease aortic segment cGMP (41); 2) -ketoisovalerate and its amino acid precursor L-valine, which do not yield acetyl CoA or glutamate but are metabolized by the same BCDC (36) that functions in the controlling reaction of the LeuGlu pathway; and 3) D-aspartate, a relatively specific and metabolically inert inhibitor of Na⁺-dependent glutamate transporters (17, 28, 42). The results demonstrate that interruption of the pathway by all three classes of compounds significantly enhances the contractile responses of aortic segments in vitro precontracted with phenylephrine or serotonin (5-hydroxytryptamine; 5-HT). The enhanced contractile responses elicited by each of the three inhibitor classes require an intact endothelium and are independent of the NO pathway, i.e., they are not influenced by inhibiting NO formation.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animals. Sprague-Dawley strain male rats weighing ~250 g were obtained from Charles River and maintained on a nutritionally complete pellet diet (Purina rat chow 50001) and water ad libitum. Rats weighed 250-450 g when used in the experiments.

Preparation and incubation of aortic slices. Rats were anesthetized rapidly by inhalation of carbon dioxide and were then exsanguinated. The aorta was excised from its origin at the aortic valve ring to the approximate level of the left renal artery, and the segment, ~7.5 cm long, was chilled immediately in 145 mM NaCl-5 mM KCl and maintained at 2°C until the onset of incubation ~10-20 min later. The adherent adventitial tissue was removed, the vessel was cut open longitudinally and sliced into horizontal segments, and each segment was bisected vertically for comparison of a control and experimental sample at a given level of the aorta. Comparable slices from 3-4 rats were usually pooled to provide sufficient tissue for analysis. To study the formation of [¹⁴C]glutamate, the slices were suspended in 2.5 ml of an incubation medium of the following general composition (in mM): 118 NaCl, 4.7 KCl, 2.0 CaCl_2, 1.2 MgSO_4, 1.2 KH₂PO₄, 25 NaHCO₃, and 10 D-glucose. The medium also contained unlabeled precursor, e.g., 50 or 100 µM L-leucine, and a radioactive compound, e.g., 0.3 µCi/ml of L-[U-¹⁴C]leucine, specific radioactivity 314 mCi/mmol (Amersham). Suspensions were generally shaken for 60 min at 37°C in an atmosphere of 5% CO₂-95% O₂. Thereafter, the tissues were removed, blotted to remove adherent medium, weighed rapidly, and frozen immediately in vials immersed in an ethanol-dry ice bath. Frozen tissues were maintained at 15°C until assayed. When appropriate, the endothelium was denuded before incubation of the slices, as previously described (1). Tissue uptake of [¹⁴C]leucine and estimation of the [¹⁴C]glutamate formed were determined as described previously (41). Uptake was quantified by the disappearance of labeled precursor from the ambient medium and [¹⁴C]glutamate formation by the preparation of aqueous extracts of the tissues, resolution of the extracts by either thin-layer chromatography and autoradiography or by anion exchange column chromatography and quantification of the radioactive glutamate fractions by liquid scintillation spectrometry. [¹⁴C]Glutamate formation (expressed as nmol/g wet weight of tissue) was calculated from radioactivity so quantified and the specific radioactivity of the precursor or of an acetyl group of the precursor [U-¹⁴C]leucine. As indicated by the studies below, one two-carbon moiety of leucine is incorporated per glutamate formed.

Estimation of ¹⁴CO₂ formed. To estimate the ¹⁴CO₂ evolved via the tissue metabolism of [¹⁴C]leucine, aortic segments prepared as described above were incubated in 2.5 ml of the following medium (in mM): 142 NaCl, 4.7 KCl, 1.0 CaCl₂, 1.2 MgSO₄, 4.0 Na⁺ phosphate (pH 7.4), 10 D-glucose, and appropriate unlabeled and radioactive ¹⁴C precursors. Tissues were incubated for 1 h in closed vessels containing a KOH trap, under an atmosphere of 100% O_2, at 37°C. ¹⁴CO₂ was estimated as the carbonate by liquid scintillation counting in ScintiSafe Plus 50% counting solution (Fisher) and the values corrected for the recovery of ¹⁴CO_2, which averaged 78.8%.

Preparation and assay of aortic homogenates. Aortic segments were homogenized in a sucrose medium containing antiproteases (2) and the suspensions centrifuged at 800 g for 10 min at 2°C to remove tissue clumps. The supernatant suspensions containing ~1.0 mg/ml of protein were then assayed for leucine--ketoglutarate transaminase and alanine--ketoglutarate transaminase activities in 0.15 ml of the following reaction mixture (in mM): 133 Tris buffer (pH 7.6), 0.13 pyridoxal phosphate, 6.58 Na⁺ -ketoglutarate, 6.58 L-leucine or L-alanine, and 0.132 µCi/ml of -keto- [1-¹⁴C]glutarate (specific radioactivity 51.8 mCi/mmol; New England Nuclear). Reaction was initiated by addition of homogenate containing 100 µg of protein and the mixtures incubated with shaking at 37°C for 20 min. Thereafter, aqueous extracts were prepared and assayed for [¹⁴C]glutamate as previously described (41).

Estimation of glutamate. Tissue content of free glutamate was estimated enzymatically using glutamate dehydrogenase, as described by Lund (31). Approximately 50-60 mg of aortic tissue were suspended and mixed in 1.0 ml of 10% perchloric acid at 2°C for 1 h. After centrifugation (11,500 g for 10 min) the supernatant solutions were neutralized with KOH, insoluble potassium perchlorate was removed by centrifugation, and the neutralized supernatants were lyophilized. The reaction mixture for enzyme assay (total volume 1.1 ml; final pH 8.7) had the following composition: 90 mM Tris buffer, 1.8 mM EDTA, 0.57 M hydrazine, 1.5 mM -nicotinamide adenine dinucleotide, 0.5 mM ADP, and an appropriate aliquot of either sample, L-glutamate standard, or water (control). Reaction at room temperature was initiated by the addition of 200 µg (20 µl) of beef liver glutamate dehydrogenase (120 U/mg; Boehringer-Mannheim) and the increase in optical absorbance at 339 nm owing to the formation of reduced nicotinamide adenine dinucleotide followed spectrophotometrically.

Contractile responses of aortic rings. After the rats were anesthetized by inhalation of carbon dioxide and were exsanguinated, the aorta was excised, adherent adventitial tissue was removed, and the vessel was placed into the isotonic incubation medium (see Preparation and incubation of aortic slices) equilibrated with 5% CO₂-95% O₂ at 23°C and containing 50 µM L-leucine and 0.1 mM ascorbate. The windkessel region, i.e., the segment from ~15-45 mm from the aortic ring, was used, unless indicated otherwise, and 7-mm rings (mean wet wt 10.1 ± 0.6 mg) were cut. Each ring was mounted by means of stainless steel hooks to record isometric tension in a water-jacketed 50-ml chamber maintained at 37°C. The incubation medium described above was equilibrated with 5% CO₂-95% O₂ and recirculated through the chamber from a 300-ml reservoir at a rate of 32 ml/min. Isometric tension was recorded via a force transducer linked to a Macintosh computer using MacLab software. The aortic segments were mounted within 10-15 min after the death of the animal. Initial tension was adjusted to 1.0 g and so maintained for 60-90 min. Thereafter, the recirculated medium was replaced with fresh medium and control observation periods begun. Segments were generally precontracted with either phenylephrine (10⁸-10⁶ M) or 5-HT (5-10 µM). Test compounds were added when plateau precontraction tension values were attained.

5

Materials. In addition to [U-¹⁴C]leucine and -keto-[1-¹⁴C]- glutarate, the following labeled compounds were used: 342 mCi/mmol L-[U-¹⁴C]isoleucine, >250 mCi/mmol L-[U-¹⁴C]- valine, 60 mCi/mmol L-[1-¹⁴C]leucine (all from New England Nuclear), 54 mCi/mmol 2-keto[1-¹⁴C]isocaproic acid (Amersham), and L-[4,5-³H]leucine (50 Ci/mmol; DuPont-New England Nuclear). The following were purchased from Sigma (St. Louis, MO): D-aspartic acid, -ketoisovalerate (3-methyl-2-oxobutanoic acid, Na⁺ salt), -ketoisocaproate (4-methyl-2-oxopentanoic acid, Na⁺ salt), L-leucinol (2-amino-4-methyl-1-pentanol), L-leucinamide hydrochloride, N-nitro-L-arginine methyl ester (L-NAME), l-norepinephrine bitartrate; L-phenylephrine hydrochloride, serotonin (5-hydroxytryptamine creatinine sulfate complex), and L-valine.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Glutamate formation from leucine, isoleucine, and valine. The formation of [¹⁴C]glutamate from L-[U-¹⁴C]- leucine by sequential aortic segments and its characteristic regional activity pattern was reported previously (41). Identical experiments were repeated using L-[U-¹⁴C]isoleucine or L-[U-¹⁴C]valine and the formation of [¹⁴C]glutamate estimated as described in Preparation and incubation of aortic slices. Each precursor was examined in each of 3 experiments and the results are illustrated in Fig. 1. Glutamate formation, whether expressed in absolute units (nmol · g wet wt¹ · h¹) (Fig. 1A) or as the ratio of tissue [¹⁴C]glutamate/[¹⁴C]precursor (Fig. 1B), a parameter independent of tissue weight varies with the extent to which each precursor is metabolized to acetoacetate and acetyl CoA, i.e., is "ketogenic." From the known metabolic pathways (34, 47) for catabolism of these three amino acids 5/6 carbons of leucine, 2/6 carbons of isoleucine, and 0/5 carbons of valine are utilized for acetyl CoA formation. On the premise that glutamate formation occurs via the acetyl CoA so formed, the relative incorporation of ¹⁴C from leucine/isoleucine/valine is therefore predicted to be 1.0:0.4:0.0. The areas under the curves in Fig. 1B yield corresponding observed values of 1.0:0.41:0.02 in close agreement.

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Formation of [14C]glutamate from [U-14C]leucine, [U-14C]isoleucine, and [U-14C]valine by sequential segments of rat aorta in vitro. Values are means ± SE for each of 3 experiments (9 rats) plotted as the final tissue content of [14C]glutamate (A) or the final tissue ratio of [14C]glutamate to [14C]- amino acid precursor (B). Aortic segments were prepared and incubated for 1 h with 50 µM amino acid precursor as described in MATERIALS AND METHODS. The values for [14C]glutamate formed in nmol/g wet weight of tissue (A) were calculated using the specific radioactivity of each amino acid precursor. The values for leucine were taken from Fig. 2 of our previous publication (41).

Formation of ¹⁴CO₂ and [¹⁴C]glutamate. The initial steps for metabolism of leucine are uptake into the endothelial cells and transamination to yield -ketoisocaproate. The keto-acid is then oxidatively decarboxylated via the BCDC to yield CO₂ from carbon-1 and isovaleryl CoA (34, 47). If the LeuGlu pathway follows these steps, a [1-¹⁴C]leucine precursor incubated with aortic segments should yield ¹⁴CO₂ but no [¹⁴C]glutamate. Accordingly, sequential aortic segments, both intact and denuded of endothelium, were incubated with [1-¹⁴C]leucine in each of six experiments. No [¹⁴C]glutamate product was detected in aqueous extracts examined by thin-layer chromatography and autoradiography. The evolution of ¹⁴CO₂ formed from carbon-1 was quantified, and the results are shown in Fig. 2A. Oxidative decarboxylation of leucine follows the regional differentiation pattern characteristic of the windkessel area and is largely eliminated by prior removal of the endothelium. Oxidative decarboxylation as a percentage of leucine uptake from the ambient medium by each of five sequential aortic segments was (means ± SE) 22.4 ± 3.5, 32.8 ± 6.0, 27.8 ± 4.6, 19.1 ± 5.1, and 16.8 ± 2.8%.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Formation of 14CO2 from [1-14C]leucine (A) and -keto-[1-14C]- isocaproate (B) by sequential, intact and endothelium-denuded segments of rat aorta in vitro. Values are means ± SE for 6 experiments (18 rats) with leucine and 4 experiments (12 rats) with -ketoisocaproate. Sequential aortic segments were prepared, split longitudinally, denuded of endothelium, and were incubated for 1 h with 100 µM precursor as described in MATERIALS AND METHODS.

View larger version (33K): [in this window] [in a new window]   Fig. 3.   Formation of [14C]glutamate and 14CO2 from [U-14C]leucine by sequential, intact, and endothelium-denuded segments of rat aorta in vitro. Values are means ± SE for 3 experiments (9 rats). Aortic segments were prepared and tested as described in the legend to Fig. 2. Values (nmol/g wet weight of tissue) for [14C]glutamate and 14CO2 formed, respectively, were calculated using the specific radioactivity of a 2-carbon and 1-carbon constituent of the precursor [U-14C]leucine. , Intact segments; , denuded segments.

Studies with -ketoglutarate. Inasmuch as the initial steps of the LeuGlu pathway are endothelial uptake of leucine and transamination by reaction with a suitable -ketoacid, such as -ketoglutarate, we studied whether the activity of the transamination reaction is rate limiting in the pathway. Successive aortic segments were incubated with [1-¹⁴C]leucine in the presence and absence of 2-5 mM -ketoglutarate and the ¹⁴CO₂ produced was quantified. Evolution of ¹⁴CO₂ in four experiments followed the activity pattern illustrated in Fig. 2A both in the absence and presence of added -ketoglutarate. Addition of -ketoglutarate did not increase significantly the ¹⁴CO₂ produced in the windkessel segment, 15-60 mm from the valve ring. Transaminase activities of homogenates prepared from five successive aortic segments were also tested as described in Preparation and assay of aortic homogenates by following the formation of [¹⁴C]glutamate from -keto-[1-¹⁴C]glutarate and leucine or alanine. The mean values observed (52.0 ± 9.3 and 75.0 ± 11.5 nmol · mg¹ · min¹, respectively) for leucine and alanine transaminases did not vary significantly with segment or exhibit the differentiation pattern of the windkessel region.

Requirement for glucose. Synthesis of glutamate via the LeuGlu pathway requires a supply of oxaloacetate for the condensation reaction with acetyl CoA, and in our experiments D-glucose was routinely used as a source of oxaloacetate. A requirement for the monosaccharide was demonstrated in two experiments, in which successive aortic segments were incubated with [U-¹⁴C]leucine in the presence and absence of 10 mM glucose. Formation of [¹⁴C]glutamate was decreased by 67.8 ± 5.4% (means ± SE of 10 trials; P < 0.001) in the absence of glucose.

Requirement for Na⁺ gradient. Endothelial cells take up leucine by both Na⁺-dependent and Na⁺-independent transport mechanisms (32). The following experiments demonstrate that leucine uptake dependent on a favorable Na⁺ gradient extracellular/intracellular is required for formation of glutamate via the LeuGlu pathway. Successive aortic segments were incubated with [U-¹⁴C]leucine either in the control Na⁺-containing medium (see Preparation and incubation of aortic slices), in the same medium containing K⁺ or Li⁺ in place of Na⁺, or in the control medium containing in addition 1.0 mM ouabain. The formation of [¹⁴C]glutamate was estimated and the results are illustrated in Fig. 4. Replacement of Na⁺ with K⁺ or Li⁺, respectively, decreased [¹⁴C]glutamate formation by 93.6 ± 0.6% (mean ± SE; 10 trials; P < 0.001) and by 61.7 ± 6.0% (5 trials; P < 0.05). Elimination of the favorable Na⁺ gradient by ouabain treatment inhibited the glutamate formation by 84.0 ± 2.3% (11 trials; P < 0.001).

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Effects of Na+ replacement or ouabain treatment on the formation of [14C]glutamate from [U-14C]leucine by sequential rat aortic segments in vitro. Values are means ± SE for 5 control experiments with 15 rats (trace A), in which the segments were prepared and incubated with 50 µM precursor leucine as described in MATERIALS AND METHODS. Ambient medium Na+ was replaced isosmotically with K+ in 3 experiments with 9 rats (trace D) and with L+ in one experiment with 3 rats (trace B). Segments were tested with 1.0 mM ouabain in 3 experiments with 9 rats (trace C). Values for [14C]glutamate formed were calculated as described in the legend to Fig. 3.

Glutamate content of aortic segments. The content of total free glutamate in five successive aortic segments (originating at the aortic valve ring and each 15 mm long) was estimated after incubation in vitro for 1 h at 37°C. The values (µmol/g wet weight) in seven experiments (30 rats) were 0.67 ± 0.11, 1.25 ± 0.09, 1.43 ± 0.15, 0.82 ± 0.08, and 0.67 ± 0.21, respectively. Glutamate content thus exhibited a regional pattern similar to that previously reported (41) for the endothelial uptake of [¹⁴C]leucine and its conversion to [¹⁴C]glutamate.

Effects of L-leucinol. L-Leucinol effectively inhibits the LeuGlu pathway and lowers the cGMP content of aortic segments in vitro (41). The effects of 5 mM leucinol on the glutamate content and BCDC activity of aortic segments were explored. In five experiments (15 rats), values of the glutamate contents of control versus leucinol-treated segments, tested as described in the preceding paragraph, were 1.17 ± 0.24 versus 0.57 ± 0.17 µmol/g (P < 0.01). The effects of 5 mM leucinol on BCDC activity were examined by incubating aortic segments with either [1-¹⁴C]leucine or -keto-[1-¹⁴C]isocaproate and estimating the ¹⁴CO₂ evolved. Leucinol decreased the ¹⁴CO₂ evolved from both [1-¹⁴C]leucine (5 experiments; control and leucinol values, respectively, 155 ± 31 and 8.6 ± 1.0 nmol/g, P < 0.01) and -keto-[1-¹⁴C]isocaproate (5 experiments; control and leucinol values, respectively, 163 ± 24 and 12.1 ± 2.1 nmol/g, P < 0.005). Inhibitory effects of leucinol could result from blocking the uptake or transamination of leucine, e.g., by blocking the mitochondrial branched chain aminotransferase, which functions as a branched-chain -ketoacid transporter (20). Direct binding to the BCDC protein is unlikely because leucinol lacks a keto group.

Other tissues. Prior results show that the LeuGlu pathway is present mainly in four tissues: the aorta, pancreas, testis, and lung (41). Inasmuch as the present evidence (Figs. 2 and 3) indicates that the differentiation of aortic segments for LeuGlu activity is correlated with leucine uptake and BCDC activity, it was of interest to examine these activities in tissues differentiated or not for the LeuGlu pathway. Segments of 11 rat tissues were incubated with either [U-¹⁴C]leucine, [1-¹⁴C]leucine, or -keto-[1-¹⁴C]isocaproate and the formation of ¹⁴CO₂ and [¹⁴C]glutamate quantified. The results listed in Table 2 confirm a high correlation ( = 0.97) between [¹⁴C]glutamate formation and ¹⁴CO₂ evolution from [U-¹⁴C]leucine. The four tissues with an active LeuGlu pathway all exhibit relatively high leucine uptake and BCDC activity assessed by ¹⁴CO₂ evolved from -ketoisocaproate. The remaining tissues have little or no LeuGlu activity, owing to a number of patterns. The liver, kidney, and heart exhibit much reduced leucine uptake and metabolism to CO₂, although the BCDC activity is high. The jejunum, and less markedly the thymus, have relatively little BCDC activity but take up leucine well. Skeletal muscle and cerebral cortex exhibit little uptake or BCDC activity.

View this table: [in this window] [in a new window]   Table 2.   Uptake and metabolism of 14C-labeled leucine and -ketoisocaproate by various rat tissues

Contractile responses: effects of L-leucinol and L-leucinamide. As indicated above, L-leucinol is an effective inhibitor of the LeuGlu pathway. The effects of leucinol (0.2-5.0 mM) on the contractile responses of segments precontracted with phenylephrine or 5-HT are shown in Figs. 5 and 6. Figure 5, A and B, illustrates prominent increases in tension owing to treatment with 5.0 mM leucinol of individual segments precontracted, respectively, with 10⁷ M phenylephrine or 5 µM 5-HT. Figure 5, A' and B', shows the marked attenuation of these responses in corresponding, adjacent segments denuded of endothelium and treated similarly. The peak increments in tension owing to leucinol were proportional to the concentration tested in the range of 0.2-5.0 mM, as shown in Fig. 6, and the values were similar for segments precontracted with either 10⁷ M phenylephrine (10 rats, 24 trials, F = 43.6, P < 0.005) or 5 µM 5-HT (8 rats, 12 trials, F = 23.6, P < 0.005). The control, mean tension value before addition of 5 mM leucinol, 312 ± 22 g/g wet wt, was increased by ~264 g/g wet wt, or 84.6%, by leucinol treatment. When four additional segments (2 rats) were denuded of endothelium and tested similarly, 5 mM leucinol elicited a much smaller increment, 50 ± 24 g/g, significantly less than the increment observed with intact segments (t = 4.66, P < 0.001).

View larger version (25K): [in this window] [in a new window]   Fig. 5.   Contractile responses of individual intact and endothelium-denuded rat aortic segments to inhibitors of the leucine-to-glutamate pathway. Each inhibitor was tested with an intact (A-F) and an adjacent denuded (A'-F') segment. Segments were precontracted at the time point indicated by arrow 1 and test inhibitor added at arrow 2. A,A': precontracted with 107 M phenylephrine and tested with 5 mM L-leucinol; B,B': precontracted with 5 µM 5-hydroxytryptamine (5-HT) and tested with 5 mM L-leucinol; C,C': precontracted with 10 µM 5-HT and tested with 10 mM L-leucinamide; D,D': precontracted with 107 M phenylephrine and tested with 10 mM -ketoisovalerate; E,E': precontracted with 106 M phenylephrine and tested with 2 mM L-valine; F,F': precontracted with 106 M phenylephrine and tested with 20 mM D-aspartate.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   Increments in tension owing to treatment of precontracted rat aortic segments with various concentrations of L-leucinol. Values are means ± SE for segments precontracted with 107 M phenylephrine (trace A) or 5 µM 5-HT (trace B). Mean tension of the precontracted segments before additon of leucinol was 312 ± 22 g/g.

7

5

Contractile responses: effects of -ketoisovalerate and L-valine. The catabolism of L-valine is initiated by its conversion to -ketoisovalerate, which is then oxidatively decarboxylated by the same BCDC, which acts on -ketoisocaproate derived from leucine (36). Inasmuch as the valine and -ketoisovalerate metabolic pathway does not yield acetyl CoA or glutamate, these compounds will competitively inhibit the functions of the LeuGlu pathway by preempting the BCDC. Accordingly, their effects on aortic contractile responses were examined in segments precontracted with phenylephrine, and the results are listed in Table 3 and illustrated for individual segments in Fig. 5, D,D' and E,E'. In tests of 17 segments (10 rats) precontracted with 10⁷ M phenylephrine, treatment with 10 mM -ketoisovalerate increased the tension values by 101 ± 12 g/g (P < 0.001), or ~42% on average, and similar increments were observed in 9 additional trials (4 rats) of segments precontracted with 10⁶ M phenylephrine (P < 0.001). By contrast, five segments (3 rats) denuded of endothelium showed no significant change in tension on treatment with 10 mM -ketoisovalerate (P < 0.001 for difference from intact segments), although they contracted normally on exposure to 10⁷ M phenylephrine (Table 3). In tests of 16 segments (6 rats) precontracted with 10⁶ M phenylephrine, treatment with 2 mM L-valine increased the tension values by 45 ± 6 g/g (P < 0.001), or ~16% on average, and significant increments (P < 0.001) were also observed in eight further trials (4 rats) after precontraction with 10⁷ M phenylephrine. Again, denuding the endothelium of three segments (2 rats) eliminated the effect of valine (P < 0.001 for difference from intact segments), although the contractile responses to phenylephrine were intact (Table 3).

View this table: [in this window] [in a new window]   Table 3.   Effects of -ketoisovalerate and L-valine on the tension of intact or endothelium-denuded aortic segments precontracted with phenylephrine

View larger version (19K): [in this window] [in a new window]   Fig. 7.   Increments in tension of successive aortic segments precontracted with phenylephrine and treated with 10 mM -ketoisovalerate (13 rats, 36 segments) or 2 mM L-valine (9 rats, 29 segments). Values are means ± SE. Segments 7-mm long were precontracted and tested as described in Table 3. Values for tension increments were plotted at the midpoint of each segment tested. The aortic arch, 0-12 mm from the aortic ring, was not tested.

6

5

7

7

Contractile responses: effects of D-aspartate. Glutamate formed via the endothelial LeuGlu pathway is transported to the underlying smooth muscle coats. Interruption of the LeuGlu pathway by D-aspartate, a relatively specific and metabolically inert inhibitor of Na⁺-dependent glutamate transporters (17, 28, 42), also enhanced contractile responses. Fifty-two segments (9 rats) precontracted with 10⁶ M phenylephrine were tested with various concentrations of D-aspartate, from 2-20 mM. Figure 5, F and 5F' illustrates responses of an intact and an endothelium-denuded segment. Figure 8 shows that increments in tension were observed throughout the concentration range tested (F = 11.36, P < 0.005) and increased from 44 ± 6 g/g (~15.6%) at 2.0 mM to 131 ± 11 g/g (~46.5%) at 20 mM. In five segments denuded of endothelium 20 mM D-aspartate treatment increased the tension by only 21 ± 2.9 g/g (Fig. 8), a decrease of ~84% compared with the intact segments (P < 0.001). Treatment of two segments with 2 mM L-NAME, followed by 20 mM D-aspartate, yielded tension increments of 131 ± 2 g/g, similar to the control values. Precontraction of six segments with 10 µM 5HT, followed by 2-20 mM D-aspartate, yielded tension increments comparable to the foregoing values.

View larger version (18K): [in this window] [in a new window]   Fig. 8.   Effects of various concentrations of D-aspartate on the increase in tension of precontracted rat aortic segments. Values are means ± SE for 52 trials (36 segments, 9 rats). Segments were precontracted with 106 M phenylephrine, and the mean tension before addition of D-aspartate was 282 ± 21 g/g. Corresponding increments in tension for 5 segments (2 rats) denuded of endothelium and tested similarly with 20 mM D-aspartate are shown; mean tension before D-aspartate was 337 ± 40 g/g.

6

Contractile responses: effects of external L-glutamate. Because inhibition of the LeuGlu pathway enhances contractile responses, glutamate added externally is expected to induce relaxation responses, providing it permeates the LeuGlu compartment. The permeation, however, is poor (41), and addition of 0.2-5.0 mM L-glutamate to aortic rings precontracted with 10⁶ M phenylephrine yielded no consistent change in tension. Raising the glutamate concentration to 10-20 mM, however, produced a biphasic response. In tests of 11 segments (4 rats) so treated versus 7 untreated adjacent segments, external glutamate elicited a rapid increase in tension (from initial values of 411 ± 28 to 506 ± 17 g/g), which persisted for 39 ± 4 min and was followed by a prolonged relaxation. The glutamate-treated segments relaxed at a rate of 2.30 ± 0.28 g · g¹ · min¹ compared with a slower rate in the untreated segments, 0.71 ± 0.12 g · g¹ · min¹ (P < 0.001). The relaxation responses to glutamate were similar and significant whether the aortic segments were precontracted with 10⁶ M phenylephrine alone (4 trials, P < 0.01), or with phenylephrine, followed by either 20 mM D-aspartate (3 trials, P < 0.025) or 20 mM Na⁺ cysteate (4 trials, P < 0.0125). Denuding the endothelium had no significant effect on the responses to glutamate.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The results above demonstrate that the aortic LeuGlu pathway proceeds via the catabolism of leucine to acetyl CoA and acetoacetate, a metabolic route established by studies of other tissues (12, 34, 47). In the rat aorta, this pathway comprises uptake of L-leucine at the luminal membrane of the endothelial cell by a Na⁺-dependent mechanism (Fig. 4); transamination to -ketoisocaproate, either in the cytosol or after transport across the mitochondrial membrane (20); oxidative decarboxylation to isovaleryl CoA by the BCDC of the mitochondrial matrix; successive dehydrogenation, carboxylation and hydration reactions to yield HMG CoA; and cleavage of HMG CoA to acetyl CoA and acetoacetate. In the presence of oxaloacetate, derived in our experiments from glucose metabolism, the acetyl CoA can be used to synthesize -ketoglutarate via tricarboxylic acid cycle reactions and this product converted to glutamate by transamination or a glutamate dehydrogenase reaction. As a consequence of the catabolic pathway [1-¹⁴C]leucine does not yield labeled glutamate, owing to the loss of carbon 1 in the BCDC reaction, and the relative efficiency of formation of [¹⁴C]glutamate from either [U-¹⁴C]leucine, [U-¹⁴C]- isoleucine, or [U-¹⁴C]valine (Fig. 1), respectively, corresponds to the number of labeled carbon atoms in each amino acid available for acetyl CoA formation, i.e., 5:2:0.

The efficiency of conversion of leucine to glutamate, the stoichiometry of the process and the controlling role of the BCDC reaction were characterized by quantification of the ¹⁴CO₂ evolved from the metabolism of [1-¹⁴C]leucine, -keto-[1-¹⁴C]isocaproate, and [U-¹⁴C]- leucine. As shown in Figs. 2 and 3, leucine metabolism to CO₂ occurs primarily in the endothelium and follows the regional differentiation pattern characteristic of the LeuGlu pathway. In the segments of peak activity ~25-40% of the leucine taken up from the ambient medium was oxidatively decarboxylated, and for each mole so metabolized there resulted 1 mol of glutamate plus 4 mol of CO₂ (Table 1). Aortic segments, therefore, are relatively efficient at utilizing acetyl CoA derived from leucine for glutamate formation. One mechanism accounting for the efficiency is the relatively direct utilization for glutamate synthesis of the -keto-[¹⁴C]glu- tarate formed from [U-¹⁴C]leucine-derived acetyl CoA. Such -keto-[¹⁴C]glutarate would be labeled at carbons 4 and 5 (34), as would the [¹⁴C]glutamate formed directly from it by transamination or the glutamate dehydrogenase reaction. If the -keto-[¹⁴C]glutarate entered the tricarboxylic acid cycle, however, the ¹⁴C would be randomized in the succinate and fumarate intermediates and label carbons 1-4 of -ketoglutarate and glutamate derived from it. The extent of such randomization can be assessed by determining the ¹⁴C content of carbon-1 in glutamate. In a prior study (41) [¹⁴C]glutamate formed via the LeuGlu pathway was treated with L-glutamate decarboxylase and the carbon 1 evolved as CO₂ contained insignificant ¹⁴C.² Thus the -ketoglutarate intermediate was converted to glutamate relatively directly. The mechanisms responsible are unknown but may involve compartmentalization in mitochondria or other organelles.

The LeuGlu, NO, and PGH₂ signal pathways share a similar regional pattern of activity in the rat aorta (1, 40, 41), suggesting that the windkessel area, site of the maximal activity, is functionally specialized for regulating contractility of the underlying smooth muscle. Hence we sought a controlling reaction in the leucine catabolic sequence whose activity would correspond to the LeuGlu regional pattern. We focused on the BCDC reaction, which controls leucine catabolic flux in a number of other tissues. As shown in Fig. 2B, endothelium-dependent oxidative decarboxylation of -keto-[1-¹⁴C]isocaproate by the BCDC in successive aortic segments clearly exhibits the regional pattern. Furthermore, we calculated the extent to which the differences in endothelium-dependent BCDC activity between the five successive segments shown in Fig. 2B could account for the corresponding segment-dependent differences in overall LeuGlu activity, i.e., glutamate formation (Fig. 3). Let a represent values of the differences between the mean, endothelium-dependent BCDC activities of the successive aortic segments, estimated as the ¹⁴CO₂ produced from -keto-[1-¹⁴C]isocaproate and expressed relative to the peak ¹⁴CO₂ producton (Fig. 2B). Let b represent the corresponding values for the differences between the mean, endothelium-dependent LeuGlu activities of the successive aortic segments, estimated as the glutamate produced and expressed relative to the peak glutamate production (Fig. 3). The ratio b/a then expresses the dependence of the segmental differences in overall LeuGlu activity on the differences in BCDC activity, and a value of 1.0 represents complete dependence (22). The observed b/a values for four segmental differences (means ± SE) are 1.01 ± 0.06, in accord with complete dependence on the BCDC activity.

In contrast with the close correspondence of the LeuGlu and BCDC regional patterns of activity, no such pattern was observed for leucine--ketoglutarate or alanine--ketoglutarate transaminase activities in aortic homogenates of successive segments. Additon of -ketoglutarate to the incubation medium composed of segments incubated with [1-¹⁴C]leucine yielded no increase in ¹⁴CO₂ production and no change in the regional pattern shown in Fig. 2A, consistent with the conclusion that leucine--ketoglutarate net transamination to yield -ketoisocaproate is not rate limiting in these segments. Two distal reactions of the leucine catabolic pathway also showed no regional segmental differences. Incubation of successive aortic segments with 1.7 µM [9,10-³H]oleate as a source of acetyl CoA in place of leucine yielded [³H]glutamate averaging 2.05 ± 0.26 nmol/g wet weight with no significant segmental differences. Incubation with -keto-[1-¹⁴C]glutarate as precursor yielded [¹⁴C]glutamate averaging 39 ± 2 nmol/g and no significant segmental differences. The evidence, therefore, favors the BCDC reaction as the determinant of the regional activity pattern of the LeuGlu pathway. Specific factors that influence this reaction and could determine the regional pattern include the amount of BCDC enzyme protein, the degree of phosphorylation determining the activity state of the enzyme (19), the availability of the -ketoisocaproate substrate owing to net transamination of leucine, the transport of -ketoisocaproate across the mitochondrial membrane via the branched chain -ketoacid transporter (20), and the concentration of cofactors required for the oxidative decarboxylation.

Estimations of the total free glutamate content of aortic segments incubated in vitro indicate that the endothelial LeuGlu pathway is a major determinant of that content. Thus the segmental values exhibit the characteristic regional pattern of the pathway, and in the peak segment (~20-40 mm from the aortic valve ring) denuding the endothelium before incubation decreased the content by ~72%. Moreover, incubation with the LeuGlu pathway inhibitor L-leucinol decreased the value by ~51%. When the segments were denuded of endothelium after incubation, the underlying coats retained the glutamate. Similarly, denuding the endothelium before incubation markedly reduced the formation of [¹⁴C]glutamate from [U-¹⁴C]leucine (41), whereas denuding after incubation did not alter the total [¹⁴C]glutamate content of the segment. These observations indicate that the final steps of the LeuGlu pathway involve efficient transport to the underlying coats. The transfer mechanism(s) responsible has not been characterized, but the transport route is highly compartmentalized and bounded by structures relatively impermeant to glutamate, inasmuch as no [¹⁴C]glutamate was detected in the ambient media after incubation of segments with [U-¹⁴C]leucine (41).

Leucine catabolism has been studied extensively in nonaortic tissues in relation to energy metabolism (12, 14), protein synthesis and degradation (5, 37), and familial disorders such as maple syrup urine disease (11, 27) and isovaleric acidemia (48), which result from genetic defects in the BCDC and isovaleric dehydrogenase, respectively. Normal diets provide the human and the rat with an excess of the branched-chain amino acids, considerably more than is required for new protein synthesis in both growing and nongrowing individuals (14). The leucine catabolic pathway therefore functions in nonaortic tissues as an essential disposal route for potentially toxic branched-chain intermediates, simultaneously providing calories for immediate use or storage as adipose tissue (14). Several variables, including starvation (13, 19), dietary protein level (19), diabetes mellitus (3), octanoate infusion (4), and insulin and growth hormone (14) modulate flux through the pathway, with the direction of change dependent on the particular tissue. Generally, the controlling enzyme in these responses is the BCDC, whose activity is decreased by covalent phosphorylation and increased by dephosphorylation (12, 19). In rat aortic endothelium the leucine catabolic pathway is utilized to supply acetyl CoA for glutamate synthesis and transfer to the underlying smooth muscle.

We tested the hypothesis that the LeuGlu pathway, along with the NO and PGH₂ pathways, functions to regulate the contractility of the aortic smooth muscle. Because leucinol and leucinamide interrupt the pathway and decrease the cGMP content of the smooth muscle, we predicted they would enhance aortic contractile responses, and the results above confirm the prediction. The enhanced responses require an intact endothelium and do not occur via reductions in NO formation or action. The specificity of action of these compounds is an important consideration because millimolar concentrations are required. Consequently we also tested two additional classes of LeuGlu pathway inhibitors, compounds that are chemically different from the leucine analogs and from each other. -Ketoisovalerate and its precursor L-valine were chosen to preempt competitively the endothelial BCDC. These compounds, too, act on the endothelium to enhance contractile responses and the effects exhibit a regional aortic pattern (Fig. 7) similar to that of the BCDC activity (Fig. 2). Specificity for the interruption of the LeuGlu pathway at the BCDC site is also indicated by the observations that -ketoisocaproate, the BCDC substrate derived from leucine, does not enhance contractile responses but does block the inhibition owing to -ketoisovalerate. -Ketoisovalerate produced contractile responses in the presence of L-NAME and did not interfere with acetylcholine-induced relaxations. Hence the effects of -ketoisovalerate are not mediated by blocking NO formation or action.

Interrupting the LeuGlu pathway by treatment with D-aspartate, a metabolically inert compound that competes with L-glutamate for transfer mediated by certain glutamate transporters, also causes enhanced contractile responses. These effects require an intact endothelium and do not result from reductions in NO formation. The results point to the involvement of glutamate transporters in the terminal steps of the pathway in which glutamate is transferred efficiently to the underlying coats without significant exchange with the extracellular medium. The mechanisms, membrane transporters, and routes responsible for this compartmentalized transport are unknown at present, and the precise locus of action of D-aspartate in the endothelium remains to be determined. In studies of nonaortic tissues, transport mechanisms with specificity for both L-glutamate and D-aspartate have been found in mitochondrial (29) and plasma membranes (10). The X family of glutamate/D-aspartate transporters (10), which mediate Na⁺-dependent transport across plasma membranes, have been particularly well characterized and a number cloned (43, 46). These function prominently, for example, in glial and neuronal cells of the central nervous system where efficient cellular uptake of glutamate is essential to maintain low extracellular concentrations and prevent indiscriminate signaling and nerve damage (17). It is reasonable to suggest that the compartmentalized transport of glutamate synthesized in the aortic LeuGlu pathway is also necessary for discrete signaling, particularly because free glutamate is relatively abundant in blood and many tissues.

Relatively high concentrations, 10-20 mM, of externally added glutamate elicited a rapid-onset increase of ~23% in tension followed, after a delay of ~39 min, by a prolonged relaxation response. The high concentrations required and the slow onset of the relaxation responses could result from the relative impermeability to glutamate of the LeuGlu pathway. In prior studies glutamate treatment (0.5-2.0 mM) of rat aortic segments did increase cGMP levels but only when the segments were denuded of endothelium and treated with 3-isobutyl-1-methylxanthine to inhibit cyclic nucleotide phosphodiesterase (41). Precontracted rabbit aortic rings were reported to exhibit relaxation responses on treatment with 100 mM glutamate or 1-10 mM glutamine (33).

Systematic studies (1, 40, 41) of sequential rat aortic segments in vitro provide comprehensive evidence of regional differentiation of the aortic endothelium. Figure 9 illustrates the patterns observed for six activities whose values peak in the windkessel area and decrease distally: constitutive formation of NO from L-arginine (1), cGMP content (1), and increase in cGMP owing to inhibition of the cyclooxygenase arm of the eicosanoid pathway and PGH₂ formation (40), endothelial uptake of [U-¹⁴C]leucine and its conversion to [¹⁴C]glutamate (41), free glutamate content, and BCDC activity (Fig. 2). The increase in values from the aortic arch (first ~12 mm from the valve ring) to the windkessel area are noteworthy in view of the ultrastructural studies utilizing morphometric methods reported by Kao et al. (23). They observed that rat endothelial cells in the aortic arch are longer and thinner and contain fewer intracytoplasmic vesicles and fewer complex interdigitating intercellular contacts than those in the windkessel area.

View larger version (36K): [in this window] [in a new window]   Fig. 9.   Patterns of regional differentiation in rat aorta for six activities dependent on the endothelium. For comparison, data are plotted as percentages of the maximal value observed for each activity. Trace A, formation of [14C]glutamate from [14C]leucine (41); trace B, tissue content of free glutamate; trace C, tissue cGMP levels (1); trace D, BCDC enzyme activity assayed by the liberation of 14CO