the purine nucleotide cycle -...

The Purine Nucleotide Cycle

A PATHWAYFORAMMONIAPRODUCTIONIN THE

RAT KIDNEY

RONALDT. BOGUSKY,LEAHM. LowENSTITN, and JoHN M. LOWENSTEIN

From the Department of Medicine, Boston University School of Medicine,Boston, Massachusetts 02118, and the Graduate Department of Biochemistry,Brandeis University, Waltham, Massachusetts 02154

A B S T R A C T Particle-free extracts prepared from kid-ney cortex of rat catalyze the formation of ammonia viathe purine nucleotide cycle. The cycle generates am-monia and fumarate from aspartate, using catalyticamounts of inosine monophosphate, adenylosuccinate,and adenosine monophosphate. The specific activitiesof the enzymes of the cycle are 1.27+0.27 nmol/mgprotein per min (SE) for adenylosuccinate synthetase,1.38±0.16 for adenylosuccinase, and 44.0±3.3 for AMPdeaminase.

Compared with controls, extracts prepared from kid-neys of rats fed ammonium chloride for 2 days showa 60% increase in adenylosuccinate synthetase and athreefold increase in adenylosuccinase activity, and agreater and more rapid synthesis of ammonia and adeninenucleotide from aspartate and inosine monophosphate.Extracts prepared from kidneys of rats fed a potassium-deficient diet show a twofold increase in adenylosuccinatesynthetase and a threefold increase in adenylosuccinaseactivity. In such extracts the rate of synthesis of am-monia and adenine nucleotide from aspartate and inosinemonophosphate is also increased.

These results show that the reactions of the purinenucleotide cycle are present and can operate in extractsof kidney cortex. The operational capacity of the cycleis accelerated by ammonium chloride feeding and potas-sium depletion, conditions known to increase renalammonia excretion.

Extracts of kidney cortex convert inosine monophos-phate to uric acid. This is prevented by addition ofallopurinol or 1-pyrophosphoryl ribose 5-phosphate tothe reaction mix,ture.

Received for publication 17 June 1974 and in rezised formn24 February 1976.

INTRODUCTIONThe kidney plays a major role in the maintenance ofacid-base balance, largely by the production of ammoniawhich diffuses into the tubular fluid where it combineswith hydrogen ions (1). The resulting ammonium ionsare nondiffusible and are excreted in the urine. Part ofthe renal ammonia is derived from the amide groupof glutamine via the glutaminase reaction (2, 3). Theremainder is currently believed to be derived largelyfrom a-amino groups of amino acids by transaminationto a-ketoglutarate (1), followed by oxidative deamina-tion of the resulting glutamate via the glutamate dehy-drogenase reaction (4). We report below that extractsof rat kidney cortex can catalyze the formation ofammonia via the purine nucleotide cycle, a sequence ofreactions that results in the production of ammonia andfumarate from aspartate (5) :'

AMP+ H20- > IMP + NH3 (a)

IMP + GTP+ aspartate >

adenylosuccinate + GDP+ Pi

Adenylosuccinate -> AMP+ fumarate.

(b)

(c)

The enzymes that catalyze the reactions of the cycle areadenylate deaminase (a), adenylosuccinate synthetase(b), and adenylosuccinase (c). The net reaction for oneturn of the cycle is:

Aspartate + GTP+ H20 *fumarate + GDP+ Pi + NH3. (d)

'Abbreviations used in this paper: GDP, guanosine di-phosphate; GMP, guanosine monophosphate; GTP, guano-sine triphosphate; IMP, inosine monophosphate; SAMP,adenylosuccinate.

The Journal of Clinical Investigation Volume 58 August 1976 -326335326

Activities of enzymes a, b, and c are repor,ted for kidneyextracts prepared from control rats and from rats sub-jected to ammonium-chloride feeding and potassium de-pletion, conditions known to increase renal ammoniaproduction (6).

METHODS

Approximately 4 g of rat kidney cortex was homogenizedin 3 vol by weight of 60 mMpotassium phosphate buffer,pH 7.7, 2.0 mMEDTA, and 1.0 mMdithiothreitol at 4°Cusing a motor-driven glass homogenizer. The homogenatewas allowed to stand on ice for 30 min and was then cen-trifuged at 31,000 g for 10 min. The supernate so obtainedwas centrifuged at 105,000 g for 45 min. The resulting high--speed supernate was subjected to gel filtration to removeendogenous substrates. 5 ml of the high-speed supernatewas placed on a column of Sephadex G-25 (coarse grade)with a gel bed of 1.5 X 15 cm, which had been equilibratedwith a mixture containing 30 mM potassium phosphatebuffer, pH 7.2, 200 mMKCl, and 1.0 mMdithiothreitol.The column was eluted with the same mixture. The first15 ml to be eluted contained no protein and was discarded.1-ml fractions were collected, and the three fractions con-taining the highest protein concentrations were pooled. Theresulting protein extract contained about 15 mg protein perml.

Unless otherwise indicated, reaction mixtures designedto demonstrate the operation of the purine nucleotide cyclecontained 0.1 ml protein extract, 50 mMKCI, 0.16 mMdithiothreitol, 0.25 mg creatine phosphokinase, 16 mMcreatine phosphate, 0.3 mMGTP, 0.5 mMIMP, 4.0 mMasparate, 5 mMpotassium phosphate buffer, pH 7.2, 27 mMimidazole-HCl buffer, pH 7.2, 8.3 mMMgCl, and 15 jugyeast hexokinase (sp act 50 U per mg) in a 0.6-ml volume.The reaction mixture was placed in a cuvette with a 1-cmlight path into which was inserted a 0.9-cm wide quartzslab that reduced the light path to 0.1 cm. The referencecuvette contained the same substances but lacked asparate.Both cuvettes were held at 370C in a water-jacketed cuvetteholder. The conversion of IMP to adenine nucleotides wasstarted by adding the protein extract. The conversion ofAMP-to IMP was initiated by adding 2-deoxyglucose in anamount equal to that of the creatine phosphate added origi-nally. The reactions were followed by repeated spectralscanning from 310 to 250 nm using a Coleman 124 split-beam spectrophotometer (Coleman Instruments Div., Oak-brook, Ill.). The recorder was zeroed at 310 nm after eachscan. Fig. 1 shows tracings of a typical set of spectralscans obtained after starting the amination sequence byadding the protein extract. After amination had ceased,deamination was initiated by adding 2-deoxyglucose. Theamounts of IMP, SAMP, and adenine nucleotide presentduring each spectral scan were calculated as described pre-viously (7) .2

2 Creatine phosphate and creatine phosphokinase are addedbecause they serve as a GTP-regenerating system. Adenylo-succinate synthetase is strongly inhibited by one of its endproducts, namely GDP. The inhibition is so strong thatlittle or no reaction occurs unless the regenerating system isadded. (No reaction may occur if the GTP that is added iscontaminated with GDP.)

The calculation of the concentration of -the three nucleo-tides is based on the absorbancies at 282, 270, and 262.5 nm.

Time(min)

0.3

0.2

0.1

oO

D 0.340

0.2

0.1

0 L.1310 282 270 262 250

X (nm)

FIGURE 1 Operation of the purine nucleotide cycle in ex-tracts of renal cortex. For details of the reaction mixturesee Methods. The amination sequence was initiated by add-ing the protein extract; the spectral scans of this sequenceare shown in the bottom panel. The deamination sequencewas initiated by adding 2-deoxyglucose; the spectral scansof this sequence are shown in the top panel. These resultswere converted into adenylosuccinate and adenine nucleo-tide formed. (The scans are the actual results obtained dur-ing the experiment shown as solid lines in Fig. 5.)

Adenylate deaminase was measured by following the de-crease in absorbance at 262.5 nm using a light path of 1 mm.

For a 1 mmlight pathAAdenylosuccinate = AA22/1.0AAdenine nucleotide = (AA=.5 - 1.14 AA282)/0.88AAdenine nucleotide = (AA27,- 1.5 AA28,)/0.60.

At 282 nm, IMP and adenine nucleotide (AMP + ADP+ATP) absorb equally, and any change in absorbance at 282nm is due only to changes in adenylosuccinate. At 262.5 and270 nm, the contribution of adenylosuccinate is first sub-tracted out by using an appropriate multiple of AA2A8. Theremainder represents the interconiversion of IMP and ade-nine nucleotide, and is divided by the respective differenceextinction coefficient to give the change in adenine nucleo-tide. The maximum difference in absorbance between IMPand AMPoccurs at 262.5 nm, and a smaller correction foradenylosuccinate has to be applied at 262.5 than at 270 nm.The second equation is therefore more precise than thethird, which is simply used as a check on the second toensure that there is no distortion of the spectra. Hypoxan-thine and uric acid formed were measured separately andwere subtracted from the purine nucleotide cycle pool be-fore making the calculations. For further details, such asextinction coefficients of the members of the purine nu-cleotide cycle, see reference 7.

Ammonia Production in the Rat Kidney 327

0.2

iE

.- 0.10

z

0

FIGURE 2 Opertracts of rat kicextraction mediuhomogenized inand 1.0 mM dihomo,genized inmMEDTA. TIsupernate was sMethods. The rimMGTP, 4.0 ntine phosphate, Cphate buffer, pHand 15 ,g crysstarted by addirwas run in a fi(equivalent to 1the arrow. Thefollowed spectroThe reference cinucleotides; 0EDTA; solid Ibuffer.)

In assay A thezole-HCl buffermM potassium

MgC12, 5 mMpotassium phosphate buffer, pH 7.2, 0.16 mM2-Deoxyglucose dithiothreitol, 0.25 mg creatine phosphokinase, 16 mMcrea-

+ tine phosphate, 0.3 mM GTP, 0.5 mM IMP, 4.0 mMaspartate, and 0.1 ml of protein extract in a final 0.6-mlvolume. The temperature was 37°C.

Adenylosuccinase was measured by following the changein absorbance at 282 nm using a light path of 1 mm. The

.rS __-< \ -complete reaction mixture contained 27 mMimidazole-HCl/4 nz \ \buffer, pH 7.2, 50 mMKCl, 8.3 mMMgCl2, 5 mMpotas-

-' 4 \sium phosphate buffer, pH 7.2, 0.16 mMdithiothreitol, 0.5/ \ mMSAMP, and 0.1 ml of kidney extract in a final 0.6-ml

/ 6 \ -volume. The temperature was 37°C.Protein was determined by the method of Lowry et al.

(8), uric acid by a modification of the method of Brown(9), hypoxanthine by the method of Kalckar (10), ammoniaby the method of Seligson and Seligson (11), lactate by

_.o the method of Hohorst (12), and fumarate and malate byo0- \ the method of Williamson and Corkey (13), except that the

+- change in absorbance was measured spectrophotometrically1 2 3 4 5 instead of fluorometrically.

Time (h) Animals were acid-loaded by administration of 20 mmolammonium chloride per kg body weight by gavage for 2

ation of the purine nucleotide cycle in ex- days. During this time the animals had access to food andlney cortex and the effect of EDTA in the water ad lib. 6-h urine samples were collected before andim. The left renal cortices of two rats were during the acid-loading period. Urinary ammonia (expressed60 mMpotassium phosphate buffer, pH 7.7, as micromoles ammonia per milligram creatinine) rose fromithiothreitol; the right renal cortices were a control level of 14.1+2.4 (SEM) (ni = 11) to 184+18 and

the same buffer which also contained 2.0 270+23 on the 1st and 2nd day of acid loading, respectivelyhe homogenates were centrifuged, and each (P <0.001). Other controls showed that an equivalentubjected to gel filtration as described under amount of ammonium carbonate produced 20.5±4.5 Amoleaction mixture contained 0.5 mMIMP, 0.3 ammonia per mg creatinine on the 1st and 2nd day ofnM aspartate, 8.3 mMMgCl2, 16 mMcrea- administration (P > 0.1).).25 mg creatine phosphokinase, 5 mMphos- Animals were made potassium-deficient by feeding them1 7.2, 27 mMimidazole-HCl buffer, pH 7.2, a diet low in potassium, <0.1%% potassium in the wholeitalline yeast hexokinase. The reaction was diet (General Biochemicals, Chagrin Falls, Ohio, catalogng the kidney extract, 1.5 mg protein, and no. 170550), and giving them a solution of 0.9% NaCl asnal 0.6-ml volume at 25°C. 2-Deoxyglucose drinking water for 10 days. In addition each rat received6 mM) was added at the point indicated by 1 mg deoxycorticosterone acetate daily by subcutaneous in-operation of the purine nucleotide cycle was jection. Under these conditions serum potassium concen-iphotometrically as described under Methods. trations fell from control values of 5.26+0.19 (SEM) (n=uvette lacked aspartate. (A and A, adenine 12) to 2.69+0.16 meq per liter after 10 days of depletionand *, adenylosuccinate; broken lines, no (P < 0.001).

lines, 2.0 mMEDTA added to extraction Male Sprague-Dawley rats (250 g) were obtained fromCharles River Breeding Laboratories (Wilmington, Mass.).Adenylosuccinate was prepared for us by Vera Schultz,

cBiochemistry Department, Brandeis University. Adenosinereaction mixture contained 27 mM imida- a,j3-methylene diphosphonate was obtained from Miles Lab-pH 7.2, 50 mMKCl, 0.5 mMAMP, 5 oratories, Inc. (Elkhart, Ind.). All other biochemicals were

phosphate buffer, pH 7.2, and 0.1 ml of obtained from Sigma Chemical Corp. (St. Louis, Mo.).protein extract, in a 0.6-ml volume. l he reaction was runat 37°C. Assays were run in this manner because the en-zyme extract contained KCI and potassium phosphate inamounts which optimized the stability of the enzymes whichwere assayed. However, these conditions were not optimumwith respect to adenylate deaminase activity. Activities closeto maximum reaction velocity were obtained in assay B inwhich the reaction mixture contained 27 mMimidazole-HClbuffer, pH 7.2, 175 mMKCI, 2.0 mMAMP, and enzyme.To assay the enzyme under these conditions, the proteinextract was subjected to gel filtration under the conditionsdescribed above, except that phosphate buffer was omitted.A comparison of assays A and B showed a ratio of activities(B/A) of 2.95.

Adenylosuccinate synthetase was measured by followingthe change in absorbance at 259 nm using a light path of1 mm. The complete reaction mixture contained 27 mMimidazole-HCl buffer, pH 7.2, 50 mM KCI, 8.3 mM

RESULTSActivities of purine nucleotide cycle enzymes. The

specific activities of adenylosuccinate synthetase andadenylosuccinase in the extracts prepared from renalcortex were 1.27+0.07 (SE) (n = 32) and 1.38+0.16(n = 30) nmol/mg protein per min, respectively. Thespecific activity of adenylate deaminase was 44.0±3.4nmlol/mg protein per min (n = 25) when EDTA was

omlitted from the extraction medium. When 2 mMEDTA was present during the extraction, the specificactivity of adenylate deaminase was 4.95±0.65 nmol(n = 14).

Effect of EDTA on operation of the purine nucleotide

328 R. T. Bogusky, L. M. Lowenstein, and J. M. Lowenstein

0.5

N ~~~~0.4N"

- 0.3

4 5 6 E

-0

FIGURE 3 Effect of 0.5 mMallopurinol on the operationof the purine nucleotide cycle. Kidney cortices of two ratswere homogenized in 60 mMpotassium phosphate buffer,pH 7.7, 2.0 mMEDTA, and 1.0 mMdithiothreitol. Thehomogenate was centrifuged and 5 ml of the resulting super-nate was subjected to gel filtration as described under Meth-ods. Another 5 ml of the supernate was also subjected togel filtration under the same conditions, except that theeluting buffer also contained 3.0 mMallopurinol. The re-action mixture was as described in the legend to Fig. 2for the extract prepared in the presence of 2 mMEDTA.The operation of the purine nucleotide cycle was followedspectrophotometrically as described under Methods. Sampleswere also taken from the reference and test cuvettes forthe determination of uric acid. (A and A, adenine nucleo-tides; 0 and 0, adenylosuccinate; O and +, uric acid intest cuvette; V and V, uric acid in reference cuvette; opensymbols, no allopurinol added to test and reference cuvettes;and solid symbols, 0.5 mMallopurinol added to reactionmixtures of both test and reference cuvettes.) Note thatno uric acid was formed when allopurinol was present.Solid, broken, and dotted lines are used to assist recognition.

cycle. Initial attempts to demonstrate the operation ofthe complete cycle in extracts of kidney cortex were

frustrated by a loss of both adenine and hypoxanthinenucleotides from the system. We therefore investigatedthe conditions required to keep these nucleotides intact.An experiment demonstrating the operation of thepurine nucleotide cycle is shown in Fig. 2. The cycle was

initiated by the addition of extract. More adenine nucleo-tide accumulated when the extract was prepared in thepresence of EDTA (solid lines) than in its absence(broken lines). This finding suggests that adenylate de-aminase activity is inhibited by EDTA, or that less ofthe enzyme is extracted in the presence of EDTA

0

z

< 0.2

0.1

than in its absence, or that EDTA reduces the loss ofIMP by side reactions.

Effect of allopurinol. During the incubation of thekidney extract with the reaction mixture, a portion ofthe IMP was converted to uric acid (Fig. 3, open sym-bols). Formation of uric acid was prevented by the addi-tion of 0.5 mMallopurinol to the reaction mixture (Fig.3, solid symbols). However, the presence of allopurinol

0 2 3 4 5 6

Time (h)

FIGURE 4 Effect of 1-pyrophosphoryl ribose 5-phosphateon the operation of the purine nucleotide cycle. Kidneycortices from two rats were homogenized in 60 mMpo-tassium phosphate buffer, pH 7.7, 2.0 mMEDTA, and 1.0mMdithiothreitol. The homogenate was centrifuged, andthe resulting supernate was subjected to gel filtration asdescribed under Methods, except that the eluting mixturealso contained 3.0 mMallopurinol. The reaction mixturewas as described in the legend to Fig. 2, except that it alsocontained 0.5 mMallopurinol. The spectrophotometric assayfor the purine nucleotide cycle was employed. In addition,samples of the reaction mixture were withdrawn at thetimes indicated and assayed for hypoxanthine. (A and A,

adenine nucleotides; 0 and 0, adenylosuccinate; O and *,

hypoxanthine in test cuvette; V and V, hypoxanthine inreference cuvette; open symbols, no 1-pyrophosphoryl ribose5-phosphate added to test and reference cuvettes; solidsymbols, 0.8 mM1-pyrophosphoryl ribose 5-phosphate addedto test and reference cuvettes.) Note that the amount ofhypoxanthine formed in the presence of 1-pyrophosphorylribose 5-phosphate was zero. Solid, broken, and dotted linesare used to assist recognition.


aE

a

0 1 2 3Time (h)

0.4

0.3

E

020

z

0.1

0 2 3 4 5 6Time (h)

FIGURE 5 Effect of ammonium chloride feeding on theoperation of the purine nucleotide cycle. A rat weighing 250g was fed 20 mmol ammonium chloride per kg body weightby gastric gavage daily for 2 days and was given drinkingwater ad lib. A sibling that did not receive NH4Cl servedas control. Both animals were fed with Purina laboratorychow (Ralston Purina Co., St. Louis, Mo.). On the 2ndday the animals were killed, and extracts of their kidneycortices were prepared as described under Methods. Thereaction mixture was as described in the legend to Fig. 2,except that both test and reference cuvettes contained 0.8mM1-pyrophosphoryl ribose 5-phosphate and 0.5 mMallo-purinol. The spectrophotometric assay for the purine nucleo-tide cycle was employed. (A and A, adenine nucleotides;O an(l *, adenylosuccinate; broken lines, control rat; andsolid lines, NH4Cl-fed rat.)

had relatively little effect on the rate of operation of thepurine nucleotide cycle or on the extent to which its inter-mediates accumulated. Both adenine and guanine nucleo-tides were completely converted to uric acid after thereaction mixture had been allowed to stand at roomtemperature for 24 h, even when allopurinol was presentin the reaction mixture (not shown in Fig. 3). In thepresence of 0.5 mMallopurinol, hypoxanthine accumu-lated in the sample and reference curvettes at differentrates (Fig. 4, dotted lines). This difference causes adistortion of the spectra.

Effect of 1-pyrophosplhoryl ribose S-phosphate. Inthe presence of 0.8 mM1-pyrophosphoryl ribose 5-phos-phate, no hypoxanthine accumulated over a 6-h period(Fig. 4, solid symbols). The presence of 1-pyrophos-phoryl ribose 5-phosphate resulted in an increased ac-cumulation of adenylosuccinate and a decreased accumu-lation of adenine nucleotide during the amination phase

of the experinment. One possible explanation for thisobservation is that 1-pyrophosphoryl ribose 5-phosphatesomewhat inhibits adenylosuccinase.

The loss of purine nucleotides to nucleosides andpurine bases was not inhibited by the addition of adeno-sine a,#-methylene diphosphonate, indicating that theloss was not initiated by a 5'-nucleotidase inhibitable bythis analogue (14).

Effect of ainmoniuml chloride feeding. Protein ex-tracts prepared from rats fed 20 mmol ammoniumchloride per kg body weight daily for 2 days showed agreater and more rapid accumulation of adenine nucleo-tide during tests for the operation of the purine nucleotidecycle than did untreated controls. Typical results ob-tained with paired rats are shown in Fig. 5. Otherexperiments showed that 3 days of ammonium chloridefeeding produced a similar effect. Measurements ofadenylosuccinate synthetase and adenylosuccinase showedthat the activities of the enzynmes were increased 1.6-and 2.6-fold, respectively, in eight paired experiments(Table I). Four rats fed an amount of ammoniumcarbonate equivalent to the ammonium chloride giventhe other rats did not show a change in adenylosuccinaseactivity.

Ammonium chloride feeding resulted in an increase inammonia production under cycling conditions (TableII). This increase was not observed in kidney extractsprepared from ammonium carbonate-fed rats.

Effect of potassium depletion. Operation of the purinenucleotide cycle in extracts prepared from kidney cortexof rats that had been kept on the potassium-deficient dietwas compared with the operation of the cycle in extractsprepared from normal rats. Results of a typical experi-ment obtained with paired rats are shown in Fig. 6.Hypokalemia caused an increase in the rate of accumu-

lation of adenine nucleotide and in the total amount

accumulated (AMP + ADP+ ATP). At the same timethere was a considerable reduction in the amount ofadenylosuccinate accumulated. Similar results were ob-tained in five other experiments. Adenyosuccinate syn-thetase and adenylosuccinase increased 1.8- and 2.7-foldin activity, respectively (Table I). Potassium depletionresulted in an increase in ammonia production undercycling conditions (Table II).

Ammonia production. Ammonia production did not

start until the deamination phase of the cycle was ini-tiated by adding 2-deoxyglucose. Six experiments, run

under the conditions described in the legend to Fig. 6,showed a production of 0.56±0.9 (SEM) nmol NHs/mgprotein per min.

As has already been stated, adenylosuccinate synthetaseis strongly *inhibited by one of its reaction products,namely GDP. This necessi.tates addition of creatinephosphate to regenerate GTP from GDPto demonstrate


TABLE I

Activities of Purine Nucleotide Cycle Enzymes in Kidney Cortex of Normal, Acidotic, and Hypokalemic Rats

Normal rats Paired rats Paired ratsEnzyme activity

measured n Normal Acidotic n P* Normal Hypokalemic n P*

Adenylate deaminase 14.94±1.1 25 11.44-2.5 8.9-42.2 6 0.1 17.9--3.8 6.2±40.7 7 0.02(44.0) (33.6) (26.3) (52.8) (18.3)

Adenylosuccinatesynthetase 1.27±-0.07 25 1.21--0.14 1.92±t0.21 8 0.02 1.14±t0.14 2.08±-0.39 8 0.02

Adenylosuccinase 1.38±t0.16 25 1.28--0.24 3.27±t0.51 8 0.005 1.02±t0.19 2.76±-0.53 13 0.001

Activities are expressed as nanomoles per milligram protein per minute, and are quoted as assayed at pH 7.2. The assay con-ditions are described in the Methods section. Adenylosuccinate synthetase and adenylosuccinase activities were determinedunder conditions yielding rates close to Vmax. Adenylate deaminase activities were determined under conditions yieldingVmax/2.95. The Vmax values are quoted in parentheses; they were obtained by multiplying the measured rates by 2.95.* Paired t test (30).

the operation of the complete cycle. Conversely, omissionof the GTP-regenerating system can be used to demon-strate that anmnmonia production by the extract dependson the operation of adenylosuccinate synthetase. TableII shows ammonia production during one complete turnof the purine nucleotide cycle. In the presence of theGTP-regenerating system, little or no ammonia pro-duction occurred during the amination phase of thecycle, which was observed by running spectral scans.However, there was a rapid release of ammonia dur-ing the deamination phase of the cycle, which was alsoobserved by running spectral scans. In controls whichlacked creatine phosphate, spectral scans showed thatneither the amination nor the deamination phase of thecycle occurred. Ammllonia analyses of the controlsshowed an accumulation of approximately 0.1 mMam-monia during the period which corresponded to theamination phase in the complete reaction mixture. Thecontrols showed little additional accumulation of am-monia during the period corresponding to the deamina-tion phase of the cycle in the complete reaction mixture.The simplest explanation for these observations is thatwhen creatine phosphate is omitted, the GDP formedprevents production of ammonia via the purine nucleo-tide cycle by inhibiting adenylosuccinate synthetase.However, under these conditions GDP is degradedfurther to GMPand guanosine, which is then deaminatedto xanthosine and ammonia. The correctness of thisinterpretation is supported by the observation that whenkidney extracts were incubated with 0.1 mMGTPalone,the amount of ammonia produced was about 0.1 mM, butthis ammonia production was prevented by addition ofcreatine phosphate (Table II).

The fate of aspartate carbon. The combined opera-*tion of adenylosuccinate synthetase and adenylosuccinaseresults in the conversion of aspartate to fumarate. Pre-vious studies with muscle extracts showed a stoichio-metric relation between adenine nucleotide formed and

malate plus funmarate formed during the reaminationphase of the purine nucleotide cycle (7). This straight-forward stoichiometry prevailed because there were noside reactions apart from the formation of malate fromfumarate. In the case of kidney extracts, additional re-actions occur which convert fumarate not only to malatebut also to lactate. Under the experimental conditionsdescribed in the legend to Fig. 6, for 1.0 mol adeninenucleotide formed from IMP there was formed 0.089mol fumarate, 0.25 mol malate, and 0.36 mol lactate(total = 0.70 mol). The 0.30 mol of asparate carbonnot accounted for presumably accumulated in other in-termediates. Lactate was probably formed from malatevia malate dehydrogenase, phosphoenolpyruvate carboxy-kinase, pyruvate kinase, and lactate dehydrogenase.NAD+ was not added to the reaction mixture, but suffi-

TABLE IIAmmonia Production by Extracts of Rat Kidney during

a Single Turn of the Purine Nucleotide Cycle

Ammonia production, mMGTP-

Rats used for regen- Amina- Deamina-preparing kidney erating tion tion

extract system phase phase

Controls + 0.01 0.21 0.20- 0.15 0.18 0.03

Acidotic + 0.04 0.48 0.44- 0.13 0.14 0.01

Hypokalemic + 0.01 0.38 0.37- 0.11 0.12 0.01

Controls, extract + 0.01 0.01 0incubated with - 0.09 0.12 0.03GTP in absence ofaspartate and IMP

The complete reaction mixture was as described under Methods, exceptthat the GTP concentration was 0.1 mM. The amination sequence was

initiated by adding the protein extract. Reaction mixtures marked with a

minus in the "GTP-regenerating system" column lacked creatinephosphate.


0.5 l quent deamidation of a-ketoglutaramate to a-ketoglutarateplus ammonia and the deamination of glutamate via the

2-Deoxyglucose glutanmate dehydrogenase reaction (17-19). Varioustheories exist as to what factors control ammonia pro-duction, but they all essentially attempt to explain how

04 glutaminase and glutamate dehydrogenase can be de-inhibited by removal of end products or inhibitors (17,20-27). Thus the currently accepted -scheme pivots onthe assumption that glutamate dehydrogenase plays aprimary role in ammonia formation from the a-amino

0.3 _ groups of amino acids.E Operation of the purine nucleotide cycle has been

demonstrated in exetracts of muscle (7, 28) and brain(29). The cycle may serve to maintain a high ratio of[ATP]/[ADP], and is involved in the control of gly-

< 0.2 / colysis (7). In addition it may serve in the control of/ \ the citric acid cycle and in the liberation of ammonia

I/' / \ \ _ fronm a-amino acids (5).The experiments reported above demonstrate the pro-

/P M. \ duction of ammonia from aspartate via the reactions of0.1 _ t, / the purine nucleotide cycle in particle-free extracts of

kidney cortex. High activities of glutamate-aspartate-,8 / < < transamiinase occur in every tissue (5), ensuring a rapid

d// \ conversion of glutamate to asp.artate and vice versa.

JO7 Adenylosuccinate synthetase i's strongly inhibited byO I 2 3 4 5 6 GDP, a product of the reaction it catalyzes, but this in-

Time (h) hibition can be avoided by enmploying a nucleosidetriphosphate-regenercating system such as one consisting

FIGURE 6 Effect of hypokalemia on the operation of the of creatierehonl)atinasean scratine p onspte(purine nucleotide cycle. A rat weighing 250 g was made Thypokalemic by a daily injection of 1.0 mg of deoxycor- Iheegenerating system also causes adenine nucleotideticosterone acetate for 10 days while being fed a low potas- to be "trapped" as the triphosplhate at the end of thesium diet and receiving 0.9% NaCl as drinking water over amiination plhase of the cycle: IMP + aspartate >the same period (serum potassium = 2.35 mequiv per liter). adeniylosucciniate > AMP-- ADP > ATP, and thisA sibling received the regular laboratory diet and water for . .

d

drinking. On the 11th day of the treatment, the animals were arevents thne deaminationphase of thoe cycle from oper-killed and extracts of their kidney cortices were prepared ating. One way in which deamination can be initiatedas described under Methods. The reaction mixture was as is by the additioll of 2-deoxyglucose, which, in con-described in the legend to Fig. 2, except that both test and junction with hexokinase and myokinase, leads to thereference cuvettes contained 0.8 mM1-pyrophosphoryl ribose conversion of ATP into ADP and AMP.5-phosphate. The spectrophotometric assay for the purinenucleotide cycle was employed. (A and A, adenine nucleo- Acid loading and hypokalemia are accompanied by antides; 0 and *, adenylosuccinate; broken lines, control increased output of ammiiiionia by the kidney. Our resultsrat; and solid lines, hypokalemic rat.) slhow that both conditions are associated with elevated

cient endogenous, enzyme-boundNAD* probably sur- activities of adenylosuccinate syntlhetase and adenylo-vived the gel filtration of the high-speed supernate to succinase (Table I). Experinmentally induced hypo-catalyze the dismutation between malate and lactate. kalenmia causes a decrease in the activity of adenylate

deaminase (Table I). However, it should be noted thatDISCUSSION even under these conditions, adenylate deaminase is still

The currently accepted scheme for ammonia production about three times more active than adenylosuccinaseby the kidney involves the conversion of glutamine to (Table I). (This comparison refers to the assays car-ammllonia and glutanmate via the glutaminase reaction, ried out at pH 7.2.) Of interest in this connection is aand the conversion of glutamate to ammonia and a-keto- report 'that primary hypokalemic periodic paralysis inglutarate via the glutamate dehydrogenase reaction (4, humiians is associated with very low levels of adenylate15-17). Another and probably minor pathway involves deamlinase in muscle (31). Adnministration of a loadthe transamination of glutamine and a-ketoglutarate to of carbolhydrate to patients suffering from this diseaseyield a-ketoglutaramate and glutamate, and the subse- can lead to sudden onsets of paralysis. In skeletal muscle


the rate of ammonia production correlates well with therate of glycolysis, although there is no stoichiometry be-tween the two processes (5).

The possible contribution of the purine nucleotidecycle to ammonia production by the kidney can be esti-mated by comparison of our results with previouslypublished data for the rat. The literature contains manyreports of rates of ammonia production by kidney basedon experiments with slices, perfused kidneys, and wholeanimals. Measurenments performed in vitro commonly in-volved the addition of glutamine to the incubation mix-ture or perfusate. In such cases the glutamine addedrepresented a large excess over the amount of ammoniaproduced during the experiment, and the amide group ofglutamine was the major source of ammonia. Under suchconditions it is difficult to assess the amounts of am-monia arising from other sources such as the a-aminogroup of amino acids. An assessment of the relative con-tributions to ammonia production made by differentnitrogen sources under physiological conditions hasbeen made only for kidney of the dog; an average of43% of the excreted ammonia was calculated to arisefrom glutamine, 22% from the a-amino group of aminoacids, and 35% from arterial ammonia (4, 32). Com-parable assessments for ra,t or man are not available(33), although it has been estimated that 63% of theammonia produced by the kidney of man is contributedby the amide nitrogen of glutamine (32).

Several investigators have tested the effect of addingglutamate or aspartate on ammonia production by kidneyslices or the perfused kidney (34-39). In these experi-nients it is difficult to judge how much ammonia arisesfrom the added substrate and how much from endogen-ous sources such as glutamine. In all cases addition ofaspartate or glutamate caused an enhancement of am-monia production. The concentrations of glutamate oraspartate that were added were very high compared totheir normal extracellular concentrations. Comparison ofnormal kidneys with kidneys from acidotic animals showsthat in all cases the latter exhibited higher rates of am-monia production (22, 34-40).

Table I slhows that the specific activities of adenylo-succinate synthetase and adenylosuccinase in kidneysof normal rats were 1.2 and 1.3 nmol/mg protein permin respectively. These activities rose to 1.9 and 3.3,respectively, in rats fed ammonium chloride. The specificactivity of adenylate deaminase in kidneys from am-monium chloride-fed rats sets an upper limit to ammoniaproduction by this enzyme of 26 nmol/mg protein permin. These activities are of the order of magnitude8 of

3 The rates found in the literature were converted tonanomoles per milligram cytoplasmic protein per minute byusing the following conversion factors. Rates expressed pergram fresh weight of kidney were converted to rates per

the rates of ammonia production reported in the litera-ture (22, 34-40). A comparison of the rates ofammonia production reported for intact rats (22, 40)with the activities of adenylosuccinate synthetase andadenylosuccinase (Table I) shows that these enzymescan account for 10-100% of ammonia production innormal rats and for 7-64% in acidotic rats. In view ofthe wide range of rates of ammonia production reportedin the literature, and the uncertainty concerning the con-tribution made by the amide group of glutamine, thesefigures indicate that in the kidney the purine nucleotidecycle may account for a substantial portion of ammoniaproduction from a-amino groups of a-amino acids.

Our results also show that kidney extracts catalyzethe conversion of IMP to inosine and uric acid (Fig. 3).The loss of purine nucleotides into nucleosides and purinebases is probably initiated by alkaline phosphatase, sinceit cannot be prevented by adenosine a,#-methylene diphos-phonate (AOPCP), a powerful inhibitor of 5'-nucleoti-dase (14). The contribution that the mammalian kidneymakes to an animal's total uric acid production has notbeen assessed previously. The conversion of IMP touric acid is inhibited by either allopurinol or 1-pyrophos-phoryl ribose 5-phosphate, while the production ofammonia via the purine nucleotide cycle is markedlyincreased by these substances (Figs. 3 and 4). The sim-plest explanation for this finding is that uric acid pro-duction constitutes a drain on the pool of purine nucleo-tides that catalyze the production of ammonia via thepurine nucleotide cycle, and that allopurinol and 1-pyrophosphoryl ribose 5-phosphate block this drain.Clinical observations have linked ammonia and uricacid production in gout (42, 43). Compared to controls,patients with primary gout may excrete 30% less am-

monia in the urine (44), and they have elevated bloodlevels of glutamate (45). Our studies indicate thatammonia and uric acid production in the kidney may belinked via the purine nucleotide cycle. IMP can either

milligram cytoplasmic, nonparticulate protein by multiplyingby 70. The amount of cytoplasmic, nonparticulate protein pergram fresh weight of kidney was determined experimentallyby a method reported previously (41), and was found tobe 70 mg protein per g fresh weight. (The enzyme activitiesmeasured in Table II are found largely and possibly ex-

clusively in the soluble cytoplasm.) The average wet weightto dry weight ratio for rat kidney was determined to be3.92; this factor was used to convert results expressed pergram dry weight of kidney to results per gram wet weight.The activities quoted in Table II are per minute, whereasmost rates of ammonia production in the literature are

quoted per hour or longer; appropriate corrections were

applied to obtain the same time base. Some results in theliterature are expressed per 100 g body weight; to convertthese to per gram fresh weight of kidney, we used a valueof 0.82 g kidney per 100 g body weight. Appropriate cor-rections were also made to convert results expressed inmillimole and micromole to nanomole.


participate catalytically in the purine nucleotide cycle orbe dissipated to form urate. The factors that controlthe relative rates at which these two pathways operatein vivo remain to be determined.

ACKNOWLEDGMENTSWe gratefully acknowledge the assistance of Keith Torn-heim, Vera Schultz, Karen Steele, and Gretchen Bean.

This work was supported in part by U. S. Public HealthService grants AM-11793 and GM-07261.

REFERENCES1. Pitts, R. F. 1972. Symposium on acid-base homeostasis.

Control of renal production of ammonia. Kidney Int. 1:297-305.

2. Van Slyke, D. D., R. A. Phillips, P. B. Hamilton, R.M. Archibald, P. H. Futcher, and A. Hiller. 1943. Glu-tamine as source material of urinary ammonia. J. Biol.Chemn. 150: 481-482.

3. Pitts, R. F., L. A. Pilkington, and J. C. M. de Haas.1965. N' tracer studies on the origin of urinary am-monia in the acidotic dog, with notes on the enzymaticsynthesis of labeled glutamic acid and glutamines. J.Clin. Invest. 44: 731-745.

4. Pitts, R. F. 1968. Physiology of the Kidney and BodyFluids. Year Book Medical Publishers Inc., Chicago.2nd edition. 202.

5. Lowenstein, J. M. 1972. Ammonia production in muscleand other tissues: The purine nucleotide cycle. Physiol.Rev. 52: 382-414.

6. Goodman, A. D., R. E. Fuisz, and G. F. Cahill, Jr. 1966.Renal gluconeogenesis in acidosis, alkalosis, and potas-sium deficiency. Its possible role in regulation of renalammonia production. J. Clin. Invest. 45: 612-619.

7. Tornheim, K., and J. M. Lowenstein. 1972. The purinenucleotide cycle. The production of ammonia fromaspartate by extracts of rat skeletal muscle. J. Biol.Chem. 247: 162-169.

8. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J.Randall. 1951. Protein measurement with the Folinphenol reagent. J. Biol. Chem. 193: 265-275.

9. Brown, H. 1945. The determination of uric acid inhuman blood. J. Biol. Chem. 158: 601-608.

10. Kalckar, H. M. 1947. Differential spectrophotometryof purine compounds by means of specific enzymes. I.Determination of hydroxypurine compounds. J. Biol.Chem. 167: 429443.

11. Seligson, D., and H. Seligson. 1951. A microdiffusionmethod for the determination of nitrogen liberated asammonia. J. Lab. Clin. Med. 38: 324-330.

12. Hohorst, H-J. 1963. L-(+)-Lactate. Determination withLactic Dehydrogenase and DPN. In Methods of En-zymatic Analysis. H-U. Bergmeyer, editor. AcademicPress, Inc., New York. 266-270.

13. Williamson, J. R., and B. E. Corkey. 1969. Assays ofintermediates of the citric acid cycle and related com-pounds by fluorometric enzyme methods. Methods En-zymol. 13: 434-513.

14. Burger, R. M., and J. M. Lowenstein. 1975. 5'-Nucleo-tidase from smooth muscle of small intestine and frombrain. Inhibition by nucleotides. Biochemistry. 14: 2362-2366.

15. Simpson, D. P. 1972. Pathways of glutamine and or-ganic acid metabolism in renal cortex in chronic meta-bolic acidosis. J. Clin. Invest. 51: 1969-1978.

16. Bourke, E., A. Fine, and J. Scott. 1971. Mechanismsof ammoniagenesis in human kidney. Biochem. J. 125:94P. (Abstr.)

17. Goldstein, L. 1967. Pathways of glutamine deaminationand their control in the rat kidney. Am. J. Physiol.213: 983-989.

18. Goldstein, L., R. Richterich-Van Baerle, and E. H.Dearborn. 1957. Kidney gluatminases. II. The glutamine-a-keto acid transamination-deamidation system of theguinea pig. Enzymologia. 18: 261-270.

19. Richterich, R. W., and L. Goldstein. 1958. Distributionof glutamine metabolizing enzymes and production ofurinary ammonia in the mammalian kidney. Amn. J.Physiol. 195: 316-320.

20. Welbourne, T. C., and S. Balagura-Baruch. 1972. Renalmetabolism of glutamine in dogs during infusion ofa-ketoglutaric acid. Am. J. Physiol. 222: 663-666.

21. Pagliara, A. S., and A. D. Goodman. 1970. Relation ofrenal cortical gluconeogenesis, glutamate content, andproduction of ammonia. J. Clin. Invest. 49: 1967-1974.

22. Alleyne, G. A. O., and G. H. Scullard. 1969. Renalmetabolic response to acid base changes. I. Enzymaticcontrol of ammoniagenesis in the rat. J. Clin. Invest.48: 364-370.

23. Preuss, H. G. 1968. Pyridine nucleotides in renal am-monia metabolism. J. Lab. Clin. Med. 72: 370-382.

24. Preuss, H. G., and F. R. Weiss. 1971. Rate-limitingfactor in rat kidney slice ammoniagenesis. Am. J. Phys-iol. 221: 458-464.

25. Preuss, H. G., F. R. Weiss, and S. Adler. 1971. Renalammonia production in the presence of citric acid cycleblockade. Proc. Soc. Exp. Biol. Med. 136: 738-741.

26. LeJohn, H. B. 1968. Unidirectional inhibition of glu-tamate dehydrogenase by metabolites. A possible regu-latory mechanism. J. Biol. Chemn. 243: 5126-5131.

27. De Haan, E. J., J. M. Tager, and E. C. Slater. 1967.Factors affecting the pathway of glutamate oxidationin rat-liver mitochondria. Biochim. Biophys. Acta. 131:1-13.

28. Tornheim, K., and J. M. Lowenstein. 1975. The purinenucleotide cycle. Control of phosphofructokinase andglycolytic oscillations in muscle extracts. J. Biol. Chem.250: 6304-6314.

29. Schultz, V., and J. M. Lowenstein. 1976. Purine nu-cleotide cycle. Evidence for the occurrence of the cyclein brain. J. Biol. Chem. 251: 485-492.

30. Snedecor, G. W., and W. G. Cochran. 1971. StatisticalMethods. Iowa State University Press, Ames. 7th edi-tion. 93.

31. Engel, A. G., C. S. Potter, and J. W. Rosevear. 1964.Nucleotides and adenosine monophosphate deaminaseactivity of muscle in primary hypokalaemic periodicparalysis. Nature (Lond.). 202: 670-672.

32. Pitts, R. F. 1973. Production and excretion of am-monia in relation to acid-base regulation. Renal physi-ology. Handb. Physiol. 8: 455-496.

33. Relman, A. S., and R. G. Narins. 1975. The control ofammonia production in the rat. Med. Clin. North Am.59: 583-593.

34. Kamm, D. E., and G. L. Strope. 1972. The effects ofacidosis and alkalosis on the metabolism of glutamineand glutamate in renal cortex slices. J. Clin. Invest. 51:1251-1263.

35. Freed, K. H., C. Bowie, 0. Vavatsi-Manos, and H. G.Preuss. 1973. Oxygen consumption and ammoniagenesisin rat kidney slices. Am. J. Physiol. 224: 268-270.


36. Vavatsi-Manos, O., D. M. Roxe, G. E. Schreiner, andH. G. Preuss. 1973. y-Aminobutyric acid shunt in renalammoniagenesis. Am. J. Physiol. 224: 154-157.

37. Hems, D. A. 1972. Metabolism of glutamine and glu-tamic acid by isolated perfused kidneys of normal andacidotic rats. Biochemn. J. 130: 671-680.

38. Pitts, R. F. 1971. Metabolism of amino acids by theperfused rat kidney. Amn. J. Physiol. 220: 862-867.

39. Welbourne, T. C. 1974. Ammonia production and path-ways of glutamine metabolism in the isolated perfusedrat kidney. Aml. J. Physiol. 226: 544-548.

40. Goorno, W. E., F. C. Rector, Jr., and D. W. Seldin.1967. Relation of renal gluconeogenesis to ammoniaproduction in the dog and rat. Am. J. Physiol. 213:969-974.

41. Spencer, A., L. Corman, and J. M. Lowenstein. 1964.Citrate and the conversion of carbohydrate into fat.A comparison of citrate and acetate incorporation intofatty acids. Biocheml. J. 93: 378-388.

42. Gutman, A. B., and T-F. Yu. 1963. On the nature ofthe inborn metabolic error(s) of primary gout. Trans.Assoc. Am. Physicians. Phila. 76: 141-151.

43. Gutman, A. B., and T. F. Yu. 1963. An abnormality ofglutamine metabolism in primary gout. Am. J. Med.35: 820-831.

44. Falls, W. F., Jr. 1972. Comparison of urinary acidifica-tion and ammonium excretion in normal and gouty sub-jects. Metab. Clitt. Exp. 21: 433-445.

45. Gutman, A. B., and T-F. Yu. 1973. Hyperglutamatemiain primary gout. Amii. J. Med. 54: 713-724.


the purine nucleotide cycle -...

Documents