41

Molecular and Cellular Biochemistry 187: 41–46, 1998.© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

NADH oxidase activity of soybean plasmamembranes inhibited by submicromolarconcentrations of ATP

D. James MorréDepartment of Medicinal Chemistry and Molecular Pharmacology, Purdue University, West Lafayette, IN, USA

Received 28 July 1997; accepted 21 November 1997

Abstract

The activity of an auxin-stimulated NADH oxidase activity from soybean hypocotyls was inhibited by submicromolarconcentrations of ATP. Auxins are plant growth regulators that increase the rate of cell enlargement in plant stems. A syntheticauxin, 2,4-dichlorophenoxyacetic acid (2,4-D), was used. The inhibition was half maximal at 1 nM ATP and was not observedwith other nucleotides and nucleosides. The inhibition was the result of an increase in the K

m for NADH from about 60 µM

to > 100 µM and was noncompetitive. The decrease in Km due to ATP was enhanced by the addition of 1 µM 2,4-D. The V

max

of the plasma membrane NADH oxidase was approximately doubled (1.5–2.8-fold) by ATP and by 1 µM 2,4-D. No furtherincrease in the V

max was observed by the combination of 1 nM to 0.1 mM ATP in the presence of 1 µM 2,4-D. The results

demonstrate a response of the NADH oxidase activity of isolated vesicles of soybean plasma membranes to ATP distinctfrom that observed previously with other nucleotide di- and triphosphates. The results are suggestive either of control of thecell surface NADH oxidase by phosphorylation or a direct response to ATP binding at nanomolar concentrations of ATP.(Mol Cell Biochem 187: 41–46, 1998)

Key words: NADH oxidase, cyclic AMP, ATP, growth inhibition, plasma membrane, plant, soybean

Abbreviations: 2,4-D – 2,4-dichlorophenoxyacetic acid; 2,3-D – 2,3-dichlorophenoxyacetic acid

Introduction

The auxins are low molecular weight plant growth regulatorsbest known for their ability to stimulate enlargement of cellsof plant stems. Both natural [e.g. indole-3-acetic acid (IAA)]and synthetic [e.g. 2,4-dichlorophenoxyacetic acid (2,4-D)]elicit similar growth responses. An NADH oxidase activitystimulated by these auxins, both natural (IAA) and synthetic(2,4-D), has been described from soybean hypocotyls [1–3]and shown to be activated by guanine nucleotides [4]. Theactivity of the oxidase correlates with rate of cell enlargement[3] and the protein is expected to play some functional role

in the growth process. In the course of these studies it wasnoted that, in the absence of other nucleotides, the activityappeared to be inhibited by ATP within the same nano-molar concentration range as GTP stimulated (Figs 1 and3 of ref. 4). The inhibition was not observed with ADP (Fig. 3of ref. 4). With both pig liver [5] and rat liver [6] plasmamembranes, nM ATP also was found to inhibit the NADHoxidase activity. The inhibition was augmented by cAMP butoccurred, as well, in its absence [5, 6]. This paper was tocharacterize further the inhibition of NADH oxidation byisolated plasma membrane vesicles from hypocotyls ofetiolated soybean seedlings.

Address for offprints: D. James Morré, Department of Medicinal Chemistry and Molecular Pharmacology, 1333 HANS Life Sciences Research Building, PurdueUniversity, West Lafayette, IN 47907-1333, USA

42

Materials and methods

Plant material and isolation of plasma membranes

Soybean seeds (Glycine max L. Merr., cv. Williams) weresoaked for 4 h in deionized water and grown in the dark(20–22°C) in moist vermiculite contained in foil-covered18 × 23 × 10 plastic boxes normally without supplementaladditions of water. After 4–6 days, 2 cm hypocotyl seg-ments, cut 5 mm below the cotyledon, were harvested underdiminished light (0.15 µE s–1 m–2) and placed in cold water.Segments (40 g) were homogenized (Waring blender) in40 ml of homogenization buffer (0.3 M sucrose, 50 mMTris-Mes [pH 7.4], 10 mM KCl, 1 mM MgCl

2, and 1 mM

PMSF). The homogenates were filtered through one layer ofMiracloth (Chicopee Mills, NY) and centrifuged for 10 minat 6,000 g (Sorvall HB-rotor). The supernatant was re-centrifuged at 60,000 g (Beckman SW 28 rotor) for 30 minand the pellets were resuspended in 0.25 M sucrose with 5mM potassium phosphate (pH 6.8). Plasma membranevesicles were prepared using a 16 g aqueous two-phasepartitioning system that yielded predominantly right side-out and sealed vesicles [7, 8]. Resuspended 60,000 g pelletswere mixed with 6.4% (w/w) Polyethylene Glycol 3350(Fisher), 6.4% (w/w) Dextran T500 (Pharmacia), 0.25 Msucrose and 5 mM potassium phosphate (pH 6.8). Aftermixing the tubes by 40 inversions in the cold, the phaseswere separated by centrifugation at 750 g for 5 min. Theupper phase was collected and the lower phase was partitioneda second time with fresh upper phase. The first and secondupper phases were combined and diluted approximately 4-fold with buffer. The plasma membranes were collected bycentrifugation at 100,000 g for 30 min and stored at –70°Cprior to assay. The yield was 4 mg of plasma membraneprotein. The purity based on morphometric analysis afterspecific staining with phosphotungstic acid at low pH [9] andassay of marker enzymes was > 95%.

NADH oxidase activity

The assay for the plasma membrane NADH oxidase was in50 mM Tris-Mes buffer (pH 7.0), 150 µM NADH in thepresence of 1 mM potassium cyanide, the latter to inhibitany mitochondrial NADH oxidases contaminating theplasma membranes. The assay was started by the additionof 0.1 mg of plasma membrane protein. The reaction wasmonitored by the decrease in the absorbance at 340 nmusing a Hitachi Model U3210 spectrophotometer. Thechange of absorbance was recorded as a function of timeby a chart recorder. The specific activity of the plasmamembrane was calculated using an absorption coefficientof 6.21 mM–1 cm–1.

Assays were initiated by addition of NADH. Following theaddition of NADH and for each subsequent addition, theassays were continued for 10 min with the steady state ratesbetween 5–10 min being reported.

Protein

Protein content was determined by the BCA procedure [10].Standards were prepared with bovine serum albumin.

Results

Effect of ATP on NADH oxidase activity

The NADH oxidase was assayed under conditions of linearitywith respect to time and protein concentration. Whenincubated for 10 min at 37°C in the presence of ATP, theactivity was inhibited by about 50% at ATP concentrationsof 0.1–1 µM. The concentration for half maximal inhibitionof NADH oxidation was about 1 nM and at the lowest NADHconcentration tested in Fig. 1 of 30 µM, an effect wasobserved at an ATP concentration as low as 0.1 nM. As the

Fig. 1. NADH oxidase activity as a function of ATP concentration at 30, 60and 150 µM NADH assayed over 10 min. Each assay contained ca 100 µgplasma membrane protein in a total volume of 2.5 ml. Values are averagesfrom 3 experiments ± S.D. among experiments.

43

ATP concentration exceeded 1 µM, the activity was stimulatedby ATP as reported previously [4].

The plasma membrane vesicles utilized in the study weresealed and largely right side-out. Latency based on ATPaseactivity induced in response to disruption of the vesiclemembranes by detergent treatment was > 85%.

The kinetic parameters of the NADH oxidase activity ofthe sealed, right sideout plasma membrane vesicles werealtered by incubation with ATP (Table 1). Analyses of doublereciprocal plots revealed a K

m for NADH of 61 ± 10 µM in

the absence of ATP and a significantly increased Km for

NADH of 134 ± 4 µM with incubation with 1 nM ATP. Atthe same time, the apparent V

max increased slightly from

3.3 ± 1.0 in the absence of ATP to 9.4 ± 2.3 in the presenceof 1 nM ATP.

The change in Km expressed as an inhibition in NADH

oxidase activity at NADH concentrations of 30–150 µMwas maximal after about 10 min at 1 nM ATP or higher.Although not statistically different from 1 nM ATP, boththe K

m and V

max appeared to be less affected by 0.1 mM

ATP than by 1 nM ATP although both were increasedcompared to no ATP (Table 1). In the presence of a 1 µMconcentration of the auxin herbicide 2,4-D, ATP alsoincreased the K

m. The K

m was increased in the presence of

2,4-D significantly above the values obtained for ATP in theabsence of 2,4-D (Table 1). The apparent V

max was increased

by the addition of 2,4-D alone to about the same extent aswith ATP alone. The V

max was not increased beyond that

observed in the presence of 2,4-D by the further addition ofATP despite the rather marked and statistically significantincrease in the K

m resulting from simultaneous 2,4-D plus

ATP addition (Table 1).The kinetics tended to be biphasic both in the presence

and in the absence of 1 nM ATP (Fig. 2). The kineticparameters were weighted most heavily on the range of ATPconcentrations between 0.015–0.24 mM NADH, the mostaccurate portion of the dose response.

Table 1. Kinetic parameters of the NADH oxidase activity of soybeanplasma membranes and response to ATP and 2,4-D

Km Vmax

Addition n (µM) (nmoles/min/mgprotein)

None 6 61 ± 10a 3.3 ± 1.0a

ATP 1 nM 6 134 ± 4b 9.4 ± 2.3b

ATP 1 µM 3 138 ± 4b 6.7 ± 1.6b

ATP 0.1 mM 3 100 ± 27b 5.1 ± 1.6ab

2,4-D 1 µM 4 94 ± 2b 5.4 ± 1.0b

ATP 1 nM + 2,4-D 1 µM 3 183 ± 17c 5.4 ± 1.2ab

ATP 1 µM + 2,4-D 1 µM 3 261 ± 67c 11.0 ± 4.6b

ATP 0.1 nM + 2,4-D 1 µM 3 183 ± 76bc 5.1 ± 1.6ab

Values are averages ± S.D. Values not followed by the same letter aresignificantly different (p < 0.01) as determined by two-tailed t-test.

Fig. 2. NADH oxidase activity in the presence of 100 mM Tris-HCl (pH 7.4)with varying concentration of NADH where activity was inhibited by 1 nMATP. Solid symbols, no ATP; open symbols, dotted lines in the presence of1 nM ATP. Averages are from 3 experiments ± S.D. among experiments. Theinset shows the double reciprocal plots calculated from the experimentalmeans.

The biphasic characteristics were enhanced at thehigher ATP concentrations. At 1 µM ATP a stronglybiphasic character was seen with the low NADH con-centrations (Fig. 3). The double reciprocal plots werenearly parallel extrapolating to a decreased K

m and V

max.

However, at 0.03 mM NADH, the slope changed abruptly toyield both an increased K

m and apparent V

max. Yet in these

experiments, NADH oxidase activity was still inhibited evenat the highest NADH concentration tested of 0.24 mM inthe presence of 1 µM ATP (See also Fig. 1).

ATP-γ-S in 2 experiments did not inhibit NADH oxidaseactivity of soybean plasma membranes over the range 10–10

to 10–6 M in the presence of either 15 and 60 µM NADH.The K

m for NADH from three experiments in the presence

of 1 µM ATP-γ-S was 42 ± 18 µM and the Vmax

was 5.6 ± 1.0nmoles/min/mg protein.

Inhibition of NADH oxidase activity by ATP was notdependent on cAMP nor did it appear to be augmented bycAMP. At optimum conditions for cAMP and ATP inhibition

44

ADP was without effect on the NADH oxidase activity overthe inhibitory concentration range for ATP of 10–10 to 10–7 M.The K

m determined in the presence of 1 nM ADP was 56 µM

compared to 167 µM for 1 nM ATP determined in parallel.

Discussion

The NADH oxidase of plasma membranes, which is of lowspecific activity, has been purified from both plants [1] andrat liver [3, 11]. The activity purifies as a complex of threemajor proteins of approximate molecular masses of 36, 55,and 72 kD [1]. The activity is of interest in that it is hormone-and growth factor-responsive in plasma membranes of soybean[2], rat liver [3, 11], and normal human keratinocytes [12]but constitutively active in plasma membranes of hepatomas[13] and hyperplastic liver nodules induced by the livercarcinogen, 2-acetylaminofluorene [14]. The activity hasbeen well characterized kinetically and with respect toproportionality with time of incubation, plasma membraneprotein content, pH optimum and cofactor requirements. TheNADH oxidase activity of the purified plasma membranes isresistant to the usual mitochondrial inhibitors includingrotenone and antimycin and the assays are carried outroutinely in the presence of 1 mM potassium cyanide.

An apparent inhibition of plasma membrane NADHoxidase activity by low concentrations of ATP [4] wasconfirmed previously for plasma membranes of pig liver [5]and in the present communication for plasma membranes ofetiolated soybean seedlings. The inhibition was the resultof a shift in the apparent K

m of the oxidase from about 60 µM

to > 100 µM NADH in the presence of 1 nM to 0.1 mM ATP.A number of plasma membrane proteins are regulated byATP [15] but few are maximally influenced by such lownucleotide concentrations as was the plasma membraneNADH oxidase activity.

The ATP concentration that resulted in inhibition of NADHoxidation of about 0.1 µM is well below the physiologicalrange of ATP concentrations for dark-grown plant tissues.Estimates of cytosolic ATP levels in plant cells range from0.2 mM (darkness) to 2 mM (light) [16].

With the pig liver plasma membranes [5] and also forrat liver plasma membranes [6], the ATP response wasaccelerated by the coaddition of cyclic AMP at concentrationsabout 0.1 those of the ATP. Cyclic AMP was not required toelicit a response. ATP alone was effective. The fact that cyclicAMP accelerated the reaction, however, and that the responsewas blocked by the selective inhibitor of cyclic AMP-dependent protein kineses, H-89 but not by staurosporine[5], suggested the possible involvement of protein phos-phorylation in the inhibitory response.

Except for the lack of response to cyclic AMP with theplant membranes, the inhibition of NADH oxidase activity

Fig. 3. NADH oxidase activity in the absence (solid symbols) or presence(open symbols, dotted lines) of 1 µM ATP. Averages are from 3 experiments± S.D. among experiments. The inset shows the double reciprocal plotscalculated from the experimental means.

of the activity in pig and rat liver (1 nM ATP + 0.1 nMcAMP), the K

m for NADH with no addition was 50 µM, for

ATP alone was 111 µM, for cAMP alone was 63 µM and forthe combination of ATP plus cAMP was 50 µM. Thecalculated V

max for each of the 4 treatments in nmoles/min/

mg protein were 3.0 for no addition, 4.8 for ATP alone, 3.6for cAMP alone and 3.6 for the combination of ATP pluscAMP. With 1 nM ATP, the inhibition occurred after a brieflag and was not maximal until after 8 min of incubation.With 1 µM ATP, inhibition appeared to be maximal withinthe first minute after ATP addition.

In single experiments, the response of the kinetic par-ameters to the redox environment was determined. In thepresence of 0.003% hydrogen peroxide, the V

max for

NADH was increased from 71–333 µM NADH by theaddition of 1 nM ATP. The V

max was increased 2.5-fold by

1 nM ATP in the presence of 0.003% H2O

2. With 1 µM

reduced dithiothreitol, the Km for NADH was increased

from 25–67 µM by 1 nM ATP and the Vmax

was reduced to0.7 the value in the absence of ATP. With both GSH and GSSGthe K

ms were increased from 66 ± 1 µM to 163 ± 37 µM

NADH by 1 nM ATP and the Vmax

was increased 1.5-fold.

45

with plasma membrane vesicles of soybean and liver weresimilar. Both occurred at submicromolar concentration ofATP and were time and concentration dependent. Bothappeared to be sensitive to redox environment. With rat liverplasma membranes 1 µM dithiothreitol resulted in a loweredK

m for NADH both in the presence and absence of ATP,

whereas 100 µM of the mild oxidizing agent N-chloro-succinamide increased the K

m for NADH both in the

presence and absence of ATP.If the response of the plasma membrane-associated

NADH oxidase activity was mediated by a protein kinase,would such low concentrations of ATP survive hydrolysis fora sufficient length of time to be effective? The plasmamembrane preparations used exhibited an ATPase activity ofabout 100 nmol/min per mg protein under optimum con-ditions. The NADH oxidase assays were in the absence ofadded Mg2+ so the activity may have been as low as 5 nmoles/min/mg protein [17]. For 100 µg protein, the ATPase activitywould have been about 0.5 nmol/min in the final assayvolume of 2.5 ml. Additionally, the concentrations of ATPwere several orders of magnitude below the K

m of the

ATPase and for the most part would probably be bound tomembrane proteins. Even so, at the lowest concentrationsof ATP observed to alter the NADH oxidase, e.g. 1 nM, theadded ATP should not have survived in solution for morethan a few seconds. It is possible, however, that a portion ofthe ATP bound to membranes may have survived longer.

There have been a number of reports in the literature thatextracellular ATP in the submillimolar range will inhibit thegrowth of transformed mammalian cells in culture but inhibitto a lesser extent the growth of non-transformed cells. Forexample, addition of ATP to cultures of transformed mousefibroblasts (Swiss Mouse 3T6 cells) resulted in growthinhibition, whereas the growth of the correspondingnontransformed counterparts of 3T6 cells, 3T3 cells, wasonly slightly affected [18].

Effects of externally supplied ATP on plants appear to belimited to the report of Hagar et al. [19] where ATP wasconvincingly demonstrated to restore auxin- (IAA-) inducedgrowth to auxin-starved hypocotyls of Helianthus annusunder anaerobic conditions. The response was to millimolarconcentrations of nucleotide and was given as well by ITP,GTP, UTP, and CTP in parallel to the nucleotide stimulationsof NADH oxidation reported previously for NADH oxidationfor both soybean [4], and rat liver [20] plasma membranes.

Were the response of a cell surface activity to ATP inplants of physiological relevance, how might the externalsurface of the plant cell come in contact with even nanomolarconcentrations of ATP? One explanation might be providedfrom the corelease of ATP during secretion from thecontents of Golgi apparatus secretory vesicles fusing withthe plasma membrane. For example, ATP is costored withcatecholamines within chromaffin granules in the ratio of 4

moles of catecholamine: 1 mole of ATP [21]. In the vasdeferens of the guinea pig, both ATP and norepinephrineare not only cosecreted but appear to function as co-transmitters based on a variety of pharmacological andelectrophysiological experimentation [22]. Even in plants,low concentration of secreted ATP might function as partof a feedback signaling system to coordinate secretion andcell enlargement.

Several examples exist of phosphorylation of ecto-proteins by externally supplied ATP presumably through theaction of protein kinases located at the extracellularsurface of the plasma membrane with aortic endothelialcells [23] the surface of brain neurons [24] and in hippo-campal slices [25]. During the induction of long-termpotentiation in the hippocampal slices, a 48/50 kD proteinduplex became phosphorylated by extracellular ATP. In thissystem, ATP is secreted into the synaptic cleft as thesource of nucleotide. A monoclonal antibody directed tothe catalytic domain of protein kinase C selectivelyinhibited the protein phosphorylation activity providingadditional evidence for the existence and operation of anecto-protein kinase activity on the surface of hippocampalpyramidal neurons [25].

Despite uncertainties of functional significance, theinhibition of an NADH oxidase activity by ATP with rightside-out plasma membrane vesicles from soybean mayindicate a new property of the plant cell surface with potentialfor a better understanding of plant cell enlargement [26].Additionally, it may provide yet another parallel between cellelongation induced by auxin and uncontrolled growth incancer involving the cell surface NADH oxidase.

References

1. Brightman AO, Barr R, Crane FL, Morré DJ: Auxin-stimulated NADHoxidase purified from plasma membrane of soybean. Plant Physiol 86:1264–1269, 1988

2. Morré DJ, Brightman AO, Wu L-Y, Barr R, Leak B, Crane FL: Role ofplasma membrane redox activities in elongation growth in plants.Physiol Plant 73: 187–193, 1988

3. Morré DJ, Brightman AO: NADH oxidase of plasma membranes. JBioenerg Biomembr 23: 469–489, 1991

4. Morré DJ, Brightman AO, Barr R, Davidson M, Crane FL: NADHoxidase activity of plasma membranes of soybean hypocotyls isactivated by guanine nucleotides. Plant Physiol 102: 595–602, 1993

5. Morré DJ, Navas P, Rodriguez-Aguilera J-C, Morré DM, Villaba JM, deCabo R, Lawrence J: Cyclic AMP-plus ATP-dependent modulation ofthe NADH oxidase activity of porcine liver plasma membranes. BiochimBiophys Acta 1224: 566–574, 1994

6. Morré DJ, Rodriguez-Aguilera J-C, Navas P, Morré DM: Redoxmodulation of the response of NADH oxidase activity of rat liver plasmamembranes to cyclic AMP plus ATP. Mol Cell Biochem, 173: 71–77, 1997

7. Kjellbom P, Larsson C: Preparation and polypeptide composition ofchlorophyll-free plasma membranes from leaves of light-grown spinachand barley. Physiol Plant 62: 501–509, 1984

46

8. Sandelius AS, Barr R, Crane FL, Morré DJ: Redox reactions of plasmamembranes isolated from soybean hypocotyls by phase partition. PlantSci 48: 1–10, 1987

9. Roland JC, Lembi CA, Morré DJ: Phosphotungstic acid-chromic acidas a selective electron-dense stain for plasma membranes of plant cells.Stain Technol 47: 195–200, 1972

10. Smith PK, Krohn RI, Hermanson GT, Mallia AK, Gartner FH,Provenzano MD, Fujimoto EK, Goeke NM, Olson BJ, Klenk DC:Measurement of protein using bicinchoninic acid. Anal Biochem 150:70–76, 1985

11. Brightman AO, Wang J, Kin-man Miu R, Sun IL, Barr R, Crane FL, MorréDJ: A growth factor- and hormone-stimulated NADH oxidase from ratliver plasma membrane. Biochim Biophys Acta 1105: 109–117, 1992

12. Morré DJ, Morré DM, Paulik M, Batova A, Broome AM, Piri L, CreekKE: Retinoic acid and calcitriol inhibition of growth and NADH oxidaseof normal and immortalized human keratinocytes. Biochim BiophysActa 1134: 217–222, 1992

13. Bruno M, Brightman AO, Lawrence J, Werderitsh D, Morré DJ:Stimulation of NADH oxidase activity from rat liver plasma membranesby growth factors and hormones is decreased or absent with hepatomaplasma membranes. Biochem J 284: 625–628, 1992

14. Morré DJ, Crane FL, Eriksson LC, Low H, Morré DM: NADH oxidaseof liver plasma membrane stimulated by diferric transferrin and neoplastictransformation induced by the carcinogen 2-acetylaminofluorene.Biochim Biophys Acta 1057: 140–156, 1991

15. Dubayk GR, Fedan JS (eds): Biological Actions of Extracellular ATP.Ann NY Acad Sci 603: 542 pp, 1990

16. Spalding E, Goldsmith MHM: Activation of K+ channels in the plasmamembranes of Arabidopsis by ATP produced photosynthetically. PlantCell 5: 477–484, 1993

17. Berczi A, Morré DJ: Partial purification of plant plasma membrane K+,Mg2+ ATPase by ion exchange chromatography. Physiol Plant 85: 207–214, 1992

18. Friedberg I, Kübler D: The role of surface protein kinase in the ATP-induced growth inhibition in transformed mouse fibroblasts. Ann NYAcad Sci 603: 513–515, 1990

19. Hager A, Menzel H, Krauss A: Versuche und Hypothese zurPrimärwirkung des Auxins beim Streckungswachstum. Planta 100: 47–75, 1971

20. Morré DJ, Davidson M, Geilen C, Lawrence J, Flesher G, Crowe R,Crane FL: NADH oxidase activity of rat liver plasma membrane activatedby guanine nucleotides. Biochem J 292: 647–653, 1993

21. Winkler H, Carmichael SW: The chromaffin granule. In: A.M. Poisner,J.M. Trifaro (eds). The Secretary Granule. Elsevier, Amsterdam, 1982,pp 3–79

22. White TD: Role of adenine compounds in autonomic neurotrans-mission. Pharmacol Ther 38: 129–168, 1988

23. Pirotton S, Boutherin-Falson O, Robaye B, Boeynaems JM: Ecto-phosphorylation on aortic endothelial cells. Exquisite sensitivity tostaurosporine. Biochem J 285: 585–591, 1992

24. Hogan MV, Pawlowska Z, Yang HA, Kornecki E, Ehrilic YH: Surfacephosphorylation by ecto-protein kinase C in brain neurons: A targetfor Alzheimer’s beta-amyloid peptides. J Neurochem 65: 2022–2030,1995

25. Chen W, Wieraszko A, Hogan MV, Yang HA, Kornecki E, Ehrlich YH:Surface protein phosphorylation by ecto-protein kinase is required forthe maintenance of hippocampal long-term potentiation. Proc Natl AcadSci USA 93: 8688–8693, 1996

26. Morré DJ: The hormone- and growth factor-stimulated NADH oxidase.Bioenerg Biomemb 26: 421–433, 1994