www.elsevier.com/locate/microc

Microchemical Journal

Applicability of microplate assay coupled to Fiske–Subbarow reducer

for the determination of phosphorous produced by in vivo human

lymphocytes: PKC is probably cross talking with ecto 5V-nucleotidase

Alejandro Martınez-Martıneza,T, Laura A. de la Rosaa, Claudia A. Jımenez-Munoza,

Angel Gabriel Dıaz-Sancheza, Jesus A. Araujo-Gonzaleza, Jose R. Peralta-Videab,

Guadalupe de la Rosac, Jorge L. Gardea-Torresdeyb,c

aDepartamento de Ciencias Basicas, Instituto de Ciencias Biomedicas, Universidad Autonoma de Ciudad Juarez, Cd. Juarez, 32310, Chih., MexicobDepartment of Chemistry, University of Texas at El Paso, El Paso, Texas 79968-0513, USA

cEnvironmental Science and Engineering PhD Program, University of Texas at El Paso, El Paso, Texas 79968-0513, USA

Received 21 January 2005; accepted 22 January 2005

Available online 7 March 2005

Abstract

In this research, the phorbol ester, phorbol 12-myristate 13-acetate (PMA), was used to assess the effect of protein kinase C (PKC)

activation on the specific activity of ecto-5V-nucleotidase (eNT) in human lymphocytes. PMA mimics the effects of diacylglycerol, a natural

compound released by the hydrolysis of the glycosilphosphatidilinositol (GPI) moiety, in activating PKC. In order to evaluate the activity of

eNT in living lymphocytes, a micro-assay method was established with a low detection limit for inorganic phosphate (Pi) of 0.94 nmol Pi

assay�1. The dephosphorylation of Adenosine monophosphate (AMP) by functional lymphocytes was evaluated and the contribution of the

eNT activity was calculated by its inhibition with the specific eNT inhibitor a,h-methylene ADP (MADP) and the use of the broad spectrum

phosphatases inhibitor (but not eNT), levamisole. Under the conditions tested, we obtained an AMPase value of 8.05F4.4 nmol Pi million

cells�1 h�1. The addition of MADP to the incubation media decreased the AMPase activity to 2.43F0.9 nmol Pi million cells�1 h�1

( pb0.05). On the other hand, when lymphocytes were incubated with PMA, an increase of 182% in the AMPase activity was observed.

However, the addition of levamisole inhibited the AMPase activity by about 17%, while the co-incubation of cells with PMA and levamisole

reduced only an 8% of the total PMA-increased AMPase activity. These results show that (1) a non-radioactive micro-method can be used to

assess the Pi production in living cells; (2) the obtained data strongly suggest that eNT is the main ecto-enzyme present on the surface of

circulating lymphocytes responsible for the hydrolysis of extracellular AMP; and (3) that PKC is cross talking with eNT.

D 2005 Elsevier B.V. All rights reserved.

Keywords: Ecto-5V-nucleotidase; 5VNT; Nucleotidase; CD73; Lymphocyte; PKC; a-h-MADP; PMA; Pi microassay

1. Introduction

Lymphocyte cluster of differentiation 73 (CD73) is an

ecto-enzyme possessing ecto-5V-nucleotidase activity (EC

3.2.3.5, afterwards eNT/CD73), which catalyzes the

dephosphorylation of purine and pyrimidine, ribo- and

deoxyribonucleoside monophosphates to their correspond-

0026-265X/$ - see front matter D 2005 Elsevier B.V. All rights reserved.

doi:10.1016/j.microc.2005.01.013

T Corresponding author. Fax: +52 656 6881836.

E-mail address: amartine@uacj.mx (A. Martınez-Martınez).

ing nucleosides [1]. This enzyme has a molecular mass of

70-kDa, it is anchored to the plasma membrane by a

glycosilphosphatidilinositol (GPI) bridge and is present on

approximately 15% of peripheral blood lymphocytes

(PBL). However, it is unevenly distributed among the

different lymphocyte subsets, being expressed on the

majority of CD8+ T cells (51%) and B cells (70%),

whereas a lesser percentage of CD4+ T cells is CD73+

(11%) [2,3].

It has been suggested that CD73 plays a role in regulating

lymphocyte adhesion to endothelium, since antibodies (Ab)

81 (2005) 92–97

A. Martınez-Martınez et al. / Microchemical Journal 81 (2005) 92–97 93

against CD73 block lymphocyte adhesion to cultured

endothelial cells [4,5]. Similarly, an anti-CD73 Ab blocked

aggregation of freshly isolated germinal center B cells and

follicular dendritic cells [6].

Cell type specific regulation of the expressed eNT/

CD73 is suggested because it is sequestrated from the

lymphocyte surface after treatment with specific anti-

bodies; however, the same treatment was ineffective for

eNT/CD73 expressed in endothelial cells [7]. Moreover,

engagement of lymphocyte eNT/CD73 to the Ab results in

tyrosine phosphorylation and dephosphorylation of intra-

cellular protein substrates, whereas eNT/CD73 on endo-

thelial cells remains resistant to the Ab treatment [7].

These differences in regulation and function suggest that

the specific physiological roles of eNT/CD73 vary with the

tissue where it is studied, and that several mechanisms

may be involved in the regulation of eNT/CD73 activity

and expression [8–10]. There is evidence that eNT/CD73

participates in the clustering of the leukocyte integrin LFA-

1 and in the consequent binding of lymphocytes to

endothelial cells [6,11]. However, little is known about

the biochemical paths that eNT/CD73 uses to cross-talk to

different effectors in lymphocyte cells. Furthermore,

because anti eNT/CD73 monoclonal antibodies (mAbs)

can stimulate processes such as the production of T cells,

secretion of IL-2 (Interleukin-2), and the expression of the

IL-2R (Interleukin-2-receptor), it has been proposed that

eNT/CD73 can act as a membrane receptor that transmits

co-stimulatory signals for human T cell proliferation in

vivo [12–15].

Phosphoinositides and diacylglicerol (DAG) are ubiq-

uitous second messengers, which may be generated from

the phospholipase C-mediated hydrolysis of the phosphati-

dylinositol moiety of GPI-anchored membrane proteins. In

addition, since eNT/CD73 is a GPI-linked molecule, it can

move rather freely on the cell surface and form clusters with

other cell surface molecules most likely in lipid rafts, which

are preformed modules enriched in signaling molecules

[13–16]. Interestingly, certain GPI-linked proteins have

been shown to co-cluster with h2 integrins [17–19].

Furthermore, engagement with the urokinase receptor

increases h2 integrin-mediated binding of neutrophils and

monocytes to the endothelium [20]. Thus, it is very likely

that eNT/CD73 has the same properties on active cells such

as lymphocytes. The activation of eNT/CD73 could be

related to microenvironment localization of several mole-

cules related to the phosphoinositides/DAG signaling

cascade, such as phospholipase C and protein kinase C that

is activated by DAG [21].

In this work, the possibility of cross talk between protein

kinase C and eNT activity in living lymphocytes was studied

through the use of the PKC activator PMA and the eNT/

CD73 specific inhibitor a,h-methylene ADP (MADP). A

sensitive spectrophotometric method for the determination

of inorganic phosphate released by enzyme activity was

used. The results are reported herein.

2. Materials and methods

2.1. Reagents

Fiske–Subbarrow reducer was purchased from Fluka,

water was MQ grade, trace metal grade sulfuric acid was

from Fisher Scientific, and the rest of the chemicals were

purchased from SIGMA with the maximum purity

available.

2.2. Procedures

2.2.1. Lymphocytes purification

Blood samples were obtained from the Instituto Mex-

icano del Seguro Social Clınica #6, Ciudad Juarez,

Chihuahua, with ethical agreement of the Institution.

Lymphocytes were purified by centrifugation through a

PercollR density gradient according to a technique modified

by de la Rosa et al. [22] after Boyum [23]. Briefly, total

blood was diluted 1:1 using a purification buffer (125 mM

NaCl, 5 mM KCl, 1 mM MgCl2, 0.5 mM glucose, 14 mM

Trizma, 1 mM CaCl2, pH 7.4 with HCl). Four milliliter of

this diluted blood was gently overlaid on 4 mL of 60%

isotonic percoll (avoiding mixing of the solutions), and

centrifuged at 3000 rpm for 25 min at room temperature

(Fisher Scientific, 8K). Lymphocytes were obtained from

the interface between the plasma and percoll (lymphocytes

form a white band) and were washed twice by resuspension

in 2 mL of purification buffer followed by centrifugation at

2000 rpm for 10 min at room temperature. Washed

lymphocytes were resuspended in 2 mL of purification

buffer. After a gentle homogenization with a disposable

plastic Pasteur pipette, 100 AL aliquots were withdrawn for

cell counting. Cell number and size were measured using a

channel analyzer ADVIA (Hematology System, BayerR).The viability of the lymphocytes was determined in a

Neubauer chamber using trypan blue as exclusion colorant

to verify the integrity of the plasma membrane.

2.2.2. Enzyme activity (AMPase)

The ecto AMPase activity was determined on living cells

trough a temporal course of inorganic phosphate production

in presence and absence of AMP as a substrate following a

protocol described by Martinez-Martinez et al. [24]. Total

AMPase was assayed in 3 mL assay medium (125 mM

NaCl, 5 mM KCl, 10 mM MgCl2, 0.5 mM glucose, 14 mM

Tris, 1 mM CaCl2, pH 7.4 with HCl 1N) containing 9�106

lymphocytes mL�1 and incubated at 37 8C. AMPase activity

was measured by adding 1 mM final AMP concentration,

and negative controls were preformed in which no substrate

was added to the incubation medium. Aliquots of 150 AL of

the mixture were withdrawn at 0, 60, 120, and 180 min and

immediately sonicated on ice for 30 s (50% power) to

disrupt cells. Sixty microliter of 50% ice-cold trichloroacetic

acid (TCA) was immediately added to stop the enzyme

activity and to precipitate the protein. Subsequently, the

Fig. 1. Calibration curve for Pi determination. Phosphate was added as

NH4H2PO4 and complexed with (NH4)6Mo7O24. Phosphomolibdic com-

plexes were reduced with Fiske–Subbarow, and absorbance at 660 nm was

recorded. The values for the curve were e=337.7 mM�1, intersect in Abs

was 0.013F0.004 (n=17; r=0.9999). (Microplate Reader BIO-RAD).

A. Martınez-Martınez et al. / Microchemical Journal 81 (2005) 92–9794

homogenized cells were centrifuged for 3 min at 10,000�g

and 4 8C. One hundred microliter of clear supernatant was

transferred to a microplate and Pi was quantified by adding

240 AL ammonium-heptamolybdate solution (0.5% w/v

(NH4)6Mo7O24, 2% w/v sodium dodecyl sulfate (SDS), 0.5

M H2SO4). Phosphomolybdic complexes were reduced with

10 AL of 0.16 g mL�1 Fiske–Subbarow solution in

deionized water [25]. After 15 min of incubation, the

absorbance was recorded at 660 nm on a Microplate Reader

BIO-RAD. No spontaneous hydrolysis of AMP was

detected under the conditions tested in all assays (n=17).

The calibration curve was prepared from a stock solution

of NH4H2PO4 giving a Pi concentration of 1 mg mL�1 and

two 1:10 serial dilutions were made to reach 0.01 mg Pi

mL�1. After that, 0, 10, 20, 40, 60, 80, and 100 AL were

added to their respective wells in the microplate, adjusting

the volume to 100 AL with MQ water. Ammonium-

heptamolybdate and Fiske–Subbarow solutions were added

as previously described to develop the phosphomolybdate

complexes that absorb at 660 nm.

2.2.3. Drug treatments

Basal AMPase activity was determine with 9�106

lymphocytes mL�1 in presence of 1 mM AMP. To stimulate

PCK, 10 ng mL�1 PMA was added to the lymphocyte

suspension and incubated for 10 min before adding 1 mM

AMP. The inhibitors MADP and levamisole were added

immediately before starting the time course, to reach a final

concentration in the assay media of 400 AMMADP and 500

AM levamisole. All experiments were carried out at least

three times in duplicate. Results were analyzed by two tails

unpaired T-test.

3. Results

3.1. Pi determination

The method for Pi determination was a modification of

the Fiske–Subbarow colorimetric detection of inorganic

phosphate [24,25]. The calibration curve for the microplate

assay was from 0 to 322.5 AM Pi (Fig. 1). The first point

tested was, 32.25 AM. The signal noise at 0 AM Pi was

0.0044F0.00164 AU. The inferior detection limit (IDL) was

set as the average of the absorbance value plus 3 standard

deviations. In the conditions tested, the IDL for the

instrument was 0.0093 AU, the IDL for the assay was

0.025 AU, and the molar extinction coefficient for the

phosphomolybdic complex was 377.7 AM�1 path cell �1.

Since this method needs only 0.1 mL sample, Pi can be

detected as low as 0.94 nmol Pi assay�1.

3.2. Lymphocytes purification

The lymphocytes purification procedure has shown to

be gentle enough to render functional cells, and is useful

for functional studies where the integrity of the plasma

membrane is required [22]. Unbroken lymphocytes were

evaluated by the trypan blue exclusion assay, which

allows the study of two parameters: (1) lymphocyte

dimension and shape; and (2) direct counting. Data about

lymphocyte dimension and shape are related to either

swelling or shrinkage. This reflects lymphocyte stress to

osmotic media and/or leakage of metabolites as a

consequence of mechanical damage to the plasma

membrane. On the other had, direct counting on the

Neubauer chamber has the advantage of providing the

proportions of intact lymphocytes versus dead or damaged

lymphocytes. Lymphocytes were also tested for its long-

term integrity. Purified lymphocytes that were kept at 4

8C for 24 h were evaluated by the trypan blue exclusion

assay and showed percentages of intact cells very similar

to those of freshly purified lymphocytes (90–95%).

Nonetheless, the assays were always performed with

freshly purified lymphocytes counted with an ADVIA

channel analyzer as described in the Materials and

methods section.

3.3. AMP is dephosphorylated when added to unbroken

lymphocytes

As stated above, lymphocytes were purified and no

leakage on their plasma membrane was detected. In such

circumstance, AMP cannot freely diffuse inside the cell

because of the charge impedance imposed by the phosphate

group. When incubated in the presence of 1 mM AMP,

lymphocytes produced inorganic phosphate at a rate of

8.1F4.4 nmol Pi million cells�1 h�1 at 37 8C (n=14). Pi

production was linear for up to 3 h (Fig. 2). On the other

hand, control lymphocytes incubated with all components

Fig. 2. AMPase activity on living human lymphocytes. AMP hydrolysis

was quantified by the Pi production (nmol Pi h�1 million cells�1). Intact

human lymphocytes obtained from circulating blood were incubated in the

presence of 1 mM AMP. After 0, 1, and 3 h, an aliquot of 10 AL of cell

suspension was extracted and assayed for Pi with a Fiske–Subbarow

reducer adapted to a microassay. Each point represents at least three

independent experiments. Bars represent standard errors.

A. Martınez-Martınez et al. / Microchemical Journal 81 (2005) 92–97 95

except AMP had a Pi production of 0.23F0.26 nmol Pi

million cells�1 h�1 (n=7). In each case, the Pi production of

control lymphocytes was subtracted from that of AMP-

containing cells.

3.4. The main AMPase activity on lymphocyte surface is

eNT

Fig. 3 shows the effect of h-methylene-ADP (MADP) (a

specific eNT inhibitor that acts by binding on the catalytic

site), on Pi production. Lymphocyte suspensions were

Fig. 3. AMPase activity on living human lymphocytes is eNT. Intact human

lymphocytes obtained from circulating blood were incubated with 1 mM

AMP in the presence and absence of MADP, a specific eNT inhibitor. After

1 h incubation, 10 AL of cell suspension was assayed for Pi production as

detailed under Materials and methods, n=4. Bars represent standard errors.

incubated with 1 mM AMP in the presence of 400 AMMADP. In such conditions, the AMPase activity decreased

from 9.47F5.99 to 2.43F0.88 nmol Pi million cells�1 h�1

(n=7; pb0.05). This suggests that 63% of the total AMPase

activity on the lymphocyte surface is due to eNT. Nonethe-

less, when lymphocytes were incubated in the presence of 1

mM AMP and 400 AM levamisole, the AMPase activity

decreased by only 17% (n=7) (Fig. 4). Since levamisole is a

broad spectrum inhibitor of unspecific phosphatases, includ-

ing alkaline phosphatases but not eNT, these results suggest

that the main AMPase activity on lymphocytes surface is

due to eNT.

3.5. Lymphocyte eNT is activated after PMA incubation

Phorbol 12-myristate 13-acetate (PMA) is an agonist of

the DAG conserved binding domain of PKC [21]. In

order to test the cross talk between eNT and PKC,

lymphocytes were incubated in presence of 10 ng mL�1

PMA for 10 min at 37 8C; after that, AMPase activity

was assessed as described in the Materials and methods

section. Fig. 4 shows that PMA treatment increased the

AMPase value from 6.63F0.97 to 12.08F2.69 nmol Pi

million cells�1 h�1 (n=7; pb0.05). This means that the

AMPase value increased about 182%. In order to discard the

participation of alkaline phosphatase as the PMA-responsive

ectoenzime, the AMPase assay included 400 AM of

levamisole. In this case, the AMPase activity decreased

from 12.08F2.69 to 11.14F0.99 nmol Pi million cells�1

h�1 (Fig. 4). This result suggests that only 8% of the

stimulated AMPase activity is probably due to alkaline

phosphatase.

Fig. 4. PMA activates lymphocyte eNT. Lymphocytes were preincubated

for 10 min with PMA and then AMP was added in presence or absence of

levamisole. PMA induced a 1.8 fold increase in AMPase activity (n=7;

pb0.05) that was not sensitive to levamisole (n=7; pb0.05). Different

numbers (1 and 2) indicate significant differences ( pb0.05). Bars represent

standard errors.

A. Martınez-Martınez et al. / Microchemical Journal 81 (2005) 92–9796

4. Discussion

4.1. AMP is hydrolyzed by intact lymphocytes

It is well documented that nucleotides cannot freely

diffuse inside cells because the phosphate charges make

them impermeable to the plasma membrane. For instance,

adenosine and inorganic phosphate from AMP are incorpo-

rated into the cell trough specific transporters after AMP is

dephosphorylated [26]. In a first attempt to quantify ecto-

AMPase activity, we tried to measure the phosphate present

in the assay media after different incubation times. However,

we failed to measure Pi production under this protocol,

probably due to phosphate transporters present in plasma

membranes that might have incorporated phosphate as soon

as it was produced by the dephosphorylation of AMP [26].

Thus, it was necessary to measure total phosphate by

disrupting the cells using sonication. In such conditions, the

basal value of Pi was shown to be 0.23F0.26 nmol Pi. million

cells�1 h�1. This value reflects the endogenous Pi produced

by lymphocyte because the extraction and the enzyme assay

media were phosphate free. Notwithstanding, this value can

be indicative of the internal hydrolysis of endogenous pool of

several phosphorylated compounds. Up to the knowledge of

the authors, there is no reference in the literature to compare

this Pi production by purified lymphocytes.

4.2. The main AMPase activity on lymphocyte surface is eNT

Previous studies have shown that eNT/CD73 is

expressed on ~15% of peripheral blood lymphocytes when

screened for its immunoreactivity CD73 [2,3]. CD73 has

been shown to be the same molecular entity to e5NT with a

gene identification number 4505466 (NM_002526) in the

NIH GeneBank (http://www.ncbi.nlm.nih.gov/entrez).

Besides eNT, other molecules also present in lymphocyte

surface possess phosphohydrolase activity, such as NTP-

diphosphohydrolases (eNTPDases), also known as CD39

[27]. However, MADP has proven to be a specific inhibitor

for eNT [1] and combined with the use of AMP as substrate

(Fig. 2), the eNT activity can be dissected. On the other

hand, the observed value for eNT activity (8.1F4.4 nmol

million cells�1 h�1) is in accordance to a value reported by

Henttinen et al. [28] from a radioactivity assay (8.3F1.1

nmol million cells�1 h�1). Since these assays where

performed on intact lymphocytes, we discard the contribu-

tion of cytosolic nucleotidases, which can be extruded upon

cell damage [29,30]. Thus, the results suggests that the eNT

activity observed represents the AMPase activity of the 15%

CD73 positive lymphocytes, which are expected to be found

among the total cells assayed.

4.3. Lymphocyte eNT is PMA activated

Lymphocyte eNT/CD73 is a glycosylphosphatidylinosi-

tol (GPI)-linked ecto enzyme, susceptible to phosphatidy-

linositol-specific phospholipase C (PI-PLC) cleavage

[15,31]. The importance of GPI involvement in trans-

membrane signaling is well documented [32]. In addition,

it has also been demonstrated that e5NT/CD73 gene

contains in its promoter the TCF/LEF consensus binding

site controlled by the Wnt/h-catenin signaling that is

linked to the metabolism of adenosine [33]. However, a

signaling link between eNT/CD73 and PKC has not been

demonstrated in these cells. To the knowledge of the

authors, this is the first study that shows the cross-talk

between PKC and eNT in human lymphocytes. In canine

muscle heart, it has been shown that PKC activation

induced an increment in NT activity, suggesting the

involvement of a PKC-dependent NT phosphorylation

[34,35]. In a physiological scheme, the adenosine produced

by eNT could bind to metabotropic P2Y receptors and

couple to PKC through either phosopholipase C or D,

PKC in turn can feed-back the signal to eNT, which is

activated [36,37]. The increase of about 182% in AMPase

activity observed in lymphocytes after treatment with the

PKC agonist PMA (Fig. 3) strongly suggests that a cross

talk between PKC and eNT is present.

Acknowledgments

Laura A. de la Rosa acknowledges UACJ and

SAGARPA-CONACyT (Sagarpa-2002-C01-0787) for

financial support. Dr. Alejandro Martinez-Martinez

acknowledges the financial support from UACJ and CON-

ACyT (J41836-Q). Anagel G Dıaz-Sanchez received a

fellowship from CONACyT. Dr. Gardea-Torresdey

acknowledges the financial support from the National

Institutes of Health (Grant S06GM8012-33), the National

Institute of Environmental Health Sciences (Grant

R01ES11367-01), and the Dudley family for the Endowed

Research Professorship in Chemistry. Drs. Gardea-Torres-

dey and Peralta-Videa also acknowledge the financial

support form HBCU/MI ETC. Guadalupe de la Rosa

acknowledges CONACyT (grant 131996) and the Univer-

sidad de Guanajuato.

References

[1] H. Zimmermann, Biochem. J. 285 (1992) 345.

[2] L.F. Thompson, J.M. Ruedi, A. Glass, G. Moldenhauer, P. Moller,

M.G. Low, M.R. Klemens, M. Massaia, H. Lucas, Tissue Antigens 35

(1990) 9.

[3] F. Rosi, A.B. Agostinho, F. Carlucci, L. Zanoni, B. Porcelli, E.

Marinello, P. Galieni, A. Tabucchi, Life Sci. 62 (1998) 2257.

[4] L. Airas, M. Salmi, S. Jalkanen, J. Immunol. 151 (1993) 4228.

[5] L. Airas, J. Hellman, M. Salmi, P. Bono, T. Puurunen, D.J. Smith, S.

Jalkanen, J. Exp. Med. 182 (1995) 1603.

[6] L. Airas, S. Jalkanen, Blood 88 (1996) 1755.

[7] L. Airas, M. Niemel7, M. Salmi, T. Puurunen, D. Smith, S. Jalkanen,

J. Cell Biol. 136 (1997) 421.

[8] F.W. Sunderman, Ann. Clin. Lab. Sci. 20 (1990) 123.

A. Martınez-Martınez et al. / Microchemical Journal 81 (2005) 92–97 97

[9] R.Y. Resta, Y. Yamashita, L.F. Thompson, Immunol. Rev. 161 (1998)

95.

[10] K. Synnestvedt, G.T. Furuta, J. Clin. Invest. 110 (2002) 993.

[11] L. Airas, J. Niemel7, S. Jalkanen, J. Immunol. 165 (2000) 5411.

[12] M. Massaia, L. Perrin, A. Bianchi, J. Ruedi, C. Attisano, D. Altieri,

G.T. Rijkers, L.F. Thompson, J. Immunol. 145 (1990) 1664.

[13] M. Moran, M.C. Miceli, Immunity 9 (1998) 787.

[14] L.F. Thompson, J.M. Ruedi, J. Immunol. 142 (1989) 1518.

[15] L.F. Thompson, J.M. Ruedi, A. Glass, M.G. Low, A.H. Lucas, J.

Immunol. 143 (1989) 1815.

[16] K. Simons, E. Ikonen, Nature 387 (1997) 569.

[17] J. Bohuslav, V. Horejsi, C. Hansmann, J. Stockl, U.H. Weidle, O.

Majdic, I. Bartke, W. Knapp, H. Stockinger, J. Exp. Med. 181 (1995)

1381.

[18] K. Krauss, P. Altevogt, J. Biol. Chem. 274 (1999) 36921.

[19] L.F. Thompson, J.M. Ruedi, M.G. Low, L.T. Clement, J. Immunol.

139 (1987) 4042.

[20] A.E. May, S.M. Kanse, L.R. Lund, R.H. Gisler, B.A. Imhof, K.T.

Preissner, J. Exp. Med. 188 (1998) 1029.

[21] S. Carrasco, I. Merida, Mol. Biol. Cell 15 (2004) 2932.

[22] L.A. de la Rosa, A.G. Cabado, M.A. Botana, J.I. Vidal, M.R. Vieytes,

L.M. Botana, Pflqgers Arch. Eur. J. Phys. 280 (2001) 518.

[23] A. Boyum, Scand. J. Clin. Lab. Invest. (Suppl. 97) (1968) 7.

[24] A. Martinez-Martinez, C. Flores-Flores, F.J. Campoy, E. Munoz-

Delgado, C. Fini, C.J. Vidal, Biochim. Biophys. Acta 1386 (1998) 16.

[25] C.H. Fiske, Y. Subbarow, J. Biol. Chem. 66 (1925) 375.

[26] P. Bottger, L. Pedersen, J. Biol. Chem. 277 (2002) 42741.

[27] H. Zimmermann, Nat. Med. 5 (1999) 987.

[28] T. Henttinen, S. Jalkanen, G.G. Yegutkin, J. Biol. Chem. 278 (2003)

24888.

[29] E.R. Lazarowski, R.C. Boucher, T.K. Harden, J. Biol. Chem. 275

(2000) 31061.

[30] C.E. Sorensen, I. Novak, J. Biol. Chem. 276 (2001) 32925.

[31] M.T. Lehto, F.J. Sharom, Biochem. J. 332 (1998) 101.

[32] R. Resta, L.F. Thompson, Cell. Signal. 9 (1997) 131.

[33] J. Spychala, J. Kitajewski, Exp. Cell Res. 296 (2004) 99.

[34] M. Kitakaze, M. Hori, T. Morioka, T. Minamino, S. Takashima, H.

Sato, Y. Shinozaki, M. Chujo, H. Mori, M. Inoue, Circulation 89

(1994) 1237.

[35] K. Przyklenk, R.A. Kloner, Basic Res. Cardiol. 91 (1996) 41.

[36] D.M. Cohen, S.R. Gullans, W.W. Chin, J. Clin. Invest. 97 (1996)

1884.

[37] E.K. Iliodromitis, T. Miki, G.S. Liu, J.M. Downey, M.V. Cohen, D.T.

Kremastinos, J. Mol. Cell. Cardiol. 30 (1998) 2201.