Metabolic pathways for activation of the antiviral agent 9-(2- phosphonylmethoxyethyl)adenine in human lymphoid cells

Article abstract of DOI:10.1128/AAC.39.10.2304

9-(2-Phosphonylmethoxyethyl)adenine (PMEA), the acyclic phosphonate analog of adenine monophosphate, is a promising antiviral drug with activity against herpesviruses, Epstein-Barr virus, and retroviruses, including the human immunodeficiency virus. In order to be active, it must be converted to the diphosphate derivative, the putative inhibitor of viral DNA polymerases. The metabolic pathway responsible for activation of PMEA is unclear. The metabolism of PMEA was investigated in human T-lymphoid cells (CEMss) and a PMEA-resistant subline (CEMss(r-1)) with a partial deficiency in adenylate kinase activity. Experiments with [³H]PMEA showed that extracts of CEMss phosphorylated PMEA to its mono- and diphosphate in the presence of ATP as the phosphate donor. No other nucleotides or 5-phosphoribosyl pyrophosphate displayed appreciable activity as a phosphate donor. Subcellular fractionation experiments showed that CEMss cells contained two nucleotide kinase activities, one in mitochondria and one in the cytosol, which phosphorylated PMEA. The PMEA-resistant CEMss mutant proved to have a deficiency in the mitochondrial adenylate kinase activity, indicating that this enzyme was important in the phosphorylation of PMEA. Other effective antiviral purine phosphonate derivatives of PMEA showed a profile of phosphorylating activity similar to that of PMEA. By comparison, phosphorylation of the pyrimidine analog (S)-1-(3-hydroxy-2- phosphonylmethoxypropyl) cytosine proceeded by an enzyme present in the cytosol. We conclude from these studies that adenylate kinase which has been localized in the intermembrane space of mitochondria is the major route for PMEA phosphorylation in CEMss cells but that another hitherto unidentified enzyme(s) present in the cytosol may contribute to the anabolism of the phosphonates.

Full text of DOI:10.1128/AAC.39.10.2304

ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Oct. 1995, p. 2304–2308
Vol. 39, No. 10
0
066-4804/95/$04.00ϩ0
Copyright ᭧ 1995, American Society for Microbiology
Metabolic Pathways for Activation of the Antiviral Agent 9-(2-
Phosphonylmethoxyethyl)Adenine in Human Lymphoid Cells
1
1
1
1,2
BRIAN L. ROBBINS, JACK GREENHAW, MICHELE C. CONNELLY, AND ARNOLD FRIDLAND
*
1
Department of Infectious Diseases, St. Jude Children’s Research Hospital, Memphis, Tennessee 38105,
2
and Department of Pharmacology, University of Tennessee, Memphis, Tennessee 38101
Received 25 April 1995/Returned for modification 21 June 1995/Accepted 9 August 1995
9-(2-Phosphonylmethoxyethyl)adenine (PMEA), the acyclic phosphonate analog of adenine monophosphate,
is a promising antiviral drug with activity against herpesviruses, Epstein-Barr virus, and retroviruses, includ-
ing the human immunodeficiency virus. In order to be active, it must be converted to the diphosphate
derivative, the putative inhibitor of viral DNA polymerases. The metabolic pathway responsible for activation
of PMEA is unclear. The metabolism of PMEA was investigated in human T-lymphoid cells (CEMss) and a
r-1
PMEA-resistant subline (CEMss ) with a partial deficiency in adenylate kinase activity. Experiments with
3
[
H]PMEA showed that extracts of CEMss phosphorylated PMEA to its mono- and diphosphate in the
presence of ATP as the phosphate donor. No other nucleotides or 5-phosphoribosyl pyrophosphate displayed
appreciable activity as a phosphate donor. Subcellular fractionation experiments showed that CEMss cells
contained two nucleotide kinase activities, one in mitochondria and one in the cytosol, which phosphorylated
PMEA. The PMEA-resistant CEMss mutant proved to have a deficiency in the mitochondrial adenylate kinase
activity, indicating that this enzyme was important in the phosphorylation of PMEA. Other effective antiviral
purine phosphonate derivatives of PMEA showed a profile of phosphorylating activity similar to that of PMEA.
By comparison, phosphorylation of the pyrimidine analog (S)-1-(3-hydroxy-2-phosphonylmethoxypropyl) cy-
tosine proceeded by an enzyme present in the cytosol. We conclude from these studies that adenylate kinase
which has been localized in the intermembrane space of mitochondria is the major route for PMEA phos-
phorylation in CEMss cells but that another hitherto unidentified enzyme(s) present in the cytosol may
contribute to the anabolism of the phosphonates.
9
-(2-Phosphonylmethoxyethyl)adenine (PMEA) is a proto-
responsible for this activation to PMEApp remains unclear.
We have recently reported (20) that the development of resis-
tance to PMEA in the human lymphoid cell line CEMss is
associated with an enhanced drug efflux and a reduced phos-
phorylation of PMEA to the diphosphate PMEApp. This latter
effect is associated with a ϳ50% decrease in the activity of
adenylate kinase (AK) (ATP:AMP phosphotransferase, EC
2.7.4.3) compared to that of the parental cells. These results
implicate AK as a key enzyme in the activation of PMEA to its
active form.
At least three different AK isoenzymes have been identified
in mammalian cells (12, 14, 21, 24–26). These isoenzymes differ
in amino acid sequence and tissue distribution (12, 24, 30).
AK1 (originally called myokinase) exists in the cytosol and is
the primary enzyme in skeletal muscle, brain, and erythrocytes,
while AK2 is localized in the outer membrane spaces of mito-
chondria and is found primarily in liver, kidney, spleen, and
heart (24). AK3, which has specificity for GTP as a phosphate
donor, is more ubiquitous in tissues and is found in the mito-
chondrial matrix (24, 26).
type of a new class of compounds, several of which are shown
in Fig. 1. The acyclic nucleoside phosphonates demonstrate
potent activity against a wide range of DNA and RNA viruses.
PMEA is active against several herpesviruses, including herpes
simplex virus types 1 and 2 (HSV-1 and HSV-2, respectively)
both in vitro and in vivo (7, 9, 10), Epstein-Barr virus (15), and
several retroviruses, including the human immunodeficiency
viruses (HIVs) (2–4, 8, 18). The in vivo antiviral efficacy of
PMEA has been documented in various animal retroviral mod-
els of AIDS, including feline immunodeficiency virus infections
in cats, murine leukemia/sarcoma virus infections in mice, and
simian immunodeficiency virus in macaques (3–5, 9, 11, 16).
PMEA has been shown to be particularly effective in pre-
exposure prophylaxis of simian immunodeficiency virus infec-
tions in macaques and found to completely suppress viremia
and disease symptoms in all the treated animals (28). PMEA is
therefore of interest as a potential drug for the treatment of
HIV infection and some of the opportunistic infections asso-
ciated with AIDS. PMEA and its lipophilic prodrug bispivaloy-
loxymethyl(bispom)-PMEA have currently entered phase I/II
clinical trials as treatment for HIV infections in AIDS patients
The isoforms of AK and their subcellular distribution in
lymphoid cells have not been determined. In this report, we
have determined and compared the subcellular distribution of
AK activities in PMEA-resistant and -sensitive human lym-
phoid cells (CEMss) and examined the specificity of the en-
zyme for PMEA and related phosphonate derivatives.
(
6).
The antiviral efficacy of PMEA is based on the capacity of its
diphosphate derivative (PMEApp) to preferentially inhibit vi-
ral DNA replication with relative sparing of host DNA synthe-
sis (2, 7, 8). Conversion of PMEA to this active metabolite is
carried out by cellular enzymes; however, the exact pathway
MATERIALS AND METHODS
Chemicals. All of the phosphonate analogs (for structures, see Fig. 1)
PMEA, (S)-1-(3-hydroxy-2-phosphonylmethoxypropyl) adenine (HPMPA),
*
Corresponding author. Mailing address: Department of Infectious
Diseases, St. Jude Children’s Research Hospital, 332 N. Lauderdale
Blvd., Memphis, TN 38105. Phone: (901) 495-3459. Fax: (901) 495-
(
R)-9-(2-phosphonylmethoxypropyl) adenine (PMPA), 9-(phosphonoylme-
thoxyethyl)-2,6-diaminopurine (PMEDAP), 9-(2,5-dihydro-5(phosphonome-
thoxy)-2-furanyl) adenine (D4AP), (R)-9-(2-phosphonylmethoxypropyl)-2,6-
1
311. Electronic mail address: Fridland@MBCF.stjude.org.
2304

VOL. 39, 1995
PHOSPHORYLATION OF PMEA IN CEMss CELLS
2305
ethyleneimine-cellulose plates were purchased from Baxter Scientific
Products (McGraw Park, Ill.). 5,5Ј-dithio-bis(2-nitrobenzoate)(DTNB), p-
hydroxy-mercuribenzoic acid (pHMB), pyruvate kinase, phosphoenol pyru-
vate (PEP), Trizma base, dithiothreitol (DTT), 5-phosphorylribose-1-pyro-
phosphate (PRPP), bovine serum albumin, and nucleotides were purchased
from Sigma Chemical Co. (St. Louis, Mo.). Media used for cell cultures were
from Biowhittaker Inc. (Walkersville, Md.). Protease inhibitors were ob-
tained as a kit from Boehringer Mannheim Corp. (Indianapolis, Ind.).
Cells and cell fractionation for the determination of AK distribution. CEMss
cells were obtained from Peter Nara through the AIDS Research and Reference
Reagent Program, Division of AIDS, National Institute of Allergy and Infections
r-1
Diseases. The parental and PMEA-resistant CEMss lines were grown as de-
scribed previously (20). The resistant cells, when grown in the presence of
PMEA, were subcultured into drug-free medium for at least 1 week before
lysates were prepared. Cells in log phase were collected by centrifugation,
washed twice with phosphate-buffered saline, and fractionated by the method of
Richter (19). The cells were suspended at 2 ϫ 108 cells per ml in MSH buffer
(
210 mM mannitol, 70 mM sucrose, 5 mM HEPES [N-2-hydroxyethylpiperazine-
NЈ-2-ethanesulfonic acid], 1 mM EDTA, pH 7.4) and homogenized in a glass
homogenizer for 12 strokes. Protease inhibitors (1 mM Pefabloc, 1 ␮g/ml apro-
tinin, 10 ␮g/ml leupeptin, 10 ␮g/ml pepstatin A, 1 mM DTT) were added, and the
cell lysate was centrifuged for 10 min at 800 ϫ g at 4ЊC to sediment the nuclei.
The resulting supernatant was centrifuged at 9,500 ϫ g for 10 min at 4ЊC to
sediment the mitochondria and at 105,000 ϫ g to sediment the microsomal
fraction. The mitochondrial pellet was washed twice with ice-cold MSH buffer
and resuspended to 800 ␮l in MSH buffer plus protease inhibitors. The suspen-
sion was subjected to three freeze-thaw cycles to lyse the mitochondria and then
stored in the presence of 10% glycerol and 3 mM DTT in liquid nitrogen until
use. The cytosolic fractions were aliquoted and stored in liquid nitrogen until use.
The microsomal pellets were suspended in 2 ml of MSH buffer and stored in
liquid nitrogen until use. The protein concentrations of all fractions were deter-
mined prior to use with the Bio-Rad protein assay kit.
Nucleotide kinase assays and the compartment markers. AMP and UMP
kinase activities were determined at 37ЊC in a reaction volume of 30 ␮l contain-
ing 0.05 M Tris buffer (pH 7.0), 3 mM DTT, 5 mM MgCl
2
, 0.5 ␮g of bovine
serum albumin per ml, 15 mM phospho(enol)pyruvate, pyruvate kinase (1 U)
and 0.05 ␮g of the mitochondrial or 1.5 ␮g of the cytosolic fraction. The reaction
3
was initiated with the appropriate radioactive substrate: [ H]AMP (50 ␮M, 0.3
3
Ci/mol) or [ H]UMP (50 ␮M, 0.3 Ci/mol). The reactions were stopped at 10 min
by addition of 10 ␮l of 0.3 M EDTA, and 4-␮l aliquots were immediately spotted
on polyethyleneimine-cellulose plates, and the various phosphorylated products
were separated by ascending thin-layer chromatography in 0.5 M ammonium
formate, pH 4.3. Standard lanes (4 ␮l of 5 mM mono-, di-, and triphosphate
derivatives) were prepared to permit identification of the nucleotide under UV
light after chromatography. The nucleotide spots were cut out from the plates
and counted by liquid scintillation counting.
In the assays for PMEA and other phosphonate derivatives the reactions were
carried out in 0.1 ml (total volume) of the same components as above. Enzyme
concentrations were 0.1 mg of the mitochondrial fraction or 1 mg of the cytosolic
fraction. The reaction was initiated by addition of [14C]PMEA (2 mM, 2.5
mCi/mol) or other phosphonate derivatives as indicated, and the mixture was
incubated at 37ЊC for 2 h. The reactions were terminated by the addition of 0.25
ml of 98.5% ice-cold methanol at pH 7. The formation of PMEA monophos-
phate (PMEAp) and PMEApp were determined by anion-exchange high-pres-
sure liquid chromatography (HPLC) using a series of linear gradients from 0.005
M (buffer A) to 0.6 M (buffer B) ammonium phosphate buffer. The gradient was
15 min from 0 to 30% buffer B, 10 min from 30 to 40% buffer B, and then 22 min
at 100% buffer B with a flow rate of 1.5 ml/min.
FIG. 1. The structural formulas of the phosphonates PMEA, PMEDAP,
PMPA, PMPDAP, HPMPC, HPMPA, and D4AP.
diaminopurine (PMPDAP), and (S)-1-(3-hydroxy-2-phosphonylmethoxy-
propyl)cytosine (HPMPC) and the phosphorylated products of these com-
pounds were kindly provided by Norbert Bischofberger, Gilead Sciences
1
4
(
Foster City, Calif.). The radioactive compounds PMEA [adenine-8- C],
3 3
PMEDAP [adenine-8- H], PMPA [adenine-2,8- H], PMPDAP [adenine-8-
3
3
3
H], HPMPA [adenine-2,8- H], HPMPC [cytidine-5- H], and D4AP [ade-
3
nine-8- H] were obtained from Moravek Biochemicals (Brea, Calif.). Poly-
r-1
a
TABLE 1. Subcellular distribution of enzyme activities in wild-type and PMEA-resistant CEMss cells
pmol/min/106 cells (% of cellular activity)
Cell type
Substrate
AMP
Mitochondria
Cytosol
Microsomes
CEMss
CEMss
CEMss
CEMss
CEMss
CEMss
CEMss
CEMss
CEMss
CEMss
CEMss
CEMss
1,836 Ϯ 155(71)
867 Ϯ 437(62)
0.50 Ϯ 0.18(56)
0.26 Ϯ 0.09(46)
2.1 Ϯ 0.76(0.5)
1.7 Ϯ 0.17(0.7)
0.86 Ϯ 0.23(1)
0.47 Ϯ 0.38(1)
990 Ϯ 300(100)
1,310 Ϯ 350(100)
2,180 Ϯ 620(100)
2,610 Ϯ 420(100)
679 Ϯ 267(26)
470 Ϯ 115(34)
0.39 Ϯ 0.074(44)
0.30 Ϯ 0.04(54)
356 Ϯ 60.0(95)
255 Ϯ 72(97)
72.8 Ϯ 15.83(99)
71.9 Ϯ 8.9(99)
Ͻ0.1
76 Ϯ 10(4)
52 Ϯ 11(4)
Ͻ0.1
r-1
r-1
r-1
r-1
r-1
r-1
PMEA
Ͻ0.1
UMP
12 Ϯ 5.80(3.5)
6.2 Ϯ 0.36(2.3)
ND
Adenosine
Cytochrome c
Glutamate
ND
ND
Ͻ0.1
ND
Ͻ0.1
Ͻ0.1
ND
ND
a
Fractions were prepared and assayed as described in Materials and Methods. The data are the means of two experiments Ϯ the range. ND, not determined.

2
306
ROBBINS ET AL.
ANTIMICROB. AGENTS CHEMOTHER.
a
TABLE 2. Inhibition of enzymatic activity by sulfhydryl reagents
soluble enzymes from mitochondrial outer membrane space
and matrix, respectively, were negligible (Ͻ0.1%) in the cyto-
plasmic supernatant (Table 1).
%
recovery of enzyme activity
Cellular
fraction
Substrate
DTNB (100 ␮M)
pHMB (100 ␮M)
Effect of sulfhydryl reagents on nucleotide kinases. AK1 and
UMP kinase are both highly susceptible to inactivation by
sulfhydryl reagents such as pHMB and DTNB, whereas both
AK2 and AK3 are insensitive to these inhibitors (12–14, 21).
To determine possible differences between the mitochondrial
and cytosolic enzyme activities, the effects of these two inhib-
itors on the phosphorylation of the various compounds by the
subcellular fractions was studied. As shown in Table 2, the
activity of UMP kinase was almost totally inhibited by DTNB
and pHMB at a concentration of 100 ␮M, whereas essentially
complete activity was retained for adenylate kinase and PMEA
phosphorylation activities in both the mitochondrial and cyto-
solic fractions of these cells in the presence of these two in-
hibitors. In contrast, when AK1 (obtained commercially) was
tested, complete enzyme inhibition was seen under the same
conditions with these two sulfhydryl inhibitors. These results
indicated that the phosphorylation of PMEA and AMP in both
the mitochondria and cytosol of CEMss cells was not due to
either AK1 or UMP kinase.
Comparison of the substrate specificity of mitochondrial
and cytosolic enzymes. The phosphate donor specificity of var-
ious nucleoside triphosphates for the phosphorylation of the
various substrates in the mitochondria and cytosol was exam-
ined. As shown in Table 3, ATP and dATP both acted as
phosphate donors for adenylate kinase and UMP kinase, but
only ATP acted as a phosphate donor for PMEA phosphory-
lation for both the mitochondrial and cytosolic enzymes. Rel-
atively little activity was obtained with any of the other triphos-
phates tested with either the cytosolic or mitochondrial enzyme
activities. These results show that none of the phosphorylating
activities represent AK3, since this enzyme has specificity for
GTP as the phosphate donor.
Mitochondria
Cytosol
AMP
PMEA
AMP
PMEA
UMP
100 Ϯ 5.0
94 Ϯ 4.0
88 Ϯ 6.0
110 Ϯ 8.0
Ͻ1
90 Ϯ 3.0
88 Ϯ 1.0
96 Ϯ 4.0
120 Ϯ 23
Ͻ1
a
Fractions were prepared and assayed as described in Materials and Methods
with the following modifications. Assay mixtures and controls were prepared by
omitting DTT and adding the indicated concentration of DTNB or pHMB 15
min prior to the addition of label. The data are the means of two experiments Ϯ
the range.
Cytochrome c oxidase, a mitochondria marker of the outer membrane, was
assayed essentially by the method of Storrie (23); oxidation of the reduced
cytochrome c was measured at 550 nm, and the specific activity was calculated
using an extinction coefficient of 19.6. Glutamate dehydrogenase, a mitochon-
drial matrix marker, was measured by the method of Schmidt (22) by following
the NADH oxidation and using an extinction coefficient of 6.22. AK was mea-
sured by a radioactive assay essentially as described in earlier studies (29).
Data analysis. Kinetic parameters were calculated using a Lineweaver-Burk
double reciprocal plot. Enzfitter, a program by Elsevier-Biosoft, was used to fit
a line to the data by using weighted least squares linear regression.
RESULTS
PMEA phosphorylation activity of parental and resistant
CEMss cells. We compared the effect of ATP and PRPP as
phosphate donors on phosphorylation of PMEA in cell extracts
r-1
of both the wild-type and PMEA-resistant CEMss cells.
CEMss cell extracts incubated with 5 mM ATP and 2 mM
PMEA formed 12.3 and 5.8 pmol of PMEAp and PMEApp,
respectively, per min per mg of protein, while extracts of the
resistant cells generated PMEAp and PMEApp half as effi-
ciently as the wild type (data not shown). This decrease in
PMEA phosphorylation paralleled the decrease in AMP ki-
nase activity we have found in the CEMss variant (20). In
contrast, there was no detectable phosphorylation of PMEA to
PMEApp when PRPP was substituted for ATP as the phos-
phate donor. These results clearly demonstrate the role of
AMP kinase in the phosphorylation of PMEA.
Table 4 shows the kinetic properties of adenylate kinase with
AMP and PMEA as the phosphate acceptors. The substrate
specificity for PMEA for the mitochondrial adenylate kinase
was ϳ73-fold lower than that for the natural substrate AMP
and the Vmax of PMEA phosphorylation for the mitochondrial
enzyme was at least 7,000-fold lower than for AMP. No major
differences were found in the specificity of the cytosolic en-
zyme, except that the specific activity for the latter is consid-
erably lower than in the mitochondrial fraction.
Intracellular distribution of AK activity. Cells were fraction-
ated, and the subcellular localization of AK and PMEA phos-
phorylating activity(ies) was determined in wild-type and re-
Substrate activity of related acyclic phosphonate. A number
of structural analogs of PMEA (Fig. 1) were tested as sub-
strates for the mitochondrial and cytosolic enzymes. Table 5
r-1
sistant CEMss cells. Approximately 70% of total AK activity
in the CEMss cells was in the mitochondria, and 26 and 4% of
the activity was in cytosolic and microsomal fractions, respec-
tively (Table 1). Of note was that in the resistant cells the
decrease in AK activity was mostly in the mitochondria. A
similar profile of phosphorylating activity was observed with
PMEA except that a somewhat higher (ϳ40%) phosphorylat-
ing activity was found in the cytosol, as with AMP. HPLC
analysis of the products also revealed that in the mitochondria
PMEAp accumulated to a greater extent that PMEApp, at a
ratio of 4:1, while in the cytosol both PMEAp and PMEApp
were formed to about the same extent, albeit at a lower rate
than in the mitochondria.
TABLE 3. Phosphate donor specificity of nucleoside
monophosphate kinase activities in mitochondria and cytosol
Activity (% of ATP control)a
Acceptor-donor pair
Mitochondria
Cytosol
AMP-dATP
AMP-UTP
AMP-CTP
AMP-GTP
PMEA-dATP
UMP-dATP
UMP-UTP
UMP-CTP
UMP-GTP
48 Ϯ 6.0
4.2 Ϯ 1.4
3.8 Ϯ 0.5
4.0 Ϯ 0.6
Ͻ0.1
100 Ϯ 2.0
37 Ϯ 4.0
35 Ϯ 2.0
35 Ϯ 1.0
71 Ϯ 3.0
0.1 Ϯ 0.03
Ͻ0.1
Ͻ0.1
For comparison, we determined the subcellular expression
of two other enzymes, UMP kinase and adenosine kinase.
Essentially all the UMP kinase and adenosine kinase activity
was associated with the cytosol or microsomal fractions (Table
Ͻ0.1
73 Ϯ 6.0
6.5 Ϯ 2.3
6.9 Ϯ 2.2
6.9 Ϯ 0.2
1
). These results indicated that the enzyme activities associated
with the mitochondria in the CEMss cells are not likely to be
due to an adherent cytoplasmic contamination. Cytochrome c
oxidase and glutamate dehydrogenase, indicators of leakage of
a
Activities are expressed as the percentage of the activity of the individual
fractions with 2 mM ATP as the phosphate donor. Values are the average of two
experiments Ϯ the range.

VOL. 39, 1995
PHOSPHORYLATION OF PMEA IN CEMss CELLS
2307
a
TABLE 4. Kinetics of AMP kinase in mitochondria and cytosol
CEMss cells demonstrated both diminished adenylate kinase
activity and decreased phosphorylation of PMEA in the intact
Subcellular fraction and
substrate
V
max (pmol/min/
V
max/K
efficiency
m
r-1
m
K (mM)
cells. Also, the PMEA-resistant cell line CEMss was just as
␮g of protein)
sensitive as the wild type to 5-amino-4-imidazole carboxamide
riboside, a compound which is anabolized via PRPP synthetase
Mitochondria, AMP
Cytosol, AMP
Mitochondria,
PMEA
0.19 Ϯ 0.06
0.14 Ϯ 0.02
14.0 Ϯ 2.8
2,400 Ϯ 780
55 Ϯ 6.4
12,631
393
0.025
(20), further indicating that this enzyme is not a primary route
0.34 Ϯ 0.06
for PMEA phosphorylation.
Subcellular localization studies showed that the major phos-
phorylating activity for PMEA and AMP in CEMss cells was
associated with the mitochondria. Moreover, in the resistant
cells the deficiency in the adenylate kinase was associated with
the mitochondrial isozyme, clearly establishing this enzyme as
the anabolic route for PMEA in this human T-cell line. The Km
of AK2 for PMEA was about 74-fold higher than the Km for
the natural substrate AMP. Despite this low substrate activity,
the amount of phosphorylation associated with AK2 can ac-
count for the level of PMEA nucleotide accumulated in cells
after incubations with inhibitory concentrations of the drug.
Thus, extrapolation to conditions of 10 ␮M PMEA (an antivi-
ral concentration) results in the potential accumulation of 1.2
Cytosol, PMEA
15.9 Ϯ 2.0
0.04 Ϯ 0.01
0.003
a
Data are the means of two experiments Ϯ the range. The kinetics of AMP
were determined at enzyme concentrations of 0.05 ␮g and 1.5 ␮g for the mito-
chondrial and cytosolic fractions, respectively. PMEA kinetics were determined
at enzyme concentrations of 0.1 and 1 mg of mitochondrial and cytosolic frac-
tions, respectively.
shows that all acyclic phosphonate derivatives of adenine
tested (i.e., PMEDAP, PMPDAP, PMPA, D4AP, and HPM-
PA) behaved as substrates for both the mitochondrial and
cytosolic enzymes, but with some significant differences. The
3
-hydroxy-2-phosphonylpropyl derivative (HPMPA) was the
6
most active substrate for both the mitochondrial and cytosolic
enzymes. The order of phosphorylating activity was HPMPAϾ
PMPAϾD4APϾPMEAϾPMPDAPϾPMEDAP. By contrast,
the pyrimidine counterpart of HPMPA (designated HPMPC),
was a substrate for the cytosolic enzyme but not for the mito-
chondrial enzyme. It is noteworthy that the phosphorylation of
HPMPC in the cytosol, unlike PMEA, was susceptible to in-
activation by the inhibitors DTNB and pHMB (data not
shown).
pmol/10 cells by 8 h of incubation with PMEA, which com-
6
pares to the actual accumulation of 0.2 pmol/10 cells seen in
the intact cells during this time period.
About 40% of the total amount of measurable PMEA phos-
phorylating activity was found in the cytosolic fraction of
CEMss cells (Table 1). The identity of this enzyme was not
associated with either AK1 or NMK, since the compounds
DNTB and pHMB, two known inhibitors of these enzymes, did
not affect this phosphorylation. A possible explanation is that
this represents some leakage of AK2 from the mitochondria to
the cytosol during the fractionation; however, we did not find
any significant leakage of marker enzymes from the mitochon-
dria. Whatever the exact identity of this enzyme, our results
with the resistant cells indicate that the cytosolic activity is not
the primary route for anabolic activation of PMEA. Neverthe-
less, both enzyme activities may contribute to the metabolism
of the analog in intact cells, but additional studies are needed
to elucidate this question.
DISCUSSION
Previous studies have shown that the antiviral activity of
PMEA is not related to preferential activation by viral en-
zymes. Balzarini et al. (1, 2) and Nave et al. (17) studied the
phosphorylation of PMEA and various other antiviral nucle-
otide analogs such as 2Ј,3Ј-dideoxyadenosine monophosphate
and 9-␤-arabinofuranosyladenine monophosphate and found
that these compounds could be converted directly to their
diphosphate derivatives by the reverse action of PRPP syn-
thetase isolated from either Escherichia coli or rat by using
PRPP as phosphate donor. They suggested that PRPP syn-
thetase could be involved in the phosphorylation of these com-
pounds in intact cells. Our results, however, showed that PRPP
synthetase was not involved in the phosphorylation of PMEA
in the human T-lymphocytic cell line CEMss. The evidence was
derived from our demonstration that (i) PRPP did not substi-
tute for ATP as the phosphate donor in the phosphorylation of
PMEA in cell extracts and (ii) a PMEA-resistant subline of
AKs obviously occupy a central role in adenine nucleotide
metabolism in cells; however, their expression in different tis-
sues can vary considerably. Tissues such as liver, kidney,
spleen, and heart contain mainly AK2 in the mitochondria,
while the cytosolic isozyme AK1 is predominant in tissue such
as muscle, brain, and erythrocytes (24). In the present studies,
the major kinase for AMP and PMEA in the human T-lym-
phocytic cell line CEMss was found in the mitochondria. More-
over, none of the kinases of PMEA or AMP, found in the
cytosol or mitochondria, were inhibited by the sulfhydryls
DTNB or pHMB. It is concluded from these results that the
r-1
a
TABLE 5. Comparison of phosphorylating activities for phosphonate analogs in wild-type and CEMss cells
fmol/106 cells (% of total)
Compound
Mitochondria
Cytosol
CEMss
CEMssr-1
CEMss
CEMssr-1
PMEA
PMEDAP
814 Ϯ 58(70)
402 Ϯ 34(43)
138 Ϯ 10(38)
168 Ϯ 3(35)
132 Ϯ 24(15)
461 Ϯ 71(35)
293 Ϯ 80(36)
1,231 Ϯ 89(80)
353 Ϯ 17(30)
534 Ϯ 37(57)
964 Ϯ 304(66)
1,189 Ϯ 240(38)
937 Ϯ 110(40)
1,028 Ϯ 30(14)
1,140 Ϯ 408(100)
223 Ϯ 15(62)
316 Ϯ 41(65)
769 Ϯ 76(85)
870 Ϯ 103(65)
518 Ϯ 61(64)
316 Ϯ 72(20)
(
R)PMPDAP
R)PMPA
497 Ϯ 49(34)
(
1,929 Ϯ 362(62)
1,435 Ϯ 81(60)
6,453 Ϯ 1,925(86)
D4AP
(
S)HPMPA
S)HPMPC
(
1,720 Ϯ 86(100)
a
These analogs were assayed under essentially the same conditions as PMEA, described in Materials and Methods. Concentrations of 2 mM of the analogs were
incubated at 37ЊC for 2 h and separated in the same manner as PMEA. Data are the means of two experiments Ϯ the range.

2
308
ROBBINS ET AL.
ANTIMICROB. AGENTS CHEMOTHER.
CEMss cells contain little, if any, expression of AK1 and that
AK2 is the major enzyme operating in the phosphorylation of
adenosine monophosphate analogs. An important question
that remains for future investigation is whether a similar pro-
file of AK activity is found in other more therapeutically rel-
evant cells such as in peripheral blood lymphocytes and mono-
cytes/macrophages.
7. Cerny, J., S. Foster, and Y. Cheng. 1992. Cell-protecting effect against herpes
simplex virus-1 and cellular metabolism of 9-(2-phosphonylmethoxyethyl)ad-
enine in HeLa S3 cells. Mol. Pharmacol. 42:537–544.
8
. De Clercq, E. 1993. Broad-spectrum anti-DNA virus and anti-retrovirus
activity of phosphonylmethoxyalkylpurines and pyrimidines. Biochem. Phar-
macol. 42:963–972.
9. De Clercq, E., A. Holy, and I. Rosenberg. 1989. Efficacy of phosphonylme-
thoxyalkyl derivatives of adenine in experimental herpes simplex virus and
vaccinia virus infections in vivo. Antimicrob. Agents Chemother. 33:185–191.
0. De Clercq, E., A. Holy, I. Rosenberg, T. Sakuma, J. Balzarini, and P. C.
Maudgal. 1986. A novel selective broad-spectrum anti-DNA virus agent.
Nature (London) 323:464–467.
Several other phosphonate analogs which were previously
shown (8, 9, 16) to be active antiviral agents were also studied
as substrates for AK. The active enantiomer S-HPMPA (3-
hydroxy-2-phosphonomethoxypropyl) adenine and PMPA (2-
phosphonylmethoxypropyl) adenine were the two most effi-
cient substrates of the analogs tested for AK2 and for the
cytosolic enzyme activity. By contrast, the two 2,6-diaminopu-
rine derivative of PME (PMEDAP) and PMP (PMPDAP)
were significantly less efficient substrates than the former com-
pounds for the mitochondrial kinase but were phosphorylated
two- to threefold better than PMEA by the cytosol. Since both
PMEDAP and PMPDAP possess an amino group in the no. 2
position, it appears that subtle changes in the purine ring affect
the ability of these compounds to act as substrates for AK2. It
is also interesting that S-HPMPA and PMPA are both more
active than PMEA against herpesviruses and retroviruses, re-
spectively (9, 27). Thus, it seems possible that the increased
antiviral activity of these two drugs is at least partly related
to their increased phosphorylation. Obviously, enzymes other
than just kinases (i.e., DNA polymerases, reverse transcriptase)
are important determinants of the action of the phosphonates.
Nevertheless, the present studies have revealed the underlying
enzymatic basis for the intracellular anabolism of these novel
antiviral agents in human lymphoid cells. The results show that
both the cytosol and mitochondria contain purine nucleotide
kinases that can convert these analogs to their active metabo-
lites. The exact contribution of these alternate pathways to the
metabolism and action of the phosphonate analogs and purine
nucleotide analogs in general remains to be elucidated.
1
11. Egberink, H., M. Borst, H. Niphuis, J. Balzarini, H. Neu, H. Schellekens, E.
De Clercq, M. Horzinek, and M. Koolen. 1990. Suppression of feline immu-
nodeficiency virus infection in vivo by 9-(2-phosphonylmethoxyethyl)ad-
enine. Proc. Natl. Acad. Sci. USA 87:3087–3091.
1
2. Frank, R., M. Trosin, A. G. Tomasselli, G. E. Schulz, and R. H. Schirmer.
1984. Mitochondrial adenylate kinase (AK2) from bovine heart homology
with the cytosolic isoenzyme in the catalytic region. Eur. J. Biochem. 141:
629–636.
1
3. Hande, K. R., and B. A. Chabner. 1978. Pyrimidine nucleoside monophos-
phate kinase from human leukemic blast cells. Cancer Res. 38:579–585.
4. Kress, C. F., V. H. Bono, and L. Noda. 1966. The sulfhydryl groups of rabbit
muscle adenosine triphosphate-adenosine monophosphate phosphotransfer-
ase. J. Biol. Chem. 241:2293–2300.
1
1
5. Lin, J. C., E. De Clercq, and J. S. Pagano. 1987. Novel acyclic adenosine
analogs inhibit Epstein-Barr virus replication. Antimicrob. Agents Che-
mother. 31:1431–1433.
16. Naesens, L., J. Balzarini, I. Rosenberg, A. Holy, and E. De Clercq. 1989. 9-(2-
Phosphonylmethoxyethyl)-2,6-diaminopurine(PMEDAP): a novel agent with
anti-human immunodeficiency virus activity in vitro and potent anti-Moloney
murine sarcoma virus activity in vivo. Eur. J. Clin. Microbiol. Infect. Dis.
8:1043–1047.
1
7. Nave, J. F., A. Eschibach, D. Wolff-Kugel, S. Halazy, and J. Balzarini. 1994.
Enzymatic phosphorylation and pyrophosphorylation of 2Ј,3Ј-dideoxyade-
nosine-5Ј-monophosphate, a key metabolite in the pathway for activation of
the anti-HIV (human immunodeficiency virus) agent 2Ј,3Ј-dideoxyinosine.
Biochem. Pharmacol. 48:1105–1112.
1
8. Pauwels, R., J. Balzarini, D. Schols, M. Baba, J. Desmyter, I. Rosenberg, A.
Holy, and E. De Clercq. 1988. Phosphonylmethoxyethyl purine derivatives: a
new class of anti-human immunodeficiency virus agents. Antimicrob. Agents
Chemother. 32:1025–1030.
1
2
9. Richter, C. 1984. Hydroperoxide effects on redox state of pyridine nucleoti-
ϩ
des and Ca retention by mitochondria. Methods Enzymol. 105:435–441.
0. Robbins, B., M. Connelly, D. Marshall, R. Srinivas, and A. Fridland. 1995.
A human T lymphoid cell variant resistant to the acyclic nucleoside phos-
phonate, 9-(2-phosphonylmethoxyethyl)adenine (PMEA), shows a unique
combination of a phosphorylation defect and increased efflux of the agent.
Mol. Pharmacol. 47:391–397.
ACKNOWLEDGMENTS
This work was supported in part by PHS grant RO1 AI27652, Can-
cer center (CORE) grant P30 CA21765 from the National Institutes of
Health, and the American Lebanese Syrian Associated Charities.
We thank Norbert Bischofberger (Gilead Sciences, Foster City, Ca-
lif.) for the acyclic phosphonate derivatives.
2
1. Schirmer, R. H., and E. Thuma. 1972. Sensitivity of adenylate kinase
isozymes from normal and dystrophic human muscle to sulfhydryl reagents.
Biochim. Biophys. Acta 268:92–97.
2. Schmidt, E. 1974. Glutamate dehydrogenase UV assay, p. 650–656. In H. U.
Bergmeyer (ed.), Methods in enzymatic analysis, 2nd ed., vol. 2. Academic
Press, Inc., New York.
2
REFERENCES
2
3. Storrie, B., and E. A. Madden. 1990. Isolation of subcellular organelles.
Methods Enzymol. 182:203–224.
1
. Balzarini, J., and E. De Clercq. 1991. 5-Phosphoribosyl 1-pyrophosphate
synthetase converts the acyclic nucleoside phosphonates 9-(3-hydroxy-2-
phosphonyl-methoxypropyl)adenine directly to their antivirally active
diphosphate derivatives. J. Biol. Chem. 266:8686–8689.
2
4. Tanabe, T., M. Yamada, T. Noma, T. Kaji, and A. Nakazawa. 1993. Tissue
specific and developmentally regulated expression of the genes encoding
adenylate kinase isozymes. J. Biochem. 113(2):200–207.
2
. Balzarini, J., Z. Hao, P. Herdewijn, D. G. Johns, and E. De Clercq. 1991.
Intracellular metabolism and mechanism of anti-retrovirus action of 9-(2-
phosphonylmethoxyethyl)adenine, a potent anti-human immunodeficiency
virus compound. Proc. Natl. Acad. Sci. USA 88:1499–1503.
2
5. Tomasselli, A. G., and L. H. Noda. 1980. Mitochondrial ATP:AMP phos-
photransferase from beef heart: purification and properties. Eur. J. Biochem.
103:481–491.
2
2
6. Tomasselli, A. G., R. H. Schirmer, and L. H. Noda. 1979. Mitochondrial
GTP-AMP phosphotransferase. Eur. J. Biochem. 93:257–262.
7. Tsai, C. C., K. Follis, T. Beck, A. Sabo, R. Grant, C. Bartz, N. Bischofberger,
and R. Benveniste. Prevention of simian immunodeficiency virus infection in
macaques by 9-(2-phosphonylmethoxypropyl)adenine (PMPA). Science, in
press.
3
4
. Balzarini, J., L. Naesens, and E. De Clercq. 1990. Anti-retrovirus activity of
9
-(2-phosphonylmethoxyethyl)adenine (PMEA) in vivo increases when it is
less frequently administered. Int. J. Cancer 46:337–340.
. Balzarini, J., L. Naesens, P. Herdewijn, I. Rosenberg, A. Holy, R. Pauwels,
M. Baba, D. Johns, and E. De Clercq. 1989. Marked in vivo antiretrovirus
activity of 9-(2-phosphonylmethoxyethyl)adenine, a selective anti-human im-
munodeficiency virus agent. Proc. Natl. Acad. Sci. USA 86:332–336.
. Balzarini, J., L. Naesens, J. Slachmuylders, H. Niphuis, I. Rosenberg, A.
Holy, H. Schellekens, and E. DeClercq. 1991. 9-(2-Phosphonylmethoxyethy-
l)adenine effectively inhibits simian immunodeficiency virus (SIV) infection
in Rhesus monkeys. AIDS 5:21–28.
28. Tsai, C. C., K. Follis, A. Sabo, R. Grant, C. Bartz, R. Nolte, R. Benveniste,
and N. Bischofberger. 1994. Preexposure prophylaxis with 9-(2-phosphonyl-
methoxyethyl)adenine against simian immunodeficiency virus infection in
macaques. J. Infect. Dis. 169:260–266.
29. Verhoef, V., J. Sarup, and A. Fridland. 1981. Identification of the mechanism
of activation of 9-␤-D-arabinofuranosyladenine in human lymphoid cells
using mutants deficient in nucleoside kinases. Cancer Res. 41:4478–4483.
30. Watanabe, K., T. Itakura, and S. Kubo. 1979. Distribution of adenylate
kinase isozymes in porcine tissues and their subcellular localization. J. Bio-
chem. 85:799–805.
5
6
. Barditch-Crovo, P., J. Toole, H. Burgee, M. Wachsman, D. Ebeling, K.
Cundy, H. Jaffe, and P. Lietman. 1995. A randomized, double-blind, place-
bo-controlled phase I/II evaluation of 9-[2-(bispivaloyloxy-methyl) phospho-
nylmethoxy]adenine (bis-pom PMEA), an orally bioavailable prodrug of the
anti-HIV nucleotide, PMEA. Antiviral Res. 26:A229.