Glutathione peroxidase protects mice from viral-induced myocarditis

Abstract

Glutathione peroxidase 1 (GPX-1) is a selenium-dependent enzyme with antioxidant properties. Previous investigations determined that mice deficient in selenium developed myocarditis when infected with a benign strain of coxsackievirus B3 (CVB3/0). To determine whether this effect was mediated by GPX-1, mice with a disrupted Gpx1 gene (Gpx1-/-) were infected with CVB3/0. Gpx1-/- mice developed myocarditis after CVB3/0 infection, whereas infected wild-type mice (Gpx1+/+) were resistant. Sequencing of viruses recovered from Gpx1(-/-)-infected mice demonstrated seven nucleotide changes in the viral genome, of which three occurred at the G residue, the most easily oxidized base. No changes were found in virus isolated from Gpx1+/+ mice. These results demonstrate that GPX-1 provides protection against viral-induced damage in vivo due to mutations in the viral genome of a benign virus.

Full text

11430892-6638/98/0012-1143/$02.25 Q FASEB
/ 382f 0007 Mp 1143 Friday Jul 17 09:22 AM LP–FASEB 0007
Glutathione peroxidase protects mice from viral-
induced myocarditis
M. A. BECK,*
,1
R. S. ESWORTHY,
†
Y.-S. HO,
‡
AND F.-F. CHU
†
*Frank Porter Graham Child Development Center, University of North Carolina at Chapel Hill,
Chapel Hill, North Carolina 27599-8180, USA;
†
Department of Medical Oncology, City of Hope
Medical Center, Duarte, California 91010, USA; and
‡
Institute of Chemical Toxicology, Wayne State
University, Detroit, Michigan 48201, USA
ABSTRACT
Glutathione peroxidase 1 (GPX-1) is a
selenium-dependent enzyme with antioxidant prop-
erties. Previous investigations determined that mice
deficient in selenium developed myocarditis when in-
fected with a benign strain of coxsackievirus B3
(CVB3/0). To determine whether this effect was me-
diated by GPX-1, mice with a disrupted Gpx1 gene
(Gpx1
0/0
) were infected with CVB3/0. Gpx1
0/0
mice
developed myocarditis after CVB3/0 infection,
whereas infected wild-type mice (Gpx1
///
) were re-
sistant. Sequencing of viruses recovered from
Gpx1
0/0
-infected mice demonstrated seven nucleo-
tide changes in the viral genome, of which three
occurred at the G residue, the most easily oxidized
base. No changes were found in virus isolated from
Gpx1
///
mice. These results demonstrate that GPX-1
provides protection against viral-induced damage in
vivo due to mutations in the viral genome of a benign
virus.—Beck, M. A., Esworthy, R. S., Ho, Y.-S., Chu,
F.-F. Glutathione peroxidase protects mice from vi-
ral-induced myocarditis. FASEB J. 12, 1143–1149
(1998)
Key Words: selenium · GPX-1 · malondialdehyde · virus titer
· cardiac pathology
S
ELENIUM
(Se)
2
has been recognized as an essential
trace element for four decades (1). There are eight
well-characterized mammalian selenoproteins, in-
cluding thioredoxin reductase and four isozymes of
glutathione peroxidase (2–5). Keshan disease, an en-
demic cardiomyopathy affecting women and chil-
dren residing in specific regions of the People’s Re-
public of China, was found to be associated with Se
deficiency (6). However, because of the seasonal and
annual incidence of the disease, an infectious cofac-
tor, in addition to a deficiency in Se, was postulated
to contribute to the development of Keshan disease
(7). A likely candidate for an infectious cofactor of
Keshan disease is coxsackievirus. Coxsackieviruses,
RNA picornaviruses of approximately 7500 nucleo-
tides, have been isolated from Keshan disease tissues
and are known etiologic agents of viral-induced myo-
carditis (8, 9).
Infection of mice with virulent strains of coxsack-
ievirus B3 (CVB3) results in heart disease similar to
human pathology (10). Se- or vitamin E-deficient
mice are more severely affected by CVB3 infection
than Se- and vitamin E-adequate mice (10, 11).
CVB3/0, a benign strain of CVB3 that does not cause
myocarditis in Se-adequate mice, induced myocardi-
tis in Se-deficient mice (12). Six point mutations in
the genome of virus isolated from the hearts of Se-
deficient mice resulted in nucleotide substitutions
that were identical to those found in virulent strains
of CVB3 (13). These mutations transform a normally
benign strain of virus into a strain that induces myo-
carditis even in mice with normal Se nutriture.
How does a deficiency in Se promote the devel-
opment of a virulent CVB3 strain from an avirulent
strain? We propose that a decrease in GPX-1 activity,
a consequence of a deficiency in Se, is the critical step
leading to the change in virulence. To test this hy-
pothesis, we infected mice with a disrupted Gpx1
gene (Gpx1
0/0
or Gpx1-KO) (14) with the avirulent
CVB3/0. We found that, similar to infected Se-defi-
cient mice, an avirulent CVB3 rapidly mutates to a
virulent genotype in Gpx1-KO mice.
METHODS
Infection of mice
Gpx1-KO mice and wild-type mice have been described else-
where (14). Mice were maintained under protocols approved
by the Institutional Animal Review Board and inoculated at
3–4 wk of age with the noncardiovirulent coxsackievirus B3/
0 (CVB3/0) obtained from Steven Tracy, University of Ne-
1
Correspondence: Departments of Pediatrics and Nutri-
tion, FPG Child Development Center, 105 Smith Level Rd.,
CB #8180, University of North Carolina at Chapel Hill, Chapel
Hill, NC 27599-8180, USA. E-mail: melinda
T
beck@unc.edu
2
Abbreviations: Se, selenium; PCR, polymerase chain reac-
tion; GPX-1, glutathione peroxidase 1; S.I., stimulation index;
IL, interleukin; MDA, malondialdehyde; CVB3, coxsackievirus
B3; CVB3/0, a benign strain of CVB3; TBARS, thiobarbituric
acid reactive substances; TCID
50
, tissue culture infectious dose-
50.

1144 Vol. 12 September 1998 The FASEB Journal BECK ET AL.
/ 382f 0007 Mp 1144 Friday Jul 17 09:22 AM LP–FASEB 0007
braska Medical Center. Virus was obtained by transfection of
HeLa cells with a plasmid containing an infectious cDNA copy
of the CVB3 genome inserted into a plasmid vector (15). Mice
were inoculated with 10
5
times the median tissue culture in-
fectious dose (TCID
50
) of CVB3/0 intraperitoneally. At 10
days postinoculation, mice were bled, killed, and their hearts,
liver, and spleen were removed.
Histopathology of hearts
One-half of the heart, cut longitudinally, was placed in for-
malin, embedded in paraffin, sectioned (6 mm), and stained
with hematoxylin and eosin. The extent of inflammatory le-
sions within the myocardium was graded without knowledge
of the other experimental variables. Grading was done semi-
quantitatively according to the relative degree (from heart to
heart) of mononuclear cell infiltration and the extent of ne-
crosis.
Virus titers
One-half of the heart was snap frozen on dry ice, weighed,
ground in a small volume of RPMI-1640, using a Tenbroek
homogenizer, and freeze-thawed three times. The ground tis-
sue was centrifuged (20001g); the supernate was recovered,
titrated on HeLa cell monolayers to TCID
50
, and reported as
the geometric mean titer per gram of tissue.
Neutralizing antibody titers
Neutralizing antibody titers were measured by inhibition of
viral cytopathic effects as described previously (10).
Spleen cell proliferation assay
The assay for spleen cell proliferative activity in response to
mitogen (concanavalin A) and antigen (CVB3 antigen) has
been described in detail elsewhere (10).
Fluorescence-activated cell sorter analysis
Mediastinal lymph nodes were isolated from CVB3/0-infected
Gpx1-KO and wild-type mice at 10 days postinoculation.
Lymph nodes were teased into single cell suspensions, stained
with anti-mouse CD4, and CD8 conjugated with fluorescein.
Stained cells were analyzed by fluorescent activated cell sorter
(Becton-Dickinson, Rutherford, N.J.).
Malondialdehyde and catalase levels
The entire heart and a fraction of each liver (of equal weight
to heart) from both Gpx1-KO and wild-type mice infected 10
days earlier with CVB3/0 were homogenized by Polytron in a
buffer composed of 0.15 M KCl, 10 mM MOPS (morpholi-
nopropane sulfonic acid) (pH 7.2), 0.02% BHT (butylated
hydroxy toluene) (16–18). A portion of each homogenate was
made 0.2% with Triton-X-100. The Triton-X-100 samples were
sonicated at ice temperatures, then centrifuged at 15,000 1 g.
The supernatant was recovered for catalase assays using the
method of Beutler (19). The protein content of the samples
was determined with the BCA assay (Pierce, Rockford, Ill.)
using bovine serum albumin as the standard. The remainder
of each homogenate was processed for malondialdehyde
(MDA) determination as follows: the 500 1 g supernatant (5
min) was centrifuged at 15,000 1 g (20 min). The pellet was
recovered, rinsed once in the homogenization buffer, and re-
suspended in 10 vol of the homogenization buffer by sonica-
tion at 47C. The second supernatant was centrifuged at
100,000 1 g for 30 min. The pellet was recovered, rinsed once,
and resuspended in homogenization buffer by sonication.
The 15,000 1 g and 100,000 1 g pellet fractions were analyzed
for MDA content by a TBARS assay (thiobarbituric acid reac-
tive substances). Protein samples (200 mg) were prepared for
the TBARS assays according to the method of Schuh et al.
(20), with slight modifications suggested by Jentzsch et al.
(17). The standard was MDA generated during the acid step
from malonaldehyde bis (dimethyl acetyl) (Sigma, St. Louis,
Mo.). The TBARS chromogen was extracted into n-butanol
(1.25 ml, equal volume to sample). The chromogen sample
was read at 535 and 572 nm. The difference in absorption at
the two wave lengths was used to quantitate the chromogen
using the standard generated from malonaldehyde bis (di-
methyl acetyl) (17).
Sequencing of virus
The details of virus sequencing (by direct sequencing of poly-
merase chain reaction [PCR] -generated DNA) have been re-
ported earlier (13). DNA was sequenced at the UNC-Chapel
Hill automated DNA Sequencing Facility on a Model 373A
DNA Sequencer (Applied Biosystems, Foster City, Calif.) using
the Taq DyeDeoxy Terminator Cycle Sequencing Kit (Applied
Biosystems). At least two separate PCR primer sets and two
separate reverse transcriptase and PCR reactions were used to
confirm sequence changes.
RESULTS
Cardiac pathology of Gpx1-KO mice postinfection
To test the effect of expression of Gpx1 encoding
GPX-1 on the development of myocarditis after viral
challenge, we inoculated both wild-type (Gpx1
///
)
mice and mice with a disrupted Gpx1 gene (14) with
CVB3/0. Other GPX isozymes, GPX3 and GPX4, are
not affected in Gpx1-KO mice. Previous work in our
laboratory demonstrated that Se-deficient mice de-
velop myocarditis 7 days after CVB3/0 infection, with
peak cardiac damage occurring on day 10 postinfec-
tion (12). Se-adequate mice do not develop any signs
of myocarditis postinfection with the avirulent
CVB3/0 strain. Therefore, we killed both wild-type
and Gpx1-KO mice at 10 days postinfection, the peak
time for cardiac damage. Histological sections of
heart were stained and examined by light micros-
copy. More than half of the Gpx1-KO mice developed
mild to moderate myocarditis after CVB3/0 infec-
tion, characterized by a mononuclear cell infiltrate
and cytolysis (Fig. 1). In contrast, neither inflamma-
tion nor cytolysis were found in the hearts of wild-
type mice infected with this virus. Thus, similar to Se-
deficient mice, Gpx1-KO mice develop myocarditis
when infected with an avirulent strain of CVB3.
Virus titer
To investigate whether the difference in pathology
between Gpx1-KO and wild-type mice was due to a

GPK-1 PROTECTS MICE FROM VIRAL-INDUCED MYOCARDITIS 1145
/ 382f 0007 Mp 1145 Friday Jul 17 09:22 AM LP–FASEB 0007
Figure 1. CVB3/0-induced inflammation. Mice were in-
jected intraperitoneally with 10
5
tissue culture infectious
dose-50 (TCID
50
) of CVB3/0. Histologic sections of hearts
stained with hematoxylin and eosin from CVB3/0-infected
wild-type (A) and Gpx1-KO mice (B) at 10 days p.i. Arrow
indicates typical inflammatory lesion found scattered
throughout the myocardium. C) Anatomically similar por-
tions of the left ventricle were prepared as described in
Methods and graded for inflammation. Inflammation was
graded on a scale of 0 to 4/. Pathologic scores: 0, no le-
sions; 1/, foci of mononuclear cell inflammation associ-
ated with myocardial cell reactive changes without
myocardial cell necrosis; 2/, inflammatory foci clearly as-
sociated with myocardial cell reactive changes; 3/, inflam-
matory foci clearly associated with myocardial cell necrosis
and dystrophic calcification; 4/, extensive inflammatory
infiltration, necrosis, and dystrophic calcification. Each bar
represents the percentage of animals assigned to each
grade for each genotype. The graph represents 25 ///
mice and 29 0/0 mice.

1146 Vol. 12 September 1998 The FASEB Journal BECK ET AL.
/ 382f 0007 Mp 1146 Friday Jul 17 09:22 AM LP–FASEB 0007
Figure 2. Serum CVB3/0 neutralizing antibody titers 10 days
postinfection. Geometric mean titers (error bars represent
standard deviation) were determined by inhibition of viral cy-
topathic effect (10). Bar for /// data represents 14 animals
and bar for 0/0 data represents 36 animals.
difference in viral titer, we determined the geometric
mean titers of virus recovered from the hearts of in-
fected animals. The mean viral titer from the hearts
of Gpx1-KO mice (3.12{0.7 TCID
50
) was indistin-
guishable from the heart virus titer of wild-type mice
(3.33{0.9 TCID
50
). Thus, at 10 days postinfection,
the development of myocarditis in the Gpx-1 KO
mice was not associated with elevated virus titers in
the heart tissue. In contrast, hearts from Se-deficient
mice infected with CVB3/0 had 10-fold higher virus
titers when compared with Se-adequate mice (12).
Immune functions of infected Gpx1-KO mice
The immune system is involved in viral clearance and
contributes to its pathology (21). Se-deficient mice
were shown to have impaired immune responses af-
ter CVB3 infection. To determine whether Gpx1-KO
mice also had impaired immune responses, we ex-
amined several immune functions in the infected
Gpx1-KO mice and compared them with responses
from wild-type mice. Serum titers of virus neutraliz-
ing antibody for both Gpx1-KO and wild-type mice
were determined at 10 days postinfection. Neutraliz-
ing antibody titers in Gpx1-KO mice were less than
20% of those found in wild-type mice (Fig. 2), indi-
cating that an impairment of the B cell response oc-
curred in the infected Gpx1-KO mice. This is in con-
trast to Se-deficient mice, in which neutralizing
antibody titers were the same as the titers found in
Se-adequate mice (12). Functional B cell activity may
possibly depend on normal GPX-1 activity during de-
velopment, as Se-deficient mice were 3 wk old before
being fed the Se-deficient diet, an age when much of
the immune system has already matured.
T cell responses were determined by measuring
[
3
H]thymidine incorporation in either mitogen
(concanavalin A) or CVB3 viral antigen-stimulated
splenocytes (12) obtained from CVB3/0-infected
Gpx1-KO and wild-type mice. No difference in the
splenocyte proliferative responses were found be-
tween Gpx1-KO mice for mitogen (stimulation in-
dex: [S.I]: 41.8 (4.2) or antigen (S.I.: 22{2.3) and
wild-type mice (mitogen S.I.: 39.4{4.6; antigen S.I.:
21{2.2). Again, this is in contrast to Se-deficient
mice, in which proliferative responses to mitogen and
antigen were significantly decreased (12). Possibly a
TH2 response, which drives B cell production of an-
tibody by T cell secretion of interleukin 4 (IL-4) and
IL-5, is favored in the Se-deficient animals. However,
a TH1 type response, which stimulates T cell activity
by secretion of IL-2 and g-interferon, might predom-
inate in the Gpx1-KO mice, blunting the TH2 re-
sponse (22). The observation that viral antibody titers
were equivalent in wild-type and knockout mice sug-
gests that the antibody response plays little role in
controlling viral replication in the infected mice.
A deficiency in Se has been shown to result in de-
creased levels of CD4/ T cells (23). To determine
whether CD4/ T cells were affected in infected
Gpx1-KO mice, we analyzed lymphocytes in medias-
tinal lymph nodes from Gpx1-KO and wild-type mice
for levels of CD4/ and CD8/ T cells by fluorescence-
activated cell sorter. No differences in percentages of
CD4/ and CD8/ T cells were found between Gpx1-
KO and wild-type mice (data not shown).
Analysis of viral sequence from Gpx1-KO and wild-
type mice
Our previous work demonstrated that CVB3/0 virus
recovered from Se-deficient mice had six point mu-
tations that were identical to those found in virulent
CVB3 strains (13). Thus, the change in viral pheno-
type from avirulent to virulent in the Se-deficient
mice was due to genetic change of the virus. To de-
termine whether replication of the benign CVB3 in
Gpx1-KO mice would also alter the viral genome, we
isolated and sequenced viral RNA from the hearts of
a number of Gpx1-KO mice with and without heart
pathology and from the hearts of wild-type mice.
Seven nucleotides in the CVB3/0 genome that rep-
licated in Gpx1-KO mice had changed (Table 1).
These changes in nucleotide sequence were found
only in Gpx1-KO mice with pathology. No sequence
changes were found in virus recovered from hearts
of Gpx1-KO mice without gross pathology nor from
the hearts of wild-type mice. The strict association be-
tween changes in the viral nucleotide sequence and
induction of cardiomyopathy in Gpx1-KO mice sug-
gests that some, if not all, of these changes are re-
quired for virulence.

GPK-1 PROTECTS MICE FROM VIRAL-INDUCED MYOCARDITIS 1147
/ 382f 0007 Mp 1147 Friday Jul 17 09:22 AM LP–FASEB 0007
TABLE 1. Nucleotide sequences obtained from virus recovered from the hearts of CVB3/0-infected Gpx1-KO and wildtype mice, with and
without heart pathology, 10 days postinfection
Mouse genotype
Heart pathology
score
Nucleotide number (5*-3*)
234 788 2271 2438 2690 3324 7334
/// 0CGAGGCC
/// 0CGAGGCC
0/0 0CGAGGCC
0/0 0CGAGGCC
0/0 2/TATCATT
0/0 3/ TATCATT
0/0 2/ TATCATT
Se-deficient
a
3/ TA T C G T T
C3H/HeJ
a
a
From ref 13.
Six of the seven changes in the viral genome ob-
tained from the hearts of Gpx1-KO mice are identical
to those found in virus recovered from the hearts of
Se-deficient mice (Table 1). Thus, it appears that the
genetic changes in the virus that occurred in the Se-
deficient mice can be attributed to a deficiency in
Gpx-1 activity.
Malondialdehyde and catalase levels in infected
mice
GPX-1 plays a role in the antioxidant defense strategy
of the host. To determine whether increased oxida-
tive stress occurred in the hearts or livers of infected
mice, we quantified lipid peroxidation (by measuring
MDA levels) and catalase activity, which may be in-
creased in Gpx1-KO mice during viral infection (24,
25). Similar levels of malondialdehyde and catalase
activity were found in the hearts and livers of Gpx1-
KO and wild-type mice 10 days after viral infection
(malondialdehyde in heart microsomes of KO:
225{18 pmol/mg protein in wild-type: 225{36
pmol/mg; heart catalase activity in KO: 21.1 { 10.8
U/mg protein in wild-type: 23.0 { 6.0 U/mg protein;
liver data not shown). Baseline levels of MDA and
catalase in the hearts of uninfected KO and wild-type
mice do not differ (14). Thus, measurements of MDA
and catalase at 10 days postinfection may not be ade-
quate to detect oxidative stress in this model.
DISCUSSION
These data demonstrate that, in a host deficient in
GPX-1 activity, an avirulent virus undergoes a phe-
notypic change to virulence due to point mutations
in the genome. Single mutations in various sites in
the CVB3 genome have been shown to be associated
with cardiovirulence (26, 27). Our previous work
demonstrating similar point mutations in Se-defi-
cient mice suggests that the mutations may be driven
by oxidative damage.
Glutathione peroxidase 1 (GPX-1), the first sele-
noprotein identified in mammals (28, 29), is present
in both cytosol and mitochondria. The activity of
GPX-1 is widespread, with highest levels in liver, kid-
ney, and heart (30, 31). Heart tissue has a very low
level of catalase activity, which is the other major en-
zyme involved in the detoxification of H
2
O
2
(32). Al-
though the importance of GPX-1 as an antioxidant
has been debated (33, 34), we hypothesized that re-
duction of GPX-1 activity in Se-deficient mice con-
tributed to the development of myocarditis postin-
fection with a benign strain of CVB3. The work
reported here with the Gpx1-KO mice confirms that
GPX-1 plays a critical role in the defense against viral-
induced myocarditis. Although many studies dem-
onstrate an antioxidant activity of GPX-1 in cell cul-
tures, there is less evidence for its function as an
antioxidant in whole organisms (14, 35). Indeed, be-
cause of its rapid response to changes in Se nutriture,
GPX-1 has been described as an Se storage protein
in rodents (34). This study in Gpx1-KO mice is the
first report demonstrating a critical protective role
for this enzyme against viral-induced pathogenesis
and that other antioxidant protective mechanisms
cannot compensate for a lack of GPX-1.
What is the mechanism that allows for the changes
in the viral genome to occur in the Se-deficient or
Gpx1-KO animal? There are several possibilities. One
explanation is direct oxidative damage of viral RNA.
Single-stranded RNA has been suggested to be more
susceptible than double-stranded DNA to damage by
free radicals. Although oxidative damage to RNA has
been studied far less than oxidative damage to DNA,
a recent report (36) suggested that the Escherichia coli
muT protein (which has mammalian homologs) is
instrumental in protecting transcriptional fidelity by
hydrolyzing oxidized guanine, which mispairs with
adenine. Guanine is the site of mutation at 3/7 sites
in the virus obtained from Gpx1-KO mice, which

1148 Vol. 12 September 1998 The FASEB Journal BECK ET AL.
/ 382f 0007 Mp 1148 Friday Jul 17 09:22 AM LP–FASEB 0007
likely reflects the high oxidative potential of this nu-
cleotide (37, 38). Unrepaired oxidized bases can
cause DNA mispairing and mutation during replica-
tion (38, 39), and similar changes may occur at oxi-
dized bases in viral RNA genomes (40). In addition,
RNA viruses lack efficient proofreading and post-
replicative repair activities (41). Thus, mutation rates
of RNA viruses are in the range of 10
03
to 10
05
sub-
stitutions per copied nucleotide, which is at least 10
3
-
fold greater than the mutation rate for cellular DNA
(42, 43). It has been suggested that RNA viruses rep-
licate near the minimal fidelity compatible with main-
taining their genetic information, although not all
mutations will be viable. Therefore, in an individual
virus population, individual genomes that differ in
one or more nucleotides will form the average or con-
sensus sequence of the population, or quasispecies.
Thus, a high mutation rate coupled with a lack of
GPX-1 antioxidant protection increases the likeli-
hood of an accelerated mutation rate of coxsackie-
virus observed in Gpx1-KO mice.
We hypothesize that the transformation of CVB3/
O from avirulence to virulence in Gpx1-KO mice in-
volves several steps: 1) oxidative RNA damage, 2) in-
creased mutagenesis of oxidized viral RNA, 3) lack of
proofreading enzymes, and 4) impairment of B and/
or T cell function. Our work clearly defines an oxi-
dative protection role for GPX-1 activity and points
to the importance of adequate oxidative defense
mechanisms of the host to protect from viral chal-
lenge. This work demonstrates that oxidative damage
can affect not only the host, but can influence the
pathogen itself, altering a normally avirulent patho-
gen into a virulent one that can then cause damage
even in mice not oxidatively challenged. Further in-
vestigations are needed to identify the critical roles
of GPX-1 in resistance to viral pathogenesis.
This work was supported by National Institutes of Health
grants DK46921 (F.-F.C.), P30 ESO6693 (Y.-S.H.), and
HL54123 (M.A.B.). We thank Q. Shi for excellent technical
assistance.
REFERENCES
1. Schwarz, K., and Foltz, C. M. (1957) Selenium as an integral part
of factor 3 against dietary liver degeneration. J. Am. Chem. Soc.
79, 3292–3293
2. Gladyshev, V. N., Jeang, K.-T., and Stadtman, T. C. (1996) Sele-
nocysteine, identified as the penultimate C-terminal residue in
human T-cell thioredoxin reductase, corresponds to TGA in the
human placental gene. Proc. Natl. Acad. Sci. USA 93, 6146–6151
3. Vendeland, S. C., Beilstein, M. A., Yeh, J.-Y., Ream, W., and
Whanger, P. D. (1995) Rat skeletal muscle selenoprotein W:
cDNA clone and mRNA modulation by dietary selenium. Proc.
Natl. Acad. Sci. USA 92, 8749–8753
4. Burk, R. F., and Hill, K. E. (1993) Regulation of selenoproteins.
Annu. Rev. Nutr. 13, 65–81
5. Chu, F.-F., Doroshow, J. H., and Esworthy, R. S. (1993) Expres-
sion, characterization and tissue distribution of a new cellular
selenium-dependent glutathione peroxidase, GSHPx-G1. J. Biol.
Chem. 268, 2571–2576
6. Gu, B. Q. (1983) Pathology of Keshan disease. A comprehensive
review. Chin. Med. J. (Engl. Ed.) 96, 251–261
7. Ge, K. Y., Xue, A., and Bai, J. (1983) Keshan disease–an endemic
cardiomyopathy in China. Virchows Arch. 401, 1–14
8. Beck, M. A. (1997) The role of nutrition in viral disease. J. Nutr.
Biochem. 7, 683–690
9. Martino, T. A., Liu, P., Petric, M., and Sole, M. J. (1977) In Hu-
man Enterovirus Infections (Rotbart, H. A., ed) pp. 291–351, Amer-
ican Society for Microbiology Press, Washington D.C.
10. Beck, M. A., Kolbeck, P. C., Rohr, L. H., Shi, Q., Morris, V.C.,
and Levander, O.A.. (1994) Increased virulence of a human en-
terovirus (coxsackievirus B3) in selenium-deficient mice. J. Infect.
Dis. 170, 351–357
11. Beck, M. A., Kolbeck, P. C., Rohr, L. H., Shi, Q., Morris, V. C.,
and Levander, O. A. (1994) Vitamin E deficiency intensifies the
myocardial injury of coxsackievirus B3 infection of mice. J. Nutr.
124, 345–358
12. Beck, M. A., Kolbeck, P. C., Rohr, L. H., Shi, Q., Morris, V. C.,
and Levander, O. A. (1994) Amyocarditic coxsackievirus be-
comes myocarditic in selenium deficient mice. J. Med. Virol. 43,
166–170
13. Beck, M. A., Shi, Q., Morris, V. C., and Levander, O. A. (1995)
Rapid genomic evolution of a non-virulent coxsackievirus B3 in
selenium-deficient mice results in selection of identical virulent
isolates. Nat. Med. 1, 433–436
14. Ho, Y.-S., Magnenat, J. L., Bronson, R. T., Cao, J., Gargano, M.,
Sugawara, M., and Funk, C. D. (1997) Mice deficient in cellular
glutathione peroxidase develop normally and show no increased
sensitivity to hypoxia. J. Biol. Chem. 272, 16644–16651
15. Chapman, N. M, Tu, Z., Tracy, S., and Gauntt, C. J. (1994) An
infectious cDNA copy of the genome of a non-cardiovirulent
coxsackievirus B3 strain—its complete sequence analysis and
comparison to the genomes of cardiovirulent coxsackieviruses.
Arch. Virol. 135, 115–130
16. Janero, D. R., and Burghardt, B. (1988) Analysis of cardiac
membrane phospholipid peroxidation kinetics as malondialde-
hyde: nonspecificity of thiobarbituric. Lipids 23, 452–458
17. Jentzsch, A. M., Bachman, H., Furst, P., and Biesalski, H. K.
(1996) Improved analysis of malondialdehyde in human body
fluids. Free Rad. Biol. Med. 20, 251–256
18. Albro, P. W., Corbett, J. T., and Schroeder, J. L. (1987) Rapid
isolates of microsomes for studies of lipid peroxidation. Lipids
22, 751–756
19. Beutler, E. (1975) Red Cell Metabolism: A Manual of Biochemical
Methods, 2nd Ed, Grune and Stratton, New York
20. Schuh, J., Fairclough G. F., Jr., and Haschemeyer, R. H. (1978)
Oxygen-mediated heterogeneity of apo-low-density lipoprotein.
Proc. Natl. Acad. Sci. USA 75, 3173–3177
21. Cook, D. N., Beck, M. A., Coffman, T., Kirby, S. L., Sheridan, J. F.,
Pragnell, I. B., and Smithies, O. (1995) Requirement of MIP-1a for
inflammatory response to viral infection. Science 269, 1583–1585
22. Abbas, A. K., Murphy, K. M., and Sher, A. (1996) Functional
diversity of helper T lymphocytes. Nature (London) 383, 787–793
23. Spallholz, J. E., Boylan, L. M., and Larsen, H. S. (1990) Advances
in understanding selenium’s role in the immune system. Ann.
N.Y. Acad. Sci. 587, 123–139
24. Akman, S. A., Forrest, G., Chu, F.-F., and Doroshow, J. H. (1989)
Resistance to hydrogen peroxide associated with altered catalase
mRNA stability in MCF7 breast cancer cells. Biochim. Biophys.Acta
1009, 70–74
25. Lin, F., Thomas, J. P., and Girotti, A. W. (1993) Hyperexpression
of catalase in selenium-deprived murine L210 cells. Arch.
Biochem. Biophys. 305, 176–185
26. Tu, Z., Chapman N. M., Hufnagel, G., Tracy, S., Romero, J. R.,
Barry, W. H., Zhao, L., Currey, K., and Shapiro, B. (1995) The
cardiovirulent phenotype of coxsackievirus B3 is determined at
a single site in the genomic 5* nontranslated region. J. Virol. 69,
4607–4618
27. Chapman, N. M., Romero, J. R., Pallansch, M. A., and Tracy, S.
(1997) Sites other than nucleotide 234 determine cardioviru-
lence in natural isolates of coxsackievirus B3. J. Med. Virol. 52,
253–261
28. Rotruck, J. T., Pope, A. L., Ganther, H. E., Swanson, A. B., Hafeman,
D. G., and Hoekstra, W. G. (1973) Selenium: biochemical role as a
component of glutathione peroxidase. Science 179, 588–590

GPK-1 PROTECTS MICE FROM VIRAL-INDUCED MYOCARDITIS 1149
/ 382f 0007 Mp 1149 Friday Jul 17 09:22 AM LP–FASEB 0007
29. Esworthy, R. S., Ho, Y.-S., and Chu, F.-F. (1997) The Gpx1 gene
encodes mitochondrial glutathione peroxidase in the mouse
liver. Arch. Biochem. Biophys. 340, 59–63
30. Chambers, I., Frampton, J., Goldfarb, P., Affara, N., McBain, W.,
and Harrison, P. R. (1986) The structure of the mouse glutathi-
one peroxidase gene: the selenocysteine in the active site is en-
coded by the ‘termination’ codon, TGA. EMBO J. 5, 1221–1227
31. Zhang, L. P., Maiorino, M., Roveri, A., and Ursini, F. (1989)
Phospholipid hydroperoxide glutathione peroxidase specific ac-
tivity in tissues of rats of different age and comparison with other
glutathione peroxidases. Biochim. Biophys. Acta 1006, 140–143
32. Kang, Y. J., Chen, Y., and Epstein, P. N. (1996) Suppression of
doxorubucin cardiotoxicity by overexpression of catalase in the
heart of transgenic mice. J. Biol. Chem. 271, 12610–12616
33. Combs, G. F., Jr., and Combs, S. B. (1984) The nutritional bio-
chemistry of selenium. Annu. Rev. Nutr. 4, 257–280
34. Burk, R. F. (1991) Molecular biology of selenium with implica-
tions for its metabolism. FASEB J. 5, 2274–2279
35. Raes, M., Michiels, C., and Remacle, J. (1987) Comparative study
of the enzymatic defense systems against oxygen-derived free
radicals: the key role of glutathione peroxidase. Free Rad. Biol.
Med. 3, 3–7
36. Taddei, F., Hayakawa, H., Bouton, M.-F., Cirinesi, A.-M., Matic,
I., Sekiguchi, M., and Radman, M. (1997) Counteraction by
MutT protein of transcriptional errors caused by oxidative dam-
age. Science 278, 128–130
37. Cheng, K. C., Cahill, D. S., Kasai, H., Nishimura, S., and Loeb,
L. A. (1992) 8-Hydroxyguanine, an abundant form of oxidative
DNA damage causes GrT and ArC substitutions. J. Biol. Chem.
267, 166–172
38. Yanagawa, H., Ogawa, Y., and Ueno, M. (1992) Redox ribonu-
cleosides. Isolation and characterization of 5-hydroxyuridine, 8-
hydroxyguanosine, and 8-hydroxyadenosine from Torula yeast
RNA. J. Biol. Chem. 267, 13320–13326
39. Shibutani, S., Takeshita, M., and Grollman, A. P. (1991) Inser-
tion of specific bases during DNA synthesis past the oxidation-
damaged base 8-oxodG. Nature (London) 349, 431–434
40. Domingo, E. (1997) RNA virus evolution, population dynamics,
and nutritional status. Biol. Trace Elem. Res. 56, 23–30
41. Steinhauer, D., Domingo, E., and Holland, J. J. (1992) Lack of
evidence for proofreading mechanisms associated with an RNA
virus polymerase. Gene 122, 281–288
42. Domingo, E., and Holland, J. J. (1994) Mutation rates and rapid
evolution of RNA viruses. In The Evolutionary Biology of Viruses
(Morse, S. S., ed) pp. 161–184, Raven Press, Ltd., New York
43. Domingo, E., Holland, J. J., Biebricher, C., and Eigen, M. (1995)
Quasispecies: the concept and the word. In Molecular Evolution
of the Viruses (Gibbs, A., Calisher, C., and Garcia-Arenal, F., eds)
pp. 171–180, Cambridge University Press
Received for publication January 23, 1998.
Revised for publication March 24, 1998.