PUBLISHED VERSION: Understanding Selenium and Glutathione as Antiviral Factors in COVID-19: Does the Viral M pro Protease Target Host Selenoproteins and Glutathione Synthesis?

Abstract

Glutathione peroxidases (GPX), a family of antioxidant selenoenzymes, functionally link selenium and glutathione, which both show correlations with clinical outcomes in COVID-19. Thus, it is highly significant that cytosolic GPX1 has been shown to interact with an inactive C145A mutant of Mpro , the main cysteine protease of SARS-CoV-2, but not with catalytically active wild-type Mpro. This seemingly anomalous result is what might be expected if GPX1 is a substrate for the active protease, leading to its fragmentation. We show that the GPX1 active site sequence is substantially similar to a known Mpro cleavage site, and is identified as a potential cysteine protease site by the Procleave algorithm. Proteolytic knockdown of GPX1 is highly consistent with previously documented effects of recombinant SARS-CoV Mpro in transfected cells, including increased reactive oxygen species and NF-κB activation. Because NF-κB in turn activates many pro-inflammatory cytokines, this mechanism could contribute to increased inflammation and cytokine storms observed in COVID-19. Using web-based protease cleavage site prediction tools, we show that Mpro may be targeting not only GPX1, but several other selenoproteins including SELENOF and thioredoxin reductase 1, as well as glutamate-cysteine ligase, the rate-limiting enzyme for glutathione synthesis. This hypothesized proteolytic knockdown of components of both the thioredoxin and glutaredoxin systems is consistent with a viral strategy to inhibit DNA synthesis, to increase the pool of ribonucleotides for RNA synthesis, thereby enhancing virion production. The resulting "collateral damage" of increased oxidative stress and inflammation would be exacerbated by dietary deficiencies of selenium and glutathione precursors.

Full text

PERSPECTIVE
published: 02 September 2020
doi: 10.3389/fnut.2020.00143
Frontiers in Nutrition | www.frontiersin.org 1September 2020 | Volume 7 | Article 143
Edited by:
Christophe Matthys,
KU Leuven, Belgium
Reviewed by:
Alan Diamond,
University of Illinois at Chicago,
United States
Anna Kipp,
Friedrich Schiller University
Jena, Germany
*Correspondence:
Ethan Will Taylor
ewtaylor@uncg.edu
Specialty section:
This article was submitted to
Nutritional Immunology,
a section of the journal
Frontiers in Nutrition
Received: 26 June 2020
Accepted: 21 July 2020
Published: 02 September 2020
Citation:
Taylor EW and Radding W (2020)
Understanding Selenium and
Glutathione as Antiviral Factors in
COVID-19: Does the Viral Mpro
Protease Target Host Selenoproteins
and Glutathione Synthesis?
Front. Nutr. 7:143.
doi: 10.3389/fnut.2020.00143
Understanding Selenium and
Glutathione as Antiviral Factors in
COVID-19: Does the Viral Mpro
Protease Target Host Selenoproteins
and Glutathione Synthesis?
Ethan Will Taylor*and Wilson Radding
Department of Chemistry and Biochemistry, The University of North Carolina at Greensboro, Greensboro, NC, United States
Glutathione peroxidases (GPX), a family of antioxidant selenoenzymes, functionally
link selenium and glutathione, which both show correlations with clinical outcomes in
COVID-19. Thus, it is highly significant that cytosolic GPX1 has been shown to interact
with an inactive C145A mutant of Mpro, the main cysteine protease of SARS-CoV-2,
but not with catalytically active wild-type Mpro. This seemingly anomalous result is
what might be expected if GPX1 is a substrate for the active protease, leading to its
fragmentation. We show that the GPX1 active site sequence is substantially similar
to a known Mpro cleavage site, and is identified as a potential cysteine protease
site by the Procleave algorithm. Proteolytic knockdown of GPX1 is highly consistent
with previously documented effects of recombinant SARS-CoV Mpro in transfected
cells, including increased reactive oxygen species and NF-κB activation. Because
NF-κB in turn activates many pro-inflammatory cytokines, this mechanism could
contribute to increased inflammation and cytokine storms observed in COVID-19.
Using web-based protease cleavage site prediction tools, we show that Mpro may
be targeting not only GPX1, but several other selenoproteins including SELENOF and
thioredoxin reductase 1, as well as glutamate-cysteine ligase, the rate-limiting enzyme
for glutathione synthesis. This hypothesized proteolytic knockdown of components of
both the thioredoxin and glutaredoxin systems is consistent with a viral strategy to
inhibit DNA synthesis, to increase the pool of ribonucleotides for RNA synthesis, thereby
enhancing virion production. The resulting “collateral damage” of increased oxidative
stress and inflammation would be exacerbated by dietary deficiencies of selenium and
glutathione precursors.
Keywords: coronavirus, COVID-19, selenium, glutathione, glutathione peroxidase 1, protease, selenoprotein,
thioredoxin reductase

Taylor and Radding Selenoproteins as SARS-CoV-2 Protease Substrates
INTRODUCTION
Several independent studies have now established a significant
association between the outcome of COVID-19 and previously
documented regional selenium (Se) status in Chinese cities (1),
and a similar relationship between serum Se and mortality in a
European cohort of COVID-19 patients (2). These observations
invite questions about the mechanisms involved, particularly
because they fit into a consistent pattern of a role for Se that has
been reported over several decades for a variety of RNA viruses
(enteroviruses, hantaviruses, and influenza A) and viruses with
an RNA stage (HIV-1 and Hepatitis B virus), as reviewed by
various authors (3–5).
Most (but not all) of the biological roles of Se, both
as selenocysteine in selenoproteins, and as redox-active Se-
containing metabolites, involve interactions with cysteine thiols
and disulfides in proteins and peptides, and their various
oxidized forms. Like Se, the essential antioxidant and free radical
scavenger glutathione (GSH), a tripeptide thiol, has also proven
to be an important factor in various viral infections, particularly
in HIV/AIDS, as reviewed in section 2.1.2. of Taylor (6), and
most recently, in COVID-19 (7,8). Given their intertwined
biochemical roles, there are likely to be common factors and
mechanisms underlying the therapeutic importance of Se and
GSH in viral infections.
A CONFIRMED MOLECULAR
INTERACTION BETWEEN SARS-CoV-2
PROTEASE AND A HUMAN
SELENOPROTEIN
Correlations between COVID-19 clinical outcomes and both
host Se and GSH status provide important context to a related
observation emerging from a proteomics-based study of possible
cellular targets of SARS-CoV-2 (SCoV2) proteins. Using affinity-
purification mass spectrometry, Gordon et al. identified high-
confidence protein-protein interactions between 26 of the SCoV2
proteins and human proteins (9). As bait for interacting human
proteins, one of the proteins they used was the SCoV2 main viral
protease Mpro, a cysteine protease also known as nonstructural
protein 5 (nsp5). The study also included a catalytically inactive
C145A Mpro mutant (lacking the active site cysteine), which
was also used as a bait protein, in order to discriminate false
positives that might bind non-specifically to the Mpro active site
cysteine via disulfide bond formation. One of the interactions
they identified involved the cytosolic form of the selenoprotein
glutathione peroxidase, GPX1, which bound strongly to the
inactive C145A Mpro mutant. However, this interaction with
GPX1 was not observed with the wild type Mpro .This is precisely
what one might expect to see if GPX1 is a protease substrate,
because its cleavage would produce two fragments that would
necessarily have reduced affinity to the enzyme relative to GPX1.
Significantly, as shown in Figure 6 in their Extended Data (9),
Gordon et al. identified one other host protein (TRMT1) that
bound only to the inactive Mpro C145A mutant, and concluded
it was a likely Mpro substrate, because they were able to identify
a putative Mpro cleavage site in the TRMT1 protein sequence
(PRLQ/ANFT), where the slash (/) represents the cleavage site.
Thus, the only piece of evidence lacking to draw a similar
conclusion for GPX1 is a candidate Mpro cleavage site, which most
algorithms will fail to find, e.g., if they are highly stringent about
requiring a Q in the P1 position (Figure 1).
As shown in Figure 1A for the 2003 SARS coronavirus (SARS-
CoV, SCoV) and in Figure 1B for the 2019 SCoV2, the consensus
logo patterns of 11 Mpro cleavage sites in the viral 1ab polyprotein
are essentially identical for the 2 viruses. Specifically, 8 of the
11 sites are fully identical over the 10 residues spanning the
P5 to P5′positions, and only two of the 11 cleavage sites have
more than a single amino acid difference within the 10 residue
span (Figure S1). This means that (as discussed below) functional
studies of the 2003 SCoV protease are highly relevant to that of
the 2019 virus. From Figures 1A,B, the most important residue
positions for Mpro recognition are, in descending order, P1, P2,
P1′, P4, and P3, with the first 3 being the major determinants,
with the minimal P2-P1′consensus being LQ/(S,A,N,G). The
downstream sequence past P1′is highly variable.
The GPx1 active site has the sequence LUG, where U is
the catalytic selenocysteine; this matches the observed Mpro
consensus core target sequence combination LQ/G in 2 of 3
positions. Furthermore, as shown in Figure 1C, the important
upstream side of the known Mpro cleavage site at the nsp13/14
junction, NVATLQ/A, is remarkably similar to that of the active
site sequence of GPx1, NVASLU/G, where the selenocysteine
(U) lines up with a glutamine (Q) in the viral sequence. These
two amino acids (U and Q) are not highly similar, but are both
midrange in size, and polar in nature, because the selenol residue
is predominantly ionized at physiological pH. The other two
“mismatches” in the important positions P3 and P1′, i.e., S vs.
T and G vs. A, both differ only slightly, by the size of a methyl
group, and in any case, having a glycine (G) in the P1′position is
a permitted residue (Figure 1A), as observed in the SCoV Mpro
cleavage site at the nsp5/6 junction (indicated by ∗in Figure 2).
Significantly, the Procleave protease cleavage prediction server
(procleave.erc.monash.edu.au) (13) identified the GPX1 active
site octameric sequence ASLU/GTTV as a highly ranked possible
cleavage site for a cathepsin S-like cysteine protease (Figure 2G).
For several reaction intermediates in the GPX1 mechanism, Se
is attached to either oxygen, or to nitrogen, as a selenenylamide.
The latter is quite stable, and may accumulate during GSH
depletion, because GSH is needed to regenerate the selenolate
(14). Selenenylamide is more similar than selenocysteine to
glutamine, so that form might be preferentially targeted by Mpro.
Because the mutation rate in cellular genes is thousands of
times slower than that for an RNA virus, if there is a physical
interaction between these two proteins as Gordon et al. observed
using the inactive mutant Mpro, it is almost certainly because
the virus has evolved to target the host for some reason, rather
than the converse. The most parsimonious hypothesis is simply
that this is a case of a viral protease targeting a cellular gene for
cleavage, for which there are abundant precedents, including the
proposed Mpro targeting of TRMT1 and HDAC2 (9), and HIV-
1 protease, which can be toxic to cells, via its action at various
cellular targets (15).
Frontiers in Nutrition | www.frontiersin.org 2September 2020 | Volume 7 | Article 143

Taylor and Radding Selenoproteins as SARS-CoV-2 Protease Substrates
FIGURE 1 | Coronavirus Mpro cleavage site consensus sequence logo plots and comparisons. Logo plots from multiple alignments of the known Mpro cleavage sites
are shown for the 2003 SCoV (A) and 2019 SCoV2 (B). The height of a letter at each position reflects its probability in the alignment; each of the logos shown
represents the consensus of 11 Mpro cleavage sites from a single virus. Plots were generated using WebLogo (weblogo.berkeley.edu/logo.cgi), from the alignments in
Figure S1.(C) Comparison of the GPX1 active site sequence containing selenocysteine (U) to the known SCoV2 Mpro cleavage site at the nsp13/14 junction; this site
is identical in the 2003 and 2019 coronaviruses. (D) Comparison of a predicted Mpro cleavage site in human selenoprotein F to a known SCoV2 Mpro cleavage site at
the nsp12/13 junction.
FUNCTIONAL EFFECTS OF MPRO IN
TRANSFECTED CELLS ARE CONSISTENT
WITH GPX1 KNOCKDOWN
As detailed above, the original 2003 SCoV Mpro has essentially
identical substrate sequence specificity as that of SCoV2. More
specifically, the 10 residue sequence of the cleavage site at
the nsp13/14 junction shown in Figure 2A, with the greatest
overall similarity to the GPX1 active site sequence, is 100%
identical in the 2003 and 2019 SARS coronaviruses (Figure S1).
Thus, the results of a 2006 study on the effects of transfecting
cells with a SCoV Mpro expression vector, in order to study
the effects of Mpro alone in the absence of intact virus
(16), are highly relevant for COVID-19. In transfected human
cells, the results were exactly what one would expect from
knockdown of GPX1: increased oxidative stress via production
of reactive oxygen species and activation of transcription factor
NF-κB (16). Both of these responses can be produced by
exposing cells to hydrogen peroxide, and inhibited by increasing
GPX1 activity, either by overexpression or via the addition of
sodium selenite to cell culture media (17,18). Cells expressing
the SCoV Mpro also became apoptotic, which is a known
consequence of NF-κB activation, especially in combination
with a high burden of ROS (19), which is even more likely
to occur if other antioxidant proteins are also degraded, as
discussed below.
POTENTIAL MPRO SITES IN OTHER
SELENOPROTEINS AND
GLUTATHIONE-RELATED PROTEINS
The question then arises, could targeting of host selenoproteins
for proteolysis be part of a more extensive viral strategy
contributing to the correlations between Se and GSH status
and COVID-19 outcome? To investigate this possibility, we
undertook a systematic search for potential Mpro cleavage sites
in the human selenoproteome and in several GSH-related
proteins including glutaredoxin, a small thioredoxin-like CXXC
protein that is recycled by GSH, rather than directly by a
reductase enzyme. For this analysis, we used three online
resources: (1) NetCorona (cbs.dtu.dk/services/NetCorona) (12),
the only resource specific for predictions of Mpro cleavage sites
in coronaviruses (although not limited to just the SARS type
coronaviruses), (2) PROSPER, the Protease Specificity Prediction
Server (prosper.erc.monash.edu.au) (11), which was rated in a
Frontiers in Nutrition | www.frontiersin.org 3September 2020 | Volume 7 | Article 143

Taylor and Radding Selenoproteins as SARS-CoV-2 Protease Substrates
FIGURE 2 | Comparison of known and predicted Mpro cleavage sites. The most relevant positions P4-P1′of the 11 known SARS coronavirus Mpro cleavage sites in
non-structural proteins are shown (for nsp4 through nsp16), aligned with predicted sites in several human selenoproteins (GPX1, SelenoF, SelenoP, and TXNRD1), as
well as glutaredoxin-1 (GLRX-1) and the rate-limiting enzyme for glutathione synthesis, glutamate-cysteine ligase catalytic subunit (GCLC, with 2 predicted cleavage
sites A,B). The known nsp cleavage sites shown are all from SARS-CoV-2, with the addition of the nsp5/6 site from SARS-CoV (marked*). The sites identified in
human proteins are displayed next to the known Mpro cleavage site to which they are most similar, highlighted in matching color. The experimentally determined
catalytic efficiencies of SARS-CoV Mpro at each of the known nsp cleavage sites are shown as relative kcat/Km(10). Three different methods for prediction of cysteine
protease cleavage sites were used: PROSPER (11), with numbers in parenthesis showing the score for the predicted site, any score >0.8 being significant; NetCorona
(12), where scores >0.5 are considered positive, and Procleave (13), where every possible P4-P4′octamer in a protein sequence is evaluated, scored and ranked.
Some aligned non-identical residues highlighted in italics are nonetheless chemically and structurally similar to residues found in the same position in the known
cleavage sites, e.g., serine (S) vs. threonine (T), Leucine (L) or isoleucine (I) vs. valine (V), and glutamine (Q) vs. asparagine (N); all of these pairs differ only by a single
carbon atom (CH3or CH2unit). The predicted Mpro cleavage sites in human proteins labeled (A–G) are discussed in the text. The computational results cited are
collated in Figure S2.
recent study as one of the most accurate resources for prediction
of HIV-1 protease sites (20), and (3) Procleave (cited above),
which was only used if either of the first 2 methods failed to
identify a site. Both Procleave and PROSPER do not specifically
predict coronavirus cysteine protease sites, but make predictions
for more generic cysteine proteases (e.g., cathepsins B, K, L,
or S). Thus, for both PROSPER and Procleave, the results had
to be screened manually for matches to the Mpro cleavage
site consensus as shown in Figures 1A,B. This search is also
predicated on the assumption that the approach of Gordon
et al. (9) might fail to identify some protease targets, because
their high affinity interactions are mostly limited to an 8–
10 residue sequence, and in some cases may be unable to
withstand the conditions needed to be isolable and observable by
mass spectrometry.
Figure 2 presents the most notable and highly ranked
potential Mpro sites identified computationally (the order A-G
is not significant, as placement in the listing was determined by
similarity to the known nsp sites). The most remarkable instance
is in selenoprotein F (SELENOF), an ER protein involved
in glycoprotein folding quality control. Over the 8 residues
spanning P4 to P4′, it has 7 of 8 identical to the Mpro site at
the nsp12/nsp13 junction, with the only mismatch at the highly
variable P3′position (Figure 1D), giving a near-perfect match to
the Mpro consensus. This was the highest scored NetCorona hit,
also highly scored by PROSPER, as shown in Figure 2F, and is
Frontiers in Nutrition | www.frontiersin.org 4September 2020 | Volume 7 | Article 143

Taylor and Radding Selenoproteins as SARS-CoV-2 Protease Substrates
significant because coronavirus assembly begins in the ER with
spike protein accumulation on the ER membrane surface.
Equally important is the hit on the selenoprotein thioredoxin
reductase 1 (TXNRD1), which also was highly scored by both
NetCorona and PROSPER (Figure 2E). Note that in a larger set
of 77 human coronaviruses, a serine (S) in P4 is very common
(12), and in addition the sequence SI in the TXNRD1 site is
isosteric to TV, the aligned residues in the nsp12/13 site, because
transfer of a methyl group from threonine to valine would result
in serine and isoleucine. Thus, SI should occupy about the
same volume as TV in the Mpro active site. Functionally, this
predicted cleavage site is only 5 residues from the C terminal of
TXNRD1, and would result in removal of the C-terminal redox
center of TXNRD1 (AGCUG), making the enzyme incapable of
regenerating reduced thioredoxin.
A predicted site in selenoprotein P (SELENOP, Figure 2C)
is at position 56, 3 residues prior to the first selenocysteine,
such that cleavage here would likely interfere with the N-
terminal redox activity, possibly without affecting the Se-
transport function of the rest of the protein.
Two possible sites were identified in the catalytic subunit of
glutamate-cysteine ligase (GCLC), the rate limiting enzyme for
GSH synthesis. The first of these, GCLC-a (Figure 2B), would
cleave after position 217, and is an exact match to P4-P1 of
nsp5/6, but with an A in P1′, a highly favorable residue for that
position, although SCoV has a G there, and SCoV2 has an S.
Site GCLC-b (Figure 2A), at position 443, is an exact match to
P4-P1 of nsp4/5, but with a G in P1′. The two coronavirus sites
with exact P4-P1 matches to GCLC are those at either end of
nsp5, i.e., Mpro itself, which have the highest catalytic efficiency
for cleavage of any of the viral Mpro targets (100 and 41%). This
suggests that Mpro might be particularly efficient at disrupting
GSH synthesis, as compared to action at some of the other target
sites predicted here, thereby contributing to the clinical findings
in COVID-19 (7).
Finally (Figure 2D), we include a non-canonical hit that
would cleave at position 90 of glutaredoxin-1 (GLRX-1). This site
was very highly scored by PROSPER, as well as Procleave, despite
the glutamine (Q) at P1′, which has an additional carbon atom
relative to the asparagine (N) found at P1′in the nsp8/9 site, to
which it is otherwise most similar.
WHY WOULD SARS-CoV-2 TARGET
COMPONENTS OF THE THIOREDOXIN
AND GLUTAREDOXIN SYSTEMS?
Along with GSH reductase, GSH and glutaredoxin comprise the
glutaredoxin system, one of two cellular redox systems involved
in maintaining reduced thiol states, protein folding and repair,
and providing electrons to ribonucleotide reductase (RNR). The
other redox system with similar roles is the thioredoxin system,
comprised of thioredoxin and thioredoxin reductase (21).
Because DNA is an “add-on” to RNA biosynthesis,
deoxyribonucleotides can only be synthesized via the reduction
of ribonucleotides by RNR, a process which is unsustainable
without the participation of one or both of the thioredoxin
and glutaredoxin systems. As a consequence of this basic fact
of biochemistry, one should expect that RNA viruses might
exploit various mechanisms to interfere with components of the
thioredoxin and glutaredoxin systems, in order to minimize the
diversion of ribonucleotides into DNA synthesis. This is simply
the inverse of a strategy used by some large DNA viruses to
maximize DNA production, by encoding their own thioredoxin-
like proteins, glutaredoxins and even entire RNR genes (22).
Consistent with this hypothesis, the results presented here
suggest coronaviral targeting of TXNRD1, glutaredoxin-1, and
GCLC, a key enzyme for GSH synthesis, for proteolytic cleavage.
This would lead to knockdown of both of the essential redox
systems required to sustain DNA synthesis, and thereby conserve
ribonucleotides to enhance RNA synthesis for virus production.
THE POTENTIAL MPRO TARGETS TXNRD1
AND GCLC ARE STRONGLY
UPREGULATED BY VITAMIN D3
In light of accruing evidence that vitamin D3 status is inversely
correlated with severity of COVID-19 (23–26), it is significant
that, via its hormone-like actions on gene expression, D3 has
been shown to be a potent activator of both TXNRD1 and
GCLC. In a microarray study, 6 h after exposure to 1,25-
dihydroxyvitamin D3, TXNRD1 mRNA was upregulated 7.1X,
and GCLC was upregulated 2.5X (27). These effects of vitamin
D3, as well as a resulting increase in GSH levels, have been
functionally documented in cell culture studies (28,29) and in
human subjects (30). Vitamin D3-mediated activation of both
TXNRD1 and GSH biosynthesis could substantially counteract the
proteolytic knockdown of TXNRD1 and GCLC predicted by our
hypothesis, which, if validated, would thus represent an important
mechanism contributing to a role for vitamin D3 in moderating
the severity of COVID-19. However, the therapeutic potential of
D3 in COVID-19 is the subject of ongoing debate (31). Perhaps
inconsistencies in the findings of various studies of this question
may be explained in part by a 20-year old observation, that the
effective upregulation of TXNRD1 by vitamin D3 requires the
presence of an adequate level of Se (28). So the full potential of
vitamin D3 vs. COVID-19 may only be seen in combination with
optimal Se intake, and possibly, vice-versa.
DISCUSSION AND CONCLUSIONS
Altogether, given the protein interaction and functional data
(9,16), it is a very strong hypothesis that GPX1 is an Mpro
substrate. Of the other sites we propose, some are more
likely than others, but those in TXNRD1 and SELENOF are
particularly convincing, and all of the predicted sites map to
surface accessible regions of the targeted proteins (Figure S3).
If validated, these results offer new insights into COVID-19
pathogenesis. If SCoV2 is targeting GSH biosynthesis as well as
TXNRD1 and GPx1 for proteolytic knockdown, in infected cells,
the resulting decreases in these critical antioxidant molecules
would contribute to increased oxidative stress, NF-κB activation
and pro-apoptotic signaling (17,18). Because NF-κB is an
Frontiers in Nutrition | www.frontiersin.org 5September 2020 | Volume 7 | Article 143

Taylor and Radding Selenoproteins as SARS-CoV-2 Protease Substrates
activator of many pro-inflammatory cytokines, including IL-6
(32), this could contribute to increased inflammation and the
cytokine storms observed in COVID-19, and be a significant
basis of pathogenic effects associated with SCoV2 infection of
various tissues, including the lung, gastrointestinal tract and
cardiovascular system. Importantly, these consequences of virus-
mediated proteolysis would be taking place in everyone who is
actively infected, regardless of their Se status, and would thus
be consistent with results suggesting that Se intakes up to
several times the minimal dietary requirement are associated
with an increase in cure rate from COVID-19 (1). But people
with suboptimal nutritional status could be particularly at
risk, because their GSH and selenoprotein levels might be
low to begin with, making them more vulnerable to the
detrimental effects of virus-induced proteolysis. Still, an infection
resulting from low dose exposure to virus and limited by a
strong immune response in a person with excellent dietary
status still might have minimal impact on cellular and patient
health, because proteolytic knockdown of host proteins is
likely to be incomplete, due to low catalytic efficiencies at
some target sites (Figure 2) and the stochastic nature of such
molecular interactions.
These results are not unprecedented—our lab has shown
that some RNA viruses target host mRNAs encoding isoforms
of thioredoxin reductase via RNA:RNA antisense interactions,
which, like proteolysis, would likely also result in host
selenoprotein knockdown, but by a different mechanism (22,
33). If our hypothesis is confirmed (i.e., some of these host
cleavage sites prove to be functional), that leaves us with an
interesting question—what evolutionary advantages are driving
some viruses to go so far as to degrade or block the synthesis of
GSH and specific host selenoproteins?
Understanding why RNA viruses may have developed such
strategies presents an interesting challenge for future research.
DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the
article/Supplementary Material.
AUTHOR CONTRIBUTIONS
This study was conceived and designed by ET, who wrote the
initial draft. ET and WR worked on the data analysis and
revisions of the manuscript. All authors contributed to the article
and approved the submitted version.
FUNDING
This work has been supported by a recurring unrestricted gift
from the Dr. Arthur and Bonnie Ennis Foundation, Decatur,
IL, USA.
SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found
online at: https://www.frontiersin.org/articles/10.3389/fnut.2020.
00143/full#supplementary-material
REFERENCES
1. Zhang J, Taylor EW, Bennett K, Saad R, Rayman MP. Association between
regional selenium status and reported outcome of COVID-19 cases in China.
Am J Clin Nutr. (2020) 111:1297–9. doi: 10.1093/ajcn/nqaa095
2. Moghaddam A, Heller RA, Sun Q, Seelig J, Cherkezov A, Seibert L, et al.
Selenium deficiency is associated with mortality risk from COVID-19.
Nutrients. (2020) 12:2098. doi: 10.3390/nu12072098
3. Steinbrenner H, Al-Quraishy S, Dkhil MA, Wunderlich F, Sies H. Dietary
selenium in adjuvant therapy of viral and bacterial infections. Adv Nutr.
(2015) 6:73–82. doi: 10.3945/an.114.007575
4. Guillin OM, Vindry C, Ohlmann T, Chavatte L. Selenium, selenoproteins and
viral infection. Nutrients. (2019) 11:2101. doi: 10.3390/nu11092101
5. Hiffler L, Rakotoambinina B. Selenium and RNA viruses interactions:
potential implications for SARSCov-2 infection (COVID-19). Front Nutr.
(2020). doi: 10.3389/fnut.2020.00164
6. Taylor EW. The oxidative stress-induced niacin sink (OSINS) model for HIV
pathogenesis. Toxicology. (2010) 278:124–30. doi: 10.1016/j.tox.2009.10.018
7. Polonikov A. Endogenous deficiency of glutathione as the most likely cause
of serious manifestations and death in COVID-19 patients. ACS Infect Dis.
(2020) 6:1558–62. doi: 10.1021/acsinfecdis.0c00288
8. Horowitz RI, Freeman PR, Bruzzese J. Efficacy of glutathione therapy in
relieving dyspnea associated with COVID-19 pneumonia: a report of 2 cases.
Respir Med Case Rep. (2020) 30:101063. doi: 10.1016/j.rmcr.2020.101063
9. Gordon DE, Jang GM, Bouhaddou M, Xu J, Obernier K, White KM, et al.
A SARS-CoV-2 protein interaction map reveals targets for drug repurposing.
Nature. (2020) 583:459–68. doi: 10.1038/s41586-020-2286-9
10. Grum-Tokars V, Ratia K, Begaye A, Baker SC, Mesecar AD. Evaluating
the 3C-like protease activity of SARS-Coronavirus: recommendations for
standardized assays for drug discovery. Virus Res. (2008) 133:63–73.
doi: 10.1016/j.virusres.2007.02.015
11. Song J, Tan H, Perry AJ, Akutsu T, Webb GI, Whisstock JC, et al.
PROSPER: an integrated feature-based tool for predicting protease substrate
cleavage sites. PLoS ONE. (2012) 7:e50300. doi: 10.1371/journal.pone.
0050300
12. Kiemer L, Lund O, Brunak S, Blom N. Coronavirus 3CLpro proteinase
cleavage sites: possible relevance to SARS virus pathology. BMC
Bioinformatics. (2004) 5:72. doi: 10.1186/1471-2105-5-72
13. Li F, Leier A, Liu Q, Wang Y, Xiang D, Akutsu T, et al. Procleave:
predicting protease-specific substrate cleavage sites by combining sequence
and structural information. Genomics Proteomics Bioinformatics. (2020)
18:52–64. doi: 10.1016/j.gpb.2019.08.002
14. Orian L, Mauri P, Roveri A, Toppo S, Benazzi L, Bosello-Travain V,
et al. Selenocysteine oxidation in glutathione peroxidase catalysis: an MS-
supported quantum mechanics study. Free Radic Biol Med. (2015) 87:1–14.
doi: 10.1016/j.freeradbiomed.2015.06.011
15. Blanco R, Carrasco L, Ventoso I. Cell killing by HIV-1 protease. J Biol Chem.
(2003) 278:1086–93. doi: 10.1074/jbc.M205636200
16. Lin CW, Lin KH, Hsieh TH, Shiu SY, Li JY. Severe acute respiratory
syndrome coronavirus 3C-like protease-induced apoptosis. FEMS Immunol
Med Microbiol. (2006) 46:375–80. doi: 10.1111/j.1574-695X.2006.00045.x
17. Sappey C, Legrand-Poels S, Best-Belpomme M, Favier A, Rentier B, Piette
J. Stimulation of glutathione peroxidase activity decreases HIV type 1
activation after oxidative stress. AIDS Res Hum Retrovir. (1994) 10:1451–61.
doi: 10.1089/aid.1994.10.1451
18. Kretz-Remy C, Arrigo AP. Selenium: a key element that controls NF-kappa
B activation and I kappa B alpha half life. Biofactors. (2001) 14:117–25.
doi: 10.1002/biof.5520140116
Frontiers in Nutrition | www.frontiersin.org 6September 2020 | Volume 7 | Article 143

Taylor and Radding Selenoproteins as SARS-CoV-2 Protease Substrates
19. Kaltschmidt B, Kaltschmidt C, Hofmann TG, Hehner SP, Droge W, Schmitz
ML. The pro- or anti-apoptotic function of NF-kappaB is determined by
the nature of the apoptotic stimulus. Eur J Biochem. (2000) 267:3828–35.
doi: 10.1046/j.1432-1327.2000.01421.x
20. Rognvaldsson T, You L, Garwicz D. State of the art prediction of
HIV-1 protease cleavage sites. Bioinformatics. (2015) 31:1204–10.
doi: 10.1093/bioinformatics/btu810
21. Sengupta R, Holmgren A. Thioredoxin and glutaredoxin-mediated redox
regulation of ribonucleotide reductase. World J Biol Chem. (2014) 5:68–74.
doi: 10.4331/wjbc.v5.i1.68
22. Taylor EW. RNA viruses vs. DNA synthesis: a general viral strategy that may
contribute to the protective antiviral effects of selenium. Preprints. (2020)
2020060069. doi: 10.20944/preprints202006.0069.v1
23. Arboleda JF, Urcuqui-Inchima S. Vitamin D supplementation: a potential
approach for coronavirus/COVID-19 therapeutics? Front Immunol. (2020)
11:1523. doi: 10.3389/fimmu.2020.01523
24. Lau FH, Majumder R, Torabi R, Saeg F, Hoffman R, Cirillo JD, et al.
Vitamin D insufficiency is prevalent in severe COVID-19. medRxiv. (2020).
doi: 10.1101/2020.04.24.20075838
25. Marik PE, Kory P, Varon J. Does vitamin D status impact mortality
from SARS-CoV-2 infection? Med Drug Discov. (2020) 6:100041.
doi: 10.1016/j.medidd.2020.100041
26. Rhodes JM, Subramanian S, Laird E, Griffin G, Kenny RA. Perspective:
vitamin D deficiency and COVID-19 severity – plausibly linked by latitude,
ethnicity, impacts on cytokines, ACE2, and thrombosis (R1). J Intern Med.
(2020). doi: 10.1111/joim.13149. [Epub ahead of print].
27. Kovalenko PL, Zhang Z, Cui M, Clinton SK, Fleet JC. 1,25 dihydroxyvitamin
D-mediated orchestration of anticancer, transcript-level effects in the
immortalized, non-transformed prostate epithelial cell line, RWPE1. BMC
Genomics. (2010) 11:26. doi: 10.1186/1471-2164-11-26
28. Schutze N, Fritsche J, Ebert-Dumig R, Schneider D, Kohrle J, Andreesen
R, et al. The selenoprotein thioredoxin reductase is expressed in peripheral
blood monocytes and THP1 human myeloid leukemia cells–regulation
by 1,25-dihydroxyvitamin D3 and selenite. Biofactors. (1999) 10:329–38.
doi: 10.1002/biof.5520100403
29. Jain SK, Micinski D. Vitamin D upregulates glutamate cysteine ligase and
glutathione reductase, and GSH formation, and decreases ROS and MCP-1
and IL-8 secretion in high-glucose exposed U937 monocytes. Biochem Biophys
Res Commun. (2013) 437:7–11. doi: 10.1016/j.bbrc.2013.06.004
30. Alvarez JA, Chowdhury R, Jones DP, Martin GS, Brigham KL, Binongo JN,
et al. Vitamin D status is independently associated with plasma glutathione
and cysteine thiol/disulphide redox status in adults. Clin Endocrinol. (2014)
81:458–66. doi: 10.1111/cen.12449
31. Brown RA. COVID-19 and vitamin-D – NICE; unthinking discrimination
against BAME, Elderly and Obese? Affirmative-action needed to mitigate
D deficiency, in high-COVID-risk groups. BMJ. (2020) 369:m2475.
doi: 10.1136/bmj.m2475
32. Liu T, Zhang L, Joo D, Sun SC. NF-kappaB signaling in inflammation. Signal
Transduct Target Ther. (2017) 2:17023. doi: 10.1038/sigtrans.2017.23
33. Taylor EW, Ruzicka JA, Premadasa L, Zhao L. Cellular selenoprotein mRNA
tethering via antisense interactions with Ebola and HIV-1 mRNAs may
impact host selenium biochemistry. Curr Top Med Chem. (2016) 16:1530–5.
doi: 10.2174/1568026615666150915121633
Conflict of Interest: The authors declare that the research was conducted in the
absence of any commercial or financial relationships that could be construed as a
potential conflict of interest.
Copyright © 2020 Taylor and Radding. This is an open-access article distributed
under the terms of the Creative Commons Attribution License (CC BY). The use,
distribution or reproduction in other forums is permitted, provided the original
author(s) and the copyright owner(s) are credited and that the original publication
in this journal is cited, in accordance with accepted academic practice. No use,
distribution or reproduction is permitted which does not comply with these terms.
Frontiers in Nutrition | www.frontiersin.org 7September 2020 | Volume 7 | Article 143

Supplementary Materials
for
Taylor & Radding, 2020
Mpro site 2003 SARS-CoV 2019 SARS-CoV-2 # Non-identical
nsp4/5 SAVLQ/SGFRK SAVLQ/SGFRK 0
nsp5/6 GVTFQ/GKFKK GVTFQ/SAVKR 4
nsp6/7 VATVQ/SKMSD VATVQ/SKMSD 0
nsp7/8 RATLQ/AIASE RATLQ/AIASE 0
nsp8/9 AVKLQ/NNELS AVKLQ/NNELS 0
nsp9/10 TVRLQ/AGNAT TVRLQ/AGNAT 0
nsp10/11 EPLMQ/SADAS EPMLQ/SADAQ 3
nsp12/13 HTVLQ/AVGAC HTVLQ/AVGAC 0
nsp13/14 VATLQ/AENVT VATLQ/AENVT 0
nsp14/15 FTRLQ/SLENV FTRLQ/SLENV 0
nsp15/16 YPKLQ/ASQAW YPKLQ/SSQAW 1
Figure S1. The 11 known Mpro cleavage sites in the 2003 and 2019 SARS coronaviruses with a
comparison of the 10-residue protein sequences spanning the cleavages sites in the two viruses.
The alignments underneath each of the virus names (2003 SARS-CoV and 2019 SARS-CoV-2)
were used to generate the logos shown in Fig. 1A and 1B. The cleavage sites and sequences for
the SARS-CoV (Genbank NC_004718.3) are given in reference (10), and the equivalent sequences
in SARS-CoV-2 were taken from its Genbank Reference Sequence, NC_045512.

6:HEVHUYHUSURWHDVHFOHDYDJHVLWHSUHGLFWLRQUHVXOWVFLWHGLQ)LJXUH
*HQEDQNVRXUFHILOHVIRUWKHSURWHLQVHTXHQFHV*&/&13B6(/(12313B
*/5;13B7;15'13B6(/(12)13B*3;13B
61HW&RURQDUHVXOWVIURPFEVGWXGNVHUYLFHV1HW&RURQD

1HW&RURQD6HUYHUSUHGLFWLRQUHVXOWV
7HFKQLFDO8QLYHUVLW\RI'HQPDUN

165 SelenoF
MVAMAAGPSGCLVPAFGLRLLLATVLQAVSAFGAEFSSEACRELGFSSNLLCSSCDLLGQFNLLQLDPDCRGCCQEEAQF 80
ETKKLYAGAILEVCGCKLGRFPQVQAFVRSDKPKLFRGLQIKYVRGSDPVLKLLDDNGNIAEELSILKWNTDSVEEFLSE 160
KLERI Selenocysteine (U) at #96 changed to C for search as NetCorona ignores U
..........................C..................................................... 80
................................................................................ 160
.....
Pos Score Cleavage
___________________________
27 0.846 ATVLQ^AVSAF SelenoF
60 0.067 none SelenoF
65 0.073 none SelenoF
75 0.066 none SelenoF
79 0.064 none SelenoF
103 0.064 none SelenoF
105 0.219 none SelenoF
120 0.136 none SelenoF
___________________________

499 TXNRD1
MNGPEDLPKSYDYDLIIIGGGSGGLAAAKEAAQYGKKVMVLDFVTPTPLGTRWGLGGTCVNVGCIPKKLMHQAALLGQAL 80
QDSRNYGWKVEETVKHDWDRMIEAVQNHIGSLNWGYRVALREKKVVYENAYGQFIGPHRIKATNNKGKEKIYSAERFLIA 160
TGERPRYLGIPGDKEYCISSDDLFSLPYCPGKTLVVGASYVALECAGFLAGIGLDVTVMVRSILLRGFDQDMANKIGEHM 240
EEHGIKFIRQFVPIKVEQIEAGTPGRLRVVAQSTNSEEIIEGEYNTVMLAIGRDACTRKIGLETVGVKINEKTGKIPVTD 320
EEQTNVPYIYAIGDILEDKVELTPVAIQAGRLLAQRLYAGSTVKCDYENVPTTVFTPLEYGACGLSEEKAVEKFGEENIE 400
VYHSYFWPLEWTIPSRDNNKCYAKIICNTKDNERVVGFHVLGPNAGEVTQGFAAALKCGLTKKQLDSTIGIHPVCAEVFT 480
TLSVTKRSGASILQAGCCG (Penultimate U changed to C as NetCorona ignores U)
................................................................................ 80
................................................................................ 160
................................................................................ 240
...............................C................................................ 320
................................................................................ 400
................................................................................ 480
.............C.....
Pos Score Cleavage
___________________________
33 0.068 none TXNRD1
72 0.104 none TXNRD1
78 0.071 none TXNRD1
81 0.123 none TXNRD1
106 0.073 none TXNRD1
133 0.071 none TXNRD1
230 0.075 none TXNRD1
250 0.070 none TXNRD1
258 0.065 none TXNRD1
272 0.505 RVVAQ^STNSE TXNRD1
323 0.061 none TXNRD1
348 0.191 none TXNRD1
355 0.064 none TXNRD1
450 0.107 none TXNRD1
464 0.074 none TXNRD1
494 0.640 ASILQ^AGCCG TXNRD1
___________________________

635263(5WKH3URWHDVH6SHFLILFLW\3UHGLFWLRQ6HUYHUUHVXOWVIURPSURVSHUHUFPRQDVKHGXDX
6%
B. GCLC-a VTFQ/A WZK^WZ;ϭ͘ϬϬͿ
nsp5/6* VTFQ/G
6'
D. GLRX-1 VSLQ/Q WZK^WZ;ϭ͘ϯϬͿ
nsp8/9 VKLQ/N


 

6(
E. TXNRD1
SILQ/A WZK^WZ;ϭ͘ϭϰͿ
nsp12/13 TVLQ/A
 
6)
F. SelenoF TVLQ/A WZK^WZ;ϭ͘ϬϮͿ
nsp12/13 TVLQ/A


63URFOHDYHUHVXOWVIURPSURFOHDYHHUFPRQDVKHGXDX
6$
A. GCLC-b AVLQ/G WƌŽůĞĂǀĞηϳͬϱϵϮĨŽƌĂƚŚĞƉƐŝŶ^
nsp4/5 AVLQ/S

 

6%

B. GCLC-a VTFQ/A WƌŽůĞĂǀĞηϭϯŽĨϱϵϮĨŽƌĂƚŚĞƉƐŝŶ>
nsp5/6* VTFQ/G


6&
C. SelenoP
ALLQ/A WƌŽůĞĂǀĞηϯͬϯϳϰĨŽƌĂƚŚĞƉƐŝŶ>
nsp7/8 ATLQ/A

6'
D. GLRX-1 VSLQ/Q WƌŽůĞĂǀĞηϭͬϵϵĨŽƌĂƚŚĞƉƐŝŶ^
nsp8/9 VKLQ/N
6*
G. GPX1 ASLU/G WƌŽůĞĂǀĞηϮͬϭϵϲĨŽƌĂƚŚĞƉƐŝŶ^
nsp13/14 ATLQ/A


)LJXUH60DSSLQJRISUHGLFWHG0SURFOHDYDJHVLWHVWRDFFHVVLEOHVXUIDFHUHJLRQVRIWKHWDUJHWSURWHLQV

















&ŝŐ͘ ^ϯͲ͘ ,ŽŵŽůŽŐLJŵŽĚĞůŽĨ'>ƐŚŽǁŝŶŐůŽĐĂƚŝŽŶŽĨ
ƐŝƚĞ'>Ͳď͕s>Yͬ';ĐŝƌĐůĞĚͿ͘dŚĞ'ůŶŝŶWϭŝƐƐŚŽǁŶ͘dŚĞ
'>ŚŽŵŽůŽŐLJŵŽĚĞůƵƐĞĚŝŶůŝĞƵŽĨĂŶyͲƌĂLJƐƚƌƵĐƚƵƌĞŝƐ
ďĂƐĞĚ ŽŶ ƚŚĞ LJĞĂƐƚ ĞŶnjLJŵĞ W ĨŝůĞ ϯŝŐϱ͘ϭ ĂŶĚ ŝƐ ĨŽƵŶĚ
ŚĞƌĞ͗ƐǁŝƐƐŵŽĚĞů͘ĞdžƉĂƐLJ͘ŽƌŐͬƌĞƉŽƐŝƚŽƌLJͬƵŶŝƉƌŽƚͬWϰϴϱϬϲ













&ŝŐ͘^ϯͲ͘EDZƐƚƌƵĐƚƵƌĞŽĨŚƵŵĂŶ'>ZyͲϭ;WĨŝůĞϭũŚďͿ͘
dŚĞƉƌĞĚŝĐƚĞĚDƉƌŽĐůĞĂǀĂŐĞƐŝƚĞƌĞŐŝŽŶ ƐŚŽǁŶŝŶ&ŝŐ͘Ϯ͕
s^>YͬY͕ŝƐĐŝƌĐůĞĚĂŶĚƚŚĞ'ůŶĂƚWϭŝƐĚŝƐƉůĂLJĞĚ͘











&ŝŐ͘ ^ϯͲ'͘ yͲƌĂLJ ƐƚƌƵĐƚƵƌĞ ŽĨ ŵŽŶŽŵĞƌŝĐ 'Wyϭ ;W ĨŝůĞ
ϮĨϴĂͿ͘dŚĞƉ ƌĞĚŝĐƚĞĚDƉƌŽĐůĞĂǀĂŐĞƐŝƚĞƌĞŐŝŽŶƐŚŽǁŶŝŶ&ŝŐ͘
Ϯ'͕^>hͬ'͕ŝƐĐŝƌĐůĞĚĂŶĚƚŚĞ>ĞƵĂƚWϮŝƐĚŝƐƉůĂLJĞĚ͘

&ŝŐ͘ ^ϯͲ͘dŚĞƐĂŵĞŚŽŵŽůŽŐLJŵŽĚĞůŽĨ'>ƐŚŽǁŝŶŐ
ůŽĐĂƚŝŽŶŽĨ ƐŝƚĞ'>ͲĂ͕sd&Yͬ;ĐŝƌĐůĞĚͿ͘ dŚĞ'ůŶŝŶWϭŝƐ
ƐŚŽǁŶ͘dŚĞƐŝƚĞŝƐĂƚƚŚĞƉƌŽƚĞŝŶƐƵƌĨĂĐĞďƵƚŝƐƌĞĐĞƐƐĞĚŝŶ
Ă ĐŚĂŶŶĞů ;ŝŶƐĞƚͿ͘ dŚĞƌĞ ŝƐ ƐŽŵĞ ƵŶĐĞƌƚĂŝŶƚLJ ŝŶ ƚŚĞ ϯ
ƐƚƌƵĐƚƵƌĞ͕ďƵƚĂŶĞdžƉĞƌŝŵĞŶƚĂůƐƚƌƵĐƚƵƌĞǁĂƐŶŽƚĂǀĂŝůĂďůĞ͘













&ŝŐ͘^ϯͲ͘yͲƌĂLJƐƚƌƵĐƚƵƌĞŽĨdyEZϭ;WĨŝůĞϮnjnjĐͿ͘dŚĞͲ
ƚĞƌŵŝŶĂů ǁŝƚŚ ƚŚĞ ƉƌĞĚŝĐƚĞĚ DƉƌŽ ĐůĞĂǀĂŐĞ ƐŝƚĞ ƌĞŐŝŽŶ
ƐŚŽǁŶŝŶ&ŝŐ͘Ϯ͕^/>Yͬ͕ŝƐĐŝƌĐůĞĚĂŶĚƚŚĞ>ĞƵĂƚWϮŝƐ
ĚŝƐƉůĂLJĞĚ͘dŚĞ'h'ͲƚĞƌŵŝŶĂůƌĞĚŽdžĐĞŶƚĞƌŽĨdyEZϭ
ǁŽƵůĚďĞƌĞŵŽǀĞĚďLJDƉƌŽĐůĞĂǀĂŐĞĂƚƚŚŝƐƐŝƚĞ͘


&ŝŐ͘ ^ϯͲ ĂŶĚ ^ϯͲ&͗  ^ƚƌƵĐƚƵƌĞƐ ǁĞƌĞ ŶŽƚ ĂǀĂŝůĂďůĞ ĨŽƌ
^>EK&ĂŶĚ^>EKW͕ďƵƚĨŽƌƚŚĞƐĞƉƌŽƚĞŝŶƐ͕ƚŚĞ
ƉƌĞĚŝĐƚĞĚ DƉƌŽ ĐůĞĂǀĂŐĞ ƐŝƚĞ ǁĂƐ ĞŝƚŚĞƌ ŶĞĂƌ ƚŚĞ EͲ
ƚĞƌŵŝŶĂů Z ƐŝŐŶĂů ƉĞƉƚŝĚĞ ĐůĞĂǀĂŐĞ ƐŝƚĞ ;ƉŽƐŝƚŝŽŶ Ϯϳ ŽĨ
^>EK&Ϳ Žƌ ƌŝŐŚƚ ďĞĨŽƌĞ Ă ƌĞĚŽdž ĐĞŶƚĞƌ ĂĐƚŝǀĞ ƐŝƚĞ
;ƉŽƐŝƚŝŽŶϱϲŽĨ^>EKWͿ͕ƐŽƚŚĂƚĂŚŝŐŚƉƌŽďĂďŝůŝƚLJŽĨ
ƐƵƌĨĂĐĞĂĐĐĞƐƐŝďŝůŝƚLJĐĂŶďĞŝŶĨĞƌƌĞĚŝŶďŽƚŚĐĂƐĞƐ͘
