Tinkering with a viral ribonucleotide reductase.

al ribonucleotide
Turin, S. Luigi Gonzaga Hospital, 10043, Orbassano, Turin, Italy
, 13353 Berlin, Germany
residues. It follows that they do not express a functional
RNR. Why do b-herpesviruses encode R1 homologue genes
devoid of enzymatic activity? This long-standing biological
enigma was recently unraveled for one b-herpesvirus: the
murine CMV (MCMV).
Review
sphate by thymidylate synthase.
viruses have been divided into three phylogenetically dis-
tinct subfamilies: the a-, b- and g-herpesvirinae. Some
avian and reptile herpesviruses also fall in the a- subfamily
[4]. Throughout this review, we refer to only the mamma-
lian herpesviruses.
a- and g-herpesviruses encode two class Ia RNR sub-
units (R1 and R2) (Box 1), thereby expressing a functional
enzyme [5] that is required for virus growth in non-dividing
cells and for viral pathogenesis and reactivation from
latency in infected hosts [6–10]. By contrast, b-herpes-
viruses, including CMV, do not harbor a gene for the R2
subunit and their R1 subunits lack important catalytic
Dihydrofolate reductase (DHFR): EC 1.5.1.3; catalyzes the NADPH-dependent
reduction of folic acid and dihydrofolic acid to tetrahydrofolic acid. Derivatives
of tetrahydrofolate are required for many facets of single-carbon metabolism,
including the biosynthesis of purines and thymidylic acid.
Folylpolyglutamate synthase (FPGS): EC 6.3.2.17; converts folates to poly-
glutamated forms. Polyglutamated folates are selectively retained within the
cell and have an increased affinity for DHFR and thymidylate synthase.
Thymidine kinase: EC 2.7.1.21; Thymidine kinase (TK) was first identified as a
deoxythymidine kinase but it phosphorylates pyrimidines and even purine
nucleosides. This is a key step in the salvage pathway of nucleosides. TK is
presumed to provide triphosphate nucleosides for viral DNA synthesis in
resting cells where the cellular enzyme is not expressed. Its broad substrate
specificity enables phosphorylation of antiviral nucleoside analogs (e.g.
acyclovir).
Thymidylate synthase: EC 2.1.1.45; converts deoxyuridine monophosphate
(dUMP) into thymidine monophosphate (TMP), a key step in deoxythymidine
triphosphate biosynthesis.
Corresponding author: Lembo, D. (david.lembo@unito.it).organization and cellular tropism, mammalian herpes-
(KSHV orHHV-8). Based on differences in genome content, lyzes dUTP to dUMP and pyrophosphate, preventing the incorporation of uracil
into DNA. Moreover, supplies dUMP for synthesis of thymidine monopho-Tinkering with a vir
reductase
David Lembo1 and Wolfram Brune2
1Department of Clinical and Biological Sciences, University of
2Division of Viral Infections, Robert Koch Institute, Nordufer 20
Ribonucleotide reductase (RNR), a crucial enzyme for
nucleotide anabolism, is encoded by all living organisms
and by large DNA viruses such as the herpesviruses.
Surprisingly, the b-herpesvirus subfamily RNR R1 sub-
unit homologues are catalytically inactive and their func-
tion remained enigmatic for many years. Recent work
sheds light on the function of M45, the murine cytome-
galovirus R1 homologue; during viral evolution, M45
apparently lost its original RNR activity but gained the
ability, via inhibiting RIP1, a cellular adaptor protein, to
block cellular signaling pathways involved in innate
immunity and inflammation. The discovery of this novel
mechanism of viral immune subversion provides further
support to the concept of evolutionary tinkering.
The b-herpesviruses ribonucleotide reductase R1
subunits: a biological enigma
Coined by Franc¸ois Jacob in 1977, the term ‘evolutionary
tinkering’ refers to evolutionary innovation through the
generation of new proteins and functions from pre-existing
modules [1]. A striking example of evolutionary tinkering
was recently discovered in cytomegalovirus (CMV): the
conversion of a ribonucleotide reductase (RNR) large sub-
unit (R1) into an inhibitor of pro-inflammatory and innate
immune signaling pathways [2,3].
RNRs synthesize the four deoxyribonucleotides (dNTPs)
that are required for DNA synthesis (Box 1); thus, RNRs
are expressed by all living organisms and by large DNA
viruses. TheHerpesviridae family is one of the best charac-
terized families of large DNA viruses. Among its members
are human pathogens including herpes simplex virus
(HSV), varicella zoster virus (VZV), CMV, Epstein-Barr
virus (EBV) and Kaposi’s sarcoma-associated herpesvirus0968-0004/$ – see front matter � 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.tibs.2008.In this review, we summarize the current knowledge on
the evolutionary gain and loss of nucleotide anabolic
enzymes in the herpesviruses; we describe a newly
acquired function identified in the R1 homologue of MCMV
and discuss how this function might be shared by other
herpesviruses.
Viral strategies to gain DNA precursors
dNTPs are precursors that living organisms and DNA
viruses need for the replication of their genomes. The
concentration of intracellular dNTPs can vary widely at
different stages of the cell cycle and between different cell
types, being very low in quiescent, terminally differen-
tiated cells and high in mitotically active cells. The
majority of cells in an adult organism are differentiated
and maintained in a quiescent state. Therefore, during
their evolution, DNA viruses faced the formidable chal-
lenge of gaining access to a sufficient supply of free dNTPs
in a largely inhospitable environment. In general terms,
they adopted three distinct strategies to ensure efficient
viral DNA replication [11]. In the first strategy, small DNA
viruses, such as parvoviruses, preferentially infect mitoti-
cally active cells (e.g. erythrocyte precursors). By contrast,
Glossary
Deoxycytidylate (dCMP) deaminase: EC 3.5.4.12; provides the substrate for
thymidylate synthase by converting dCMP to dUMP.
Deoxyuridine triphosphate pyrophosphatase (dUTPase): EC 3.6.1.23; hydro-09.008 Available online 5 November 2008 25

the synthesis of dNTPs. A third strategy is taken by large
DNA viruses, for example, herpesviruses and poxviruses.
Owing to the size of their double-stranded DNA genomes
(up to 350 kbp), these viruses have sufficient genetic
capacity to harbor several genes encoding functional hom-
ologues of cellular nucleotide anabolic enzymes. Sim-
ilarities between viral and cellular enzymes indicate
that the former were captured from the host and then
retained because they increased viral fitness [12,13]. This
adaptation made large DNA viruses less dependent on the
enzymatic machinery of the host, enabling their genome
replication to occur independently of the host cell-cycle
Box 1. Ribonucleotide reductases
RNRs catalyze the substitution of the 20OH- of a ribonucleoside di- or
triphosphate by a hydrogen atom, providing a balanced supply of
the four dNTPs required for DNA replication and repair [23]. They
are present in all living organisms, in some bacteriophages [52,53]
and in large DNA viruses that infect eukaryotes [54].
Although all RNRs share a common catalytic mechanism invol-
ving protein radicals, they are grouped into three classes according
to their subunit composition, their interaction with oxygen and the
mechanism used for radical generation [23].
Class I RNRs
Class I reductases have an a2b2 structure consisting of two
homodimeric subunits of proteins R1 (a2) and R2 (b2). The large
Review Trends in Biochemical Sciences Vol.34 No.1R1 subunit contains the catalytic site and the allosteric site for
substrate specificity. The small R2 subunit is a radical storage device
containing an iron center-generated tyrosyl free radical, which is
continuously shuttled to a cysteine of the R1 subunit where it
generates a thiyl radical that is required for the activation of the
substrate. The reaction requires oxygen and an external electron
donor. Class I enzymes are grouped into two subclasses (Ia and Ib).in the second strategy, adenoviruses, polyomaviruses and
papillomaviruses encode proteins that disrupt cell growth-
control pathways governed by the tumor suppressors pRb
(retinoblastoma) and p53, thereby driving quiescent cells
to enter the S phase of the cell cycle. These two strategies
rely closely on the host cellular enzymatic machinery for
The electron donor is thioredoxin or glutaredoxin for class Ia and
the glutaredoxin-like NrdH protein for class Ib. Most of the class Ia
enzymes possess an N-terminal ATP cone codifying an allosteric site
that regulates the overall activity of the enzymes. By contrast, class
Ib lacks the ATP cone and, therefore, is not inhibited by dATP.
Class Ia has been found in aerobic eubacteria, in few arche-
abacteria, in all eukaryotes with one exception (i.e. Euglena gracilis),
in bacteriophages and in large DNA viruses. Class Ib occurs in a
large spectrum of eubacteria.
Class II RNRs
These enzymes contain a single subunit and have simple structure
(a or sometimes a2). They use thioredoxin or glutaredoxin as
electron donors and generate their thiyl radical with the aid of
adenosylcobalamin. Class II enzymes are indifferent to oxygen and
occur in both aerobic and anaerobic eubacteria and archeabacteria
and in some bacteriophages.
Class III RNRs
Like class I RNRs, class III enzymes have an a2b2 structure. The b2
dimer contains a redox-active iron-sulfur center that, together with
S-adenosyl methionine and reduced flavodoxin, generates a glycyl
radical at the C terminus of the a subunits. They use formate as an
electron donor and rely on anaerobiosis. Class III RNRs occur in
facultative anaerobic and in strictly anaerobic organisms.
Table 1. Distribution of dNTPs anabolic enzymes among herpesvi
Subfamily Virus Thymidine
kinase
DHFR
a HSV-1 + �
a HSV-2 + �
a VZV + �
b CMVsd � �
b HHV-6 � �
b HHV-7 � �
g EBV + �
g KSHV + +
aAbbreviations: NF: non functional.
bInactive in enzymatic assays.
cDeduced by sequence analysis.
dCMVs: human, chimpanzee, rat and mouse cytomegalovirus.
26phase. Consistent with this view is the observation that
virus-encoded DNA metabolic enzymes, although often
dispensable for viral growth in proliferating cultured cells,
are important determinants of viral replication in different
tissues in vivo [6,8–10,14].
Distribution of nucleotide anabolic enzymes in the
Herpesviridae family
Analyses of sequenced genomes of the a- and g-herpes-
viruses showed the universal existence of genes encoding
functional equivalents of cellular nucleotide anabolic
enzymes [13,15]. This finding clearly indicates that they
have adopted the third replicative strategy, albeit with
different nuances (Table 1). Surprisingly, b-herpesviruses
seem to have abandoned the strategy of supplying nucleo-
tide anabolic enzymes because their genomes lack most of
the genes involved in the DNA precursor biosynthesis,
namely, a thymidine kinase (see Glossary), a thymidylate
synthase, a dihydrofolate reductase and the R2 subunit. By
contrast, the genes for the R1 subunit and the deoxyuridine
triphosphate pyrophosphatase (dUTPase) harbor
mutations and encode catalytically inactive proteins [16–
19]. The MCMV R1 homologue, the M45 protein, is a
paradigm of the latter case. Owing to the species specificity
of the b-herpesviruses, M45 attracted much attention.
Indeed, because the use of established animal models is
prerequisite for addressing the in vivo relevance of R1
proteins, it follows that MCMV is an attractive b-herpes-
virus member for such studies.
The MCMV M45 protein is not a functional R1 subunit
CMV genomes harbor an open reading frame (ORF) that
shares homology with the R1 subunits of other herpes-
viruses. This ORF was termed UL45 in human CMV
(HCMV) and M45 in MCMV, respectively [20,21]. The idea
rus subfamiliesa
Thymidylate
synthase
dUTPase RNR
R1 R2
� + + +
� + + +
+ + + +
� NFb NFb �
� NFc NFc �
� NFc NFb �
� + + +
+ + + +

that CMVs encode R1 subunits was based on several
considerations. Basic local alignment search tool (BLAST)
searches of the SWISS-PROT protein database (http://
blast.ncbi.nlm.nih.gov/Blast.cgi) identified that the M45
C-terminal region is homologous to class Ia R1 subunits.
The closest matches occur with other herpesvirus R1 sub-
units (e.g. 27% amino acid identity to the HCMV R1
subunit, 24% to the HHV-6 R1 subunit and 22% to the
HSV-1 and HSV-2 R1 subunits). Moreover, an in silico
analysis of the HCMV genome confirmed that the UL45
protein was structurally similar to an R1 subunit [22].
Finally, the genomic positions of the UL45 and the M45
genes are conserved, that is, they are encoded at an ana-
logous position to the R1 homologues of other herpes-
viruses. By contrast, the lack of an apparent R2 subunit
homologue in the b-herpesviruses genomes led to the
hypothesis that their R1 homologues are not functional
equivalents of a R1 subunit. This was further supported by
an amino acid sequence comparison between M45 and the
Escherichia coli R1 subunit, which represents the class Ia
prototype and the HSV and EBV R1 proteins, chosen as
representatives of a- (HSV-1 and -2) and g-herpesviruses
(EBV) (Figure 1). This sequence alignment revealed that
the residues known to have a catalytic role in theE. coli R1
subunit [23] are highly conserved in the HSV and EBV R1
proteins and in the human R1. These include the E. coli
cysteines 225, 439, 462, 754, 759 and the two tyrosines at
positions 730 and 731. The first two cysteines provide
sulphydryl groups at the active site, whereas the two at
the very C terminus are involved in the transfer of reducing
power to the active site (the presence of a unique C-term-
inal extension does not enable a clear identification of the
two latter cysteines in the R1 proteins of g-herpesviruses).
The two tyrosines have a role in a reversible radical
electron-transfer path between the tyrosyl radical in the
R2 subunit and the active site in R1. Of note, in the M45
sequence, only Cys814 (corresponding to Cys439 in E. coli)
and the Tyr1163 (corresponding to Tyr730 in E.coli) are
conserved. Moreover, the proposed GxGxxG nucleotide-
binding site, which is well conserved in all a- and g-
herpesviruses [24], is only partially conserved in the
M45 sequence (xxGxxG at residues 890–895). Sequence
alignments with the R1 proteins of all the b-herpesviruses
for which sequence information is available show that the
lack of important catalytic residues is a general feature of
this subfamily (Figure 1).
Given that M45 and UL45 bear sequence features of
class Ia R1 subunits, they were expected to complex with
ture
een
1 pr
nge
(Gx
RNR
r pro
-crys
Review Trends in Biochemical Sciences Vol.34 No.1Figure 1. Organization and alignment of herpesvirus R1 proteins. Sequence fea
compared with those of the functional R1 proteins of a- (blue) or g-herpesviruses (gr
and HSV-2) and g-herpesviruses (EBV), respectively. The E. coli and the human R
subunits, respectively. The C-terminal region of the b-herpesvirus R1 proteins (ora
residues known to have a direct catalytic role. The proposed nucleotide-binding site
cone (yellow), which turns on-off the activity of mammalian and prokaryotic class Ia
represent N-terminal extensions of variable size, which are absent in mammalian o
shown as follows: the putative protein kinase (PK; dark blue) domain and the a
herpesviruses sequences (short gray bar in the EBV sequence). HCMV, CCMV, RCMV
respectively. Abbreviations: TuHV-1, tupaia herpesvirus 1; HHV-6 and -7, human herpes of the enzymatically inactive R1 homologues of b-herpesviruses (orange) are
). The HSV-1, HSV-2 and EBV R1 proteins are shown as representatives of a- (HSV-1
oteins are shown as prototypes of active prokaryotic and mammalian class Ia R1
bars) share homology with the cellular or viral counterparts but lacks most of the
GxxG) the catalytically active five cysteines and two tyrosines are shown. The ATP
by binding allosteric effectors, is absent in herpesvirus R1 proteins. The gray bars
karyotic R1 proteins. Specific functional domains in the N-terminal extensions are
tallin (ac; turquoise) domain. Short C-terminal extensions are present in the g-
, and MCMV are cytomegaloviruses from humans, chimpanzees, rats and mice,
sviruses 6 and 7.
27

an R2 subunit to form an active enzyme. This idea formed
the basis for the hypothesis that M45 or UL45might usurp
the place of the cellular R1 subunit, thereby forming an
inducible hybrid version of the RNR enzyme together with
the cellular R2 subunit, the expression of which is upre-
gulated by MCMV and HCMV infection [17,18]. However,
neither recombinant nor native immunopurifiedM45 seem
to be a functional equivalent of a class Ia R1 subunit,
because RNR activity assays indicated a lack of activity
in conjunction with the cellular R2 subunit. Moreover, M45
alone lacks activity [18], thus, excluding the possibility
that it might behave as a homomeric RNR, such as the
class II enzymes expressed by Eubacteria and Archeabac-
teria (Box 1). Similarly to M45, two other b-herpesviral R1
homologues, the HCMV UL45 and the human herpesvirus
7 (HHV-7) U28 proteins are also catalytically inactive
[16,17]. These findings are consistent with the view that
b-herpesviruses differ from the a- and g- subfamilies in
their strategy to satisfy their need for DNA precursors.
Several studies reported that CMVs developed the ability
to force a quiescent cell to express the cellular enzymes
involved in thymidylate biosynthesis (folylpolyglutamate
synthase, dihydrofolate reductase, deoxycytidylate
[dCMP] deaminase, thymidylate synthase and thymidine
kinase) and in ribonucleotide reduction (R1 and R2 sub-
units), thereby expanding all four dNTP pools. Moreover,
these events are essential for viral replication in resting
cells [17,18,25]. This strategy differs from the one exploited
by adenoviruses, polyomaviruses and papillomaviruses
because bothHCMVandMCMVhave evolvedmechanisms
to prevent cellular DNA synthesis [26–28], thereby provid-
ing the viral DNA polymerase with competition-free access
to the cellular dNTP pools.
The emergence of this particular replicative strategy
during the course of CMV evolution has enabled the se-
lective elimination from the viral genome of many genes
involved in nucleotide metabolism or, alternatively, per-
mitted them to mutate and gain new functions, as seen for
the RNR R1 genes. In this scenario, molecular tinkering
with anRNR could occur without gene duplication, because
the function of the original gene was replaced by a virus-
induced cellular counterpart.
A new function for the MCMV M45 protein
Before a function for M45 was identified, several consider-
ations indicated that it is not a vestigial gene encoding a
useless function. First, given their compact genomes and
the strong evolutionary pressures they face, it is unlikely
that viruses harbor useless genes. Because the M45 read-
ing frame has remained open, it was reasonable to presume
that it encodes a functional protein. Second, its abundant
expression at late times (24–48 h post-infection) and its
association with the viral particle [18] indicated that M45
was involved in virus maturation or in a function immedi-
ately following virus entry, the time when it is delivered to
the cytoplasm of newly infected cells. Third, MCMV
mutants carrying an inactivated M45 gene do not replicate
in endothelial cell lines and grow poorly on macrophages
(cell types that have key roles in CMV dissemination) but
Reviewalmost normally on fibroblasts, and M45 protects these
cells from rapid virus-induced cell death [29]. Thus, viral
28gene products that regulate growth in macrophages and
endothelial cells should have a considerable affect on
MCMV replication and dissemination in vivo. Indeed,
M45 mutant viruses replicate poorly and are avirulent
in severe combined immunodeficient (SCID) mice [18],
even though these mice are highly susceptible to MCMV
owing to their defective immune system. These findings
pointed to a direct role for M45 in viral pathogenesis.
Nevertheless, the exact function of M45 remained enig-
matic until very recently.
One way to elucidate an action mechanism of a protein
is to search for interacting proteins. In a recent study from
our laboratories, the affinity purification of M45 from
MCMV-infected fibroblasts identified the cellular adaptor
protein receptor-interacting protein 1 (RIP1) as an inter-
action partner [2]. Another study fromEdwardMocarski’s
laboratory identified a RIP homotypic interaction motif
(RHIM) inM45 and used this finding as a starting point for
further investigations [3]. RIP1 is a serine–threonine
kinase that functions as an essential mediator of cellular
stress [30,31]. By receiving signals from receptors at the
cell surface and from intracellular stress and damage
sensors, RIP1 mediates, on the one hand, the expression
of pro-inflammatory genes by activating the transcription
factors nuclear factor (NF)-kBand activator protein 1 (AP-
1) and triggers, on the other hand, pathways leading to
apoptosis or necrosis (Figure 2). The crucial role of RIP1 at
the intersection of various signaling pathwaysmakes it an
attractive target for viral interference. Hence, the discov-
ery of the M45–RIP1 interaction immediately indicated
that M45 functions as a RIP1 inhibitor. This notion is
supported by the results of two recent studies that ana-
lyzed the impact of M45 on different RIP1-dependent
processes [2,3] (Figure 2). The first study showed that
M45 blocks RIP1-dependentNF-kBactivation after tumor
necrosis factor (TNF) receptor 1 (TNF-R1) or Toll-like
receptor 3 (TLR3) stimulation. Moreover, M45 prevents
RIP1 ubiquitylation, a process that is required for NF-kB
activation. M45 also inhibits p38 mitogen-activated
protein kinase (MAPK) activation and blocks caspase-
independent programmed cell death (i.e. necrosis or
necroptosis) after TNF-R1 stimulation [2]. The second
study focused on cell death inhibition by M45, showing
that M45 inhibits programmed cell death triggered by
TNFa, by overexpression of TIR domain-containing
adaptor inducing interferon b (TRIF) or by infection with
anM45-deficientMCMV. The study also showed thatM45
interacts with RIP3 in a RHIM-dependent manner and
further indicated that M45 might interact directly with
TRIF [3]. Taken together, these results indicate that M45
functions as a viral inhibitor of RIP1-mediated signaling
(vIRS) and that it blocks many (if not all) RIP1-dependent
signals.
Which parts of M45 are crucial for its function?
M45 is a 1174 amino acid protein consisting of a C-terminal
R1 homology domain and a unique N-terminal domain
lacking discernable similarity to any protein of known
function. The protein is sensitive to proteolytic cleavage
Trends in Biochemical Sciences Vol.34 No.1between amino acids 277 and 278 and is present in both
cleaved and uncleaved forms in infected cells [18].

ReviewThe two recent reports on the function of M45 also
assessed which parts ofM45 are required for its interaction
with RIP1 and its function as a vIRS (Figure 3b). One study
showed that, whereas a large portion of the N-terminal
domain is dispensable, the C-terminal portion of the R1
homology domain is required for inhibiting RIP1-mediated
NF-kB activation [2]. The C-terminal (but not the N-term-
inal) end was also required for the interaction with
endogenous RIP1 inmurine fibroblasts [2]. The other study
attributed the ability of M45 to inhibit TNFa- or TRIF-
induced cell death to the unique N-terminal portion of the
protein, particularly to the RHIM domain near the N
terminus [3]. The RHIM domain, but not the C-terminal
R1 homology domain, is required for the M45–RIP1 inter-
action in transfected 293T cells [3].
Both recent publications agree in that M45 interacts
with RIP1 and inhibits RIP1-dependent signaling path-
ways. However, the two studies come to different con-
clusions regarding the requirement of the M45 N- or the
C-terminal portions for a physical interaction with RIP1 in
co-immunoprecipitation experiments. The reason for this
discrepancy is unclear and should be addressed in future
Figure 2. The role of RIP1 in diverse signaling pathways. The adaptor protein RIP1
transcription factors involved in innate immunity and inflammation. RIP1 can also trigg
(light blue) [46], of Toll-like receptors 3 and 4 (TLR3 and TLR4, blue) [32,47], the RNA heli
associated protein 5 (Mda5; blue) [48,49] and by the p53-inducible death domain-contai
(usually of viral origin), whereas TLR4 is activated by bacterial lipopolysaccharide (LPS
complexes also comprise the TNF receptor-associated death domain (TRADD, orange),
(purple) and the CARD domain-containing adaptor inducing interferon b (Cardif, blue). R
RIP homotypic interaction motif (RHIM, brown bar). The activation of RIP1-dependen
depicted in (b). These include the induction of apoptosis via Fas-associated death doma
involving the production of reactive oxygen species (ROS), and the activation of transcri
AP-1 together induce the expression of proinflammatory cytokines and NF-kB and IRF3
inhibit several of the RIP1-mediated pathways (shown by red lines) and probably inhib
kinase 1.Trends in Biochemical Sciences Vol.34 No.1studies. It is possible that the experimental conditions of
the co-immunoprecipitation experiments influenced the
results. Once mutated versions of the M45 gene have been
integrated into the MCMV genome, it should be possible to
determine which interactions occur under physiological
conditions in MCMV-infected cells.
The definition of a ‘RIP homotypic interaction motif’
might indicate that interactions between RHIM-contain-
ing proteins and downstream signal transductions are
strictly RHIM-dependent. However, it is unclear if this
is really the case. A previous study showed that the TRIF
RHIM domain is necessary for interaction with RIP1 and
NF-kB activation [32], but two other studies detected
TRIF-induced NF-kB activation even when the TRIF
RHIM was mutated or the RHIM-containing C terminus
was deleted [33,34]. Hence, it seems that not all down-
stream signals are controlled by RHIM-dependent inter-
actions.
Available data indicate functions for both the N- and the
C-terminal part of M45 (Figure 3b). The RHIM-containing
N-terminal portion is required for inhibition of apoptosis
induced by TNFa, TRIF overexpression or by the viral
is recruited by different receptors and sensors and mediates the activation of
er different forms of cell death. (a) RIP1 (red) is activated by stimulation of TNF-R1
cases retinoic acid inducible gene I (RIG-I; dark blue) and melanoma differentiation-
ning protein (PIDD, dark green) [50]. TLR3, RIG-I and Mda5 are activated by dsRNA
) or certain viral glycoproteins (gp) [51]. Some of the receptor-activated signaling
members of the TNF receptor-associated factor (TRAF, bright green) family, TRIF
IP1 interacts with other proteins through its death domain (hatched segment) or a
t signaling complexes triggers a variety of downstream effects (1–5), which are
in (FADD, pink) and caspase-8 (casp-8, dark purple), caspase-independent necrosis
ption factors NF-kB, AP-1 and interferon regulatory factors (IRF) 3 and 7. NF-kB and
and/or IRF7 synergize to induce the expression of interferon b. M45 was shown to
its others as well. Abbreviations: IKK, inhibitor of kB kinase; TBK1, TANK-binding
29

Reviewinfection itself [3]. The carboxyterminal portion, by con-
trast, is necessary for blocking NF-kB activation induced
by TNF receptor or TLR3 stimulation [2]. However, the
M45 C terminus also seems to have a role in preventing
rapid endothelial cell death triggered by MCMV infection.
Six mutant MCMVs carrying transposon insertions in
different regions of the M45 ORF showed the same phe-
notype; they all induced endothelial cell death, even if the
transposon was inserted near the 30 end of the ORF [29].
This result indicates that only the full-length M45 protein,
but not C-terminally truncated versions, can efficiently
prevent endothelial cell death. A role of the C terminus
can also be inferred from the observation that the N-
terminal portion of M45 (expressed by transient transfec-
tion) inhibited virus-induced endothelial cell death to a
lesser degree than the full-length protein [3]. Alterna-
tively, it is possible that the C terminus is required for
packaging of the M45 protein into the virus particle and,
thus, for an immediate availability of the protein at the
time of infection.
R1 homologues in other herpesviruses
The new function identified forMCMVM45 points towards
the possibility that R1 homologues of other herpesviruses
might perform similar functions. On the amino acid
Figure 3. Functional domains of the M45 protein. The 1174 amino acid M45 protein con
RIP homotypic interaction motif (RHIM). (a) The presence of a RHIM in herpesviral R1 ho
isolates, respectively, contain a RHIM similar to the one found in the cellular proteins
sequences are also present in the HSV-1 and HSV-2 R1 proteins ICP6 and ICP10, res
respectively. (b) Schematic view of mutant M45 proteins; truncated versions, mutations
mutants (black triangles). The mutant M45 proteins were tested in different assays (I-VI)
tested activity. In assays I-V, the mutant M45 proteins were expressed by transient tran
Assays: I, inhibition of TNFa-induced NF-kB activation; II, inhibition of NF-kB activation u
inhibition cell death induced by TRIF overexpression; V and VI, inhibition of virus-indu
terminal extension is shaded in gray. Epitope tags (HA or myc) at the C terminus are
insertion mutants (i.e. each mutant carries only one Tn insertion). However, each Tn in
30Trends in Biochemical Sciences Vol.34 No.1sequence level, M45 shows the highest homology to R45
and E45, the R1 homologues encoded by the Maastricht
isolate and the English isolate of rat CMV, respectively.
R45 and E45 also contain a RHIM motif (Figure 3a). Thus,
it seems likely that the two proteins function in a manner
similar to M45. By contrast, the R1 homologues of human
and other primate CMVs contain a considerably shorter N-
terminal domain (Figure 1) and lack a RHIM motif. The
HCMV UL45 protein differs from M45 in that it is dis-
pensable for viral replication in endothelial cells [35],
indicating that at least some of the functional features
of M45 are not shared by UL45. However, another study
showed that cells infected with a UL45-deficient HCMV
were more sensitive to cell death induced by stimulation of
Fas (a member of the TNF receptor family) compared with
cells infected with the wild-type virus [17], similar to
observations of M45-deficient MCMV [2]. To date, insuffi-
cient data are available to draw clear conclusions on what
the function of UL45 might be. It is possible that UL45
shares only a subset of the activities of M45. Alternatively,
evolutionary tinkering might have yielded yet another
function for this R1 homologue. Beyond the knowledge
that the HHV-7 R1 homologue does not encode a functional
R1 subunit [16], nothing is known about the function of the
R1 homologues of other b-herpesviruses.
sists of an R1 homology domain and an extended N-terminal domain containing a
mologues. M45 of MCMV and R45 and E45 of the rat CMV Maastricht and English
RIP1, RIP3 and TRIF (the murine versions of these proteins are shown). RHIM-like
pectively. Highly conserved and similar residues are shaded in black and grey,
of the RHIM (crimson), a frame-shift mutant ( fs) and six transposon (Tn) insertion
for biological activities [2,3,29]. ‘+’ indicates the presence and ‘�’ the absence of the
sfection. In assay VI, the mutant proteins were expressed from the viral genome.
pon stimulation of Toll-like receptor 3; III, inhibition of TNFa-induced cell death; IV,
ced endothelial cell death. The R1 homology domain is shown in orange, the N-
shown as black bars. Note that the last row represents six individual transposon
sertion mutant showed the same phenotype [29]. Abbreviations: nd, not done.

References
1 Jacob, F. (1977) Evolution and tinkering. Science 196, 1161–1166
2 Mack, C. et al. (2008) Inhibition of proinflammatory and innate
immune signaling pathways by a cytomegalovirus RIP1-interacting
protein. Proc. Natl. Acad. Sci. U. S. A. 105, 3094–3099
3 Upton, J.W. et al. (2008) Cytomegalovirus M45 cell death suppression
requires receptor-interacting protein (RIP) homotypic interactionmotif
(RHIM)-dependent interaction with RIP1. J. Biol. Chem. 283, 16966–
16970
4 Pellet, P.E. and Roizman, B. (2007) The family of Herpesviridae: a brief
introduction. In Fields Virology (5th edn) (Knipe, D.M. and Howley,
P.M., eds), pp. 2479–2499, Lippincott Williams & Wilkins
5 Conner, J. et al. (1994) Ribonucleotide reductase of herpesviruses. Rev.
Med. Virol. 4, 25–34
6 Goldstein, D.J. and Weller, S.K. (1988) Herpes simplex virus type
1-induced ribonucleotide reductase activity is dispensable for virus
growth and DNA synthesis: isolation and characterization of an ICP6
lacZ insertion mutant. J. Virol. 62, 196–205
7 Cameron, J.M. et al. (1988) Ribonucleotide reductase encoded by
herpes simplex virus is a determinant of the pathogenicity of the
virus in mice and a valid antiviral target. J. Gen. Virol. 69, 2607–2612In contrast to the b-herpesviruses, a- and g-herpes-
viruses express functional RNR R1 homologues. Remark-
ably, some a-herpesvirus R1 proteins display features
reminiscent of M45. The best studied example is infected
cell protein 10 (ICP10), the HSV-2 R1 protein encoded by
ORF UL39. In addition to its C-terminal R1 domain, the
protein also comprises an extended N-terminal domain
(Figure 1). The N-terminal domain was reported to have
protein kinase (PK) activity, but this finding remains
controversial [36,37]. More recent work showed that
ICP10 has chaperone-like activity, perhaps because its
N terminus contains a stretch exhibiting similarity to
the a-crystallin domain of small heat shock proteins
[38]. Importantly, ICP10 possesses anti-apoptotic activity;
it blocks TNFa-induced apoptosis [39] and also cell death
resulting from osmotic stress, growth-factor withdrawal or
excitotoxicity [40–42]. The mechanism(s) by which ICP10
inhibits cell death remains incompletely defined. Indeed,
different mechanisms have been proposed; one group
showed that ICP10 blocks TNFa-induced cell death at,
or upstream of, caspase-8 activation [39,43], whereas
another group presented evidence that ICP10 prevents
cell death by activating extracellular signal-regulated
kinase (ERK)-dependent survival pathways, resulting in
an upregulation of the anti-apoptotic protein Bcl-2-associ-
ated athanogene 1 (Bag-1) and stabilization of Bcl-2 [40].
Moreover, it has been suggested that the a-crystallin
domain has a role in the anti-apoptotic activity of ICP10
[38,44]. However, an a-crystallin domain is also present in
the HSV-1 R1, ICP6 [38], even though ICP6 is not an anti-
apoptotic protein.
It is currently unclear whether the N-terminal PK
domain is sufficient for cell death inhibition [40,41] or
whether the C-terminal R1 domain is also required [43].
The reason for this apparent discrepancy might lie in the
fact that different assay systems were used to evaluate the
anti-apoptotic activity of ICP10; indeed, more than one
mechanism could be involved.
Comparing the known properties of MCMV M45 and
HSV-2 ICP10 reveals remarkable similarities; both are R1
homologues with an extendedN-terminal domain, both are
packaged into viral particles [18,36] and both protect cells
from cell death induced by the viral infection or by TNFa.
Moreover, ICP10 possesses a motif similar to the RHIM
motif identified in M45 (Figure 3a). Hence, it would be
surprising if these two proteins did not share at least some
similarities in their mechanism of action.
Concluding remarks and future perspectives
Recent studies revealed a novel function of M45, providing
the first answer to the long-standing question surrounding
the biological roles of the b-herpesviruses R1 homologues.
Evolutionary tinkering with a viral RNR yielded a protein
with a potent function in viral immune subversion, which
can simultaneously block pro-inflammatory and innate
immune signaling pathways by interacting with a central
mediator molecule of the host cell. Further studies are
required to determine whether this is a shared function
among other b-herpesviruses R1 homologues and whether
Reviewadditional functions exist. Several questions are still unan-
swered, providing a rich area for future research (Box 2).The origin and the evolutionary relationship between
the b-herpesvirus R1 homologues and those of the a- and g-
herpesviruses remain unclear and require further inves-
tigations. Similarities between the herpesvirus R1 genes
and the cellular counterparts indicate that the former were
captured from the host [12,13]. The b-herpesvirus R1 genes
are most likely descended from a common herpesvirus R1
ancestor gene encoding a functional RNR R1 subunit and
have subsequently lost their RNR activity. However, it is
unclear when additional functions were acquired by her-
pesvirus R1 homologues. Were they already present in an
ancestral R1 or were they acquired later during herpes-
virus evolution? The observation that both ICP10 andM45
exert a cell-death-inhibiting activity indicates that this
additional function might have arisen before the diver-
gence of the a- and b-herpesvirus subfamilies 200 million
years ago [45] and before the loss of RNR activity in the b-
herpesvirinae.
Acknowledgements
Work in the authors’ laboratories was supported by grants from the
Turin University Research Fund and the Regione Piemonte (Progetto
Ricerca Sanitaria Finalizzata 2008) to D.L. and the Deutsche
Forschungsgemeinschaft (SFB 421 TP B14) to W.B.
Box 2. Outstanding questions
� Does the MCMV M45 gene encode additional functions beyond
RIP1 inhibition? Which parts of the protein are involved?
� Which of the function(s) of M45 are shared by HCMV UL45? Or has
evolutionary tinkering yielded a completely different function for
this protein?
� What is the biological function of the catalytically inactive R1
subunits of other b-herpesviruses, such as HHV-6 and HHV-7?
� What is the molecular mechanism of the anti-apoptotic function of
HSV-2 ICP10? Is it related to the action mechanism of M45?
� Do g-herpesvirus R1 proteins fulfill additional functions beyond
their role as RNR subunits? What is the function of their N-
terminal extension?
� Do R1 homologues of other viruses or living organisms also
possess functions unrelated to ribonucleotide metabolism?
� Which other viruses encode inhibitors of RIP1-mediated signaling
(vIRS)?
Trends in Biochemical Sciences Vol.34 No.18 Jacobson, J.G. et al. (1989) A herpes simplex virus ribonucleotide
reductase deletion mutant is defective for productive acute and
31

reactivatable latent infections of mice and for replication in mouse
cells. Virology 173, 276–283
9 de Wind, N. et al. (1993) Ribonucleotide reductase-deficient mutants of
pseudorabies virus are avirulent for pigs and induce partial protective
immunity. J. Gen. Virol. 74, 351–359
10 Heineman, T.C. and Cohen, J.I. (1994) Deletion of the varicella-zoster
promoter in the Toll-like receptor signaling. J. Immunol. 169, 6668–
6672
34 Kaiser,W.J. and Offermann,M.K. (2005) Apoptosis induced by the toll-
like receptor adaptor TRIF is dependent on its receptor interacting
protein homotypic interaction motif. J. Immunol. 174, 4942–4952
35 Hahn, G. et al. (2002) The human cytomegalovirus ribonucleotide
Review Trends in Biochemical Sciences Vol.34 No.1virus large subunit of ribonucleotide reductase impairs growth of virus
in vitro. J. Virol. 68, 3317–3323
11 Ball, A.L. (2007) Virus replication strategies. In Fields Virology (5th
edn) (Knipe, D.M. and Howley, P.M., eds), pp. 119–139, Lippincott
Williams & Wilkins
12 Shackelton, L.A. and Holmes, E.C. (2004) The evolution of large DNA
viruses: combining genomic information of viruses and their hosts.
Trends Microbiol. 12, 458–465
13 McGeoch, D.J. et al. (2006) Topics in herpesvirus genomics and
evolution. Virus Res. 117, 90–104
14 Tenser, R.B. (1994) The role of herpes simplex virus thymidine kinase
expression in pathogenesis and latency. In Pathogenesis of Human
Herpes Viruses Due to Specific Pathogenicity Genes (Becker, Y. and
Darai, G., eds), pp. 68–86, Springer
15 Holzerlandt, R. et al. (2002) Identification of new herpesvirus gene
homologs in the human genome. Genome Res. 12, 1739–1748
16 Sun, Y. and Conner, J. (1999) The U28 ORF of human herpesvirus-7
does not encode a functional ribonucleotide reductase R1 subunit. J.
Gen. Virol. 80, 2713–2718
17 Patrone, M. et al. (2003) The human cytomegalovirus UL45 gene
product is a late, virion-associated protein and influences virus
growth at low multiplicities of infection. J. Gen. Virol. 84, 3359–3370
18 Lembo, D. et al. (2004) The ribonucleotide reductase R1 homolog of
murine cytomegalovirus is not a functional enzyme subunit but is
required for pathogenesis. J. Virol. 78, 4278–4288
19 Caposio, P. et al. (2004) Evidence that the human cytomegalovirus 46-
kDa UL72 protein is not an active dUTPase but a late protein
dispensable for replication in fibroblasts. Virology 325, 264–276
20 Chee, M.S. et al. (1990) Analysis of the protein-coding content of the
sequence of human cytomegalovirus strain AD169. Curr. Top.
Microbiol. Immunol. 154, 125–169
21 Rawlinson,W.D. et al. (1996) Analysis of the complete DNA sequence of
murine cytomegalovirus. J. Virol. 70, 8833–8849
22 Novotny, J. et al. (2001) In silico structural and functional analysis of
the human cytomegalovirus (HHV5) genome. J. Mol. Biol. 310, 1151–
1166
23 Nordlund, P. and Reichard, P. (2006) Ribonucleotide reductases.Annu.
Rev. Biochem. 75, 681–706
24 Willoughby, K. et al. (1997) Sequences of the ribonucleotide reductase-
encoding genes of felid herpesvirus 1 and molecular phylogenetic
analysis. Virus Genes 15, 203–218
25 Landolfo, S. et al. (2003) The human cytomegalovirus. Pharmacol.
Ther. 98, 269–297
26 Bain, M. and Sinclair, J. (2007) The S phase of the cell cycle and its
perturbation by human cytomegalovirus. Rev. Med. Virol. 17, 423–434
27 Sanchez, V. and Spector, D.H. (2008) Subversion of cell cycle regulatory
pathways. Curr. Top. Microbiol. Immunol. 325, 243–262
28 Wiebusch, L. et al. (2008) Cell cycle independent expression of
immediate early gene 3 results in a G1 and G2 arrest in MCMV-
infected cells. J. Virol. 82, 10188–10198
29 Brune, W. et al. (2001) A ribonucleotide reductase homolog of
cytomegalovirus and endothelial cell tropism. Science 291, 303–305
30 Meylan, E. and Tschopp, J. (2005) The RIP kinases: crucial integrators
of cellular stress. Trends Biochem. Sci. 30, 151–159
31 Festjens, N. et al. (2007) RIP1, a kinase on the crossroads of a cell’s
decision to live or die. Cell Death Differ. 14, 400–410
32 Meylan, E. et al. (2004) RIP1 is an essential mediator of Toll-like
receptor 3-induced NF-k B activation. Nat. Immunol. 5, 503–507
33 Yamamoto, M. et al. (2002) Cutting edge: a novel Toll/IL-1 receptor
domain-containing adapter that preferentially activates the IFN-b32reductase homolog UL45 is dispensable for growth in endothelial
cells, as determined by a BAC-cloned clinical isolate of human
cytomegalovirus with preserved wild-type characteristics. J. Virol.
76, 9551–9555
36 Smith, C.C. and Aurelian, L. (1997) The large subunit of herpes
simplex virus type 2 ribonucleotide reductase (ICP10) is associated
with the virion tegument and has PK activity. Virology 234, 235–242
37 Langelier, Y. et al. (1998) The R1 subunit of herpes simplex virus
ribonucleotide reductase is a good substrate for host cell protein
kinases but is not itself a protein kinase. J. Biol. Chem. 273, 1435–1443
38 Chabaud, S. et al. (2003) The R1 subunit of herpes simplex virus
ribonucleotide reductase has chaperone-like activity similar to
Hsp27. FEBS Lett. 545, 213–218
39 Langelier, Y. et al. (2002) The R1 subunit of herpes simplex virus
ribonucleotide reductase protects cells against apoptosis at, or
upstream of, caspase-8 activation. J. Gen. Virol. 83, 2779–2789
40 Perkins, D. et al. (2003) The herpes simplex virus type 2 R1 protein
kinase (ICP10 PK) functions as a dominant regulator of apoptosis in
hippocampal neurons involving activation of the ERK survival
pathway and upregulation of the antiapoptotic protein Bag-1. J.
Virol. 77, 1292–1305
41 Perkins, D. et al. (2002) Expression of herpes simplex virus type 2
protein ICP10 PK rescues neurons from apoptosis due to serum
deprivation or genetic defects. Exp. Neurol. 174, 118–122
42 Golembewski, E.K. et al. (2007) The HSV-2 protein ICP10PK prevents
neuronal apoptosis and loss of function in an in vivo model of
neurodegeneration associated with glutamate excitotoxicity. Exp.
Neurol. 203, 381–393
43 Chabaud, S. et al. (2007) The ribonucleotide reductase domain of the R1
subunit of herpes simplex virus type 2 ribonucleotide reductase is
essential for R1 antiapoptotic function. J. Gen. Virol. 88, 384–394
44 Gober, M.D. et al. (2005) Herpes simplex virus type 2 encodes a heat
shock protein homologue with apoptosis regulatory functions. Front.
Biosci. 10, 2788–2803
45 McGeoch, D.J. et al. (1995) Molecular phylogeny and evolutionary
timescale for the family of mammalian herpesviruses. J. Mol. Biol.
247, 443–458
46 Kelliher, M.A. et al. (1998) The death domain kinase RIP mediates the
TNF-induced NF-kB signal. Immunity 8, 297–303
47 Cusson-Hermance, N. et al. (2005) Rip1 mediates the Trif-dependent
toll-like receptor 3- and 4-induced NF-kB activation but does not
contribute to interferon regulatory factor 3 activation. J. Biol. Chem.
280, 36560–36566
48 Balachandran, S. et al. (2004) A FADD-dependent innate immune
mechanism in mammalian cells. Nature 432, 401–405
49 Michallet, M.C. et al. (2008) TRADD protein is an essential component
of the RIG-like helicase antiviral pathway. Immunity 28, 651–661
50 Janssens, S. et al. (2005) PIDD mediates NF-kB activation in response
to DNA damage. Cell 123, 1079–1092
51 Kawai, T. and Akira, S. (2006) Innate immune recognition of viral
infection. Nat. Immunol. 7, 131–137
52 Young, P. et al. (1994) Intron-containing T4 bacteriophage gene sunY
encodes an anaerobic ribonucleotide reductase. J. Biol. Chem. 269,
20229–20232
53 Hatfull, G.F. and Sarkis, G.J. (1993) DNA sequence, structure and gene
expression of mycobacteriophage L5: a phage system for mycobacterial
genetics. Mol. Microbiol. 7, 395–405
54 Tidona, C.A. and Darai, G. (2000) Iridovirus homologues of cellular
genes–implications for the molecular evolution of large DNA viruses.
Virus Genes 21, 77–81