High yield synthesis, purification and characterisation of the RNase L activators 5'-triphosphate 2′-5′-oligoadenylates

Article abstract of DOI:10.1016/j.antiviral.2010.06.003

Upon viral infection, double-stranded viral RNA is detected very early in the host cell by several cellular 2′-5′ oligoadenylate synthetases, which synthesize 2′-5′ adenylate oligonucleotides that activate the cellular RNase L, firing an early primary antiviral response through self and non-self RNA cleavage. Transfecting cells with synthetic 2′-5′ adenylate oligonucleotides activate RNase L, and thus provide a useful shortcut to study the early steps of cellular and viral commitments into this pathway. Defined 2′-5′ adenylate oligonucleotides can be produced in vitro, but their controlled synthesis, purification, and characterisation have not been reported in detail. Here, we report a method suitable to produce large amounts of 2-5As of defined lengths in vitro using porcine OAS1 (pOAS) and human OAS2 (hOAS). We have synthesized a broad spectrum of 2-5As at the milligram scale and report an HPLC-purification and characterisation protocol with quantified yield for 2-5A of various lengths.

Full text of DOI:10.1016/j.antiviral.2010.06.003

Antiviral Research 87 (2010) 345–352
Contents lists available at ScienceDirect
Antiviral Research
journal homepage: www.elsevier.com/locate/antiviral
High yield synthesis, purification and characterisation of the RNase L activators
5ꢀ-triphosphate 2ꢀ–5ꢀ-oligoadenylates
B. Morina, N. Rabaha, J. Boretto-Solera, H. Toloub, K. Alvareza,∗, B. Canarda,∗
a Architecture et Fonction des Macromolécules Biologiques, CNRS and Universités d’Aix-Marseille I et II, UMR 6098, ESIL Case 925, 13288 Marseille, France
b Unité de Virologie Tropicale, IMTSSA, BP 46, 13998, Marseille Armées, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Upon viral infection, double-stranded viral RNA is detected very early in the host cell by several cel-
lular 2ꢀ–5ꢀ oligoadenylate synthetases, which synthesize 2ꢀ–5ꢀ adenylate oligonucleotides that activate
the cellular RNase L, firing an early primary antiviral response through self and non-self RNA cleavage.
Transfecting cells with synthetic 2ꢀ–5ꢀ adenylate oligonucleotides activate RNase L, and thus provide a
useful shortcut to study the early steps of cellular and viral commitments into this pathway. Defined
2ꢀ–5ꢀ adenylate oligonucleotides can be produced in vitro, but their controlled synthesis, purification,
and characterisation have not been reported in detail. Here, we report a method suitable to produce
large amounts of 2–5As of defined lengths in vitro using porcine OAS1 (pOAS) and human OAS2 (hOAS).
We have synthesized a broad spectrum of 2–5As at the milligram scale and report an HPLC-purification
and characterisation protocol with quantified yield for 2–5A of various lengths.
Received 7 April 2010
Received in revised form 1 June 2010
Accepted 7 June 2010
Keywords:
Antiviral
2ꢀ–5ꢀ Oligoadenylate
Oligoadenylate synthetase
Ribonuclease L
HPLC
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
tant role in the endogenous antiviral pathway (Silverman, 2007b).
In recent years, several correlations between viral infections and
The mammalian cell response to a viral infection is mainly medi-
ated by type I interferon-regulated pathways (Sen, 2001; Servant
et al., 2002). One of the main and early effects is to activate the 2ꢀ–5ꢀ
oligoadenylate synthetase (OAS)/ribonuclease L (RNase L) path-
way (Roberts et al., 1976), where double-stranded non-self RNA
produced during infection by the viral invader is recognized by
one or several OAS isoforms. This enzyme is allosterically acti-
vated by dsRNA to synthesize small specific RNA molecules which
are oligomers of adenosine with the particularity to be linked by
2ꢀ–5ꢀ phosphodiester bonds instead of 3ꢀ–5ꢀ as in “classic” RNAs
(Hovanessian and Justesen, 2007). These oligomers (2–5As) are
known to bind RNase L (Dong and Silverman, 1995), and the 2ꢀ–5ꢀ
phosphodiester bonds have been shown to be essential for the
specific binding. The structural basis of this activation has been
reported in the crystallographic structure of the 2–5A:RNase L com-
plex (Tanaka et al., 2004), currently the only example of protein
crystallized in complex with 2–5As. Once activated, RNase L cleaves
both cellular and viral single-stranded RNA, thus playing an impor-
the OAS/RNase L pathway have been reported (Behera et al., 2002;
Kajaste-Rudnitski et al., 2006; Sawicki et al., 2003; Smith et al.,
2005). Conversely, the ability of some viruses to counteract the
antiviral action of RNase L has also been demonstrated (Han et al.,
2007; Min and Krug, 2006; Xiang et al., 2002). Precise mechanisms
remain ill-defined, though, and mechanisms of indirect or direct
interaction between viral proteins and the OAS/RNase L pathway
remain to be discovered.
No direct interaction of a viral protein with 2–5As has been
reported yet. Conceptually, however, these 2–5As represent early
triggers ofthe non-specificinnate immune response, and thus, early
viral interference with this pathway would be expected to effi-
ciently dampen the host response to viral invasion. Indeed, direct
induction of the RNase L response with these purified oligomers has
been achieved successfully through transfection of 2–5A (Malathi
et al., 2007; Silverman, 2007a). Likewise, activation of RNase L by
small organic molecules has been shown to efficiently launch an
antiviral state in the host cell (Thakur et al., 2007). Thus, it would
be logic that viruses had evolved a way to interfere early with the
innate immune response which involves the induction of more than
a thousand actors in a potent and complex antiviral response.
With this idea in mind, we set up to obtain high amounts of
pure and structurally characterised 2–5A in order to study their
interaction with potential viral targets before they reach RNase L
in the cell cytoplasm.
Abbreviations: DP, degree of polymerisation; OAS, oligoadenylate synthetase;
pOAS1, porcine OAS 1; hOAS2, human OAS 2; 2–5A, 2ꢀ–5ꢀ oligoadenylate; RNase L,
ribonuclease L; CIP, calf alkaline phosphatase; VCE, vaccinia virus capping enzyme;
RNase T2, ribonuclease T2.
∗
Corresponding authors. Tel.: +33 4 91 82 86 44; fax: +33 4 91 82 86 46.
E-mail addresses: karine.alvarez@afmb.univ-mrs.fr (K. Alvarez),
Several 2–5A synthesis protocols have been described, using
either chemical or enzymatic method (Imai and Torrence, 1981;
bruno.canard@afmb.univ-mrs.fr (B. Canard).
0166-3542/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.antiviral.2010.06.003

346
B. Morin et al. / Antiviral Research 87 (2010) 345–352
Justesen et al., 1980; Kitade et al., 1991; Rusch et al., 2001; Thakur
et al., 2007). However, the chemical synthesis of a 5ꢀ-triphosphate,
a sine qua non condition to retain RNase L activation, is not easy
and has been reported in an eight-step synthesis. In addition,
neither easy nor large-scale production nor detailed purification,
reporting yield, size distribution, and individual fractionation of
size-controlled 2–5A, has been documented.
The enzymatic synthesis of 2–5A has been reported using puri-
fied OASs. Human OASs belong to a family of enzymes encoded
by 3 closely linked genes with the following order: small (OAS1),
medium (OAS2) and large (OAS3) isoforms, corresponding to pro-
teins of 40/46, 69/71 and 100 kDa respectively. They are composed
of one, two or three repeats of the basic OAS module, correspond-
ing to their size (Hovanessian and Justesen, 2007). At their optimum
activity conditions, OAS1 and OAS2 have the capacity to synthesize
oligomers of “high” molecular weight, whereas OAS3 has the ten-
dency to synthesize preferentially dimeric oligonucleotides. Some
of these enzymes have been used to synthesize 2–5As (Rebouillat
and Hovanessian, 1999), but no detailed protocol, precise purifi-
cation, nor characterisation of the synthesized products has been
described.
Here we report a protocol to produce large amount of differ-
ent 2–5As in vitro. Two OASs, the porcine OAS1 (pOAS1) and the
human OAS2 (hOAS2), were used and compared to synthesize a
broad spectrum of 2–5As at the milligram scale with quantified
yield for 2–5A of various lengths. We show that the Vaccinia virus
mRNA capping enzyme is active as an RNA 5ꢀ-triphosphatase on the
2–5As, providing a simple and robust method to prepare 2–5A with
a 5ꢀ-diphosphate. The purified 2–5As are able to induce rRNA degra-
dation in transfected cell, as expected for an RNase L-mediated
antiviral response.
eluted at 0.32 M NaCl. Fractions were collected and applied to an
immobilized metal affinity chromatography column (Ni-NTA, GE
Healthcare) in 50 mM NaPO4 buffer pH 7.5, 45 mM imidazole, 10%
glycerol, 0.1% Igepal CA630, 1 mM EDTA, and 0.32 M NaCl. After
washing with the same buffer, the protein was eluted in the same
buffer supplemented with 240 mM imidazole (final concentration).
The collected protein was dialysed against the same buffer without
imidazole. The last dialysis buffer change constituted the storage
buffer, i.e., 50 mM NaPO4 pH 7.5, 50% glycerol, 0.3 M NaCl, 0.5 mM
EDTA and 5 mM ␤-mercaptoethanol.
2.3. HPLC set-up
A Waters model 600 gradient HPLC system equipped with two
600 pumps, a 717 plus Autosampler injector, and a 996 photo-
diode array detector. An in-line degasser AF was employed for
reverse phase chromatography. The separation column (Reverse
Phase C18) and a pre-column (Reverse Phase C18) were installed
in parallel on a two 7000 Rheodyne valve system (Interchim) for
the online-cleaning procedure. A filter insert protected the column
assembly. As buffer stock solution a 1 M solution of triethylam-
monium bicarbonate (TEAB) was prepared by adding dry-ice to a
1 M triethylamine solution until the pH reached 7.4 and filtered
through 0.22 ␮M GV-type membranes (Millipore). HPLC eluents
were freshly prepared. Eluent A was a 0.05 M solution of TEAB (pH
7.4) and eluent B was a 1:1 mixture of acetonitrile (HPLC grade,
SDS, Peypin, France) and TEAB (final concentration 0.05 M, pH 7.4).
The applied gradients for analytical and preparative separation of
2–5As are described in the corresponding sections (see below).
For HPLC coupled mass spectrometry, a Waters model Alliance
2790 system equipped with a photodiode array detector was
employed. Mass spectra were recorded on an analyzer Quadrupole
time-of-flight “Qtof” mass spectrometer (Waters). Conditions were
accelerating potential 20 V, capillary potential 3000 V, source tem-
perature 150 ◦C and nebulization temperature 250 ◦C.
2. Materials and methods
2.1. Expression and purification of recombinant human OAS2 in
insect cells
2.3.1. Analytical scale
20 ␮l of sample was mixed with 180 ␮l of TEAB (0.05 M) and
analysed. The column assembly consists of a pre-column (Delta-
pak C18 100 Å, 5 ␮m, 3.9 × 20 mm) and a separation column (Nova-
pak C18, 4 ␮m, 3.9 × 150 mm). Separations were run at a flow rate
of 1 ml/min and started with a 5 min elution (100% eluent A) on
the pre-column to remove proteic material. The analytical gradient
started after 5 min at 100% eluent A with an increase to 8% eluent
B after 15 min, to 15% eluent B after 30 min, to 20% eluent B after
35 min and to 100% eluent B after 40 min.
The human OAS2 cDNA cloned in pFastBac vector (Invitrogen
Bac-to-Bac Baculovirus Expression System) as a N-terminal His-
tag fusion protein was a kind gift of Saumendra N. Sarkar and
Games C. Sen from The Lerner Research Institute, Cleveland, USA.
This construction was transformed into DH10-Bac Escherichia coli
to produce bacmid DNA. The presence of recombinant bacmids
was verified by PCR, and bacmids were used for the generation
of human OAS2 recombinant baculovirus using Invitrogen Bac-to-
Bac Baculovirus Expression System. All experiments were carried
out according to the manufacturer’s protocol in insect Sf9 cells. Sf9
insect cells were maintained in a spinner or monolayer cultures
at 27 ◦C in Insect-XPress medium (BioWhittaker, Lonza). For large-
scale production of the hOAS2 isoenzyme, 3 × 108 Sf9 cells were
infected. 72 h post-infection, cells were harvested and the hOAS2
protein was purified as described in Sarkar and Sen (1998).
2.3.2. Preparative scale
250 ␮l of sample was mixed with 400 ␮l of TEAB (0.05 M) and
analysed. The column assembly consists of a pre-column (XTerra
prep MSC18, 10 ␮m, 10 × 10 mm) and a separation column (XTerra
prep MSC18, 10 ␮m, 10 × 250 mm). Separations were run at a flow
rate of 5 ml/min and started with a 5 min elution (100% eluent A) on
the pre-column to remove proteic material. The analytical gradient
started after 5 min at 100% eluent A with an increase to 8% eluent
B after 15 min, to 15% eluent B after 30 min, to 20% eluent B after
35 min and to 100% eluent B after 40 min.
2.2. Expression and purification of recombinant porcine OAS1 in
E. coli
The optimized gene of porcine OAS1 cloned into pET-15b as
an N-terminal His-tag fusion protein was purchased from Gen-
eart (Germany). The protein was expressed in BL21 (DE3) pLysS
(Stratagene, U.S.A.) at 25 ◦C overnight after induction with 0.5 mM
IPTG in LB Broth. The protein was purified in two steps, first with
cation exchange chromatography (Cellulose Phosphate P11, What-
man) in 50 mM NaPO4 buffer pH 7.5, 10% glycerol, 1 mM DTT, 0.1%
Igepal CA630 (Sigma–Aldrich), 1 mM EDTA, and 50 mM NaCl. A gra-
dient from 50 mM to 0.5 M NaCl was applied, and the protein was
2.4. Mass spectrometry analysis
MALDI-TOF mass spectra were recorded on a Voyager DE
mass spectrometer (Perseptive Biosystems, Framingham, MA, USA)
equipped with an N2 laser (337 nm). MALDI conditions were
accelerating potential, 24,000 V; guide wire, 0.05% of accelerating
voltage; grid voltage, 94% of accelerating voltage; delay extraction
time, 550 ns. Spectra were obtained in negative mode and were

B. Morin et al. / Antiviral Research 87 (2010) 345–352
347
not smoothed. The oligonucleotides (100 pmol) were suspended in
10 ␮l of water and desalted using drop dialysis through a mem-
brane filter WSWP 0.025 ␮m, 13 mm (Millipore) floating on a 0.1 M
ammonium citrate solution for 30 min. When drop dialysis did not
allow the removal of all salt traces, the samples were further treated
with a few beads of DOWEX 50W X8 resin (ammonium form) before
spotting them on the MALDI target. Samples of 0.5 ␮l were mixed
with 0.5 ␮l of the matrix 2,4,6-trihydroxyacetophenone (THAP,
45 mg, ammonium citrate, 4 mg in 500 ␮l acetonitrile/water, 1:1,
v/v) and the mixtures were spotted on the stainless steel MALDI
target and left to dry under air before MALDI analysis.
reaction was stopped by heating at 95 ◦C for 5 min and analysed by
HPLC.
2.10. RNase T2 phosphodiesterase activity
After reaction, purification, separation and collection (using
large-scale conditions), the sample corresponding to the 5ꢀ-
triphosphate trimer was lyophilised two times and re-suspended
in water. The purity was verified by HPLC, its concentration deter-
mined by measuring OD260 values and the molecular mass was
verified by mass spectrometry (MALDI-TOF). 50 nmols of 2–5A
trimer was then treated with 1 unit of CIP and injected into the
HPLC system for purification (using large-scale conditions). The
peak corresponding to the 5ꢀ-hydroxyl 2–5A trimer was collected,
lyophilised two times and re-suspended in water. 20 nmols of the
2–5A and 3–5A (Dharmacon) trimer having 5ꢀ-hydroxyl ends were
treated with 23 units of RNase T2 (Gibco BRL) at 37 ◦C for 1 h. The
reaction was stopped by heating at 95 ◦C for 5 min, loaded onto the
HPLC column and analysed by mass spectrometry.
2.5. HPLC coupled mass spectrometry
Samples were previously filtered on Microcon (size exclusion
>3000 Da) (Millipore) to remove proteic material. 20 ␮l of sam-
ple was mixed with 20 ␮l of TEAB (0.05 M) and analysed. The
column assembly consists of a pre-column (Delta-pak C18 100 Å,
5 ␮m, 3.9 × 20 mm) and a separation column (Nova-pak C18, 4 ␮m,
3.9 × 150 mm). Separations were run at a flow rate of 1 ml/min. The
analytical gradient started after 5 min at 100% eluent A with an
increase to 8% eluent B after 15 min, to 15% eluent B after 30 min,
to 20% eluent B after 35 min and to 100% eluent B after 40 min. The
flow was split in tandem UV and MS. Mass spectra were obtained
in negative mode (electrospray ion source).
2.11. Cell transfections and RNAse L-mediated rRNA cleavage
Hela cells (human cervical epithelial cells) were grown in
DMEM supplemented with 10% heat-inactivated fetal bovine serum
(Biotech GmBH) and streptomycin–penicillin (Gibco). Cells were
plated 1 day before transfection in order to be 50–60% conflu-
ent by the time of transfection. 2–5As were transfected using
Oligofectamine Reagent (Invitrogen) according to the manufac-
turer’s protocol. Cells were harvested 3 h after transfection and
washed with ice cold PBS. Total RNA was extracted from trans-
fected cells using Qiagen RNeasy isolation kit (Qiagen) according to
the manufacturer’s instructions. rRNA degradation was analysed
using agarose gel electrophoresis. RNA samples (2 ␮g) were heat-
denatured 10 min at 70 ◦C in RNA loading buffer (50% formamide,
6% formaldehyde and 5% glycerol in MOPS buffer), then separated
on 1% agarose gel in MOPS buffer.
2.6. 2–5A synthetase activity assay
The 2–5A synthetase activity assay was performed by incubating
the recombinant protein (2 ␮M of pOAS1 or 1 ␮M of hOAS2) in a
reaction mixture containing 20 mM Tris/base pH 7.5, 20 mM MgCl2,
2.5 mM DTT, 50 ␮g/ml poly (I:C), 10 mM ATP as a substrate in a
final volume of 50 ␮l at 30 ◦C. The reaction was stopped by heating
at 95 ◦C for 5 min and analysed on HPLC. In the case of radiolabel-
containing reactions, 0.4 ␮Ci/␮l of (␥-32P) ATP was added to the
mixture. The reaction was stopped by the addition of a standard gel-
loading formamide Blue Dye/EDTA solution and heating at 95 ◦C for
5 min. The 2–5A products were then analysed by electrophoresis on
14% polyacrylamide gels containing 7% urea in TTE buffer (89 mM
Tris–HCl, pH 8.0, 28 mM taurine, 0.5 mM EDTA). For reactions using
2 substrates, 7 mM of ATP and 3 mM of the other (ADP or AMP) was
added to the mixture, and incubated for 2 h.
3. Results and discussion
3.1. Kinetics and product distribution during hOAS2- and
pOAS1-mediated synthesis of 2–5A oligomers
2.7. Large-scale synthesis of triphosphate 2–5A
In order to cover a wide degree of polymerisation (DP) of 2–5A
products, we chose the porcine OAS1 (pOAS1) and the human
OAS2 (hOAS2) oligoadenylate synthetase as catalysts in enzymatic
reaction synthesizing 2–5As from ATP. An OAS3 enzyme was not
included in the study since such protein catalyses mainly the pro-
duction of 2ꢀ–5ꢀ dimers (Rebouillat et al., 1999). The enzymes were
expressed and purified from recombinant cloned genes in E. coli and
insect cells for pOAS1 and hOAS2 (Hartmann et al., 2003; Kodym
et al., 2009; Sarkar and Sen, 1998), respectively (see Section 2). We
obtained ∼450 ␮g of purified hOAS2 and 60 mg of purified pOAS1
per liter of culture, both with a ∼90% purity as judged by denaturing
PAGE. For enzyme characterisation, we tested published exper-
imental conditions for these proteins (Marie et al., 1997; Sarkar
and Sen, 1998), except that a high concentration of substrate (ATP
10 mM) was used in an attempt to maximize the overall yield. How-
ever, at concentrations above 10 mM, substrate inhibition became
prevalent (data not shown). We first selected the lowest concen-
tration of each protein giving a wide range of oligomer products, as
detected in PAGE separated products containing (␥-32P) AMP in 5ꢀ
(i.e., using (␥-32P) ATP as a substrate, Fig. 1). The radiolabel incor-
porated in each band product was counted in the region of the gel
corresponding to each of the 2–5A oligomers, and plotted in a graph
(Fig. 1).
A preparative reaction (1 ml) was set up using the same condi-
tions as described above, with 2 ␮M of pOAS1 for 2 h, and samples
of 250 ␮l reaction mixture were injected into the HPLC system. The
different peaks were collected, and fractions corresponding to each
oligomer were pooled together. Finally, the pooled fractions were
lyophilised two times and the concentration was determined. The
ε260 values (M−1 cm−1) of 2–5As used were 30,800, 46,200, 61,600,
77,000 and 92,400 for the dimer, trimer, tetramer, pentamer and
hexamer respectively.
2.8. Vaccinia virus 5ꢀ-RNA triphosphatase activity
25 ␮l of the synthetase reaction mixture was treated with 80
units of Vaccinia virus capping enzyme (Epicentre Biotechnologies)
at 37 ◦C for 2 h. The reaction was stopped by heating at 95 ◦C for
5 min and analysed by HPLC.
2.9. CIP alkaline phosphatase activity
100 ␮l of the synthetase reaction mixture was treated with 10
units of Calf alkaline phosphatase (CIP) (Biolabs) at 25 ◦C for 1 h. The

348
B. Morin et al. / Antiviral Research 87 (2010) 345–352
Fig. 1. Time course reaction of 2–5A synthesis. The OAS activity of both the porcine (A) and human (B) OAS was assayed using (␥-32P) ATP. Reaction conditions were as
described in Section 2. The reaction was stopped at different times and the 2–5A products were analysed by PAGE, and the gel was auto-radiographed. The radioactivity of
the spots corresponding to each oligomer was counted and the values were expressed as a percentage of total products formed at the end of the reaction. The values are
shown in the graphic below the corresponding autoradiogram. ꢀ: dimer, ꢁ: trimer, ꢂ: tetramer, ×: pentamer, ♦: hexamer, ᭹: heptamer, and |: octamer.
As expected from enzymes from type 1 and type 2 families, the
size distribution of the products is well over the dimer. The prod-
uct DP varies between these two polymerases, in accordance with
the reported difference in processivity (Hovanessian and Justesen,
2007). The reaction reaches equilibrium much faster in pOAS1
(1 h) (Fig. 1A) than in hOAS2 (>12 h) (Fig. 1B). A clear prefer-
ence for the trimer is observed for both at equilibrium, although
pOAS1 synthesizes before equilibrium (after 10 min) a maximum
of dimer in absolute. However, the pOAS1 is less stable and pre-
cipitates during the reaction in contrast to the hOAS2 which does
not. We did not investigate on this observation further, although
it should prove helpful if the dimer synthesis yield is to be opti-
mized. In the case of pOAS1 (Fig. 1A), the trimer is the major
oligomer in the mixture at equilibrium and represents 40% of
the total products, while dimer and tetramer represent around
22% of the total products formed. Longer oligomers represent
less than 10%, which renders their production more problem-
atic.
In the case of the hOAS2, the reaction is much slower than but
similar to that of pOAS1 in the early stage (Fig. 1B). After 14 h of
reaction, the size distribution of the products formed during the
reaction is stable and in contrast to pOAS1, a very low concen-
tration of dimer remains, as if it had been used in the reaction
and converted to products of higher DP. The final proportions of
trimer and tetramer are around 30% of total products, followed by
the pentamer (22%) and hexamer (10%). At the end of the reaction,
the amount of pentamer produced using hOAS2 is close to that of
tetramer produced using pOAS1. Overall, the OASs are polymerases
of low processivity compared to e.g., RNA polymerases, allowing
appropriate control of product size.
3.2. Production of 5ꢀ triphosphate, 5ꢀ diphosphate and 5ꢀ hydroxyl
2–5A
In order to produce 2–5A oligomers with various 5ꢀ-ends, the
synthesized reaction products were treated with phosphatase to
generate two different types of 5-ends: a 5ꢀ-diphosphate and 5ꢀ-
hydroxyl. The commercial Vaccinia Virus capping enzyme (VCE)
is known to possess all the activities essential for RNA cap-
ping which are the triphosphatase, the guanyltransferase and the
methyltransferase (Shuman, 1990). In our experiment, the triphos-
phatase activity was used in order to generate oligomers with a
5ꢀ-diphosphate end. Calf intestine phosphatase (CIP), known to
remove all 5ꢀ-phosphates (Kim and Wyckoff, 1991) was also used
to generate 5ꢀ-hydroxyl RNAs. First, a 2–5A synthesis reaction was
performed as described in Section 2, with 2 ␮M of pOAS1 during
15 min. The reaction was stopped, split in two, and two separate
subsequent reactions were performed: one with VCE and the other
with CIP. Then, the three reactions were analysed by HPLC. Mass
spectrometry analysis was performed in order to identify every
peak product (Fig. 2). Hence, analysis of the synthesis reaction on
HPLC shows 4 major peaks which correspond respectively to the
residual ATP, the dimer, the trimer and the tetramer (Fig. 2, lane
1). Upstream of each of these peaks has small peaks corresponding
to each 5ꢀ-diphosphate oligomers generated by the formation of
ADP during the reaction. After incubation with the VCE (Fig. 2, lane

B. Morin et al. / Antiviral Research 87 (2010) 345–352
349
Fig. 2. HPLC profiles of synthesis and phosphatase reaction experiments. A synthesis reaction was performed for 15 min using condition described in Section 2. The reaction
mixture was devised in two samples, one load on the HPLC system (lane 1, in black), the second used for phosphatase experiments. Two phosphatase experiments, using
VCE (lane 2, in blue) and CIP enzymes (lane 3, in green), were performed to cleave 5ꢀ triphosphatase to generate a diphosphate or a hydroxyl 5ꢀ-end, respectively. Reaction
conditions are given in Section 2. The reaction products were then loaded on an HPLC system (Delta-pak C18 100 Å, and Nova-pak C18). The gradient started after 5 min at
100% eluent A with an increase to 8% eluent B after 15 min, to 15% eluent B after 30 min, to 20% eluent B after 35 min and to 100% eluent B after 40 min. The corresponding
composition of each peak is indicated above them. 5ꢀ-Triphosphate ends are labeled 3P, 5ꢀ-diphosphate end are labeled 2P, and 5ꢀ-hydroxyl ends are labeled 5ꢀ-OH. The two
arrows indicate the synthetase reaction mixture treated by both VCE and CIP (see Section 2 for details). (For interpretation of the references to colour in this figure legend,
the reader is referred to the web version of the article.)
2), we observe a decrease of the retention time for the different
major peaks. In fact, the peaks corresponding to the 5ꢀ triphos-
phate oligomers are transformed into those corresponding to 5ꢀ
diphosphate oligomers. Likewise, when the cleavage reaction was
performed using CIP (Fig. 2, lane 3), we observed an increase of
the retention time of the different products correlated to their size.
In contrast to the cleavage with the VCE, the CIP cleaves all the
phosphates at the 5ꢀ end of the 2–5A and generates products less
charged than the triphosphate 2–5A, which explain the differences
observed in retention time. Altogether, these results validate a con-
venient method of 2–5A oligomer synthesis, either 5ꢀ-triphosphate,
5ꢀ-diphosphate, or 5ꢀ-hydroxyl. We note that it is possible to pro-
duce small quantities of 5ꢀ-diphosphate and 5ꢀ-monophosphate
2–5A, using reaction mixture with two substrates: ATP and ADP, or
ATP and AMP (data not shown). The synthesis of 5ꢀ-monophosphate
2–5A was not investigated further.
using pOAS1. The entire volume of the reaction was divided into
four fractions and injected independently on HPLC. All the peaks
corresponding to the triphosphate oligomers were collected and
lyophilised two times to remove the volatile buffer solution. Com-
pounds were characterised by mass spectrometry and isolated
yields of each oligomer are given in Table 1.
Hence, masses of various oligomers, as measured by mass spec-
trometry, are in agreement with the theoretical values, namely 835
for the dimer; 1165.76 for the trimer; 1495.30 for the tetramer;
1823.7 for the pentamer; and 2152.05 for the hexamer (Table 1).
Based on the mass of recovered products, the best yield (55.5%)
was obtained for the trimer. In fact, the amount of trimer recov-
ered after purification is relatively high, close to 1 mg. Yields of
two of the other oligomers are also fairly good with a better yield
for the tetramer (24.8%) than the dimer (13.9%), and the amount
recovered is in the order of several hundreds of ␮g. At last, but in
less proportion, pentamer and hexamer are also recovered during
the purification on HPLC. Quantities of product recovered are also
significant with more than 80 ␮g for the pentamer, and more than
10 ␮g for the hexamer. It is also interesting to note the total amount
of product recovered at the end of the purification, which repre-
3.3. Synthesis of triphosphate 2–5A on a preparative scale
To quantitate the yield of 2–5A oligomer after purification of
a 2–5A synthesis, a large-scale 1 ml reaction for 3 h was set up
Table 1
Large-scale synthesis of triphosphate 2–5A oligomers. A preparative reaction was set up with pOAS. For experimental conditions and MALDI-TOF characterisation, see Section
2. Yields are given as isolated values (percentage of formed product, nmol or ␮g product produced and percentage of ATP conversion) based on absorbance measurements
of pure product after lyophilisation. Given retention times correspond to the use of the preparative gradient as described in Section 2.
Size
HPLC retention time (min)
Yields
m/z exp. negative mode
m/z calc.
% Formed product
nmol
Mg
%/ATP
2
3
4
5
5
23.2
25.4
27.5
29.1
30.7
13.9
55.5
24.8
5.1
324.0
867.4
289.4
47.4
269.4
1005.8
430.6
86.1
6.5
26
11.6
2.4
0.3
835.00
836.39
1165.76
1495.30
1823.70
2152.05
1165.59
1494.80
1824.00
2153.21
0.7
5.2
11.2

350
B. Morin et al. / Antiviral Research 87 (2010) 345–352
Fig. 3. HPLC profiles of RNase T2 experiments. A preparative reaction was set up and products were purified as described in Section 2. The purified 2–5A trimer fraction was
then incubated with the CIP to generate a hydroxyl 5ꢀ-end. Commercial 5ꢀ-hydroxyl 3–5A trimer (AAA RNA) and the purified 5ꢀ-hydroxyl 2–5A trimer were then treated with
RNase T2 under conditions described in Section 2. 2–5A trimer (lane 1), 2–5A trimer treated with RNase T2 (lane 2), 3–5A trimer (lane 3) and 3–5A trimer treated with RNase
T2 (lane 4) samples, were injected into an HPLC system (Delta-pak C18 100 Å, and Nova-pak C18). The gradient started after 5 min at 100% eluent A with an increase to 8%
eluent B after 15 min, to 15% eluent B after 30 min, to 20% eluent B after 35 min and to 100% eluent B after 40 min (see Section 2 for details). The corresponding composition
of each peak is indicated above them, and the retention times are indicated under the corresponding peaks.
sents almost 50% of the total ATP used initially for the reaction. This
large-scale production of 2–5A oligomers in vitro shows that signif-
icant amounts of products (mg scale) can be easily recovered after
purification of the reaction products by HPLC. It is noteworthy that
the HPLC method used here (based on hydrophobic interactions)
has the advantage of removing residual salts present in the reaction
in contrast to previously reported studies using an ion exchange
purification procedure (Budowsky et al., 1994; Xiao et al., 1996).
the RNase T2 (Rushizky and Sober, 1963). As expected, this result
shows that the 3–5A trimer is totally degraded by the treatment
with the RNase T2. On the contrary, the same experiment made
with the 2–5A trimer shows little change in retention time after
treatment with RNase T2, indicating that phosphodiester bonds
present in the trimer synthesized by the pOAS1 are different from
those of the commercial trimer. The variation of retention time
between 2–5A and 3–5A trimers and the resistance to the RNase
T2 of the 2–5A trimer are both in agreement with the presence
of 2ꢀ–5ꢀ phosphodiester bonds in oligomers produced by these
OASs.
3.4. Characterisation and structural analysis of 2–5A products
We then set up to characterise the precise structure of 2–5A
products in order to design standardized purification protocols. The
chemical type of the internucleotidic bonds of the 2–5A products
was assayed using phosphodiesterases.
3.5. Analysis of rRNA degradation in transfected cells
Finally, to further characterise the purified 2–5As we tested
their effect on the activation of endogenous RNase L in cell trans-
fection experiments. To that end, we transfected Hela cells with
increasing concentrations of 2–5A (1–10 ␮M) and monitored rRNA
degradation after 3 h following transfection. As a positive con-
trol, synthetic dsRNA poly (I:C), known to induce the OAS/RNase L
response (Silverman, 2007b). Poly (I:C) binds to endogenous OASs
which activate latent RNase L. The latter cleaves ribosomal RNA
thereby leading to 28S and 18S degradation (Fig. 4, and (Silverman,
2007b)). As shown in Fig. 4, the transfection of cells with either
oligofectamine alone (mock) or the 2–5A dimer at various concen-
trations does not induce rRNA degradation since 28S and 18S rRNA
bands remain intact. The 2–5A dimer is known to neither bind nor
activate RNase L (Xiang et al., 2003). However, the introduction of
trimer into cells leads to the appearance of rRNA degradation prod-
ucts (Fig. 4). This degradation is weak at 1 ␮M and increases with
increasing amount of 2–5A. A similar but slightly weaker effect is
obtained with tetrameric 2–5A. Functional and structural studies
revealed that RNAse L contains three domains namely the N-
terminal ankyrin repeat domain, responsible for 2–5A binding, the
kinase-like domain and the C-terminal catalytic domain, respon-
sible for catalytic activity (Dong and Silverman, 1997; Nakanishi
RNase T2 is unable to hydrolyse oligomers having 2ꢀ–5ꢀ phos-
phodiester bonds (Budowsky et al., 1994; Rushizky and Sober,
1963). Two types of trimer were used to assay RNase T2 cleavage: a
commercial 3–5A trimer (5ꢀAAA RNA trimer) with a 5ꢀ-hydroxyl
end, and a 2–5A trimer purified after pOAS1-mediated synthe-
sis. In order to compare trimers, the 2–5A trimer was treated
with CIP to release a 5ꢀ-hydroxyl end. After the treatment of
both 2–5A and 3–5A 5ꢀ hydroxyl-trimers with RNase T2, reac-
tions were analysed using HPLC (Fig. 3). The first characterisation
of the difference between 3ꢀ–5ꢀ and 2ꢀ–5ꢀ phosphodiester bonds
is the variation of retention times between the two substrates,
∼34 min and ∼41 min, respectively (Fig. 3, lane 1 and lane 3).
Indeed, the 5ꢀAAA RNA trimer has a retention time increased rel-
ative to that of the 2–5A trimer, as described previously (Lesiak
et al., 1983). After treatment with the RNase T2, the retention
time observed for the entire peak corresponding to the 3–5A
trimer was considerably lower (∼24 min). Thus, the products of
cleavage contained in the peak were characterised by mass spec-
trometry and two different products were identified: 3ꢀAMP and
adenosine. Thus, these two cleavage products had already been
identified after the treatment of a 3–5A trimer of adenosine with

B. Morin et al. / Antiviral Research 87 (2010) 345–352
351
Budowsky, E.I., Kayushina, E.N., Tarasov, A.K., Orlenko, S.A., Cherkasov, I.A., Gavrilov,
A.E., Strelenko, Y.A., 1994. Preparation of cyclic 2ꢀ,3ꢀ-monophosphates of
oligoadenylates (A2ꢀp)nA > p and A3ꢀp(A2ꢀp)n-1A > p. Eur. J. Biochem. 220 (1),
97–104.
Dong, B., Silverman, R.H., 1995. 2–5A-dependent RNase molecules dimerize during
activation by 2–5A. J. Biol. Chem. 270 (8), 4133–4137.
Dong, B., Silverman, R.H., 1997. A bipartite model of 2–5A-dependent RNase L. J. Biol.
Chem. 272 (35), 22236–22242.
Han, J.Q., Townsend, H.L., Jha, B.K., Paranjape, J.M., Silverman, R.H., Barton, D.J.,
2007. A phylogenetically conserved RNA structure in the poliovirus open read-
ing frame inhibits the antiviral endoribonuclease RNase L. J. Virol. 81 (11),
5561–5572.
Fig. 4. Analysis of rRNA cleavage in mammalian cells following 2–5A transfection.
Hela cells were transfected with oligofectamine reagent alone (Mock), Poly (I:C) at
0.5 and 2 ␮g/ml or various 2–5As at increasing concentrations (1–10 ␮M). Total RNA
was extracted 3 h after transfection and analysed using agarose gel electrophoresis
(see Section 2 for experimental details). M: RNA 0.5–10 kb Ladder; 2 (2–5A): dimer;
3 (2–5A): trimer; 4 (2–5A): tetramer. Arrows indicate rRNA degradation products.
Results are representative of three independent experiments.
Hartmann, R., Justesen, J., Sarkar, S.N., Sen, G.C., Yee, V.C., 2003. Crystal structure of
the 2ꢀ-specific and double-stranded RNA-activated interferon-induced antiviral
protein 2ꢀ-5ꢀ-oligoadenylate synthetase. Mol. Cell. 12 (5), 1173–1185.
Hovanessian, A.G., Justesen, J., 2007. The human 2ꢀ-5ꢀoligoadenylate synthetase
family: unique interferon-inducible enzymes catalyzing 2ꢀ-5ꢀ instead of 3ꢀ-5ꢀ
phosphodiester bond formation. Biochimie 89 (6–7), 779–788.
Imai, J., Torrence, P.F., 1981. An efficient chemical synthesis of adenylyl(2ꢀ goes to
5ꢀ)adenylyl(2ꢀ goes to 5ꢀ)adenosine [(2ꢀ-5ꢀ)-oligo(A)]. Methods Enzymol. 79 (Pt
B), 233–244.
Justesen, J., Ferbus, D., Thang, M.N., 1980. 2ꢀ5ꢀ oligoadenylate synthetase, an
interferon induced enzyme: direct assay methods for the products, 2ꢀ5ꢀ oligoad-
enylates and 2ꢀ5ꢀ co-oligonucleotides. Nucleic Acids Res. 8 (14), 3073–3085.
Kajaste-Rudnitski, A., Mashimo, T., Frenkiel, M.P., Guenet, J.L., Lucas, M., Despres, P.,
2006. The 2ꢀ,5ꢀ-oligoadenylate synthetase 1b is a potent inhibitor of West Nile
virus replication inside infected cells. J. Biol. Chem. 281 (8), 4624–4637.
Kim, E.E., Wyckoff, H.W., 1991. Reaction mechanism of alkaline phosphatase based
on crystal structures. Two-metal ion catalysis. J. Mol. Biol. 218 (2), 449–464.
Kitade, Y., Nakata, Y., Hirota, K., Maki, Y., Pabuccuoglu, A., Torrence, P.F., 1991. 8-
Methyladenosine-substituted analogues of 2–5A: synthesis and their biological
activities. Nucleic Acids Res. 19 (15), 4103–4108.
et al., 2005; Tanaka et al., 2004). The ankyrin domain consists
of eight complete ankyrin repeats which will bind the two 2–5A
ends (AMP moieties) at repeats 2 and 4. The central AMP moi-
ety is oriented towards the solvent and even if it is involved in
weak interactions with the protein it does not interact directly
with the ankyrin region (Tanaka et al., 2004), thus pointing out
the importance of 2–5A length for RNase L binding and activation
(Dong and Silverman, 1997; Nakanishi et al., 2005; Tanaka et al.,
2004; Xiang et al., 2003). In fact short 2–5A dimers are not able
to interact with both ankyrin repeat sites at the same time and
thus are not able to bind nor to activate RNAse L (Tanaka et al.,
2004; Xiang et al., 2003). Hence, the physiological effect of the
purified 2–5As corresponds to what was previously reported, i.e.,
the dimer is inactive while the trimer and the tetramer show the
expected RNase L activation effect (Tanaka et al., 2004; Xiang et al.,
2003).
In conclusion, we have made use of two mammalian OASs to
produce a variety of 2–5As that were subsequently HPLC puri-
fied and individually characterised. The method described herein is
relatively straightforward and does not require a specialized equip-
ment other than an HPLC set-up. The large amount and purity
of these 2–5As should prove useful to study the yet ill-defined
pathway of viral innate immunity involving 2ꢀ5ꢀOAS and RNase
L.
Kodym, R., Kodym, E., Story, M.D., 2009. 2ꢀ-5ꢀ-Oligoadenylate synthetase is activated
by a specific RNA sequence motif. Biochem. Biophys. Res. Commun. 388 (2),
317–322.
Lesiak, K., Imai, J., Floyd-Smith, G., Torrence, P.F., 1983. Biological activities of phos-
phodiester linkage isomers of 2–5A. J. Biol. Chem. 258 (21), 13082–13088.
Malathi, K., Dong, B., Gale Jr., M., Silverman, R.H., 2007. Small self-RNA generated by
RNase L amplifies antiviral innate immunity. Nature 448 (7155), 816–819.
Marie, I., Blanco, J., Rebouillat, D., Hovanessian, A.G., 1997. 69-kDa and 100-kDa
isoforms of interferon-induced (2ꢀ–5ꢀ) oligoadenylate synthetase exhibit differ-
ential catalytic parameters. Eur. J. Biochem. 248 (2), 558–566.
Min, J.Y., Krug, R.M., 2006. The primary function of RNA binding by the influenza A
virus NS1 protein in infected cells: Inhibiting the 2ꢀ-5ꢀ oligo (A) synthetase/RNase
L pathway. Proc. Natl. Acad. Sci. U. S. A. 103 (18), 7100–7105.
Nakanishi, M., Tanaka, N., Mizutani, Y., Mochizuki, M., Ueno, Y., Nakamura, K.T.,
Kitade, Y., 2005. Functional characterization of 2ꢀ,5ꢀ-linked oligoadenylate bind-
ing determinant of human RNase L. J. Biol. Chem. 280 (50), 41694–41699.
Rebouillat, D., Hovanessian, A.G., 1999. The human 2ꢀ,5ꢀ-oligoadenylate synthetase
family: interferon-induced proteins with unique enzymatic properties. J. Inter-
feron Cytokine Res. 19 (4), 295–308.
Rebouillat, D., Hovnanian, A., Marie, I., Hovanessian, A.G., 1999. The 100-kDa 2ꢀ,5ꢀ-
oligoadenylate synthetase catalyzing preferentially the synthesis of dimeric
pppA2ꢀp5ꢀA molecules is composed of three homologous domains. J. Biol. Chem.
274 (3), 1557–1565.
Roberts, W.K., Hovanessian, A., Brown, R.E., Clemens, M.J., Kerr, I.M., 1976.
Interferon-mediated protein kinase and low-molecular-weight inhibitor of pro-
tein synthesis. Nature 264 (5585), 477–480.
Rusch, L., Dong, B., Silverman, R.H., 2001. Monitoring activation of ribonuclease L by
2ꢀ,5ꢀ-oligoadenylates using purified recombinant enzyme and intact malignant
glioma cells. Methods Enzymol. 342, 10–20.
Rushizky, G.W., Sober, H.A., 1963. Studies on the specificity of ribonuclease T2. J.
Biol. Chem. 238, 371–376.
Sarkar, S.N., Sen, G.C., 1998. Production, purification, and characterization of recom-
binant 2ꢀ, 5ꢀ-oligoadenylate synthetases. Methods 15 (3), 233–242.
Sawicki, D.L., Silverman, R.H., Williams, B.R., Sawicki, S.G., 2003. Alphavirus minus-
strand synthesis and persistence in mouse embryo fibroblasts derived from mice
lacking RNase L and protein kinase R. J. Virol. 77 (3), 1801–1811.
Sen, G.C., 2001. Viruses and interferons. Annu. Rev. Microbiol. 55, 255–281.
Servant, M.J., Grandvaux, N., Hiscott, J., 2002. Multiple signaling pathways leading
to the activation of interferon regulatory factor 3. Biochem. Pharmacol. 64 (5–6),
985–992.
Acknowledgments
We wish to thank Jean-Jacques Vasseur (IBMM, Montpellier)
for MALDI-TOF spectrometry analysis, Gilles Valette (Plateau tech-
nique, Institut des Biomolécules Max Mousseron (IBMM)) for others
mass spectrometry analysis, and Saumendra N. Sarkar and Games C.
Sen from the Lerner Research Institute, Cleveland, USA for the gen-
erous gift of the human OAS2 cDNA clone. The help and advice of
Stéphane Canaan and Jean Claude Bakala (IBSM-EIPL, Marseille) in
baculovirus expression is gratefully acknowledged. This work was
supported by a grant from the Agence Nationale de la Recherche
(ANR), programme “Maladies infectieuses et leur environnement”,
project CHIKVIRULENCE, and a grant from the Direction Générale
de l’Armement (contract 09cà40).
Shuman, S., 1990. Catalytic activity of vaccinia mRNA capping enzyme subunits
coexpressed in Escherichia coli. J. Biol. Chem. 265 (20), 11960–11966.
Silverman, R.H., 2007a. A scientific journey through the 2–5A/RNase L system.
Cytokine Growth Factor Rev. 18 (5–6), 381–388.
Appendix A. Supplementary data
Silverman, R.H., 2007b. Viral encounters with 2ꢀ,5ꢀ-oligoadenylate synthetase and
RNase L during the interferon antiviral response. J. Virol. 81 (23), 12720–12729.
Smith, J.A., Schmechel, S.C., Williams, B.R., Silverman, R.H., Schiff, L.A., 2005.
Involvement of the interferon-regulated antiviral proteins PKR and RNase L in
reovirus-induced shutoff of cellular translation. J. Virol. 79 (4), 2240–2250.
Tanaka, N., Nakanishi, M., Kusakabe, Y., Goto, Y., Kitade, Y., Nakamura, K.T., 2004.
Structural basis for recognition of 2ꢀ,5ꢀ-linked oligoadenylates by human ribonu-
clease L. Embo J. 23 (20), 3929–3938.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.antiviral.2010.06.003.
References
Behera, A.K., Kumar, M., Lockey, R.F., Mohapatra, S.S., 2002. 2ꢀ-5ꢀ Oligoadenylate
synthetase plays a critical role in interferon-gamma inhibition of respiratory
syncytial virus infection of human epithelial cells. J. Biol. Chem. 277 (28),
25601–25608.
Thakur, C.S., Jha, B.K., Dong, B., Das Gupta, J., Silverman, K.M., Mao, H., Sawai, H., Naka-
mura, A.O., Banerjee, A.K., Gudkov, A., Silverman, R.H., 2007. Small-molecule

352
B. Morin et al. / Antiviral Research 87 (2010) 345–352
activators of RNase L with broad-spectrum antiviral activity. Proc. Natl. Acad.
Sci. U. S. A. 104 (23), 9585–9590.
ated with prostate cancer on apoptosis induced by 2ꢀ,5ꢀ-oligoadenylates. Cancer
Res. 63 (20), 6795–6801.
Xiang, Y., Condit, R.C., Vijaysri, S., Jacobs, B., Williams, B.R., Silverman, R.H., 2002.
Blockade of interferon induction and action by the E3L double-stranded RNA
binding proteins of vaccinia virus. J. Virol. 76 (10), 5251–5259.
Xiao, W., Player, M.R., Li, G., Zhang, W., Lesiak, K., Torrence, P.F., 1996. Synthesis and
characterization of composite nucleic acids containing 2ꢀ, 5ꢀ-oligoriboadenylate
linked to antisense DNA. Antisense Nucleic Acid Drug Dev. 6 (4), 247–258.
Xiang, Y., Wang, Z., Murakami, J., Plummer, S., Klein, E.A., Carpten, J.D., Trent, J.M.,
Isaacs, W.B., Casey, G., Silverman, R.H., 2003. Effects of RNase L mutations associ-