RNA (2001), 7 :1486–1495+ Cambridge University Press+ Printed in the USA+

Copyright © 2001 RNA Society+

DOI: 10+1017+S1355838201014078

METHOD

Synthesis and properties of mRNAs containing

the novel “anti-reverse” cap analogs
7-methyl(39-O -methyl)GpppG and
7-methyl(39-deoxy)GpppG

JANUSZ STEPINSKI,1 CLINT WADDELL,2 RYSZARD STOLARSKI,1

EDWARD DARZYNKIEWICZ,1 and ROBERT E. RHOADS 2
1
Department of Biophysics, University of Warsaw, 02-089 Warsaw, Poland
2
Department of Biochemistry and Molecular Biology, Louisiana State University Health Sciences Center,
Shreveport, Louisiana, 71130-3932, USA

ABSTRACT
The ability to synthesize capped RNA transcripts in vitro using bacteriophage polymerases has been of considerable
value in a variety of applications. However, Pasquinelli et al. [RNA (1995) 1:957–967] found that one-third to one-half
of the caps are incorporated in the reverse orientation, that is, with the m 7 G moiety of m 7 GpppG linked by a 39-59
phosphodiester bond to the first nucleotide residue of the RNA chain. Such reverse caps are unlikely to be recognized
by eIF4E, based on previous studies, and thus complicate any comparison of the translational efficiencies of in
vitro-synthesized mRNAs. We therefore designed two novel cap analogs, P 1 -39-deoxy-7-methyguanosine-59 P 3 -
guanosine-59 triphosphate and P 1 -39-O,7-dimethylguanosine-59 P 3 -guanosine-59 triphosphate, that are, theoretically,
incapable of being incorporated in the reverse orientation. The key reactions of pyrophosphate bond formation were
achieved in anhydrous dimethylformamide solutions employing the catalytic properties of zinc salts. Structures were
proven by 1 H NMR. Transcripts produced with SP6 polymerase using “anti-reverse” cap analogs (ARCAs) were of the
predicted length and indistinguishable in size and homogeneity from those produced with m 7 GpppG or GpppG.
Analysis of the transcripts with RNase T2 and tobacco acid pyrophosphatase indicated that reverse caps were formed
with m 7 GpppG but not with ARCAs. Both of the ARCAs inhibited cell-free translation with a KI similar to that of
m 7 GpppG. Finally, the translational efficiency of ARCA-capped transcripts in a rabbit reticulocyte lysate was 2.3- to
2.6-fold higher than that of m 7 GpppG-capped transcripts. This suggests the presence of reverse caps in conventional
in vitro-synthesized mRNAs reduces their translational efficiency.
Keywords: bacteriophage polymerases; cap analog inhibition; cap-dependent translation; in vitro transcription;
ribonuclease T2; tobacco acid pyrophosphatase; translational efficiency

INTRODUCTION ribonucleoside triphosphates and a cap dinucleotide

such as m 7 G(59)ppp(59)G+ The polymerase initiates
Because all cellular, nonorganeller mRNAs are capped
transcription with a nucleophilic attack by the 39-OH of
(Shatkin, 1987), as well as most of the U-type snRNAs
the Guo moiety in m 7 GpppG on the a-phosphate of the
(Mattaj, 1986), the ability to synthesize capped RNA
next templated nucleoside triphosphate, resulting in the
transcripts in vitro is a useful research tool+ The most
initial product m 7 GpppGpN+ Suppression of the alter-
frequently employed technique is to transcribe a DNA
native, GTP-initiated product pppGpN is achieved by
template with either a bacterial (Contreras et al+, 1982)
setting the ratio of m 7 GpppG to GTP to between 5 and
or bacteriophage (Konarska et al+, 1984; Yisraeli & Mel-
10 in the transcription reaction mixture+
ton, 1989) RNA polymerase in the presence of all four
Somewhat unexpectedly, Pasquinelli et al+ (1995)
found that bacteriophage polymerases also utilize the
Reprint requests to: Robert E+ Rhoads, Department of Biochem- 39-OH of the 7-methylguanosine moiety of m 7 GpppG
istry and Molecular Biology, Louisiana State University Health Sci-
ences Center, Shreveport, Louisiana 71130-3932, USA; e-mail: to initiate transcription, resulting in one-third to one-half
rrhoad@lsuhsc+edu+ of the products containing a “reverse cap,” that is,
1486
mRNAs containing anti-reverse caps 1487

Gpppm 7GpN+ These authors demonstrated that reverse- dokura et al+, 1997) instead of Mn 21 , and using anhy-
capped pre-U1 RNA transcripts, when injected into Xen- drous dimethylformamide (DMF) instead of water as a
opus laevis nuclei, are exported more slowly than natural solvent, allowed us to obtain high yields (see Materials
transcripts+ Similarly, cytoplasmic reverse-capped U1 and Methods)+ In a similar reaction, the synthesis of
RNAs fail to be imported into the nucleus+ The pres- CH3pppG from GDP and the imidazolide of methyl phos-
ence of a cap on mRNA has a strong stimulatory effect phate in DMF, Kadokura et al+ (1997) obtained a yield
on translation (reviewed by Shatkin, 1985; Rhoads, of 39% without ZnCl2 compared with 98% with ZnCl2 +
1985)+ The ability of reverse-capped mRNAs to be trans- Furthermore, carrying out the reactions in aqueous
lated was not tested by Pasquinelli et al+ (1995), but Mn 21 resulted in yields of 42% and 37%, respectively,
based on what is known about recognition of the cap for the coupling of Im(m 7 GDP) plus GMP (Sawai et al+,
structure by eIF4E (Marcotrigiano et al+, 1997; Matsuo 1999) and Im(m 7 GMP) plus GDP (Sawai et al+, 1991),
et al+, 1997; Cai et al+, 1999), one would not expect whereas we obtained yields of 78% and 88%, respec-
reverse-capped mRNAs to be translated any more ef- tively, for the same reactions using DMF and ZnCl2 +
13
ficiently than uncapped RNAs+ C NMR and UV spectra for intermediates were
The presence of a significant proportion of reverse- in good agreement with the predicted structures (data
capped RNAs in the products of in vitro transcription not shown)+ The 1 H NMR assignments of the protons
should decrease the overall translational activity of the in both ARCAs confirmed their chemical structures
RNA preparation+ In addition, heterogeneous popula- (Table 1)+ Two sets of sugar 1 H signals in each spec-
tions of RNA molecules would compromise the deter- trum point to dinucleotides+ The presence of the methyl
mination of translational efficiency and other functional signals at 4+068 ppm (compound 10) and 4+027 (com-
properties of mRNA+ For these reasons, we designed pound 9) together with the disappearance of the H(8)
two cap analogues that, theoretically, cannot be incor- resonances, due to exchange for solvent deuterium
porated in the reverse orientation+ We found that these (Darzynkiewicz et al+, 1990), testify to the presence of
“anti-reverse” cap analogs (ARCAs) are, in fact, in- 7-methylguanine+ In the case of compound 10, the ad-
corporated exclusively in the normal orientation+ Fur- ditional methyl group was observed at 3+483 ppm, ac-
thermore, they behave like the natural caps in their companied by a characteristic upfield shift of the H39
interactions with the translational machinery, resulting signal+ Lack of the 39-hydroxyl in compound 9 gives the
in mRNAs that are translationally more active+ characteristic “deoxy” pattern of H39/H30 at 2+086–
2+148 ppm with further scalar couplings to H49 and H29+
Information regarding conformational parameters is
RESULTS AND DISCUSSION
gathered in Table 2+ This shows populations of the N
Transcription by bacteriophage RNA polymerases in the form in the N m S dynamic equilibrium of the sugar
presence of m 7 GpppG is initiated with a nucleophilic ring, populations of the 1sc (gauche-gauche) con-
attack by the 39-OH of either the m 7 Guo or Guo moiety former about C49-C59, and populations of the ap
on the electrophilic a-phosphate of the first templated (gauche9-gauche9) conformer of the phosphate group+
nucleoside triphosphate+ The basic strategy of our The 7-substituted Guo moieties show the character-
approach was to eliminate one of these two 39-OH istic preference of the N conformer, up to 100% in
groups+ In the case of P 1 -39-deoxy-7-methyguano- the case of m 2 7, O 39 Guo, as opposed to Guo, in which
sine-59 P 3 -guanosine-59 triphosphate (Fig+ 1, com- the S conformer dominates+ The preference of 1sc is
pound 9, henceforth abbreviated m 7 39dGpppG), the
substitution was 2H for 2OH+ In the case of P 1 -
39-O,7-dimethylguanosine-59 P 3 -guanosine-59 triphos- TABLE 1+ 1 H NMR chemical shifts in parts per million (60+001) ver-
phate (Fig+ 1, compound 10, henceforth abbreviated sus internal sodium 3-trimethylsilyl-[2,2,3,3- 2 H4 ]-propionate+
m 2 7, O 39 GpppG), the substitution was 2OCH3 for 2OH+ m 7 39dGpppG (9) m 2 7, O 39 GpppG (10)

m 7 39dG G m 2 7, O 39 G G
Synthesis of ARCAs
a a
H8 — 8+016 — 7+990
Chemical synthesis of the ARCAs was performed by H19 5+796 5+776 5+864 5+785
a methodology that differed in some respects from H29 4+587 4+650 4+682 4+687
methodologies reported previously (Stepinski et al+, H39 2+148 4+473 4+109 4+473
H30 2+086 — — —
1995; Jankowska et al+, 1996)+ To avoid preparation of
H49 4+728 4+346 4+428 4+339
imidazole derivatives from 7-methylated substrates, H59 4+460 4+26 b 4+384 4+278
the activation of which is often difficult, we developed H50 4+196 4+26 b 4+219 4+239
a new coupling strategy involving guanosine 59- CH3 4+027 — 4+068 (N7) —
phosphorimidazolide and the modified 7-methylated nu- 3+483 (39O)
cleoside diphosphate (Fig+ 1)+ Moreover, carrying out a
Deuterated+
the coupling reaction in the presence of ZnCl2 (Ka- b
Signal overlapping+
1488 J. Stepinski et al.

FIGURE 1. Synthesis of “anti-reverse” cap analogs+ ImGMP is guanosine 59-imidazolide monophosphate+

also more pronounced in the 7-substituted guano- rate experiments, the yield of product in the presence
sines (Darzynkiewicz et al+, 1990; Wieczorek et al+, of ARCAs was not statistically different from the yield
1997)+ The conformation of the Guo moiety of AR- in the presence of m 7 GpppG+
CAs is similar to that of Guo in normal caps, 64% of
the S form (36% N) and 63% of the 1sc form (Las-
sota et al+, 1984)+ Thus, m 27, O 39Guo and m 739dG show Analysis of cap orientation in ARCA-capped
conformational features characteristic of m 7Guo rather RNA transcripts
than Guo+ From the structure of the ARCAs, it should not be pos-
sible for them to be incorporated in the reverse orien-
tation+ We verified this experimentally by digestion with
Synthesis of ARCA-capped RNA transcripts
RNase T2 and tobacco acid pyrophosphatase (TAP)+ To
The ARCAs were next tested in an in vitro trans- obtain a higher proportion of radioactivity in the cap ver-
cription system+ A template DNA was generated by sus internal positions, a shorter DNA template was pro-
digesting the plasmid pSP-luc1 with Eco RI+ The theo- duced by cleaving pSP-luc1 with Nco I instead of Eco RI+
retical size of an RNA transcript from this template This was expected to yield an RNA product of 43 nt (plus
was 1,706 nt+ This was near the size of the prod- a cap)+ The size of this product was confirmed by poly-
ucts observed in reactions carried out in the pres- acrylamide gel electrophoresis in Tris/borate/EDTA/urea
ence of [a- 32 P]ATP and either GpppG, m 7 GpppG, (data not shown)+ RNase T2 digests RNA with no base
m 7 39dGpppG, or m 2 7, O 39 GpppG (Fig+ 2)+ In six sepa- specificity+ Thus, it will generate mostly 39-NMPs from this
mRNAs containing anti-reverse caps 1489

TABLE 2+ 1 H- 1 H and 1 H- 31 P coupling constants in hertz (60+2), and RNA+ Those nucleotide residues located 59 to an A res-
conformer populations (65%) in the dynamic equilibria N m S of the idue will acquire a 32 P-labeled 39-phosphate by nearest
sugar ring, and about C49-C59 (% 1 sc) and C59-O59 (% ap) bonds+
neighbor transfer+ The pyrophosphate bonds in the cap,
m 7 39dGpppG (9) m 2 7, O 39 GpppG (10) however, are not susceptible to RNase T2+ Because the
m 7 39dG G m 2 7, O 39 G G first nucleotide residue after the cap in the synthetic RNA
is A, the a-phosphate of [a- 32 P]ATP will be transferred
J(19,29) 0+0 a 6+2 4+0 6+3 to the cap upon RNase T2 digestion+ Thus, for RNAs
J(29,39) 4+5 5+2 5+0 5+1
J(29,30) 0+0 a — — —
initiated in the normal orientation with m 7 GpppG, the
J(39,30) 14+2 — — — product will be m 7 GpppGp* (where p* represents 32 P)+
J(39,49) 10+4 3+7 5+1 3+6 RNase T2-digestion products expected from RNAs
J(30,49) 5+1 — — — initiated with GTP or with each of the four cap analogs
J(49,59) 3+0 b 4+0 b 3+0 4+1 in either normal or reverse orientations are shown in
J(49,50) 2+7 4+0 b 2+6 4+2
J(59,50) 11+6 b
11+5 11+8
Table 3+
J(59,P) 5+0 6+0 b 4+4 5+4 The RNase T2-digestion products of normal and
J(50,P) 5+8 6+0 b 5+9 6+5 reverse m 7 GpppG-capped RNAs (m 7 GpppGp* and
J(49,P) 1+0 b 1+0 b 1+0 b 1+0 b Gpppm 7 Gp*, respectively) have identical masses and
%N 100 37 56 36 charges; they would therefore be expected to elute from
% 1 sc c 80 55 b 80 54
% ap d 72 66 b 74 66
an anion exchange column at nearly the same time+
However, digestion of normal and reverse-capped
a
Less than the line width, ;1 Hz+ mRNAs with TAP yields two alternate labeled products,
b
Approximate value+
c
1synclinal, that is, O59 in gauche orientation to O49 and C39+ pGp* and pm 7 Gp*, that differ in both charge and mass,
d
Antiperiplanar, that is, P59 in trans orientation to C49+ because the m 7 group confers a positive charge on G
(Hendler et al+, 1970)+ The nucleotides pm 7 39dGp* and
pm 2 7, O 39 Gp* have the same charge as pm 7 Gp*+ Thus,
whereas RNAse T2 digestion alone would not be ex-
pected to distinguish between normal and reverse ori-
entations, the combination of RNAse T2 and TAP would
do so (see Table 3)+
RNAs were synthesized from the short DNA tem-
plate in the presence of [a- 32 P]ATP, the other three
nonradioactive NTPs, and either no cap analog, GpppG,
m 7 GpppG, m 7 39dGpppG, or m 2 7, O 39 GpppG+ The prod-
ucts were digested with RNase T2 and resolved by
anion exchange HPLC (Fig+ 3)+ Uncapped RNA yielded
primarily 39-NMPs (Fig+ 3A, 20–30 min) with a small
amount of material that may have been the partially
degraded product ppGp* (Fig+ 3A, 76 min)+ The ex-
pected product pppGp* was not observed+ Based on
its high negative charge, it may not have eluted from
the column+ Its presence, however, is likely, as RNase
T2 plus TAP digestion yielded pGp* (Fig+ 3B, 56 min)
where none existed previously (Fig+ 3A)+
In the case of GpppG-capped RNAs, RNase T2 alone
yielded a structure eluting at 89 min (Fig+ 3C) that is likely
to be GpppGp* (the presence of a second phosphate es-
ter reduces the charge relative to pppGp*)+ The minor
peak at 77 min may be the partially degraded product
ppGp*+ Consistent with these assignments, both com-
pounds disappeared upon TAP digestion simultaneously
FIGURE 2. Electrophoretic migration of capped luciferase mRNAs with the appearance of a new peak corresponding
transcribed as described in Materials and Methods from Eco RI- to pGp* at 56 min (Fig+ 3D)+ No pm 7 Gp* (42 min) was
digested pSP-luc 1 with SP6 polymerase and [a- 32 P]ATP in the formed, as expected+
presence of the cap dinucleotides shown+ Samples were run on a
1% agarose gel containing 0+12 M formaldehyde in 0+4 M 3-(N - The major highly charged RNase T2-resistant prod-
morpholino)propanesulfonic acid, pH 7+0, 0+1 M sodium acetate, uct from m 7 GpppG-capped RNA eluted at 73 min
and 0+01 M EDTA at 70 mA for 5 h (Greenberg, 1994)+ A phos- (Fig+ 3E) and is likely to be m 7 GpppGp*+ This com-
phorimage obtained on a Molecular Dynamics Storm 860 instru-
ment is shown+ The positions and sizes in kilobases of rabbit 28S pound elutes earlier than the peak at 89 min in Fig-
rRNA, 18S rRNA, and b-globin mRNA are shown+ ure 3C, tentatively assigned the structure GpppGp*,
1490 J. Stepinski et al.

FIGURE 3. Analysis of in vitro-synthesized RNAs by enzymatic digestion and anion exchange HPLC+ mRNAs were gen-
erated by transcription of Nco I-digested pSP-luc 1 with [a- 32 P]ATP and either no cap dinucleotide (A,B), GpppG (C,D),
m 7 GpppG (E,F), m 7 39dGpppG (G,H), or m 2 7, O 39 GpppG (I,J)+ Aliquots of 5 to 13 ng of RNA were digested with RNase T2
(left panels) or RNase T2 plus tobacco acid pyrophosphatase (TAP; right panels) as described in Materials and Methods+
Nucleotides and caps were separated on a Partisil 10SAX/25 column developed with a gradient of potassium phosphate,
pH 3+5, as described in Materials and Methods+ Fractions of 1 mL were collected and the Cerenkov radiation determined with
a Beckman LS 6500 scintillation counter+ The elution times of the following standard compounds, detected by UV absorption,
are shown: 39-CMP, 39-UMP, 39-AMP, 39-GMP, 59-GDP, 59-GTP, 39,59-GDP (pGp), 39,59-m 7 GMP (pm 7 Gp), and GpppG+
mRNAs containing anti-reverse caps 1491

TABLE 3+ Predicted and observed cap structures from in vitro-synthesized mRNAs after enzymatic digestion

Possible 59-end-labeled
digestion products Product observed

Cap dinucleotide Orientation a

Possible transcription products b
RNase T2 RNase T2 1 TAP RNase T2 1 TAP

None N/A pppGp*Ap(Np)40C pppGp* pGp* 100%

GpppG N/A GpppGp*Ap(Np)40C GpppGp* pGp* 100%
m 7 GpppG Normal m 7 GpppGp*Ap(Np)40C m 7 GpppGp* pGp* 67%
Reverse Gpppm 7 Gp*Ap(Np)40C Gpppm 7 Gp* pm 7 Gp* 33%
m 7 39dGpppG Normal m 7 39dGpppGp*Ap(Np)40C m 7 39dGpppGp* pGp* 100%
Reverse Gpppm 7 39dGp*Ap(Np)40C Gpppm 7 39dGp* pm 7 39dGp* 0%
m 2 7,O39 GpppG Normal m 2 7, O 39 GpppGp*Ap(Np)40C m 2 7, O 39 GpppGp* pGp* 100%
Reverse Gpppm 2 7, O 39 Gp*Ap(Np)40C Gpppm 2 7, O 39 Gp* pm 7, O 39 Gp* 0%
a
Normal orientation is when the 39-OH of Guo in the structure m 7 G(59)ppp(59)G (or its structural analogs) is attached to the first nucleotide
residue in the RNA chain by a 39-59 phosphodiester linkage+ Reverse orientation is when the 39-OH of m 7 Guo is the point of attachment+
b
Radioactive atoms ( 32 P) are indicated by *+

because of the additional positive charge+ The minor is well correlated for a variety of different cap analog
peak at 77 min may be the reverse cap Gpppm 7 Gp*, structures (Carberry et al+, 1990)+ We tested GpppG,
suggesting that the proximity of the 39-P to the pos- m 7GpppG, and the two ARCAs for their ability to com-
itive charge of the m 7 G ring may influence the charge pete with natural globin mRNA for recognition by the
on P+ These assignments are strengthened by the translational machinery and thereby inhibit translation
fact that TAP digestion converted these products to in a rabbit reticulocyte lysate (RRL) system (Fig+ 4)+
two labeled compounds eluting earlier, pGp* (56 min) GpppG was not an inhibitor, and in fact slightly stim-
and pm 7 Gp* (42 min) (Fig+ 3F)+ The ratios of pGp* ulated protein synthesis at low concentrations+ The
and pm 7 Gp* suggest that they were derived from the two ARCAs, on the other hand, were equally inhibi-
73- and 77-min peaks of Figure 3E, respectively+ tory as m 7 GpppG+
With the ARCA m 739dGpppG, an RNase T2-resistant It is possible to compare cap analogs quantitatively as
product was observed at 78 min that is likely to be inhibitors by fitting a theoretical curve to the translation
m 7 39dGpppGp* (Fig+ 3G)+ It elutes at nearly the same data (Cai et al+, 1999)+ The curve is derived from an
time as the compound thought to be m 7 GpppGp* equation that treats cap analogs as competitive inhibi-
(78 min versus 77 min for Fig+ 3G and E, respectively)+
Interestingly, whereas there were two peaks in this re-
gion for RNA synthesized with m7GpppG (Fig+ 3E), there
was only one for RNA synthesized with the ARCA
(Fig+ 3G), consistent with the inability of the ARCA to be
incorporated in the reverse orientation+ When digested
with TAP, this peak at 78 min disappeared and a new
one appeared at the elution time of pGp* (Fig+ 3H,
56 min)+ The fact that no pm 7 Gp* appeared at 42 min
with the ARCA (Fig+ 3H), whereas it did with m 7 GpppG
(Fig+ 3F), is further proof that the ARCA can be incor-
porated in only one orientation+
The products observed upon digestion of RNA
synthesized with the second ARCA, m 2 7, O 39 GpppG
(Fig+ 3I,J), eluted the same as those obtained with
m 7 39dGpppG-capped RNA+

Inhibition of protein synthesis by ARCAs

One criterion for assessing cap analog interaction with
FIGURE 4. Inhibition of translation by ARCAs compared with
the translational machinery is inhibition of protein syn- m 7 GpppG and GpppG+ Natural rabbit globin mRNA was translated at
thesis (Canaani et al+, 1976; Hickey et al+, 1976)+ The 5 mg/mL in a RRL system, and globin synthesis was detected by
binding of cap analogs to eIF4E, measured in vitro incorporation of [ 3 H]Leu into protein as described in Materials and
Methods+ The following cap analogs were included during translation
with purified components, and the inhibition of protein at the indicated concentrations: GpppG: circles; m 7 GpppG: squares;
synthesis, measured in a cell-free translation system, m 7 39dGpppG: triangles; m 2 7, O 39 GpppG: diamonds+
1492 J. Stepinski et al.

tors of mRNA binding to eIF4E; the value of the disso- GpppG-capped mRNAs indicates that the RRL system
ciation constant for the cap analog•eIF4E complex, employed was highly dependent on the cap+ In the ex-
KI , is varied until the best least-squares fit is obtained+ periment shown, the m27, O 39GpppG-and m 739dGpppGp-
The curves in Figure 4 for m 7 GpppG, m 7 39dGpppG, capped mRNAs were 2+3- and 2+6-fold more efficient
and m 2 7, O 39 GpppG are, in fact, theoretical curves cor- than m 7 GpppG-capped mRNA, respectively+ In six ex-
responding to KI values of 27+8 6 12+6, 27+8 6 7+1, and periments employing four separate batches of in vitro-
14+3 6 1+9 mM, respectively+ Although it appears in this synthesized mRNAs, the mRNAs produced with ARCAs
experiment that the m 27, O 39GpppG compound was more were reproducibly more active than with m 7 GpppG+
inhibitory, in a replicate of this experiment, the KI values Pasquinelli et al+ (1995) found that reverse capping
for the ARCAs did not differ statistically from those of varied between 28 and 48%, depending on the pH of
m 7 GpppG+ This is also in agreement with a study show- the in vitro-transcription reaction+ In the experiment
ing that m 2 7, O 39 GMP and m 7 GMP were equally inhibi- shown in Figure 3 and summarized in Table 3, reverse
tory for translation (Darzynkiewicz et al+, 1985)+ capping was approximately 33%+ If ARCAs and normal
caps stimulate translation to the same degree, which
seems likely based on the inhibition data (Fig+ 4), one
Translation of ARCA-capped mRNAs
would expect the (homogeneous) preparation of ARCA-
Based on the results presented thus far, namely, that capped mRNA to be only 1+5-fold more active than the
one-third to one-half of m 7 GpppG is incorporated into (heterogeneous) preparation of m 7 GpppG-capped
RNA in reverse orientation (Pasquinelli et al+, 1995; mRNAs+ The fact that the RNA was even more active
Fig+ 2; Table 3), that ARCAs are incorporated exclusively (2+3- to 2+6-fold) suggests that reverse-capped mRNAs
in the normal orientation (Fig+ 3), and that ARCAs are may actually be inhibitory rather than merely neutral in
recognized to the same extent as m 7 GpppG by the their effect on translation+
translational machinery (Fig+ 4), one would predict that These results indicate that ARCAs are very similar to
the homogeneous population of in vitro-synthesized normal cap analogs+ The modifications at the 39-O -
ARCA-capped mRNAs would be more active transla- position of m 7 Guo do not appear to affect conformation
tionally than m 7 GpppG-capped mRNAs+ We tested this (Table 2) or interaction with the translational machinery
with luciferase mRNAs that were either uncapped or (Figs+ 4 and 5)+ This is in agreement with the X-ray-
capped with GpppG, m 7 GpppG, or the two ARCAs derived structure of an NH2-terminally truncated frag-
(Fig+ 5)+ The fact that all the m 7 G-containing mRNAs ment of mouse eIF4E in complex with m 7 GDP, which
were translated more efficiently than the uncapped or shows that the 39-OH of the m 7 Guo moiety projects
away from the protein and is accessible to solvent (Mar-
cotrigiano et al+, 1997)+ Consequently, the ARCAs have
the advantage of being incorporated exclusively in the
normal orientation but no apparent disadvantages+ To
our knowledge, the degree to which m 7 G is incorpo-
rated in place of G in internal positions of a synthetic
RNA chain by bacteriophage polymerases has not been
rigorously determined+ Regardless of the level of this
misincorporation, the ARCAs would be incapable of
donating m 7 G internally as well as at the 59 end+ A
different type of ARCA, for example, m 4 2,2,7, O 39 GpppG
or m 3 2,2,7 39dGpppG, would be useful for in vitro syn-
thesis of U-type snRNAs with 100% normal cap
orientation+

MATERIALS AND METHODS

Column chromatography
FIGURE 5. Translational activity of ARCA-capped mRNAs+ Lucifer-
Both final products (9 and 10) and intermediate nucleotides
ase mRNAs were synthesized in vitro using SP6 RNA polymerase in
the presence of all four NTPs and either no cap analog (circles), (3–8) were isolated from reaction mixtures by column chro-
GpppG (squares), m 7 GpppG (diamonds), m 2 7, O 39 GpppG (inverted matography on DEAE-Sephadex (A-25, HCO3 2 form) using
triangles), or m 7 39dGpppG (triangles)+ The RNAs were translated for a linear gradient of triethylammonium bicarbonate (TEAB),
60 min in a RRL system as described in Materials and Methods, and pH 7+5, in water+ Fractions were collected, and products peaks
luciferase activity was measured in triplicate by luminometry (RLU:
relative light units)+ Translational efficiency for each mRNA was es-
(monitored at 260 nm) were pooled and evaporated to dry-
timated from the slopes of the curves of luciferase activity versus ness, with ethanol added repeatedly to remove the TEAB
mRNA concentration+ buffer+ Thus, the products were obtained as TEA salts+
mRNAs containing anti-reverse caps 1493

The purity of intermediates and products was monitored at the precipitate was dried in a vacuum desiccator over P4O 10 +
260 nm by analytical HPLC using a Spectra-Physics SP8800 The imidazolide thus obtained was dissolved in 1+2 mL of
apparatus on a 25-cm LC-18-T reverse phase column (Su- DMF, and 200 mg of tris(triethylammonium)phosphate was
pelco)+ The mobile phase was a linear gradient of methanol added+ The latter was prepared from TEA and phosphoric
from 0 to 25% in 0+1 M KH2PO4 , pH 6, over 15 min with flow acid followed by drying over P4O10 in a desiccator to obtain
rate of 1+3 mL/min+ semicrystalline mass+ Finally, 80 mg of ZnCl2 were added,
Mono- and dinucleotides obtained by enzymatic digestion and the reaction mixture was stirred at room temperature for
of in vitro-synthesized RNAs were analyzed by HPLC using a 6+5 h, poured into a beaker containing a solution of 250 mg
Waters 625 instrument with a 996 PDA detector on a 4+5 3 EDTA in 15 mL water, and neutralized with 1 M NaHCO3 +
250 mm Partisil 10SAX/25 column (Whatman)+ The program Chromatographic isolation on DEAE-Sephadex using a lin-
for elution of nucleotides consisted of water for the first 5 min, ear gradient of 0–1 M TEAB gave 5 (yield: 41 mg, 66%)+
a linear gradient of 0 to 87+5 mM KH2PO4 , pH 3+5, over
35 min, a linear gradient of 87+5 to 500 mM of KH 2PO4 over 3 9-O-Methyguanosine 5 9-diphosphate (6)
30 min, and isocratic elution at 500 mM KH 2PO4 for 21 min,
all at a flow rate of 1 mL/min+ This compound was obtained by a procedure analogous to
that of 5 except starting from 58 mg of 4 (yield: 32 mg, 49%)+
Spectroscopy
1
3 9-Deoxy-7-methyguanosine 5 9-diphosphate (7)
H NMR and 13 C NMR spectra were recorded on a Varian
UNITYplus 500 MHz instrument in dimethylsulfoxide-d6 (for Compound 5 (34 mg, 0+055 mmol) was mixed with 1 mL of
nucleoside intermediates) and 2 H 2O (for mono- and dinucle- dimethylsulfoxide, 1 mL of DMF, and 100 mL of methyl iodide
otides)+ The absorption spectra were obtained on a Cary 3E at room temperature+ After 3 h, the reaction mixture was treated
spectrophotometer+ with 30 mL of cold water and extracted three times with 10-mL
1
H NMR spectra of 9 and 10 were run at 25 8C at 1+4 portions of diethyl ether+ Chromatography of the aqueous
mg/0+7 mL and 0+4 mg/0+7 mL in 2 H2O, respectively+ Confor- phase, after neutralization with NaHCO3 , on DEAE-Sephadex
mations of the sugar moieties were derived from the vicinal using a linear gradient of 0 to 0+8 M TEAB gave 7 (yield:
1
H- 1 H coupling constants (Haasnoot et al+, 1980)+ Conforma- 10 mg, 28%)+
tions of the phosphate groups were determined from the 1 H-
31
P coupling constants (Lankhorst et al+, 1984)+ 3 9-O, 7-Dimethylguanosine 5 9-diphosphate (8)

Synthesis of mono- and dinucleotides This compound was obtained by a procedure analogous to
that of 7, except starting from 66 mg of 6 (yield: 64 mg, 95%)+
3 9-Deoxyguanosine 5 9-monophosphate (3)
P 1 -3 9-Deoxy-7-methyguanosine-5 9
39-Deoxyguanosine (1, commercial product from Sigma, 50 P 3 -guanosine-5 9 triphosphate (9)
mg, 0+19 mmol) was stirred overnight with trimethylphos-
phate (2 mL) and phosphorus oxychloride (53 mL, 0+57 mmol) GMP (purchased from Sigma, converted to the TEA salt, 29
at 6 8C+ Adding 20 mL of water and neutralizing with 1 M mg, 0+05 mmol), imidazole (17 mg, 0+25 mmol), and 2,29-
NaHCO3 quenched the reaction+ DEAE-Sephadex chroma- dithiodipyridine (22 mg, 0+1 mmol, from Aldrich) were mixed
tography using a linear gradient of 0–0+9 M TEAB afforded 3 in anhydrous DMF (1+2 mL) and TEA (7 mL)+ Triphenylphos-
(yield: 45 mg, 43%)+ phine (26 mg, 0+1 mmol) was added, and the mixture was
stirred for 5 h at room temperature+ The mixture was placed
in a centrifuge tube, and sodium perchlorate (25 mg, anhy-
3 9-O-Methyguanosine 5 9-monophosphate (4)
drous) dissolved in acetone (6 mL) was added+ The proce-
This compound was obtained by a procedure analogous to dure for washing the precipitate with acetone and drying over
that of 3, but starting from 59 mg of 39-O -methylguanosine P4O10 was the same as for 5+ The imidazolide of GMP thus
(2), which was prepared as previously described (Kusmierek obtained was dissolved in DMF (1+2 mL), and 7 (10 mg, TEA
& Shugar, 1978) (yield: 80 mg, 69%)+ salt, 0+015 mmol) was added+ Next, ZnCl2 (40 mg) was added,
and the mixture was stirred at room temperature overnight,
poured into a beaker containing a solution of 125 mg of EDTA
3 9-Deoxyguanosine 5 9-diphosphate (5)
in 15 mL of water, and neutralized with 1 M NaHCO3 + Chro-
Compound 3 (55 mg, TEA salt, 0+1 mmol), imidazole (34 mg, matographic isolation as for 5 gave 9 (13 mg, 88% based on
0+5 mmol), and 2,29-dithiodipyridine (Aldrich, 44 mg, 0+2 mmol) the amount of 7 used)+
were mixed in anhydrous DMF (1+2 mL) and TEA (14 mL)+
Triphenylphosphine (52 mg, 0+2 mmol) was added, and the P 1 -3 9-O, 7-Dimethylguanosine-5 9
mixture was stirred for 5 h at room temperature+ The mixture P 3 -guanosine-5 9 triphosphate (10)
was placed in a centrifuge tube, and sodium perchlorate
(49 mg, anhydrous) dissolved in acetone (6 mL) was added+ This compound was prepared from GMP and 8 (34 mg) by a
After cooling for 2 h in a refrigerator, the mixture was centri- procedure analogous to that for 9 (yield: 23 mg, 78%)+
fuged and the supernatant was discarded+ The precipitate The final products (9 and 10) were converted to their Na 1
was ground with a new portion of acetone, cooled, and cen- salts by ion exchange using a small column of Dowex 50Wx8
trifuged again+ The process was repeated once more, and (Na 1 form), evaporation of the eluates to a small volume,
1494 J. Stepinski et al.

precipitation with ethanol, and centrifugation to give amor- some cases, the added mRNA was natural rabbit globin mRNA
phous white powders+ Analyses of purity were essentially the and protein synthesis was measured by incorporation of
same as described earlier (Darzynkiewicz et al+, 1990)+ Pa- [ 3 H]Leu into trichloroacetic acid-precipitable form+ In other
rameters from the 1 H NMR spectra of 9 and 10 are shown in cases, the added mRNAs were luciferase mRNAs (the long
Tables 1 and 2+ form), synthesized in vitro as described above, and protein
synthesis was measured by assaying luciferase activity using
beetle luciferin (Promega) as substrate and a Monolite 2010
7-Methylguanosine 3 9,5 9-diphosphate. luminometer to detect light emission+
The ability of cap analogs to inhibit cell-free translation in
Guanosine 39,59-diphosphate was methylated to make the
the RRL system programmed with globin mRNA was mea-
chromatographic standard pm 7 Gp (Fig+ 3) by the same pro-
sured as described previously (Cai et al+, 1999)+ Data were fit
cedure as used for 7+
by least squares minimization to a theoretical rate equation+
The concentrations of cap analog solutions were measured
by UV absorption at pH 7+0 using the following parameters
In vitro synthesis of RNA
(l, eM ): GpppG, 251 nm, 25+5 3 10 3 ; m 7GpppG, m 739dGpppG,
RNAs that were either capped with various types of cap and m 2 7, O 39 GpppG, 255 nm, 22+6 3 10 3 +
analogs or uncapped were synthesized by in vitro transcrip-
tion in two lengths+ The DNA template used for both lengths
of RNA was pSP-luc 1 (Promega), which contains an SP6 ACKNOWLEDGMENTS
bacteriophage promoter+ To generate the short RNAs (43 nt This project was supported by Grant No+ 6 P04A 055 17 from
exclusive of the cap), the plasmid was digested with Nco I+ the Polish Committee for Scientific Research (KBN), Grant
To generate the long RNAs (1,706 nt, containing the entire No+ MEN/NSF-98-337 from the U+S+-Polish Maria Sklodowska-
luciferase coding region), the plasmid was digested with Curie Joint Fund II, and Grant No+ GM20818 from the Na-
Eco RI+ A typical in vitro-transcription reaction mixture con- tional Institute of General Medical Sciences+ The authors wish
tained, in 20 mL, 40 mM Tris-HCl, pH 7+9, 6 mM MgCl2 , to express their gratitude to the Laboratory of Nuclear Mag-
2 mM spermidine, 10 mM DTT, 2 mg BSA, 20 U of RNasin netic Resonance, Institute of Biochemistry and Biophysics,
(Promega), 0+5 mM ATP, 0+5 mM CTP, 0+5 mM UTP, 0+1 mM Polish Academy of Sciences, for access to the NMR spec-
GTP, 1 mM cap analog (GpppG, m 7 GpppG, m 7 39dGpppG, trometer, and to Dorota Haber and Aneta Powierza for tech-
or m 2 7, O 39 GpppG), 0+2–1+0 mg DNA, and 20 U of SP6 nical assistance+
polymerase (Promega)+ Reactions to synthesize the short
RNAs also contained 28 mCi of [a- 32 P]ATP (ICN), whereas
Received June 11, 2001; returned for revision
those to synthesize the long RNAs contained 0+8 mCi of
June 22, 2001; revised manuscript received
[a- 32 P]CTP (ICN)+ Reaction mixtures were incubated for
July 9, 2001
60 min at 37 8C, extracted with phenol and chloroform, the
solution made 2 M in sodium acetate, the nucleic acids
precipitated with 3 vol of ethanol on dry ice for 5 min, and REFERENCES
the solution centrifuged at 14,000 rpm for 30 min+ The pel-
Cai A, Jankowska-Anyszka M, Centers A, Chlebicka L, Stepinski J,
let was dissolved in water, the solution made 0+2 M in so- Stolarski R, Darzynkiewicz E, Rhoads RE+ 1999+ Quantitative as-
dium acetate, and the nucleic acid was again precipitated sessment of mRNA cap analogs as inhibitors of in vitro transla-
with 2+5 vol of ethanol at 4 8C for 30 min and centrifuged at tion+ Biochemistry 38 :8538–8547+
14,000 rpm for 30 min+ The pellet was allowed to air dry Canaani D, Revel M, Groner Y+ 1976+ Translational discrimination of
“capped” and “non-capped” mRNAs: Inhibition by a series of chem-
and then dissolved in diethylpyrocarbonate-treated water+
ical analogs of m 7 GpppX+ FEBS Lett 64 :326–331+
Carberry SE, Darzynkiewicz E, Stepinski J, Tahara SM, Rhoads RE,
Goss DJ+ 1990+ A spectroscopic study of the binding of N-7-
Enzymatic digestion of RNAs substituted cap analogs to human protein synthesis initiation fac-
tor 4E+ Biochemistry 29 :3337–3341+
The short RNAs were digested with 67 U RNase T2 (Life Contreras R, Cheroutre H, Degrave W, Fiers W+ 1982+ Simple, effi-
Technologies) in 15 mL of 0+14 M sodium acetate, pH 4+6, at cient in vitro synthesis of capped RNA useful for direct expression
of cloned eukaryotic genes+ Nucleic Acids Res 10 :6353+
37 8C for 60 min+ In some cases, the RNAs were subjected to
Darzynkiewicz E, Ekiel I, Tahara SM, Seliger LS, Shatkin A+ 1985+
a two-step digestion instead+ The first digestion was with 10 Chemical synthesis and characterization of 7-methylguanosine
U TAP (Epicentre Technologies) in 5 mL of 50 mM sodium cap analogues+ Biochemistry 24 :1701–1707+
acetate, pH 6+0, 1 mM EDTA, 0+1% b-mercaptoethanol, and Darzynkiewicz E, Stepinski J, Tahara SM, Stolarski R, Ekiel I, Haber
0+01% Triton X-100 at 37 8C for 60 min+ The digestion was D, Neuvonen K, Lehikoinen P, Labadi I, Lönnberg H+ 1990+ Syn-
thesis, conformation and hydrolytic stability of 1,P3-dinucleoside
continued for 60 min at 37 8C with 67 U RNase T2 in a final triphosphates related to mRNA 59-cap, and comparative kinetic
volume of 16 mL of 0+12 M sodium acetate, pH 4+6+ Samples studies on their nucleoside and nucleoside monophosphate an-
were analyzed without further treatment by anion exchange alogs+ Nucleosides and Nucleotides 9 :599– 618+
HPLC as described above+ Greenberg ME+ 1994+ Preparation and analysis of RNA+ In: Ausubel
FM, Brent R, Kingston RE, Moore DD, Seidman JG, Smith JA,
Struhl K, eds+ Current protocols in molecular biology, Vol+ 1+ New
York: John Wiley and Sons, Inc+ p 4+9+3+
Cell-free translation Haasnoot CAG, deLeeuw FAAM, Altona C+ 1980+ The relationship
between proton–proton nmr coupling constants and substituent
A micrococcal nuclease-treated RRL system was used for in electronegativities+ Tetrahedron 36 :2783–2792+
vitro translation as described previously (Cai et al+, 1999)+ In Hendler S, Fürer E, Srinivasan PR+ 1970+ Synthesis and chemical
mRNAs containing anti-reverse caps 1495

properties of monomers and polymers containing 7-methylguanine m 7 GDP and interaction with 4E-binding protein+ Nature Struct
and an investigation of their substrate or template properties for Biol 4 :717–724+
bacterial DNA or RNA polymerases+ Biochemistry 9 :4141– 4153+ Mattaj IW+ 1986+ Cap trimethylation of U snRNA is cytoplasmic and
Hickey ED, Weber LA, Baglioni C+ 1976+ Inhibition of initiation of dependent on U snRNP protein binding+ Cell 46 :905–911+
protein synthesis by 7-methylguanosine-59-monophosphate+ Proc Pasquinelli AE, Dahlberg JE, Lund E+ 1995+ Reverse 59 caps in RNAs
Natl Acad Sci USA 73 :19–23+ made in vitro by phage RNA polymerases+ RNA 1:957–967+
Jankowska M, Stepinski J, Stolarski R, Wieczorek Z, Temeriusz A, Rhoads RE+ 1985+ The cap structure of eukaryotic messenger RNA
Haber D, Darzynkiewicz E+ 1996+ 1 H NMR and fluorescence stud- and its interaction with cap-binding protein+ Prog Mol Subcell Biol
ies of new mRNA 59-cap analogues+ Collect Czech Chem Com- 9 :104–155+
mun 61:S197–S202+ Sawai H, Wakai H, Nakamura-Ozaki A+ 1999+ Synthesis and reac-
Kadokura M, Wada T, Urashima C, Sekine M+ 1997+ Efficient synthe- tions of nucleoside 59-diphosphate imidazolide+ A nonenzymatic
sis of g-methyl-capped guanosine 59-triphosphate as a 59-terminal capping agent for 59-monophosphorylated oligoribonucleotides
unique structure of U6 RNA via a new triphosphate bond forma- in aqueous solution+ J Org Chem 64 :5836–5840+
tion involving activation of methyl phosphorimidazolidate using Sawai H, Wakai H, Shimazu M+ 1991+ Facile synthesis of cap portion
ZnCl2 as a catalyst in DMF under anhydrous conditions+ Tetra- of messenger RNA by Mn(II) ion catalyzed pyrophosphate for-
hedron Lett 38 :8359–8362+ mation in aqueous solution+ Tetrahedron Lett 32 :6905–6906+
Konarska MM, Padgett RA, Sharp PA+ 1984+ Recognition of a cap Shatkin AJ+ 1985+ mRNA cap binding proteins: Essential factors for
structure in splicing in vitro of mRNA precursors+ Cell 38 :731–736+ initiating translation+ Cell 40 :223–224+
Kusmierek J, Shugar D+ 1978+ A new route to 29(39)-O -alkyl purine Shatkin AJ+ 1987+ mRNA caps—old and newer hats+ Bioessays
nucleosides+ Nucleic Acids Res Special Publ No 4 :s73–s77+ 7 :275–277+
Lankhorst PL, Haasnoot CAG, Erkelens C, Altona C+ 1984+ Carbon-13 Stepinski J, Bretner M, Jankowska M, Felczak K, Stolarski R, Wiec-
NMR in conformational analysis of nucleic acid fragments+ Re- zorek Z, Cai A-L, Rhoads RE, Temeriusz A, Haber D, Darzynk-
parametrization of the Karplus equation for vicinal NMR coupling iewicz E+ 1995+ Synthesis and properties of P 1 , P 2 , P 3 - and P 1 ,P 4 -
constants in COOP and HCOP fragments+ J Biomol Struct Dyn dinucleoside di-, tri- and tetraphosphate mRNA 59-cap analogues+
1:1387–1405+ Nucleosides and Nucleotides 14 :717–721+
Lassota P, Stolarski R, Shugar D+ 1984+ Conformation about the gly- Wieczorek Z, Zdanowski K, Chlebicka L, Stepinski J, Jankowska M,
cosidic bond and susceptibility to 59-nucleotidase of 8-substituted Kierdaszuk B, Temeriusz A, Darzynkiewicz E, Stolarski R+ 1997+
analogues of 59-GMP+ Z Naturforsch 39c :55– 63+ Fluorescence and NMR studies of intramolecular stacking of
Marcotrigiano J, Gingras A-C, Sonenberg N, Burley SK+ 1997+ Co- mRNA cap-analogues+ Biochim Biophys Acta 1354 :145–152+
crystal structure of the messenger RNA 59 cap-binding protein Yisraeli JK, Melton DA+ 1989+ Synthesis of long, capped transcripts in
(eIF4E) bound to 7-methyl-GDP+ Cell 89 :951–961+ vitro by SP6 and T7 RNA polymerases+ Methods Enzymol 180 :42–
Matsuo H, Li H, McGuire AM, Fletcher CM, Gingras A-C, Sonenberg 50+
N, Wagner G+ 1997+ Structure of translation factor eIF4E bound to