The EMBO Journal vol.15 no.12 pp.3144-3152, 1996
The rpsO mRNA of Escherichia coli is
polyadenylated at multiple sites resulting from
endonucleolytic processing and exonucleolytic
Jeanette Haugel-Nielsen, Eliane Hajnsdorf
and Philippe Regnier1
Institut de Biologie Physico-Chimique, 13 rue Pierre et Marie Curie,
75005 Paris, France
TherpsO monocistronic messenger, encoding ribosomal
occurring at the hairpin structure of the transcription
terminator tl. We report that mRNA fragments differ-
ing from the monocistronic transcript by their 3'
termini are also polyadenylated in the absence of
polynucleotide phosphorylase and RNase
of these 3' extremities result from endonucleolytic
cleavages by RNase E and RNase III and from exonu-
cleolytic degradation. Most of these mRNA fragments
are destabilized upon polyadenylation with the excep-
tion ofthe RNA species generated by RNase III. RNase
E appears to reduce the amount of poly(A) added at
the transcription terminator tl.
Keywords: 3' to 5' exonucleases/mRNA polyadenylation/
mRNA stability/poly(A) polymerase I/RNase E
Post-transcriptional addition of poly(A) tails to the 3' ends
of mRNAs, which was first identified in eukaryotic cells,
now appears to be a characteristic of all living organisms
(Brawerman, 1981; Littauer and Sorek, 1982; Cohen,
1995; Hajnsdorf et al., 1995; O'Hara et al., 1995). There
is considerable evidence that the long poly(A) tails of
eukaryotic mRNAs are structural determinants of their
stabilities (Baker, 1993) while, in contrast, the shorter
oligo(A) tails found at the 3' ends of some bacterial
mRNAs (Sarkar et al., 1978; Karnik et al., 1987; Cao and
Sarkar, 1992a, 1993) have been recently shown to promote
their degradation (Hajnsdorf et al., 1995; O'Hara et al.,
1995). The pcnB gene coding for poly(A) polymerase I
(PAP I) of Escherichia coli has been cloned and sequenced
(Liu and Parkinson, 1989; March et al., 1989; Cao and
Sarkar, 1992b), and a second poly(A) polymerase, PAP
II, has been identified in a pcnB deletion strain (Kalapos
et al., 1994).
The rpsO mRNA, which encodes ribosomal protein
S15 of E.coli, is one of the RNAs destabilized upon
polyadenylation (Hajnsdorf et al., 1995). The rpsO gene
forms an operon with the downstream pnp gene, which
encodes polynucleotide phosphorylase (PNPase) (Regnier
and Portier, 1986; Regnier et al., 1987), one of the
exonucleases involved in mRNA degradation (Figure 1)
(Ehretsmann et al., 1992a). The most abundant rpsO
transcript, hereafter referred to as P1-tl, is monocistronic
34©Oxford University Press
and extends from the promoter P1 to the Rho-independent
transcriptional terminator tl located just downstream of
rpsO (Hajnsdorf et al., 1994b) (Figure 1). The RNase E
cleavage occurring in the dicistronic rpsO-pnp transcript
at the M site just downstream of tl (Figure 1), produces
a mRNA species which cannot be distinguished from Pl-
tl on a Northern blot (Regnier and Hajnsdorf, 1991;
Hajnsdorf et al., 1994b). Another endonucleolytic cleavage
made by RNase III in the dicistronic transcript gives rise
to a PI-RIII rpsO mRNA that is 82 nucleotide longer
(Regnier and Portier,1986; Hajnsdorf et al., 1994b)
(Figure 1). The initial and limiting step in the degradation
of the rpsO message is the endonucleolytic cleavage made
by RNase E at the M2 site, between the rpsO coding
sequence and the hairpin of the transcriptional terminator,
which protects the message from the exonucleolytic attack
(Figure 1) (Regnier and Hajnsdorf, 1991; Hajnsdorf et al.,
1994b). Subsequently, the 3' to 5' exonucleases PNPase
and presumably RNase II, carry out the rapid processive
degradation of the PI-M2 RNase E processed molecule
(Hajnsdorf et al., 1994b; Braun et al., 1996). Two inter-
mediary products of this degradation pathway, fragments
b and c, which differ from P1-M2 by their 3' termini
(Figure 1) are detected when there is no PNPase in the
cell (Braun et al., 1996).
Simultaneous inactivation of the three ribonucleases
implicated in the RNase E-dependent degradation pathway
of the rpsO mRNA (RNase E, PNPase and RNase II)
does not completely abolish the degradation of the rpsO
transcript (Hajnsdorf et al., 1994b, 1995). It has been
demonstrated that a poly(A) tail added post-transcrip-
tionally downstream of tl renders the P1-tl mRNA sus-
ceptible to the attack of unidentified RNases (Hajnsdorf
et al., 1995). Polyadenylation of P1-tI requires the activity
of PAP I (Hajnsdorf et al., 1995). Moreover, characteriz-
ation of a mRNA-poly(A) junction internal to the rpsO
coding sequence suggested that polyadenylation could take
place at multiple sites and contribute to the degradation of
other rpsO mRNAs.
We show here that polyadenylation of rpsO occurs at
many locations including the RNase E, RNase III and
exonuclease cleavage sites. Our data also indicate that
mRNAs processed by RNase E at M2 are polyadenylated
more efficiently than the P1-tl species harbouring the 3'
terminal hairpin, and that polyadenylation contributes to
the degradation of PI-M2 and other mRNA fragments.
The rpsO mRNA is polyadenylated at several sites
in addition to tl
The locations of polyadenylation sites at the 3' ends
of the rpsO mRNAs were determined previously by
amplification and cloning of cDNAs initiated with an
rpsO mRNA polyadenylation in Ecoli
oligo(dT18) primer hybridized to poly(A) tails (Hajnsdorf
et al., 1995). Identification of one clone containing an
mRNA-poly(A) junction at nucleotide 282, in the coding
sequence of rpsO, in addition to 32 clones harbouring
poly(A) tails at tl (Hajnsdorf et al., 1995) suggested that
the rpsO mRNA is polyadenylated at sites other than tl.
In order to test this possibility, we cloned, in bulk, the
PCR fragments amplified from cDNA initiated with an
oligo(dT18) primer. Our experimental procedure did not
include gel purification of the major cDNA band, a step
presumably favouring identification of the polyadenylation
site at tl (Hajnsdorf et al., 1995). We assumed that the
amounts of cDNAs obtained by reverse transcription and
amplification and subsequently, the numbers of the cDNA
P1-ti (420 nt)
PI-RIII (502 nt)
P1-M2 (385 nt)
b (320 nt)
Si nuclease probe
Fig. 1. Structure of rp.sO transcripts. Transcription of the rpsO-pn)p
operon from the promoter P1 gives rise to the primary transcript P1-tl.
as well as to the processed transcripts Pl-RIII. P1-M2. b and c- which
differ from each other by their 3' termini (Regnier and Portier, 1986:
Hajnsdorf et at., 1994b: Braun et al.. 1996). These RNAs are shown
by wavy lines beneath the genetic map and their lengths in nucleotides
(nt) are indicated in parenthesis. Positions of the Rho-independent
transcription terminator tI (420), of the RNase E cleavage sites M2
(385) and M (423). of the proximal RNase III cleavage site RIII (502)
and of the Pstl site (220) used for cloning of amplified cDNAs are
indicated (Regnier and Portier. 1986: Regnier and Hajnsdorf. 1991).
Numbering starts at the first nucleotide of the rpsO transcripts initiated
at the upstream transcription start site (Regnier and Portier. 1986). The
uniformly labelled DraI-BgIl RNA probe used for Northern blots and
the 3'-labelled probe used for SI nuclease mapping are shown beneath
the transcripts. The dot indicates the 3' labelled HpaII site of the
nuclease probe. The primer (PCR primer) used in combination with
an oligo(dT15) primer for cDNA amplification is indicated by an arrow
showing its polarity.
clones containing the different mRNA-poly(A) junctions
would be proportional to the relative abundance of the
rpsO mRNAs polyadenylated at the corresponding sites,
if these mRNAs are at similar concentrations in the same
Strain SK5704(pFB 1) harbours thepnp7 allele encoding
inactive PNPase, and the rnbS00ts and ams 1ts alleles
encoding thermosensitive RNase II and RNase E, respect-
ively (Arraiano et al., 1988). This strain, deficient for
exonucleases (PNPase and RNase II) and RNase E at
440C, is referred to as RNase E- exo-. It overproduces
the rpsO mRNA due to the presence of the pFB 1 plasmid
harbouring rpsO and accumulates the polyadenylated
form of P1-tl at the non-permissive temperature (44°C)
(Hajnsdorf et al., 1995). Total RNA extracted 15 min after
the temperature shift was used as template for reverse
transcription initiated with the oligo(dT18) primer which
contains a BamHI cloning site upstream of 18 T residues.
Theoligo(dT18)primer and a second primer in the rpsO
coding region (PCR primer, Figure
amplify the rpsO cDNAs and the products were cleaved
at the PstI and BamHI sites located in rpsO and the
oligo(dT) primer, respectively and ligated into a cloning
vector. Sequencing of the rpsO-poly(A) junctions of 104
clones allowed us to determine the relative abundance of
mRNAs polyadenylated at 28 sites in rpsO (Table I).
Because the transcription start site is the only 5' terminus
identified upstream of the PstI cloning site, we assume
that each of these mRNAs has its 5' end at this location.
The majority of mRNAs (52%) are polyadenylated at
the site previously characterized just downstream of tl
2). Moreover, poly(A)
between nucleotides 275 and 291 (23% of the 104 clones)
include the site previously characterized at position 282.
They might result from polyadenylation of the RNA
fragment c whose 3' ends map in this region (Braun et al.,
1996). In addition, 26 clones harbour poly(A) tails fused at
15 sites distributed throughout rpsO and the intercistronic
region of the rpsO-pnp mRNA between positions 235
and 542 (Figure 2).
The poly(A) sequence downstream of position 243
identified in one clone (Figure 2) presumably resulted
from annealing of the oligo(dT18) primer to the stretch of
six As encoded by rpsO, and therefore probably does not
correspond to a polyadenylation site (see below).
1), were used to
tails fused to mRNA
Table I. Localization of mRNA-poly(A) junctions in oligo(dT) primed rpsO-pnp cDNAs
Number of clonesa
RNase E- exo-
G235(1): C241(l): C243(1); T275(4); G276(1); C277(2): C279(l); T281(2); G282(4); G283(2);
T285(1); C287(2); C289(4); G291(1): T302(2); G315(2); C343(l); C372(l); T373(l); C382(6);
C389(l); G406(1); T419(1): C420(53); T463(1): G502(5); T519(1); T542(l)
C236(l); C243(2); G250(1): G269(1); T281(1); G283(1); T284(5); T285(l); C289(l); C313(1);
G315(2); T321(2); G325(1); C343(2); G352(3); T365(1); C367(1); G368(1); C372(5); T373(5);
C378(2); T380(l); G381(2): C382(9); G383(32); T386(3): T407(l); C420(3); G466(1); G502(2)
C243(9); C258(l); C274(l): C277(I); T417(l); T419(1); C420(71); G602(15)
C243(7); G399(1): C420(3); G429(1); G502(1): T599(6)
RNase E+ exo-
RNase E- exo+
RNase E+ exo+
aThe total number of clones containing amplified cDNA sequenced to identify mRNA-poly(A) junctions is indicated. The numbers in parenthesis
indicate those which do not contain rpsO sequences. or which contain several rpsO sequences because of multiple inserts.
bmRNA-poly(A) junctions indicate the position of the last nucleotide (G. T or C) identified upstream of the poly(A) sequences in clones of
amplified cDNA. The number of clones harbouring stretches of As fused at the same position is indicated in parenthesis.
J.Haugel-Nielsen, E.Hajnsdorf and RRegnier
Fig. 2. Localization of polyadenylation sites in transcripts of the
rpsO-pnp operon. The number of clones containing the different
mRNA-poly(A) junctions listed in Table I have been plotted as a
function of their positions in the rpsO transcript. The range of
nucleotide positions on the abscissa extends from the PstI cloning site
at nucleotide 220 to the most 3' polyadenylation site identified at
nucleotide 602. The regions indicated (275-291) and (313-325)
correspond to the 3' extremities of the processed c and b rpsO
fragments, respectively (Braun et al., 1996). The M2 RNase E
maturation site is downstream nucleotide 385. Position 420
corresponds to the Rho-independent transcription terminator tI and
position 502 to the proximal RNase III maturation site RIII (Regnier
and Portier, 1986). Positions 243, 599 and 602 which presumably
reflect initiation of oligo(dT) primed reverse transcription at encoded
A-rich regions are also indicated (see Discussion). Locations of c, b,
M2, tl and RIII 3' extremities are indicated on the diagrams together
with the percentage of rpsO cDNA clones harbouring stretches of As
fused at ti, M2 and RIII.
Two new polyadenylation sites located at the proximal
RNase III site (nucleotide 502), and three nucleotides
upstream of the RNase E maturation site M2 (nucleotide
382) were identified in this study, and correspond to 5%
and 6% of all rpsO clones, respectively (Figure 2). This
raises the possibility that the 3' mRNA termini resulting
from endonucleolytic cleavages by RNase III and RNase
E can be polyadenylated.
The mRNA cleaved by RNase E at site M2 is
If polyadenylation detected in the vicinity of M2 in the
RNase E- exo- strain takes place at the 3' end generated
by RNase E, we should obtain more clones containing the
M2-poly(A) junction by using as template RNA from the
SK5726(pFB1I) (pnp7 rnbSOOt5) strain (Arraiano et al.,
1988) (referred to as RNase E+ exo-), which accumulates
the P1-M2 mRNA
(Hajnsdorf et al., 1994b) (Figure 3).
Total RNA was isolated from the RNase E+ exo- strain
15 min after the shift to 44°C, which inactivates RNase
II and induces the accumulation of P1-M2. Then, poly-
adenylation sites were analysed as described above for
the RNase E- exo- strain (Table I). It is striking that 66%
of the 94 rpsO clones characterized contain stretches of
As fused at M2 or a few nucleotides upstream, between
positions 365 and 386 (Figure 2), demonstrating that the
3' end generated by RNase E is a substrate for PAP I and/
or PAP II. These data also imply that the 3' terminus
Fig. 3. Analysis of rpsO transcripts in strains of Ecoli deficient for
RNase E and/or exonucleases. Strains MG1693 (wt; RNase+ exo+),
SK5665 (amns; RNase E- exo+), SK5726 (pnp rnb; RNase E+ exo-)
and SK5704 (ams pnp rnb; RNase E- exo-) transformed with pFB
were grown at 30'C toOD650= 0.25 before being shifted to 44°C.
Five micrograms of RNAs extracted from aliquots withdrawn 15 min
after the shift were analysed on a Northern blot probed with the Dral-
Bgll RNA probe. The positions of the three main rpsO mRNA species
(P1-RIII, Pl-tl and PI-M2) are indicated beside the gels. The lane
containing RNA from the RNase E+ exo- strain was exposed for a
longer time than the other lanes. Relative amounts of these three
mRNAs in each strain are shown in Table II. The band which migrates
slightly more rapidly than PI-M2 in the RNase E+ exo+, RNase E-
exo+ and RNase E- exo- strains has not been considered in our
analysis because the nature of the corresponding mRNA is not known.
resulting from the endonucleolytic cleavage might be
slightly resected before being polyadenylated. To confirm
this hypothesis we carried out Si nuclease protection
analysis which shows that processing at M2 gives rise to
a set of 3' termini spread over at least 20 nucleotides
upstream of M2 (Figure 4). These mRNAs
presumably polyadenylated. SI nuclease mapping does
not detect the A residues added post-transcriptionally at
the 3' ends of mRNAs. The large quantity of the mRNA
3' termini mapping at tl presumably prevented detection
of mRNA-poly(A) junctions corresponding to mRNA 3'
ends mapping in the vicinity of this site (Figure 4).
As in the RNase E- exo- strain, poly(A) sequences were
also fused in the region extending from nucleotides 275
to 291 (10% of the clones), at the RNase III site (2% of
the clones) and at the upstream nucleotide 243 (2% of the
clones) (Table I and Figure 2). Moreover, 16 clones
contained stretches of As fused at 11 sites which were
mostly (10 out of 11) different from those identified in
the RNase E- exo- strain (Table I). Some of them between
nucleotides 313 and 325 might result from polyadenylation
of the b decay intermediate (Braun et al., 1996) (Figure 1).
Strikingly, mRNA polyadenylated at tI represented only
3% of the total polyadenylated rpsO mRNA in the RNase
E+ exo- strain, in constrast to 52% in the RNase E- exo-
strain (Figure 2). We wished to verify that the relative
numbers of clones harbouring poly(A) sequences at tl
and M2 reflected the amounts of cDNAs initiated at the
poly(A) tails of P1-tl and P1-M2 in the RNase E+ exo-
strain. For this purpose, oligo(dT18) primed cDNAs were
amplified using the 5' labelled internal primer as above,
and the radioactive PCR products were analysed in a
polyacrylamide urea gel (Figure 5). Quantification of this
gel using a PhosphorImager showed that P1-M2 gives rise
to ~16 times more amplified cDNA than P1-tl. This is in
agreement with the data in Table I showing that cloning
of amplified cDNAs yielded
~20 times more clones
rpsO mRNA polyadenylation in Ecoli
Fig. 5. Amplification of oligo(dT) primed rpsO cDNA. Aliquots of
RNAs from strains MG1693 (wt; RNase+ exo+), SK5665 (ams;
RNase E- exo+), SK5726 (pnp rnb; RNase E+ exo-) and SK5704
(ams pnp mb; RNase E- exo-) analysed in the Northern blot of
Figure 3 were reverse transcribed with theBH2TI8oligo(dT) primer.
Then, this primer was used in combination with the 5' end labelled
internal PCR primer indicated in Figure 1 for amplification of the
resulting cDNAs. The radioactive products were analysed in an 8%
polyacrylamide-urea gel together with radioactive DNA fragments
whose lengths are indicated on the left of the autoradiograph. DNA
fragments presumably resulting from the amplification of cDNAs
initiated at the polyadenylated 3' end of Pl-tl, Pl-M2 and P1-RIII, or
at encoded A-rich regions located downstream of nucleotides '243'
and '602', are indicated on the right. Amounts of amplified cDNA
fragments corresponding to Pl-M2 and PI-tl in the RNase E+ exo-
strain were quantified with a Phosphorlmager.
Fig. 4. Location of mRNA termini generated upon RNase E cleavages.
A culture of strain SK5003 (pnp rnb; RNase E+ exo-) was grown at
300C to OD650 = 0.25 and shifted to 440C. Twenty micrograms RNA
extracted from an aliquot withdrawn 30 min after the temperature shift
were mixed with the 3-labelled S1 probe shown in Figure 1,
denatured, hybridized, digested with S1 nuclease and analysed in a 8%
polyacrylamide-urea gel (lane RNase E+ exo-). Sequence ladders
generated by chemical cleavages of the 3-labelled probe at G+A and
C+T are shown. Positions of the RIII, tl and M2 3' extremities are
harbouring poly(A) at M2 (between positions 365 and
386) than at tl (positions 419 and 420). This result might
reflect either the large difference of the relative amounts
of the P1-tl mRNA species in the two strains (Figure 3),
or preferential polyadenylation at M2 in the RNase E+
The M2 3' termini are more efficiently
polyadenylated than tl in the RNase E+ exo- strain
To evaluate the efficiencies of polyadenylation at tl, M2
and RIII in the RNase E- exo- and the RNase E+ exo-
strains, we compared the numbers of clones originating
from mRNAs polyadenylated at these sites to the amounts
of the P1-tl, P1-M2 and PI-RIII mRNA species detected
on a Northern blot (Table II). Because the Northern blot
shown in Figure 3 allows quantification of only these
three major mRNA species, we only took into account
those clones harbouring poly(A) tails at tl, M2 or RIII in
the calculations shown in Table II. The large amount of
P1-tl mRNA in the RNase E- exo- strain (88% of the
three major mRNA species) suggests that the prevalence
of mRNA polyadenylated at tl in this strain [83% of
poly(A) fused at tl, M2 and RIII] simply reflects the
relative abundance of this mRNA species (Table II).
Accordingly, mRNA of this strain mostly gives rise to
an amplified cDNA fragment corresponding to P1-tl
In contrast, the strong discrepancy observed in the
RNase E+ exo- strain between the relative amounts of
mRNAs polyadenylated at M2 and tl (93% and 4% of
the clones, respectively) and the relative abundances of
the P1-M2 and P1-tl mRNAs (58% and 28% respectively,
Table II), suggest that P1-M2 transcripts contain a much
higher proportion (at least 10 times) of polyadenylated
molecules than the P1-tl mRNAs (Table II).
The isolation of clones polyadenylated at M2 in the
RNase E- exo- strain (Figure 2), in which P1-M2 cannot
be detected (Figure 3), might result from the high yield
of PCR amplification of molecules present at very low
concentration in the RNA preparation. Alternatively, this
might also indicate that M2 is efficiently polyadenylated.
The P1-M2 transcript is stabilized in a strain
deficient for PAP I
Because polyadenylation was reported
stability of several transcripts including the rpsO P1-tl
mRNA (Xu et al., 1993; Hajnsdorf et al., 1995; O'Hara
to reduce the
J.Haugel-Nielsen, E.Hajnsdorf and P.Regnier
Table II. Polyadenylation of P1-ti, PI-RIII and PI-M2 mRNA species
RNase E- exo-
RNase E+ exo-
a[mRNA(%)] is the ratio of the amount of one of the three mRNA
species (P1-tl, Pl-M2 or PI-RIII) to the total amount of rpsO mRNA
present in the bands of P1-ti+Pl-M2+Pl-RIII mRNAs quantified
with a Phosphorlmager from the Northern blot in Figure 3.
b[poly(A) mRNA(%)] represents the proportion of P1-tl, P1-M2 or
PI-RIII in the population of polyadenylated rpsO mRNAs. It is the
ratio of the number of clones containing stretches of As fused at tI
(downstream positions 419 and 420), M2 (between positions 365 and
386) or RIII (downstream position 502) to the total number of clones
corresponding to mRNAs polyadenylated at tl, M2 and RIII (Table I).
et al., 1995), we investigated whether the P1-M2 mRNA
detected in both IBPC674 (RNase E- exo- PAP I') and
IBPC673 (RNase E- exo- PAP I-) (Hajnsdorf et al., 1995)
at permissive temperature is more stable in the PAP- strain
than in the isogenic PAP+ bacteria. These two strains
harbour the rnb5OOts and amsltl mutations. The Northern
blot shown in Figure 6 demonstrates that P1-M2
detectable for at least 20 min after inhibition of transcrip-
tion by rifampicin in the RNase E- exo- PAP I- cells
shifted to 44°C at the time of antibiotic addition, to
inactivate RNase E and RNase II. In contrast this RNA
species is no longer detectable 8 min after addition of the
antibiotic to the RNase E- exo- PAP I' cells. These data
suggest that P1-M2 is destabilized upon polyadenylation
by PAP I. The observation that the RNA fragment c and,
to a smaller extent, the fragment b also appear to be
stabilized in the absence of PAP I, is consistent with the
finding that they are both polyadenylated (Figure 2).
Interestingly, this experiment also shows that the decay
rate of P1-RIII is not altered in a PAP I deficient strain
(Figure 6), indicating that polyadenylation at its 3' end
generated by RNase III does not modify its stability.
Most poly(A) tails of rpsO mRNAs are detected at
ti when cells contain exonucleases
All of the mRNA-poly(A) junctions described above were
determined using RNA from exonuclease deficient strains.
These strains contain a higher fraction of polyadenylated
mRNAs than strains containing exonucleases, presumably
because RNase II and PNPase compete against the elonga-
tion of poly(A) tails (Cao and Sarkar, 1992a; Hajnsdorf
et al., 1995). We therefore analysed the mRNA poly(A)
junctions in strains containing PNPase and RNase II, in
order to investigate whether exonucleases have an effect
on the abundance of different poly(A) mRNA species.
SK5665(pFB1) (amslts) (Arraiano et al., 1988) (referred
to as RNase E+ exo+ and RNase E- exo+, respectively)
were grown for 15 min at 44°C before RNA preparation,
as described above.
Nearly all the clones of amplified cDNAs corresponding
to polyadenylated rpsO mRNA isolated from the RNase
_i ' r
:. ::' .. :_ . ....
Fig. 6. PAP I destabilizes the processed P1-M2 rpsO transcript.
(A) Strains IBPC674 (SK5704 pcnB+; RNase E- exo- PAP I +) and
IBPC673 (SK5704 pcnB; RNase E- exo- PAP I -) were grown at
30°C until the OD6st reached 0.25. Then, rifampicin was added to a
final concentration of 500 ,ug/ml and the culture was shifted to 44°C.
RNAs extracted at the times indicated in min above each lane were
analysed on a Northern blot probed with the DraI-BgIl RNA probe.
The positions of the Pl-RIII, P1-tl, PI-M2, b and c transcripts are
shown beside the autoradiograph. The Pl-M2 band detected in this
experiment is not visible in the RNase E- exo- PAP I+ strain 15 and
30 min after the temperature shift (Figure 3, right lane). (B and C)
Northern blots of (A) were quantified with a Phosphorlmager and the
relative amounts of Pl-RIII (B) (Z1, *), PI-tl (C) (0, 0) and P1-M2
(C) (A, A) remaining in the pcnB+ (filled symbols) and pcnB (open
symbols) strains were plotted as a function of time. Radioactivity in
bands b and c was too low to be quantified.
E+ exo+ and RNase E- exo+ strains had poly(A) tails
fused at tl (Table
polyadenylated at the proximal RTIII site, and two clones
corresponding to polyadenylation at new sites located
in the hairpin of the transcription terminator and ~10
nucleotides downstream tl, were also found in the wild-
type strain (Figure 2). Moreover, three clones isolated
from the RNase E- strain harboured poly(A) tails in the
vicinity of the 3' end of the c fragment.
A large part of the clones obtained in these experiments
with mRNAs of the RNase E- exo+ and RNase E+ exo+
cells (24% and 68%, respectively) contain poly(A) tails
fused downstream of nucleotides 243, 599 or 602 (Figure
2), presumably due to the presence in the message of an
encoded A-rich sequence able to direct annealing of the
oligo(dT18) primer (Figure 7). In fact, it was possible to
amplify an oligo(dT18) primed cDNA fragment beginning
at the stretch of six As located downstream nucleotide
I and Figure 2). Only one clone
rpsO mRNA polyadenylation in E.coli
Fig. 7. Oligo(dT) primed reverse transcription may be initiated at
encoded A-rich sequences. The encoded A residues of rpsO and groEL
which presumably hybridize with theoligo(dT18)primer to initiate
reverse transcription are underlined. The arrows starting from the
junctions between the stretches of As and the sequences of rpsO and
groEL (upstream) identified in several cDNA clones indicate the
putative initiation sites of cDNA synthesis primed by the oligo(dT18)
primers. The numbers in quotation marks indicate the positions of the
last encoded rpsO nucleotide identified upstream of the poly(A)
sequence. Nineteen clones containing the groEL-poly(A) sequences
were obtained by reverse transcription and amplification of RNase E+
exo+ RNA. Preferential cloning of this DNA fragment is presumably
due to the presence upstream of a PstI site, of a sequence able to
anneal 13 of the 20 nucleotides of the internal PCR primer at 36 and
45°C. These temperatures were found to optimize amplification of
rpsO DNA fragments. The high yield of PCR fragments obtained
under these conditions probably results from a better annealing of the
oligo(dT,g)primer to poly(A) tails. However, the internal primer
whose melting temperature is 62°C probably forms non-specific
hybrids at these temperatures.
243 using an in vitro synthesized rpsO transcript as a
template (data not shown). In addition, it was interesting
to point out that we obtained many more of these clones
resulting from reverse transcription primed with oligo-
(dT18) at nucleotides 243, 599 or 602 with RNA from
exonuclease-replete strains which contain a low fraction
of polyadenylated mRNAs, than from RNA of the exo-
nuclease deficient strains which exhibit long and/or abund-
ant poly(A) tails. These data, which imply that these
clones do not reflect polyadenylation of mRNA, confirm
that they result from reverse transcription priming at
internal A-rich sequences. This conclusion was re-inforced
by the observation that numerous clones lacking rpsO
sequences, obtained from RNase E+ exo+ RNAs (Table
I), contained fragments of the groEL gene, which were
also reverse transcribed from an encoded A-rich region
A striking discrepancy between the RNase E- exo+ and
RNase E+ exo+ strains is the proportion of clones which
contains rpsO cDNA (Table II). The large difference in
P1-tl rpsO mRNA concentrations in the two strains
(Figure 3) presumably explains why we obtained a much
higher proportion of rpsO clones from the RNase E- exo+
strain (100 out of 110 clones analysed) than from the
RNase E+ exo+ strain (19 out of 97 clones analysed)
(Table I). However, our data indicate that this discrepancy
might also result from a reduction of the proportion of
molecules polyadenylated at tl in the RNase E+ strain.
This proportion can be estimated in each strain by compar-
ing the number of clones containing a tl poly(A) junction,
which reflects the amount of P1-tl mRNA harbouring a
poly(A) tail at tl, with the number of clones containing a
mRNA-poly(A) junction at nucleotide 243, which reflects
the total amount of the P1-tl mRNA able to anneal the
primer at the internal stretch of six As. Table III shows
that the ratio of clones corresponding to polyadenylated
P1-tl to the number of clones corresponding to total P1-
RNase E+ strain, suggesting that the tl termini harbour
is much higher in the RNase E- strain than in the
Table III. Internal versus terminal initiation of reverse transcription of
the P1-ti rpsO mRNA in the RNase E+ and RNase E- strainsa
RNase E+ exo+
RNase E- exo+
aThe number of clones harbouring stretches of As at the terminator tl
and downstream nucleotide 243 in each strain were taken from
less poly(A) tails in the presence of active RNase E. This
negative effect of RNase E might also in part explain
why P1-tl contains a lower proportion of polyadenylated
molecules than P1-M2 in the RNase E+ exo- strain
This paper presents the first evidence that polyadenylation
can occur at processing sites in E.coli mRNA. We demon-
strated that polyadenylation occurs at the M2 RNase E
cleavage site upstream of the hairpin transcription termin-
ator of rpsO if this message is processed by RNase E
in vivo, while nearly no poly(A) tails are detected at this
site in an RNase E deficient strain. This suggests that
poly(A) tracts fused upstream of transcription termination
sites of Ipp mRNA (Cao and Sarkar, 1992a), colEl RNA
I of E.coli (Xu et al., 1993) and hag mRNA of Bacillus
subtilis (Cao and Sarkar, 1993), reflect polyadenylation
at processing sites. However, our data, showing that a
substantial number of mRNA-poly(A) clones can result
from reverse transcription initiated at encoded A-rich
regions, indicate that these Ipp and hag polyadenylated
mRNAs might also result from hybridization of the
oligo(dT) primer to the stretches of As encoded in these
genesjust upstream oftheir respective transcription termin-
ators. Poly(A) tails were also detected at the position of
the RNase III cleavage downstream of rpsO, in agreement
with previous data suggesting that post-transcriptional
addition of A residues occurs at an RNase III processing
site of the phage T7 early transcript (Rosenberg et al.,
1974). The 3' ends of the decay intermediates (fragments
b and c), generated exonucleolytically in the absence of
PNPase (Braun et al., 1996), also appear to be polyadenyl-
ated. The scattering of polyadenylation sites over 17
nucleotides in the vicinity of the c fragment 3' end
(between nucleotides 275 and 291) is consistent with the
heterogeneity of the 3' ends of this exonucleolytically
generated fragment (Braun et al., 1996) and with the
persistence of residual exonucleolytic activity of RNase
II and other RNases in the exonuclease deficient strains
at 44°C (Donovan and Kushner, 1986; Hajnsdorf et al.,
1994b; O'Hara et al., 1995). Probably, this residual activity
also accounts for the heterogeneity of the me-dependent
mRNA-poly(A) junctions mapping between two and 20
nucleotides upstream of M2.
As observed previously for the P1-tI mRNA (Hajnsdorf
et al., 1995), we show here that P1-M2 is more stable in
a PAP I deficient strain also devoid of PNPase, RNase II
and RNase E. These data indicate that PAP I participates in
the polyadenylation ofP1-M2, and that the polyadenylated
species can be degraded by unidentified exo and/or endo-
J.Haugel-Nielsen, E.Hajnsdorf and P.Regnier
gives rise to a smear ofheterogeneous extended molecules
not detectable in Northern blots of RNase E- exo- PAP
l' RNA (Figure 6). Polyadenylation appears also to be
involved in the degradation of the b and c small decay
intermediates. It is interesting to point out that lack of
polyadenylation in the RNase E- exo- PAP I- (Figure 6)
and RNase E- exo+ PAP I- strains (data not shown) has
no effect on the stability of PI-RIII. This might reflect
the fact that PI-RIII is polyadenylated by another enzyme
(e.g. PAP II), or that PI-RIII is degraded independently
of its poly(A) tail even in the absence of PNPase, RNase
II and RNase E. This also suggests that sequences just
upstream of the polyadenylation site can modulate the
destabilizing effect of poly(A) tails.
We deduced from the proportions of amplified cDNAs
harbouring mRNA-poly(A) junctions at tl and M2 that
P1-tl contains a significantly lower proportion of poly-
adenylated molecules than P1-M2 in RNase E+ exo-
strains, and that P1-tl seems to be less efficiently poly-
adenylated in the presence of RNase E. A possible
explanation is that the binding of PAP I to P1-tl mRNA
is impeded by RNase E bound at the M2 site (Hajnsdorf
et al., 1994b), or associated in a complex with the
Exonuclease Impeding Factor bound at the tl hairpin
(Causton et al., 1994). Alternatively, the lower proportion
ofpolyadenylated P1-tI might reflect the greater instability
of these molecules in the RNase E+ strain. Moreover, it
is also possible that the preferential amplification of M2
poly(A) junctions compared with those at tl in the RNase
E+ strain is due to the presence of longer poly(A) tails
which exhibit a higher affinity for the primer. Similarly,
the small number of rpsO clones polyadenylated at tl
obtained with the RNA of the RNase E+ exo+ strain
might reflect the fact that poly(A) tails are shorter in the
presence of RNase E.
Finally, we wish to stress the limits of the technique
employed and the significance of the data obtained.
Comparisons between Northern blots and polyadenylation
patterns detect a reliable difference of polyadenylation
frequency between P1-tl and P1-M2 in the exonuclease
deficient strains (Table II), which mainly yields clones
containing rpsO cDNA. Probably, one has to be cautious
in assigning polyadenylation frequencies from the few
rpsO clones obtained with RNA ofRNase E' exo+ strains,
which yields numerous artefacts. Attention must also be
drawn to the fact that our method allows us to compare
proportions of polyadenylated molecules for each mRNA
species, but does not give any indication as to the
abundance of polyadenylated molecules. In fact, Northern
blots fail to distinguish between tail-less and (A)tail-
containing mRNAs, and patterns of mRNA-poly(A) junc-
tions do not relate to their concentrations.
In spite of the amino acid sequence homology with
terminal transferase (Masters et al., 1990) suggesting that
PAP I might have a preferential affinity for stably folded
3' termini such as transcription termination hairpins,
polyadenylation at M2 and RIII demonstrates that PAP I
also recognizes unstructured termini and therefore, like
mammalian poly(A) polymerases, presumably does not
exhibit specificity for particular mRNA motifs (Keller,
1995). The latter conclusion is very strongly supported
by the observation that polyadenylation can occur at 45
Polyadenylation of P1-M2 presumably
of the 185 nucleotides of rpsO mRNA extending between
positions 235 and 420 (tl) (Table I).
The fact that many polyadenylation sites are spread
throughout rpsO in cells deficient for exonucleases and
RNase E indicates that the rpsO mRNA is cleaved at
many sites despite the inactivation of PNPase, RNase II
and RNase E.
It is striking that, with the exception of tl, most
polyadenylated mRNA species, including M2, could be
identified in the exonuclease deficient strains but not
in strains containing PNPase and RNase II. An initial
explanation might be that these termini do not become
accessible to PAP I due to the processivity of 3' to 5'
exonucleases and because PNPase is associated with
RNase E (Carpousis et al., 1994; Py et al., 1994) and
might therefore be delivered to the M2 3' end immediately
after the processing step. A second possibility is that once
polyadenylated these mRNAs are degraded so rapidly by
exonucleases that they cannot be detected by the method
we have used. Accordingly, poly(A) tails were proposed
to improve the affinity of PNPase for RNA I of ColE1
plasmids, and to facilitate processive degradation through
secondary structures (Xu and Cohen, 1995). Moreover,
the pre-eminence of RNase II which represents 90% of
the poly(A) degrading activity in crude extracts (Deutscher
and Reuven, 1991), the higher velocity ofpoly(A) degrada-
tion by this nuclease compared with other homopolymers
(Singer and Tolbert, 1965) and the absence of detectable
elongated (polyadenylated) rpsO mRNAs in a PNPase
deficient strain containing RNase II (Hajnsdorf et al.,
1994b) strongly suggest that RNase II could contribute to
the degradation ofpoly(A) tails and carry out the complete
degradation of the polyadenylated mRNA fragments in
the absence ofimpeding secondary structures (Gupta et al.,
1977; McLaren et al., 1991).
In contrast, detection of mRNAs polyadenylated at tl
in strains containing exonucleases suggests that the stable
hairpin of the terminator offers a barrier to nucleases even
when the mRNA is polyadenylated. It has been proposed
that RNase II stalls at tl (Hajnsdorf et al., 1994b; Pepe
et al., 1994). It can also be hypothesized that RNase II
removes oligo(A) tails and dissociates when it reaches tl
(Coburn and Mackie, 1996). The mRNA molecule could
undergo several rounds of polyadenylation and de-adenyl-
ation by PAP I and RNase II before being completely
degraded by PNPase or other ribonucleases. Such a de-
adenylating function of RNase II might also explain why
mRNAs terminated by a hairpin are destabilized upon
RNase II inactivation (Hajnsdorf et al., 1994b; Pepe et al.,
1994). This dynamic view of the structure of mRNA 3'
end is reminiscent of the exchanges of the 3' terminal
adenosine nucleotide of tRNAs catalysed by ribonuclease
T and nucleotidyl transferase (Deutscher, 1990).
We have demonstrated here that bacterial mRNAs are
polyadenylated at 3' termini generated by RNase E, RNase
III and exonuclease cleavages. In this respect, bacterial
poly(A) polymerase(s) are similar to eukaryotic enzymes
which polyadenylate mRNA at processing sites (Keller,
1995). A question arising is whether Ecoli PAPs have
affinity for processing complexes as in mammalian cells
(Keller, 1995), or for other features of mRNA sequence
or structure. On the other hand, there are now many
indications that mRNA fragments such as the P1-M2, P1-
rpsOmRNApolyadenylation in Ecoli
RIII, b and cfragmentsof therpsOmRNA and the RNase
Eprocessed RNA I of ColE1 plasmids (Xu etal., 1993)
arepolyadenylated.The destabilizationresultingfrom this
polyadenylation indicates thatpoly(A) tails play a major
role in the rapid elimination ofintermediary products of
mRNAdecayin addition to its role in the degradation of
Materials and methods
Isogenic ribonuclease deficient strains SK5704 (ainslts pnp7 rnb5OOts),
SK5726[pnp7 rnb500tspDK39(CmR rnb500's)] and SK5665 (amslts)
(Arraiano et al., 1988) were transformed withplasmid pFBI (Hajnsdorf
et al., 1995) and grown in LB medium supplemented with thymine
(40,g/mI) andampicillin (100 gg/ml).The isogenic wild-type strain
MG1693 transformed withpFBI wasgrown in the same medium, and
strain SK5003 [pnp7 rnb5OOst pDK39(CmR rnb5OOts)] (Donovan and
Kushner, 1986) was grown in LB mediumsupplemented with 20 ,ug/ml
chloramphenicol. IBPC673 (SK5704 pcnB7 pRS415) and IBPC 674
(SK5704 pcnB+ pRS415) (Hajnsdorf et al., 1995) were grown
LB medium supplemented with thymine (40 ,ug/ml) and ampicillin
Total RNA was prepared as described previously (Hajnsdorf et al.,
1994b) from 10 mlaliquotsof culturesgrownas mentioned in thefigure
legends. RNA preparations were analysed by Northern blotting as
described (Hajnsdorfet al., 1994b) and probed with antisense rpsO
RNAsynthesized by int vitro transcription from the pEHa4 plasmid
containingtherpsO DraI-BgIl fragment, extending from the translation
initiation codon to thetranscriptionterminatort I ofrpsO (Hajnsdorfetal.,
1994a) (Figure 1). Northern blots werequantifiedwith aPhosphorlmager
(Hajnsdorfet al., 1994b) with the 3-labelled probe shown in Figure 1.
cDNAs weresynthesizedfrom the BH2T1 primer aspreviously described
(Hajnsdorfetal., 1995) exceptthat reversetranscription was carried out
in 75 mM Tris-HCI pH 8.3, 15 mMMgCl2, 4 mM DTT, 112 mM KCI
25% of thesynthesized cDNA (5 gl)was amplified with the BH,T18
primerand the internal PCR primer (Figure 1) as described previously
(Hajnsdorfetal., 1995). The PCRfragmentswere cleaved with PstI and
BamHI and cloned, in bulk, into the pT3T718U vector. Single-stranded
DNAsproducedin JMIOITR weresequenced.
Pl-tl RNA was transcribed in vitroby SP6 RNA Polymerase under
conditions described by
generated template extendingfrom nucleotides I to 420 ofrpsO amplified
from thepB15.6 plasmid.Primers used for theamplificationwere 5'-ATT-
AAGGTGACACTATAGCCGCTTAACGTCGCGTAAATTG-3' and 5'-
GAAAAAAGGGGCCACTCAGG-3', the underlined sequence corres-
pondingto thesequenceof the SP6 promoter.
1 mM of each of the four dNTPs.
the manufacturer (Promega) from
We thankM.Grunberg-ManagoandM.Springer forencouragement and
discussions, G.Mackie for communication of resultspriortopublication,
and D.Stern andM.Dreyfusfor criticalreadingof the manuscript. This
work was funded by CNRS (URA1139), University Paris 7 and the
European Union (Human Capital and Mobility ERB CHRX CT 93-
0263).J. H.-N. issupported byagrantfrom Forskerakademiet(Denmark).
Arraiano,M.C., Yancey,S.D. and Kushner,S.R. (1988) Stabilization of
discrete mRNA breakdownproductsinamns pnp rnb multiple mutant
of Escherichia coli K 12. J. Bacteriol., 170, 4625-4633.
Baker,E.H. (1993) In Belasco,J.G. and Brawerman,G. (eds), Control of
mRNA Stability. Control of poly(A) length. Academic Press Inc., New
York, pp. 367-415.
phosphorylase is required for the rapid degradation of the RNase E
processed rpsO mRNA of Escherichia coli devoid of its 3' hairpin.
Mol. Microbiol., 19, 997-1005.
Brawerman,G. (1981) The role of the poly(A) sequence in mammalian
messenger RNA. CRC Crit. Rev. Biochem., 10, 1-38.
Cao,G.-J. and Sarkar,N. (1992a) Poly(A) RNA in Escherichia coli:
Nucleotide sequence at the junction of the /pp transcript and the
polyadenylate moiety. Proc. Natl Acad. Sci. USA, 89, 7546-7550.
Cao,G.-J. and Sarkar,N. (1992b) Identification of the gene for an
Escherichia coli poly(A) polymerase. Proc. Nati Acad. Sci. USA, 89,
Cao,G.-J. and Sarkar,N. (1993) Poly(A) RNA
Identification of the polyadenylation site of flagellin mRNA. FEMS
Microbiol. Lett., 108, 281-286.
Carpousis,A.G., Van Houwe,G., Ehretsmann,C. and Krisch,H.M. (1994)
Co-purification of E. coli RNase E and PNPase: Evidence for aspecific
association between two enzymes important in RNA processing and
degradation. Cell, 76, 889-900.
Py,E., McLaren,R.S. and Higgins,C.F. (1994) mRNA
degradation in Escherichia coli: a novel factor which impedes the
exoribonucleolytic activity of PNPase at stem-loop structures. Mol.
Microbiol., 14, 731-741.
Coburn,G.A. and Mackie,G.A. (1996) Overexpression, purification and
properties of Escherichia coli ribonuclease II. J. Biol. Chem., 271,
Cohen,S.N. (1995) Surprises at the 3' end of prokaryotic RNA. Cell,
Deutscher,M. (1990) Ribonucleases, tRNA nucleotidyltransferase and
the 3' processing of tRNA. Prog. Nucleic Acid Res. Mol. Biol., 39,
Deutscher,M.P. and Reuven,N.B. (1991) Enzymatic basis for hydrolytic
versus phosphorolytic mRNA degradation in Escherichia coli and
Bacillus subtilis. Proc. Natl Acad. Sci USA, 88, 3277-3280.
Donovan,W.P. and Kushner,S.R. (1986) Polynucleotide phosphorylase
and ribonuclease II are required for cell viability and mRNA turnover
in Escherichia coli K- 12. Proc. Natl Acad. Sci. USA, 83, 120-124.
degradation in prokaryotes. FASEB J., 6, 3186-3192.
Gupta,R.S., Kasai,T. and Schlessinger,D. (1977) Purification and some
novel properties of Escherichia coli RNase II. J. Biol. Chem., 252,
inactivation and degradation of the RNase III processed pnp message
encoding polynucleotide phosphorylase of Escherichia coli. J. Mol.
Biol., 239, 439454.
Hajnsdorf,E., Steier,O., Coscoy,L., Teysset,L. and Regnier,P. (1994b)
Roles of RNase E, RNase II and PNPase in the degradation of the
rpsO transcript of Escherichia coli: Stabilizing function and evidence
for efficient degradation in an ams-pnp-rnb mutant. EMBO J., 13,
Polyadenylylation destabilizes the rpsO mRNA of Escherichia coli.
Proc. Natl Acad. Sci. USA, 92, 3973-3977.
Identification of a second poly(A) polymerase in Escherichia coli.
Biochem. Biophvs. Res. Commun., 198, 459-465.
Karnik,P., Taijanidisz,J., Sasvari-Szekely,M. and Sarkar,N. (1987) 3'-
terminal polyadenylate sequence of Escherichia coli tryptophane
synthetase a-subunit messenger RNA. J. Mol. Biol., 196, 347-354.
Keller,W. (1995) No end yet to messenger RNA 3' processing. Cell, 81,
Littauer,U.Z. and Sorek,H. (1982), The regulatory function of poly(A)
and adjacent 3' sequences in translated RNA. Prog. Nucleic Acid Res.
Mol. Biol., 27, 53-83.
Liu,J. and Parkinson,J.S. (1989) Genetics and sequence analysis of the
pcnB locus, an Escherichia coli gene involved in plasmid copy number
control. J. Bacteriol., 171, 1254-1261.
March,J.B., Colloms,M.D., Hart-Davis,D., Oliver,I.R. and Masters,M.
(1989) Cloning and characterization of an Escherichia coli gene,
pcnB, affecting plasmid copy number. Mol. Microbiol., 3, 903-910.
Masters,M., March,J.B., Oliver,I.R. and Collins,J.F. (1990) A possible
role for the pcnB gene product of Escherichia coli in modulating
RNA:RNA interactions. Mol. Gen. Genet., 220, 341-344.
in Bacillus subtilis.
J.Haugel-Nielsen, E.Hajnsdorf and RRegnier Download full-text
McLaren,R.S., Newbury,S.F., Dance,G.S.C., Causton,H.C. and Higgins,
C.F. (1991) mRNA degradation by processive 3-5' exonucleases
in vitro and the implication for prokaryotic mRNA decay in vivo.
J. Mol. Biol., 221, 81-95.
O'Hara,E.B., Chekanova,J.A., Ingle,C.A., Kushner,Z.R., Peters,E. and
Kushner,S.R. (1995) Polyadenylylation helps regulate mRNA decay
in Escherichia coli. Proc. Natl Acad. Sci. USA, 92, 1807-1811.
Pepe,C.M., Maslesa-Galic,S. and Simons,R.W. (1994) Decay of the ISIO
antisense RNA by 3' exonucleases: evidence that RNase II stabilizes
RNA-OUT against PNPase attack. Mol. Microbiol., 13, 1133-1142.
Py,B., Causton,H., Mudd,E.A. and Higgins,C.F. (1994) A protein complex
mediating mRNA degradation in Escherichia coli. Mol. Microbiol.,
Regnier,P. and Hajnsdorf,E. (1991) Decay of mRNA encoding ribosomal
protein S15 of Escherichia coli is initiated by an RNase E-dependent
endonucleolytic cleavage that removes the 3' stabilizing stem and
loop structure. J. Mol. Biol., 217, 283-292.
Regnier,P. and Portier,C. (1986) Initiation, attenuation and RNase III
processing of transcripts from the E. coli operon encoding ribosomal
protein S15 and polynucleotide phosphorylase. J. Mol. Biol., 187,
sequence of the pnp gene of E.
phosphorylase; homology of the primary structure of the protein with
the RNA binding domain of ribosomal protein S1. J. Biol. Chem.,
Rosenberg,M., Kramer,R.A. and Steitz,J.A. (1974) T7 early messenger
RNAs are the direct products of ribonuclease III cleavage. J. Mol.
Biol., 89, 777-782.
Sarkar,N., Langley,D. and Paulus,H. (1978) Isolation and characterization
of polyadenylate-containing RNA from Bacillus brevis. Biochemistry,
Singer,M.F. and Tolbert,G. (1965) Purification and properties of a
potassium-activated phosphodiesterase (RNase II) from Escherichia
coli. Biochemistry, 4, 1319-1330.
Xu,F. and Cohen,S. (1995) RNA degradation in Escherichia coli regulated
by 3' adenylation and 5' phosphorylation. Nature, 374, 180-183.
Xu,F., Lin-Chao,S. and Cohen,S.N. (1993) The Escherichia coli pcnB
gene promotes adenylation of antisense RNAI of ColEl type plasmids
in vivo and degradation of RNAI decay intermediate. Proc. Natl Acad.
Sci. USA, 90, 6756-6760.
coli encoding polynucleotide
Received on Januarv 30, 1996; revised on March 21, 1996