Cx32 mRNA in rat liver: effects of inflammation on
poly(A) tail distribution and mRNA degradation
NICHOLAS G. THEODORAKIS1 AND ANTONIO DE MAIO1,2
of Pediatric Surgery and 2Department of Physiology,
The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205
1Division
gap junctions; cellular communication; connexins; endotoxin;
hepatocytes; mRNA stability; gene regulation
GAP JUNCTIONS ARE PROTEIN channels between adjacent
cells through which ions and other small molecules (⬍1
kDa) can pass (4, 5). These channels are composed of
two hemichannels opposed to each other on adjacent
cells. These hemichannels are termed connexons, each
is composed of a hexameric arrangement of six identical protein subunits, named connexins (Cx). In some
cells, such as hepatocytes, gap junctions are found
aggregated in large latticelike arrays containing hundreds of gap junctions (4). The expression of Cx genes is
tissue specific (16). For example, Cx32 is the most
abundant Cx expressed in normal liver (19). The liver
also contains a second gap junction protein, Cx26. In
rat liver, Cx26 is expressed at a lower level than Cx32;
moreover, whereas Cx32 has a relatively homogeneous
distribution throughout the liver, Cx26 is found predominantly in the periportal region (12).
The costs of publication of this article were defrayed in part by the
payment of page charges. The article must therefore be hereby
marked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734
solely to indicate this fact.
Gap junctional intercellular communication is important for proper organ function. For example, in the
heart, electrical conductance is propagated through
gap junctions (18) and, in the liver, they are believed to
play an important role in hepatic metabolism (13).
Alterations in hepatic gap junction communication and
expression have been observed after different pathophysiological conditions. Thus expression of Cx32 is
reduced in the liver during acute inflammatory states
induced by the administration of bacterial lipopolysaccharide (LPS), a bacteria wall component (11). A similar
decrease in Cx32 expression was observed after a
regional hepatic ischemia-reperfusion injury (9), partial hepatectomy (15, 23), and hepatic cholestasis (23).
The reduction of Cx32 expression in the liver during
LPS-induced inflammation is not caused by a reduction
in Cx32 gene transcription, but rather is due to increased degradation of the Cx32 mRNA (11). To further
understand the mechanism of Cx32 mRNA degradation
after administration of LPS, the shortening of the Cx32
mRNA poly(A) tail was examined, because this is
considered the first step in the degradation of many
eukaryotic mRNAs.
METHODS
Animal preparation. Adult male Sprague-Dawley rats were
fasted overnight and allowed water ad libitum before treatment with intravenous injection of LPS (1 mg/kg) or intraperitoneal injection of actinomycin D (2 mg/kg) or cycloheximide
(6 mg/kg) as described previously (11). This dose of LPS does
not result in mortality, but induces a rapid acute inflammatory response (24). At the indicated times, the rats were killed
by pentobarbital injection and liver samples were removed.
Probes. A rat Cx32 cDNA was obtained from David Paul
(Harvard University). This clone encodes a nearly full-length
rat Cx32 cDNA cloned into the EcoR I site of pGEM-3, in an
orientation such that transcription with T7 polymerase yielded
an antisense RNA probe. To prepare a template for RNase
protection, a Kpn I fragment was removed to construct a
plasmid that contains the final 260 nucleotides of the rat
Cx32 cDNA.
Transcription and RNA analysis. Rat liver nuclei were
isolated, and transcription in isolated nuclei was performed
as described (10). Total RNA was prepared by guanide
thiocyanate-acid phenol extraction (6) and analyzed by Northern blot as described (11). Cx32 mRNA levels were also
analyzed by an RNA protection assay. RNA (5–10 µg) was
incubated in 30 µl of hybridization buffer (80% formamide,
0.4 M NaCl, 40 mM PIPES, 20 mM EDTA) at 42°C for 16 h
with a radiolabeled antisense RNA probe that corresponds to
the final 260 nucleotides of the rat Cx32 gene (see Fig. 1). Ten
volumes of RNase digestion buffer (10 mM Tris, pH 7.5, 0.3 M
NaCl, 5 mM EDTA, 40 µg/ml RNase A) were added, and the
mixture was incubated at 25°C for 60 min. The RNase
digestion was terminated by the addition of 20 µl of 20% SDS
0363-6119/99 $5.00 Copyright r 1999 the American Physiological Society
R1249
Downloaded from http://ajpregu.physiology.org/ by 10.220.33.1 on June 18, 2017
Theodorakis, Nicholas G., and Antonio De Maio. Cx32
mRNA in rat liver: effects of inflammation on poly(A) tail
distribution and mRNA degradation. Am. J. Physiol. 276
(Regulatory Integrative Comp. Physiol. 45): R1249–R1257,
1999.—Previous studies showed that the expression of connexin 32 (Cx32), the polypeptide subunit component of the
major hepatic gap junction, is reduced in liver by changes in
mRNA stability during bacterial lipopolysaccharide (LPS)induced inflammation. In this study, we examined the distribution of Cx32 mRNA poly(A) tail lengths during LPSinduced inflammation, because this is considered the first
step in the degradation of many mRNAs. During LPS treatment the first detectable change in Cx32 mRNA was a
gradual shortening of its poly(A) tail, which reached a final
size of ⬃20 nucleotides. However, the poly(A) tail did not
disappear entirely before the bulk of Cx32 mRNA was degraded. Treatment with actinomycin D, which blocks the
degradation of Cx32 mRNA after LPS administration, resulted in the appearance of a completely deadenylated mRNA,
which otherwise could not be detected. On the contrary,
treatment with cycloheximide resulted in a decrease in the
stability of Cx32 mRNA without an apparent change of the
poly(A) tail size. The effect of cycloheximide on Cx32 mRNA
stability seems to be due indirectly to the induction of an
inflammatory response by this drug. These results suggest
that, similar for many mRNAs, shortening of the poly(A) tail
is one of the first steps in the degradation of Cx32 mRNA
during inflammation.
R1250
CX32
MRNA
POLY(A) TAIL DEGRADATION
RESULTS
and 10 µl of 10 mg/ml proteinase K followed by incubation at
37°C for 30 min. Samples were extracted with phenolchloroform and precipitated with ethanol. The protected
fragments were resuspended in loading buffer [90% formamide, 0.5⫻ Tris-borate-EDTA (TBE)] and separated on a
6% polyacrylamide gel in 1⫻ TBE containing 8 M urea at
250 V.
Distribution of poly(A) tail. Poly(A) tail distribution was
determined using a modification of the oligonucleotidedirected RNase H digestion assay (25). Briefly, RNA (20–30
µg) was resuspended in 28 µl of 1 mM EDTA and denatured at
70°C for 10 min. Then 1 µl of 3 M KCl and 0.2 A260 units of an
oligonucleotide that hybridizes 300 nucleotides from the 38
end of Cx32 mRNA (see Fig. 1) was added. The samples were
heated to 70°C and slowly cooled to room temperature to
anneal the oligonucleotide. In samples for which the poly(A)
tail was to be removed completely, oligo(dT) (1 µg) was also
added. The reaction was mixed with 30 µl of 2⫻ RNase H
digestion buffer (80 mM Tris, pH 7.4, 120 mM MgCl2, 200 mM
KCl, 16 U/ml of RNase H) and incubated at 37°C for 30 min.
The digestion was terminated by the addition of 5 µl of 0.5 M
EDTA followed by extraction with phenol-chloroform and
precipitation with ethanol. The digestion products were separated by electrophoresis on a 6% polyacrylamide-urea-TBE
gel at 250 V and electroblotted to charged-modified nylon
membranes (GeneScreen Plus) in 0.5⫻ TBE buffer in a
Hoeffer TransBlot apparatus at 20 V for 3 h. Blots were
hybridized to a radiolabeled antisense RNA probe that hybridizes to the 38 end of the Cx32 mRNA (see Fig. 1) at 65°C.
Membranes were washed at 65°C in two changes each of wash
I (1 M NaCl, 50 mM Tris, pH 7.4, 2 mM EDTA, 1% SDS), 2⫻
SSC (1⫻ SSC is 0.15 M NaCl and 0.015 M sodium citrate, pH
7.0) containing 0.2% SDS, and 0.2⫻ SSC containing 0.2%
SDS.
Calculations of poly(A) tail size. Autoradiograms were
imaged and analyzed using a Kodak DC120 digital camera
and Kodak ds1D software (Kodak Scientific Imaging Systems, New Haven, CT). Molecular weights were determined
by comparison to 32P-labeled DNA fragments run in an
adjacent lane on the same gel. The poly(A) tail size was then
determined by subtracting the calculated size of the deadenylated fragment. To calculate the average poly(A) tail size, the
intensity of the signal at each position along the lane in the
autoradiogram was determined. The average mobility was
determined as ⌺I · M/⌺(I), where I is the signal intensity, and
M is the mobility. The mobility was then converted into
molecular weight using a linear relation of the mobility as a
function of the log of the molecular weight.
Downloaded from http://ajpregu.physiology.org/ by 10.220.33.1 on June 18, 2017
Fig. 1. Schematic diagram of connexin 32 (Cx32) mRNA and probes
used. Solid black bar indicates protein-coding region of Cx32 mRNA;
untranslated regions are shown by thin lines. Region of mRNA
protected by antisense RNA probe and binding site and sequence of
antisense oligonucleotide (Cx32AS1) are shown above and below
mRNA, respectively. KpnI restriction site used to generate RNA
antisense probe is indicated.
Cx32 mRNA poly(A) tail distribution during normal
and inflammatory states. Previous studies showed a
decrease in the stability of hepatic Cx32 mRNA during
an acute inflammatory state induced by the administration of LPS (11). Because shortening of the poly(A) tail
is the first step in the degradation of many eukaryotic
mRNAs (3, 7), the distribution of Cx32 mRNA poly(A)
tails in rat liver was analyzed during LPS-induced
acute inflammation. Rats were intravenously injected
with LPS, and liver samples were taken at 0, 2, 3, 4, or
5 h postinjection. Total RNA was isolated, and poly(A)
tail distribution of Cx32 mRNA was analyzed by an
oligonucleotide-directed RNase H cleavage assay (Fig.
2A). The poly(A) tail distribution was compared with
the level of Cx32 mRNA determined by an RNase
protection assay (Fig. 2B). In liver samples of untreated rats, Cx32 mRNA poly(A) tails displayed a
heterogeneous distribution, with sizes of ⬃20–160 bases
long (Fig. 2A, lane 1). One particular feature is the
conspicuous periodicity of ⬃20–25 bases in the distribution of Cx32 mRNA poly(A) tail sizes that was calculated as described in METHODS and presented in Fig. 2C.
The average size of the poly(A) tail decreased between 2 and 3 h after LPS treatment but maintained
the periodic distribution (Fig. 2A, lanes 2 and 3, and C).
At this time little degradation of Cx32 mRNA was
observed in the RNase protection assay (Fig. 2B, lanes
3 and 4). The Cx32 mRNA poly(A) tail lengths changed
from 20–160 bases at time zero to 20–60 bases 5 h after
LPS injection. This reflects a change in the average of
poly(A) tail distribution from 78 bases (time zero) to 42
bases (5 h after LPS injection). Thus Cx32 mRNA
poly(A) tail lengths became smaller concurrent with
the disappearance of the message after LPS administration (Fig. 2, A–C). These data suggest that shortening
of Cx32 mRNA poly(A) tail precedes degradation of the
bulk of the message during inflammation. No completely deadenylated mRNA can be detected in liver
samples of either untreated or LPS-treated rats. This
can be seen by comparison of the size of the 38-terminal
fragment of Cx32 mRNA when it has been completely
deadenylated by treatment with oligo(dT) plus RNase
H (Fig. 2A, lane 6) with the size distribution of the
fragments that still have poly(A) tails (Fig. 2A, lanes
1-5).
One possible explanation for the heterogeneous distribution of Cx32 mRNA poly(A) tail is that the different
lengths represent mRNAs of different ages. Thus newly
synthesized Cx32 mRNA may have a long poly(A) tail,
which would get shorter as the message ages. Previous
studies showed that after Cx32 mRNA levels decline
after LPS treatment there is a reappearance of the
message during recovery from LPS injection, presumably due to new RNA synthesis (11). Accordingly, the
poly(A) tail distribution of Cx32 mRNA during recovery
from LPS administration was examined (Fig. 3). The
poly(A) tail length of Cx32 mRNA was 15–190 bases in
samples of untreated animals; it decreased to 15–90
bases after 8 h of LPS injection and recovered to
CX32
MRNA
POLY(A) TAIL DEGRADATION
R1251
Fig. 2. Analysis of Cx32 mRNA levels and poly(A) tail distribution.
Rats were injected with saline or with 1 mg/kg lipopolysaccharide
(LPS). At indicated times animals were killed and their livers were
removed for purification of RNA. A: Cx32 poly(A) tail distribution in
LPS-treated rats. Cx32 mRNA poly(A) tail distribution was determined by RNase H mapping using an antisense oligodeoxynucleotide
that hybridizes 300 nucleotides from the 38 end of the message.
Hybrids were treated with RNase H; cleavage products were separated on a polyacrylamide gel containing urea and electroblotted on
nylon membranes. Fragments were hybridized to an antisense RNA
probe and visualized by autoradiography. Lane 1: no RNA addition.
Lanes 2-6: Cx32 mRNA 38-end fragments from rat livers treated with
LPS for indicated times. Lane 7: Cx32 mRNA 38-end fragment from
rat liver RNA that was deadenylated with oligo(dT) plus RNase H. B:
Cx32 mRNA levels in LPS-treated rats. Cx32 RNA levels were
determined by RNase protection using an antisense RNA probe that
hybridizes to the 38 end of the message. Protected fragments were
analyzed by electrophoresis on a polyacrylamide gel containing urea and
visualized by autoradiography. Lane 1: markers. Lane 2: fragments
protected by hybridization of the probe to tRNA. Lanes 3-7: fragments
protected by hybridization of the probe to RNA from rat livers treated
with LPS for indicated times. C: densitometric tracing of Cx32 poly(A)
tail. Autoradiograph shown in A was scanned using a laser densitometer
(Molecular Dynamics). A linear tracing of the density of the poly(A) tails of
Cx32 mRNA from rats treated with LPS for 0, 2, or 3 h is shown.
15–195 bases within 12 h of LPS treatment (Fig. 3A,
compare lane 8 with lanes 2–7). The appearance of
Cx32 mRNA with a longer poly(A) tail corresponded
with a rise in the level of Cx32 mRNA 12 h after LPS
treatment (Fig. 3B, lane 8).
Fig. 3. Cx32 poly(A) tail distribution during inflammation and
recovery. Rats were injected with LPS, and their livers were removed
at indicated times for analysis of Cx32 poly(A) tail length or Cx32
mRNA levels as described. A: Cx32 poly(A) tail distribution. Lane 1:
Cx32 mRNA 38-end fragment from rat liver RNA that was deadenylated with oligo(dT) (dT) plus RNase H. Lanes 2-8: Cx32 mRNA 38-end
fragments from rat livers treated with LPS for indicated times.
B: Cx32 mRNA levels. Lane 1: fragments protected by hybridization
of probe to tRNA. Lanes 2-8: fragments protected by hybridization of
probe to RNA from rat livers treated with LPS for indicated times.
Downloaded from http://ajpregu.physiology.org/ by 10.220.33.1 on June 18, 2017
Effect of actinomycin D on Cx32 mRNA poly(A) tail
distribution. Previous studies showed that the LPSinduced degradation of Cx32 mRNA could be prevented
by the administration of the transcriptional inhibitor
actinomycin D (11). We were interested in determining
how treatment with this drug perturbed the degradation of Cx32 mRNA poly(A) tail. Therefore the poly(A)
tail distribution of Cx32 mRNA was examined in liver
samples obtained from rats treated with actinomycin
D, LPS, or combinations of both (Fig. 4A). Administration of actinomycin D alone resulted in a shortening of
the poly(A) tail of Cx32 mRNA (Fig. 4A, lane 2),
although degradation of the message was not evident
(Fig. 4B). The degree of poly(A) tail shortening after
actinomycin D treatment was less than that observed
after administration of LPS alone (Fig. 4A, lanes 3 and
4). When actinomycin D was given before LPS treatment, the degree of poly(A) shortening of Cx32 mRNA
was reduced compared with LPS treatment alone (Fig.
4A, lanes 5 and 6). Treatment with actinomycin D
resulted in the accumulation of a completely deadenylated Cx32 mRNA species, irrespective of whether it
was given in combination with LPS or not (Fig. 4A,
lanes 2, 5, and 6). To further investigate the effect of
actinomycin D on Cx32 mRNA, the poly(A) tail distribution of Cx32 mRNA was analyzed at different times
R1252
CX32
MRNA
POLY(A) TAIL DEGRADATION
Fig. 4. Deadenylated Cx32 mRNA accumulates in
livers of actinomycin D (Act D)-treated rats. Rats were
injected intraperitoneally with 2 mg/kg actinomycin D
(⫹) or ethanol (⫺) followed by injection with 1 mg/kg
LPS (⫹) or saline (⫺) 1 h later. After 6 h, animals were
killed and their livers were removed for analysis of
Cx32 poly(A) tail length or Cx32 mRNA levels as
described. A: Cx32 poly(A) tail distribution. Lanes 1-6:
Cx32 mRNA 38-end fragments from livers of untreated
rats (lane 1) or from livers of rats treated with LPS
(lanes 3 and 4), actinomycin D (lane 2), or both (lanes 5
and 6). Lane 7: Cx32 mRNA 38-end fragment from rat
liver RNA that was deadenylated with oligo(dT) plus
RNase H. B: Cx32 mRNA levels. Lane 1: fragments
protected by hybridization of probe to tRNA. Lanes 2-7:
fragments protected by hybridization of probe to RNA
from rat livers treated with LPS, actinomycin D, or
both as described above.
(Fig. 7). Cycloheximide pretreatment did not prevent
the LPS-induced degradation of Cx32 mRNA (Fig. 7B).
However, whereas LPS treatment caused Cx32 mRNA
poly(A) tails to decrease in size, cycloheximide treat-
Fig. 5. Deadenylated Cx32 mRNA accumulation in rat livers during
an actinomycin D time course. Rats were injected intraperitoneally
with 2 mg/kg actinomycin D. Animals were killed, and their livers
were removed for purification of RNA. Cx32 mRNA levels and poly(A)
tail lengths were determined as described. A: Cx32 poly(A) tail
distribution. Lane 1: Cx32 mRNA 38-end fragment from rat liver RNA
that was deadenylated with oligo(dT) plus RNase H. Lanes 2-10:
Cx32 mRNA 38-end fragments from rat livers treated with actinomycin D for indicated times. B: Cx32 mRNA levels. Lane 1: fragments
protected by hybridization of probe to tRNA. Lanes 2-10: fragments
protected by hybridization of probe to RNA from rat livers treated
with actinomycin D for indicated times.
Downloaded from http://ajpregu.physiology.org/ by 10.220.33.1 on June 18, 2017
after treatment with actinomycin D. The presence of
deadenylated Cx32 mRNA could be detected within 4 h
of treatment with the transcriptional inhibitor, which
continued to accumulate throughout 8 h of treatment
(Fig. 5A). A gradual shortening of the poly(A) tail
occurred concomitantly with the accumulation of the
deadenylated mRNA, even though no appreciable Cx32
mRNA degradation was observed (Fig. 5B).
The results described above could be due to a direct
effect of actinomycin D on the degradation process or by
an indirect effect, such as preventing the synthesis or
activation of an LPS-induced factor that causes the
degradation of Cx32 mRNA. To distinguish between
these possibilities, we performed an experiment in
which rats were injected with LPS first, followed by an
actinomycin D injection after 1 h. If an LPS-induced
factor was required to cause the degradation of Cx32
mRNA, we reasoned that it should be synthesized
within this time. However, if actinomycin D has a direct
effect on the degradation of Cx32 mRNA, then it should
block Cx32 degradation even if it was added after LPS.
Deadenylated Cx32 mRNA was observed in samples
from rats injected with actinomycin D, either with or
without prior treatment with LPS (Fig. 6A). Nevertheless, LPS induced the degradation of Cx32 mRNA even
if actinomycin D was added afterward (Fig. 6B). These
results suggest that actinomycin D affects the activation of an LPS-induced factor necessary for degradation
of Cx32 mRNA. Furthermore, actinomycin D treatment
had a second effect: the accumulation of a deadenylated
mRNA that was not previously detected in the absence
of the drug.
Effect of cycloheximide on Cx32 mRNA poly(A) tail
distribution. In contrast to actinomycin D, the translation inhibitor cycloheximide did not prevent the LPSinduced degradation of Cx32 mRNA; moreover, injection of cycloheximide alone was sufficient to cause the
degradation of Cx32 mRNA (11). To further explore
these effects, we examined the effect of the translational inhibitor cycloheximide on Cx32 mRNA poly(A)
tails. Rats were injected with cycloheximide or saline
and injected with LPS or saline 1 h later. After 6 h,
livers were harvested for determination of Cx32 mRNA
poly(A) tail lengths and message levels as described
CX32
MRNA
POLY(A) TAIL DEGRADATION
ment resulted in an average increase in the length of
Cx32 mRNA poly(A) tails, irrespective of the addition of
LPS (Fig. 7A). The Cx32 mRNA remaining after cycloheximide treatment had long poly(A) tails, indicating
that probably only newly synthesized Cx32 mRNA
remained.
We next examined Cx32 mRNA poly(A) tail lengths
during a time course of cycloheximide treatment. Injection of cycloheximide caused a loss of Cx32 mRNA (Fig.
8B). Although the kinetics of Cx32 mRNA degradation
are similar after either LPS or cycloheximide injection,
they result in different distributions of poly(A) tails.
Whereas LPS treatment results in shorter Cx32 mRNA
poly(A) tails, cycloheximide treatment results in a
longer average Cx32 mRNA poly(A) tail (Fig. 8A). This
distribution appears to be the result of a preferential
loss of Cx32 mRNA with short poly(A) tails.
The degradation of Cx32 mRNA induced by cycloheximide implied that cycloheximide treatment might be
inducing an inflammatory response in the liver. To test
this, we examined the transcription rates and mRNA
levels for genes that encode mRNAs known to change
after inflammation. Rats were injected with LPS or
cycloheximide; liver samples were harvested at various
times after injection for isolation of nuclei for run-off
transcription analysis and total RNA for analysis by
Northern blot. Cycloheximide treatment resulted in an
elevation of the transcription of the genes encoding
␣2-macroglobulin and ␤-fibrinogen, which are acute
phase response genes in rats (Fig. 9A). The kinetics and
levels of transcription were similar to those observed
after injection of LPS. We also examined transcription
of the gene encoding phosphoenolpyruvate carboxykinase (PEPCK), which is known to be reduced after
inflammation (24). Cycloheximide treatment attenuated PEPCK gene transcription with similar kinetics to
that observed after LPS treatment (Fig. 9A). These
results were confirmed by analyzing RNA levels by
Northern blot (Fig. 9B). In general, RNA levels were
commensurate with transcription rates; both ␣2macroglobulin and ␤-fibrinogen mRNAs became ele-
Fig. 7. Cycloheximide (CHX) treatment accelerates the degradation of Cx32 mRNA species with
short poly(A) tails. Rats were injected intraperitoneally with 6 mg/kg cycloheximide (⫹) or not (⫺),
followed by injection with 1 mg/kg LPS (⫹) or
saline (⫺) 1 h later. After 6 h, animals were killed,
and their livers were removed for purification of
RNA. A: Cx32 mRNA poly(A) tail distribution.
Cx32 poly(A) tail length was determined as described. Lane 1: Cx32 mRNA 38-end fragment
from rat liver RNA that was deadenylated with
oligo(dT) plus RNase H. Lanes 2-7: Cx32 mRNA
38-end fragments from rat livers treated with LPS
or CHX. B: Cx32 mRNA levels. Lane 1: fragments
protected by hybridization of probe to tRNA.
Lanes 2-6: fragments protected by hybridization
of probe to RNA from rat livers treated with LPS
and/or CHX.
Downloaded from http://ajpregu.physiology.org/ by 10.220.33.1 on June 18, 2017
Fig. 6. Cx32 mRNA poly(A) tails in livers of actinomycin D-treated
rats after LPS treatment. Rats were injected or not with 1 mg/kg
LPS; 1 h later they were injected intraperitoneally with 2 mg/kg
actinomycin D. After 0, 3, or 6 h after actinomycin D injection,
animals were killed, and their livers were removed for purification of
RNA. Cx32 poly(A) tail length and mRNA levels were determined as
described. A: Cx32 mRNA poly(A) tail distribution. Lanes 1 and 12:
Cx32 mRNA 38-end fragment from rat liver RNA that was deadenylated with oligo(dT) plus RNase H. Lanes 2-11: Cx32 mRNA 38-end
fragments from rat livers treated with LPS or actinomycin D. B: Cx32
mRNA levels. Cx32 RNA levels were determined by Northern blot as
described. Lanes 1-10: Cx32 mRNA levels in rat livers treated with
LPS or actinomycin D. Con, control.
R1253
R1254
CX32
MRNA
POLY(A) TAIL DEGRADATION
vated after either cycloheximide or LPS treatment.
Another gene, the expression of which parallels the
acute phase response, metallothionein, was also elevated after administration of cycloheximide. PEPCK
mRNA was reduced after LPS injection but not by
cycloheximide injection, even though both treatments
reduced the transcription rate of its gene. These results
suggest that cycloheximide treatment results in an
inflammatory or inflammatory-like response in the
liver. The effects on Cx32 mRNA levels and poly(A) tail
lengths could be due to two effects: a rapid shortening of
poly(A) tails of messages that are already committed to
degradation and a block in poly(A) shortening and
degradation of messages that have long poly(A) tails.
DISCUSSION
In normal rat liver, the expression of Cx32 is characterized by a low rate of transcription and the presence
of a very stable mRNA. During inflammation the level
of Cx32 mRNA is altered by an increase in message
degradation (1, 11). Consequently, changes in Cx32
mRNA stability play an important role in the expression of this gene within the liver. In the present study,
the changes in Cx32 mRNA stability during LPS-
induced inflammation were correlated by changes in
the distribution of Cx32 mRNA poly(A) tail. The first
detectable change in Cx32 mRNA is the shortening of
the poly(A) tail, which occurs before the degradation of
the bulk of Cx32 mRNA. Over time, the poly(A) tail
becomes shorter as the message disappears. No completely deadenylated Cx32 mRNA was detected after
any time point after the administration of LPS.
One of the most prominent features of the Cx32
mRNA poly(A) tail is a distinct periodicity of ⬃20–25
bases. This distribution is consistent with the distribution of bulk poly(A) tails of cytoplasmic mRNAs in
cultured cells as described by Baer and Kornberg (2). In
addition, a periodic distribution of poly(A) tail lengths
has been observed in rabbit and mouse globin mRNA
(14, 17). The periodic distribution of Cx32 poly(A) may
be related to the physical properties of the poly(A)
binding protein (PABP), which binds to poly(A) tails of
25–30 nucleotides (1). Our data suggest that the degradation of Cx32 mRNA proceeds by sequential loss of
20–25 nucleotide blocks of poly(A) tail from the 38 end.
It is possible that the loss of these 20–25 nucleotide
blocks of poly(A) tail is influenced by the binding of
PABP. There are several possible explanations of how
Downloaded from http://ajpregu.physiology.org/ by 10.220.33.1 on June 18, 2017
Fig. 8. Comparison of time course of LPS and cycloheximide treatments. Rats were injected intravenously with 1 mg/kg LPS or intraperitoneally with 6
mg/kg cycloheximide; at the indicated times livers
were removed for purification of RNA. Cx32 mRNA
levels and poly(A) tail lengths were determined as
described. A: Cx32 poly(A) tail distribution. Lanes 1
and 9: Cx32 mRNA 38-end fragment from rat liver
RNA that was deadenylated with oligo(dT) plus RNase
H. Lanes 2-8: Cx32 mRNA 38-end fragments from rat
livers treated with LPS for indicated times. Lanes
10-16: Cx32 mRNA 38-end fragments from rat livers
treated with cycloheximide for indicated times.
B: Cx32 mRNA levels. Cx32 RNA levels were determined by Northern blot as described. Lanes 1-7: Cx32
mRNA 38-end fragments from rat livers treated with
LPS for indicated times. Lanes 8-14: Cx32 mRNA
38-end fragments from rat livers treated with cycloheximide for indicated times.
CX32
MRNA
POLY(A) TAIL DEGRADATION
this can occur. For example, there might be 3-exonuclease that degrades the poly(A) tail, which could be
physically impeded by the binding of the PABP. Alternatively, a poly(A)-specific nuclease that requires the
presence of the PABP could cleave the poly(A) tail in
discrete units dictated by the binding of the PABP.
Another observation is that the size distribution of
Cx32 mRNA poly(A) tail seems to reflect the age of the
message. Thus newer mRNA has a long poly(A) tail,
whereas old messages have short poly(A) tails.
The effects of actinomycin D on Cx32 mRNA are
complex. Previous studies have shown that administration of actinomycin D before LPS injection resulted in
the stabilization of Cx32 mRNA (11). In the present
study, we showed that actinomycin D does not stabilize
the message if LPS is given first. One interpretation of
this finding is that LPS affects the activation of a factor
involved in the degradation of Cx32 mRNA. It is
curious that administration of actinomycin D results in
the appearance of a completely deadenylated mRNA
species, regardless of whether it is given before, after,
or in the absence of LPS treatment. Moreover, if
actinomycin D is administered either in the absence of
or before LPS treatment, a reduction in the length of
the Cx32 mRNA poly(A) tail still occurs. However, this
reduction is not as great as that which occurs if LPS is
given before or in the absence of actinomycin D. We
speculate that treatment with actinomycin D alone
reveals the normal rate of Cx32 mRNA poly(A) shortening that occurs in the absence of stress. This rate
appears to be ⬃25–50 residues/h, as estimated from
the data in Figs. 4–6. If poly(A) tail shortening is the
rate-limiting step in Cx32 mRNA degradation, then it
would take ⬃25–50 h to degrade a message if the initial
poly(A) tail size is ⬃200 nucleotides. This is in reasonably close agreement with data obtained from Kren and
coworkers (15), who reported that the half-life of Cx32
mRNA in the presence of actinomycin D was 10.9 h in
normal liver. Although we detected little or no degradation of Cx32 mRNA in the presence of actinomycin D in
normal livers, we do note that a completely deadenylated species accumulates. The apparent discrepancy
between the results of Kren and coworkers and ours
could be due to the fact they analyzed poly(A)-selected
mRNA, rather than total RNA. Therefore, the accumulated deadenylated mRNA that we observed would not
have been detected in those studies. The appearance of
deadenylated mRNA that occurs after actinomycin D
treatment suggests that degradation of nonpolyadenylated mRNA might be inhibited under these conditions,
although it is unclear how this could occur. One possibility is that actinomycin D treatment might result in an
alteration in messenger ribonucleoprotein particle structure or transport that occurs when transcription is
inhibited, as has been reported by Dreyfuss and coworkers (8, 20). It is interesting to note that there are other
mRNAs, such as the c-fos mRNA, for which the degradation is altered by treatment with actinomycin D (22).
The block in degradation of these mRNAs might also be
due to an inhibition of the degradation of the deadenylated species.
Downloaded from http://ajpregu.physiology.org/ by 10.220.33.1 on June 18, 2017
Fig. 9. Effects of cycloheximide on transcription rate and mRNA
levels in rat livers. Rats were injected with saline, cycloheximide (6
mg/kg), or LPS (1 mg/kg). Livers were isolated at the indicated times,
and nuclei and total RNA were prepared. A: transcription in isolated
nuclei. Nuclei isolated from rats treated with cycloheximide or LPS
were incubated in presence of radiolabeled UTP. Radioactive runoff
transcripts were purified and hybridized to filter-bound plasmids
containing ␤-fibrinogen (␤-fib), ␣2-macroglobulin (␣2 Mac), and phosphoenolpyruvate carboxykinase (PEPCK) sequences. Filters were washed
and exposed to X-ray film. B: Northern blots. Total RNA isolated from rats
injected with LPS or cycloheximide was separated on formaldehyde/
agarose gels, blotted to Nylon membranes, and probed for labeled DNAs
containing ␤-fibrinogen, ␣2-macroglobulin, and PEPCK sequences.
Filters were washed and exposed to X-ray film. MT, metallothionein.
R1255
R1256
CX32
MRNA
POLY(A) TAIL DEGRADATION
Fig. 10. Model for degradation of Cx32 mRNA. Transcribed region of
mRNA is represented by wavy line; poly(A) tail is indicated by ovals
representing blocks of 25–30 adenylate residues (An ). knormal, ratelimiting step in degredation; kLPS, rate of poly(a) tail removal; k?,
removal of final block of poly(A); kdegredation, after all of poly(A) tail
removed, rapid degredation of deadenylated mRNA species. See
DISCUSSION for more information.
detected during actinomycin D treatment, suggesting
that actinomycin D inhibits the degradation of the
deadenylate mRNA. We suggest that after the poly(A)
tail of Cx32 mRNA is removed, the subsequent degradation steps, which are rapid, are likely to be common to
the degradation of many, if not most, mRNAs. We
suggest that actinomycin D and cycloheximide inhibit
different steps of the degradation pathway. Cycloheximide inhibits poly(A) tail shortening of mRNAs at an
early step (but nevertheless accelerates poly(A) tail
shortening of mRNA already committed to the degradation process), whereas actinomycin D inhibits the degradation of completely deadenylated mRNA.
In summary, our data suggest that shortening of the
poly(A) tail of Cx32 mRNA is the first step in the
degradation of this message during LPS-induced inflammatory states. It is also likely that shortening of the
poly(A) tail is the rate-limiting step in the degradation
of Cx32 mRNA. Finally, there is an apparent correlation between the size of the poly(A) tail and the age of
the message.
Perspectives
Changes in gene expression are an important component in the cellular response to different disease conditions such as inflammation. Although transcriptional
regulation has been the major focus of research in gene
expression, it is becoming increasingly clear that posttranscriptional mechanisms also play an important
role. Indeed, the concentration of any message is determined equally by its rate of synthesis (transcription
rate) and its rate of degradation. Accordingly, mechanisms of mRNA degradation have gained attention
recently. In some cases, shortening of the poly(A) tail
has been observed to be the first step in mRNA decay,
Downloaded from http://ajpregu.physiology.org/ by 10.220.33.1 on June 18, 2017
Curiously, the translation inhibitor cycloheximide
has the opposite effect as actinomycin D. Cx32 mRNA
levels decrease after administration of this drug. It is
likely that cycloheximide is mimicking at least some
aspects of the acute phase response. Indeed, we have
observed that injection of cycloheximide causes an
elevation in the transcription rate of ␤-fibrinogen and
␣2-macroglobulin genes, both of which are acute phase
response genes in rodents and part of the inflammatory
response (Fig. 9). Alternatively, cycloheximide may
directly accelerate the degradation of Cx32 mRNA, as it
does for ␤-tubulin (17a). We also note that after cycloheximide treatment Cx32 mRNA poly(A) tails have a
longer average length, which is probably due to the loss
of the species with the shorter tails. Perhaps a message
needs some time before it is committed to the process of
poly(A) shortening and degradation. In this scenario,
these ‘‘older’’ mRNAs that have shorter poly(A) tails
can be rapidly degraded after cycloheximide treatment.
However, the ‘‘newer’’ mRNAs that have longer poly(A)
tails have not yet been committed to the degradation
pathway and are thus resistant to the effects of cycloheximide. It is perhaps not surprising that cycloheximide treatment might have multiple effects; for example PEPCK mRNA is apparently stabilized by
cycloheximide treatment (Fig. 9B).
A standard model for the degradation of eukaryotic
mRNA has been derived from studies in yeast (3, 7). In
this model, poly(A) tail shortening is the first step in
the degradation of the message and is the step that
apparently determines the half-life of the mRNA. After
the poly(A) tail has been shortened to a minimal length,
the 58 cap structure of the mRNA is removed, and the
mRNA is rapidly degraded by a 58-exonuclease. The
degradation of Cx32 mRNA has many features in
common with this standard model, such as the shortening of the poly(A) tail being the first and rate-limiting
step. However, we believe that the poly(A) tail of Cx32
mRNA is completely removed before the subsequent
degradation steps. We propose that the degradation of
Cx32 mRNA comprises the following steps, which are
summarized in Fig. 10. The first step is a commitment
of the mRNA to the degradation process. This step may
be inhibited by cycloheximde. The second step is a
shortening of the poly(A) tail, which is the rate-limiting
step in degradation. During inflammation, the rate of
poly(A) tail removal increases. After the poly(A) tail is
reduced to short length, the removal of the final block of
poly(A) may or may not occur at the same rate as
poly(A) shortening. Although we have no evidence to
suggest that the removal of the last block of poly(A) is
kinetically different from poly(A) shortening, evidence
from other laboratories suggests that it might not be
the case. For example, purified yeast poly(A) nuclease
cannot completely remove all of the poly(A) tail of a
synthetic substrate (21). After all of the poly(A) tail has
been removed, the deadenylated mRNA species is degraded very rapidly (Fig. 6). If the rate of degradation is
very rapid compared with the rate of poly(A) removal,
then the deadenylated mRNA species would not normally be detected. However, deadenylated mRNA is
CX32
MRNA
POLY(A) TAIL DEGRADATION
whereas in others, an endonucleolytic cleavage occurs
first. Previous studies showed that the expression of
Cx32 is regulated by changes in mRNA stability. Data
presented in this study suggest shortening of the
poly(A) tail of Cx32 mRNA is the first step in degradation of the message. This information, in addition to
other studies on the effect of inflammation at the
transcriptional level suggest that this disease condition
has multiple effects on cellular homeostasis. Attempts
to control the inflammatory response require a profound knowledge of the mechanisms involved in gene
expression.
Received 5 June 1998; accepted in final form 19 January 1999.
REFERENCES
1. Baer, B. W., and R. D. Kornberg. The protein responsible for
the repeating structure of cytoplasmic poly(A)-ribonucleoprotein.
J. Cell Biol. 96: 717–721, 1983.
2. Baer, B. W., and R. D. Kornberg. Repeating structure of
cytoplasmic poly(A)-ribonucleoprotein. Proc. Natl. Acad. Sci.
USA 77: 1890–1892, 1980.
3. Beelman, C. A., and R. Parker. Degradation of mRNA in
eukaryotes. Cell 81: 179–183, 1995.
4. Bennett, M. V. L., L. C. Barrio, T. A. Bargiello, D. C. Spray,
E. Hertzberg, and J. C. Saez. Gap junctions: new tools, new
answers, new questions. Neuron 6: 305–320, 1991.
5. Bennett, M. V. L., and D. A. Goodenough. Gap junctions,
electrotonic coupling and intercellular communication. Neurosci.
Res. Program Bull. 16: 373–486, 1978.
6. Chomczynski, P., and N. Sacchi. Single-step method of RNA
isolation by acid guanidinium thiocyanate-phenol-chloroform
extraction. Anal. Biochem. 162: 156–159, 1987.
7. Decker, C. J., and R. Parker. Mechanisms of mRNA degradation in eukaryotes. Trends Biochem. Sci. 19: 336–340, 1994.
8. Dreyfuss, G., S. A. Adam, and Y. D. Choi. Physical changes in
cytoplasmic ribonucleoprotien in cells treated with inhibitors of
mRNA transcription. Mol. Cell. Biol. 4: 415–423, 1984.
9. Gingalewski, C., and A. De Maio. Differential decrease in
connexin 32 expression in the ischemic and nonischemic regions
of rat liver during ischemia/reperfusion. J. Cell. Physiol. 171:
20–27, 1997.
10. Gingalewski, C., N. G. Theodorakis, J. Yang, S. C. Beck,
and A. De Maio. Distinct expression of heat shock and acute
phase genes during regional hepatic ischemia/reperfusion. Am. J.
Physiol. 271 (Regulatory Integrative Comp. Physiol. 40): R634–
R640, 1996.
11. Gingalewski, C., K. Wang, M. G. Clemens, and A. De Maio.
Posttranscriptional regulation of connexin 32 expression in liver
during acute inflammation. J. Cell. Physiol. 166: 461–467, 1996.
12. Gumucio, J. Hepatocyte heterogeneity: the coming of age from
the description of a biological curiosity to a partial understanding of its physiological meaning and regulation. Hepatology 9:
154–160, 1989.
13. Hooper, M. L., and J. H. Subak-Sharpe. Metabolic cooperation between cells. Int. Rev. Cytol. 69: 45–104, 1981.
14. Kelly, J. M., and R. A. Cox. Periodicity in the length of
38-poly(A) tails from native globin mRNA of rabbit. Nucleic Acids
Res. 10: 4173–4179, 1982.
15. Kren, B. T., N. M. Kumar, S. Wang, N. B. Gilula, and C. J.
Steer. Differential regulation of multiple gap junction transcripts and proteins during rat liver regeneration. J. Cell Biol.
123: 707–718, 1993.
16. Kumar, N. M., and N. B. Gilula. The gap junctional communication channel. Cell 84: 381–388, 1996.
17. Mansbridge, J. N., J. A. Crossley, W. G. Lanyon, and R.
Williamson. The poly(adenylic acid) sequence of mouse globin
messenger RNA. Eur. J. Biochem. 44: 261–269, 1974.
17a.Pachter, J. S., T. J. Yen, and D. W. Cleveland. Autoregulation
of tubulin expression is achieved through specific degradation of
polysomal tubulin mRNAs. Cell 51: 283–292, 1987.
18. Page, E., and C. K. Manjunath. Communicating junctions
between cardiac cells. In: The Heart and Cardiovascular Systems, edited by H. A. Fozzard, E. Haber, A. M. Katz, and H. E.
Morgan. New York: Raven, 1986, p. 573–600.
19. Paul, D. Molecular cloning of cDNA for rat liver gap junction
protein. J. Cell Biol. 103: 123–134, 1986.
20. Pinol-Roma, S., and G. Dreyfuss. Transcription-dependent
and transcription-independent nuclear transport of hnRNP proteins. Science 253: 312–314, 1991.
21. Sachs, A. B., and J. A. Deardorff. Translation initiation
requires the PAB-dependent poly(A) ribonuclease in yeast. Cell
70: 961–973, 1992.
22. Shyu, A.-B., M. E. Greenberg, and J. B. Belasco. The c-fos
transcript is targeted for rapid decay by two distinct mRNA
degradation pathways. Genes Dev. 3: 60–72, 1989.
23. Traub, O., P. M. Drüge, and K. Willecke. Degradation and
resynthesis of gap junction protein in plasma membranes of
regenerating liver after partial hepatectomy or cholestasis. Proc.
Natl. Acad. Sci. USA 80: 755–759, 1983.
24. Wang, K., C. S. Deutschman, M. G. Clemens, and A. De
Maio. Reciprocal expression of acute phase genes and phosphoenolpyruvate carboxykinase (PEPCK) during acute inflammation. Shock 3: 204–209, 1995.
25. Zanger, R. C., M. Hernandez, T. A. Kocarek, and R. F.
Novak. Determination of the poly(A) tail lengths of a single
mRNA species on total hepatic RNA. Biotechniques 18: 465–469,
1995.
Downloaded from http://ajpregu.physiology.org/ by 10.220.33.1 on June 18, 2017
We thank Dr. David Paul for the gift of Cx32 cDNA, Maria Lourdes
Mooney for technical assistance, and the members of the De Maio
laboratory for critically reading the manuscript.
This research was supported by the Robert Garrett Research
Foundation (to A. De Maio).
Present address for N. G. Theodorakis: Dept. of Surgery,
Georgetown University Medical Center, Washington, DC 20007.
Address for reprint requests and other correspondence: A. De
Maio, Division of Pediatric Surgery, Johns Hopkins Univ., School of
Medicine, 720 Rutland Ave., Baltimore, MD 21205 (E-mail:
[email protected]).
R1257