Báo cáo khoa học: Phân tích ảnh hưởng của thay đổi độ ổn định mRNA đến mức độ ổn định cyclin mRNA trong chu kỳ tế bào động vật có vú

Analysis of the contribution of changes in mRNA stability

to the changes in steady-state levels of cyclin mRNA in the

mammalian cell cycle

Anna Penelova

, Larry Richman

, Barbara Neupert

, Viesturs Simanis

and Lukas C. Ku

¨hn

1 Genetics Unit, Swiss Institute for Experimental Cancer Research (ISREC), Epalinges, Switzerland

2 Cell Cycle Control Laboratory, Swiss Institute for Experimental Cancer Research (ISREC), Epalinges, Switzerland

Introduction

Cyclin-dependent kinases (cdks) are central to the pro-

gression and control of the mammalian cell cycle [1–3].

Their activity is regulated positively by interaction with

cyclins and negatively by cdk-inhibitors that bind to

cdk-cyclin complexes. Cyclin-dependent kinases are

also regulated by phosphorylation. The protein levels

of cdk activators and inhibitors are tightly controlled

by the rate of their synthesis and by speciﬁc phos-

phorylation events that initiate ubiquitination and

degradation by proteasomes, thus limiting expression

to a speciﬁc cell cycle phase. D-type cyclins (D1, D2

and D3) are highest in early G

phase, when they acti-

vate cdk4 and cdk6. E-type cyclins (E1 and E2) peak

in late G

and associate with cdk2 to complete G

and

initiate S phase. Cyclin A2 accumulates during S phase

with highest levels in late S and G

. It associates with

cdk2 during S phase and subsequently with cdk1

(cdc2) to pass the S ⁄G

boundary. Finally, progression

through G

and mitosis require cyclins B1 and B2 that

associate with cdk1.

Because the expression of cyclins plays a large part

in controlling cell cycle progression, it is important to

understand the transcriptional and post-transcriptional

mechanisms that inﬂuence cyclin levels. Indeed, recent

microarray data demonstrate signiﬁcant variations of

cyclin mRNA levels in human ﬁbroblasts after release

from serum starvation (G

phase) [4] or a double

thymidine block (late G

phase) [5]. Transcription of

Keywords

cell cycle; cyclin; elutriation; fluorescence

activated cell sorter; mRNA stability

Correspondence

L. C. Ku

¨hn, Swiss Institute for Experimental

Cancer Research, Genetics Unit, Chemin

des Boveresses 155, CH-1066 Epalinges,

Switzerland

Fax: +4121 652 69 33

Tel: +4121 692 58 36

E-mail: lukas.kuehn@isrec.ch

(Received 29 June 2005, accepted 16

August 2005)

doi:10.1111/j.1742-4658.2005.04918.x

Cyclins are the essential regulatory subunits of cyclin-dependent protein

kinases. They accumulate and disappear periodically at speciﬁc phases of

the cell cycle. Here we investigated whether variations in cyclin mRNA

levels in exponentially growing cells can be attributed to changes in mRNA

stability. Mouse EL4 lymphoma cells and 3T3 ﬁbroblasts were synchron-

ized by elutriation or cell sorting. Steady-state levels and degradation of

cyclin mRNAs and some other cell cycle related mRNAs were measured at

early G

, late G

, S and G

⁄M phases. In both cell lines mRNAs of cyclins

C, D1 and D3 remained unchanged throughout the cell cycle. In contrast,

cyclin A2 and B1 mRNAs accumulated 3.1- and 5.7-fold between early G

and G

⁄M phase, whereas cyclin E1 mRNA decreased 1.7-fold. Mouse

cyclin A2 and B1 genes, by alternative polyadenylation, gave rise to more

than one transcript. In both cases, the longer transcripts were the minor

species but accumulated more strongly in G

⁄M phase. All mRNAs were

rather stable with half-lives of 1.5–2 h for cyclin E1 mRNA and 3–4 h for

the others. Changes in mRNA stability accounted for the accumulation in

⁄M phase of the short cyclin A2 and B1 mRNAs, but contributed only

partially to changes in levels of the other mRNAs.

Abbreviations

cdk, cyclin dependent kinase; DMEM, Dulbecco’s modified Eagle medium; DRB, 5,6-dichloro-1-b-D-ribofuranosylbenzimidazole; FACS,

fluorescence activated cell sorter; FBS, fetal bovine serum; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; UTR, untranslated region.

FEBS Journal 272 (2005) 5217–5229 ª2005 FEBS 5217

D-type cyclin mRNA is certainly induced by mitogenic

signals that trigger G

⁄G

transition [6], whereas tran-

scription of cyclin E1 starts in late G

[7]. Likewise,

A- and B-type cyclin mRNA were reported to be

induced in S and G

⁄M phase as a consequence of

events in G

phase [4,8–10]. In addition, several studies

concluded that cyclin, cdk and cdk-inhibitor mRNA

stability can vary throughout the cell cycle [11–15]. Cer-

tain transacting proteins such as HuR were proposed

as regulators of changes in mRNA stability during the

cell cycle [15]. In this context it is of interest that during

vertebrate evolution many of the cyclin mRNAs show

a rather high phylogenetic conservation of their 3¢un-

translated regions (3¢UTR) suggesting that speciﬁc ele-

ments in the 3¢UTR might contribute to control RNA

half-life [16]. On the other hand a recent study with

human MOLT-4 cells showed no change in cyclin

mRNA half-lives throughout the cell cycle [17].

While most studies on cyclin mRNA stability in the

cell cycle have been carried out with human cells,

essential regulatory steps are likely to be conserved in

evolution and thus amenable to genetic analysis in the

mouse. We therefore examined mRNA expression and

stability in synchronized mouse lymphoma EL4 cells

and 3T3 ﬁbroblasts. We analyzed the mRNA steady-

state level and half-life of mouse cyclins and a selection

of other cell cycle related genes for which important

cell cycle-related changes were reported in microarray

studies [4,5]. We show that mRNAs for cks2, cyclin

A2, B1 and E1 vary in the cell cycle but that mRNA

half-life changes contribute only partially to these vari-

ations.

Results

Steady-state levels of cyclin mRNA in the cell

cycle

In a ﬁrst series of experiments we determined whether

mRNA steady-state levels of cyclins and several cell

cycle-related mRNAs change at different positions in

the cell cycle. To achieve this, about 5 ·10

logarith-

mically dividing mouse EL4 lymphoma cells were sep-

arated by elutriation into 12–15 fractions. EL4 cells

are particularly well suited for this separation method

as they are not adherent and grow to high density. An

aliquot of each fraction was analysed on a ﬂuorescence

activated cell sorter (FACS) for the proﬁle of DNA

content after propidium iodide staining. Pooled frac-

tions of cells highly enriched in early G

, late G

and G

⁄M phase were selected for further analysis

(Fig. 1A). Steady-state mRNA levels were analysed by

real-time PCR. By taking the early G

cells as a refer-

ence, mRNA levels of cyclins C, D1 and D3, as well

as c-myc, RanGTPase and RanBP1 were unchanged

(Fig. 1B). Cyclin D2 was not expressed in EL4 cells.

Cyclin E1 mRNA increased slightly in late G

and

then diminished about 2-fold in G

⁄M phase. The

clearest induction in G

⁄M compared to early G

cells

was observed for cyclin A2 mRNA (3.1-fold) and

cyclin B1 mRNA (5.7-fold). Cks2 mRNA was three-

fold higher in S phase and 2.4-fold higher in G

⁄M

and very similarly the control histone H4 mRNA

showed a threefold increase in S phase. Thus, changes

in mRNA occur parallel to changes in protein expres-

sion [18,19], but cannot account for strong differences

of cyclin protein levels that are modulated post-trans-

lationally [20,21]. Overall we observed smaller differ-

ences in RNA steady-state levels than those reported

by others for human cells [11,12,15].

The relatively small changes in mRNA levels made

us wonder whether there was any problem with the

separation procedure. To verify this, we separated EL4

cells in logarithmic growth by the FACS according to

cellular DNA content revealed by Hoechst 33342

(Fig. 1C). This method gave highly enriched cell popu-

lations with sufﬁcient amount of mRNA for real-time

PCR measurements, but could not distinguish early

and late G

cells. The results were qualitatively very

similar to the measurements obtained with elutriated

cells, although somewhat less pronounced because we

took the average G

cells as a reference. We found

again that mRNA levels for cyclins C, D1 and D3 as

well as for c-myc, RanGTPase and RanBP1 showed

no changes in the cell cycle (Fig. 1D). Cyclin E1

mRNA decreased from G

to G

⁄M by a factor of

1.7-fold, whereas the mRNA of cyclin A2, cyclin B1

and cks2 increased 1.9-, 3.2- and 2.4-fold, respectively.

We found similar results with mouse 3T3 cells that

were either sorted by the FACS or synchronized by a

double thymidine block. They showed no change in

steady-state levels for most mRNAs, with the excep-

tion of a two- to 2.5-fold increase between G

and

⁄M for mRNAs of cks2, cyclins A2 and B1 (data

not shown).

Next we wanted to be sure that cells were fully

viable after elutriation. To test this, elutriated cell frac-

tions were brought back into cell culture for 2–8 h, at

which time their DNA content was analysed by the

FACScan (Fig. 2). EL4 cells advanced synchronously

in the cell cycle without signiﬁcant delay (Fig. 2). The

ﬁrst fractions of cells harvested in the elutriation pro-

tocol behaved like early G

cells. They enter S phase

only after about 4 h of culturing, whereas later frac-

tions comprise G

cells that resumed S phase almost

immediately and that we considered therefore as late

Cell cycle regulation of mouse cyclin mRNAs A. Penelova et al.

5218 FEBS Journal 272 (2005) 5217–5229 ª2005 FEBS

cells. The FACS proﬁles allowed us to estimate the

total cycle to about 13 h of which about 6 h corres-

pond to G

phase, about 3.5 h to S phase and another

3.5 h to G

phase and mitosis. This correlated well

with the estimated doubling time of EL4 cells in log-

arithmic growth.

No major variation in the mRNA half-life of cyclin

mRNAs in the cell cycle

In order to test whether changes in mRNA steady-

state levels correlate with any changes in mRNA sta-

bility, we carried out half-life measurements on the

different elutriated cell fractions. Transcription was

inhibited with 5,6-dichloro-1-b-d-ribofuranosylbenzimi-

dazole (DRB) and mRNA measured at 0, 30, 60, 120

and 180 min by real-time PCR (Fig. 3). Half-life

measurements showed no strong differences in mRNA

degradation rates in the different cell cycle phases.

Only mRNA of cyclins A2 and B1 showed at most a

1.6-fold higher stability in S and G

⁄M phases. As a

positive control, we found as expected a rapid degra-

dation with a half-life of less than 1 h for the unstable

c-myc mRNA, indicating that the transcription block

by DRB was effective. Similar data were also obtained

with actinomycin D or with EL4 cells enriched in spe-

ciﬁc cell cycle phases by the FACS (data not shown).

Northern blot analysis of mRNA from fractions

of elutriated EL4 cells

We needed to conﬁrm the real-time PCR data by nor-

thern blots of RNA from elutriated cells. The cell cycle

distribution of the cell fractions is shown in Fig. 4.

Fig. 1. Steady-state levels of cyclin mRNAs in enriched cell cycle fractions of mouse EL4 cells. Cells were separated either by elutriation or

by cell sorting into G

,SandG

⁄M phase fractions. (A) Cells were separated by elutriation into about 15 fractions. The DNA content was

measured after Hoechst staining by the FACS. Representative fractions showed a strong enrichment for cells in early G

(a), late G

(b),

S (c) or G

⁄M phase (d). (B) The mRNA content of these fractions was quantified by real-time PCR and normalized to mARP0 mRNA. Values

in early G

cells were set as 1. Results are the average of at least four experiments ± SD. (C) Typical FACS profile of the DNA content of

logarithmically growing EL4 cells stained by Hoechst 33342. Cells were separated by FACS sorting into three fractions as indicated. (D) In

each fraction, mRNAs were quantified by real-time PCR. Values are normalized to mARP0 mRNA. The amount of each mRNA in G

phase

cells is set as 1. Results are the average of two experiments.

A. Penelova et al. Cell cycle regulation of mouse cyclin mRNAs

FEBS Journal 272 (2005) 5217–5229 ª2005 FEBS 5219

Northern blot hybridizations were carried out for

genes that had shown differences of steady-state levels

in the cell cycle (cyclin A2, B1 and E1). The invariant

mRNAs of cyclin D3 and glyceraldehyde-3-phosphate

dehydrogenase (GAPDH) were analysed as controls

(Fig. 4A). These blots revealed that there is more than

one transcript from mouse cyclin A2 and B1 genes. In

the case of mouse cyclin A2, besides the more abun-

dant mRNA of 1.8 kb, there is a minor species of

3.0 kb. For mouse cyclin B1 we see in addition to the

most abundant 1.7 kb mRNA, a 2.5 kb mRNA and a

very minor 2.1 kb mRNA. Based on EST database

searches taking into account all 5¢and 3¢ends of iden-

tiﬁed cDNAs, we concluded that these mRNA hetero-

geneities arise from alternative polyadenylation. This

was conﬁrmed by control hybridizations with 3¢UTR

probes downstream of the ﬁrst polyadenylation site

that consistently revealed only the longer transcripts

(Fig. 5). Longer transcripts of cyclins A2 and B1 were

also visible in mouse 3T3 cells and in mouse thymus

and spleen, but were less clearly detectable in tissues

with fewer proliferating cells (data not shown). In tes-

tes two cyclin A2 transcripts and only the shorter

cyclin B1 transcript were visible, in agreement with

previous reports [22,23]. The hybridizations with a

coding region probe, after normalization to GAPDH

expression, showed that the 1.8 kb cyclin A2 mRNA

accumulated about 1.8-fold in S and G

⁄M phase com-

pared to early G

phase, while the 1.7 kb cyclin B1

mRNA increased at most threefold (Fig. 4B). More

strikingly, the longer mRNA variants of both cyclins

accumulated much more than the short ones and

reached 35–45% of the total amount in late S and

⁄M cells. The 2.5 kb mRNA of cyclin B1 showed

reproducibly a strong, up to 10-fold increase, while the

magnitude of the 3.0 kb cyclin A2 mRNA increase

Fig. 2. Cell cultures of mouse EL4 cell fractions after elutriation. Immediately after elutriation selected cell fractions enriched in a given cell

cycle phase (as indicated) were put back into culture for 2, 4, 6 or 8 h. The cell cycle progression was analysed by FACS profiles of the DNA

content of propidium iodide stained aliquots of cells.

Cell cycle regulation of mouse cyclin mRNAs A. Penelova et al.

5220 FEBS Journal 272 (2005) 5217–5229 ª2005 FEBS

showed some variation between experiments (Fig. 4

and Fig. 6A). The reason for this is unclear. At the

same time cyclin E1 declined 2.2-fold.

Based on northern blot analysis, as already deter-

mined by real-time PCR, changes in steady-state levels

were not associated with strong modiﬁcations in the

mRNA half-life in different cell cycle phases (Fig. 6B).

Half-lives were about 2 h for cyclin E1 mRNA and 3–

4.5 h for the other transcripts in most cell cycle phases.

These values are close to those obtained by real-time

PCR. Only in the case of the short cyclin A2 and B1

mRNAs were the half-lives signiﬁcantly prolonged in

⁄M phase. This change fully accounts for the accu-

mulation of these transcripts in G

⁄M phase. For the

long cyclin A2 and B1 mRNAs we found also a minor

stability change that cannot account for their strong

accumulation in G

⁄M phase.

Given the clear accumulation of cyclin A2 and B1

mRNAs in G

⁄M compared to early G

and the

reports on human cells that demonstrated a strong dif-

ference in mRNA stability in these phases [11,12,15], it

seemed important to verify carefully mRNA half-lives

at the transition between G

⁄M and early G

phase.

For this, EL4 cells were arrested in mitosis by noco-

dazole and then released for 0, 30, 60, 90, 120 or

180 min. At least 75% of the arrested cells completed

mitosis and divided within 3 h (Fig. 7A). mRNA half-

lives were measured at each time-point (Fig. 7B). The

results indicated no signiﬁcant changes in mRNA half-

life for most transcripts except the long transcript of

cyclin B1 which appeared to decay quite rapidly at the

time of the release. Notably with the nocodazole arres-

ted cells we did not ﬁnd the prolonged half-life seen

before for cyclin A2 and B1 mRNAs in enriched

⁄M fractions. Based on these data it seems unlikely

that changes in steady-state levels can be attributed to

transient changes in half-life.

Discussion

The purpose of the present study was to analyse the

contribution of post-transcriptional mechanisms in the

cell cycle regulation of cyclin mRNAs. Previous studies

on HeLa cells [11,12] and colorectal carcinoma RKO

cells [15] had found strong mRNA stability changes.

We reasoned that such a feature, if it was physiologi-

cally important, should be conserved between human

and mouse. We therefore analysed the steady-state lev-

els and mRNA stability at different points in the cell

cycle of mouse 3T3 and EL4 cell lines. The general

conclusion of our analysis is that, in contrast to these

earlier studies, but in agreement with a recent publica-

tion on human MOLT 4 cells [17], most cyclin

Fig. 3. Half-life of cyclin mRNAs in EL4 cell fractions enriched by

elutriation. Cell fractions were put back into cell culture for about

30 min and incubated for 0, 30, 60, 120 or 180 min with DRB prior

to the isolation of total mRNA. Remaining mRNA was measured by

real-time PCR and normalized to mARP0 mRNA. The short-lived

c-myc mRNA served as a control. (A) The mRNA half-life was calcu-

lated from linear regression on semi-logarithmic plots. Results are

the average of three to four experiments ± SD. (B) Alternatively,

data of decay of mRNAs showing the strongest changes in steady-

state levels (Fig. 1) were plotted on a semi-logarithmic scale and a

single regression line calculated. The intercept of the regression

line at log

of 50% ¼1.699, corresponds to the half-life. The lower

and upper 95% confidence limits were at 0.75 and 1.5 times the

half-life.

A. Penelova et al. Cell cycle regulation of mouse cyclin mRNAs

FEBS Journal 272 (2005) 5217–5229 ª2005 FEBS 5221

Báo cáo khoa học: Analysis of the contribution of changes in mRNA stability to the changes in steady-state levels of cyclin mRNA in the mammalian cell cycle

Chủ đề:

Tài liệu liên quan

Tài liêu mới

AI tóm tắt

Giới thiệu tài liệu

Đối tượng sử dụng

Từ khoá chính

Nội dung tóm tắt

Hỗ trợ

Phương thức thanh toán

Theo dõi chúng tôi