| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,1
,1,23
* Institute of Molecular Biology,
Institute of Molecular Pathology, Vienna Biocenter, University of Vienna, Dr. Bohr-Gasse, A-1030 Vienna, Austria; and
Department of Immunology and Oncology, Centro Nacional de Biotecnologia-CSIC, Campus de Cantoblanco de la Universidad Autonoma, E-28049 Madrid, Spain
3Correspondence: Ernst W. Müllner, E-mail: em{at}mol.univie.ac.at Jose A. Garcia-Santz, E-mail: jasanz{at}cnb.uam.es
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: T cell activation • translational control • array hybridiza-tion
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. It later proved
to be a regulatory mechanism involved in a wide range of cellular
processes, including growth stimulation (2
, 3)
, cell cycle
progression (4)
, and differentiation (5
, 6)
.
Translational control affects either the global rate of protein
synthesis or is restricted to individual mRNAs or small subsets of
transcripts (7)
. Multiple signal pathways have been
implicated in global translational control and may lead either to
changes in the steady-state levels or the phosphorylation status of
eukaryotic initiation factors (eIFs; i.e., eIF-2
, eIF-4G, or eIF-4E)
(8
9
10)
or the activity of their specific regulatory
proteins (i.e., 4E binding proteins) (11)
. eIF-mediated
modulation of translational efficiency thus contributes massively to
the regulation of cell metabolism and proliferation.
Based on a limited but growing number of examples, a variety of
mechanisms have been described to account for translational control of
individual or small groups of mRNAs (7
, 12)
. These
mechanisms may involve interactions of regulatory elements within the
5'- or 3'-untranslated regions (UTRs) of the mRNA with specific RNA
binding proteins (13)
or even small RNAs (14
, 15)
. In addition, subcellular localization of the mRNA and
cytoplasmic modulation of the poly(A+)-tail
length may also regulate translation (16
17
18)
. Despite
this variety, almost all of these mechanisms act at the level of
translation initiation, which is the rate-limiting step in protein
synthesis (7
, 12)
.
Activation of T lymphocytes is central to the development of an immune
response against viral infections, intracellular parasites, allogeneic
cells, as well as for tumor surveillance (19)
. Resting T
lymphocytes are in the G0 phase of the cell cycle
and have a very low metabolic rate. T cell activation induces a series
of biochemical changes, including activation of the protein tyrosine
kinase pathway, mobilization of Ca2+ from
intracellular stores, and protein kinase C activation
(20)
. After these initial biochemical changes, there is a
seven- to tenfold increase of protein synthesis (21)
that
precedes the massive (30- to 40-fold) increase in mRNA synthesis
(J. A. Garcia-Sanz and E. W. Müllner, unpublished
results). As a consequence, the cells enter the cell cycle and undergo
limited proliferation for up to six to eight divisions
(22)
. The initial burst of protein synthesis correlates
with an increased availability of translation initiation factors
(8
, 23
, 24)
and with the mobilization of single ribosomes
to polyribosomes (25
, 26)
. It is a key event in the
activation process since a partial inhibition of translation delays and
depresses T cell activation (27)
.
In addition to this global change in translational efficiency, some
mRNAs are targets for specific translational control. For instance,
genes such as p56lck and Il2Rec chain contain
upstream open reading frames (5'UTR-ORF), which are responsible for
their low translation efficiencies (28
, 29)
. Furthermore,
some mRNAs have been shown to change their translation efficiency
between resting and activated T cells (rpL32; rpL7) (30
, 31)
or between T cells activated by different stimuli (IL-2)
(32)
. These cases do not represent isolated examples,
since it has recently been estimated that ~13% of the mRNA species
are translationally regulated upon T cell activation, including both
activated and repressed transcripts (33)
.
Here we describe an approach that allows the direct isolation of transcripts translationally regulated upon T cell activation. The method is based on a combination of sucrose gradient centrifugation to separate free mRNPs from polysome-bound mRNAs and a differential screening technique using either cDNA libraries or cDNA arrays. The high selectivity of the technique was confirmed by analysis of the differential ribosomal recruitment of representative mRNAs.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
.
Small resting T cells recovered from the 50–60% Percoll interphase
were resuspended (106 cells/ml) in DMEM-10
(Dulbecco’s modified Eagle’s medium supplemented with 10% heat
inactivated fetal calf serum, 20 mM HEPES, pH 7.0, 0.050 mM
ß-mercaptoethanol, and 2 mM glutamine). Cells were activated (48 h,
37°C, 5% CO2 in air) with a combination 10
µg/ml plate-bound anti-CD3 monoclonal antibody (mAb) (clone 66.1,
kindly provided by Dr. A. Lanzavecchia, Basel, Switzerland) and soluble
anti-CD28 mAb (mAb 248, 10-5 ascites dilution,
kindly provided by Dr. D. Olive, Marseille, France). Cells were then
transferred to new plates and cultured for an additional 48 h with
100 U/ml human recombinant interleukin-2 (hrIL-2, provided by Dr. A.
Lanzavecchia). Phenotypic analysis and activation status of the cell populations were monitored cytofluorometrically with mAbs specific for CD3, CD25, and CD69. The cell cycle was analysis was carried out by cytofluorometry of permeabilized cells stained with propidium iodide, and the proportion of cells in the various cell cycle phases were calculated using a software package from Coulter (Miami, Fla.).
Sucrose gradient fractionation and RNA analysis
Sucrose gradient fractionation was performed essentially as
described (34)
. Extracts from resting and activated cells
were prepared by lysis at 4°C in extraction buffer (10 mM Tris-HCl,
pH 8.0, 140 mM NaCl, 1.5 mM MgCl2, 0.5%
Nonidet-P40 and 500 U/ml RNAsin), and nuclei were removed by
centrifugation (12000 g, 10 s, 4°C). The supernatant
was supplemented with 20 mM dithiothreitol, 150 µg/ml cycloheximide,
665 µg/ml heparin and 1 mM phenylmethylsulfonyl fluoride and
centrifuged (12000 g, 5 min, 4°C) to eliminate
mitochondria. The supernatant was layered onto a 10 ml linear sucrose
gradient (15–40% sucrose [w/v] supplemented with 10 mM Tris-HCl, pH
7.5, 140 mM NaCl, 1.5 mM MgCl2, 10 mM
dithiothreitol, 100 µg/ml cycloheximide, and 0.5 mg/ml heparin) and
centrifuged in a SW41Ti rotor (Beckman, Palo Alto, Calif.) (38000 rpm,
120 min, 4°C) without brake. Fractions (550 µl) were collected and
digested with 100 µg proteinase K in 1% sodium dodecylsulfate (SDS)
and 10 mM EDTA (30 min, 37°C). RNAs were then recovered by
phenol-chloroform-isoamyl alcohol extraction, followed by ethanol
precipitation.
RNAs were either used as templates for cDNA synthesis and radiolabeling
or analyzed by electrophoresis on denaturing 1.2% formaldehyde agarose
gels and subsequent Northern blotting. After RNA transfer to nylon
membranes (GeneScreen, NEN, Boston, Mass.) and UV cross-linking, the
distribution of 18S and 28S rRNAs was visualized by methylene blue
staining of the filters (35)
. The membranes were
sequentially hybridized with various
[-32P]dCTP-labeled random-primed cDNA probes
or antisense-[-32P]CTP-labeled RNA probes.
After washing and autoradiography, signals were quantified by
PhosphorImaging (Molecular Dynamics, Sunnyvale, Calif.).
First strand cDNA synthesis and reverse strand priming labeling
From each sucrose gradient, fractions 1–9 (containing free
mRNPs and tRNAs) and fractions 12–19 (corresponding to polysome-bound
RNAs) were pooled. Poly(A+) mRNA was isolated
quantitatively from these pools using Oligotex-dT beads (Qiagen,
Hilden, Germany) as described (36)
. First strand cDNA
synthesis was performed using 40 µg/ml poly(A+)
mRNA (preheated at 65°C, 5 min) in 40 µg/ml oligo dT(12–18), 50 mM
Tris-HCl (pH 8.3), 75 mM KCl, 3 mM MgCl2, 10 mM
dithiothreitol, 1 mM dNTPs, and 2000 U/ml avian myeloblastosis virus
reverse transcriptase (Promega, Madison, Wis.) by incubation for 1 h at 37°C. The RNA strand within the DNA-RNA duplex was degraded with
0.3 M NaOH (1 h, 65°C), followed by neutralization of the sample with
0.3 M HCl, 10 mM Tris-HCl (pH 8.0) and 1 mM EDTA (pH 8.0)
(37)
. The reaction products were purified on a Sephadex
G-50 spin column (Pharmacia) and the first strand cDNA was quantitated
by fluorometry. The integrity and quality of the cDNA pools were
analyzed in agarose-TBE gels (data not shown). Twenty nanograms of
first strand cDNA were used to generate
[-32P]dCTP-labeled second strand cDNA, as
described (37)
. These probes were used for differential
screening of cDNA libraries and cDNA arrays.
cDNA library construction and differential hybridization
Poly(A+) mRNA (5 µg) from resting human
T lymphocytes was used to construct a cDNA library in lambda-ZAP XR
according to the manufacturer’s instructions (Stratagene, La Jolla,
Calif.). In total, 1.8 x 106 independent
phages were obtained from the packaging of cDNA inserts ligated to the
vector arms, of which 9 x 105 phages were
amplified to a titer of 3.3 x 1010
plaque-forming units (pfu)/ml. The average size of the bulk cDNA
inserts, tested by polymerase chain reaction, was in the range of 2.0
kb.
Aliquots from this library were seeded at a density of 1.2 x
103 pfu per plate (145x145 mm), incubated (12 h,
37°C), and phage DNA was then transferred to nylon membranes
(GeneScreen) by serial replica plating. To equalize DNA loading onto
the filters, transfer times were optimized to 30 s for the first
filter, 1 min, 2 min, and 8 min, leading to a deviation in DNA content
of less than 1.7-fold (data not shown). After UV cross-linking, the
membranes were hybridized with the
[-32P]dCTP-labeled second strand cDNAs
(generated from ribosome-free and polysome-bound mRNAs by RSP) in a
solution containing 5x Denhardt’s solution, 50% deionized formamide,
1% SDS, 10 mM EDTA, 12.5 mM
Na2HPO4, 12.5 mM
NaH2PO4, 5x sodium saline
citrate (SSC), and 200 µg/ml denatured and sonicated salmon sperm DNA
(16 h, 42°C). After washing sequentially in 2x SSC/1% SDS/5x
Denhardt’s at room temperature, in 2x SSC/1% SDS, and in 0.2x
SSC/1% SDS at 65°C, the filters were exposed to X-ray film at
-80°C using intensifying screens. If required, phages of interest
were purified to homogeneity in a second and/or third round of
differential hybridization. The 5' end of each selected cDNA was
sequenced after excision of the phagemid pBluescript SK(±)
(Stratagene). Approximately 200 bases were compared with EMBL and
GenBank databases using the FASTA sequence analysis program.
cDNA array differential hybridization and computer analysis
Gene expression analysis using four ATLAS human cDNA array
filters (Clontech, Palo Alto, Calif.) was performed essentially
according to the manufacturer’s protocol, except that RSP was used to
generate the radiolabeled probes from ribosome-free and polysome-bound
RNA pools. Hybridization signals for each cDNA (spot volumes) were
analyzed and quantified by PhosphorImaging (Molecular Dynamics).
Three algorithms were used to identify translationally controlled
transcripts. In the first one, the ribosome-free/polysome-bound ratio
in resting and activated T cells was calculated for each spot.
Transcripts with changes 
2 or 
0.5 in the ratio were considered as
translationally controlled candidates. Similar results were obtained
when ribosome-free/total ratio was calculated (the total is the sum of
ribosome-free and polysome-bound fractions).
In the second algorithm, a conditional test was performed in which cDNAs with a total(resting)/total(activated) ratio <0.9 were excluded due to transcriptional control. For the others, the ratios bound(activated)/bound(resting), free(resting)/bound(resting), and bound(activated)/free(activated) were calculated. cDNA clones with a value >1.0 for all three ratios were considered as translationally activated candidates. Conversely, when all three ratios were <1.0, the corresponding cDNAs were considered as translationally repressed candidates.
The third algorithm was similar to the second one, but normalized values were used. Each hybridization signal was normalized with the average intensity signal of the housekeeping genes from the same filter. Transcriptional and translational regulation with normalized values were calculated as described above. Candidates fulfilling the criteria from at least two algorithms were considered translationally controlled.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, and 2) primary resting T cells
can be activated in vitro in an in vivo-like
fashion with a combination of anti-CD3 and anti-CD28 mAbs
(38)
.
A typical starting population contained 80% T lymphocytes, as
determined by surface expression of the T cell receptor/CD3 complex
(TCR/CD3+). These cells were small resting T
cells, since they were devoid of activation markers (i.e., CD25, CD69),
and in the G0 phase of the cell cycle (Fig. 1
). T cells were activated in vitro with a combination of
anti-CD3 and anti-CD28 mAbs, which unlike antigen lead to a large
proportion of responding cells. Indeed, 16 h after stimulation,
although >99% of the cells were still in the
Go/G1 phase of the cycle,
81.7% of the cells expressed the CD69 activation marker and 57.3%
expressed CD25 (Fig. 1)
; virtually all had an increased volume (data
not shown). At later activation times, i.e., 96 h after
stimulation, most cells were activated since >95% were CD25 positive
and were actively proliferating, as assessed by cytofluorometry,
whereas at this time CD69 expression had returned to basal levels (Fig. 1)
. Cell populations from 0 h and 96 h after stimulation thus
represent two highly purified, distinct homogeneous populations
corresponding respectively to resting and activated T lymphocytes.
These populations were used for further molecular analysis.
|
Comparing resting and activated T cells, those transcripts that
redistribute in sucrose gradients are translationally controlled. For
this analysis, translationally inactive transcripts (free mRNP
particles or bound to a single ribosome, hereafter referred to as
ribosome-free) were separated from translationally active mRNAs
(polysome-bound) by centrifugation through linear 15–40% sucrose
gradients (34)
from both resting and activated T cells.
This fractionation yielded four different mRNA pools.
Poly(A+) mRNA was prepared from each pool, used
to synthesize oligo(dT)-primed first strand cDNA, and subsequently
radiolabeled by random priming. These complex probes were then used to
hybridize simultaneously four identical replicas of a cDNA library in a
differential screening procedure outlined in Fig. 2
. Translationally activated transcripts should give positive
hybridization signals with the replicas hybridized with the
ribosome-free pool of resting T cells and the polysome-bound pool of
activated T cells. The replica filters hybridized with the other two
probes should be negative or yield signals of lower intensity (Fig. 2)
.
The opposite hybridization pattern should account for translational
repression of an mRNA after T cell activation (Fig. 2)
. The same
differential screening approach was subsequently applied to cDNA array
screening.
|
Isolation of cDNAs coding for translationally controlled mRNAs by
differential screening of a cDNA library
Ribosome-free and polysome-bound transcripts from resting (0 h)
and activated (96 h after stimulation) T cells were used as substrates
to hybridize a human resting T cell library in the differential
screening described above. A representative hybridization is depicted
in Fig. 3
. Most of the clones were positive when hybridized with the four probes,
but a few code for translationally controlled mRNAs. Of 6000 phages
screened, the clones fulfilling the criteria for translational control
were isolated. Partial sequencing of these clones led to the
identification of 12 different cDNAs (see Table 1
).
|
|
The isolated clones encoded either mRNAs already known to be
translationally regulated, or mRNAs that had not been previously
described to be regulated at this level. The former group includes
cDNAs coding for translation elongation factor 1- (EF1; three
different clones) and ribosomal protein L31. Both cDNAs belong to the
family of 5'-polypyrimidine track-containing mRNAs, which are
translationally correlated through a common mechanisms
(39)
. As shown in Fig. 4D
, these transcripts are also translationally controlled
after T cell activation. Two clones encoding ferritin light chain,
which is known to be iron-regulated (40
, 41)
, were also
isolated. Other cDNAs belonging to this first group include thymosin
ß4 and HLA class II DR-ß. Thymosin
ß4 is translationally regulated upon thymocyte
stimulation and during the cell cycle (42)
; the mouse HLA
class II ß chain homologue IAß, as well as
its heterodimeric counterpart IA, are both
translationally controlled in macrophages and B cells
(43)
.
|
For several additional transcripts identified as being translationally
controlled, such regulation has not previously been described—namely,
mRNAs coding for the HLA class I molecules HLA-B and HLABw41 and their
counterpart, ß2-microglobulin; the
TNF--inducible protein A20; edg-2; ß-actin, and a 23 kDa human
basic protein (Table 1)
. Analysis of ß-actin mRNA redistribution in
polysome gradients showed mobilization of a peak of ß-actin mRNA
present in the ribosome-free fractions in resting T cells toward the
polysome-bound fractions in stimulated T cells (Fig. 4A, C
).
The 23 kDa human basic protein mRNA also redistributes from the
ribosome-free fractions toward the polysome-bound after T cell
activation (Fig. 4E
).
Identification of translationally controlled mRNAs by differential
screening of cDNA arrays
The same strategy was applied to differentially screen cDNA
arrays. These arrays (Clontech) contain ~600 filter-bound cDNAs of
several classes and allowed high through-put expression analysis.
Probes obtained from ribosome-free and polysome-bound mRNAs from
both resting and activated T cells were simultaneously hybridized to
four cDNA arrays. The four hybridization signals for each cDNA were
analyzed by PhosphorImaging. After discarding signals too close to
background, 472 cDNAs were further analyzed (the raw signal data can be
obtained at http://www.cnb.uam.es/~jasanz/Mikulits1999a). An
analysis of the intensity of the four hybridization signals of each
cDNA represented in the array identified 52 cDNAs that fulfilled the
criteria for translational regulation (Table 2
).
|
The translationally regulated transcripts encode proteins implicated in
a variety of cell functions. Redistribution of the cyclic AMP-dependent
activating transcription factor-4 (ATF-4, also known as DNA binding
protein TAXREB 67) on polysome gradients was analyzed in resting and
activated T cells. More than 40% of the ATF-4 mRNA migrates more
slowly than the 40S ribosomal subunit in resting T cells; this peak is
mobilized into polysomes in activated T cells (Fig. 5B, F
). In resting T cells, most of the
mitogen-activated protein kinase p38 (MAPK p38) mRNAs accumulate in the
ribosome-free fractions. After activation, there is a mobilization of
nearly 20% of the mRNAs toward the heavy polysome fractions (Fig. 5E, F
). Transcripts coding for transducin-ß2,
GADD153, heat shock protein 27, and neurotrophin-4 also redistribute on
polysome gradients after T cell activation (Fig. 5C, D, F
).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, but only a
few had been shown to be regulated at the translational level
(28
29
30
, 32
, 45)
. A recent analysis of the overall
contribution of translational control to the regulation of gene
expression demonstrated that whereas 36% of the mRNAs were
transcriptionally activated, 13% were translationally regulated during
antigenic activation of a T cell clone (33)
. The procedure
used in this study did not permit identification of the genes involved.
Here we demonstrate that transcripts undergoing changes in mRNA loading
of ribosomes after T cell activation can be identified directly by a
combination of polysome gradient fractionation and differential
screening of cDNA libraries or cDNA arrays (see Fig. 2
). Fifty-two cDNA clones of 472 analyzed from a cDNA array were identified as being translationally regulated. Twelve additional clones were isolated by screening of an aliquot of a cDNA library. These clones encode proteins with a broad range of functions, including transcription factors, metabolic enzymes, and cell signaling effectors, further demonstrating the prevalence of translational control as a regulatory mechanism of gene expression.
One premise of translational regulation is that the mRNA should be
expressed in both resting and activated T cells, and activation should
be accompanied by a redistribution of the transcript between
ribosome-free and polysome-bound fractions. Thus, we screened a resting
T cell library to exclude possible signal noise from those genes
transcriptionally induced after T cell activation. The screening of an
aliquot of the cDNA library represented a successful proof-of-principle
since several cDNAs encoding translationally regulated transcripts were
isolated. Nonetheless, this approach has several drawbacks. First, the
identification of the isolated cDNAs requires partial sequencing of the
clones. More important, the isolated clones represent only transcripts
of high or moderate abundance. The screening could be extended, at
least in theory, to cover a higher number of clones, but this would
lead mainly to an increase in the number of repeated isolations of the
same cDNA (ferritin L-chain was isolated twice and eEF1 three times
from 6000 random clones) rather than identification of novel, less
expressed genes. Isolation of rare transcripts present only in a few
copies per cell, and thus under-represented in any cDNA library, would
not be feasible unless an efficient normalization of the library was
carried out.
The same strategy, when applied to the screening of cDNA arrays, allowed us to overcome these drawbacks. PhosphorImager analysis allowed signal quantification over a range of four orders of magnitude, thereby allowing a high through-put analysis. In addition, the changes in total mRNA levels could be simultaneously determined. The main obvious drawback of the arrays is the choice of the cDNAs contained. This should not represent a major problem since genome-wide arrays should become available in the near future.
Quantitative analysis of the hybridization signals for each cDNA in the
array allowed us to determine the ribosome-free/polysome-bound ratio
for both resting and activated T cells. In the subsequent analysis,
three independent criteria were used to asses translational control.
They allowed us to overcome problems arising from 1) the
different complexities of the mRNA populations (the ribosome-free pool
represents 10–20% of mRNA species, whereas the polysome-bound
represents 80–90%) (33)
, and 2) the lack of
appropriate internal controls in the array (since cDNAs that are
supposed to serve as controls, such as ß-actin and 23 kDa highly
basic protein, are translationally regulated). The criteria are based
on changes in either the free/bound signal ratio or the free/total
ratio between resting and activated T cells using either the raw data
or values corrected for the total hybridization signal of the filter.
Only cDNAs fulfilling at least two of these criteria were considered as
translationally controlled (Table 2)
. One-third of these 52 identified
cDNAs fulfilled all three criteria.
Several of the 64 cDNAs isolated encode transcripts already known to be
translationally controlled, validating the reliability of the
differential screening approach. These include the eEF1 and
ribosomal protein L31, which contain a TOP (track of pyrimidines)
element in their 5'UTR. TOP mRNAs are translationally co-regulated
through a common mechanism during development, hormonal, and growth
stimulation (39)
and, as shown in Fig. 4D
,
after T cell activation. Ferritin-L chain mRNA was also found to be
translationally controlled upon T cell activation. The molecular
mechanism responsible for this regulation is unknown, but changes in
the activity of the mRNA binding protein IRP by activation and growth
factor deprivation of T cells have been reported (46
, 47)
.
These observations are compatible with the iron-dependent mechanism for
translational control (40
, 41)
. Thymosin
ß4, a major cytoplasmic protein (0.1–1% of
total cellular protein), binds monomeric actin to inhibit its
polymerization (48)
, and is translationally controlled
after thymocyte stimulation and during the cell cycle
(42)
.
HLA class I molecules and their counterpart
ß2-microglobulin, as well as HLA class II ß
chain cDNAs were also isolated. Translational control of the class II
mouse homologues IA and
IAß mRNAs has been recently shown upon
interferon 
(IFN) treatment of primary macrophages and B cell
lines (43)
. Immune cells up-regulate the expression
levels of major histocompatibility complex (MHC) proteins upon
activation. Once an infection is resolved, however, overproduction of
those molecules may lead to disease. Translational control may provide
a faster means than transcriptional repression to stop the synthesis of
MHC proteins, and thus is of great immunological interest.
Translational control had not yet been described for the vast majority
of identified cDNAs. Some of these cDNAs encode proteins known to play
an important role on T cell activation. For others, involvement in T
cell activation has not been reported. Among the identified cDNAs that
encode proteins with a clear role in T cell activation, there is the
MAPK p38, for which we could demonstrate translational control on T
cell activation through redistribution in sucrose gradients (Fig. 5E, F
). MAPK p38 is activated on TCR triggering
(49
50
51)
. Under those conditions, it has been implicated
in the phosphorylation of the eIF4E (52)
. The MAPK p38
signaling pathway involves several effectors, including MAPK-activated
protein kinase 3 (53)
and members of the c-Jun
NH2-terminal protein kinase family (JNK).
JNKs activate the transcription factors from the activating protein-1
(AP-1) family (50)
. In particular JNK2 plays an essential
role during T cell activation (54)
. In addition to MAPK
p38, both MAPKAP kinase 3 and JNK2 were identified as translationally
regulated (Table 2)
. This allows us to speculate that translational
control regulates not only individual transcripts, but also subsets of
transcripts encoding functionally related proteins.
Our data also show translational regulation of ß-actin (Fig. 4A, C
) and of the heat shock protein 27 kDa (HSP 27) mRNAs
(Fig. 5F
). One HSP 27 function is to regulate the
polymerization of actin microfilaments (55)
. This activity
is regulated by the MAPK p38 pathway (56)
. Translational
control of these molecules and thymosin
ß4 might be linked to the reorganization of the
actin cytoskeleton, known to play an important role in lymphocyte
activation (57)
.
Transducin ß-1 and ß-2 mRNAs were identified as translationally
regulated; translational activation was directly demonstrated for
transducin ß-2 (Fig. 5C, F
). These mRNAs encode two
of the ß subunits that are part of the G-protein heterotrimeric
complex (58)
. The ß subunits are responsible for
coupling of the G-protein to the seven-domain transmembrane receptors,
and specificity of the interaction may depend on the ß subunit
identity (59)
. A differential expression of the and
ß subunits has been suggested during lymphocyte development
(60)
. The induction of several transmembrane receptors,
including chemokine receptors (61)
, on T cell activation
might explain the cell’s need to increase the expression level of
G-proteins.
Gadd 153 mRNA was also shown to redistribute in polysome gradients
after T cell activation (Fig. 5D, F
). It encodes CHOP-10, a
protein related to the CCAAT/enhancer binding protein (C/EBP) family,
and regulates gene transcription through heterodimerization with other
C/EBP family members (62
, 63)
. CHOP-10 is strongly induced
after growth arrest and DNA damage (64
, 65)
.
Transcriptional regulation of Gadd 153 mRNA involves members from the
C/EBP family (66)
and the AP-1 family (67
, 68)
, including ATF-4 (69)
. In addition, CHOP-10
might be phosphorylated by the MAPK p38 (70)
. It is
involved in apoptosis and in the inhibition of cell cycle progression
(71
, 72)
. More recently, it has also been shown to play a
role in erythropoiesis (73)
. However, the role of CHOP-10
in T cell activation is not obvious. Similarly, the neurotrophic factor
neurotrophin-4, for which a role in inflammatory processes has been
suggested (74)
, is also translationally controlled on T
cell activation (Fig. 5F
).
Finally, translational control was also demonstrated for the ATF-4 mRNA
(Fig. 5B, F
), which belongs to the AP-1 transcription
factor family (75)
. Little is known about its function.
ATF-4 has been implicated in fibroblast response to anoxia, the
formation of eye lens and in the regulation of cyclin A
(76
77
78)
. Taken together, screening for translational
targets provided also new data concerning expression of these genes,
which might clarify their role during T cell activation.
Several other clones were analyzed for redistribution of their mRNAs in
polysome gradients after T cell activation. Unfortunately, in many
cases, including TFIIIC box B binding subunit, GAP-43, IFN,
transforming growth factor (TGFß) receptor type III, their expression
level—especially in resting T cells—was too low to allow a reliable
signal quantification after sucrose gradient fractionation. It is
noteworthy, however, that all mRNAs for which the analysis was feasible
redistribute in sucrose gradients, thereby validating our differential
screening strategy. The different profiles of the mRNAs analyzed (Figs. 4
, 5)
further suggest that these transcripts are regulated by different
mechanisms, supporting the widespread occurrence of translational
control.
Many genes are transcriptionally activated on T cell activation. In addition to the identification of translationally regulated genes, this approach allowed the simultaneous analysis of transcriptionally regulated mRNAs. As expected, genes like c-myb, the DNA-repair protein RAD52 homologue, the cAMP-response element binding protein CRE-BP1, the TGF-ß receptor III ERK2, c-Rel, Bax ß, leukosialin CD34, or thymosin ß 10 were transcriptionally activated (see further information on-line).
T lymphocyte activation was used as model system to validate the
screening approach. This technique can be applied to analyze other
physiological situations, such as cell cycle progression, oocyte
fertilization, differentiation and apoptosis. At the moment, cDNA
arrays are available only for a few species (mice, humans, rats, yeast,
etc.). This differential screening to identify translationally
regulated mRNAs in other organisms could also be carried out using
SSH (suppression subtractive hybridization) (79)
or SAGE
(serial analysis of gene expression) (80)
.
Most screening techniques, including subtractive hybridization
(81)
, differential display reverse
transcription-polymerase chain reaction (82
, 83)
, SSH
(79)
, and array analysis, have until now been applied
exclusively to identify mRNAs subject to massive variations in
abundance due to changes in either mRNA synthesis (transcription) or
cytoplasmic degradation of the transcript. Frequently, however,
regulation of the transcript level is not in itself sufficient to
account for variations in the levels of the corresponding protein. For
example, in yeast cells, changes in protein levels of up to 20-fold
were detected without corresponding alterations in mRNA abundance.
Conversely, 30-fold changes in mRNA levels were observed without
changes in protein levels (84)
. Thus, there is a
surprisingly poor correlation between protein and transcript levels in
eukaryotic systems (including yeast and mammalian cells)
(85)
. Yet it is the amount of protein and its activity
that determine cell phenotype. Taken together, this supports the notion
that the study of polysome-bound mRNAs rather than total mRNA levels
should be considered as the relevant parameters in ‘functional
genomics’ studies.
|
ACKNOWLEDGMENTS |
|---|
Note added in proof: Upon submission of this manuscript, a similar differential screening technique using cDNA arrays has been reported (Zong, Q., Schummer, M., Hood, L., and Morris, D. R. (1999) Proc. Natl. Acad. Sci. USA, 96, 10632–10636). The authors report that upon serum stimulation of resting fibroblasts, less than 1% of the mRNAs were translationally regulated. The discrepancy in their percentage of translationally regulated genes and the results presented here (>11%) are unclear, but may be due to the different cell systems used or the method of analysis.
|
FOOTNOTES |
|---|
2 Both authors contributed equally to this work and both should be regarded as senior co-authors.
Received for publication September 22, 1999.
Revision received January 26, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
expression in mitogen-activated primary T lymphocytes. EMBO J 9,3831-3837[Medline]
. J. Immunol. 161,1837-1843
