(Received for publication, November 19, 1996)
From the Department of Pathology and the § Department of
Oncology, The Institute of Medical Science, The University of Tokyo,
4-6-1 Shirokanedai, Minato-ku, Tokyo 108, Japan and the
Department of Immunology, Juntendo University, School of
Medicine, Core Research for Evolutional Science and Technology, Japan
Science and Technology Corporation, 2-1-1 Hongo, Bunkyo-ku,
Tokyo 113, Japan
Signals emanated from CD30 can activate the
nuclear factor B (NF
B). The two conserved subdomains, D1 and D2,
in the C-terminal cytoplasmic region of CD30 were tested for
interaction with two tumor necrosis factor receptor-associated factor
(TRAF) proteins with NF
B activating capacity, TRAF2 and TRAF5. TRAF5
is the newest member of the TRAF family that binds to lymphotoxin
receptor and CD40. TRAF5, as well as TRAF2, interacted with the D2
subdomain of CD30 in vitro and in vivo.
Deletion analysis by the yeast two-hybrid system revealed that the
C-terminal 22 and 30 amino acid residues are dispensable for
interaction of TRAF5 and TRAF2 with CD30, respectively. Substitution of
alanine for threonine at 463 abolished the interaction with TRAF2.
Overexpression of the TRAF domain of TRAF2 or TRAF5 showed a dominant
negative effect on CD30-mediated NF
B activation. Simultaneous
expression of these TRAF domains further suppressed the NF
B
activation, suggesting an interplay of these TRAF proteins. Expression
of TRAF2 and TRAF5 mRNA was demonstrated in T- and B-cell lines
that express CD30. Taken together, our results indicate that TRAF2 and
TRAF5 directly interact with CD30 and are involved in NF
B activation
by CD30 signaling.
CD30 is a member of TNF1 receptor
superfamily that comprises a group of cysteine-rich receptor proteins
such as CD27, CD40, and Fas antigen (1-4). Biochemical studies of CD30
as well as functional studies of the ligand for CD30 (CD30L) provided
strong evidence to support regulatory roles for CD30 in lymphocytes
(5-7). CD30L induces various biological effects on human CD30-positive cell lines such as activation, proliferation, differentiation, and cell
death, depending on cell type, stage of differentiation, transformation
status, and the presence of other stimuli (7). Recently, it was
reported that CD30-deficient mice showed impaired negative selection in
the thymus (8) and that CD30 is involved in signaling TCR-mediated cell
death of T-cell hybridoma (9). As for signal transduction of CD30,
Ellis et al. reported the induction of Ca2+
influx by cross-linking CD30 on Jurkat cells (10), and signals mediated
by CD30 were seen to regulate gene expression through activation of
NFB (11, 12).
Because the cytoplasmic tail of receptors of the TNF receptor family
does not have intrinsic catalytic activity such as kinase activity, it
was considered that molecules that associate with these receptors
mediate signal transduction. Putative signal transducing proteins that
associate with TNF receptor type II were cloned and named TNF
receptor-associated factor (TRAF) 1 and 2 (13). Subsequently, TRAF3 or
CRAF1 (CD40 bp, CAP1, LAP1) was identified as the CD40 signal
transducing molecule (Refs. 14-17; reviewed in Ref. 18). We have
recently cloned TRAF5 that associates with lymphotoxin receptor and
CD40 and mediates NF
B activating signals (19, 20).
We have found that an approximately 100-amino acid sequence of the
C-terminal region of the CD30 cytoplasmic region was highly conserved
among human, rat, and mouse CD30 protein, and in this region there are
two subdomains with a higher conservation (D1 and D2) (21). Recently,
the association of TRAF1, 2, and 3 with the cytoplasmic tail of CD30
has been reported, and the binding sites were characterized (22).
However, it has remained to be determined whether and how these TRAF
proteins are involved in NFB activation by CD30. To better
understand the CD30 signal transduction pathway that leads to NF
B
activation, we asked whether TRAF5, as well as TRAF2 are involved in
CD30 signaling. We found that TRAF5, as was shown for TRAF2, directly
associates with the D2 subdomain of the CD30 cytoplasmic tail, and both
TRAF2 and TRAF5 are involved in the signal transduction leading to
NF
B activation.
GST fusion protein was prepared using the pGEX series of GST fusion protein vectors (Pharmacia Biotech Inc.). The cDNA fragments encoding regions covering both or either of the two conserved subdomains (D1 and D2) of the CD30 cytoplasmic tail of human (2), mouse (23), and rat (21) were prepared by PCR using primers with extra sequences for restriction enzyme digestion. These were named GST-hCD30 (D1+D2) (amino acids 459-595), GST-mCD30 (D1+D2) (amino acids 388-498), GST-rCD30 (D1+D2) (amino acids 383-493), GST-rCD30D2 (amino acids 426-493), and GST-rCD30D1 (amino acids 383-445). GST and GST fusion proteins were prepared by standard methods (24). In vitro transcription and translation was performed using the TNT in vitro translation system (Promega) according to the manufacturer's instruction. In vitro binding analysis was done as described previously (19). The gels were dried and analyzed on a BAS 2000 image analyzer (Fuji Film).
In Vivo Binding AssayTransient cotransfection and
coprecipitation analyses were done using 293T cells. For these
experiments, an expression vector for human CD30 (pME-hCD30) was
prepared using an SR-driven pME18S plasmid (a gift from Dr. K
Maruyama). Expression vectors for human CD30 with C-terminal deletions
of 47, 95, and 132 amino acids were also prepared using PCR and pCR3TM
eukaryotic TA cloning vector (In vitrogen), and named pCR-hCD30(
D2),
pCR-hCD30(
D1+D2), and phCD(
132), respectively. These deletion
mutants lack the D2 subdomain or both D1 and D2 subdomains at the C
terminus (21), respectively. Those for mouse TRAF2 and TRAF5 tagged
with FLAG were prepared using pME18S and PCR-amplified cDNA
fragments encompassing the protein coding region. The resultant
plasmids were named pMEFLAG-TRAF2 and pMEFLAG-TRAF5, respectively. The
293T cells were cotransfected with 1 µg each of expression vectors of
these human CD30 and FLAG-tagged TRAF5 or TRAF2, using LipofectinTM
reagent (Life Technologies, Inc.). The cell lysates were first
immunoprecipitated with anti-CD30 monoclonal antibody BerH2 (DAKO), and
the immnunoprecipitates were analyzed by immunoblotting with anti-FLAG
M2 antibody (Eastman Kodak) or immunoprecipitated with anti-FLAG M2
antibody and detected by BerH2 using an ECLTM chemiluminescence
detection kit (Amersham Corp.).
For cotransformation assays in a
yeast two-hybrid system, DNA encoding the rat CD30 intracellular domain
(amino acids 308-493) (21) was cloned into the yeast Gal4-DNA binding
domain vector pAS1-CYH2 (a gift from Dr. J. W. Harper) (25)
(pAS-rCD30IC) and used to test interaction with the TRAF family
proteins. Bait plasmids with a series of C-terminal deletion of the
CD30 cytoplasmic region were prepared to map the binding regions. These
plasmids were named pAS-rCD3022, pAS-rCD30
30, pAS-rCD30
36,
pAS-rCD30
51, pAS-rCD30
67, pAS-rCD30
81, and pAS-rCD30
97,
where the numbers following
indicate the number of deleted amino
acids. Gal4 activation domain plasmids having TRAF domains of TRAF2 or
TRAF5 were pACT-TRAF2D (amino acids 274-501) and pACT-TRAF5D (amino
acids 223-558).
Expression plasmids of mouse TRAF2 and
TRAF5 were prepared using the pCRTM3 eukaryotic TA cloning vector
(Invitrogen) and PCR-amplified cDNA fragments encompassing the
protein coding region. The resultant plasmids were named pCR-TRAF2 and
pCR-TRAF5, respectively. Using the same method, those for truncated
TRAF proteins retaining only the TRAF domain were prepared. The
resultant plasmids (pCR-TRAF2D and pCR-TRAF5D) encode amino acids
269-501 and 233-558, respectively. For CAT assays, 1 µg of the
reporter plasmid, [B]6TK-CAT or
[
Bm]6TK-CAT (20), 1 µg of
-galactosidase
expression vector driven by the
-actin promoter (
-actin-
-gal),
and 1 µg of pME-hCD30, pCR-TRAF2, pCR-TRAF2D, pCR-TRAF5, or
pCR-TRAF5D were cotransfected with LipofectinTM reagent (Life
Technologies, Inc.). The total amount of DNA transfected was always
adjusted to 5 µg with control expression vectors, pME18S, or pCRTM3.
40 h after transfection, cell extracts were prepared by three
cycles of freeze-thawing followed by centrifugation.
-Galactosidase
activity was used to standardize transfection efficiency. CAT assays
were performed using a fluor diffusion method (26).
Signaling from CD30 was shown to activate NFB; however,
transducers of this signal have not been well characterized. On the other hand, two TRAF proteins, TRAF2 and TRAF5, were proven to have
NF
B activating capacity. Therefore, study of the involvement of
TRAF2 and TRAF5 in NF
B activation by CD30 is needed to better understand the biological functions of CD30. We reported that an
approximately 100-amino acid sequence is highly conserved between rat
and human CD30, and we have tentatively divided this region into two
subdomains, D1 and D2 (21). These observations also held true for the
mouse CD30 (23), thereby suggesting the functional importance of these
subdomains (Fig. 1A). Moreover, it was shown that TRAF1, 2, and 3 interact with CD30 through binding sites in
C-terminal 36 amino acids (22), which signifies that these TRAF
proteins bind to the region containing the C-terminally located D2
subdomain. To determine if TRAF5 binds to CD30 and if so which of the
subdomains are involved in the binding, we examined interactions of
TRAF5, along with TRAF2, with CD30 in vitro and in
vivo. First we examined in vitro binding using the CD30
cytoplasmic region fused to glutathione S-transferase and
in vitro translated TRAF proteins. We constructed GST fusion
proteins of CD30 having both or either one of them (Fig.
1A). GST fusion proteins with two subdomains were
GST-hCD30(D1+D2) that has the C-terminal 136 amino acids of human CD30,
and GST-mCD30(D1+D2) and GST-rCD30(D1+D2) that have C-terminal 110 amino acids of mouse and rat CD30, respectively. GST-rCD30 (D2) and
GST-rCD30 (D1) have either of the subdomains of the rat CD30. Affinity
purified GST fusion proteins were tested for their potential to
precipitate 35S-labeled in vitro translated
full-length mouse TRAF2 and TRAF5. As shown in Fig. 1B, both
TRAF2 and TRAF5 were precipitated by GST-CD30 proteins having both D1
and D2 subdomains of human, mouse, or rat origin, they interacted with
the CD30 GST fusion protein possessing the D2 subdomain but not with
that having the D1 subdomain. Thus, it was suggested that TRAF5, as
well as TRAF2, interacts with CD30 through the binding site(s) in the
D2 subdomain.
To confirm this proposal, we next examined the interaction of CD30 with
TRAF5 and TRAF2 by cotransfection and coprecipitation analysis. Human
CD30 and deletion mutants were coexpressed with FLAG-tagged TRAF5 or
TRAF2 in 293T cells. CD30 proteins were immunoprecipitated from cell
lysates using the BerH2 anti-CD30 monoclonal antibody (DAKO), and the
immune complexes were analyzed for the presence of FLAG-TRAF5 or
FLAG-TRAF2 by immunoblotting using anti-FLAG M2 antibody (Eastman
Kodak). The BerH2 anti-CD30 antibody co-immunoprecipitated TRAF2 and
TRAF5 from the cells transfected with wild type CD30 cDNA. However,
deletion of the D2 subdomain or both D1 and D2 subdomains abrogated the
precipitation of either TRAF2 or TRAF5 (Fig. 2,
upper panel). Conversely, the anti-FLAG M2 antibody
co-immunoprecipitated only the wild type CD30 but not the deletion
mutants with co-expressed TRAF2 or TRAF5 (Fig. 2, lower
panel). Expression of the transfected CD30 proteins, as well as
FLAG-tagged TRAF 2 and TRAF5, was confirmed by immunoblotting with
BerH2 or anti-FLAG M2 antibody (data not shown). Taken together, these
results indicated that TRAF5 can interact with the CD30 cytoplasmic
tail through the C-terminal region containing the D2 subdomain in
vivo.
To further map the binding sites of CD30, deletion studies were done
using the yeast two-hybrid system. The yeasts were cotransformed with
plasmids that express wild type and various C-terminal deletion mutants
of rat CD30 along with TRAF2 or TRAF5 expression vector and then
assayed for their potential to grow on medium lacking leucine and
tryptophan and for -galactosidase activity. The TRAF domain of TRAF5
as well as that of TRAF2 interacted with the cytoplasmic domain of CD30
(Fig. 3A). However, deletion of the
C-terminal 30 amino acids that removes one-third of the D2 subdomain
abolished the binding to TRAF5 protein, whereas this construct retained the potential to interact with TRAF2 protein. Deletion of 36 amino acids that removes half of the D2 subdomain, as well as deletion of the
whole D2 subdomain, abolished the interaction with both the TRAF
proteins. Gedrich et al. identified two binding sites of the
human CD30 for TRAF1, 2, and 3, the PEQET and EEEGKE sequences, that
are located at 462-466 and 480-485, respectively (22). Thus, the
above results indicated that the PEQET sequence alone is sufficient for
interaction with TRAF2 but not with TRAF5 and suggested that the
EPPLGSC sequence (464-470) is required for TRAF5 binding. A similar
sequence PXQXT is also present in the binding
domain of CD40 to TRAF3.
The threonine residue at position 254 of CD40 was shown to be important
for binding to TRAF3 (14) and for biological functions. We next
examined the effect of amino acid substitution at 463 for interaction
with TRAF2. Substitution of alanine for threonine deprived the mutant
protein CD3030A of the potential to interact with TRAF2 (Fig.
3B), suggesting that the threonine residue in the PEQET
binding sequence is essential for biological activity of CD30 and that
this threonine residue may have functional similarities with threonine
254 in CD40 (27).
Because we have previously demonstrated that TRAF5 can activate NFB
(19, 20), the results in the present study indicated that both the two
TRAF proteins with NF
B activating capacity, TRAF2 and TRAF5,
interact with CD30. Because signals from CD30 have been shown to
activate NF
B and induce gene expression (11, 12), we asked if TRAF2
and TRAF5 are involved in NF
B activation. We performed a reporter
gene assay using [
B]6TK-CAT and
[
Bm]6TK-CAT plasmids and expression vectors for CD30
and full size or N-terminally truncated TRAF2 or TRAF5 that lacks ring
finger and zinc finger domains. We first confirmed that overexpression
of TRAF2 or TRAF5 alone resulted in a 4.1- or 4.8-fold induction of CAT
activity, in a
B site-dependent manner (Fig.
4A). On the other hand, the overexpression of
CD30 alone showed about a 9-fold induction of CAT activity, also in a
B site-dependent manner (Fig. 4B).
Cotransfection of CD30 and the TRAF domain of TRAF2 or TRAF5 suppressed
CD30-mediated NF
B activation by 66 or 39%, respectively (Fig.
4B), indicating the dominant negative effect of TRAF domain
overexpression. Simultaneous expression of TRAF domains of TRAF2 and
TRAF5 further suppressed CD30-mediated NF
B activation by 80%,
thereby suggesting that both of these proteins are involved in signal
transduction of CD30.
CD30 is expressed in activated T-cells as well as lymphocytes infected with human T-cell lymphotrophic virus-1 or EBV; therefore, we examined the expression of TRAF2 and TRAF5 in lymphocytes by Northern blot analysis. A panel of B- and T-lymphocyte cell lines, some of which are infected with human T-cell lymphotrophic virus-1 or EBV, were included in the study. We found that TRAF2 and TRAF5 transcripts were expressed in all the cell lines tested. CD30 mRNA was also expressed in these cell lines, except for EBV noninfected BJAB and Ramos cells (data not shown). Colocalization of expression provided further supportive evidence for the involvement of TRAF5 and TRAF2 in CD30 signal transduction.
These results demonstrate for the first time the involvement of TRAF2
and TRAF5 in NFB activation by CD30. Direct interaction of TRAF5
with the C-terminal end of the CD30 protein and involvement of TRAF2
and TRAF5 in CD30-mediated NF
B activation underline the complexity
of interplay between the TRAF family proteins in the signal
transduction of CD30 and may explain the pleiotropic biological
activity mediated by CD30-CD30L interaction. Stoichiometric studies of
the interaction will be needed to analyze the interplay of TRAF
proteins and determine the roles played by individual TRAF proteins
during CD30 signal transduction.
We thank M. Ohara for helpful comments on the manuscript and Carl F. Ware for sharing unpublished results.