Mydj2 as a potent partner of hsc70 in mammalian cells
Petros Bozidis, Ioannis Lazaridis, Gerassimos N. Pagoulatos and Charalampos E. Angelidis
Laboratory of General Biology, Medical School, University of Ioannina, Greece
Dj2 is a member of the DnaJ family of proteins, which
regulate the chaperoning function of the hsp70s. We isolated
a m onkey c DNA d j2 clon e c orresponding to the large
mRNA species encoded by t he gene. T his mRNA differs
from the small mRNA produced by the same gene i n that it
contains a l ong 3¢ untranslated region. Both messages were
found to be equally stable and t o produce the same protein,
which is s usceptible t o f arnesylation. Studies in mouse tissues
and various cell lines revealed that these messages and their
products are differentially expressed. Surprisingly, we found
that only the nonfarnesylated form of dj2 is capable of
translocating to the cell nucleus, especially after h eat shock.
Finally, b ased on protein interaction studies, our results
indicate that dj2 is a specific partner for hsc70 and not
for h sp70.
Keywords: DnaJ homologue; dj2; h eat s hock; cochaperone;
nanomachine.
It is widely accepted today that the eukaryotic DnaJ
homologs consist a family of proteins which in combination
withthehsp70familymembersmakeupthebasicmolecular
chaperone machinery of the mammalian cell [1]. The
members of the above g ene families work together in a
variety of cellular p rocesses, including p rotein folding during
which the hsp70s bind unfolded, partially folded or dena-
tured polypep tide substrates and assist their renaturation
through a cycle o f binding and release regulated b y their
DnaJ cochaperones [2,3]. Based on the existense of three
dinstict domains, namely: the highly conserved J domain

facilitates the early steps of transmembrane receptor
biogenesis in cystic fibrosis [20] and is mobilized to the
nucleus in order to refold m isfolded receptors into biolo-
gically active conformation states [17]. Finally, overexpres-
sion of dj2 was recently found to decrease aggregate
formation caused by expanded polyglutamine tracts, a
hallmark of neurodegenerative diseases [21,22].
In the present study, we isolated and characterized a
cDNA clone from Cercopithecus aethiops (monkey) cells,
Mydj2, whic h is similar to mouse Hsj2 [23]. This cDNA
corresponds to an ortholog of the DNAJA1 gene according
to the nomenclature suggested by Ohtsuka & Hata [24] it
should be named caDjA1. Comparison of Mydj2 and Hsj2
with the HDJ2 [7] and hdj2 [8] cDNAs, showed that our
clone although similar to Hsj2, has an extended 3¢ noncod-
ing region of 981 bp. To determine w hether the additional
sequences influence the stability of the RNA, as previously
reported for other RNAs [25] studies addressing this
question were performed and showed that there is no
difference in the stability of t he two d j2 mRNAs. Further-
more, w e investigated the in vivo properties of the endo-
genous Mydj2 in mammalian cells. We found that only the
nonfarnesylated form of d j2 translocates to the nu cleus
especially after heat shock and that dj2 binds only to the
constitutive form of hsp70, namely hsc70.
MATERIALS AND METHODS
CDNA library screening
A cDNA library, prepared using RNA isolated from COS
cells was obtained from Stratgene (monkey COS cell line
cDNA k ZAP

treated by immersing the culture dishes in a water bath set at
the desired temperature.
Sub-confluent control or heat treated cells were harves-
ted, washed with NaCl/P
i
and resuspended in 300 lLRIPA
buffer (50 m
M
Tris/HCL, 150 m
M
NaCl, 1% Triton X-100,
1% sodium deoxycholate, 0.1% SDS) with 1 lgÆmL
)1
pepstatin, 1 lgÆmL
)1
leupeptin, 1 m
M
phenylmethanesulfo-
nyl fluoride and 10 UÆmL
)1
apyrase. Lysates were prepared
after incubation of the cell s uspension on ice for 1 0 m in.
During this period the lysates were homogenized by passing
five or six times t hrough a 21-guage needle, f ollowed by
centrifugation at 12 900 g for 10 min at 4 °C (Eppendorf).
The supernatants were mixed with SDS-sample buffer at
final concentration (62.5 m
M
Tris/HCl, pH 6.8, 5%
2-mercaptoethanol, 3% S DS, 10% glycerol, 0.1% bromo-

frame with 6 ·His, into the BamHI and SalIsitesofthe
pQE-32 plasmid (Qiagen, GmbH, Germany) resulting in
the generation of the pQE-32-Mydj2 plasmid. The accuracy
of the resu lting construct ( pQE-32-Mydj2) was v erified by
DNA sequencing, and the plasmid was subsequently used to
overexpress M ydj2–6·His in Escherichia coli.
Expression and purification of histidine-tagged Mydj2:
antibody production
The p QE-32-Mydj2 plasmid was used to overexpress the
Mydj2 protein. Overnight cultures of E. co li JM109 carry-
ing pQE-32-Mdj2 plasmid were diluted 10-fold and cultured
for 1 h. After isopropyl thio-b-
D
-galactoside induction
(2 m
M
)for2hat37°C, the cells were collected by brief
centrifugation and cell lysates were prepared by sonication.
The recombinant prote in was purified from the cell lysates
using a Ni/nitrilotriacetic acid column and imidazole elution
(50–250 m
M
) as described by the manufacturers (Qiagen).
Anti-Mydj2 antibodies were obtained by injecting a male
rabbit with purified Mydj2 protein [30].
Protein–protein interactions experiments
For immunoprecipitations, cell lysates prep ared in RIPA
buffer were incubated overnight at 4 °C by end-over-end
rocking with 5 lL of hsc70-specific antibody (SPA-815,
StressGene), 3 lL hdj2-specific antibody, o r 3 lLanti-

i
and fi xed
for 1 0 m in, at room temperature i n 2% p araformaldehyde.
The cells were washed three times with cold NaCl/P
i
and permeabilized by incubating in ice cold, absolute
methanol for 3–5 min at 20 °C. Then, the cells were washed
three times with cold NaCl/P
i
and i ncubated i n 3% BSA in
1554 P. Bozidis et al. (Eur. J. Biochem. 269) Ó FEBS 2002
NaCl/P
i
to prevent nonspecific staining. After 1 h of
incubation with blocking medium, the cells were washed
three times with NaCl/P
i
and incubated for 1 h with
20–30 lL of the primary antibody diluted as indicated in
NaCl/P
i
/3% BSA. Following washings (three times),
20–30 lL of the secondary antibody, fluorescein isothio-
cyante-conjugated goat anti-(rabbit IgG) Ig diluted 1 : 25 in
NaCl/P
i
, was added. After 1 h of incubation the cells on the
coverslip were washed three times with NaCl/P
i
, placed at

i
and total RNA w as prepared. Northern b lot analysis and
detection of both messages, large dj2 and small dj2 mRNAs,
were performed as described previously [34].
RESULTS
Isolation of a cDNA clone encoding for Mydj2
Using the methodology described above we isolated a
2.3-kb full length monkey cDNA which encodes a DnaJ
homologue (GenBank accession number: AF395203) and
more specifically the ortholog of the human dj2 protein. As
a c DNA source we used the premade by Stratagene cDNA
library of monkey COS cells which was screened using the
entire hum an dj2 cDNA clone [8]. Among several positive
clones, one clone (5aI) with a 2.2-kb insert was isolated.
Nucleotide sequence analysis revealed that 5aI clone had a
single open reading frame of 397 amino-acids beginning
with the A TG codon at nucleotide 36–38 and terminating
with the TAG codon at nucleotide 1228–1230. Comparison
of our clone ( 5aI) with the human dj2 cDNA [7,8] showed
that clone 5aI stems from a larger m RNA having 0.9-kb
additional sequences at the 3¢ UTR ( Fig. 1A).
Further sequence comparison between the isolated mon-
key dj2 cDNA (clone 5aI) a nd the human dj2 cDNA
showed an identity of 99% for the J domain, 100% for the
G/F region, 99% for the cysteine rich domain (cysteine rich
region) and 97% for the C-terminus (Fig. 1A).
Differential expression of large and small dj2 mRNAs
The distribution of dj2 RNAs in different cell lines was
initially studied. We used as DNA probes the entire coding
region of the Mydj2 gene (Fig. 1A), which can hybridize to

or tissue t ype. The significanse of the above finding remains
to be clarified.
We then examined the distribution of t he corresponding
dj2 protein in the s ame rat and mice tissues (Fig. 2B,C,D).
For this study we used two d j2 specific antibo dies, one wh ich
Fig. 1. Characterization and c omparison of l arge and small dj2 mRNAs.
(A) Schematic re presentation and sequence comparison of t he m onkey
large dj2 cDNA and human small dj2 cDNA (B). Northern blot
analysis of the two forms o f d j2 m RNAs, using RNAs from c ontrol ( –)
or heat treated for 90 min at 43 °C and 90 min recovery to 37 °C(+)
monkey COS, human HeLa, mouse teratocarcinoma F9 and human
histocytic lymphoma U-937 cell lines. The po sition of 18S rRNAs is
shown at the bottom.
Ó FEBS 2002 Monkey dj2, a specific partner for hsc70 (Eur. J. Biochem. 269) 1555
recognizes the entire monkey dj2 protein (see Materials and
methods) a nd another that recognizes only the N-terminal
end (1–179 amino a cids) o f t he human dj2 protein
(Neomarkers, cat. no. MS225P). Samples from rat (Fig. 2B)
and mouse tissues (Fig. 2C,D) contained the same amounts
of protein were used for this study. Dj2 protein l evels were
found to be particularly high in testis, brain, kidney and
liver. On th e other h and, tissues like the heart, muscle and
lung revealed lower but detectable amounts of dj2 proteins
(Fig. 2B,C,D). Unexpectedly, under our experimental con-
ditions, the identification of the farnesylated and the
nonfarnesylated dj2 forms was not possible regardless of
the type of antibodies and tissues used. However, one
particular feature of dj2 expression was the possible
existence of farnesylated forms or isoforms of dj2 in testis,
which were identified only w ith our anti-dj2 Ig (Fig. 2B,C).

approximately 46–53 kDa, as expected (data not shown).
The dj2 protein produced was found to be susceptible to
farnesylation and the inhibition of farnesylation in CV1
Fig. 2. Differential expression of dj2 in mouse and rat tissues. (A) RNA
blot analysis was performed for dj2 in mouse tissues. Total RNAs
(20 lg)fromlung,brain,testis,kidney,heart,spleen,muscle,liverand
from COS and HeLa cells, were analyzed by Northern blotting, using
as prob e the
32
P-labeled c DNA for monkey dj2. The position of 18S
rRNAs is shown at the bottom. (B,C,D) Dj2 protein distribution in
tissues of Wistar rats (B) and mice (C,D). Tissue total cell extracts were
obtained as described in Materials and m ethods. Equal amo unts of
proteins were analyzed by immunoblotting using specific antibodies
for the entire dj2 protein (B,C) and for the N-terminal fragment of dj2
(D). P , denotes the purified rec ombinant Mydj2 prot ein. The lower
band observed in testis probably represents a testis spe cific dj2
ortholog or a modified form of dj2.
Fig. 3. Both large and small dj2 mRNAs are stable molecules. COS and
HeLa cells were treated with 10 lgÆmL
)1
actinomycin-D for 0, 2, 4, 16
and 24 h. (A) Parallel cultures were treated in the same way, with
actinomycin D and exp osed in heat shock for 90 m in at 43 °C.
(B) Then, RNAs were prepared and 2 0 lg of each sample were
subjected to RNA blot analysis using the entire Mydj2 cDNA, radio-
labelled with [c
32
P]dCTP, as a probe. Integrity o f RNAs was verified
by the a ppare ntly identical intensities of 1 8S rRNAs.

Therefore, we concluded, that under our experimental
conditions, dj2 is diffused in the entire cell and only t he
nonfarnesylated form is translocated to the nucleus.
Association of cochaperones Mydj2 with Myhsc70
The partner selectivity between chaperones and cochaper-
ones is not entirely clear. To further define how hsc70 and
DnaJ-like proteins interact, we decided to use two meth-
odological approaches, one involving immunoprecipitations
and another involving a modified pull-down assay.
CV1 or COS cells were exposed at 42.5 °C for 90 min and
recovered at 37 °C for 90 min. RIPA cell extracts from
control or heat-shocked cells were then prepared and used in
pull-down experiments. More specifically, the lysate from
2.5 · 10
6
cells was mixed with Mydj2-His purified recom-
binant protein immobilized on Ni-nitrilotriacetic acid resin
(Qiagen). After incubation, extended washings and centri-
fugation of the c oprecipitated proteins, all fractions were
subjected to SDS/PAGE and Western blotting analysis. As
shown in Fig. 6, hsc70 was found to bind to immobilized
Mydj2-His protein (lanes 3, 3¢). In contrast, no binding
between the immobilized dj2 protein and the inducible hsp70
protein was observed (Fig. 6, lanes: 3, 3¢). Interestingly, the
same results were ob tained when lysates from heat shock ed
cells were used (Fig. 6, lanes: 6 , 6¢) despite the fact that in this
case the levels of hsp70 were substantially elevated (compare
lanes 1 with 4 and 1 ¢ with 4¢ in Fig . 6) as it was expected.
Fig. 5. The farnesylated d j2 form translocates to the cell nucleus. CV1
cells growing under physiological conditions (37 °C) or heat treated f o r

The a bove results clearly demonstrate t hat Mydj2 binds
specifically to hsc70 and that this binding is not susceptible
to changes under elevated temperatures. Because the Mydj2
is associated only with the constitutive Myhsc70 we suggest
that these proteins constitute possible partners in the
construction of a cellular chaperoning functional unit
referred to as a chaperoning nanomachine.
DISCUSSION
It is known that members of the different chaperone families
are interweaved or combined in order to organize nanoma-
chines. For example, the hsp70 family requires cofactors for
specifying its functions. A major group of these partners
belong to the DnaJ family [5,8]. Little is known about the
specific c ombination and regulation of all these nanoma-
chines. However, we know that chaperones play an essential
role in various cellular functions, such a s the acquisition o f
thermotolerance and ce ll survival [26,36,37], the protection
from ischemic injury [38] and in various human disorders
[39–43].
The data presented in this report describe the features of a
member of the 4 0-kDa hsp family, the monkey dj2 protein.
This member has all the appropriate domains that classify it
as a member of the orthodox DnaJ subfamily. The isolated
clone 5aI g ives an open reading frame of 1191 bp, which is
able to produce a polypeptide of 397 amino acids. Com-
paring our clone with the reported human dj2 [7,8] and
mouse dj2 [23] clones, clone 5aI appeared to be similar to
the mouse dj2 clone. Using clone 5aI as a p robe and RNAs
from cell lines, w e identified two different in size mRNAs.
The s teady-state levels of both messages were examined in

the binding substrate. Cell extracts, washes and eluted protein samples
were analyzed by Western blotting using specific a ntibodies against
hsp70, hsc70 and Mydj2. 1,4,1¢,4¢:cellextracts,2,5,2¢,5¢: third wash-
ings, 3,6,3¢,6¢: eluted p roteins.
1558 P. Bozidis et al. (Eur. J. Biochem. 269) Ó FEBS 2002
According to our findings only the nonfarnesylated form of
dj2 is a ble to t ranslocate into the nucleus. In contrast, the
farnesylated form remains localized to the cytosolic fraction.
Moreover during heat shock (90 min at 4 3 °Cand60min
recovery to 37 °C), the farnesylated dj2 protein translocates
mostly to the nucleus suggesting that t his migration is related
to the facilitation of t he folding of the he at denatu red nuclear
proteins. This r esult is i n agreement with previous observa-
tion which s uggests t hat t he HDJ2 protein is mobilized to the
nucleus in response to the presence of inappropriate folded
mutated receptors [17].
Recent studies revealed that hsc70 and dj2 constitute a
potent chaperone pair that is required for mitochondrial
import of preornithine transcarbamylase and refolding o f
denatured l uciferase [10] or unfolded mutated receptor [17].
In order to verify, in vitro and semi in vivo, the existence of
this functional pair, we performed pull-down and immu-
noprecipitation experiments. In pull-d own assays using
recombinant Mydj2 fused t o 6·His and immobilized to
Ni-nitrilotriacetic acid agarose beads, the binding of
Myhsc70 with Mydj2 was obtained. Under the same
conditions hsp70 did not coprecipitate with d j2, which
means that only hsc70 and dj2 can be combined to form a
functional pair.
Having identified the Myhsc70 as the direct interaction

hsp40, a human homologue of bacterial DnaJ. Biochem. Biophys.
Res. Commun. 19 7, 235–240.
6. Raabe, T. & Manley, J.L. (1991) A human homologue of the
Escherichia coli DnaJ heat shock protein. Nucleic Acids Res. 19,
6645.
7. Chellaiah. A., Davis, A. & Mohanakumar, T. (1993) Cloning of a
unique human homologue of the Escherichia coli DNAJ heat
shock protein. Biochim. Biophys. Acta 1174, 11–113.
8. Oh, S., Iwahori, A. & K ato, S. (1993) Human c DNA encoding
DnaJ protein homologue. Biochem. Biophys. Acta. 1174, 114–116.
9. Cheetham, M.E., Brian, J.P. & A nderton, B.H. ( 1992) Human
homologues of the bacterial heat-shock protein DnaJ a re pre-
ferentially expressed in neurons. Biochem. J. 284, 469–476.
10. Terada, K. & M ori, M. (2000) Characterization of human DnaJ
homologs dj2 and dj3, and bag-1 are positive cochaperones of
hsc70. J. Biol. Chem. 275 , 24728–24734.
11. Kanazawa, M., Terada, K., Kato, S. & Mori, M. (1997) HSDJ, a
human homolog of DnaJ, is f arn esylated and is involved in protein
import into mitochondria. J. Biochem. 121, 890–895.
12. Szabo, A.R., Korszun, F.U., Hartl & Flanagan, J . (1996) A zinc
finger-like domain of the molecular chaperone DnaJ is involved in
binding to denatured protein s ubstrates. EMBO J. 15 , 408–417.
13. Karzai, A.W. & McMacken, R. (1996) A bipartite signaling
mechanism involved in D naJ-mediated activation o f the Escheri-
chia coli DnaK protein. J. Biol. Chem. 271, 11236–11246.
14. Cheetham, M.E. & Caplan, A.J. (1998) S tructure, function and
evolution of DnaJ: conservation a nd adaptation of chaperone
function. Cell Stress Chaperones 3, 28–36.
15. Caplan, A.J., Tsai, J., Casey, P.J. & Douglas, M.G. (1992) Far-
nesylation of YDJ1p is required for function at elevated growth

approximately 24 cM f rom the centromere,. Genomics 53,415.
24.Ohtsuka,K.&Hata,M.(2000)MammalianHSP40/DNAJ
homologs: cloning a novel cDN A and a proposal for the ir clas-
sification and nomenclature. Cell Stress Chaperones 5, 98–112.
25. Kritzik, M.R., Ziober, A.F., Dicharry, S., Conrad, D.J. & Sigal, E.
(1997) Characterization and sequence of an additional 15-lipoxy-
genase transcript and of the human gene. Biochim. Biophys. Acta
1352, 267–281.
26. Angelidis, C .E., Nova, C., Lazaridis, I., Kontoyiannis, D., Kollias,
G. & Pagoulatos, G.N. (1996) Overexpression of HSP70 in
transgenic mice results in increased cell thermotolerance. Trans-
genics 2, 111– 117.
Ó FEBS 2002 Monkey dj2, a specific partner for hsc70 (Eur. J. Biochem. 269) 1559
27. Angelidis, C .E., Lazaridis, I. & Pagoulatos, G.N. (1988) Specific
inhibition of simian virus 40 protein synthesis by heat and arsenite
treatment. Eur. J. Biochem. 179, 27–34.
28. Tanguay, R.M., W u, Y. & K handjian, E.W. (1993) Tissue-specific
expression o f heat shock proteins of the mouse in t he absence o f
stress. Dev. Genet. 14, 112–118.
29. Angelidis, C.E., Lazaridis, I. & Pagoulatos, G .N. (1999)
Aggregation of h sp70 and hsc70 in vivo is distinct and tempera-
ture-dependent and their chaperone function is directly related to
non-aggregated forms. Eur. J. Biochem. 259, 505–512.
30. Davis, A.R., Alevy, Y .G., Chellaiah, A., Quinn, M.T. &
Mohanakumar, T. (1998) Characterization of HDJ-2, a human 40
kD heat shock p rotein. Int. J. Bioc hem. Cell Biol. 30, 1203–1221.
31. Hohfeld, J., Minami, Y. & Hartl, F U. (1995) Hip, a novel
cochaperone involved in the eukaryotic Hsc70/Hsp40 reaction
cycle. Cell 83 , 589–598.
32. Zeiner, M., Gebauer, M. & Gehring, U. (1997) Mammalian

M., Kawai, T., Konda, E., Morinaga, H., Nikawa, T. & Kishi, K.
(1998) Implications of heat shock/stress proteins for medicine and
disease. J. Med. Invest. 44, 137–147.
42. Latchman, D. ( 1998) Heat shock proteins: Protective effect and
potential therapeu tic use. Int. J. Mol. Med. 2, 375–381.
43. Chai, Y., Koppenhafer, S.L., Bonini, N.M. & Paulson, H.L.
(1999) Analysis of the role of heat shock protein (Hsp) molecular
chaperones in polyglutamine disease. J. Neurosci. 19, 10338–
10347.
1560 P. Bozidis et al. (Eur. J. Biochem. 269) Ó FEBS 2002