(Received for publication, June 4, 1996, and in revised form, August 30, 1996)
From the Biomedical Research Laboratories, Sankyo Co., Ltd., 2-58 Hiromachi 1-chome, Shinagawa-ku, Tokyo 140, Japan
A cDNA clone coding for a novel
oxidoreductase was cloned from a human bone marrow-derived stromal cell
line KM-102. We screened a cDNA library constructed from the
mRNA of KM-102 cells stimulated with phorbol 12-myristate
13-acetate and calcium ionophore A23187 using a 32P-labeled
15-mer synthetic oligonucleotide (5-TAAATAAATAAATAA-3
) probe. This
probe was designed as a complementary sequence to the three reiterated
AUUUA sequences, which are contained in the 3
-untranslated regions of
cytokine and some proto-oncogene mRNAs and correlate with rapid
mRNA turnover. Then, we obtained one cDNA clone, and further
sequence analysis revealed that it coded for a new protein exhibiting
30 to ~40% homology with glutathione reductase. By fusion protein
analysis, this protein showed reducing activities on
2,6-dichlorophenol-indophenol and 5,5
-dithio-bis(2-nitrobenzoic acid)
but only a weak reducing activity on oxidized glutathione. Although it
lacked a stretch of hydrophobic amino acids in its N terminus, it was
secreted by monkey kidney-derived COS-1 cells when we introduced the
expression plasmid into them and also secreted by a human lung
carcinoma cell line A549. Northern blot analysis revealed that the
mRNA turnover of this protein was regulated by inflammatory stimuli
in KM-102 cells. These results show that this protein may have
scavenging enzyme properties and has its mRNA expression regulated
in a similar fashion to cytokine genes or proto-oncogenes. Thus, we
named it KDRF (KM-102-
erived
eductase-like
actor), and KDRF may play a role in scavenging reactive
oxygen intermediates, which are possibly toxic to cells, in response to
inflammatory stimuli.
Molecular oxygen, utilized in the form of oxidizing agents in the
metabolic pathway within cells, is reduced to form water (H2O) after passing through a reactive oxygen intermediate
(ROI)1. Through the activation of a variety
of ROI-producing enzymes in response to exogenous stimuli, ROIs are
produced in all cells. The representative ROIs are superoxide
anion (O2), hydrogen peroxide
(H2O2), and hydroxyl radical (-OH·)(1-3). Because excessive accumulation of ROIs is toxic (1, 4), the
intracellular level of ROIs is tightly regulated by several small
antioxidant molecules (e.g. reduced glutathione), which contain sulfhydryl groups, and ROI-scavenging enzymes (e.g.
superoxide dismutase, catalase, glutathione peroxidase, and glutathione
reductase). Both prokaryotic and eukaryotic cells have inducible
defenses to counter oxidative damage (5-13). In fact, it has been
reported that the mRNA of glutathione reductase (GR) or manganous
superoxide dismutase was induced in human cells and cell lines by
stimulation with lectin, tumor necrosis factor, or interleukin-1
(IL-1)(14, 15). But the mechanism(s) by which cells receive and respond
to exogenous stimuli, including the induction mechanism(s) of the mRNAs of ROI-scavenging enzymes, has not yet been elucidated.
On the other hand, the control of transcription of many transiently
expressed genes, such as cytokine genes and proto-oncogenes, has
extensively been studied using exogenous stimuli. These experiments have shown that this class of mRNAs rapidly increased in cells by a
signal(s) with the exogenous stimuli because it gives rise to a
temporary block in specific mRNA degradation through protein synthesis (16). But the accumulation of the mRNAs is not maintained for a long time, and after receiving the signal, the amount of mRNAs in the cells rapidly decreases. A common feature of these rapid turnover mRNAs is that they all have reiterated or dispersed AUUUA sequences in their 3-untranslated regions (17), and several studies have demonstrated that the presence of the reiterated or
dispersed AUUUA sequences in the 3
-untranslated regions of the
mRNAs correlated with rapid mRNA degradation (17-20). Thus, this class of mRNAs is posttranscriptionally regulated, and the AUUUA sequences are believed to be involved in the selective
degradation of transiently expressed mRNAs.
In the case of scavenging enzymes, although mRNAs of GR and
manganous superoxide dismutase also transiently increase with exogenous
stimuli, their mRNAs do not have such reiterated or dispersed AUUUA
sequences in their 3-untranslated regions (21, 22), except that
mRNA of human catalase has three dispersed AUUUA sequences in its
3
-untranslated region (23). Therefore, the relationship between such
scavenging enzymes and the AUUUA sequences has yet to be discussed.
In this report, we describe the isolation and characterization of a
cDNA clone encoding a novel oxidoreductase from a human bone
marrow-derived stromal cell line KM-102 (24) using a complementary probe to the AUUUA sequences. We named it KDRF
(KM-102-erived
eductase-like
actor) and demonstrated that this molecule may have
scavenging enzyme properties and has its mRNA expression regulated
in a similar fashion to cytokine genes or some proto-oncogenes also.
The findings suggest that KDRF may play a role in scavenging ROIs,
especially in response to inflammatory stimuli.
Human stromal cell line KM-102 was a kind gift from Drs. K. Harigaya and H. Handa. KM-102 cells were cultured with Iscove's modified minimum essential medium (Boehringer Mannheim) supplemented with 10% heat-inactivated fetal bovine serum in plastic culture dishes
with a diameter of 15 cm. After growing the cells confluently, phorbol
12-myristate 13-acetate (PMA) (Sigma) and calcium
ionophore A23187 (Sigma) were added to the culture at
10 ng/ml and 0.2 µM, respectively. After cultivation for
3, 6, or 14 h, total RNA for each was extracted by the guanidine
thiocyanate extraction method (25). Six hundred µg each of the total
RNAs obtained in this manner were mixed, followed by oligo(dT)
cellulose Type 7 (Pharmacia Biotech Inc.) column chromatography to
purify poly(A)+ RNA. Using this procedure, approximately
100 µg of poly(A)+ RNA were obtained. This
poly(A)+ RNA was used for construction of a cDNA
library and for Northern blot analysis. Separately,
poly(A)+ RNAs were also prepared from KM-102 cells after
stimulation with PMA (used at 10 ng/ml) + A23187 (used at 0.2 µM), human IL-1 (used at 25 units/ml, Genzyme Co.,
Cambridge, MA), or lipopolysaccharides (LPS) from Escherichia
coli (used at 1 µg/ml, Sigma) for 0.5, 1, 4, or
17 h. The poly(A)+ RNAs were used for Northern blot
analysis.
A cDNA library was prepared in the
Okayama-Berg expression vector (pcDV1 Oligo(dT)-Tailed Plasmid Primer
and pL1 Oligo(dG)-Tailed Linker, Pharmacia Biotech Inc.) from 5 µg of
the above-mentioned poly(A)+ RNA (26). The library was
screened with a 32P-labeled synthetic 15-mer
oligonucleotide (5-TAAATAAATAAATAA-3
, prepared with a Perkin-Elmer
380B Automatic DNA Synthesizer, Perkin-Elmer, Norwalk, CT) according to
the method as described (27). Among the 6500 colonies screened, 33 positive clones were identified. A single clone, pcD-31, was selected
and subjected to the second screening. For the second screening, a
cDNA library was prepared from 3.3 µg of the poly(A)+
RNA using a
gt10 system (Amersham Corp.). Either a 292-bp
PstI-StuI fragment (this PstI site was
derived from pL1 Oligo(dG)-Tailed Linker) or a 223-bp
EcoT22I-StuI fragment (PstI,
StuI, and EcoT22I, Takara Shuzo Co., Ohtsu,
Japan) was prepared from pcD-31 as a probe and labeled with
[
-32P]dCTP (Amersham Corp.) using a Multiprime DNA
labeling system (Amersham Corp.), and plaque hybridization was
performed according to the standard procedure as described (25). Among
the 2 × 105 plaques screened, one positive clone was
obtained (
31-7). Following digestion with EcoRI (Takara
Shuzo Co.), the insert cDNA of
31-7 (3.9 kilobase pairs) was
recloned in the EcoRI site of pUC18, and the resultant
plasmid was named pUCKM31-7. The nucleotide sequence of this clone was
determined in both strands by the dideoxy chain termination method
(28).
Five µg of poly(A)+
RNA of stimulated or unstimulated KM-102 cells were fractionated in a
1.0% agarose gel and blotted onto a nylon membrane
(Hybond-N+, Amersham Corp.)(25). The blot was hybridized
with the 32P-labeled cDNA fragment prepared from the
plasmid pcD-31 (a 292-bp PstI-StuI fragment) or
pUCKM31-7 (a 1006-bp SmaI-XbaI fragment; SmaI and XbaI, Takara Shuzo Co.), and the
hybridization was performed overnight at 42 °C in a solution of 50%
formamide, 5 × SSPE (1 × SSPE = 150 mM
NaCl, 10 mM sodium dihydrogenphosphate monohydrate, and 1 mM EDTA, pH 7.4), 5 × Denhardt's solution (1 × Denhardt's = 0.2 g/liter bovine serum albumin
(Sigma), 0.2 g/liter polyvinylpyrrolidone (Sigma), and 0.2 g/liter Ficoll
(Sigma)), 2% SDS, and 100 µg/ml of denatured salmon
sperm DNA (Sigma). Membranes were washed once at room
temperature in 2 × SSC (1 × SSC = 150 mM
NaCl and 15 mM sodium citrate) containing 0.05% SDS for 30 min, and once at 50 °C in 0.1 × SSC containing 0.1% SDS for
1 h. They were exposed to x-ray films at 70 °C with
intensifier screens. The size of mRNA was estimated using the
relative mobilities of RNA standards purchased from Life Technologies,
Inc. Northern blots containing 2 µg of poly(A)+ RNA of
other human cell lines and multiple human tissues were purchased from
Clontech Laboratories (Palo Alto, CA) and used for hybridization with
the probe. The same membranes were reprobed with human
-actin to
verify the amount of poly(A)+ RNA in each lane.
After digestion of the pUCKM31-7 plasmid DNA with
HindIII (Takara Shuzo Co.) and isolation of the 3003-bp
fragment containing the cDNA insert, the terminals were blunted
using a DNA blunting kit (Takara Shuzo Co.). A high expression plasmid
pcDL-SR296 (29) was digested with PstI and
KpnI (Takara Shuzo Co.). After similarly blunting the
terminals and following dephosphorylation with alkaline phosphatase
(derived from E. coli, Takara Shuzo Co.), it was coupled
with the above-mentioned cDNA fragment in a reaction using T4 DNA
ligase (Pharmacia Biotech Inc.). E. coli DH5
(Life
Technologies) was then transformed with this DNA, and the strain in
which the direction of the cDNA transcription was identical to the
direction of the SR
promoter was selected, and this plasmid was
named pSR
31-7. Monkey kidney-derived COS-1 cells were transfected
with pSR
31-7. COS-1 cells were cultured in a flask (150 cm2) with Dulbecco's modified Eagle medium (DMEM, Nissui
Pharmaceutical Co., Tokyo, Japan) supplemented with 10% fetal bovine
serum. COS-1 cells were collected by trypsin-EDTA
(Sigma) from the flask in which they were grown to
semiconfluence and were washed twice with phosphate-buffered salts
(PBS(
) buffer, Takara Shuzo Co.). Next, the cells were suspended in
PBS(
) buffer at 2 × 107 cells/ml. The plasmid DNA
prepared using the cesium chloride method (25) was dissolved in PBS(
)
buffer at 200 µg/ml. Twenty µl each of the above-mentioned cell
suspension and DNA solution were mixed, and the mixture was placed in a
chamber containing electrodes at intervals of 2 mm (FTC-13, Shimadzu
Co., Kyoto, Japan). Transfection of COS-1 cells was performed with the
electroporation method (30) using the GTE-1 (Shimadzu Co.) gene
introduction device available, and 600-V/30 µs pulses were then
applied twice at 1-s intervals. After cooling the chamber at 4 °C
for 5 min, the cell-DNA mixture inside was added to 10 ml of DMEM
containing 10% fetal bovine serum. Then, the cells were cultured
overnight at 37 °C in a 5% CO2 atmosphere after being
transferred to a Petriplate (
90 × 20 mm). After culturing
overnight, the culture supernatant was removed. The cells were then
washed with serum-free DMEM, and 10 ml of serum-free DMEM were added
followed by culturing for an additional 3 days. The cell supernatant
was then collected from the culture.
pUCKM31-7 DNA was digested with
HindIII. After isolation and purification of the 3003-bp
fragment containing the cDNA insert, the terminals were blunted
using a DNA blunting kit. The fragment was then further digested with
XbaI, and a fragment containing an open reading frame was
purified. A fusion expression vector pMAL-c (New England BioLabs,
Beverly, MA) was also digested with XbaI and
StuI, followed by dephosphorylation with alkaline
phosphatase, and the above-mentioned HindIII
(modified)-XbaI fragment containing the cDNA was coupled
with this expression vector in a reaction using T4 DNA ligase. E. coli TB-1 strain (New England BioLabs) was transformed with this
DNA, and the plasmids were analyzed. The strain in which the direction
of the cDNA transcription was identical to the direction of the
promoter was selected, and this plasmid was named pMAL31-7. Next, we
tried to reconstruct pMAL31-7 as another type of fusion plasmid that
starts with Lys49 in pUCKM31-7 using polymerase chain
reaction (PCR) as illustrated in Fig. 1. This Lys49
corresponds to the first amino acid in the shortest type of this protein which is secreted by COS-1 cells. We synthesized five oligonucleotides (oligo i to v) with a Perkin-Elmer Automated DNA Synthesizer 394: i, 5-GATGACAAA1AGCTACTA-3
; ii,
5
-CATAGGATGCTCCAACAA-3
; iii,
5
-TCGAGCTCGGTACCCGATGACGATGACAAAA1AGCTACTA-3
; iv,
5
-TAGTAGCTT1TTTGTCATCGTCATCGGGTACCGAGCTCGA-3
; and
v, 5
-TCGAGCTCGGTACCCGAT-3
.
In oligos i, iii, and iv, A1 and T1 correspond to A241 in the pUCKM31-7 cDNA sequence (Fig. 2b). Oligo ii is a complementary sequence of the pUCKM31-7 cDNA (from 836 to 853 in Fig. 2b). Oligos iii and iv are complementary sequences to each other including the enterokinase recognition and KpnI sites (Fig. 1).
The first round of PCR was carried out in a reaction mixture containing 100 pmol of each of the primers i and ii, 1 µg of the pUCKM31-7 DNA, 10 µl of 10-fold concentration Taq polymerase reaction buffer solution (containing 500 mM Tris-HCl, pH 8.3, 500 mM KCl, 15 mM MgCl2, 100 mM deoxynucleotide triphosphates, and 2 mg/ml gelatin (BRL)), and 5 units of Taq polymerase (Takara Shuzo Co.) in a total reaction volume of 100 µl. The PCR reaction was carried out at 72 °C for 3 min, then a reaction consisting of 94 °C for 1 min, 55 °C for 2 min, and 72 °C for 3 min was repeated for 30 cycles, and finally a reaction was carried out at 72 °C for 10 min (we called this amplified fragment, fragment A). All of the PCR reactions were conducted using a Perkin-Elmer DNA Thermal Cycler 480. Three µg each of oligos iii and iv were annealed in annealing buffer (7 mM Tris-HCl, pH 7.5, 0.1 mM EDTA, 20 mM NaCl, and 7 mM MgCl2) at 70 °C for 10 min and cooled down to room temperature (this annealed fragment was called fragment B). The second PCR was carried out under the same cycle conditions in a reaction mixture containing 100 pmol of each of the primers ii and v, 5 µg of fragment A and fragment B as templates, 10 µl of 10-fold concentration Taq polymerase reaction buffer solution, and 5 units of Taq polymerase in a total reaction volume of 100 µl. Following PCR, 8% polyacrylamide gel electrophoresis (PAGE) was performed to purify the amplified DNA product. After purification of this amplified fragment, we digested it with KpnI and SmaI. pMAL31-7 DNA were also digested with KpnI and SmaI, followed by dephosphorylation with alkaline phosphatase. The amplified fragment was then coupled with the dephosphorylated pMAL31-7 DNA in a reaction with T4 DNA ligase. E. coli DH10B (BRL) was transformed with this DNA, and the transformed strains were analyzed. We selected one strain and named this plasmid pMAL-K. It was confirmed that no abnormalities existed in the portion, in which the fragment was inserted, by analyzing the nucleotide sequence of this portion of pMAL-K. The fusion proteins produced in E. coli were purified by amylose resin affinity chromatography as recommended by the supplier (New England BioLabs). This cloning procedure is shown in Fig. 1.
Western Blot AnalysisSDS-PAGE was performed as described by Laemmli (31). Western blot analysis was conducted using ECL Western blotting detection reagent (Amersham Corp.). Rabbit antisera were raised using the affinity purified pMAL31-7 fusion protein. These antisera were used for Western blot analysis. Molecular weight markers were purchased from Bio-Rad.
Cultivation of A549 CellsA human lung carcinoma cell line A549 was purchased from Dainihon Pharmaceutical Co. (Osaka, Japan), and A549 cells were cultured with DMEM supplemented with 10% fetal bovine serum. When the cells were grown to semiconfluence, the culture supernatant was removed, and the cells were washed with serum-free DMEM. Then, serum-free DMEM was added, and the serum-free culture supernatant was collected after 3 days of cultivation.
Purification of Recombinant Protein and Amino Acid Sequence AnalysisSerum-free conditioned medium (10 liters) of COS-1 cells
transfected with the pSR31-7 expression plasmid DNA was dialyzed against 10 mM Tris-HCl, pH 9.0. The dialyzed sample was
applied to 10 ml of DEAE-Sepharose Fast Flow (Pharmacia Biotech Inc.) contained in an XK16/20 column (
2.0 × 20 cm, Pharmacia
Biotech Inc.) and eluted with a 0-0.5 M linear gradient of
NaCl in the 10 mM Tris-HCl, pH 9.0, buffer. After
collecting fractions (these fractions were monitored by Western blot)
which were eluted at NaCl concentrations from 0.1 to 0.4 M,
they were combined and dialyzed against 0.1 M Tris-HCl, pH
7.6, 5 mM EDTA, and 1 mM 2-mercaptoethanol (Sigma). Next, we performed an affinity chromatography
using 10 ml of 2
,5
-ADP Sepharose 4B (Pharmacia Biotech Inc.)
contained in an XK16/20 column, and fractions were eluted with a 0-10
mM linear gradient of NADPH in the 0.1 M
Tris-HCl, pH 7.6, 5 mM EDTA, and 1 mM
2-mercaptoethanol buffer. The fractions were subjected to SDS-PAGE
under reducing conditions, and detection of bands was performed with
silver staining and Western blot analysis. After collecting fractions
that contained the predicted pSR
31-7 coding protein, they were
combined, and we electrophoresed it on SDS-PAGE under reducing
conditions and transferred it electrophoretically to a polyvinylidine
difluoride film (ProBlot, Perkin-Elmer). The areas corresponding to the
three bands of pSR
31-7 coding proteins were cut out for protein
microsequencing. Amino-terminal sequence determination was carried out
on a gas-phase protein sequencer (PPSQ-10, Shimadzu Co.).
Fusion protein derived from pMAL-K and protein from pMAL-c,
which expresses only MBP, were purified with affinity chromatography and dialyzed against 0.01 M sodium phosphate, pH 7.5, and 5 mM EDTA (0.01 M phosphate buffer, subsequently
referred to as PB). The concentrations of purified proteins were
determined by Bio-Rad protein assay (Bio-Rad). All assays were done at
room temperature using a Beckman DV 7500 spectrophotometer (Beckman
Instruments, Fullerton, CA). Reducing activities were defined as absorbance/min/mg of protein.
Ten µl of 10 mM GSSG (Boehringer Mannheim) in 0.01 M PB were added to 300 µl of 0.01 M PB containing the purified fusion protein, MBP, or yeast GR (yGR, Boehringer Mannheim). Then, 10 µl of 6.4 mM NADPH (Boehringer Mannheim) in 5% sodium bicarbonate were added to the mixture, and absorbance at 340 nm was monitored at room temperature for 3 min.
DCIP Reduction AssayThree hundred µl of 50 µM DCIP (Sigma) in 0.01 M PB were added to 100 µl of 0.01 M PB containing the fusion protein, MBP, or yGR. Then, 10 µl of 6.4 mM NADPH in 5% sodium bicarbonate were added to the mixture, and absorbance at 600 nm was monitored at room temperature for 3 min.
DTNB Reduction AssayFifty µl of 10 mM DTNB (Wako Pure Chemical Industries, Osaka, Japan) in 0.01 M PB were added to 200 µl of 0.01 M PB containing the fusion protein, MBP, or yGR. Then, 10 µl of 6.4 mM NADPH in 5% sodium bicarbonate were added to the mixture, and absorbance at 412 nm was monitored at room temperature for 3 min.
An Okayama-Berg expression
library constructed from poly(A)+ RNA of KM-102 cells,
which had been stimulated with PMA (10 ng/ml) + A23187 (0.2 µM) for 3, 6, or 14 h, was screened with a synthetic oligonucleotide probe (5-TAAATAAATAAATAA-3
). Among the 6500 clones
screened, 33 clones were hybridized with this probe. We selected one
clone, pcD-31, and subjected it to further analysis. The pcD-31 clone
was 491 bp in length excluding poly(A). However, Northern blot analysis
of the poly(A)+ RNA from KM-102 cells revealed that the
full-length mRNA of pcD-31 was approximately 3.9 kb. To obtain the
full-length cDNA clone of pcD-31, we conducted a second screening.
Among the 2 × 105 plaques screened, one positive
clone was obtained (
31-7). Following digestion of the
31-7 DNA
with EcoRI, the insert was recloned in the EcoRI
site of pUC18, and the resultant plasmid was named pUCKM31-7. This
cDNA clone was 3815 bp in length excluding a poly(A) tail and had a
single open reading frame. The first ATG codon located at nucleotide 97 from the 5
end was followed by a 1647-nucleotide long open reading
frame ending with an in-frame termination codon TGA at position 1744, and the open reading frame was, therefore, able to code for a 549-amino
acid protein (Fig. 2b). The presumptive polyadenylation
site, ATTAAA, was located 1851 nucleotides downstream from this TGA
codon. A sequence, 5
-TTATTTATTTATTT-3
, that is complementary to the
synthetic probe was located 1580 nucleotides downstream from the
termination codon (Fig. 2b). Such reiterated copies of AUUUA
sequences are located in the 3
-untranslated regions of many cytokine
and some proto-oncogene mRNAs, as mentioned previously.
Computer search on
the SWISS-PROT and NBRF (PIR) protein sequence data banks showed the
deduced amino acid sequence had 30 to ~40% identity with human
glutathione reductase (hGR; EC1.6.4.2)(21) and other species' GRs. A
computer-aided comparison revealed 43.5% identity with probable GR of
Caenorhabditis elegans (EC1.6.4.2)(32), 37.4% with
Haemophilus influenzae (EC1.6.4.2)(33), 36.1% with Arabidopsis thaliana (EC1.6.4.2)(34), 36.3% with
Pisum sativum (EC1.6.4.2)(35), 35.5% with human, 35.4%
with Pseudomonas aeruginosa (EC1.6.4.2)(36), and 34.7% with
Trypanothione reductase of Trypanosoma brucei
(EC1.6.4.8)(37)(Fig. 3). An alignment of the amino acid sequence of hGR
with pUCKM31-7 coding protein indicated an overall sequence identity of
35.5% (Fig. 4). The homology was absolute over a well conserved
10-amino acid sequence
Leu107-Gly-Gly-Thr-Cys-Val-Asn-Val-Gly-Cys116,
referred to as the active site, and a 9-amino acid sequence Ile70-Gly-Gly-Gly-Ser-Gly-Gly-Leu-Ala78,
referred to as the FAD binding site. Rossman fold structure (38, 39),
which is the consensus sequence of the NADPH binding domains of GRs,
was conserved in pUCKM31-7 coding protein
Gly249 -Ala-Ser-Tyr-Val-Ala-Leu-Glu-Cys-Ala-Gly-Phe-Leu-Ala-Gly-Ile-Gly265.
Arg218 and Arg224, which are important amino
acid residues in determining the specificity for NADPH in hGR (39),
were also conserved in the pUCKM31-7 coding protein (Arg273
and Arg278, respectively). Karplus et al. (40)
identified six regions in the polypeptide chain of hGR which are
involved in binding GSSG; these are residues 30-37, 59-64, 110-117,
339-347, 467-476, and 406. From an alignment of hGR with the
pUCKM31-7 coding protein, the regions 30-37, 59-64, and 339-347 in
hGR revealed high homology with the corresponding regions of the
pUCKM31-7 coding protein, but the regions 110-117 and 467-476 in hGR
revealed less homology, and the residue 406 was converted to
Leu461 in the pUCKM31-7 coding protein (Fig. 4).
Arg37, Arg38, and Arg347 in hGR are
important residues for interaction with GSSG (39). Although
Arg347 in hGR was conserved as Arg403 in the
pUCKM31-7 coding protein, Arg37 and Arg38 in
hGR were converted to Lys81 and Glu82 in the
pUCKM31-7 coding protein, respectively. Human GR consists of two
identical polypeptide chains, and Cys90 in hGR is the
residue forming a disulfide bridge with Cys90 in the other
peptide chain (41). But the residue corresponding to Cys90
in hGR was converted to Glu143 in the pUCKM31-7 coding
protein (Fig. 4). Although the pUCKM31-7 coding protein had a long
N-terminal amino acid sequence in contrast to hGR, it lacked a stretch
of hydrophobic amino acids in this region as well as hGR.
Expression in Mammalian Cells
To express the pUCKM31-7
coding protein in mammalian cells, we constructed the plasmid
pSR31-7 using the high expression plasmid pcDL-SR
296 and
introduced it into COS-1 cells. The serum-free conditioned medium was
collected and subjected to Western blot analysis with the rabbit
antiserum. As can be seen in Fig. 5, although pSR
31-7 coding protein
lacked a stretch of hydrophobic amino acids in N terminus, it was
secreted in three different molecular weight products by COS-1 cells
(Mr 55,000 to ~60,000). Next, to determine
their N-terminal amino acid sequences, we collected 10 liters of
serum-free conditioned medium of COS-1 cells transfected with
pSR
31-7. The purification scheme was described under "Experimental Procedures." As a result, we were able to determine the two
sequences: I, Val24-Val-Phe-Val-Lys-Gln; and II,
Lys49-Leu-Leu-Lys-Met.
These results indicate that the pSR31-7 coding protein was processed
into its mature forms by cleavages between Pro23 and
Val24 and between Gln48 and Lys49,
which corresponded to the proteins of 526 (Mr
58,000) and 501 (Mr 55,000) amino acids,
respectively. By Western blot analysis, three forms of pSR
31-7
coding protein were detected in serum-free conditioned medium of COS-1
cells transfected with pSR
31-7 as described previously (Fig. 5).
Judging from their molecular weights, the 526 and 501 amino acid forms
of pSR
31-7 coding protein probably corresponded to the second and
third highest molecular weight products, respectively. With regard to
the product with the highest molecular weight, we were unable to
determine its N-terminal amino acid sequence.
Because the amount of pSR31-7 coding protein purified
from the COS-cell conditioned medium in the manner described above was
extremely small, it could not be used for other purposes, such as the
measurement of its reducing activity. Therefore, we attempted to
express it in E. coli as a fusion protein with maltose binding protein (MBP). The procedures for the construction of the
expression plasmid and purification of the fusion protein are shown in
Fig. 1 and under "Experimental Procedures." Because the main function of hGR is dithiol-disulfide oxidoreductase enzymatic activity, we tested whether the fusion protein has a reducing activity
on GSSG, DCIP, or DTNB in the presence of NADPH. As shown in Table
I, the fusion protein derived from pMAL-K had reducing activities on DCIP and DTNB, whereas it revealed only a weak reducing activity on GSSG. To compare the reducing activities of the fusion protein pMAL-K with those of GR, we also measured a reducing activity of yGR on GSSG, DCIP, or DTNB. Reducing activities of yGR were 0.6
absorbance/min/mg of protein on DCIP and 4.1 on DTNB, whereas those of
the fusion protein pMAL-K were 5.3 and 0.9, respectively (Table I). In
addition, a reducing activity of yGR on GSSG was 267.6, in contrast
with 0.3 for pMAL-K. These results suggest that the protein encoded by
pSR
31-7 has a reductase activity, but its substrate specificities
were different from those of GR. We, therefore, named this protein KDRF
(KM-102-
erived
eductase-like
actor).
|
Expression
of KDRF mRNA in various human cell lines and tissues was studied by
Northern blot analysis. As shown in Fig. 6a, KDRF mRNA
was present in all of the cell lines that we examined. Whereas it was
found at low levels in HL-60, HeLa, K-562, MOLT-4, Raji, SW480, and
G361 cells, it was present at a high level in A549 cells. By Western
blot analysis, we detected a single secreted form of KDRF in the
conditioned medium of A549 cells (Fig. 5). As shown in Fig.
6a, KDRF mRNA was present in all of the tissues that we
examined, but in particular, it was found at high levels in testis,
ovary, placenta, and fetal liver. High levels of expression in heart
and skeletal muscle might be derived from large amounts of mRNA in
these lanes.
As shown in Fig. 2b, there is a
5-ATTTATTTATTT-3
sequence in the 3
-untranslated region of the KDRF
cDNA. Using Northern blot analysis, we next investigated KDRF
mRNA expression levels in KM-102 cells during stimulation with
various agents. As shown in Fig. 6b, the level of KDRF
mRNA in unstimulated KM-102 cells was low. However, by stimulation
with PMA (10 ng/ml) + A23187 (0.2 µM), human IL-1
(25 units/ml), or LPS (1 µg/ml), the amount of KDRF mRNA increased
and reached to the maximum level 2 to 3 times higher than the basal
level after 4 h. Although the level of KDRF mRNA decreased
slowly thereafter in the case of IL-1
or LPS stimulation, decrease
of the mRNA was not seen, even after 17 h of stimulation with
PMA + A23187 (Fig. 6b). These results indicate that KDRF
mRNA was regulated in KM-102 cells by some inflammatory
stimuli.
In this report, we described the cloning and characterization of a
novel oxidoreductase KDRF from a human bone marrow-derived stromal cell
line KM-102. The cDNA encoding this human protein was cloned using
a unique synthetic oligonucleotide probe 5-TAAATAAATAAATAA-3
, which
is complementary to the three reiterated copies of AUUUA sequences.
From computer search analysis, KDRF exhibited 30 to ~40% identity with hGR and other species' GRs. An alignment of its amino acid sequence with hGR showed an overall sequence identity of 35.5% (Fig. 3). The active site, the FAD binding site, and the Rossman fold structure (38, 39) were all conserved in KDRF (Fig. 4). Although two other important amino acid residues, Arg218 and Arg224, were conserved in KDRF as Arg273 and Arg278, respectively (Fig. 4), only three of the six regions that are involved in binding GSSG in hGR revealed high homology with KDRF. In particular, Arg37 and Arg38 in hGR were converted to Lys81 and Glu82 in KDRF, respectively (Fig. 4). By fusion protein analysis, KDRF revealed reducing activities on GSSG, DCIP, and DTNB but had a weak reducing activity on GSSG in contrast to yGR (Table I). We, therefore, suggest that many conversions of amino acids in KDRF, which are important for binding or interacting with GSSG, are the possible reason why KDRF showed only a weak reducing activity on GSSG in contrast to yGR. As shown in Table I, substrate specificities of KDRF were different from those of yGR. These results suggest that KDRF may have a natural substrate(s) other than GSSG. However, we cannot exclude the possibility that the MBP domain in the KDRF fusion protein might affect the activity. Therefore, we are now preparing for the measurement of the reducing activities on such substrates with recombinant KDRF protein, not a fusion protein, produced in E. coli.
From Northern blot analysis, mRNA coding for KDRF was detected in all of the tissues that we tested, although the expression levels varied in each tissue (Fig. 6a). We also examined the KDRF mRNA expression level in various human cell lines (Fig. 6a). The results showed low levels of expression in HL-60, HeLa, K-562, MOLT-4, Raji, SW480, and G361 cells but a high level in A549 cells. As described earlier, mRNA coding for KDRF was found to be at a low level in the lung. A549, which is a human lung carcinoma cell line, might have acquired the high expression mechanism(s) of KDRF in the process of carcinogenesis.
In the experiment in which the KDRF high expression plasmid was
introduced into COS-1 cells, KDRF was secreted by the cells as proteins
with three different N-terminal amino acid lengths, despite the fact
that KDRF lacked a stretch of hydrophobic amino acids, called a signal
peptide, in its N terminus (Fig. 5). We also detected a
single secreted form of KDRF in the conditioned medium of A549 by
Western blot analysis (Fig. 5). From these results, we concluded that
KDRF might function as a reducing enzyme both inside and outside of the
cell. Proteins with a defined extracellular function but lacking a
signal sequence have been identified, which include IL-1 (42), basic
fibroblast growth factor (43), adult T cell leukemia-derived factor
(44), and ciliary neurotrophic factor (45). Although little is known
about the mechanism(s) that allows their selective release or the KDRF
secretion pathway, KDRF might also be a member of this class.
We confirmed that mRNA of KDRF had six AUUUA sequences, including
two reiterated copies of AUUUA sequences, in its 3-untranslated region
(Fig. 2b). Northern blot analysis showed that its mRNA expression level in KM-102 cells increased 2 to 3 times after 4-h
activation by various agents (PMA + A23187, human IL-1
, or LPS) and
decreased thereafter, except in the case of PMA + A23187 (Fig.
6b). The turnover rate of KDRF mRNA in
KM-102 cells was slow, but it has been reported that the reiterated
AUUUA sequences might exert their effects through interaction with
cell-specific factors and that some cell-specific differences exist in
this mechanism (46). Thus, it is possible that mRNA of KDRF might more rapidly turn over in other cells than in KM-102 cells. However, it
will be necessary for us to examine whether the expression of KDRF
mRNA is regulated through the AUUUA sequences.
Considered together, KDRF is a new type of human protein that may have scavenging enzyme properties and has its mRNA expression regulated in a similar fashion to cytokine genes or proto-oncogenes. However, it is important to identify its inherent substrate(s), including not only small compounds but also peptides and proteins. It is necessary to elucidate the regulatory mechanism of its mRNA expression and its physiological role(s) as well. More detailed investigations should provide a deeper insight into these issues.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) D88687[GenBank].
We thank Dr. Y. Takebe for the kind gift of
pcDL-SR296 expression vector, C. Unozawa-Fujikawa for technical
assistance, and K. Kimura-Abiko for peptide sequencing.