Cloning and Characterization of a Novel Oxidoreductase KDRF from a Human Bone Marrow-derived Stromal Cell Line KM-102*

(Received for publication, June 4, 1996, and in revised form, August 30, 1996)

Ryuta Koishi , Ichiro Kawashima , Chigusa Yoshimura , Mie Sugawara and Nobufusa Serizawa Dagger

From the Biomedical Research Laboratories, Sankyo Co., Ltd., 2-58 Hiromachi 1-chome, Shinagawa-ku, Tokyo 140, Japan

ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
Acknowledgments
REFERENCES


ABSTRACT

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-<UNL>d</UNL>erived <UNL>r</UNL>eductase-like <UNL>f</UNL>actor), and KDRF may play a role in scavenging reactive oxygen intermediates, which are possibly toxic to cells, in response to inflammatory stimuli.


INTRODUCTION

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-<UNL>d</UNL>erived <UNL>r</UNL>eductase-like <UNL>f</UNL>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.


EXPERIMENTAL PROCEDURES

Preparation of Poly(A)+ RNA from KM-102 Cells

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-1beta (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.

cDNA Cloning

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 lambda 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 [alpha -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 (lambda 31-7). Following digestion with EcoRI (Takara Shuzo Co.), the insert cDNA of lambda 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).

Northern Blot Analysis

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 beta -actin to verify the amount of poly(A)+ RNA in each lane.

Construction of an Expression Plasmid and DNA Transfection

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-SRalpha 296 (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 DH5alpha (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 SRalpha promoter was selected, and this plasmid was named pSRalpha 31-7. Monkey kidney-derived COS-1 cells were transfected with pSRalpha 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 (phi  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.

Construction and Purification of Maltose Binding Protein (MBP) Fusion Protein

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'.


Fig. 1. Construction of pMAL-K using polymerase chain reaction. Construction of pMAL31-7 is performed as described under "Experimental Procedures." Synthetic oligonucleotides i and ii, which correspond to the sequence of pUCKM31-7 cDNA in this figure, were used in the first PCR as primers using the pUCKM31-7 plasmid DNA as a template. This amplified fragment is described as A. After annealing the synthetic oligonucleotides iii and iv (shown as B in this figure), the second PCR was carried out with the synthetic oligonucleotides ii and v as primers using the fragments A and B as templates. After digestion of the amplified fragment with restriction enzymes KpnI and SmaI, the digested fragment containing the enterokinase cleavage site was ligated to the pMAL31-7 plasmid DNA, which had also been digested with KpnI and SmaI. All of the PCR conditions are shown under "Experimental Procedures." The predicted first methionine codon in pUCKM31-7 cDNA is described as ATG1. EK, enterokinase; pUCKM31-7 (KDRF) is the cDNA sequence containing the pUCKM31-7 coding protein.
[View Larger Version of this Image (25K GIF file)]


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).



Fig. 2. The restriction maps, nucleotide sequence, and predicted amino acid sequence of the pUCKM31-7 cDNA. a, restriction endonuclease cleavage maps of the cDNAs of pUCKM31-7 and pcD-31. The darkened region represents the open reading frame. b, combined nucleotide sequence and deduced amino acid sequence of the insert cDNA of pUCKM31-7 and pcD-31. The numbers to the left and below each line refer to the nucleotide positions and amino acid positions, respectively. Comparison of the pUCKM31-7 cDNA sequence with that of the pcD-31 cDNA sequence revealed that the former sequence lacks a poly(A) tail. The probable methionine initiation codon is used for numbering the amino acids, and amino acids in the open reading frame following this methionine are indicated. The valine residue at position 24 and the lysine residue at position 49 are at the N-terminals of the proteins secreted by transfected COS-1 cells and are indicated by the open and closed triangles, respectively. AUUUA motifs are indicated by double underlines. The vertical arrow marks the 5' end of clone pcD-31. The presumptive polyadenylation recognition site is marked with a single underline.
[View Larger Versions of these Images (45 + 65K GIF file)]


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 Analysis

SDS-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 Cells

A 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 Analysis

Serum-free conditioned medium (10 liters) of COS-1 cells transfected with the pSRalpha 31-7 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 (phi  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 pSRalpha 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 pSRalpha 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.).

Oxidized Glutathione (GSSG), 2,6-Dichlorophenol-indophenol (DCIP), and 5,5'-Dithio-bis(2-nitrobenzoic acid) (DTNB) Reduction Assay

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 Delta  absorbance/min/mg of protein.

GSSG Reduction Assay

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 Assay

Three 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 Assay

Fifty 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.


RESULTS

Isolation of a cDNA Clone

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 (lambda 31-7). Following digestion of the lambda 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.

Analysis of Deduced Amino Acid Sequence

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.


Fig. 3. Comparison of amino acid sequence identity between different oxidoreductases. Percentages were obtained from the ratio of the number of identical residues to the number of aligned residues for each pair of sequences.
[View Larger Version of this Image (26K GIF file)]



Fig. 4. Comparison of the amino acid sequence of the pUCKM31-7 coding protein (KDRF, upper strand) with human glutathione reductase (hGR, lower strand). Sequence data were obtained from the SWISS-PROT data base. Identical residues are indicated by *. The amino acid sequence of KDRF corresponding to the redox active site in hGR is boxed by a bold line, and the corresponding FAD binding site is boxed by a dotted line. The consensus sequence of the NADPH binding domain of GRs is boxed by a thin line. The amino acids that are important for determining the specificity for NADPH in hGR are indicated by the closed triangles, and the amino acids that are involved in binding with GSSG are underlined. The important residues for interaction with GSSG in hGR are indicated by double underlines. Cys90 in hGR is indicated by an open triangle.
[View Larger Version of this Image (75K GIF file)]


Expression in Mammalian Cells

To express the pUCKM31-7 coding protein in mammalian cells, we constructed the plasmid pSRalpha 31-7 using the high expression plasmid pcDL-SRalpha 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 pSRalpha 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 pSRalpha 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.


Fig. 5. Western blot analysis of the pUCKM31-7 coding protein (KDRF) secreted by transfected COS-1 cells and A549 cells. Serum-free culture supernatant from COS-1 cells transfected with a negative control plasmid, pcDL-SRalpha 296 (lane 1), or with pSRalpha 31-7 (lane 2), and serum-free culture supernatant from A549 cells (lane 3) were concentrated using trichloroacetic acid precipitation and analyzed under reducing conditions on 8% SDS-PAGE followed by Western blot analysis using rabbit antiserum. The serum was raised against the recombinant fusion protein for the pUCKM31-7 coding protein (KDRF). Bands of KDRFs with different molecular weights are indicated by arrows and the migration of molecular weight markers is shown on the right (in thousands).
[View Larger Version of this Image (29K GIF file)]


These results indicate that the pSRalpha 31-7 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 pSRalpha 31-7 coding protein were detected in serum-free conditioned medium of COS-1 cells transfected with pSRalpha 31-7 as described previously (Fig. 5). Judging from their molecular weights, the 526 and 501 amino acid forms of pSRalpha 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.

Recombinant Fusion Protein Purification and Measurement of Reducing Activity

Because the amount of pSRalpha 31-7 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 Delta  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 pSRalpha 31-7 has a reductase activity, but its substrate specificities were different from those of GR. We, therefore, named this protein KDRF (KM-102-<UNL>d</UNL>erived <UNL>r</UNL>eductase-like <UNL>f</UNL>actor).

Table I.

Reducing activities of pMAL-K on oxidized glutathione (GSSG), 2,6-dichlorophenol-indophenol (DCIP), or 5,5'-dithio-bis(2-nitrobenzoic acid) (DTNB)


 Delta absorbance/min/mg of proteina
Substrate pMAL-K yGR

GSSG 0.3 267.6
DTNB 0.9 4.1
DCIP 5.3 0.6

a  Assay mixtures lacking enzymes served as controls. Reducing activities of pMAL-c that expresses only MBP were the same as the controls.

mRNA Expression and Transcriptional Regulation

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.


Fig. 6. Northern blot analysis of KDRF mRNA. a, KDRF mRNA expression in various human tissues and cell lines. PBL represents peripheral blood lymphocytes. b, KDRF mRNA expression during stimulation of KM-102 cells. Poly(A)+ RNAs were prepared from the KM-102 cells without stimulation (-) and after stimulation with PMA (10 ng/ml) + A23187 (0.2 µM), IL-1beta (25 units/ml), or LPS (1 µg/ml) for 0.5, 1, 4, or 17 h. The ratios indicate the KDRF mRNA expression level in comparison with the level without stimulation. The KDRF mRNA was detected as a single band with 3.9 kb in length. Hybridization with a human beta -actin probe is shown as a control for the integrity of the samples.
[View Larger Version of this Image (33K GIF file)]


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-1beta (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-1beta 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.


DISCUSSION

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-1beta (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-1beta , 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.


FOOTNOTES

*   The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) D88687[GenBank].


Dagger    To whom correspondence should be addressed. Tel.: (81-3) 3492-3131; Fax: (81-3) 5436-8565.
1    The abbreviations used are: ROI, reactive oxygen intermediate; GR, glutathione reductase; yGR, yeast GR; hGR, human GR; IL, interleukin; PMA, phorbol 12-myristate 13-acetate; LPS, lipopolysaccharide(s); bp, base pair(s); DMEM, Dulbecco's modified Eagle's medium; MBP, maltose binding protein; PCR, polymerase chain reaction; PAGE, polyacrylamide gel electrophoresis; GSSG, oxidized glutathione; DCIP, 2,6-dichlorophenol-indophenol; DTNB, 5,5'-dithio-bis(2-nitrobenzoic acid).

Acknowledgments

We thank Dr. Y. Takebe for the kind gift of pcDL-SRalpha 296 expression vector, C. Unozawa-Fujikawa for technical assistance, and K. Kimura-Abiko for peptide sequencing.


REFERENCES

  1. Fridovich, I. (1978) Science 201, 875-880 [Medline] [Order article via Infotrieve]
  2. Cross, A. R. (1990) Free Radical Biol. Med. 8, 71-93 [CrossRef][Medline] [Order article via Infotrieve]
  3. Cross, A. R., and Jones, O. T. G. (1991) Biochim. Biophys. Acta 1057, 281-298 [Medline] [Order article via Infotrieve]
  4. Behl, C., Davis, J. B., Lesley, R., and Schubert, D. (1994) Cell 77, 817-827 [Medline] [Order article via Infotrieve]
  5. Imlay, J. A., and Linn, S. (1987) J. Bacteriol. 169, 2967-2976 [Medline] [Order article via Infotrieve]
  6. Chan, E., and Weiss, B. (1987) Proc. Natl. Acad. Sci. U. S. A. 84, 3189-3193 [Abstract]
  7. Demple, B., and Halbrook, J. (1983) Nature 304, 466-468 [Medline] [Order article via Infotrieve]
  8. Christman, M. F., Morgan, R. W., Jacobson, F. S., and Ames, B. N. (1985) Cell 41, 753-762 [Medline] [Order article via Infotrieve]
  9. Storz, G., Tartaglia, L. A., and Ames, B. N. (1990) Science 248, 189-194 [Medline] [Order article via Infotrieve]
  10. Keyse, S. M., and Tyrrell, R. M. (1989) Proc. Natl. Acad. Sci. U. S. A. 86, 99-103 [Abstract]
  11. Kim, I. H., Kim, K., and Rhee, S. G. (1989) Proc. Natl. Acad. Sci. U. S. A. 86, 6018-6022 [Abstract]
  12. Autor, A. P. (1982) J. Biol. Chem. 257, 2713-2718 [Free Full Text]
  13. Hass, M. A., Frank, L., and Massaro, D. (1982) J. Biol. Chem. 257, 9379-9383 [Abstract/Free Full Text]
  14. Arnold, H. H., and Heinze, H. (1990) FEBS Lett. 267, 189-192 [CrossRef][Medline] [Order article via Infotrieve]
  15. Wong, G. H. W., and Goeddel, D. V. (1988) Science 242, 941-944 [Medline] [Order article via Infotrieve]
  16. Thorens, B., Mermod, J-J., and Vassalli, P. (1987) Cell 48, 671-679 [Medline] [Order article via Infotrieve]
  17. Shaw, G., and Kamen, R. (1986) Cell 46, 659-667 [Medline] [Order article via Infotrieve]
  18. Meijlink, F., Curran, T., Miller, A. D., and Verma, I. M. (1985) Proc. Natl. Acad. Sci. U. S. A. 82, 4987-4991 [Abstract]
  19. Wilson, T., and Treisman, R. (1988) Nature 336, 396-399 [CrossRef][Medline] [Order article via Infotrieve]
  20. Jones, T. R., and Cole, M. D. (1987) Mol. Cell. Biol. 7, 4513-4521 [Medline] [Order article via Infotrieve]
  21. Tutic, M., Lu, X., Schirmer, R. H., and Werner, D. (1990) Eur. J. Biochem. 188, 523-528 [Abstract]
  22. Ho, Y-S., and Crapo, J. D. (1988) FEBS Lett. 229, 256-260 [CrossRef][Medline] [Order article via Infotrieve]
  23. Bell, G. I., Najarian, R. C., Mullenbach, G. T., and Hallewell, R. A. (1986) Nucleic Acids Res. 14, 5561-5562 [Medline] [Order article via Infotrieve]
  24. Harigaya, K., and Handa, H. (1985) Proc. Natl. Acad. Sci. U. S. A. 82, 3477-3480 [Abstract]
  25. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY
  26. Okayama, H., and Berg, P. (1983) Mol. Cell. Biol. 3, 280-289 [Medline] [Order article via Infotrieve]
  27. Woods, D. (1984) Focus 6, 1-3
  28. Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Sci. U. S. A. 74, 5463-5467 [Abstract]
  29. Takebe, Y., Seiki, M., Fujisawa, J., Hoy, P., Yokota, K., Arai, K., Yoshida, M., and Arai, N. (1988) Mol. Cell. Biol. 8, 466-472 [Medline] [Order article via Infotrieve]
  30. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A., and Struhl, K. (1989) Current Protocols in Molecular Biology, pp. 1.8.6-1.8.7, Greene Publishing Associates and Wiley-Interscience, New York
  31. Laemmli, U. K. (1970) Nature 227, 680-685 [Medline] [Order article via Infotrieve]
  32. Sulston, J., Du, Z., Thomas, K., Wilson, R., Hillier, L., Staden, R., Halloran, N., Green, P., Thierry-Mieg, J., Qiu, L., Dear, S., Coulson, A., Craxton, M., Durbin, R., Berks, M., Metzstein, M., Hawkins, T., Ainscough, R., and Waterston, R. (1992) Nature 356, 37-41 [CrossRef][Medline] [Order article via Infotrieve]
  33. Fleischmann, R. D., Adams, M. D., White, O., Clayton, R. A., Kirkness, E. F., Kerlavage, A. R., Bult, C. J., Tomb, J.-F., Dougherty, B. A., Merrick, J. M., McKenney, K., Sutton, G., Fitzhugh, W., Fields, C. A., Gocayne, J. D., Scott, J. D., Shirley, R., Liu, L.-I., Glodek, A., Kelley, J. M., Weidman, J. F., Phillips, C. A., Spriggs, T., Hedblom, E., Cotton, M. D., Utterback, T. R., Hanna, M. C., Nguyen, D. T., Saudek, D. M., Brandon, R. C., Fine, L. D., Fritchman, J. L., Fuhrmann, J. L., Geoghagen, N. S. M., Gnehm, C. L., McDonald, L. A., Small, K. V., Fraser, C. M., Smith, H. O., and Venter, J. C. (1995) Science 269, 496-512 [Medline] [Order article via Infotrieve]
  34. Kubo, A., Sano, T., Saji, H., Tanaka, K., Kondo, N., and Tanaka, K. (1993) Plant Cell Physiol. 34, 1259-1266
  35. Creissen, G., Edwards, E. A., Enard, C., Wellburn, A., and Mullineaux, P. (1991) Plant J. 2, 129-131
  36. Perry, A. C. F., Ni Bhriain, N., Brown, N. L., and Rouch, D. A. (1991) Mol. Microbiol. 5, 163-171 [Medline] [Order article via Infotrieve]
  37. Aboagye-Kwarteng, T., Smith, K., and Fairlamb, A. H. (1992) Mol. Microbiol. 6, 3089-3099 [Medline] [Order article via Infotrieve]
  38. Rossman, M. G., Liljas, A., Brändén, C.-I., and Banaszak, L. J. (1975) in The Enzymes (Boyer, P. D., ed), 3rd Ed., Vol. 11, pp. 61-102, Academic Press, New York
  39. Perham, R. N., Scrutton, N. S., and Berry, A. (1991) BioEssays 13, 515-525 [Medline] [Order article via Infotrieve]
  40. Karplus, P. A., Pai, E. F., and Schulz, G. E. (1989) Eur. J. Biochem. 178, 693-703 [Abstract]
  41. Untucht-Grau, R., Schirmer, R. H., Schirmer, I., and Krauth-Siegel, R. L. (1981) Eur. J. Biochem. 120, 407-419 [Abstract]
  42. Rubartelli, A., Cozzolino, F., Talio, M., and Sitia, R. (1990) EMBO J. 9, 1503-1510 [Abstract]
  43. Abraham, J. A., Whang, J. L., Tumolo, A., Mergia, A., Friedman, J., Gospodarowicz, D., and Fiddes, J. C. (1986) EMBO J. 5, 2523-2528 [Abstract]
  44. Tagaya, Y., Maeda, Y., Mitsui, A., Kondo, N., Matsui, H., Hamuro, J., Brown, N., Arai, K., Yokota, T., Wakasugi, H., and Yodoi, J. (1989) EMBO J. 8, 757-764 [Abstract]
  45. Negro, A., Tolosano, E., Skaper, S. D., Martini, I., Callegaro, L., Silengo, L., Fiorini, F., and Altruda, F. (1991) Eur. J. Biochem. 201, 289-294 [Abstract]
  46. Schuler, G. D., and Cole, M. D. (1988) Cell 55, 1115-1122 [Medline] [Order article via Infotrieve]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.