Vol. 46 No. 2
SCIENCE IN CHINA (Series C)
April 2003
The RFA regulatory sequence-binding protein in the promoter of prostate-specific antigen gene CHEN Weiwen (чฤำ)1, ZHANG Jianye (ߙྜ)1, Charles Y F Young2, ZHANG Lianying (॔)1, CHEN Liucun (чঝӉ)1 & ZHAO Jian (ვ ߒ)3 1. Institute of Biochemistry and Molecular Biology, School of Medicine, Shandong University, Jinan 250012, China; 2. Department of Urology Research of Mayo Clinic/Foundation, MN 55905ˈUSA˗ 3. Department of Thoracic Surgery, Qilu Hospital of Shandong University, Jinan 250012, China Correspondence should be addressed to Zhang Jianye (email:
[email protected]) Received December 14, 2001; revised April 10, 2002
Abstract To assure what sequence associated with the androgen regulation, a 15 bp region at the upstream of the ARE of prostate-specific antigen (PSA) promoter, termed RFA, was found indispensable for androgen receptor (AR)-mediated transactivation of PSA promoter. In transfection and CAT assays, some nucleotides substitution in RFA could significantly decrease the androgen inducibility for PSA promoter. The in vitro DNA binding assay demonstrated that RFA bound specifically with some non-receptor protein factors in prostate cell nucleus, but the mutant type of RFA lost this ability, so RFA might be a novel accessory cis-element. The RFA-binding proteins were isolated and purified by affinity chromatography using RFA probes. SDS-PAGE and preliminary protein identification showed these proteins possessed sequence high homology with multifunctional protein heterogeneous nuclear ribonucleoprotein A1, A2 (hnRNP A1, A2). RFA-binding proteins possibly cooperate with AR-mediated transactivation for PSA promoter as coactivator. The study results will facilitate further understanding the mechanism and tissue specificity of PSA promoter. Keywords: prostate-specific antigen, promoter, heterogeneous nuclear ribonucleoprotein A1, A2.
The prostate-specific antigen (PSA), which plays an important role during the liquefaction process of semen, is a differentiation marker for human prostate. It has become the most sensitive marker for monitoring and detecting prostate cancer. PSA as a serine protease can activate some growth factors that might be related to the advancement of prostate cancer by hydrolyzing growth factor-binding protein. The PSA gene is expressed specifically in prostate epithelial cells. The expression of PSA gene is induced by androgens at the transcriptional level. AR is a transcription factor that belongs to the steroid hormone/thyroid receptor superfamily. Ligand-activated AR can bind to a specific DNA sequence (ARE) in order to enhance gene transcription. Further study has demonstrated that the androgen responsive element (ARE) is necessary but not sufficient for inducing the gene expression by androgen. In addition to ARE, the function of AR may largely depend on the other non-receptor binding sites in the promoter of a particular gene, and non-DNA binding proteins might influence the function of receptor via protein-protein interaction[1]. However, much less is known about the function of AR in this regard than that of other steroid receptors, for
No. 2
RFA REGULATORY SEQUENCE-BINDING PROTEIN
185
fewer AR regulating genes have been characterized. The study on androgen regulating genes of rat has indicated that expression of these genes requires a combinational effect of ARE and other transcription factor/binding elements[2]. When Riegman et al. characterized the ARE of PSA promoter, they found that the sequence (539 to 320) upstream ARE was necessary for androgen induction[3]. We also found that there were two regions at 406 to 371 (36 bp) and 340 to 326 (15 bp) upstream the ARE of PSA promoter enhancing the androgen induction for PSA gene[4,5]. No doubt, further defining these non-receptor factor/binding elements will help us better understand the mechanism of androgen action. When examining the 36 bp region (406ü 371) upstream the ARE of PSA promoter, our interest was aroused by a 15 bp sequence (5Ą -396CAGGGATCAGGGAGT382-3Ą ) within this region, which contains two direct repeats (i.e., CAGGGA) with one nucleotide as a spacer named RFA. We speculated that the RFA might cooperate with the ARE in AR-mediated transactivation. In this report, preliminary studies were performed on RFA and its specific binding protein using co-transfection, in vitro competitive binding assay, and affinity chromatography, etc. 1 1.1
Materials and methods Plasmid constructs Expression vector pBLCAT3 and primers were provided by laboratory of Dr. Young. 5Ą -end
624 bp DNA of PSA promoter was used as a template for PCR to produce a series of different fragments of PSA promoter. All these DNA fragments including one wild type and four mutation types of RFA were initiated from 406. The sequences of the used primers are listed in table 1. Table 1 Primer sequences for PSA promoter-CAT constructs by PCR Primers for PSA promoter construction 3Ą Primer 5Ą Primer
(+31
+6)
5Ą GGATCCCTCTCCGGGTGCAGGTGGTAAGCTT3Ą
406
5Ą TCTAGACAGAGTGGTGCAGGGATCAGGG3Ą
406(A-1)
5Ą TCTAGACAGAGTGGTACAGGGGTCAAGGGGTC3Ą
406(A-2)
5Ą TCTAGACAGAGTGGTGCAGGGGTCAAGGGGTC3Ą
406(A-3)
5Ą TCTAGACAGAGTGGTGCAGGGGTCAAGGAGTC3Ą
406(A-4)
5Ą TCTAGACAGAGTGGTGCAGGGGTCAGGGGGTC3Ą
Both 3Ąand 5Ąprimers contain BamHĉand Xbaĉrestriction sequence at their respective 5Ąends. 406 (A-1)ü406 (A-4) are mutant types, whose substituted nucleotides in/out RFA sequence are underlined.
The PCR products above contain Xba I and BamH I restriction sites at 5Ąand 3Ąends, respectively. They were digested with BamH I and Xba I restriction enzymes, agarose gel purified, and ligated into the vector pBLCAT3 pre-cut with BamH I and Xba I enzymes. The constructs above were confirmed by DNA sequencing. The pBLCAT3 contains a minimal thymidine kinase (tk) promoter and promoterless chloramphenicol acetyltransferase (CAT) gene. 1.2
Cell culture & nuclear extracts PC-3 and LNCaP human prostate cancer cells are conserved in our laboratory, which grew in
186
SCIENCE IN CHINA (Series C)
Vol. 46
5% fetal bovine serum (FBS) RPMI 1640 medium at 37ć with 5% CO2. In addition, 10 nmol/L testosterone was required to LNCaP cells. Nuclear extracts from LNCap and PC-3 cells were prepared as described[6]. 1.3
Transfection experiments and CAT assay When PC-3 cells reached 50%ü70% confluency, they were co-transfected with plasmids
constructed above and a human AR expression vector, using either DEAE-dextran-chloroquine or lipofectamine according to the manufacturer’s instructions (purchased from GIBCO/BRL). Parental vectors pBLCAT3 were used as controls in the assays. Following transfections, cells were incubated with 1% charcoal-stripped FBS RPMI 1640 with or without 3.2 nmol/L mibolerone (Mib; a synthetic androgen) for 24 or 48 h. Cells were then collected and extracted for use in protein assay and a two-phase fluor diffusion CAT assay as described previously[7, 8]. All groups of cells were prepared in duplicate for transfections, which were performed at least three times. 1.4
Electrophoretic mobility shift assay (EMSA) The sequences for these oligonucleotide probes used in EMSA are shown in table 2. Table 2 Oligonucleotide sequences used for EMSA The oligonucleotide sequences containing wild/mutant type monomeric RFA (396ü382): Wild type:
5Ą GATCCAGGGATCAGGGAGTC3Ą
A-2:
GATCCAGGGGTCAAGGGGTC3Ą 5Ą
A-3:
GATCCAGGGGTCAAGGAGTC3Ą 5Ą
A-4:
GATCCAGGGGTCAGGGGGTC3Ą 5Ą
3Ą GTCCCTAGTCCCTCAGCTAG5Ą
3Ą GTCCCCAGTTCCCCAGCTAG5Ą
3Ą GTCCCCAGTTCCTCAGCTAG5Ą
3Ą GTCCCCAGTCCCCCAGCTAG5Ą PSA ARE:
5Ą GATCCTTGCAGAACAGCAAGTGCTAGCTG3Ą 3Ą GAACGTCTTGTCGTTCACGATCGACCTAG5Ą
NF-NB :
5Ą AGTTGAGGGGACTTTCCCAGGC3Ą 3Ą TCAACTCCCCTGAAAGGGTCCG5Ą
NF-țB sequence was purchased from Promega. (A-2)ü(A-4) are mutant RFAs corresponding to 406 (A-2)ü406 (A-4) in transfection experiment respectively, whose nucleotide substitutions in RFA sequence are underlined.
The two complementary single-strands of oligonucleotide were annealed to form double-stranded probes. Two methods were used for labeling probes and development. (i) Radiolabel by Klenow enzyme. The probes were labeled with [D-32P]dCTP (3000 Ci/mmol; purchased from Amersham Corp.) at 5Ą -end by Klenow enzyme to a specific activity of 8×107ü8×108 cpm/Pg. (ii) Digoxigenin label. The probes were labeled with DIG-11-ddUTP at 3Ą -end by terminal transferase according to the manufacturer’s instruction of DIG gel shift kit (purchased from Roche). In vitro DNA binding was performed by incubating the above nuclear extract (5ü8 Pg) in a
No. 2
RFA REGULATORY SEQUENCE-BINDING PROTEIN
187
buffer containing 20 mmol/L HEPES, pH 7.9, 100 mmol/L KCl, 5 mmol/L MgCl2, 0.1 mmol/L EDTA, 12% glycerol, 4 mmol/L DTT and 1 ȝg poly[d(A-T)]/poly[d(I-C)] with labeled probes for 30 min at 25ć. The DIG-labeled probes used were 2ü3 folds of
32
P-labeled probes (20ü30
fmol). Finally, the above reaction mixtures were subjected to 6% native-PAGE at 250 V for 1ü1.5 h. By the first method, gels were dried and performed autoradiograph. And by the second method, the bands in gels were transferred to the positively charged nylon membrane according to the manufacturer’s instruction of DIG gel shift kit, then the membrane was baked to fix the bands and incubated with blocking reagent, anti-DIG-AP, Fab fragments and CSPD in turn. Finally, the membrane was sealed into hybridization bag and exposed to X-ray film in darkroom. 1.5
Affinity chromatography The synthesized oligonucleotide containing double-copy RFA was used in affinity chromatography in order that protein binding capacity could be increased. The sequence of oligonucleotide is as follows: 5Ą -XGGTGCAGGGATCAGGGAGTCTCAGGTGCAGGGATCAGGGAGTCTCA-3Ą ; 5Ą -TGAGACTCCCTGATCCCTGCACCTGAGACTCCCTGATCCCTGCACC-3Ą , X = biotin. Both DynalbeadsTM M-280 Streptavidin and magnetic particles concentrator (MPC) were purchased from Dynal. According to the manufacturer’s instruction, two complementary single-strands of oligonucleotide were annealed to form ds-probes, with biotin at the 5Ą -end of one strand. The DynalbeadsTM M-280 Streptavidin were mixed with probes in a buffer containing 10 mmol/L Tris-HClˈpH 7.5, 1 mmol/L EDTAˈ1.0 mol/L NaCl for 15 min at room temperature. The efficiency of coupling was visualized by 8% native-PAGE of samples taken before and after coupling. The 100 PL magnetic beads with probes were incubated in the same binding buffer as described in sec. 1.4 with 200 PL nuclear protein, 10 Pg poly[d(A-T)] and 1 Pg poly L-lysine at 25 ć for 30 min. After separating the solid phase and liquid phase with MPC, the unbound proteins were removed with 3 washes of binding buffer. The RFA-binding proteins were released from the magnetic bead-streptavidin-probes with binding buffer containing 1 mol/L KCl. The collected elution was diluted to the concentration of 0.1 mol/L K+ with binding buffer without KCl for detecting the function of purified protein by EMSA and molecular weight of protein by SDS-PAGE. 1.6
SDS-PAGE & amino acid sequencing The purified proteins were run in 10% SDS-PAGE. Coomassie blue-staining band of target protein was excised from gel, subjected to tryptic digestion, HPLC, amino acid sequence and mass spectrometric analysis. 2 2.1
Result The function of RFA The expression vector pBLCAT3s were inserted with wild PSA promoter DNA fragment and
188
SCIENCE IN CHINA (Series C)
Vol. 46
a series of fragments of the PSA promoter that contained mutant type of RFA. Then the constructs were co-transfected with an AR expression vector. As shown in fig. 1(b), three of four mutation sites are in the direct repeat (i.e., CAGGGA), but the fourth out of it. As seen in fig. 1(a), nucleotide substitutions in RFA significantly diminished the androgen induction for transcription of reporter gene (CAT). The only difference of nucleotide substitution between 406 (A-1) and 406 (A-2) is that the former has an additional mutation site outside the RFA, but the CAT activities between the two ones are not significantly different, which suggests that the RFA is indispensable for AR-mediated transactivation, moreover, some nucleotides within RFA are necessary for its cooperation. In addition, 1 copy of wild RFA, 3 copies of wild RFA and 3 copies of mutant RFA were inserted separately at the upstream of the tk promoter of expression vector pBLCAT2, then the constructs were co-transfected with AR expression vector into PC-3 cells. As shown in fig. 2, 3 copies of wild RFA seem to exert androgenic inducibility for a heterologous promoter (tk promoter) without ARE, but 3 copies of mutant RFA do not, which suggests that the action of RFA is independent of ARE.
Fig. 1. The effect of mutations in RFA of PSA promoters on androgen inducibility. (a) PC-3 cells in duplicate plates were cotransfected with designated PSA promoter-pBLCAT3 constructs (4 Pg/plate) and a human AR expression vector (0.2 Pg/plate) using Lipofectamine (12 Pg/plate). Cells were then treated with or without 3.2 nmol/L Mib for 24 h. Cell extracts were prepared and used for protein and CAT activity assays. The results are expressed in (CPM/min)/mg protein. Error bars indicate the standard error of the mean of three separate experiments. (b) The actual nucleotide substitutions in the correspondence constructs shown in (a).
No. 2
RFA REGULATORY SEQUENCE-BINDING PROTEIN
189
Fig. 2. Androgenic induction of RFA constructs. Transfections were performed in PC-3 cells either in the presence (+) or absence () of 3.2 nmol/L Mib. The left panel represents the schematic promoter structure of the constructs. One or three copies of the designated RFA were inserted before tk promoter. Parental pBLCAT2 was used as a control in the transfections. The results of the CAT activity are expressed in (CPM/min)/mg protein. Error bars indicate the standard error of the mean of three separate transfections.
2.2
The specificity of RFA binding with protein The nuclear extracts from PC-3 and LNCaP cells were used for in vitro binding assay with
ds-probes containing monomeric RFA, including wild type RFA and a series of mutant RFA (A-2, A-3, A-4) (seen in table 2) probes. As seen in fig. 3, the RFA can form RFA-protein complex with nuclear protein, which can be competitively inhibited by 100-fold molar excess of unlabeled RFA itself. However, the mutant RFA loses the ability for competitively binding with the protein. The results suggest that the 396ü382 DNA fragment is necessary for nuclear protein binding. Both PC-3 and LNCaP nuclear extracts produce the same band shift patterns. In addition, PSA ARE cannot block the band formation, which suggests that the RFA-binding protein is non-AR
32
P-labeled double-stranded monomeric RFA was incubated with LNCaP or PC-3 cell
Fig. 3.
EMSA of RFA and prostatic nuclear extracts.
nuclear extracts which had been pre-incubated for 10ü30 min with or without a 100-fold molar excess of competing DNA for 20ü30 min.
190
SCIENCE IN CHINA (Series C)
Vol. 46
factor. Using the Genetics Computer Group (GCG) program for DNA pattern search, an NF-NB-like sequence was found within the 396/382 area, so it seemed likely that nuclear protein binding to RFA might be an NF-NB-like factor. As seen in fig. 3, however, the NF-NB sequence was not able to inhibit the formation of probe-protein complexes. Thus, this possibility is excluded. 2.3
The purification of the RFA-binding protein by affinity chromatography The RFA-binding proteins were isolated using ds-probes encompassing 2 copies of RFA. To most sequence-specific DNA-binding proteins, 1 mol/L KCl is enough to release the protein from DNA, even some protein factors lose the ability of binding to DNA only with 0.6 mol/L KCL, but a few protein factors require 2 mol/L Na+/K+ to be released[9]. To assure the salt concentration of elution for purifying target protein by affinity chromatography, EMSA was performed to observe the effect of different concentration of K+ on target protein binding to RFA. As shown in fig. 4, 1 mol/L, 1.5 mol/L and 2 mol/L KCl all made the retarded band disappear. The elution salt concentration used in this report is 1 mol/L KCl, for the exorbitant salt concentration is inconvenient for the next procedure. Whether the elution contained desired protein should be demonstrated by EMSA. As seen in fig. 5, diluted elution generated the same pattern band in the same place as total nuclear protein, which can be completely blocked by 125-fold molar excess of unlabeled monomeric RFA probes (wild type). The results indicate that the elution contains target protein.
Fig. 4. EMSAüüthe influence of different concentra-
Fig. 5.
tion of KCl on the RFA binding with protein: 1, 0.1 mol/L KCl; 2, 1 mol/L KCl; 3, 1.5 mol/L KCl; 4, 2 mol/L KCl.
RFA-binding proteins. 1, Probe without protein; 2, probe with nuclear protein of PC-3 cells; 3, probe with purified RFA-binding protein; 4, probe with purified RFA-binding protein and 125 folds of unlabeled probe.
EMSAüüthe function detection of purified
No. 2
RFA REGULATORY SEQUENCE-BINDING PROTEIN
191
2.4 SDS-PAGE & identification of the purified protein After detecting the function of purified protein by EMSA, the rest elution was run in 10% SDS-PAGE. As shown in fig. 6, purified proteins seem to be a single 36-kD protein, which indicates that relatively high purity of target proteins was obtained. Amino acid sequence and mass spectrometric analysis show the band contains two proteins that are highly homologous with hnRNP A1, A2 respectively. 3
Discussion
Our previous study indicated that deletion of the Fig. 6. SDS-PAGE of purified RFA-binding protein. 1, 36 bp fragment encompassing RFA at the upstream BSA; 2, marker; 3, purified RFA-binding protein; 4, ARE of PSA promoter can significantly diminish the PC-3 nuclear protein. AR-mediated transactivation for PSA gene[4]. The present study indicates that nucleotide substitution in RFA can significantly diminish the AR-mediated transactivation, so RFA may be a coregulatory element. The EMSA shows RFA can bind specifically with some nuclear protein of prostate cells, which can be competitively blocked by RFA self, but not by some known cis-element such as ARE, NF-NB. The result suggests that RFA-binding protein is a non-receptor protein factor. It should be noted that nucleotide substitution within RFA not only significantly diminishes the AR induction, but also makes itself lose the ability to bind with protein. Moreover, the tk promoter inserted with 3 copies of RFA at its upstream can be activated by ligand-bound AR without ARE, all of which suggest that RFA/binding protein might be cooperatively involved in androgen inducing for PSA gene via direct/indirect interaction with AR. So the RFA-binding protein may be a novel AR-associated coactivator. The preliminary protein identification shows that RFA-binding proteins are highly homologous with hnRNP A1, A2. hnRNP A1, A2 are components of heterogeneous nuclear ribonucleoprotein (hnRNP) superfamily. hnRNP are hnRNA-binding proteins among the most abundant proteins in the nucleus. All of A proteins (Mr = 34000ü38000 by SDS-PAGE, pI = 8.4ü9.1) have a similar general structure, i.e., they contain two RNA-binding domains and a glycine-rich auxiliary domain at the carboxyl terminus. So far there have been several homologous hnRNP A-like proteins being characterized. The full range of functions and mechanism of action of hnRNP proteins are still unknown. Initial study on hnRNP proteins indicated that its main function is forming complex with hnRNA and snRNPs in nucleus, involved in post-transcriptional processing and transporting to cytoplasm of pre-mRNA. Several recent reports show that some members of hnRNP are multifunctional proteins involved in several regulatory pathways, which not only bind
192
SCIENCE IN CHINA (Series C)
Vol. 46
to single-strand RNA, but also bind to the specific single/double strand DNA and interact with other protein factors. The hnRNP K, for example, can bind and transactivate a cis-element found within the human c-myc promoter, also interact with transcription factor C/EBPE to inhibit the expression of agp gene that depends on the transactivation of C/EBPE, and be involved in expression control of many other genes[10,11]. Chen et al.[12] found that homologous protein of hnRNP A1, A2—VDRE-BP1/2 bind specifically to the single/double-strand oligonucleotides harboring the Vitamin D response element (VDRE), which acts to inhibit the expression of target gene as dominant-negative regulator by competing with the Vitamin D receptor for binding to the VDRE. Two proteins, termed as p38 and p40 that belong to the hnRNP A/B subgroup, can bind specifically to the Rat spi 2 gene GAGA box and enhance its transactivation[13]. Moreover, recently the hnRNP A/B subgroup proteins have been found overexpressed in many kinds of tumor tissue[14,15]. All of these studies suggest hnRNP A1, A2 are multifunctional proteins involved in many metabolism pathways and related to the occurrence and development of tumor. From the above we can infer that RFA-binding proteins as homologous proteins of hnRNP A1, A2 may cooperate with the AR-mediated transactivation as coactivator to regulate the expression of PSA gene. Our study is still in preliminary stage, and the property and mechanism of RFA-binding protein need to be further studied. References 1.
Woodward, R. N., Li, M., Holland, L. J., Novel accessory factor-binding site required for Glucocorticoid regulation of the gamma-fibrinogen subunit gene from xerropus Laevis, Mol. Endocrinol., 1997, 11(5): 563ü576.
2.
Ho, K. C., Marche, K. B., Tan, J. A. et al., A complex response element in intron 1 of the androgen-regulated 20-kDa protein gene displays cell type-dependent androgen receptor specificity, J. Biol. Chem., 1993, 268: 27226ü27235.
3.
Riegman, P. H., Vlietstra, R. J., van der Korput, J. A. et al., The promoter of the prostate-specific antigen gene contains a
4.
Zhang, J. Y., Pang, W. Q., Zhang, L. Y. et al., Primary analysis of a regulation sequence in the promoter region of human
functional androgen responsive element, Mol. Endocrinol., 1991, 5: 1921ü1930.
prostate specific antigen gene, Chin. J. Med. Genet. (in Chinese), 1997, 14(1): 24ü27. 5.
Zhang, J. Y., Pang, W. Q., Wang, X. et al., A sequence associated with the regulation of androgen in PSA gene promoter, Prog. Biochem. Biophys. (in Chinese), 1997, 24: 228ü232.
6.
Bagchi, M. K., Tsai, S. Y., Tsai, M. J. et al., Identification of a functional intermediate in receptor activation in progester-
7.
Murtha, P., Tindall, D. J., Young, C. Y. F., Androgen induction of a human prostate-specific kallikrein, hKLK2: characteri-
one-dependent cell-free transcription, Nature, 1990, 345: 547ü550.
zation of an androgen response element in the 5Ąpromoter region of the gene, Biochemistry, 1993, 32(25): 6459ü6464. 8.
Renie, P. S., Bruchovsky, N., Leco K, J. et al., Characterization of two cis-acting DNA elements involved in the androgen regulation of the probasin gene, Mol. Endocrinol., 1993, 7: 3ü26.
9.
Marshak, D. R., Kadonaga, T. K., Burgess, R. R. et al., Strategies for Protein Purification and Characterization (in Chinese, translated by Zhu Houchu), Beijing: Science Press, 1997, 101ü117.
10.
Tomonaga, T., Levens, D., Heterogeneous nuclear ribonucleoprotein K is a DNA-binding transactivator, J. Biol. Chem., 1995, 270: 4875ü4881.
No. 2 11.
RFA REGULATORY SEQUENCE-BINDING PROTEIN
193
Miau, L. H., Chang, C. J., Shen, B. J. et al., Identification of heterogeneous nuclear ribonucleoprotein K (hnRNP K) as a repressor of C/EBP beta-mediated gene activation, J. Biol. Chem., 1998, 273: 10784ü10791.
12.
Chen, H., Hu, B., Allegretto, E. A. et al., The vitamin D response element-binding proteinüü A novel dominant-negative regulator of vitamind-directed transactivation, J. Biol. Chem., 2000, 275(45): 35557ü35564.
13.
Leverrier, S., Cinato, E., Paul, C. et al., Purification and cloning of type A/B hnRNP proteins involved in transcriptional
14.
Kamma, H., Fujimoto, M., Fujiwara, M. et al., Interaction of hnRNP A2/B1 isoforms with telomeric ssDNA and the in vi-
activation from the Rat spi 2 gene GAGA box, Biol. Chem., 2000, 381(11): 1031ü1040.
tro function, Biochem. Biophys. Res. Commun., 2001, 280(3): 625ü630. 15.
Man, Y. G., Martinez, A., Avis, I. M. et al., Phenotypically different cells with heterogeneous nuclear ribonucloprotein A2/B1 overexpression show similar genetic alterations, Am. J. Respir. Cell Mol. Biol., 2000, 23(5): 636ü645.