This Article Abstract Full Text (PDF) Data Supplement All Versions of this Article: 95/11/1058    most recent 01.RES.0000149166.33833.08v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Thiele, B.-J. Articles by Regitz-Zagrosek, V. Search for Related Content PubMed PubMed Citation Articles by Thiele, B.-J. Articles by Regitz-Zagrosek, V. Related Collections Biochemistry and metabolism Other myocardial biology Physiological and pathological control of gene expression

(Circulation Research. 2004;95:1058.)
© 2004 American Heart Association, Inc.

Molecular Medicine

RNA-Binding Proteins Heterogeneous Nuclear Ribonucleoprotein A1, E1, and K Are Involved in Post-Transcriptional Control of Collagen I and III Synthesis

Bernd-Joachim Thiele^*, Anke Doller^*, Thilo Kähne, Reinhard Pregla, Roland Hetzer, Vera Regitz-Zagrosek

From the Institut für Physiologie (B.-J.T.) and Center for Cardiovascular Research (A.D., V.R.-Z.), University-Medicine Berlin; Deutsches Herzzentrum Berlin (R.P., R.H., V.R.-Z.); and Institut für Experimentelle Innere Medizin (T.K.), Otto-von-Guericke-Universität Magdeburg, Germany.

Correspondence to Dr Anke Doller, Hessische Str.3-4, 10115 Berlin, Germany. E-mail anke.doller{at}charite.de

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
Collagen types I and III, coded by COL1A1/COL1A2 and COL3A1 genes, are the major fibrillar collagens produced by fibroblasts, including cardiac fibroblasts of the adult heart. Characteristic for different cardiomyopathies is a remodeling process associated with an upregulation of collagen synthesis, which leads to fibrosis. We report identification of three mRNA-binding proteins, heterogeneous nuclear ribonucleoprote (hnRNP) A1, E1, and K, as positive effectors of collagen synthesis acting at the post-transcriptional level by interaction with the 3'-untranslated regions (3'-UTRs) of COL1A1, 1A2, and 3A1 mRNAs. In vitro, binding experiments (electromobility shift assay and UV cross-linking) reveal significant differences in binding to CU- and AU-rich binding motifs. Reporter gene cell transfection experiments and RNA stability assays show that hnRNPs A1, E1, and K stimulate collagen expression by stabilizing mRNAs. Collagen synthesis is activated via the angiotensin II type 1 (AT₁) receptor. We demonstrate that transforming growth factor-ß1, a major product of stimulated AT₁ receptor, does not activate solely collagen synthesis but synergistically the synthesis of hnRNP A1, E1, and K as well. Thus, post-transcriptional control of collagen synthesis at the mRNA level may substantially be caused by alteration of the expression of RNA-binding proteins. The pathophysiological impact of this finding was demonstrated by screening the expression of hnRNP E1 and K in cardiovascular diseases. In the heart muscle of patients experiencing aortic stenosis, ischemic cardiomyopathy, or dilatative cardiomyopathy, a significant increase in the expression of hnRNP E1, A1, and K was found between 1.5- and 4.5-fold relative to controls.

Key Words: collagen • cardiac fibrosis • posttranscriptional regulation • mRNA-binding proteins • 3'-UTR

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
Deposition of collagen fibers in the myocardial interstitium is a landmark for the remodeling process seen in a variety of cardiovascular diseases of quite different etiology, such as ischemic heart disease, cardiac hypertrophy, arterial hypertension, or the infarcted heart.

In the failing heart, several humoral, autocrine, and paracrine systems are activated and have been identified as potentially important mediators of collagen synthesis^1,2 and degradation by collagenases (metalloproteinases).³ Many studies suggest that fibrosis is linked to an activation of the local renin-angiotensin system.^4,5 It is known that angiotensin II (Ang II) directly stimulates collagen synthesis via Ang II type 1 (AT₁) receptors. Ang II is an inducer of transforming growth factor-ß1 (TGF-ß1) or the local aldosteron system, both positive effectors of collagen synthesis.^6,7

Collagen types I and III (Col I and Col III) are the major fibrillar collagens produced by fibroblasts, including cardiofibroblasts in the adult heart.⁸ The triple-helical collagen molecules are assembled by independently coded 1(I), 2(I), and 1(III) chains. Expression of the collagen genes (COL1A1, COL1A2, and COL3A1) is regulated at the transcriptional and post-transcriptional levels. Much is known about the complex transcription regulation of the collagen genes⁹ and post-translational modification¹⁰; however, less is known about post-transcriptional processes focusing on mRNA stability and translation. From studies in hepatic fibrosis, the importance of mRNA-specific post-transcriptional control in collagen synthesis was suggested.^11,12 Molecular targets for these processes are 5'-untranslated regions (5'-UTRs) and 3'-UTRs of the collagen mRNAs interacting with RNA-binding regulative proteins.^12–14 Stem-loop structures in the 5'-UTR of all three mRNAs are involved in function of COL1A1, COL1A2, and COL3A1 mRNAs.¹² Whether common 3'-UTR–mediated principles also play a role in coordinate expression is unknown. For COL1A1 3'-UTR, a CU-rich element binding CP (synonyms CP1, PCBP, heterogeneous nuclear ribonucleoprote [hnRNP] E1) has been described that is responsible for mRNA stabilization.^13,15,16 Moreover, CU-rich elements involved in mRNA stabilization and translational control have been found in several mRNAs such as those coding for the globins,^17,18 15-lipoxygenase,^19,20 poliovirus,²¹ erythropoietin,²² tyrosine hydroxylase,²³ folate receptor,²⁴ androgen receptor,²⁵ or renin.²⁶

In this study, we describe the identification of three RNA-binding proteins, hnRNP A1, E1, and K, interacting differently with 3'-UTRs of collagen 1A1, 1A2, and 3A1 mRNAs and investigate their role in coordinate expression of collagen I and III genes. Furthermore, data of patients with cardiac diseases are shown proving that fibrotic alteration of the myocardium is associated with an upregulation of the expression of collagen mRNA-stabilizing proteins.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
Plasmids
3'-UTR transcription templates were generated by polymerase chain reaction (PCR) from human genomic DNA and cloned into the pCRII-TOPO vector (Invitrogen) or in pGL3 promoter (Promega). The 283-nt 3'-UTR1 collagen 1A1 mRNA sequence was amplified by PCR with primers 5'-ACTCCCTCCATCCCAACCT-3' and 5'-ACCAAGCTTCCTTTTTTAAAAAC-3'; the 312-nt 3'-UTR1 collagen 1A2 mRNA sequence with primers 5'-ATGAACTCAATCTAAATTA-3' and 5'-GCAAATGTTTATGGTTTTTATT-3'; and the 284-nt 3'-UTR1 collagen 3A1 mRNA sequence with primers 5'-ACCAAACTCTATCTGAA-3' and 5'-GTTATCAAAAGAGTGTTGAAGTTT-3'. pGS5-E1, pGS5-A1, and pGS5-K were gifts from A. Ostareck-Lederer (Heidelberg). Expression cloning and affinity chromatography of hnRNP E1 and K have been described previously.²⁷

Cell Culture
Human cardiac fibroblasts were isolated from explanted, end-stage failing adult human hearts, as described previously.²⁸ The resulting cell culture consisted almost exclusively of fibroblasts (96% to 98%). Primary fibroblast or HT1080 cells were cultured in DMEM using standard techniques.

Protein Extracts
HT1080 cells were resuspended in lysis buffer containing 10 mmol/L Tris, pH 7.5, 140 mmol/L NaCl, 1 mmol/L EDTA, 25% glycerol, 0.1% sodium dodecyl sulfate, 0.5% Nonidet P-40, 0.1 mmol/L dithiothreitol, and the "complete mini" protease inhibitor cocktail (Roche). S100 extracts were prepared as described previously.²⁶

RNA Preparation and RT-PCR
Total RNA of cardiac fibroblast was isolated by the guanidine thiocyanate method following the manufacturer protocol (RNAzol; Biozol). Total RNA was treated with RNase-free DNase I using a standard protocol. The purity of the RNA was controlled by the A₂₆₀/A₂₈₀ quotient. The integrity of the purified RNA was controlled by gel analysis. First-strand cDNA was synthesized in a final volume of 25 µL containing first-strand buffer, 500 ng human cardiac fibroblast RNA, 600 ng random hexamer primer (Gibco), and Superscript reverse transcriptase (20 U/µg RNA; Gibco). The mRNA levels for collagen I and III, hnRNP E1, hnRNP A1, hnRNP K, and the reference gene GAPDH (primer data available in an online supplement, available at http://circres.ahajournals.org) were determined by using "hot start" real-time PCR procedure with SYBR green that was validated with respect to the reproducibility and linearity within the measuring range. PCR was performed in duplicate with the Taqman (ABI 7000; Perkin-Elmer).

To correct for potential variances between samples in mRNA extraction and RT efficiency, the mRNA content of collagen I and III, hnRNP E1, hnRNP A1, and hnRNP K was related to the mRNA content of the stable expressed reference gene GAPDH from the same aliquot.

In Vitro Transcription, Electromobility Shift Assay, UV Cross-Linking, Blotting Methods, and Matrix-Assisted Laser Desorption Ionization– Time-of-Flight–Mass Spectrometry
In vitro transcription, electromobility shift assay (EMSA), UV cross-linking, Northern blotting and matrix-assisted laser desorption ionization–time-of-flight–mass spectrometry (MALDI-TOF-MS) of affinity-purified RNA-binding proteins were performed as described previously.²⁶ For each experiment, 10 000 cpm (EMSA) or 100 000 cpm (UV cross-linking) ³²P-labeled transcript and 70 µg HT1080 cytoplasmic extract was used. For Western blotting, the membranes were probed with anti-hnRNP E1, anti-hnRNP A1 antibody (a gift from G. Dreyfuss, University of Pennsylvania, Philadelphia), or collagen I antibody (DPC Biermann). Primary antibodies were used at 1:25000 (hnRNP E1), 1:1000 (hnRNP A1), or 1:2000 (collagen) in Tris-buffered saline Tween 20 (TBS-T) with 5% Chemiblocker overnight at 4°C. Immunoblots were incubated with horseradish peroxidase–conjugated antibody in TBS-T to visualize the immunoreactive bands with an enhanced chemiluminescence system (Amersham Pharmacia).

Transient Transfection and Luciferase Assay
HT1080 cells were grown to 80% confluence before transfection with 25 ng pGL3promoter collagen–3'-UTR constructs and different amounts of the eukaryotic expression plasmid coding for the hnRNP A1, E1, or K proteins using FuGENE 6 transfection reagent (Roche). Cells were cotransfected simultaneously with 0.5 µg pSV–ß-galactosidase (ß-gal; Promega) as a transfection control. The transfected cells were fed complete growth medium for 48 hours and lysed in reporter lysis buffer (Promega). Luciferase assays were performed in a Berthold Lumat LB9501 luminometer. ß-Gal activity was assayed as described in the protocol of the manufacturer (ß-gal enzyme assay system; Promega). Reporter gene activities were normalized to cotransfected controls.

TGF-ß Stimulation
Cardiac fibroblasts were grown until 90% to 95% confluence. Before treatment, they were serum-deprived for 48 hours. TGF-ß1 (Calbiochem) was used in a range at 10 ng/mL (triplicates from each fibroblast preparation). In control experiments, cells received equivalent amounts of vehicle (1% BSA/4 mmol/L HCl). Cells were grown under these conditions for 6 and 12 hours.

Data Analysis
All reported values are mean±SEM. Statistical analysis was performed using Student t test, with a P value of <0.05 regarded as significant.

Patients
Left ventricular myocardial samples from 18 patients with aortic stenosis (AS), 26 patients with heart failure attributable to dilatative cardiomyopathy (DCM), 6 patients with ischemic cardiomyopathy (ICM), and 20 control subjects were analyzed (patient data available in the online supplement). Small tissue samples from patients with AS were obtained by biopsy of the left ventricular septum at elective aortic valve replacement surgery. Tissue samples from patients with DCM or ICM were obtained as left ventricular septum biopsies at cardiac catheterization or at orthotopic heart transplantation. The control group was composed of donor hearts that were not used for logistic reasons. They had normal systolic cardiac function; cardiac history was absent, and postmortem histology was normal. Written informed consent was obtained from all patients. Experimental protocols were accepted by the local ethical committee. The investigation conforms with the principles outlined in the declaration of Helsinki.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
Interaction of HT1080 Cytosolic Proteins With Collagen 3'-UTRs
To investigate the influence of mRNA-binding proteins on coordinate expression of the collagen genes, the 3'-UTRs of the mRNAs coding for the human collagen 1(I), 2(I), and 1(III) chains were cloned into the expression vector TOPO to allow in vitro transcription (Figure 1). The three collagen type I and III genes are transcribed into more than one mRNA. They differ in the length of their 3'-UTRs by use of alternative polyadenylation signals.^29–31 For our study, those three collagen I and III mRNAs were chosen ending with the first of the alternative poly-A signals. These transcripts represent abundant collagen mRNAs in fibroblasts when probed by Northern blotting. The sequences of the 3'-UTRs, including putative binding sites for RNA-binding proteins, are schematically depicted in Figure 1.

View larger version (47K): [in this window] [in a new window]   Figure 1. Structure of collagen I and III mRNA 3'-UTRs. A, A schematic representation of the 3'-UTRs of human collagen 1A1, 1A2, and 3A1 cDNAs ending with the first polyadenylation signal (aataaa) is shown. The hatched areas represent potential binding sites for hnRNPK (light gray), hnRNPE1 (dark gray), and hnRNPA1 (black). B, Sequences of the cloned human collagen 1A1, 1A2, and 3A1 cDNA 3'-UTR, schematically depicted in A, are shown. CU-rich structures with the consensus for hnRNPE1 binding are underlined, and those with the consensus for hnRNPK binding are boxed. AU-rich elements appear in italics and are boxed. Gene bank accession numbers are: NM000088 COL1A1 nucleotides 4512–4798, NM000089 COL1A2 nucleotides 4238–4550, and NM000090 COL3A1 nucleotides 4501–4785. COL 3'-UTRs were either cloned into the pCR-TOPO vector and in vitro transcribed for EMSA (Figure 2) or UV cross-linking (Figure 3), or were cloned 3' to the luciferase coding region of pGL3 vector used in reporter gene assays (Figure 4).

To examine the ability of the collagen mRNA 3'-UTRs to interact with potential regulatory proteins, EMSAs were performed. To this purpose, ³²P-labeled transcripts were incubated with HT1080 cytosolic proteins, and the generation of mRNP complexes was analyzed. All three transcripts formed stable mRNP complexes (Figure 2). For identification of regulatory proteins potentially involved in turnover of collagen mRNAs, we also probed for candidate RNA-binding proteins hnRNP E1 and hnRNP K. hnRNP E1 was shown previously to bind to a CU-rich 3'-UTR element of COL1A1 mRNA, thereby stabilizing the mRNA.¹³ hnRNP K, another K-domain protein, could be also a good candidate because of a similar oligo-C binding preference.³² To this end, binding of recombinant hnRNP E1 and K to all three transcripts was examined (Figure 2). The 3'-UTR transcripts of COL1A1 and COL1A2 bound hnRNP E1 and K. However, COL3A1 3'-UTR interacted only with hnRNP K but not with E1.

View larger version (105K): [in this window] [in a new window]   Figure 2. Binding of cytoplasmic HT1080 proteins and recombinant hnRNP E1 and K to collagen mRNA 3'-UTRs analyzed by EMSA. 32P-labeled COL mRNA 3'-UTR transcripts (Figure 1) were incubated with HT1080 cytoplasm or recombinant proteins hnRNP E1 or K and examined by EMSA. Because of peculiar interaction properties, the COL1A2 transcript forms a dimer, which runs together with the monomer as a double band.

Identification of hnRNPs A1, E1, and K as Trans-Factors Binding to Collagen 3'-UTRs
For more detailed information about the nature of the bound HT1080 cytosolic protein, UV cross-linking experiments were used. Labeled protein patterns of relatively low complexity were obtained (Figure 3A, lanes 2 through 4). These were compared with the cross-linking properties of recombinant E1 and K (Figure 3A, lanes 5 through 10). Identical electrophoretic migration of two dominant 39-kDa and 66-kDa bands (COL1A1, lane 1) strongly indicated that hnRNP E1 and K were the main cytosolic binding partners interacting with 3'-UTR of COL1A1 mRNA. This finding was verified by immunological techniques: antibodies against E1 and K immobilized on Sepharose beads recognized cross-linked E1 and K in cytosolic extracts (data not shown). The cross-linking pattern of COL1A2 and COL3A1 3'-UTR transcripts was consistent with the gel shift experiments: both bound hnRNP K, although with lower efficiency, in particular COL3A1. hnRNP E1 also interacted with COL1A2 but did not bind to COL3A1.

View larger version (42K): [in this window] [in a new window]   Figure 3. Cross-linking of HT1080 cytosolic proteins and recombinant hnRNP E1 and K to collagen 3'-UTRs. A, As described in the Figure 2 legend, labeled COL 3'-UTR transcripts were generated and UV cross-linked to cytosolic HT1080 proteins or recombinant E1 and K. Transfer-labeled proteins were separated by SDS-PAGE and visualized by autoradiography. B, Identification of hCOL mRNA 3'-UTR–binding proteins by MALDI-TOF-MS. After tryptic digestion of affinity-purified proteins, peptides were identified by MALDI-TOF-MS analysis (MascotSearch).

In contrast to COL1A1, the COL1A2 and COL3A1 transcripts bound a few additional polypeptides: COL1A2, a main band of 79 kDa, and COL3A1, three of 35, 62, and 85 kDa. To verify the hnRNPs E1 and K data and to search for the identity of the other cross-linked polypeptides, an RNA affinity chromatography approach was applied and coupled with MALDI-TOF-MS for their identification as described previously.²⁶ Apart from the verification of hnRNP E1 (39 kDa) and K (66 kDa), two additional collagen mRNA-binding proteins were identified: P85 as the Zn-finger protein ZNF224,³³ P62 as a proteolytic fragment of hnRNP K, and P35 as hnRNP A1 (Figure 3B). The 79-kDa band gave no unambiguous MALDI-TOF-MS spectra.

Functional Influence of hnRNPs A1, E1, and K on Collagen Expression
According to our results, A1 and A2 chains of COLI seem to be regulated in a similar way. Therefore, we decided to concentrate on the differences between type I and type III chains, studying COL1A1 mRNA as representative for type I collagen. To investigate the influence of hnRNP proteins A1, E1, and K, which have a well-documented impact on mRNA metabolism and function, we performed reporter gene assays. Because of the poor data on ZNF224, which is primarily known as a transcription factor, it was omitted from our analysis. Hybrid plasmids were constructed carrying the luciferase reporter gene as coding and the 3'-UTRs of three human collagen genes as regulatory unit, as described in Figure 1. HT1080 cells were transfected with the LucCOL–3'-UTR constructs and cotransfected with plasmids expressing human hnRNP proteins E1, K, or A1. The concentration-dependent transfection was optimized previously to overexpress hnRNP proteins 3-fold over their endogenous levels in HT1080 cells using Western blots (data not shown). The results of the cotransfection experiments are summarized in Figure 4. The expression of reporter luciferase plasmid carrying the COL1A1–3'-UTR (LucCOL1A1) was stimulated 2- to 5-fold when cotransfected with hnRNP expression plasmids E1, K, or A1 (E1 2x, K 2x, and A1 5x) in a concentration-dependent manner (Figure 4, lanes 1 through 9). The same experiment was repeated with LucCOL3A1–3'-UTR. hnRNP K and A1 stimulated the expression 4- and 6-fold, whereas E1 had no influence (lanes 10 through 15). To investigate the capability of hnRNPs E1 and K to act as RNA stabilizers in vitro, RNase degradation assays²⁶ were performed using luciferase transcripts carrying collagen 3'-UTRs and recombinant hnRNP E1 and K (Figure 4B). The experiments show that the addition of E1 and K resulted in an 4-fold increase in in vitro half-life time of LucCOL1A1 reporter transcript (7.5 versus 30 minutes). The stability of LucCOL3A1 transcript increased 2-fold in the presence of K (27.5 versus 50 minutes), whereas E1 had no effect. This demonstrates that collagen mRNA 3'-UTRs can mediate mRNA stabilization via the action of different independently acting RNA-binding proteins.

View larger version (31K): [in this window] [in a new window]   Figure 4. A, Influence of hnRNP A1, E1, or K on the expression of luciferase-COL3'-UTR reporter genes. A total of 105 HT1080 cells were transiently transfected with 0.025 µg of Luc reporter construct containing the COL1A1 3'-UTR (lanes 1 through 9) or COL3A1 3'-UTR (lanes 10 through 18; Figure 1) and 0.5 µg of pSV–ß-gal as a transfection control. In addition, 0.025 µg (lanes 1 and 10), 0.25 µg (lanes 2 and 11), or 2.5 µg (lanes 3 and 12) of pSG5-E1; 0.025 µg (lanes 4 and 13), 0.25 µg (lanes 5 and 14), or 2.5 µg (lanes 6 and 15) of pSG5-K; 0.025 µg (lanes 7 and 16), 0.25 µg (lanes 8 and 17), or 2.5 µg (lanes 9 and 18) of pSG5-A1 expression plasmids were cotransfected. Luciferase activity was determined luminometrically. Transfections with the Luc plasmid alone were set at 100%. Data (means±SD) represent the results of three independent experiments (data are shown in online supplement; S3). P values were 0.05 and 0.01 vs control conditions. *P<0.01; *P<0.05. B, Influence of hnRN proteins E1 and K on stability of LucCOL3'-UTR transcripts in vitro. 32P-labeled LucCOL3'-UTR capped and polyadenylated transcripts were generated and subjected to in vitro degradation as described previously.26 For each experiment, 2·104 cpm transcript was incubated with 20 µg HT1080 cytoplasm as RNase source, and with recombinant E1 or K (for COL IA1 20 ng E1/K, each; and for COL IIIA1 20 ng E1 or K). Aliquots were removed after the times indicated and analyzed by agarose electrophoresis, blotting onto membranes, and autoradiography. Stability of transcripts was determined by quantifying autoradiographic signals using the program Molecular Analyst and graphical estimation of half-life times.

Synergistic Induction of Collagens and hnRNPs A1, E1, and K by TGF-ß1
TGF-ß1 is a known effector in the signaling pathway of collagen expression and the synthesis of other connecting tissue proteins involved in fibrotic processes.^5,34 We asked whether the induction of collagens I and III by TGF-ß1 is likewise accompanied by an upregulation of collagen mRNA-binding proteins. The well-known effect of TGF-ß1 on the induction of the connective tissue growth factor (CTGF) served as a positive control.³⁵ Collagens I and III as well as CTGF accumulated under the action of TGF-ß1 for 6 to 12 hours by a factor of 2. Under the same conditions, the level of the mRNAs of all three hnRNP proteins A1, E1, and K increased significantly between 40% and 70% compared with the controls without TGF-ß1 (Figure 5A). Immunological quantification of RNA-binding proteins for which antibodies were available revealed that hnRNPs A1 and E1 were upregulated by factors 5.2 and 3.4, respectively. As controls, COLI and GAPDH were also quantified. As expected, COLI was induced 4.3-fold, whereas GAPDH remained constant. This demonstrated that COL mRNA-binding proteins are synergistically induced like extracellular matrix proteins and may contribute significantly to the observed post-transcriptional upregulation of collagen synthesis.

View larger version (51K): [in this window] [in a new window]   Figure 5. Induction of mRNAs coding for Col I, Col III, hnRNP E1, hnRNP K, hnRNP A1, and CTGF by TGF-ß1. Human cardiac fibroblasts were cultured in the presence or absence of TGF-ß1 for the times indicated. A, Total RNA was extracted and mRNAs were quantitated by real-time RT-PCR. Data of nonstimulated cells are shown in black and of stimulated in gray. Values are normalized to GAPDH mRNA quantification. **P<0.01; *P<0.05. B, Analysis of collagen and collagen mRNA-binding proteins. Cytosolic proteins were isolated according to Materials and Methods and analyzed by Western blots. Comparable results were obtained in three independent experiments (data are available in online supplement; S4).

Selected Cardiomyopathies Are Characterized by an Increased Expression of COL mRNA-Binding Proteins
We were able to show in vitro and in vivo that COL mRNA-binding proteins contribute to the process of upregulation of collagen synthesis. Thus, their increase could be of pathophysiological significance for the mechanism of fibrosis generation. To test this hypothesis, we analyzed heart biopsy specimens of patients experiencing different heart diseases, such as AS, DCM, or ICM, for the mRNA level of COL mRNA-binding proteins hnRNP E1, A1, and K by real-time PCR (Figure 6). Indeed, particularly in AS, a striking increase in the E1, A1, and K mRNA level by the factors 3.5, 6.5, and 4.5 was observed. In our previous studies, samples of a comparable patient cohort with AS, partially overlapping with that analyzed in this study, were characterized by an increased collagen expression.⁴ In ICM, the E1/K expression was elevated 2-fold. hnRNP E1, A1, and K were not significantly altered in DCM.

View larger version (13K): [in this window] [in a new window]   Figure 6. Expression of the genes coding for hnRNP E1 and K in hearts from patients experiencing different cardiac diseases. Total RNA was extracted from patient biopsy specimens of the left ventricle and mRNAs of hnRNPs A1, E1, and K were quantified by real-time RT-PCR. Values were normalized to GAPDH mRNA concentrations. Controls n=20; AS n=18; DCMs n=26; ICMs n=6. **P<0.001; *P<0.05.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

 
We present molecular and clinical data that hnRNP proteins binding to collagen mRNAs are involved in the mechanism of induction of collagen synthesis and in development of cardiac fibrosis. Several lines of evidence indicate that expression of the genes coding for the two collagens mainly elevated in organ fibrosis, collagens type I and III, is controlled predominantly post-transcriptionally at the mRNA level.¹³ Stem-loop structures in the 5'-UTR of the three COL1A1, 1A2, and 3A1 mRNAs are involved in translational control.¹² Moreover, the 3'-UTR of at least one member of the family, COL1A1 mRNA, has been shown to be a target for mRNA-binding protein CP (hnRNP E), which is involved in mRNA stabilization.¹¹ The coordinated expression of type I and III collagens let us presume further common post-transcriptional mechanisms targeting at the mRNA 3'-UTR level.

An inspection of the sequences after consensus motifs for RNA-binding proteins predicted hnRNP E1 binding to COL1A1 and COL1A2, hnRNP A1 binding to COL3A1, and hnRNP K binding to all three 3'-UTRs (Figure 1). Experimentally, we can show that indeed, all three behaved accordingly and fulfilled this prediction (Figures 2 and 3). The consensus motifs for hnRNP E1 and K binding have been examined thoroughly using the SELEX method.³² That revealed, briefly for K, a motif centering in the sequence UCCCA/UA/U, and for E1, a cluster of 2 to 3 U of C_{3 to 4} in a CU-rich environment of 20 to 30 nucleotides. We previously defined the so-called "differentiation control element" a CU-rich, repetitive structure in lipoxygenase and other mRNAs binding hnRNP E1 and K^19,27 with a consensus C₄A/GC₃UCUUC₄AAG.^35,36 Other work^32,37 and ours suggests that the binding regions for K and E1 may be fairly variable. But all have in common that hnRNP K requires at least a C_{3 to 4} core flanked by A or U and is G-poor, and that hnRNP E1 needs a more extended CU-rich binding region containing at least two C_{3 to 4} units.²⁰ The experimental results obtained with the collagen 3'-UTRs confirm our experiences: all three, COL1A1, 1A2, and 3A1 3'-UTRs, contain hnRNP K–like consensus motifs and do actually bind it. COL1A1 contains a classical CU-rich hnRNP E1–binding site as demonstrated previously.¹⁵ According to our results, COL1A2 has the same feature: it is also able to bind hnRNP E1, although not COL3A1, which lacks a more extended CU-rich region. Thus, E1 is a common factor associated to COL1A1 and 1A2 mRNA 3'-UTR, which could facilitate coordinate expression of the two subunits of type I collagen. The same is true for hnRNP K with respect to a coordinate expression of collagens I and III.

The situation with hnRNP A1 is more ambiguous. The binding preference of A1 seems to be not very distinctly marked.³⁸ Because of its more general RNA-binding properties, it is often called a "core" hnRNP protein. Nevertheless, there are reports showing that hnRNP A1 prefers AU-rich regions of the type AUUUA, which otherwise bind preferentially AU-rich element binding proteins such as AUF1 (hnRNP D) or HuR.^39,40 Despite its low-sequence preference, in our cross-linking assays, A1 bound only to COL3A1, which indeed contains two AUUUA motifs (Figure 1). This is in a certain way incompatible with our A1 transfection experiments (Figure 4). The expression of the LucCOL3'–UTR reporter gene constructs is considerably stimulated when cotransfected with an A1 expression plasmid. This argues in favor of an active interaction of hnRNP A1 with all three UTRs. However, it cannot be excluded that in vitro cross-linking conditions are not fully compatible with in vivo transfection conditions. Another explanation may be that A1 did not interact directly with the RNA but could be coordinated by protein/protein interactions during mRNP assembly, which is not reflected by RNA/protein cross-linking patterns.

Functionally of importance is the fact that those hnRNP proteins, which bind to COL I and III mRNAs, are described as positive effectors of mRNA stability. In our hands, the half-life time of luciferase transcript carrying COL IA1 3'-UTR increased in vitro by recombinant hnRNP E1 and K, and that carrying COL IIIA1 3'-UTR by K, although not by E1 (Figure 4B). This concept of differential action on IA1/IIIA1 3'-UTRs was strengthened further by the negative result of the LucCOL3A1 E1 cotransfection experiment: hnRNP E1, which did not bind to COL3A1 3'-UTR (Figure 2, gel shift), also failed consequently to stimulate Luc reporter gene expression (Figure 4, lanes 10 and 11). The stimulation rates for E1 and K action in the cotransfection experiments are in good accordance with the TGF-ß1 stimulation results (Figure 5). TGF-ß1 is a well-known mediator of the signaling pathway governing collagen expression via the AT₁ receptor.^4,7 We show that hnRNP A1, E1, and K expression is synergistically upregulated in the same range as the collagens and CTGF by TGF-ß1. Thus, RNA-binding hnRNP proteins are important cytoplasmic control molecules defining the fibrotic status of a tissue. A pathophysiological relevance of this finding was confirmed by screening patients experiencing different heart diseases associated with fibrosis. They were analyzed for their potential to express hnRNP proteins E1, A1, and K in the left ventricle. hnRNP E1 and K mRNA was significantly increased in the myocardium of patients with AS or ICM, as well as hnRNP A1 mRNA, which was also found to be most increased in AS, and thus elevated that hnRNPs may play a role in collagen content regulation in AS.

There are several reports underlining the importance of 3'-UTR–mediated mRNA stability in cardiovascular pathophysiology.⁴⁰ Scenarios can be envisaged in which the altered level of mRNA stabilizing/destabilizing proteins or mutations in the 3'-UTR–binding motifs of mRNAs of cardiovascular important proteins may constitute molecular events on the way of development of different cardiovascular diseases. The expression of important signaling proteins in the heart as ß₁-adrenergic receptor or AT₁/AT₂ receptors was shown to be modulated by alterations of the level of RNA-binding proteins such as AUF1, HuR, or hnRNP A1 in the context of heart failure.^41–43 hnRNP E1 (-CP1, PCBP1) is known to be involved in post-transcriptional response to hypoxia. It is a regulatory protein that binds to a hypoxia-inducible CU-rich binding site in erythropoietin or tyrosine hydroxylase mRNA. This has been shown to be of pathophysiological significance.^22,23

In conclusion, the data from the present study suggest that mRNA-binding proteins hnRNP A1, E1, and K substantially participate in the coordinative upregulation of collagen I and III synthesis in cardiac fibrosis. This is caused by an increase in their expression via the AT₁ receptor way in which TGF-ß1 plays a critical role. Mechanistically, the RNA-binding proteins may act as stabilizers of the collagen mRNAs, effectively promoting collagen synthesis in this way.

	Acknowledgments

 
The authors acknowledge the support of the Deutsche Forschungsgemeinschaft to V.R.Z. and B.J.T.

	Footnotes

 
*Both authors contributed equally to this work.

Original received April 1, 2004; revision received October 7, 2004; accepted October 18, 2004.

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 
1. Sun Y, Weber KT. RAS and the connective tissue in the heart. Int J Biochem Cell Biol. 2003; 35: 919–931.[CrossRef][Medline] [Order article via Infotrieve]

2. Bishop JE, Lindahl G. Regulation of cardiovascular collagen synthesis by mechanical load. Cardiovasc Res. 1999; 42: 27–44.[Free Full Text]

3. Li YY, McTiernan CF, Feldman AM. Interplay of matrix metalloproteinases, tissue inhibitors of metalloproteinases and their regulators in cardiac matrix remodeling. Cardiovasc Res. 2000; 46: 214–224.[Abstract/Free Full Text]

4. Fielitz J, Hein S, Mitrovic V, Pregla R, Zurbrugg HR, Warnecke C, Schaper J, Fleck E, Regitz-Zagrosek V. Activation of the cardiac renin-angiotensin system and increased myocardial collagen expression in human aortic valve disease. J Am Coll Cardiol. 2001; 37: 1443–1449.[Abstract/Free Full Text]

5. Gonzalez A, Lopez B, Querejeta R, Diez J. Regulation of myocardial fibrillar collagen by angiotensin II. A role in hypertensive heart disease? J Mol Cell Cardiol. 2002; 34: 1585–1593.[CrossRef][Medline] [Order article via Infotrieve]

6. Delcayre C, Silvestre JS, Garnier A, et al. Cardiac aldosterone production and ventricular remodeling. Kidney Int. 2000; 57: 1346–1351.[CrossRef][Medline] [Order article via Infotrieve]

7. Dostal DE. Regulation of cardiac collagen. Angiotensin and cross-talk with local growth factors. Hypertension. 2001; 37: 841–8444.[Free Full Text]

8. Namba T, Tsutsui H, Tagawa H, Takahashi M, Saito K, Kozai T, Usui M, Imanaka-Yoshida K, Imaizumi T, Takeshita A. Regulation of fibrillar collagen expression and protein accumulation in volume-overloaded cardiac hypertrophy. Circulation. 1997; 95: 2448–2454.[Abstract/Free Full Text]

9. Ghosh AK. Factors involved in the regulation of type I collagen expression: implication in fibrosis. Exp Biol Med. 2002; 227: 301–314.[Abstract/Free Full Text]

10. Myllyharju J, Kivirikko KI. Collagens and collagen-related diseases. Ann Med. 2001; 33: 7–21.[Medline] [Order article via Infotrieve]

11. Lindquist JN, Marzluff WF, Stefanovic B. Fibrinogenesis III. Posttranscriptional regulation of type I collagen. Am J Physiol Gastrointest Liver Physiol. 2000; 279: G471–G476.[Abstract/Free Full Text]

12. Stefanovic B, Brenner DA. 5' stem-loop of collagen 1(I) mRNA inhibits translation in vitro but is required for triple helical collagen synthesis in vivo. J Biol Chem. 2003; 278: 927–933.[Abstract/Free Full Text]

13. Stefanovic B, Hellerbrand C, Holcik M, Briendl M, Liebhaber SA, Brenner DA. Posttranscriptional regulation of collagen 1(I)mRNA in hepatic stellate cells. Mol Cell Biol. 1997; 17: 5201–5209.[Abstract]

14. Määttä A, Penttinen RPK. A fibroblast protein binds the 3'-untranslated region of pro-1(I) collagen mRNA. Biochem J. 1993; 295: 691–698.[Medline] [Order article via Infotrieve]

15. Holcik M, Liebhaber SA. Four highly stable eukaryotic mRNAs assemble 3' untranslated region RNA-protein complexes sharing cis and trans components. Proc Natl Acad Sci U S A. 1997; 94: 2410–2414.[Abstract/Free Full Text]

16. Lindquist JN, Kauschke SG, Stefanovic B, Burchardt ER, Brenner DA. Characterization of the interaction between CP2 and the 3'-untranslated region of collagen 1(I) mRNA. Nucleic Acids Res. 2000; 28: 4306–4316.[Abstract/Free Full Text]

17. Chkeidze AN, Lyakhov DL, Makeyev AV, Morales J, Kong J, Liebhaber SA. Assembly of the -globin mRNA stability complex reflects binary interaction between the pyrimidine-rich 3' untranslated region determinant and poly(C) binding protein CP. Mol Cell Biol. 1999; 19: 4572–4581.[Abstract/Free Full Text]

18. Yu J, Russell JE. Structural and functional analysis of an mRNP complex that mediates the high stability of human ß-globin mRNA. Mol Cell Biol. 2001; 21: 5879–5888.[Abstract/Free Full Text]

19. Ostareck DH, Ostareck-Lederer A, Wilm M, Thiele BJ, Mann M, Hentze MW. mRNA silencing in erythroid differentiation: hnRNP K and hnRNP E1 regulate 15-lipoxygenase translation from the 3' end. Cell. 1997; 89: 597–606.[CrossRef][Medline] [Order article via Infotrieve]

20. Reimann I, Huth A, Thiele H, Thiele BJ. Repression of 15-lipoxygenase synthesis by hnRNP E1 is dependent on repetitive nature of LOX mRNA 3'UTR control element DICE. J Mol Biol. 2002; 315: 965–974.[CrossRef][Medline] [Order article via Infotrieve]

21. Gamarnik AV, Andino R. Two functional complexes formed by KH domain containing proteins with the 5' noncoding region of poliovirus RNA. RNA. 1997; 3: 882–892.[Abstract]

22. Czyzyk-Krzeska MF, Bendixen AC. Identification of the poly (C) protein in the complex associated with the 3' untranslated region of erythropoietin messenger RNA. Blood. 1999; 93: 2111–2120.[Abstract/Free Full Text]

23. Paulding WR, Czyzyk-Krzeska MF. Regulation of tyrosine hydroxylase mRNA stability by protein-binding pyrimidine-rich sequence in the 3' untranslated region. J Biol Chem. 1999; 274: 2532–2538.[Abstract/Free Full Text]

24. Xiao X, Tang YS, Mackins JY, Sun XL, Jayaram HN, Hansen DK, Antony AC. Isolation and characterization of a folate receptor mRNA-binding trans-factor from human placenta. Evidence favoring identity with heterogeneous nuclear ribonucleoprotein E1. J Biol Chem. 2001; 276: 41510–41517.[Abstract/Free Full Text]

25. Yeap BB, Voon DC, Vivian JP, McCulloch RK, Thomson AM, Giles KM, Czyzyk-Krzeska MF, Furneaux H, Wilce MC, Wilce JA, Leedman PJ. Novel binding of HuR and poly(C)-binding protein to a conserved UC-rich motif within the 3'-untranslated region of the androgen receptor messenger RNA. J Biol Chem. 2002; 277: 27183–27192.[Abstract/Free Full Text]

26. Skalweit A, Doller A, Huth A, Kähne T, Persson PB, Thiele BJ. Posttranscriptional control of renin synthesis: identification of proteins interacting with renin mRNA 3'-untranslated region. Circ Res. 2003; 92: 419–427.[Abstract/Free Full Text]

27. Thiele BJ, Berger M, Huth A, Reimann I, Schwarz K Thiele H. Tissue-specific translational regulation of alternative rabbit 15-lipoxygenase mRNAs differing in their 3'-untranslated regions. Nucleic Acids Res. 1999; 27: 1828–1836.[Abstract/Free Full Text]

28. Neuss M, Regitz-Zagrosek V, Hildebrandt A, Fleck E. Human cardiac fibroblasts express an angiotensin receptor with unusual binding characteristics which is coupled to cellular proliferation. Biochem Biophys Res Commun. 1994; 204: 1334–1339.[CrossRef][Medline] [Order article via Infotrieve]

29. Chu ML, de Wet W, Bernard M, Ramirez F. Fine structural analysis of the human pro-1(I) collagen gene. J Biol Chem. 1985; 260: 2315–2320.[Abstract/Free Full Text]

30. Chu ML, Weil D, de Wet W, Bernard M, Sippola M, Ramirez F. Isolation of cDNA and genomic clones encoding human pro-1(III) collagen. J Biol Chem. 1985; 260: 4357–4363.[Abstract/Free Full Text]

31. Myers JC, Dickson LA, de Wet WJ, Bernard MP, Chu ML, Di Liberto M, Pepe G, Sangiorgi FO, Ramirez F. Analysis of the 3'end of the human pro-2(1) collagen gene. J Biol Chem. 1983; 258: 10128–10135.[Abstract/Free Full Text]

32. Thisted T, Lyakhov D, Liebhaber SA. Optimized targets of two closely related triple KH domain proteins, heterogeneous nuclear ribonucleoprotein K and CP-2KL, suggest distinct modes of RNA recognition. J Biol Chem. 2001; 276: 17484–17496.[Abstract/Free Full Text]

33. Medugno L, Costanzo P, Lupo A, Monti M, Florio F, Pucci P, Izzo P. A novel zinc finger transcriptional repressor, ZNF224, interacts with the negative regulatory element (AldA-NRE) and inhibits gene expression. FEBS Lett. 2003; 534: 93–100.[CrossRef][Medline] [Order article via Infotrieve]

34. Villarreal FJ, Dillmann WH. Cardiac hypertrophy-induced changes in mRNA levels for TGF-beta 1, fibronectin, and collagen. Am J Physiol. 1992; 262: H1861–H1866.[Medline] [Order article via Infotrieve]

35. Chen MM, Lam A, Abraham JA, Schreiner GF, Joly AH. CTGF expression is induced by TGF-beta in cardiac fibroblasts and cardiac myocytes: a potential role in heart fibrosis. J Mol Cell Cardiol. 2000; 32: 1805–1819.[CrossRef][Medline] [Order article via Infotrieve]

36. Fleming J, Thiele BJ, Chester J, O’Prey J, Janetzki S, Aitken A, Anton IA, Rapoport SM, Harrison PR. The complete sequence of the rabbit erythroid cell-specific 15-lipoxygenase mRNA; comparison of the predicted amino acid sequence of the erythrocyte lipoxygenase with other lipoxygenases. Gene. 1989; 79: 181–188.[CrossRef][Medline] [Order article via Infotrieve]

37. Ostrowski J, Wyrwicz L, Rychlewski L, Bomstyk K. Heterogeneous nuclear ribonucleoprotein K protein associates with multiple mitochondrial transcripts within the organelle. J Biol Chem. 2002; 277: 6303–6310.[Abstract/Free Full Text]

38. Abdul-Manan N, Williams KR. hnRNP A1 binds promiscuously to oligoribonucleotides: utilization of random and homo-oligonucleotides to discriminate sequence from base-specific binding. Nucleic Acids Res. 1996; 24: 4063–4070.[Abstract/Free Full Text]

39. Hamilton BJN, Burns CM, Nichols RC, Rigby WFC. Modulation of AUUUA response element binding by heterogeneous nuclear ribonucleoprotein A1 in human T lymphocytes. J Biol Chem. 1997; 272: 28732–28741.[Abstract/Free Full Text]

40. Misquitta CM, Iyer VR, Werstiuk ES, Grover AK. The role of 3'-untranslated region (3'-UTR) mediated mRNA stability in cardiovascular pathophysiology. Mol Cell Biochem. 2001; 224: 53–67.[CrossRef][Medline] [Order article via Infotrieve]

41. Pende A, Tremmel KD, DeMaria CT, Blaxall BC, Minobe WA, Sherman JA, Bisognano JD, Bristow MR, Brewer G, Port J. Regulation of the mRNA-binding protein AUF1 by activation of the beta-adrenergic receptor signal transduction pathway. J Biol Chem. 1996; 271: 8493–8501.[Abstract/Free Full Text]

42. Haywood GA, Gullestad L, Katsuya T, Hutchinson HG, Pratt RE, Horiuchi M, Fowler MB. AT1 and AT2 angiotensin receptor gene expression in human heart failure. Circulation. 1997; 95: 1201–1206.[Abstract/Free Full Text]

43. Blaxall BC, Pellett AC, Wu SC, Pende A, Port JD. Purification and characterization of ß-adrenergic receptor mRNA-binding proteins. J Biol Chem. 2000; 274: 4290–4297.

This article has been cited by other articles:

				H. Li, W. Chen, Y. Zhou, P. Abidi, O. Sharpe, W. H. Robinson, F. B. Kraemer, and J. Liu Identification of mRNA binding proteins that regulate the stability of LDL receptor mRNA through AU-rich elements J. Lipid Res., May 1, 2009; 50(5): 820 - 831. [Abstract] [Full Text] [PDF]

				N. K. Mukhopadhyay, J. Kim, B. Cinar, A. Ramachandran, M. H. Hager, D. Di Vizio, R. M. Adam, M. A. Rubin, P. Raychaudhuri, A. De Benedetti, et al. Heterogeneous Nuclear Ribonucleoprotein K Is a Novel Regulator of Androgen Receptor Translation Cancer Res., March 15, 2009; 69(6): 2210 - 2218. [Abstract] [Full Text] [PDF]

				J. S. Orringer, C. Hammerberg, T. Hamilton, T. M. Johnson, S. Kang, D. L. Sachs, G. Fisher, and J. J. Voorhees Molecular Effects of Photodynamic Therapy for Photoaging Arch Dermatol, October 1, 2008; 144(10): 1296 - 1302. [Abstract] [Full Text] [PDF]

				M. Fahling, R. Mrowka, A. Steege, G. Nebrich, A. Perlewitz, P. B. Persson, and B. J. Thiele Translational Control of Collagen Prolyl 4-Hydroxylase-{alpha}(I) Gene Expression under Hypoxia J. Biol. Chem., September 8, 2006; 281(36): 26089 - 26101. [Abstract] [Full Text] [PDF]

				L. D. Kosturko, M. J. Maggipinto, G. Korza, J. W. Lee, J. H. Carson, and E. Barbarese Heterogeneous Nuclear Ribonucleoprotein (hnRNP) E1 Binds to hnRNP A2 and Inhibits Translation of A2 Response Element mRNAs Mol. Biol. Cell, August 1, 2006; 17(8): 3521 - 3533. [Abstract] [Full Text] [PDF]

				K. Linker, A. Pautz, M. Fechir, T. Hubrich, J. Greeve, and H. Kleinert Involvement of KSRP in the post-transcriptional regulation of human iNOS expression-complex interplay of KSRP with TTP and HuR Nucleic Acids Res., August 26, 2005; 33(15): 4813 - 4827. [Abstract] [Full Text] [PDF]

				Y. Zhang, D. E. Handy, and J. Loscalzo Adenosine-Dependent Induction of Glutathione Peroxidase 1 in Human Primary Endothelial Cells and Protection Against Oxidative Stress Circ. Res., April 29, 2005; 96(8): 831 - 837. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Data Supplement All Versions of this Article: 95/11/1058    most recent 01.RES.0000149166.33833.08v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Thiele, B.-J. Articles by Regitz-Zagrosek, V. Search for Related Content PubMed PubMed Citation Articles by Thiele, B.-J. Articles by Regitz-Zagrosek, V. Related Collections Biochemistry and metabolism Other myocardial biology Physiological and pathological control of gene expression