HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 275, Issue 52, 40967-40973, December 29, 2000

This Article Abstract Full Text (PDF) All Versions of this Article: 275/52/40967    most recent M004134200v1 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 Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Grover, J. Articles by Roughley, P. J. Search for Related Content PubMed PubMed Citation Articles by Grover, J. Articles by Roughley, P. J. Social Bookmarking                      What's this?

Analysis of the Human Lumican Gene Promoter^*

Judy Grover, Chia-Yang Liu^§, Winston W.-Y. Kao^§, and Peter J. Roughley^¶

From the  Genetics Unit, Shriners Hospital for Children and Department of Surgery, McGill University, Montreal, Quebec H3G 1A6, Canada and ^§ Department of Ophthalmology, University of Cincinnati, Cincinnati, Ohio 45267

Received for publication, May 15, 2000, and in revised form, September 7, 2000

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The human lumican gene was shown to possess one major transcription start site, resulting in exon 1 of the gene giving rise to the first 74 base pairs (bp) of the 5'-untranslated region. About 1.6 kilobase pairs of upstream promoter sequence were sequenced and analyzed to identify elements responsible for gene expression. No typical TATAA sequence was identified in the vacinity of the transcription start site, but an atypical TATCA sequence residing 41 bp upstream was shown to be necessary for transcription, although it was incapable of supporting transcription by itself. A GC box residing 74 bp upstream of the transcription start site also was essential for the initiation of transcription. Sp3 was identified as the transcriptional activator binding to the GC box. No additional elements that significantly modulated transcription were noted in the promoter sequence analyzed, when using human adult chondrocytes as the cell source for transfection in reporter assays. In contrast, reporter assays carried out in human fetal lung fibroblasts, where lumican expression is deplete, revealed the presence of a repressor element located between 384 and 598 bp upstream of the transcription start site. A GATA-binding site located between bp 386 and 391 was identified as being necessary for repression of transcription. The mouse lumican promoter does not possess an equivalent site, and this may explain why the lumican gene is expressed in fetal murine cartilage but not in fetal human cartilage.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Lumican belongs to the family of leucine-rich repeat glycoproteins present in the extracellular matrix of connective tissues (1). This protein family is characterized by the presence of adjacent leucine-rich regions bearing the motif LXXLXLXXNXL flanked by disulfide-bonded domains and includes dermatan sulfate proteoglycans, keratan sulfate proteoglycans, and glycoproteins devoid of any glycosaminoglycan chain. Lumican belongs to the subfamily of keratan sulfate proteoglycans that also includes fibromodulin (2) and keratocan (3), and in common with these proteoglycans it contains 10 leucine-rich repeats within its central region.

To date the lumican gene has been cloned in the chick (4), bovine (5), human (6), and mouse (7), and in each case amino acid sequence data revealed the presence of four potential sites for the substitution by N-linked keratan sulfate or oligosaccharides within the leucine-rich repeat region. However, it appears that not all of the sites can serve as acceptors for keratan sulfate synthesis, and in the chick cornea only three of the four sites are so substituted (8). Although lumican was initially described as a corneal proteoglycan, it is now known to be expressed in a variety of tissues, including artery (9), lung (10), and articular cartilage (11). In these tissues lumican may exist in a glycoprotein form, being substituted with short oligosaccharides or unsulfated polylactosamine chains rather than keratan sulfate. In the case of human articular cartilage, the structure of the carbohydrate substituents of lumican varies with age (11).

At the genomic level, the lumican gene has been shown to be composed of three exons and two introns, spanning approximately 7-9 kpb¹ of the genome, depending on the species examined (11-13). The first exon contains only 5'-untranslated sequence, the second exon contains most of the coding sequence, and the third exon contains the remainder of the coding sequence and the 3'-untranslated region. This arrangement results in all the leucine-rich repeats being encoded by a single exon. Lumican gene expression can vary considerably between different tissues during development, with expression being evident early during embryonic development in the chick cornea (14) but not until after birth in human cartilage (11). In this latter tissue, lumican message levels are also considerably enhanced in the adult (15). However, little is known about the promoter elements that regulate lumican gene expression, although presumptive TATA boxes residing upstream of the transcription initiation site have been described in both the chick (12) and mouse genes (13). These elements are presumably involved in the regulation of basal transcription.

To date there have been no reports concerning the structure and function of the human lumican gene promoter or the presence of repressor elements that may be responsible for regulating the level of lumican gene transcription in any species. The aim of the present work is to address these deficits.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Human Promoter Sequence-- A BAC clone containing the human lumican gene was obtained from PCR screening of a human BAC library (Research Genetics Inc., Huntsville, AL) using the primer pair 5'-GAGGATGCTGTTTCAGCT and 5'-AGGACAGATCCAGCTCAA within exon 2. The sequence of the promoter region was determined by direct sequencing of the BAC clone using the double-stranded DNA Cycle Sequencing System (Life Technologies, Inc.), commencing with primers from the published cDNA sequence (GenBank^TM accession number U18728) and then using subsequent primers generated from the extended sequence.

Mouse Promoter Sequence-- Mouse lumican genomic DNA was isolated from a BAC genomic DNA library (Research Genetics Inc., Huntsville, AL) by PCR screening. A 6-kbp¹ SalI-XbaI fragment containing 3.5-kbp 5'-flanking region, exon 1, intron 1, and part of exon 2 of the lumican gene was cloned in a pBluescriptSK() vector, as described previously (16). The nucleotide sequence of the lumican promoter was determined on both strands with the T3 primer and walk-in primers by the DNA core facility in the Department of Molecular Genetics at the University of Cincinnati.

Primer Extension-- Primer extension was performed as described previously (11), using an oligonucleotide primer in the 5'-untranslated region of the cDNA (bp 54-74, Fig. 1) and 5 µg of total RNA from adult chondrocytes as template.

Reporter Gene Constructs-- The pGL2 series of luciferase reporter gene plasmids (Promega, through Fisher) were used for all lumican promoter/reporter gene constructs, using standard protocols (17). Various PCR-amplified human promoter 5'-deletion fragments were used in reporter gene constructs to identify transcriptionally active elements as follows: region 1, bp 41 to 74; region 2, bp 54 to 74; region 3, bp 113 to 74; region 4, bp 384 to 74; region 5, bp 598 to 74; region 6, bp 759 to 74; and region 7, bp 398 to 74. All nucleotide numbering refers to Fig. 1. In addition, a construct containing the mouse promoter region, bp 1262 to 80 (GenBank^TM accession number AF186467), was prepared.

Mutation of Reporter Constructs-- A synthetic oligonucleotide, 5'-GCGTGACTGTTCTGGGCTCT, was used as upstream primer to introduce mutations (shown in bold) into the GC box (region 3A). The downstream primer in the mutagenesis amplification was 5'-CTGCCTTGACCGACGGTCTAA. Mutation in the putative TATA box (region 3B) was introduced using the synthetic oligonucleotide 5'-CAGCACTCAGAATCTGGCAGCCAG together with 5'-CAAGAGCTGAAGGGGG as the downstream primer. Synthetic oligonucleotides 5'-GATCTTTAGACAAACATGATAG, 5'-GATCTTTAGATAAACATGACAG, and 5'-GATCTTTAGACAAACATGACAG were used as upstream primers to introduce mutations into the GATA-binding sites, GBS_A, (bp 391 to 388, region 7A), GBS_B (bp 382 to 377, region 7B), and GBS_A/B (region 7C), respectively. The common downstream primer in all GBS mutagenesis amplifications was 5'-CCTTACTGTCTTGACACTGCTT. The amplified fragments containing the mutated sequences were cloned into the luciferase reporter gene pGL2B plasmid, as above.

Source of Tissue and Cells-- For chondrocyte isolation, human articular cartilage was collected from the distal femur at the time of autopsy and within 20 h of death. The specimens were from individuals aged 33, 48, 63, and 67 years. In all cases the knee joints appeared macroscopically normal, and there was no clinical evidence of a connective tissue abnormality. Chondrocytes were isolated and grown in monolayer culture, as described previously (18). A human fetal lung fibroblast cell line, HFL-1 (153-CCL, American Type Culture Collection, Manassas, VA), was maintained in culture in Dulbecco's modified Eagle's medium in the presence of 10% fetal calf serum. All tissue culture materials were from Life Technologies, Inc. For histology and immunohistochemistry, human tissue was obtained from individuals aged 17 weeks gestation (fetus), 2 months, 7, 17, 27, 56, and 63 years, and murine tissue was obtained from C57B mice aged e14.5, e17.5, newborn, 7 days, 1 month, 3 months, and 1 year.

Northern Blots-- Total cellular RNA was extracted from cultured chondrocytes or HFL-1 cells by the acid guanidinium thiocyanate/phenol/chloroform method (19). 10 µg of total RNA per cell type were blotted and probed as described previously for both lumican and human glyceraldehyde-3-phosphate dehydrogenase expression (11).

Cell Transfections-- HFL-1 cells or chondrocytes with three passages or less in culture were used for transfection, as described previously (20).

Luciferase Assay-- The dual-luciferase reporter assay system (Promega, through Fisher) was used as described by the manufacturer, except that assays were performed 72 h post-transfection as described previously (20). The activity of each construct was calculated as follows: (experimental luciferase activity  background)/(control luciferase activity  background). Results are expressed as the average and standard deviation from three separate experiments and are normalized with respect to the activity of reporter construct region 4 (100%) and the control vector (0%). Confidence levels for observed changes were calculated using the Student t test (p values).

Electrophoretic Mobility Shift Assay (EMSA)-- Double-stranded DNA probes spanning bp 63 to 94 (probe GCB) and bp 382 to 395 (probe GBS_A) were prepared by hybridization of the synthetic oligonucleotides 5'-ACTGGCGTGACTGGGCTGGGCTCTCCCC and 5'-GGGTGGGGAGAGCCCAGCCCAGCCCAGTCAC and 5'-CTTTAGATAAACAT and 5'-ATGTTTATCTAAAG, respectively, and radiolabeled (17). A third probe, spanning bp 316 to 398 (probe GBS), was prepared by PCR using the upstream primer 5'-GATCTTTAGATAAACATGATAG and downstream primer 5'-CTTACAAAGCCTCTTTACATCTGT, and bacteriophage T4 polynucleotide kinase was used to radiolabel the 5' terminus of this probe (17). Labeled probes were purified on a 5% non-denaturing acrylamide gel. Nuclear extracts were prepared from HFL-1 cells and adult chondrocytes (21). 5 µg of nuclear extract was incubated with 50,000 cpm radiolabeled probe, with or without a competitor, in a 20-µl reaction containing 12 mM Hepes/NaOH (pH 7.9), 4 mM Tris-HCl (pH 7.9), 60 mM KCl, 1 mM dithiothreitol, 6% glycerol, and 5 µg of poly(dI-dC). Competitor was either a double-stranded DNA fragment in 100-fold excess of labeled probe concentration or a specific antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA) used following the manufacturer's instructions for gel supershift analysis. The antibodies recognizing GATA-1, -2, -3, -4, or -6 were raised against peptides mapping at the carboxyl terminus of their respective proteins. After 30 min of incubation at room temperature, reactions were analyzed on a 5% non-denaturing acrylamide gel, containing 89 mM Tris-HCl (pH 8.0), 89 mM boric acid, and 2 mM EDTA. Following electrophoresis, gels were dried and autoradiographed.

Histology and Immunohistochemistry-- Tissue samples were fixed in 4% fresh paraformaldehyde in phosphate-buffered saline, pH 7.2, at 4 °C overnight. Samples with bone were decalcified for 7-10 days in 10% EDTA, 0.1 M Tris-HCl (pH 7.4), at 4 °C. Specimens were embedded in paraffin (Paraplast X-tra, Fisher) and sectioned at 6-8 µm. For histological analysis, sections were stained with iron hematoxylin-fast green-Safranin O. For immunohistochemistry, tissue sections were pretreated for 1 h with chondroitinase ABC (0.25 unit/ml; Sigma) at 37 °C, and then immunostaining was performed using the Vectastain ABC Elite kit (Vector Laboratories, Burlingame, CA) following the manufacturer's instructions. The primary antibody was an anti-peptide antiserum specific either to human lumican (11) or mouse lumican. The mouse-specific peptide contained the carboxyl-terminal sequence LRVANEITVN, and the antiserum was raised in rabbit as described previously for its human counterpart (11).

Immunoblotting-- Protein extracts from the femoral heads of newborn (3 day) or adult (3 month) mice were prepared and treated with endo--galactosidase before analysis by SDS-polyacrylamide gel electrophoresis and immunoblotting, as described previously (11), using the mouse-specific antipeptide antiserum described above.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

PCR-based techniques were used to isolate a BAC clone containing the human lumican gene, and a 1.7-kbp segment of the gene encompassing the first exon and upstream promoter region was characterized by nucleotide sequence analysis (Fig. 1). Primer extension analysis of lumican message revealed a single major transcription start site (Fig. 2) toward the 3'-end of the gene sequence analyzed. Based on the previously reported site of the first intron (11), this analysis indicates that the first exon of the human lumican gene spans 74 bp. Thus, the three exons of the gene have sizes of 74, 883, and 770 bp, which is compatible with the observation of a single message of 1.8 kb (11).

View larger version (84K): [in this window] [in a new window]   Fig. 1.   DNA sequence of the promoter region of the human lumican gene. The promoter region is numbered from 1616 to 1 and the transcribed sequence of the first exon is numbered from +1 to +74. The position of the transcription start site was determined by primer extension analysis (Fig. 2). The positions of promoter elements potentially involved in the regulation of gene transcription are indicated, including TATAAA sequences (boxed), GATA-binding sites (double underline), Ets-binding sites (single underline), and a GC box and atypical TATA box (TATCA) (dashed underline).

View larger version (25K): [in this window] [in a new window]   Fig. 2.   Analysis of the transcription start site of the lumican gene. The location of the transcription start site was analyzed by primer extension using an oligonucleotide within the first exon of the gene (lane 1). The site of the major product is marked by an arrow. The adjacent lanes (G, A, T, and C) depict the corresponding genomic sequence obtained using the same oligonucleotide as a primer. The sequences and locations of the GC box and the atypical TATA box are indicated.

Analysis of the promoter sequence revealed the presence of five TATAAA sequences within the distal region of the lumican promoter that could theoretically be used to locate the transcription machinery, although all are more than 790 bp away from the transcription start site and are unlikely to be involved in the generation of the major transcription product. There is no conventional TATA element within 100 bp of the transcription start site, but there is an atypical TATCA sequence residing 41 bp upstream from the transcription start site. Such a sequence has been postulated to be responsible for transcription regulation of the mouse lumican gene (13). The human promoter sequence is also notable for the presence of a single GC box, residing 74 bp prior to the transcription start site. Such elements have been associated with the initiation of transcription in promoters lacking conventional TATA elements (22).

In order to study the promoter elements responsible for regulation of gene transcription, reporter gene constructs containing various lengths of the promoter were transfected into adult human chondrocytes, a cell type that shows high levels of lumican gene expression (11). Initially, a variety of constructs of increasing length, and sharing a common downstream terminus within the first exon of the gene (Fig. 3), were used to determine the length of promoter required to support basal transcription (Fig. 4). The two shortest constructs (regions 1 and 2, Fig. 3) possessing 41 and 54 bp of promoter sequence were unable to support transcription of the reporter gene, whereas constructs that contained greater than 113 bp of promoter sequence (regions 3-6) could support transcription. The atypical TATA element (TATCA) resides between bp 46 to 42 in the promoter and therefore is present within the second shortest construct containing 54 bp of promoter sequence. The inability of this construct to support transcription therefore indicates that the TATCA sequence is unable by itself to initiate transcription. However, when the GC box, which is located between bp 82 to 75, is included in the constructs, promoter activity is achieved. The importance of the GC box in achieving promoter activity was assessed by site-directed mutagenesis, in which 2 bases within the binding site were mutated (region 3A). The reporter construct possessing this mutation was unable to support transcription (Fig. 4). To determine whether the atypical TATA element is necessary for the action of the GCB, the activity of a construct containing a mutated TATA element (region 3B) was assessed. This reporter construct also showed no transcriptional activity (Fig. 4), indicating that both the TATCA sequence and GC box are necessary.

View larger version (11K): [in this window] [in a new window]   Fig. 3.   Reporter gene constructs used in analysis of the lumican gene promoter. The location of the promoter regions (1-7) used in generating lumican promoter/luciferase reporter gene constructs are indicated. Numbering refers to the genomic DNA sequence data (Fig. 1). The locations of the transcription start site, GC box (GCB), and repressor GATA-binding site (GBS) are also indicated.

View larger version (12K): [in this window] [in a new window]   Fig. 4.   Reporter gene analysis of the lumican gene promoter. Reporter gene constructs were transfected into adult human chondrocytes, and luciferase activity was monitored. Results are reported as percentage induction relative to the control vector for constructs possessing promoter sequence with different 5'-ends but sharing a common 3'-end within the first exon (regions 1-4, Fig. 3). In one construct the promoter sequence spanning the GC box has been mutated (region 3A), and in another the putative TATA element has been mutated (region 3B). * indicates p < 0.01.

Transcription factors binding to the GC box were examined by EMSA using nuclear proteins isolated from adult chondrocytes and a radiolabeled probe spanning the GC box (Fig. 5). Retardation of the probe was observed when mixed with the nuclear proteins, and this retardation could be prevented by the presence of excess unlabeled probe but not by the presence of excess unlabeled probe in which the GC box was mutated. This illustrates that the nuclear protein that binds to the probe does so specifically via the GC box. To determine whether a member of the Sp1 family of transcription factors was interacting with the GC box, analysis was repeated in the presence of antibodies recognizing different family members. This should result in either a supershift in retardation if probe interaction is maintained in the presence of the antibody or elimination of retardation if the antibody prevents probe interaction. Whereas the antibody against Sp1 itself showed no effect on gel retardation, that against Sp3 prevented retardation. Taken in conjunction with the promoter activity studies, this suggests that Sp3 binding to the GC box in the proximal lumican promoter is essential for basal transcription of the human lumican gene.

View larger version (80K): [in this window] [in a new window]   Fig. 5.   EMSA analysis of the GC box region of the lumican gene promoter. The electrophoretic mobilities of the radiolabeled probe are shown in the absence (lane 1) or presence (lanes 2-6) of nuclear proteins extracted from adult human chondrocytes. Probe plus nuclear extract were incubated without (lane 2) or with anti-Sp1 (lane 3) or anti-Sp3 (lane 4) antibody, or with 100-fold excess of unlabeled probe (lane 5) or unlabeled probe containing a mutated GC box (lane 6).

In order to investigate the presence of repressor elements within the human lumican gene promoter, the reporter constructs containing different promoter lengths were transfected into human fetal lung fibroblasts. These cells were chosen because in common with neonatal and fetal human chondrocytes, they exhibit low levels of lumican gene expression in comparison to the adult chondrocytes (Fig. 6). In contrast to the adult chondrocytes, the fetal lung fibroblasts were unable to support transcription from the reporter constructs when promoter lengths of greater than 384 bp were present (regions 5 and 6, Fig. 3) (Fig. 7A). This would suggest that a repressor element resides within the region between bp 598 to 384 of the lumican promoter, which is common to the two largest constructs.

View larger version (40K): [in this window] [in a new window]   Fig. 6.   Northern blot analysis of lumican gene expression. Total cellular RNA from cultured human fetal lung fibroblasts (lane 1) or adult chondrocytes (lane 2) were analyzed. Blots were analyzed using cDNA probes corresponding to lumican (A) or glyceraldehyde-3-phosphate dehydrogenase (B). The migration position of the 18 S ribosomal RNA is indicated.

View larger version (17K): [in this window] [in a new window]   Fig. 7.   Reporter gene analysis of the distal lumican gene promoter. Reporter gene constructs were transfected into adult chondrocytes (open bars) or fetal lung fibroblasts (filled bars), and luciferase activity was monitored. Results are reported as percentage induction relative to the control vector. A, analysis of the human promoter performed with constructs possessing promoter sequence with different 5'-ends but sharing a common 3'-end within the first exon (regions 4-6, Fig. 3). B, analysis of the mouse promoter. * indicates p < 0.01.

This repressor region is notable for the presence of two GATA-binding sites (GBS) spanning the junction between promoter regions 4 and 5 (Fig. 3). To determine whether these GBS are involved in repression, an additional promoter region/luciferase reporter gene construct was prepared to include the two GBS at its 5' terminus (region 7, Fig. 3). This construct did not give rise to transcription when transfected into fetal lung fibroblasts (Fig. 8). To verify that the GBS were involved in this repression, they were mutated individually or in tandem in the promoter/reporter construct. Mutation of both GBS (region 7C) relieved the repression and restored transcription, as did mutation of the distal GBS (region 7A) alone. Mutation of the proximal GBS (region 7B) had no effect on the repression of transcription. Thus, the GATA-binding element located between bp 390 to 385 appears critical in the repression of lumican gene expression observed in fetal lung fibroblasts.

View larger version (11K): [in this window] [in a new window]   Fig. 8.   Reporter gene analysis of GBS elements in the lumican gene promoter repressor region. Reporter gene constructs were transfected into human fetal lung fibroblasts, and luciferase activity was monitored. Results are reported as percentage induction relative to the control vector, for constructs containing lumican gene promoter regions 4 or 7 (Fig. 3). The GBS elements in region 7 are either intact or contain mutations in the distal element (region 7A), the proximal element (region 7B), or both elements (region 7C). * indicates p < 0.01.

To verify that the fetal lung fibroblasts or adult chondrocytes possessed transcription factors able to interact with the GBS, EMSA was performed (Fig. 9A). The radiolabeled probe spanning the two adjacent GBS in the repressor region was retarded in the presence of nuclear proteins isolated from either cell type. This retardation could be eliminated in the presence of excess unlabeled probe but not by excess unlabeled probes in which the GBS had been mutated, confirming that the GBS present in the repressor region of the human lumican promoter is indeed a functional GATA-binding site. In an attempt to identify the protein binding to the GBS, electrophoretic mobility shift assays were performed in the presence of antibodies to different GATA molecules. However, antibodies raised against GATA-1- to -4 or -6 failed to affect the gel retardation of the probe (data not shown). In order to determine whether there were any differences in the proteins from the two cell types interacting with the GBS, EMSA was repeated with a probe (GBS_A) spanning only the active site (Fig. 9B). This illustrated that differences did exist, with the fibroblast proteins causing greater retardation of the probe than those from the chondrocytes.

View larger version (90K): [in this window] [in a new window]   Fig. 9.   EMSA analysis of the GBS elements in the lumican gene promoter repressor region. The electrophoresis profile of labeled probes GBS (A) and GBSA (B) are shown in the absence (lane 1) or presence of nuclear proteins from fetal lung fibroblasts (lanes 2-4) or adult chondrocytes (lanes 5-7). Probe plus nuclear extract were incubated without (lanes 2 and 5) or with 100-fold excess of unlabeled probe (lanes 3 and 6), or with 100-fold excess of unlabeled probe containing mutated GBSA and GBSB (lanes 4 and 7).

As a prelude to using transgenic mice to study the effect of lumican gene overexpression on the development of tissues where gene expression is normally deficient in the fetus, the sequences of the human and mouse lumican gene promoters were compared (Fig. 10). Whereas the GC box associated with basal transcription of the human gene was conserved in the mouse, the GBS associated with repression of the human gene was not. Furthermore, no GATA-binding sites could be identified within the mouse promoter sequence analyzed, and reporter gene analysis using a mouse promoter construct gave the same transcriptional activity in both human fetal lung fibroblasts and adult chondrocytes (Fig. 7B). This could imply that repression of the mouse lumican gene does not follow the same temporal regulation as the human gene. To address this issue, lumican expression was compared in cartilage from both mice and humans. Immunoblotting analysis showed that lumican was present in both neonatal and adult mouse cartilage (Fig. 11A), in contrast to the deficit previously reported in neonatal human cartilage (11). Immunohistochemical analysis confirmed the expression difference (data not shown) and showed that in the mouse neonatal cartilage lumican was expressed throughout the cartilage matrix (Fig. 11B).

View larger version (52K): [in this window] [in a new window]   Fig. 10.   Comparison of promoter sequences from the human and mouse lumican genes. The sequences of the promoter region of the human (hlumpr) and mouse (mlumpr) lumican genes are compared. The GATA-binding sites (double underline) and the GC box (boxed) involved in the repression and activation of transcription, respectively, of the human gene are indicated, as is the location of the atypical TATA box (boxed).

View larger version (65K): [in this window] [in a new window]   Fig. 11.   Analysis of lumican expression in mouse cartilage. A, immunoblot analysis of lumican present in extracts of neonatal (lane 1) and adult (lane 2) cartilage. The migration position of ovalbumin (Mr 46,000) is indicated. B, immunohistochemical analysis of lumican present in mouse neonatal cartilage (magnification, × 100).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The present work provides information on the transcription start site of the human lumican gene, the proximal elements in its promoter involved in basal transcription, and the more distal elements involved in cell-specific repression of transcription. The human lumican gene possesses a single transcription start site residing 95 bp prior to the translation start site in the published cDNA sequence (11). This makes the human gene similar to the mouse lumican gene, which also possesses a single transcription start site and a 5'-untranslated region of 88 bp in its cDNA (13). It is, however, different from the chick lumican gene, which exhibits several transcription start sites and gives rise to a 5'-untranslated region of greater than 300 bp (12).

In the case of the chick gene, the most 5' transcription start site is preceded by two potential TATA boxes that are thought to be involved in controlling transcription (12), whereas in the mouse a more atypical TATA box (TATCA) precedes the transcription start site (13), although at present there is no evidence that these sites are active. In the present work it was shown that the human lumican gene also possesses the TATCA sequence preceding its transcription start site. Although this element by itself cannot initiate transcription, it is essential for transcription to occur in conjunction with an upstream GC box which binds the transcription initiation factor Sp3. This GC box (TGGGCTGGG) resides 74 bp prior to the transcription start site, and it is interesting to note that the mouse gene contains a related sequence (TGGGCTGGT) residing 80 bp prior to its transcription start site (13), which could also be involved in initiation of transcription. Whether this sequence is also an Sp3-binding site or interacts with another member of the Sp1 family needs to be established. At present 16 Sp1 family members capable of binding to a GC box have been described (23). Based upon the limited published promoter sequence (12), it is not clear whether a GC box could also play a role in the chick promoter.

The human PRELP gene, which is closely related to the keratan sulfate proteoglycan subfamily containing lumican (18), also utilizes a GC box in regulating basal transcription (20). This gene also shares with the lumican gene the feature of low expression in human articular cartilage prior to birth, and therefore a common regulatory mechanism could be postulated as operating in the two gene promoters. In the case of the PRELP promoter, repression of gene transcription involves an Ets-binding site residing 497 bp upstream of the transcription start site (20). The human lumican gene also contains several Ets-binding sites within the 1.6-kbp promoter sequence analyzed. One of these resides within the region found to be responsible for repressor activity (Fig. 1), which is at a similar location to its counterpart in the PRELP gene. It is, however, apparent from the present work that the mechanism of repression is not identical for the two genes, as repression of the lumican gene involves a GATA-binding site rather than an Ets-binding site.

Although it is clear that the GATA-binding site is involved in the repression of transcription, it is less clear how occupation of this site leads to repression, as both adult chondrocytes (which do not exhibit repression) and fetal lung fibroblasts (which do exhibit repression) possess GATA proteins capable of interacting with the site. The GATA family of transcription factors consists of at least six members (24, 25), and it is possible that a different GATA-binding protein is present in the two cell populations. If such a scenario exists, the proteins involved do not appear to be GATA-1 to -4 or -6. It is also unlikely that GATA-5 is involved, as its expression has so far been reported only in differentiated heart and gut tissues (25). Therefore, the proteins binding to the GBS may be novel members of the GATA protein family. Alternatively, it is possible that the same GATA protein is involved in both cases but that an additional cell-specific cofactor that interacts with the GATA protein is necessary to elicit repression. The different gel retardations between the two cell types of a probe containing only the active GBS would support such a scenario, as the GATA proteins themselves have similar molecular sizes. Such cofactors have been described and in mammalian cells have been designated by the acronym FOG (friend of GATA) (26). GATA proteins possess two zinc finger-like domains (24) that could potentially interact with adjacent GBS, as found in the repressor region of the human lumican gene. However, these domains have different binding affinities for the GBS and only one need interact for a functional effect. In this case the other domain can be involved in interaction with cofactors in order to regulate transcription (27). In the fetal lung fibroblasts, it appears that only one GBS is involved in repression, and hence it is plausible that the negative regulation of transcription is mediated by a FOG protein. Such interactions would not be expected in adult chondrocytes but might be expected in neonatal human chondrocytes where transcriptional repression occurs.

The question also remains as to why lumican gene expression should be down-regulated in cartilage in the human fetus and neonate but not in mice. This presumably relates to the functional properties of lumican, some of which may not be essential for fetal development of cartilage. At present, lumican is known to interact with a variety of other extracellular matrix proteins, including types I and VI collagen in the cornea (28, 29). The presence of such interactions appears essential for normal development of some connective tissues, as mice that are homozygous for a null mutation in the lumican gene develop major abnormalities in the collagen architecture of their skin and cornea (30, 31). It is likely that lumican is able to interact with the type II collagen present in cartilage, and one therefore presumes that the regulation of collagen fibril diameter imposed by such interaction is not necessary in developing human cartilage. Alternatively, it is possible that other members of the leucine-rich repeat proteoglycan family, which can also interact with the fibrillar collagens, can compensate for this deficit in the human. In this respect it is unclear whether all members of this proteoglycan family can interact at distinct sites on the collagen molecules. Thus while decorin and fibromodulin clearly interact at distinct sites (32), the location of the site of interaction for lumican has yet to be determined. If it is the same as that for fibromodulin, it is possible that this closely related leucine-rich repeat proteoglycan could compensate for the absence of lumican. Indeed, it has been postulated that the absence of major phenotypic changes in the fibromodulin knock-out mouse could be due to compensation by increased expression of lumican (33).

	ACKNOWLEDGEMENTS

We thank N. Nikolajew for typing the manuscript and G. Bédard for the artwork involved in preparing the figures. We are indebted to the Pathology Departments at the Royal Victoria Hospital and the Montreal General Hospital for providing access to tissue.

	FOOTNOTES

^* This work was supported by research grants from the Medical Research Council of Canada (to P. J. R.), the Shriners of North America (to P. J. R.), and National Institutes of Health Grant EY 11845 (to W. W.-Y. K.).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 GenBank^TM/EMBL Data Bank with accession number(s) AF239660 (human) and AF186467 (mouse).

^¶ To whom correspondence should be addressed: Genetics Unit, Shriners Hospital for Children, 1529 Cedar Ave., Montreal, Quebec H3G 1A6, Canada. Tel.: 514-842-5964; Fax: 514-842-5581; E-mail: proughley@shriners.mcgill.ca.

Published, JBC Papers in Press, October 2, 2000, DOI 10.1074/jbc.M004134200

	ABBREVIATIONS

The abbreviations used are: kbp, kilobase pairs; bp, base pairs; BAC, bacterial artificial chromosome; EMSA, electrophoretic mobility shift assay; FOG, friend of GATA; GCB, GC box; GBS, GATA-binding site; HFL, human fetal lung; PCR, polymerase chain reaction; PRELP, proline- and arginine-rich end leucine-rich repeat protein.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This Article Abstract Full Text (PDF) All Versions of this Article: 275/52/40967    most recent M004134200v1 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 Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Grover, J. Articles by Roughley, P. J. Search for Related Content PubMed PubMed Citation Articles by Grover, J. Articles by Roughley, P. J. Social Bookmarking                      What's this?