J. Biochem. 2014;156(1):21–28

doi:10.1093/jb/mvu014

A candidate enhancer element responsible for high-level expression of the aggrecan gene in chondrocytes Received September 23, 2013; accepted February 6, 2014; published online February 18, 2014

Yuichi Ikeda1,2, Kazuo Ito3, Yuichi Izumi2 and Tamayuki Shinomura1,* 1 Tissue Regeneration; 2Periodontology, Department of Bio-Matrix, Tokyo Medical and Dental University, Tokyo 113-8549, Japan; and 3 Department of Orthopaedic Surgery, Tohoku University School of Medicine, Sendai, Miyagi 980-8574, Japan

*Tamayuki Shinomura, Tissue Regeneration, Department of BioMatrix, Tokyo Medical and Dental University, Tokyo 113-8549. Tel: þ81 3 5803 5450, Fax: þ81-3-5803-5450, email: [email protected]

Aggrecan is the most abundant proteoglycan in cartilage. It contains a lot of negatively charged glycosaminoglycan chains along the core protein, providing a large osmotic swelling pressure within the cartilage. Therefore, the biomechanical properties of cartilage, such as its compressive load-bearing capacity, are highly dependent on the presence of abundant aggrecan in the cartilage matrix. To elucidate the transcriptional mechanism that leads to abundant expression of aggrecan by chondrocytes, we screened for enhancer elements in 130 kb of the aggrecan gene, Agc1, using a reporter assay system that we previously developed. The system is based on co-transfection of candidate enhancer elements and reporter constructs into Swarm rat chondrosarcoma chondrocytes that retain a high level of aggrecan expression. We found an element that might be involved in high-level expression of Agc1 gene in chondrocytes in vivo. The element is located 30 kb upstream of the Agc1 transcription start site. Keywords: Aggrecan/cartilage/extracellular matrix/ transcriptional enhancer/type II collagen. Abbreviations: Agc1, a gene encoding aggrecan; bgyg, a b-galactosidase-hygromycin phosphotransferase fusion gene; bp, base pairs; Col2a1, a gene encoding type II collagen alpha 1 chain; kb, kilobase pairs; LTC-RCS, an established cell line from the Swarm rat chondrosarcoma; PCR, polymerase chain reaction; X-gal, 5-bromo-4-chloro-3-indolylb-D-galactoside.

Cartilage is a highly specialized tissue that contains abundant extracellular matrix (1), and the biomechanical function of cartilage is highly dependent on the integrity of that matrix. Matrix integrity is thought to be maintained by well-organized production, assembly and degradation of macromolecules in the matrix (2, 3). The major components of cartilage matrix are collagen type II and a proteoglycan, aggrecan. Type II collagen forms a dense fibrillar network that provides

structural stiffness to the tissue and a high concentration of aggrecan is retained within it. An aggrecan molecule contains about a hundred polyanionic glycosaminoglycan chains. Therefore, aggrecan draws cations and thus water into the extracellular space, creating a hyperosmotic pressure in the tissue. As a result, cartilage matrix provides dynamic resistance to compressive forces. Thus, abundant production of type II collagen and aggrecan is essential for the formation and proper function of cartilage tissue. Recently, the transcriptional regulation of type II collagen and aggrecan has been extensively studied (4, 5) and SOX9 is now widely recognized as a key transcription factor that controls their specific expression in cartilage. However, the underlying mechanisms that control abundant production of type II collagen and aggrecan in cartilage remain unclear. Despite the functional importance of the extracellular matrix in cartilage, the quantitative aspects of matrix synthesis have not been well studied. We recently investigated the transcriptional mechanism that controls high-level expression of type II collagen in chondrocytes and identified an element having strong enhancer activity (6). In this study, we similarly examined aggrecan expression at the transcriptional level in chondrocytes, and we report here results that suggest the presence of an element critical for highlevel aggrecan expression.

Materials and Methods Cell culture An established cell line from the Swarm rat chondrosarcoma (LTC-RCS chondrocytes) was used in this study (7). The cells were cultured at 37 C under 5% CO2 in DMEM10: Dulbecco’s modified Eagle’s medium (Wako Pure Chemical, Osaka, Japan) supplemented with 10% heat-inactivated foetal bovine serum (Biowest, Nuaille, France). Cloning of genomic DNA fragments encoding rat aggrecan gene A rat Bac clone (RNB1-168J05) encoding rat aggrecan gene, Agc1 was provided by RIKEN BRC, which is participating in the National Bio-Resource Project of the MEXT, Japan (8). Bac DNA containing a 167,458-bp genomic DNA fragment was isolated by alkaline lysis (9) and then digested with MfeI and/or SalI. The resultant DNA fragments were cloned into the pGEM-3Zf() vector (Promega, Madison, WI, USA). Each clone was characterized by direct sequencing with forward and reverse M13 primers, 50 -CGC CAGGGTTTTCCCAGTCACGAC-30 and 50 -CACAGGAAACAG CTATGACCATG-30 , using the BigDye terminator cycle sequencing ready reaction kit (Applied Biosystems, Foster City, CA, USA). As shown in Fig. 1A, about 130 kb of the genomic DNA region containing the Agc1 gene is covered by 17 plasmid clones (Clones I—XVII). The subfragments of plasmid clone V shown in Fig. 4A (see below) were generated by digestion with appropriate restriction enzymes and then were cloned into the pGEM-3Zf() vector. Prior to assays of enhancer activity using our own co-transfection reporter assay system described below, each plasmid DNA was linearized with AatII, AhdI, EcoRI, HindIII, KpnI or XhoI.

ß The Authors 2014. Published by Oxford University Press on behalf of the Japanese Biochemical Society. All rights reserved

21

Y. Ikeda et al.

A

1st exon A1

I

III

II

VII

V

IV

VI

IX

VIII

A3

A2

XIII

XI

X

XII

Hapln3

XV

XIV

XVII

XVI

10 kb

B

Aggrecan promoter SD

3’N

SA

bgyg

(a)

a

b

A1 bgyg

(b)

A2 bgyg

(c)

A3 (d)

bgyg

Fig. 1 Rat aggrecan gene, Agc1, and reporter constructs. (A) Diagram of the rat Agc1 gene and the arrangement of fragments covering the 134 kb segment of genomic DNA. The horizontal line represents rat genomic DNA and the location of the 18 exons of the aggrecan gene are shown by white rectangles. The first exon is indicated by the down-pointing arrow. White circles indicate the regions which contain evolutionarily conserved sequences in the aggrecan gene, named A1, A2 and A3, as described by others (5). A flanking gene encoding link protein 3, Hapln3, is represented by the white arrow (indicating the direction of transcription). The grey rectangles labelled with Roman numerals, I—XVII, represent the position and relative size of the DNA fragments that were used in transfection experiments to search for new enhancer elements. The thick black line, bottom left, shows the length of a 10 kb segment of DNA. (B) Reporter constructs, pAgc-bgyg-30 N (a), pA1-Agc-bgyg-30 N (b), pAgcA2-bgyg-30 N (c) and pAgc-A3-bgyg-30 N (d), are schematically represented. A fragment containing rat Agc1 promoter and its first exon and a fragment containing a partial sequence of the second exon are indicated by the dotted box and the grey box, respectively. A SD sequence from the rat Agc1 first exon and a SA sequence from the rat Agc1 second exon are indicated by black triangles. A fusion gene containing bgalactosidase and hygromycin phosphotransferase is indicated by arrows labelled bgyg. A fragment containing poly A signal from rat Agc1 gene (30 N) is indicated by the diagonally striped box. DNA segments encoding A1, A2 and A3 (described above) were ligated to the positions indicated by arrows, a and b. Construction of reporter plasmids A vector, pAgc-bgyg-3’N [Fig. 1B(a)] was first constructed as follows. A small fragment containing the splice acceptor sequence (SA) of rat Agc1 second exon was amplified from clone XI by polymerase chain reaction (PCR) using the specific primers, 50 -GAGGCTGTG GTTCTCCAGAGCTGCAGAGAC-30 and 50 -CTCGTCGACAGA GTTCAGCTGGAAGAGAC-30 . In this fragment, the intrinsic start codon in the second exon was replaced with CTG. The resulting fragment was inserted into the pGEM-3Zf() cloning vector. Then, a fusion gene containing coding sequences for b-galactosidase and hygromycin phosphotransferase (bgyg) was excised from the prvPtrap plasmid (10) and ligated next to the splice acceptor sequence described above. This new plasmid vector was tentatively named pSA-bgyg. Next, for construction of pAgc-bgyg-30 N, two

22

DNA fragments were prepared as follows: (i) a fragment containing a 534-bp rat Agc1 promoter, exon 1 and a splice donor sequence (SD) was excised from plasmid clone VIII and (ii) a 0.9-kb DNA fragment encoding the last exon and 30 -non-coding region (30 N) of rat Agc1 gene was excised with PstI and BglII from the Bac DNA. Each DNA fragment was then sequentially ligated to pSA-bgyg described above. The nucleotide sequences around the junction of each ligated DNA fragment were verified by DNA sequencing. The whole sequence of the reporter plasmid, pAgc-bgyg-30 N was submitted to the DDBJ (accession number AB842177). Next, for the construction of reporter plasmids containing the evolutionarily conserved sequences, A1, A2, and A3 in Agc1 gene (5), three DNA fragments were prepared as follows: (i) fragments containing A1 and A2 were excised from plasmid clones VIII and

Aggrecan gene expression

**

A

Table I. List of the primers used to prepare the PCR fragments shown in Fig. 4B.

**

** Colonies / Plate

250 200 150 100 50 0

A-

B

A-

A1

A2

A3

A1

Primer

Sequence

S1 S2 S3 S4 S5 A1 A2 A3 A4 A5 M1F M1R M2F M2R

50 -GCATGCACATGCACACACACCACGTATG-30 50 -CTGGTACCTGCCTCCGGGAAGACTGCTCC-30 50 -CCCTCCCTTGTCCCCTCCTGATATTTCCAG-30 50 -GCTGCCACAGCCCTAATTATGTGTGAAATC-30 50 -CTCAAAGCCCTGTACGTAATGAGGCCAC-30 50 -GGATCCACACAAAAGGAGAGAGCTGACTCC-30 50 -GAGGTACCTCTTTCCTGTGGCCTCATTAC-30 50 -GGCACCTCAGTGGGCAGGGAGGCAGCTTG-30 50 -GTAGCCTGAGGTAACCTGGCAGCAGCCAG-30 50 -CGTTCAAGGTGGGGAGCAGTCTTCC—30 50 -CCAAAGCAAAGTTACACACACAGAGAAGCCC—30 50 -CTCTGTGTGTGTAACTTTGCTTTGGTTTAAAGC—30 50 -AACGTTAGCAAACAAACACACACAGAGAAGC—30 50 -CTGTGTGTGTTTGTTTGCTAACGTTTAAAGCTTT—30

KpnI recognition sequences are underlined. Mutated nucleotides are indicated by bold letters.

A2

A3

Fig. 2 Expression of reporter constructs in LTC-RCS chondrocytes. (A) Number of colonies generated by transfection of pAgc-bgyg-30 N (A), pA1-Agc-bgyg-30 N (A1), pAgc-A2-bgyg-30 N (A2) and pAgcA3-bgyg-30 N (A3). Data are mean  SD (n ¼ 3). **P50.01. (B) Typical colonies generated in the transfection experiments. Most of the colonies were negative after 4 h staining with X-gal.

XI, respectively, and (ii) a fragment containing A3 was amplified from plasmid clone XIV by PCR using the specific primers, 50 -CA GGGTACCTCTGGAACAACCAGATGCCAC-30 and 50 -CCGAG TCTTAGGTACCTTCCTCAACCGGTG-30 . Each DNA fragment was then subcloned into cloning site a or b of pAgc-bgyg-30 N, as shown in Fig. 1B(a). The resultant reporter plasmids were named pA1-Agc-bgyg-30 N, pAgc-A2-bgyg-30 N and pAgc-A3-bgyg-30 N, respectively [Fig. 1B(b), (c) and (d)]. The nucleotide sequences around the junction of each ligated DNA fragment were verified by DNA sequencing. The whole sequences of these reporter plasmids were submitted to the DDBJ (accession numbers AB842178, AB842179 and AB842180, respectively). The specificity of a newly identified enhancer element was confirmed by comparison with pCol2(P/int1)-bgyg-3N0 and pSV-bgyg, which we constructed previously (6).

Preparation of PCR fragments As shown in Fig. 4B, small fragments derived from the fragment V7 were generated by PCR using various sets of specific primers (Table I, S1—S5 and A1—A5). For example, fragment V7c, which was the smallest one showing strong enhancer activity, was amplified by PCR using S2 and A2 primers. To confirm the specificity of V7c, transversion mutations were introduced to PCR fragments (Fig. 5B, V7cM1 and V7cM2) using specific primers (Table I, M1F and M1R or M2F and M2R) following the procedure of Kriegler (11). To analyse the exchangeability of enhancer elements newly found in Agc1 and Col2a1 genes in our reporter assays (Fig. 6), a 481-bp fragment (hereinafter abbreviated as E2) from intron 7 of rat Col2a1 gene that contains an enhancer element (6) was amplified from the rat genomic DNA clone (RDB-2673) by PCR using specific primers 50 -CGACGCGTGCGGCTTCTAAATTGCTACTC-30 and 50 -GCT CTAGACCTCCAAATGGCACACC-30

All PCR fragments were purified with a QIAquick PCR Purification Kit (Qiagen, Valencia, CA, USA) and used directly in the co-transfection experiments described below.

Co-transfection experiments A total of 1.0  107 LTC-RCS chondrocytes were transfected in 0.8 ml of phosphate-buffered saline (PBS) with 2.5 pmol of different reporter constructs (Fig.1B) that had been linearized by digestion with AhdI. Briefly, for reporter assays, 2.5 pmol of linearized pA1Agc-bgyg-30 N was first mixed with a 2-fold molar excess of linearized plasmid DNA that contained rat aggrecan genomic DNA fragments to be tested or with a 2-fold molar excess of the PCR products described above. The resultant mixtures were then transfected into LTC-RCS chondrocytes by electroporation using a Gene Pulser (Bio-Rad, Hercules, CA, USA) with 250 V and a capacitance of 950 mF. After electroporation, the cells were cultured for 2 days in DMEM10, and then the culture media were replaced with the medium containing 0.2 mg/ml hygromycin B (Wako Pure Chemical). For assay of reporter activity, the cells were cultured for 12 days with replacement of the hygromycin-containing medium every other day. Resultant hygromycin-resistant colonies were stained with 5-bromo-4-chloro-3-indolyl-b-D-galactoside (X-gal, Wako Pure Chemical), as described previously (12). Briefly, colonies were rinsed three times with PBS, fixed for 15 min with 1% glutaraldehyde and rinsed four times with PBS before staining with X-gal for 4 h at 37 C. After X-gal staining, colonies composed of more than 50 cells were counted and each colony was classified semiquantitatively according to the proportion of X-gal-positive cells in it as follows: (), 510%; (), 10—80%; (þ), 480% (Fig. 3A).

Gadolinium chloride treatment We previously reported that the chondrogenic phenotype of LTCRCS chondrocytes was specifically inhibited by gadolinium chloride (0.4 mM) (12, 13). Therefore, stable transformants resulting from the co-transfection with the reporter construct pA1-Agc-bgyg-30 N and B1 enhancer element (see ‘Results’ section for details) were plated at 1  104 cells/cm2 and cultured in DMEM10 containing 0.4 mM gadolinium chloride for 5 days, with replacement of the medium every other day, as described previously (12). The effects of gadolinium chloride on the reporter gene expression were monitored by X-gal staining for 4 h, as described above.

Statistical analysis All experiments were performed in triplicate and statistical analyses were conducted using an unpaired t-test. All error bars represent the standard error of the mean. Differences were considered significant at P-values 50.05 (*P50.05 and **P50.01).

23

BglII

BamHI KpnI

A

+

+/-

XmaI

-

SphI

A

KpnI

Y. Ikeda et al.

V V1

V2

B

V3

(Fold)

V4

*

Colony number

5

V5

4

V6

3

V7

2

1 kb

B

XmaI

1

V7

0 XVII

Fig. 3 Co-transfection reporter assay. (A) Typical colonies generated in the co-transfection experiments. LTC-RCS chondrocytes were transfected with pA1-Agc-bgyg-30 N in the presence and absence of the genomic DNA fragments shown in Fig. 1A. After selection with hygromycin, X-gal staining revealed the presence of various colonies that exhibited various intensities of b-galactosidase staining, as described in ‘Materials and Methods’ section: negative (), partially positive () and clearly positive (þ). (B) Number of colonies generated in co-transfection experiments (white, negative; grey, partially positive; black, clearly positive for X-gal staining). Fragment V significantly enhanced both colony formation and the expression level of reporter gene. The level of enhancement was quantified relative to the reporter construct, pA1-Agc-bgyg-30 N alone (Cont). On the other hand, none of the other 16 similarly transfected and tested genomic DNA fragments showed any activity enhancement. Data are presented as mean  SD (n ¼ 3). P-values are evaluated versus Cont. *P50.05.

V7c V7d V7e V7f V7h

C

(Fold)

7

We previously reported an enhancer element that might be responsible for regulating the expression level of type II collagen gene (Col2a1) in chondrocytes (6). In that study, we developed a new reporter assay system based on co-transfection of candidate enhancer elements and a reporter construct into Swarm rat chondrosarcoma chondrocytes that allowed their stable expression. The system was very effective for identifying enhancer activities localized in a wide range of genomic DNAs. So, in this study, we examined the transcriptional mechanism that controls the expression level of the aggrecan gene (Agc1) in chondrocytes by using that stable transformation assay system. Before screening for possible, yet-unidentified enhancer elements, we re-examined the enhancer activity of three previously reported regulatory elements, A1, A2 and A3 (5). As shown in Fig. 1B(a), we first constructed a reporter mini-gene pAgc-bgyg-30 N, that contained the basal regulatory elements. Then, using this mini-gene, one copy of A1, A2 or A3 element was ligated to the positions indicated in Fig. 1B(b—d) so as to resemble endogenous Agc1 gene. These mini-genes were then linearized and transfected into LTC-RCS chondrocytes. After selection with hygromycin, the number of generated colonies differed among the constructs containing A1, A2 and A3 (Fig. 2A). The results confirm that A1 is a potent enhancer, as previously reported (5). However, 24

**

**

6 5

**

*

4 3

**

*

2 1 0 Cont

Results

100 bp

V7g

Colony number

XV

XVI

XIV

XII

XIII

X

XI

IX

VII

VIII

V

VI

III

IV

I

II

Cont

V7a V7b

V7

V7a

V7b

V7c

V7d

V7e

V7f

V7g

V7h

Fig. 4 Locations of possible enhancer elements and their effects on expression of the reporter gene. (A) Alignment of overlapping restriction fragments derived from fragment V. To identify enhancer elements in fragment V, seven subfragments were prepared by digestion with appropriate restriction enzymes, as indicated by the arrows. The subfragments shown in black were positive for enhancer activity. On the other hand, subfragments shown in white gave no enhancement. The size of a 1 kb length is shown at the bottom right for comparison. V7 was the smallest subfragment showing enhancer activity. (B) Diagram of the relative lengths and positions of PCR products tested for enhancer activity. The PCR products shown in black and striped bars were clearly and partially positive for enhancer activity, respectively. Fragments shown in white gave no enhancement. The arrow indicates the XmaI restriction site. A scale representing 100 bp is shown at the bottom right. (C) Number of colonies generated in co-transfection experiments using V7 and its PCR products (V7a through V7h) and their intensities of X-gal staining (white, negative; grey, partially positive; black, clearly positive). Fragments V7 and V7c significantly enhanced both colony formation and the expression level of the reporter gene compared with the reporter construct, pA1-Agc-bgyg-30 N, alone (Cont). Fragments V7d, V7e, V7f and V7g showed reduced enhancer activity, whereas V7a, V7b and V7h show no enhancer activity. Data are presented as mean  SD (n ¼ 3). P-values are evaluated versus Cont. *P50.05, **P50.01.

X-gal staining revealed that most colonies were palecoloured, indicating a low expression level of the reporter gene (Fig. 2B). In other words, the expression level of the A1-containing reporter construct was unexpectedly low even though the endogenous aggrecan gene is highly active in LTC-RCS chondrocytes (7). These data thus strongly suggested the presence of

Aggrecan gene expression Rat Hom Bos Sus Gal

A

B

CTGCC-TCCG GAGCC-TCCT NNNNNNNNNN CTGCC-TCTG CTGCCCTCCT

GGAAGACTGC GTAAG-CTGC NNNNNNNNNN TGAAAGCGGC –GAACAGT-T

T---CCC-CA T---TCTGCG NNNNNNNNNG T---TCT-CT TCTTTCTTGC

-CCTTGAACG -CCCTGAACG TCC-TGAACA GCCCTGAACA AGCTTGAACA

CTGCCCGTTG CTGCCTGTTG CTGCCTGTTG CTGCCTGTTG CTCTGCATTG

TCCAGTCCTT TCCCCTCCAG TCCGATCCAC TCCCATCCGT TCC-TTCCC-

CCAGCTCCTC --AGCGCCTC GCAGCCTGGG GCAGCCCGGG ---------C

AC-------GC-------GC-------GC-------AC-CATTTTG

---------T ------GGCT ---------T ---------T CAAGCAAGGT

CCTCCTCCTT CCTCCTCCTC CCTCCTCCAG CCTCCTCCTG CCTTCTCCTG

CTCACAACCT CTCCCCGCCC CTCCTAGCCC CTTCCAGCCC GTT----GCT

AG-CTGGAAA GG-CTGGAAG CGGCTGGAAG CGGCTGGAAG AGTGTT-AAC

AGGAACACAA AGGAACACAG AGGAACACAG AGGAACACAG ACAGGTT---

CCCTCCCTTG CCCTCCCTTG TCCTGCCTTG ACGTCCCTTG ---TCCCTTG

TCCC--CTCC TCCCTCCTCTCCCTCCTCTCCCTCCTCT-TCT-GTCC

--TGATATTT -TTGATATTT -TTGATATTT -TTGATATTT CTTGAAATTT

CCAGTAAAAG CCAGTAAAAG CCAGTAAAAG CCAGTAAAAG CCAGTAAA-G

TACCTTGAAC CACCTCGAAC CACCTCGAAC CACCTCGAAC CGCCTTTAAT

ACAATCTCGC ACAATCTCGC ACAATCTCAC ACAATCTCGC ACAATCTCTC

TGCCACAGCC TGCCACAACC TGCCACAGCC TGCCACAGCC TGTGGCTGCT

CTAATTATGT CTAATTATGT CTAATTATGT CTAATTATGT CTAATTATGT

GTGAAATCGG GTGAAATCAT GTGAAATCAT GTGAAATCAT GTGAAAT-AT

TTTTTTTTTT TTTTTTTT-TTTTTTTTTT TTTTTTTTTTTTTT-----

TTAA--AGCT ------AGCT TTTTTTAGCT ------AGCT ----CCCCCC

TTAAACCAAA TGAAACCAAA CGAAACCAAA CGAAACCAAA ATAAACCAAA

GCAAACAAAC GCAAACAAGC GCAAACAAGC GCAAACAAGC GAAAACAAAT

ACACACAGAG GCGCACAGAG GCACACAGAG GGACACAGAG CAACACAGT-

AAGC-CCATT AAGC-CCATT AAGC-CCATT AAGC-CCATT AAGATCCTTT

CTC-CAAGCC CTC-CGCGGC CTCTGCGGCC CTCCGAGGCC CTCTCAGGCT

CGGGAGCAGC C-GGCGCGGC CAGGAGGAGC CAGGAGGAGC GTTTGGCAGG

AGGGCCTGGC A--GCCTGGC --GGTCTCGC A--GCCTCGC TCTCACGGCC

TGCTGCCAGG CGCTG--TGG TCCTGTTGGG TCCTGTTGGG TTCAGAAAAG

TTACCTCA-G GTAGCTCAGG TTACCTCA-G TTACCTCA-G TTACCA-A-G

G-CTACACAG G-ACGCACAG GCCTGCACAG GCCTGCACAG GTCCACACAG

AGGCCCAGCT AGGCCCGGCT AGACCCCGCT AGACCCCGCT TCTCT--GCT

G--TTT-CAC G--TTT-CGC GTGTTTTGCA GTGTTTCGCA GTCTTTTCAC

AAGCTGC-CT AAGCTGC-CC AAGCTGC-CC A-GCTGC-CC AACA-GCACT

CCCTGC--CC GCCTGG--CC GCCTGG--CC GCCCGG--CC TTTGGGTTCC

ACTG--AGGT A-----AGGT A-----AGGT A-----AGGT TGTAAAAGGA

GCCTCCCCTG GCCTCCCAGG GCCTCCCAGG GCCTCCCAGG CCCAACAAAG

CCTG-ACCCG CCTG-CCCGG CCTG-ACCAG CCTGACCAGG ----------

TGAGGCCGCC GGAGGTCGCC GGAAGCGGTC GAAAGCGGTC GAAAGGG-TC

CA-CCC---CACCCC---CA-CCCCCCA CA-CCC---AAACAAATCA

-CTCAGAGGA -CTCGGAGGA ACTCGGGGGA TCTCGGGGGG CAAAGCCCTG

AAGAGGTGGAAGA-GTGGG AAGA-GTGGG AAAGAGTGGG AACGCCT-AT

TCTCAGTGTG TTTCAGTTTG TTTCAGTTTG TTTCAGTTTG TTT-------

T----GTCTCCTGCACT TTCTGCACT TTCTGCACG ---------

TTCTACACAATTCCACGCAGCTCCACACAGCCCCTCACAGCCCACCTGAAT

GCTTACCTGG-TTGACGGG G-TTTCCAGG GCTT-ACAGG G-TTCACAGG

AGACTCTCAA AGCCTCACAC AGCCTCACAC AGCCTCACAC ---CTGA---

AGCCCT-GTA AGGCCTGGCA AGGCCT-GCAGGCCAGGTA AGGG------

CGTAAT--GA CACTCT--GA TGTATTGCGA TACTGT-GAA --------AA

GGC------GGC------GGC------GGC------GGCTGTCTGC

-------CA -------CA -------CA -------CT TGAGTGACA

CAGGAAAGAG CAGGAAAACG CAGGAAAGGC ATAGAAAGGC GAACAAAGGC

V7c

AAAGCTTTAAACCAAAGCAAACAAACACACACAG

V7cM1

AAAGCTTTAAACCAAAGCAAAGTTACACACACAG

V7cM2

AAAGCTTTAAACGTTAGCAAACAAACACACACAG

C

(Fold)

**

7

*

Colony number

6 5 4 3 2 1 0 Cont

V7c

V7cM1

V7cM2

Fig. 5 Nucleotide sequence of the smallest fragment exhibiting enhancer activity. (A) Nucleotide sequence of the newly identified B1 enhancer element corresponding to fragment V7c shown in Fig. 4B. The sequence was aligned with the corresponding sequences of human (Hom), cow (Bos), pig (Sus) and chicken (Gal) genes. Individual nucleotides that do not match with the rat nucleotide are indicated with bold letters. Underlined sequences represent the highly homologous regions among the different animal species. The XmaI restriction site shown in Fig. 4A is indicated with bold underlining. Two possible SOX9-binding sites are indicated by double underlining. (B) A partial nucleotide sequence in the B1 enhancer element (V7c) and its mutated sequences (V7cM1 and V7cM2) are shown. The possible SOX9-binding sites are underlined in V7c. Mutated nucleotides are indicated in bold. (C) Number of colonies generated by co-transfection of pA1-Agc-bgyg-30 N with mutant B1 elements (white, negative; grey, partially positive; black, clearly positive for X-gal staining). Mutant fragments V7cM1 and V7cM2 caused clear reduction in colony formation compared with the wild-type V7c fragment, although both fragments still had some enhancer activity. The level of enhancement was quantified relative to the reporter construct, pA1-Agc-bgyg-30 N, alone (Cont). Data are presented as mean  SD (n ¼ 3). P-values are evaluated versus Cont. *P50.05, **P50.01.

25

Y. Ikeda et al.

A (Fold)

*

8

**

4 0 Cont

B1

E2

pA1-Agc-βgyg-3’N B (Fold)

*

Colony number

20 16

*

12 8 4 0 Cont

B1

E2

pCol2(P/int1)-βgyg-3’N C (Fold) 4 0 Cont

B1

E2

pSV-βgyg Fig. 6 Functional interchangeability of newly identified enhancer elements in Agc1 and Col2a1 genes. The newly identified enhancer elements, B1 in Agc1 and E2 in Col2a1 were co-transfected into LTC-RCS chondrocytes together with three reporter constructs, pA1-Agc-bgyg-30 N, pCol2(P/int1)-bgyg-30 N and pSV-bgyg. Although B1 and E2 had no effect on the expression of pSV-bgyg (C), both elements had significant effects on the expression of pA1-Agc-bgyg-30 N (A) and pCol2(P/int1)-bgyg-30 N (B). Both colony formation and the expression level of the reporter genes were significantly enhanced in comparison to each reporter construct alone (Cont). Data are presented as mean  SD (n ¼ 3). P-values are evaluated versus Cont. *P50.05, **P50.01.

yet-unidentified enhancer elements that enable highlevel expression of the endogenous Agc1 gene in chondrocytes. To identify such enhancer elements, linearized pA1-Agc-bgyg-30 N [Fig. 1B(b)] was co-transfected into LTC-RCS chondrocytes with plasmid DNA containing fragments of rat Agc1 genomic DNA (i.e. plasmid clones I-XVII shown in Fig. 1A). Fragment V significantly enhanced colony formation (Fig. 3B). X-gal staining was also enhanced, with 465% of the colonies clearly positive for X-gal staining. This enhancer activity was restricted to fragment V, whereas the other fragments had either no effect (fragments IV, XI and XVI) or an inhibitory effect (fragments I—III, VI—X, XII—XV and XVII) on the colony number (i.e. expression of the reporter gene). 26

To identify the enhancer element(s) contained in fragment V, we prepared seven subfragments by digesting it with appropriate restriction enzymes. The subfragments were then recloned into the plasmid vector, pGEM3Zf() (Fig. 4A), and tested for enhancer activity by co-transfection with pA1-Agc-bgyg-30 N. V7 was the smallest subfragment having enhancer activity (Fig. 4A and C), but it is still large in size (854 bp). To narrow down the location of the enhancer activity in V7, we generated eight smaller DNA fragments by PCR and then used them directly with pA1Agc-bgyg-30 N in co-transfection experiments (Fig. 4B and C). A 475-bp fragment, V7c, was the smallest one to have enhancer activity comparable to that of subfragment V7. Fragments 5475 bp in size showed either clearly decreased colony formation (fragments V7d—V7g) or no enhancer activity (fragment V7h). These data suggested that this 475 bp segment located 30 kb upstream of the transcription start site of the aggrecan gene is a novel enhancer element, and we named it ‘B1’ (Fig. 5A). Next, we performed a computer-based sequence homology search between rat B1 element and the corresponding elements in the human, cow, pig and chicken. The search revealed that the underlined sequence is highly conserved among different animal genomes. Furthermore, computer-based sequence analyses revealed that this highly conserved sequence contains possible SOX9 interaction sites. However, the nucleotide sequences of those sites do not seem appropriate for the binding site of SOX9 homodimer. To confirm the importance of those sites, we introduced transversion mutations into the sequence of V7c fragment at the sites indicated by double underlining (Fig. 5A). These mutations each caused a clear decrease in enhancer activity, as shown by the colony number data in Fig. 5B and C. In a previous study, we identified an enhancer element (hereinafter referred to as E2) in intron 7 of the rat Col2a1 gene (6). Although the E2 element has no sequence homology to rat aggrecan B1 element, both genes are specifically and abundantly expressed in chondrocytes and their activation is similarly regulated by transcription factor SOX9 (5, 13). We therefore tested whether these two elements have similar effects on the expression of aggrecan and type II collagen reporter constructs. For this assay, three different reporter constructs, pA1-Agc-bgyg-30 N, pCol2(P/ int1)-bgyg-30 N (6), and pSV-bgyg (6) were used to co-transfect LTC-RCS chondrocytes along with B1 or E2 element (Fig. 6). The results clearly indicated that each element has apparent enhancer activity for both aggrecan and type II collagen reporter constructs. Furthermore, the results showed the specificity of their enhancement of reporter activity, because they did not enhance expression of pSV-bgyg at all (Fig. 6C). To further confirm the specificity of the B1 element, we examined whether its enhancer activity was closely related to the chondrogenic phenotype of LTC-RCS chondrocytes, as previously reported for E2 element (6). Stable transformants resulting from co-transfection of LTC-RCS chondrocytes with pA1-Agc-bgyg30 N and B1 element were treated with gadolinium chloride (12, 13). As shown in Fig. 7, that treatment

Aggrecan gene expression

A

B

Fig. 7 Effect of gadolinium chloride on expression of the reporter gene. The enhancer activity of B1 element (fragment V7c) is closely related to the chondrogenic phenotype of LTC-RCS chondrocytes. Stable transformants containing B1 element were clearly positive after 4 h of staining with X-gal (A). However, gadolinium chloride treatment caused clear reduction in X-gal staining, together with dramatic morphological changes (B). Bar ¼ 100 mm (A, B).

suppressed expression of b-galactosidase and led to dramatic morphological changes. These data suggest that the B1 enhancer element expresses its activity in differentiated chondrocytes. This was further supported by the fact that no stable transformants were generated when we co-transfected pA1-Agc-bgyg-30 N and B1 element into two murine osteogenic progenitor cell lines, Kusa A1 (14) and MC3T3-E1 (15) (data not shown). Hence, the data indicated that the enhancement of reporter gene expression occurs in a chondrocyte-specific manner.

Discussion Aggrecan is the major component of cartilage and its transcriptional activation was reported to be controlled by the Sox trio (5). However, we found that the reporter constructs containing known cis-acting elements such as SOX9-binding enhancer elements were still not enough to achieve high-level endogenous Agc1 gene expression in Swarm rat chondrosarcoma chondrocytes (Fig. 2B). Therefore, we next screened for yet-unidentified enhancer elements that could stimulate the expression of the Agc1 gene. As a result, we identified an enhancer element, B1, that is located 30 kb upstream of the transcription start site of Agc1 gene. Recently, this same 30 kb upstream region was predicted to contain a potential enhancer element by chromatin immunoprecipitation and high-throughput sequencing (ChIP-Seq) analyses of the functional genome in the mouse embryonic limb (16). Therefore, the B1 element might have enhancer activity in chondrocytes in vivo. Initially, we expected that the A1 and B1 enhancer elements would act synergistically to stimulate the reporter gene expression. However, the enhancer activity of B1 element is so strong that we could not detect any synergistic effects (data not shown). Although B1 enhancer worked without A1 element in LTC-RCS chondrocytes, we cannot rule out the possibility that the A1 element is important in vivo. Further work will be required to elucidate the physiological significance of these two enhancers.

Colonies recovered as stable transformants showed a variety of staining intensities with X-gal (Fig. 3A). As discussed earlier (6), the variation might be due to heterogeneous concatemer formation by the DNAs used, some kind of chromosomal position effect on the reporter gene expression, etc. In this study, the molar ratio between the reporter construct and enhancer elements exerted a significant effect on the reporter gene expression (data not shown). Therefore, we used a constant molar ratio for the reporter construct and test DNA fragments and were able to obtain reproducible results in co-transfection experiments. Furthermore, it became clear that the enhancer activity of B1 element gradually decreased as its size become smaller, as shown in Fig. 4C. As we had seen no size dependency for the enhancer activity of the E2 enhancer fragment of the Col2a1 gene (6), this finding for B1 element suggests that some higher order chromatin structure may be important for its enhancer activity or that multiple transcription factors may bind to it. On the other hand, B1’s enhancer activity completely disappeared when DNA fragments were digested with a restriction enzyme, XmaI (V2 and V3 fragments shown in Fig. 4A). Therefore, the sequence around the XmaI restriction site would appear to be essential for the enhancer activity. This is supported by the fact that the nucleotide sequence adjacent to the XmaI site is highly conserved across various animal species (Fig. 5A). Although we could not find any characteristic sequences, such as palindromic sequences, inverted repeat sequences or high GC content, in this region, the mutant fragments V7cM1 and V7cM2 showed significantly less activity than the wild-type fragment V7c (Fig. 5C). Therefore, the sequence of this conserved region may be important for formation of transcriptional complex with some factors. Although SOX9 may be involved in this complex (17), further work is needed to identify transcription factors that bind to the B1 element. The transcriptional regulation of the aggrecan and type II collagen genes has been extensively studied (5, 18), leading to the widely accepted belief that SOX9 is the key transcription factor that controls the expression of both genes in cartilage. However, the 27

Y. Ikeda et al.

nucleotide sequences of SOX9-binding enhancer elements in Agc1 and Col2a1 genes are neither identical nor homologous to each other. Nevertheless, it was reported that aggrecan A1 enhancer element activated the Col2a1 minimal promoter (5), which suggests that the cis-acting elements in Agc1 and Col2a1 genes are functionally similar to each other. That led us to examine the interchangeability of the B1 and E2 enhancer elements in the activation of aggrecan and type II collagen reporter genes containing SOX9-binding enhancer elements (Fig. 6). The results clearly showed that both elements effectively stimulated expression of the Agc1 and Col2a1 mini-genes, but the enhancer activity of B1 element was less potent than that of E2 element. These findings suggest that the expression levels of the aggrecan and type II collagen genes are regulated differently in vivo. In this study, we screened 130 kb of the Agc1 gene and found an enhancer element that might be responsible for up-regulation of the gene’s expression level. However, a region far upstream of the flanking gene encoding interferon-stimulated exonuclease gene 20, Isg20 (a region between 60 and 150 kb from the Agc1 transcription start site), remains to be analysed. Although a high degree of sequence conservation between rat and human sequences was not observed in this region, some enhancer might have been hidden. In addition, there was almost no overlapping of the genomic DNA fragments used for the co-transfection experiments. Therefore, some enhancers may have been missed because they were destroyed by the restriction enzyme digestion used to generate the DNA fragments. Further work will be required to determine whether or not additional enhancer elements are present.

5.

6.

7.

8.

9.

10.

11.

12.

Acknowledgements We are grateful to Dr Masaki Yanagishita for critical reading of this manuscript. We also thank Dr James H. Kimura and Dr Magnus Ho¨o¨k for their continuous encouragements.

13.

14. Funding This work was supported in part by Grants-in-Aid for Scientific Research (23659713) from the Ministry of Education, Science, Sports and Culture of Japan.

15.

Conflict of Interest None declared.

16.

References 1. Hall, B.K. (2005) Bone and Cartilage: Developmental and Evolutionary Skeletal Biology. Elsevier Academic Press, London 2. Morris, N.P., Keene, D.R., and Horton, W.A. (2002) Morphology and chemical composition of connective tissue: cartilage in Connective Tissue and Its Heritable Disorders (Royce, P.M. and Steinmann, B., eds.) pp. 41—65, Wiley-Liss Inc., New York 3. Heinega˚rd, D. (2009) Proteoglycans and more — from molecules to biology. Int. J. Exp. Pathol. 90, 575—586 4. Zhou, G., Lefebvre, V., Zhang, Z., Eberspaecher, H., and de Crombrugghe, B. (1998) Three high mobility group-like sequences within a 48-base pair enhancer of 28

17.

18.

the Col2a1 gene are required for cartilage-specific expression in vivo. J. Biol. Chem. 273, 14989—14997 Han, Y. and Lefebvre, V. (2008) L-Sox5 and Sox6 drive expression of the aggrecan gene in cartilage by securing binding of Sox9 to a far-upstream enhancer. Mol. Cell. Biol. 28, 4999—5013 Shinomura, T., Ito, K., Ho¨o¨k, M., and Kimura, J.H. (2012) A newly identified enhancer element responsible for type II collagen gene expression. J. Biochem. 152, 565—575 King, K.B. and Kimura, J.H. (2003) The establishment and characterization of an immortal cell line with a stable chondrocytic phenotype. J. Cell. Biochem. 89, 992—1004 The STAR Consortium. (2008) SNP and haplotype mapping for genetic analysis in the rat. Nat. Genet. 40, 560—566 Coyne, C.J., McClendon, M.T., Walling, J.G., Timmerman-Vaughan, G.M., Murray, S., Meksem, K., Lightfoot, D.A., Shultz, J.L., Keller, K.E., Martin, R.R., Inglis, D.A., Rajesh, P.N., McPhee, K.E., Weeden, N.F., Grusak, M.A., Li, C.M., and Storlie, E.W. (2007) Construction and characterization of two bacterial artificial chromosome libraries of pea (Pisum sativum L.) for the isolation of economically important genes. Genome 50, 871—875 Shinomura, T., Nakamura, S., Ito, K., Shirasawa, S., Ho¨o¨k, M., and Kimura, J.H. (2008) Adsorption of follicular dendritic cell-secreted protein (FDC-SP) onto mineral deposits. J. Biol. Chem. 283, 33658—33664 Kriegler, M. (1990) PCR-based expression in Gene Transfer and Expression - A Laboratory Manual (Kriegler, M., ed.) pp. 165—172, Stockton Press, New York Shirasawa, S., Yoshimi, M., Kamochi, H., Muneta, T., Kimura, J.H., Ho¨o¨k, M., and Shinomura, T. (2005) Gene trap screening for cell surface and extracellular matrix molecules produced by chondrocytes. J. Biochem. 137, 79—85 Perkins, G.L., Derfoul, A., Ast, A., and Hall, D.J. (2005) An inhibitor of the stretch-activated cation receptor exerts a potent effect on chondrocyte phenotype. Differentiation 73, 199—211 Umezawa, A., Maruyama, T., Segawa, K., Shadduck, R.K., Waheed, A., and Hata, J. (1992) Multipotent marrow stromal cell line is able to induce hematopoiesis in vivo. J. Cell. Physiol. 151, 197—205 Sudo, H., Kodama, H.A., Amagai, Y., Yamamoto, S., and Kasai, S. (1983) In vitro differentiation and calcification in a new clonal osteogenic cell line derived from newborn mouse calvaria. J. Cell Biol. 96, 191—198 Shen, Y., Yue, F., McCleary, D.F., Ye, Z., Edsall, L., Kuan, S., Wagner, U., Dixon, J., Lee, L., Lobanenkov, V.V., and Ren, B. (2012) A map of the cis-regulatory sequences in the mouse genome. Nature 488, 116—120 Nakamura, Y., Yamamoto, K., He, X., Otsuki, B., Kim, Y., Murao, H., Soeda, T., Tsumaki, N., Deng, J.M., Zhang, Z., Behringer, R.R., de Crombrugghe, B., Postlethwait, J.H., Warman, M.L., Nakamura, T., and Akiyama, H. (2011) Wwp2 is essential for palatogenesis mediated by the interaction between Sox9 and mediator subunit 25. Nat. Commun. 2, 251 Lefebvre, V., Huang, W., Harley, V.R., Goodfellow, P.N., and de Crombrugghe, B. (1997) SOX9 is a potent activator of the chondrocyte-specific enhancer of the proa1(II) collagen gene. Mol. Cell. Biol. 17, 2336—2346