Unique expression patterns of matrix metalloproteinases in regenerating newt limbs

Authors

  • Tomoko Kato,

    1. Department of Ophthalmology and Visual Science, Division of Frontier Medical Science, Biomedical Research, Graduate School of Biomedical Science, Hiroshima University, Hiroshima, Japan
    2. Developmental Biology Laboratory, Department of Biological Science, Graduate School of Science, Hiroshima University, Higashihiroshima, Japan
    Search for more papers by this author
    • Drs. Kato and Miyazaki contributed equally to this work.

  • Koyomi Miyazaki,

    1. Yoshizato MophoMatrix Project, Exploratory Research for Advanced Technology, Japan Science and Technology Corporation, Higashihiroshima, Japan
    Current affiliation:
    1. Clock Cell Biology Group, Institute of Molecular and Cell Biology, National Institute of Advanced Industrial Science and Technology, Tsukuba, Japan
    Search for more papers by this author
    • Drs. Kato and Miyazaki contributed equally to this work.

  • Keiko Shimizu-Nishikawa,

    1. Regenerative Biology Laboratory, Department of Biological Science, Graduate School of Science, Hiroshima University, Higashihiroshima, Japan
    2. Precursory Research for Embryonic Science and Technology, Japan Science and Technology Corporation, Higashihiroshima, Japan
    Current affiliation:
    1. Department of Biological Science, Faculty of Life and Environmental Science, Shimane University, Matsue, Japan
    Search for more papers by this author
  • Kazuko Koshiba,

    1. Yoshizato MophoMatrix Project, Exploratory Research for Advanced Technology, Japan Science and Technology Corporation, Higashihiroshima, Japan
    Current affiliation:
    1. Department of Molecular Biology, Nara Institute of Science and Technology, Ikoma, Japan
    Search for more papers by this author
  • Masanobu Obara,

    1. Developmental Biology Laboratory, Department of Biological Science, Graduate School of Science, Hiroshima University, Higashihiroshima, Japan
    2. Tissue Regeneration Project, Hiroshima Prefecture Collaboration of Regional Entities for the Advancement of Technological Excellence, Japan Science and Technology Corporation, Hiroshima Prefectural Institute of Science and Technology, Higashihiroshima, Japan
    Search for more papers by this author
  • Hiromu K. Mishima,

    1. Department of Ophthalmology and Visual Science, Division of Frontier Medical Science, Biomedical Research, Graduate School of Biomedical Science, Hiroshima University, Hiroshima, Japan
    Search for more papers by this author
  • Katsutoshi Yoshizato

    Corresponding author
    1. Developmental Biology Laboratory, Department of Biological Science, Graduate School of Science, Hiroshima University, Higashihiroshima, Japan
    2. Yoshizato MophoMatrix Project, Exploratory Research for Advanced Technology, Japan Science and Technology Corporation, Higashihiroshima, Japan
    3. Tissue Regeneration Project, Hiroshima Prefecture Collaboration of Regional Entities for the Advancement of Technological Excellence, Japan Science and Technology Corporation, Hiroshima Prefectural Institute of Science and Technology, Higashihiroshima, Japan
    • Developmental Biology Laboratory, Department of Biological Science, Graduate School of Science, Hiroshima University, Kagamiyama 1-3-1, Higashihiroshima, Hiroshima 739-8526, Japan
    Search for more papers by this author

Abstract

The process of regeneration of urodele limbs includes a drastic remodeling of extracellular matrices (ECMs) that is induced by matrix metalloproteinases (MMPs) and is thought to be one of the triggers of the regeneration. We studied this remodeling in limbs of Japanese newt, Cynops pyrrhogaster, by using five genes of newt MMPs (nMMPs) as probes: nMMP9, nMMP3/10-a, nMMP3/10-b, and nMMP13 that had been characterized previously, and nMMPe that was newly cloned in the present study. nMMPe was 502 amino acid residues long and showed a low homology to other known vertebrate MMPs. Reverse transcriptase-polymerase chain reactions analysis localized the transcript of nMMPe in the apical epidermal cap (AEC) and the non–blastemal wound epidermis but not in the blastemal mesenchyme or the normal epidermis. Northern blot analysis localized the transcripts of nMMP9, nMMP3/10-a, and nMMP13 in the bone of regenerating limbs, whereas those of nMMP3/10-b in AEC. mRNA in situ hybridization experiments identified the nMMP-expressing cells. nMMP9 gene was strongly expressed in chondrocytes of the cartilage of epiphysis. Of interest, basal cells of AEC, but not those of the normal skin, expressed nMMP3/10-b intensely. Immunohistochemical analysis showed that the nMMP9 proteins synthesized by chondrocytes were secreted and distributed widely in the basement membrane of bone and ECMs of the amputation plane. These nMMPs characterized in the present study might cooperatively work to remodel ECMs of regenerating limbs. Developmental Dynamics 226:366–376, 2003. © 2003 Wiley-Liss, Inc.

INTRODUCTION

Among vertebrates, the urodele uniquely shows an ability to regenerate their limbs when amputated (Wallace, 1981; Brockes, 1997). Several events consecutively take place after the amputation (Stocum, 1995). The wounded surface is covered within 2 days by the thin transparent wound epidermis composed of newly formed one to three layers of epidermal cells. As the regeneration proceeds, the wound epidermis develops in approximately 1 week after amputation into the thick apical epidermal cap (AEC) composed of multilayers of differentiated epidermal cells. On the other hand, drastic histolysis takes place in the mesenchyme near the amputation site. Parts of the muscle, bone, and connective tissue are decomposed and lose their well-organized structures. The dedifferentiation of cells of muscle and bone is thought to occur concomitantly with their histolysis, which leads to the formation of a mass of undifferentiated cells, the blastema (Wallace, 1981). Several lines of evidence indicate that AEC is necessary for the growth of blastema (Wolsky, 1988) and for the maintenance of blastemal cells in an undifferentiated state (Globus et al., 1980). Amputation of the limb triggers the breakdown of extracellular matrices (ECMs) at the wounded sites and the nearby tissues.

ECM exerts a profound influence on a variety of biologically important phenomena such as growth, differentiation, migration, apoptosis, and maturation of cells in animal tissues (Hay, 1993). ECM is a large family containing many types of molecules that have different functions (Talhouk et al., 1992). For example, collagens and fibronectins work as adhesive molecules for cell adhesion and migration, whereas tenascin and some proteoglycans work as antiadhesive molecules (Ayad et al., 1998). Thus, the control of synthesis and degradation of ECM is key for normal functions of cells. It has been reported that the composition and distribution of ECM change drastically and dynamically during the limb regeneration (Mailman and Dresden, 1979; Gulati et al., 1983; Onda et al., 1990). The basement membrane is not formed beneath the AEC until late bud stage (Neufeld and Day, 1996). Tenascin and fibronectin are intensely expressed in the blastema (Gulati et al., 1983; Onda et al., 1990). As the blastemal cells redifferentiate, the synthesis of laminin and type IV collagen reinitiates to reconstitute the basement membrane (Gulati et al., 1983), whereas that of tenascin and fibronectin becomes suppressed (Gulati et al., 1983; Onda et al., 1990). As one of the events associated with the remodeling of ECM, the regenerating limb begins to produce collagenolytic enzymes (Grillo et al., 1968).

Matrix metalloproteinases (MMPs) play a major role in degrading ECMs (Birkedal-Hansen, 1995). To date, more than 20 MMPs have been isolated in mammals that share several structural and functional properties (Nagase and Woessner, 1999). According to the substrate specificity and structural characteristics, they are grouped into collagenases (MMP1, 8, 13), gelatinases (MMP2, 9), stromelysins (MMP3, 7, 10, 11), and membrane-type MMPs (MT-MMP1, 2, 3, 4, 5). MMPs are thought to be involved in the tissue remodeling during newt limb regeneration (Grillo et al., 1968; Yang and Bryant, 1994). Actually, we previously cloned four newt cDNAs for MMPs from a cDNA library prepared from regenerating newt limb tissues: newt gelatinase b (nMMP9), stromelysin-a, -b (nMMP3/10-a, 3/10-b), and collagenase 3 (nMMP13; Miyazaki et al., 1996). Genes of these newt MMPs are not expressed in normal limb tissues. As expected, their expressions were markedly up-regulated in the regenerating tissues, suggesting their active participation in the regeneration-associated tissue remodeling (Miyazaki et al., 1996).

The present study was performed to address the following two issues. Previously, we isolated four nMMP genes and characterized them (Miyazaki et al., 1996). The present study was done to seek for additional new nMMP genes that are involved in the limb regeneration. As a result, we cloned and characterized a novel cDNA of newt MMP (nMMPe). Several biochemical and molecular biological studies demonstrated the presence of distinct types of MMPs of urodele limbs and their up-regulation during limb regeneration (Grillo et al., 1968; Yang and Bryant, 1994; Miyazaki et al., 1996). However, there have been very few studies that identified the MMP-producing axolotl cells in the process of regeneration (Yang et al., 1999). In the present study, we made detailed studies on the MMP-expressing cells in regenerating newt limbs focussing on nMMP9 and nMMP3/10-b. mRNA in situ hybridization experiments identified chondrocytes and epidermal cells as the cells in the regenerating limb that express nMMP9 and nMMP3/10-b, respectively. Furthermore, immunohistochemical analysis using specific antibodies against recombinant nMMP9 proteins revealed its distribution in the cartilage and the nearby connective tissues.

RESULTS

Cloning of a Novel Newt MMP Gene

We previously cloned four newt MMP genes (nMMP9, nMMP3/10-a, nMMP3/10-b, nMMP13; Miyazaki et al., 1996), which did not contain the nMMP1 gene. The present study tried to clone cDNAs of newt MMP1 by reverse transcriptase-polymerase chain reaction (RT-PCR) using total RNA isolated from 5-day blastemas as templates and the same degenerate primers as before (Miyazaki et al., 1996). The PCR products were electrophoresed, which gave a band with approximately 250 bp. cDNAs in this band were subcloned, which gave more than 200 clones. Approximately 120 clones were randomly selected from them and sequenced. Five and four clones were fragments of nMMP9 and nMMP13, respectively. Six clones were 262-bp long, and their sequences were identical, thus, these clones were dubbed clone 1-24-21. The DNA sequences of the remaining clones did not show any homology to MMPs. These sequencing data led us to characterize further clone 1-24-21. The full-length cDNA of clone 1-24-21 was obtained by nested 5′- and 3′–rapid amplifications of cDNA end (RACE; Fig. 1A). As described below, the cloned gene was expressed in blastemal and non–blastemal wound epidermis. Northern blot analysis was carried out by using the cloned full-length cDNA against total RNA collected from the wound epidermis of newt limbs (Fig. 1B). The single transcript was detected below the newt 18 S ribosomal RNA.

Figure 1.

Schematic representation of the strategy for cloning cDNA of the full-length coding region of the novel newt matrix metalloproteinase (MMP) and its Northern hybridization. A: 5′-rapid amplification of cDNA ends (RACE) was performed by using first-strand cDNA prepared from poly (A)+ RNA as templates, and primers of MMP-AS31 (antisense) and sense primers included in a kit for an initial polymerase chain reaction (PCR), and primers of MMP-GSP1 (antisense) and AP2 (sense) for nested PCR. Similarly 3′-RACE was performed using primers of MMP-S31 (sense) and supplied antisense primers for an initial PCR and primers of GSP2 (sense) and supplied antisense primers for nested PCR. The product of 5′-RACE and 3′-RACE was approximately 600 bp and 1.2 kb, respectively. A region containing the open reading frame was obtained by PCR using the above cDNA as templates and primers of MMP-TAIL and MMP-HEAD. Arrows indicate the position and the direction of the indicated primers. The horizontal line at the bottom represents a measure of the length in kilobases of cDNA. B:Northern blot analysis of mRNAs of the novel newt MMP. Total RNA was extracted from non–blastemal wound epidermis at 5 days after the skin removal, 20 μg of which was electrophoresed and subjected to Northern blot analysis. Closed circle and square indicate the position of 28 S and 18 S ribosomal RNA, respectively.

The deduced amino acid sequence of the RACE product is shown in Figure 2. The cloned gene had amino acid sequences characteristic of an MMP. The amino acid sequences showed basic common structures of MMPs: a signal peptide, a propeptide, a catalytic domain, a hinge domain, and a hemopexin-like domain. Collagenases, stromelysins, and metalloelastase are the prototype MMPs. Gelatinase has the fibronectin type II repeat in addition to this prototype that works to bind to gelatin. Membrane-type MMPs have the transmembrane domain and cytoplasmic domain in addition to the prototype structure. The matrilysin-type MMPs (matrilysin and endometase) lack the hinge and hemopexin-like domain of the prototype. According to these criteria, the cloned newt MMP is grouped into the prototype MMP. MMP11 is included in the prototype MMP and unique in that it contains the 11-residues insert in the propeptide that harbors a RXKR recognition motif for furin. The cloned gene was not a homologue of MMP11 because it did not have this motif. A homology search of the sequence data bank, GenBank revealed that the deduced amino acid sequence of the newt MMP gene showed similar levels of homology to different types of MMPs, the highest homology being 26.7% to human MMP13. The three amino acid residues around the zinc-binding motif are well conserved in all mammalian collagenases (Tyr-214, Asp-235, and Gly-237 in the case of human MMP13; Freije et al., 1994). However, the newt MMP did not contain two of them: aspartic acid was conserved, but tyrosin and glycine were replaced with isoleucine and serine, respectively. nMMP3/10-a and -b show more than 50% homology to both human MMP3 and MMP10, whereas the cloned MMP showed a 24.9% and 23.3% homology to each of them, respectively. Therefore, the cloned gene could not be given the type number of MMPs according to Yang and Kurkinen (1998) from the currently available homology data. As shown below, this gene is expressed in the wound epidermis. These comparisons in the structure of the cloned gene with the known MMP genes and its expression specificity led us to dub the protein encoded by the gene nMMPe. MMPs are distinguished from other metalloproteinases by the unique sequence motifs called the cystein switch and the zinc-binding domain. The cystein switch PRCG[V/N]PD is located in the propeptide. The zinc-binding domain HEXGHXXGXXH is located in the catalytic domain (van Wart and Birkedal-Hansen, 1990). Two other cysteins are also highly conserved in the hemopexin-like domain. These characteristic motifs are highly conserved among the known MMPs and also in nMMPe (Fig. 2). The sequence of methionine turn ALMYP was seen in nMMPe. nMMP3/10-a, nMMP3/10-b, and nMMP9 have 9- to 18-residue-long threonine-rich insertions, a unique feature of newt MMPs (Miyazaki et al., 1996). This insertion may be unique among urodele MMPs, because axolotl MMP9 has a similar threonine-rich insertion (Yang et al., 1999). Similarly, nMMPe had a 42-residue-long threonine-rich insertion. nMMPe represented a MMP with 502 amino acid residues. Its molecular weight was approximately 55.5K in the latent form and 44K in the activated one.

Figure 2.

Comparison of amino acid sequences of the cloned newt matrix metalloproteinase (nMMP) with those of the known MMPs. Amino acid sequences of 5 MMPs, the cloned nMMP (nMMPe), human MMP13 (hMMP13), nMMP13, nMMP3/10-a, and nMMP3/10-b were aligned by using DNASIS-Mac software. Dots represent amino acid residues identical to the nMMP cloned in this study (nMMPe). Dashes indicate that the gene lacks the corresponding residue. nMMPe includes all of the characteristic MMP domains. Putative signal peptides were predicted using an on-line server (Nielsen et al., 1997) and are marked above the region. The regions of cystein switch, calcium-binding domain, zinc-binding domain, and methionine turn are boxed. The region of a threonine-rich insertion is shadowed for nMMPe sequences. Three conserved cysteins are marked with asterisks. Nucleotide sequence of nMMPe has been deposited in the DDBJ under the accession no. AB092571.

Expression of nMMPs

The level and the site of the expression of nMMPe mRNA were determined by RT-PCR using total RNA extracted from regeneration newt forelimb tissues. First, we examined its expression level in the blastema. RNAs were extracted from blastemas including the stump collected at day 1 when the amputated surface was covered by the thin wound epidermis through 30 days after amputation when the regenerates were at late bud stage (Fig. 3A). The significant expression began at 2 days, then gradually increased, reaching a plateau around 1 week, and decreased gradually thereafter. Second, the site of nMMPe mRNA expression was examined for regenerating limb tissues collected at 2, 4, and 7 days after amputation. The tissues were separated into blastemas and stumps, the latter being further separated into tissues of skin, bone, and muscle (Fig. 3B). Amplified products of nMMPe cDNA were obtained in the blastemas, but not any other tissues examined to date. We asked whether nMMPe is expressed in the AEC or the mesenchyme of the blastema. Blastemas at the late bud stage (26 days after amputation) were collected, because blastemas before this stage were not practically separable into the two tissues. Total RNA was extracted from the separated AEC and the mesenchymal tissue. The RT-PCR product was obtained only in the AEC (Fig. 3C). We could not perform this analysis on blastemas at earlier stages as described above. Thus, the possibility remains that nMMPe might be expressed in the mesenchyme at earlier stages of regeneration. The cut surface was covered by the wound epidermis in the early phase of regeneration, suggesting that nMMPe is expressed also in the non–blastemal wound epidermis. The non–blastemal wound epidermis was made on a part of the forearm skin with a razor blade in a small square form whose area was similar to the cut surface of amputation of made in the forearm in regeneration experiments. The expression level in the wound epidermis was determined as above during the regeneration up to 20 days after amputation (Fig. 3D). As predicted, the non–blastemal wound epidermis expressed nMMPe gene, the pattern and intensity of its expression being similar to those in the blastema with the peak at 7 days after amputation. nMMPe mRNA was hardly detectable in the non–blastemal wound epidermis at 20 days while it was in the blastema. RT-PCR products from RNA isolated from the blastema and the non–blastemal wound epidermis were cloned and randomly sequenced, which showed that the products of the two tissues were identical to nMMPe.

Figure 3.

Expression of the gene for the newt matrix metalloproteinase cloned in this study (nMMPe). The level of the expression of nMMPe gene was determined by reverse transcriptase-polymerase chain reaction during regeneration of forelimbs. A: Expression in blastemas. Total RNA was extracted from blastemas collected at the indicated days after amputation. B: Tissue specificity of the expression. Regenerating limbs were collected and separated into blastema, skin, bone, and muscle for the total RNA extraction at the indicated days after amputation. C: Comparison of the expression between the apical epidermal cap (AEC) and mesenchymal tissues of the blastema at late bud stage. Blastemas were collected and separated to AEC and mesenchymal tissues for total RNA extraction. NS, normal skin; AEC, AEC at late bud stage; MT, mesenchymal tissues at late bud stage. D: Comparison of the expression between the blastema and the non–blastemal wound epidermis (WE). Tissues were collected for total RNA extraction at the indicated days after amputation or skin removal. NS, normal skin. EF-1α in A through D indicates the elongation factor-1α internal control used to normalize the expression level of nMMPe.

We previously reported that genes of nMMP9, nMMP3/10-a, and nMMP3/10-b are not expressed in normal limbs, but expressed at least at 2 days after the amputation (Miyazaki et al., 1996). Northern blot analysis in the present study showed that the expression of nMMP13 gene was also induced at least at 2 days after the amputation (data not shown). To identify the cells that responded to the amputation and began to express these genes, limbs at 6 days after amputation were separated into bone, muscle, and skin, each of which was subjected to Northern blotting. The regenerates at day 6 were used for this analysis, because we previously showed that the expression level of nMMPs reached to the plateau at day 5 (Miyazaki et al., 1996). The bone fraction was richest in transcripts of nMMP9, 3/10-a, and 13, whereas those of nMMP3/10-b were strongly expressed in the skin rather than in muscle and bone (Fig. 4). We previously showed that the former three genes are expressed in the early phase of regeneration, whereas the latter continued to be expressed through the palette stage. There seemed to be cell-type and regeneration-phase specificity in the expression of MMP genes. A 90-kDa protein with gelatinolytic activity was detected at the wounded region of the flank of axolotls (Yang and Bryant, 1994), suggesting that just a wounding induces the expression of MMP genes. We also carried out Northern blot analysis and in situ hybridization by using nMMP9 cDNAs as a probe for the skin whose epidermis had been peeled off 1 week before the analysis. nMMP9 transcripts were undetectable in our case (data not shown), suggesting that nMMP9 might respond to regeneration-associated wounding but not to the mere wounding. However, the report by Carinato et al. (2000) should be noted that the wounding made on the epidermis of Xenopus larvae induced the expression of MMP9 mRNA during the first 2 days. This early expression ceased thereafter. Thus, it is possible that the expression is induced at an early phase of epidermal wounding also in nMMP9 mRNA. In the present study, we did not test this possibility.

Figure 4.

Northern blots of newt matrix metalloproteinases (nMMPs) for tissues of regenerating limbs. Newt limbs at 6 days after amputation were microdissected observing through a microscope into the skin (S), muscle (M), and bone (B). Total RNA was extracted and 10 μg were subjected to Northern blot analysis. The filters were stained for rRNA with methylene blue to confirm equal loading. The RNAs on the filters were hybridized with the radiolabeled fragments of cDNA indicated.

In Situ Hybridization for Genes of nMMP9 and nMMP3/10-b

Forelimbs were amputated and in situ hybridization was performed for mRNA of nMMP9 at 3 days after amputation (Fig. 5). Hybridization signals were observed neither with the sense probes in the regenerating limbs (Fig. 5A) or with the antisense probes in normal limbs (data not shown). The section shown in Figure 5B was cut obliquely for proximodistal axis, and it illustrates a circular profile for the epiphysis of ulna. Chondrocytes in the cartilage of epiphysis expressed a high level of nMMP9 mRNA, but the other cells such as cells of muscle, wound epidermis, stump epidermis, and interstitial connective tissues did not express it at a significant level (Fig. 5B). Other bone-containing sections showed that osteoblasts or osteoclasts also did not express the gene (data not shown). In Figure 5C, the nMMP9 expression was restricted along the proximodistal axis to the epiphyses and was not expressed in the perichondrium. Figure 5C shows the in situ hybridization for a region containing undamaged carpi apart from the amputation site that was near the photo's left edge. Of interest, chondrocytes in this region also expressed the gene, which indicates that the cells in the undamaged region respond to the wounding and express nMMP9 gene. Similar results were obtained for limbs at 6 days after amputation (data not shown).

Figure 5.

Localized expression of newt matrix metalloproteinase (nMMP) -9 genes in chondrocytes. Limbs at 3 days after amputation were collected, fixed, decalcified, frozen, and sectioned 10 μm thick. The sections were hybridized with digoxigenin-labeled sense (A) or antisense probes (B,C) of nMMP9 cDNAs. The arrow represents the amputation site. A,B: A region proximally adjacent to the amputation site. C: A region apart from the amputation site. Carpi and distal end of ulna are shown. The amputation site was approximately 2.8 mm leftward apart from the distal end of ulna and is indicated by a solid thin line at the left side out of the photo. WE, wound epidermis; M, muscle; C, carpus; U, ulna. Scale bars = 300 μm in A,B, and 600 μm in C.

We previously reported that nMMP3/10-b mRNA was expressed from 2 days through pallet stage in the regenerating limbs, whereas those of nMMP9 and nMMP3/10-a were observed from 2 days and declined after 15 days (Miyazaki et al., 1996). Blastemas including AEC at mid bud stage (21 days) were collected and nMMP mRNA expressions were analyzed by Northern blot analysis (Fig. 6A). nMMP3/10-b gene was uniquely expressed in the blastema at a high level. The expression levels of other MMPs were comparatively low. Thus, we made in situ hybridization experiments for nMMP3/10-b mRNA to identify blastemal cells that express this gene at mid bud stage (Fig. 6B–D). At this stage, AEC consisted of four-layered epidermal cells. The cells in the basal layer were found to express the gene (Fig. 6C). When the border of the trunk epidermis and the AEC was viewed (Fig. 6D), the transcripts were seen in the AEC, but not in the secretory gland-containing trunk normal epidermis. This provided an additional piece of evidence supporting the unique characteristics of AEC. nMMP3/10-b protein might be secreted from the epidermis into the underlying connective tissues such as dermis and blastema and plays an important role in remodeling the newly synthesized basement membrane and suppressing the blastemal formation.

Figure 6.

Unique expression of newt matrix metalloproteinase (nMMP) 3/10-b in AEC. A: Blastemas were collected from limbs at 21 days after amputation (mid bud stage) and subjected to Northern blot analysis as in Figure 4. nMMP3/10-b mRNA expression was outstanding in blastemas among the nMMPs tested. Square indicates the expected signal at 4.0 kb for nMMP9, 1.8 kb for nMMP3/10-a, 1.7 kb for nMMP3/10-b, and 4.4 kb for nMMP13. B–D: Regenerates at the mid bud stage were obtained, sectioned as in Figure 5, and hybridized with digoxigenin-labeled probes of nMMP3/10-b cDNAs. B,C: Photomicrographs of the most distal region of a blastema (BL). D: Photomicrograph of the boundary region between the apical epidermal cap (AEC) and normal skin. The arrow points to the amputation site. To the right of the arrow is the regenerate, and to the left part is the stump. A secretory gland (S) is seen in the stump side. The blastema consisted of four-layered epidermal cells and the mesenchyme contained undifferentiated cells. Sections were hybridized with sense probes (B) and antisense probes (C,D). The R in B indicates a nonsignal reaction product in in situ hybridization. In C and D, meaningful signals are purple colored, some representatives of which are indicated by arrowheads. Pigments are brown-colored and indicated by P. NE, normal epidermis. Scale bar = 300 μm.

Localization of nMMP9 Protein in Cartilage and Connective Tissues of Regenerating Limbs

The unique localization of nMMP9 transcripts in the cartilage cells led us to investigate the location of the nMMP9 enzyme. Recombinant nMMP9 protein was expressed in Escherichia coli (E. coli) and used as an antigen to raise its polyclonal antibodies in mice. Immunoblot analysis was performed by using the antibody on unamputated control limbs and regenerating limbs at 6 and 13 days after amputation. Regenerating but not normal limbs contained a 74-kDa protein reactive to the antibody (Fig. 7A). The size of the latent and the activated form of axolotl MMP9 is 90K and 72K, respectively (Yang and Bryant, 1994). Therefore, this 74-kDa protein was concluded as the activated nMMP9. A band corresponding to the latent form was not detected.

Figure 7.

Immunoblotting and immunohistochemistry of newt matrix metalloproteinase (nMMP) -9. A: Immunoblotting of nMMP9 of normal and regenerating newt limbs. Forelimbs were amputated and their regenerates were separated at 3 mm from the amputation plane at 6 and 13 days after amputation together with the tissue (0), which was separated from the corresponding region of normal limbs. The tissues were solubilized in sodium dodecyl sulfate sample buffer, and their proteins (100 μg) were subjected to Western blot analysis. Proteins on the transfer filter were visualized with alkaline phosphatase–conjugated second antibodies. The arabic numerals at the left side are molecular masses in K determined by the migration of standard proteins. B,D,E: Immunohistochemistry of nMMP9 in regenerating newt limbs. Tissues of limbs at 6 days after amputation were removed from a region near the amputation site and were subjected to immunohistochemistry by using antibodies against recombinant nMMP9 proteins as the first antibody and fluorescein isothiocyanate–labeled anti-rabbit immunoglobulin G antibodies as the second. A section beneath the amputation plane is presented in B. Stains are observed in the entire vision field except the epidermis. Especially, signals around chondrocytes and of the basement membrane on the surface of cartilage are intense. The cartilage region of B is enlarged in D. A stain with nonimmune sera is shown in E. C: Hoechst staining of a regenerating limb. A region corresponding to that shown in B was obtained on a section prepared from a blastema at 5 days after amputation and was subjected to Hoechst staining. E, epidermis; S, secretory gland; Mu, muscle; BM, basement membrane; Ca, cartilage. Scale bars = 50 μm in B–E.

Immunohistochemistry was done on the sections prepared from a region near the amputation site of 6-day-regenerating limbs by using the same antibody to locate nMMP9 protein (Fig. 7B,D). It was found that the cartilage of epiphysis and connective tissues were intensely stained, whereas the epidermis was not positive (Fig. 7B). The positive stains on the basement membrane of the cartilage and subepidermal connective tissues surrounding muscles were meaningful signals, because the nonimmune sera did not produced such stains (Fig. 7E). A higher magnification clearly showed that chondrocytes secreted nMMP9 around them (Fig. 7D). It seemed that nMMP9 proteins were synthesized by chondrocytes and might degrade ECMs surrounding them.

DISCUSSION

The present study added a new member to the MMP family: the nMMPe gene whose expression is activated in AEC of blastemas and the wound epidermis. Previously cloned nMMPs are all homologous to the corresponding mammalian MMPs to a significant extent. By contrast, nMMPe showed a low homology to the known MMPs of both mammalian and amphibian origins. CMMP and XMMP were cloned from fibroblasts of chicken embryos and Xenopus laevis embryos, respectively, and showed a low homology to the known mammalian MMPs (Yang et al., 1997; Yang and Kurkinen, 1998). nMMPe also showed a low homology to both CMMP and XMMP. In the present study, we did not determine whether nMMPe protein actually exhibits MMP activity.

The nMMPe gene started to be activated in AEC from 2 days after amputation and gradually increased until 7 days, then gradually decreased thereafter until 30 days. The activation was also observed in the non–blastemal wound epidermis that covered the wound on forearms. In situ hybridization demonstrated that nMMP3/10-b gene was also expressed in AEC, exclusively in its basal cells. In the present study, we could not identify the type of epidermal cells in AEC that expressed nMMPe, because of unsuccessful trials of in situ hybridization for this gene. In situ hybridization of nMMP3/10-a and nMMP13 also did not produce convincing results in the present study, despite repeated trials.

Previous Northern blot analysis showed that nMMP9 gene became detectable at 2 days after amputation (Miyazaki et al., 1996). The present mRNA in situ hybridization experiment indicates that this expression is in the cartilage fractions. A previous study demonstrated that axolotl MMP9 was expressed biphasically (Yang et al., 1999). The first phase began 2 hr after amputation and continued for 2 days. The expression was confined to the wound epidermis. The second phase began a few days later when a small blastema was formed and the expression was in the mesenchyme, localized to cells around the cut skeletal elements. Although we did not examine the expression of nMMP9 before 2 days after amputation, it is likely that our observation on the expression of nMMP9 gene in the cartilage corresponds to the expression of the second phase in the above-cited study.

Immunohistochemistry of nMMP9 showed that all the tissues except epidermis were immunoreactive. The negative staining with nonimmune sera did not produce such stains. Thus, it can be said that the observed staining pattern is antigen-specific and of biological significance. nMMP9 mRNA revealed by in situ hybridization was observed only in chondrocytes, but the immunohistochemical signals were observed not only in chondrocytes but also on muscles and connective tissues. This apparent contradiction is explainable by the interpretation that the enzyme produced by the chondrocytes spread out to everywhere except epidermis. Thus, it appears that cartilage tissues play an important role(s) in the tissue remodeling during newt limb regeneration. Western blotting of nMMP9 detected only one band with 74 kDa, the activated form of nMMP9. The latent form with approximately 90 kDa was not detected in the present study. It is most likely that the latent form was processed to the activated form in the tissue or during the tissue preparation for Western blotting. Oofusa and Yoshizato (1991) made Western blots of MMP1 on tadpole tail tissues, which detected its activated, but not the latent, form, as in the present study.

The expression site of nMMP3/10-b was unique. The basal epidermal cells of AEC were the cells that express this gene. It is noteworthy that the basal epidermal cells of the stump adjacent to the AEC never switch this gene on. Several lines of evidence have been reported that the epidermis of AEC shows characteristics different from those of the normal epidermis (Globus et al., 1980; Wolsky, 1988). We propose that nMMP3/10-b gene is an additional useful marker to characterize AEC. The biological significance of the expression of this gene in AEC remains unknown. One intriguing suggestion is that nMMP3/10-b is responsible for the absence of morphologically discernible structures of the basement membrane in AEC (Neufeld and Day, 1996). The nMMP3/10-b enzymes synthesized by the epidermal basal cells could be secreted beneath the AEC and suppress the formation of the basement membrane. This suggestion might be supported by the substrate specificity of mammalian MMP3 and MMP10. These enzymes attack proteoglycan, fibronectin, laminin, and type IV collagen, which are major components of the basement membrane (Woessner, 1991). Furthermore, structurally incomplete basal laminae emerge at late bud stage, which are composed of loose meshworks, and the complete basement membrane reforms after the palette stage (Neufeld and Day, 1996). The period of expression of nMMP3/10-b gene is uniquely prolonged to the palette stage (Miyazaki et al., 1996), which is coincident with that of the absence of the structurally complete basement membrane. It awaits a future study to understand how various types of nMMPs reported in the present study concertedly work in degrading ECMs of regenerating newt limb.

EXPERIMENTAL PROCEDURES

Animals and Their Operations

Adult Japanese newts, Cynops pyrrhogaster, were obtained from a local supplier. Animals were anesthetized in a 0.1% solution of MS 222 (Sigma, St. Louis, MO), and their forelimbs were amputated two-thirds distal to the elbow bilaterally with a razor blade. Newts were fed three times a week and maintained in water at 24°C. RT-PCR analysis was performed for blastema, skin, muscles, and bones separated from regenerating limbs and for AEC and mesenchymal cells isolated from blastemas at medium bud stage. The non–blastemal wound epidermis (3 × 3 mm) was obtained from the forelimb.

Isolation of RNA From Regenerating Newt Limbs

Total RNA for RT-PCR and poly (A)+ RNA for Northern analysis of nMMPe were isolated with ISOGEN (Nippon Gene, Tokyo, Japan) and with a Quickprep Micro mRNA purification kit (Amersham Pharmacia Biotech, Uppsala, Sweden), respectively, according to the manufacturer's instructions. Total RNA for Northern analysis of nMMP9, nMMP3/10-a, nMMP3/10-b, and nMMP13 was prepared from frozen tissues of newt limbs by using a Trizol reagent (GIBCO BRL, Gaithersburg, MD) according to the manufacturer's instruction.

Cloning of MMP cDNA

cDNAs were synthesized from the mixture of the same amount of total RNA of blastema 2, 5, 10, 15, and 20 days after amputation with random 9-mers. PCR was carried out by using the above cDNAs as temples and two degenerate primers (a sense primer MMP-S3 and an antisense primer MMP-AS3; Miyazaki et al., 1996), with 30 cycles of denaturation (94°C, 20 sec), annealing (40°C, 30 sec), and extension (74°C, 30 sec). 5′- and 3′-RACE (Frohman et al., 1988) was performed by using cDNAs as templates synthesized from poly (A)+ RNA of blastemas obtained at 5 days after amputation with a marathon cDNA amplification kit (Clontech, Palo Alto, CA) and with a RNA PCR kit (Takara, Tokyo, Japan), respectively. 5′-RACE was made as follows. An initial PCR was carried out with 35 cycles of denaturation (94°C, 30 sec), annealing (54°C, 30 sec), and extension (72°C, 1 min) by using MMP-AS31 (5′ GGATAAGGATCTTGGTGATC 3′) and the supplied anchored primer AP1. Nested PCR was made with 25 cycles of denaturation (94°C, 30 sec), annealing (60°C, 30 sec), and extension (72°C, 1 min) by using MMP-GSP1 (5′ TAAGAGGACTTGCGCTGCTCCAGACACC 3′) and the supplied anchored primer AP2. 3′-RACE was made as follows. An initial PCR was carried out with 35 cycles of denaturation (94°C, 30 sec), annealing (54°C, 30 sec), and extension (72°C, 1 min) by using primers of MMP-S31 (5′ GGACACCAATTTAAAATGGG 3′) and the supplied anchored primer M13 primer M4. Nested PCR was carried out with 25 cycles of denaturation (94°C, 30 sec), annealing (60°C, 30 sec), and extension (72°C, 1 min) by using primers of MMP-GSP2 (5′ GGCTATAGCGAGTGCCTTAGGTGTCTGG 3′) and the supplied M13 primer M4. The full-length coding region was obtained by PCR with 30 cycles of denaturation (94°C, 30 sec), annealing (55°C, 30 sec), and extension (72°C, 30 sec). Its reaction mixtures contained cDNAs as templates synthesized from poly (A)+ RNA of blastemas obtained at 5 days after amputation, primers of MMP-HEAD (5′ AGTAACCCAAGTGAGATCCG 3′) and MMP-TAIL (5′ CGTATTGATACATTCTTGGTG 3′), and DNA polymerase (LA Taq, Takara).

RT-PCR

Total RNAs were extracted from regenerating limb tissues, treated by Rnase-free DNase I (GIBCO BRL), and used as templates of RT. RT was performed with AMV RTase XL and oligo-dT primers included in an RNA PCR kit (Takara). PCR for nMMPe and EF-1α was performed with LA Taq according to the manufacturer's instruction. PCR for nMMPe was carried out with 30 cycles of denaturation (94°C, 30 sec), annealing (55°C, 30 sec), and extension (72°C, 30 sec) by using MMP-S31 and MMP-AS31 as primers. PCR for EF-1α was done with 23 cycles of denaturation (94°C, 30 sec), annealing (55°C, 30 sec), and extension (72°C, 30 sec) by using 5′ ATCGACAAGAGAACCATCGA 3′ and 5′ CGGTGATCATGTTCTTGATG 3′ as primers (Takabatake et al., 1996).

Northern Hybridization

Total RNA for Northern hybridization of nMMPe was extracted from tissues of non–blastemal wound epidermis that was made by peeling off a part of the forearm skin with a razor blade in a small square form. The wound epidermis was collected at 5 days after the operation. Total RNA was separated by electrophoresis on a 1.0% agarose gel containing formaldehyde and transferred to nylon membranes (Hybond N+, Amersham Pharmacia Biotech). The digoxigenin (DIG)-labeled RNA probes were prepared according to manufacturer's instruction (DIG RNA labeling kit, Roche, Basel, Switzerland). Hybridization and signal detection were performed as described by Shimizu-Nishikawa et al. (1999). Northern hybridization for nMMP9, nMMP3/10-a, nMMP3/10-b, and nMMP13 was performed as follows. RNA was electrophoresed on 1.4% agarose formaldehyde gels in denaturing conditions and transferred to nylon membranes (GeenScreen Plus; Du Pont, Wilmington, DE). The blots were stained for rRNA with 0.04% methylene blue in 0.5 M sodium acetate, pH 5.2, to assess RNA integrity and equal loading. RNAs on the membranes were hybridized with cDNAs of newt MMPs that had been labeled with 32P-dCTP by random priming by using an oligolabeling kit (Pharmacia, Uppsala, Sweden). Prehybridization, hybridization, and posthybridization were carried out according to Miyazaki et al. (1996).

In Situ Hybridization

Fragments of cDNAs of nMMP9 (nucleotide 1274-2214), nMMP3/10-a (nucleotide 1238-1780), nMMP3/10-b (nucleotide 837-1425), and nMMP13 (nucleotide 794-1662) were inserted in pBSII plasmid vectors (Stratagene, La Jolla, CA). These vectors were linearized and transcribed by T3 or T7 RNA polymerase by using a DIG-RNA labeling kit (Roche) to obtain sense and antisense DIG-labeled RNA probes. Normal or regenerating newt limbs were excised and fixed overnight in 10% formaldehyde in MEM buffer at 4°C, and decalcified in 0.5 M ethylenediaminetetraacetic acid (EDTA; pH 7.0) for 1 week at room temperature. After removing lipids by incubating in serially diluted ethanol (25, 50, 75,100, 75, 50, and 25%), the tissues were embedded in Tissue Tek OCT Compound (Miles, Elkhart, IN) and 10-μm-thick cryosections were spread on poly-L-lysine–coated glass slides. Hybridization was carried out according to Wilkinson (1992). In brief, the sections were pretreated with 1 μg/ml proteinase K (Roche), acetylated with 0.25% acetic anhydride in 0.1% triethanolamine buffer, and fixed in 4% paraformaldehyde. The hybridization was carried out with either sense or antisense riboprobes overnight at 50°C in hybridization solution consisting of 50% formamide, 1× Denhardt's solution, 10% dextran, 0.3 M NaCl, 2.5 mM EDTA, 1 mg/ml E. coli tRNA, and 20 mM Tris-HCl, pH 8.0. Slides were washed in 50% formamide and 2× standard saline citrate (SSC) at 45°C for 1 hr, and treated with 20 mg/ml RNase A for 30 min at 37°C. The slides were further rinsed with 50% formamide and 2× SSC for 1 hr at 45°C, with 50% formamide/1× SSC for 1 hr at 45°C, and with the same buffer for 30 min at room temperature. The sections were rinsed three times with DIG buffer 1 (150 mM NaCl, 0.1% Tween 20, and 100 mM Tris-HCl, pH 8.0) for 15 min at room temperature and blocked with DIG buffer I containing 0.5% blocking reagent (Roche) for 30 min at room temperature. The sections were then incubated with alkaline phosphatase-labeled anti-DIG antibodies at room temperature for 1 hr, rinsed with DIG buffer 1 for 15 min three times, and with DIG buffer 2 (100 mM NaCl, 50 mM MgC12, and 100 mM Tris-HCl, pH 9.5) for 5 min. Finally, the sections were sequentially incubated with BM purple at 4°C overnight and for additional 6 hr at room temperature.

Expression of Recombinant nMMP9 in E. coli

A 1.8-kb fragment of cDNA of nMMP9 containing the entire coding sequence was produced by digesting with PstI/XbaI (nucleotide 291-2214), and ligated to an expression vector pMal-c2 by using MBP-expression systems (New England Biolab, Beverly, MA). The resulting plasmid was transferred into E. coli strain JM109 (Toyobo, Tokyo, Japan). The transformants were grown in LB broth containing 50 μg/ml ampicillin at 37°C until the turbidity of the medium reached 0.5. Then isopropyl-1-thio-β-D-galactopyranoside was added to the medium until a final concentration of 0.3 mM. The cells were incubated for additional 3 hr, collected by centrifugation, resuspended in buffer of 0.2 M NaCl, 1 mM EDTA, 1 mM sodium azide, and 20 mM Tris-HCl, pH 7.4, lysed by sonication, and centrifuged at 8,000 rpm for 15 min at 4°C. The recombinant protein was purified from the supernatant by an affinity column of amylose resin (New England Biolab) according to the manufacturer's protocols.

Antiserum Production

Recombinant nMMP9 proteins (1.1 μg) were dissolved in 0.5 ml of solution containing 100 mM NaCl, 10 mM maltose, and 20 mM Tris-HCl, pH 7.5, homogenized with an equal volume of complete Freund's adjuvant, and injected intraperitoneally into Balb/c mice. Immunizations were performed five times every 2 weeks with the same preparation using incomplete Freund's adjuvant instead of complete Freund's adjuvant. One week after the last injection, antisera were collected.

Immunoblotting

Normal and regenerating limbs at 6 and 13 days after amputation were homogenized with sodium dodecyl sulfate (SDS) sample buffer under the reducing conditions. The homogenate was centrifuged at 10,000 × g for 15 min at room temperature. The supernatant was determined for its protein concentration by a Protein Assay kit (Bio-Rad, Hercules, CA) with immunoglobulin G (IgG) as a standard protein. Aliquots containing 100 μg of proteins were electrophoresed on gels of 0.2% SDS and polyacrylamide with a 4–20% gradient (Dai-ichi Kagaku, Tokyo, Japan) and transferred onto nitrocellulose membranes. The membranes were blotted with polyclonal anti-nMMP9 antisera or normal mouse sera, washed extensively with phosphate buffered saline (PBS) containing 0.5% Tween 20, and incubated with anti-mouse antibodies labeled with alkaline phosphatase. Immunoblots were visualized by using an alkaline phosphatase substrate kit II (Vectastain, Vector Laboratories, Burlingame, CA).

Immunohistochemistry

Normal and 6-day-regenerating limbs were fixed, decalcified with 10% trichloroacetic acid, infused with successively increasing concentrations of sucrose (10, 15, and 25%) for more than 4 hr at each time, embedded in Tissue Tek OCT Compound, and frozen. Sections were cut 10 μm thick as in the In Situ Hybridization section described above. Nonspecific protein binding sites were blocked with10% bovine serum albumin (BSA) in PBS for 30 min. Sections were incubated with anti-nMMP9 mouse antisera for 1 hr and rinsed three times with PBS containing 0.5% Tween 20 for 10 min. Tissue sections were further incubated with fluorescein isothiocyanate–conjugated goat anti-mouse IgG in PBS containing 10% BSA. The sections were rinsed as described above and mounted with Mobiol (Harlow and Lane, 1988). When necessary in immunohistochemical studies, nuclei were stained with Hoechst 33342 to show the location of cells.

Acknowledgements

We thank Ms. H. Kawabata-Fukui, Ms. K. Sekiguchi, Ms. H. Kohno, and Ms. A. Kamada for their excellent technical assistance. We also thank Drs. N. Noro and K. Suzuki for helpful discussions and technical advice.

Ancillary