-
Proc. Natl. Acad. Sci. USAVol. 93, pp. 6819-6824, June
1996Developmental Biology
Cloning and characterization of cDNAs for
matrixmetalloproteinases of regenerating newt limbsKoYoMI
MIYAZAKI*, KOHJI UCHIYAMA*t, YUTAKA IMOKAWA*, AND KATSUTOSHI
YOSHIZATOt§¶*Yoshizato MorphoMatrix Project, Exploratory Research
for Advanced Technology, Research Development Corporation of Japan,
Tohkohdai, Tsukuba,Ibaraki 305, Japan, and tHiroshima-Techno-Plaza
and §Department of Biological Science, Faculty of Science,
Hiroshima University, Kagamiyama,Higashihiroshima, Hiroshima 739,
Japan
Communicated by Jerome Gross, Harvard Medical School,
Charlestown, MA, March 13, 1996 (received for review January 22,
1996)
ABSTRACT Matrix metalloproteinases (MMPs) of re-generating
urodele limbs have been suggested to play crucialroles in the
process of the dedifferentiation of cells in thedamaged tissues and
the ensuing blastema formation becausethe activation of MMPs is an
early and conspicuous eventoccurring in the amputated limb. MMP
cDNAs were cloned asproducts of the reverse transcription-PCR from
cDNA librar-ies of newt limbs, and their structures were
characterized.Three cDNAs encoding newt MMPs (2D-1, 2D-19, and
2D-24)have been cloned from second day postamputation regener-ating
limbs, and a cDNA (EB-1) was cloned from earlybud-stage
regenerating limbs. These cDNAs included thefull-length coding
regions. The deduced amino acid sequencesof 2D-1, 2D-19, 2D-24, and
EB-1 had a homology with mam-malian MMP9, MMP3/10, MMP3/10, and
MMP13, respec-tively. The basic motif of these newt MMP genes was
similarto mammalian counterparts and contained regions encodinga
putative signal sequence, a propeptide, an active site withthree
zinc-binding histidine residues, a calcium-binding do-main, a
hemopexin region, and three key cysteine residues.However, some
unique molecular evolutionary features werealso found in the newt
MMPs. cDNAs of 2D-19 and 2D-24contained a specific insertion and
deletion, respectively. Theinsertion of 2D-19 is threonine-rich,
similar to the threoninecluster found in the collagenase-like sea
urchin hatchingenzyme. Northern blot analysis showed that the
expressionlevels of the newt MMPs were dramatically increased
afteramputation, suggesting that they play an important role(s)
intissue remodeling of the regenerating limb.
Limbs of adult urodele exhibit a remarkable ability to
restoremissing parts when they are accidentally lost and have
pro-vided investigators with an ideal experimental model to
studythe mechanism of the complete restoration of original
pattern(1, 2). Generally, limb regeneration proceeds through
fivesteps: (i) formation of wound epidermis, (ii)
dedifferentiationof cells under the wound epidermis, (iii)
formation of blast-ema, (iv) growth and differentiation of the
blastema, and (v)pattern reformation. Dedifferentiation and
blastema forma-tion are unique features of urodele limbs and are of
primeimportance in the initial phase of regeneration (1, 2).
Thebreakdown of interstitial connective tissues, cartilages,
bones,and muscles under the wound epidermis seems to be a triggerof
the dedifferentiation of liberated cells, because these cellsstart
to lose the morphologic characteristics of their differen-tiated
state concomitantly with the tissue demolition (2).
It has been generally accepted that extracellular matrix(ECM)
molecules rapidly turn over during processes involvingtissue
remodeling, such as wound healing, metamorphosis, andregeneration
(2-4). ECM is thought to stabilize the differen-tiated state of
cells and support the expression of their normal
tissue-specific phenotypes (5, 6). Consequently, the
degrada-tion of the ECM destabilizes the differentiated state and
wouldbe a trigger of the dedifferentiation. The dedifferentiated
cellsproliferate and form a blastema (1, 2). Matrix
metal-loproteinases (MMPs) play a major role in the degradation
ofthe ECM. Grillo and coworkers (7) assayed collagenolyticactivity
at different stages of newt limb regeneration. Theactivity
increased as histolysis progressed and was highest inthe region
just proximal and distal to the blastema-stumpjunction. Recently,
Yang and Bryant (8) demonstrated thepresence of gelatinolytic
enzymes with molecular weights of90,000, 73,000, 60,000, 55,000,
and 52,000 in regeneratingMexican axolotl limbs using the
zymographic technique (8).Because the distribution of ECM
components such as
collagen, fibronectin, laminin, tenascin, and
proteoglycancomplexed with hyaluronic acid and condroitin sulfate
in limbtissues is drastically changed during regeneration
(9-12),MMPs should be expected to play crucial roles in the
processof tissue remodeling. So far, 11 types of MMPs have
beenidentified among mammals and chickens (13). These enzymesshare
some common structural motifs and constitute a singleprotein
superfamily. They are synthesized as a proenzymeform, and contain
Ca2+- and Zn2+-binding domains. Theydiffer in specificity of
substrate. For example, MMP1 (inter-stitial collagenase) and MMP13
(collagenase 3) degrade typeI, II, and III collagens, and
proteoglycan. MMP2 (gelatinase a)and MMP9 (gelatinase b) hydrolyze
gelatin, proteoglycan, andcollagens (types IV, V, VII, and X). MMP3
(stromelysin 1) andMMP10 (stromelysin 2) decompose fibronectin,
laminin, pro-teoglycan, and type II, IV, V, IX, and X collagens.
MammalianMMPs have been shown to function in ECM remodelingduring
embryonic development (14, 15) and wound healing(4), and in the
metastasis of transformed cells (16).To our knowledge, only three
amphibian MMP cDNAs,
MMP1 (17) from Rana catesbeiana, MMP11 (stromelysin 3)(18), and
MMP13 (M. E. Fini, S. Scott, Z. Wang, and D. D.Brown, unpublished;
GenBank accession no. L49412) fromXenopus laevis have been cloned
hitherto. These are uniqueanimals in that they show dramatic tissue
remodeling duringmetamorphosis. However, no clones have been
isolated forcDNAs of newt MMPs that function in tissue
remodelingduring limb regeneration.The purpose of the present study
was to isolate and char-
acterize cDNAs encoding MMPs that are involved in the
limbregeneration of urodela. Newt MMP cDNAs were cloned asproducts
of the reverse transcription (RT)-PCR and used asprobes to clone
MMP cDNAs from cDNA libraries of newt
Abbreviations: ECM, extracellular matrix; MMP, matrix
metal-loproteinase; RT, reverse transcription; nMMP, newt MMP.Data
deposition: The sequences reported in this paper have beendeposited
in the GenBank data base [accession nos. D82052 for 2D-1(nMMP9);
D82053 for 2D-19 (nMMP3/10-a); D82054 for 2D-24(nMMP3/10-b); and
D82055 for EB-1 (nMMP13)].TPresent address: Department of
Bioscience, Kitasato University,Sagamihara, Kanagawa 228, Japan.ITo
whom reprint requests should be addressed.
6819
The publication costs of this article were defrayed in part by
page chargepayment. This article must therefore be hereby marked
"advertisement" inaccordance with 18 U.S.C. §1734 solely to
indicate this fact.
Dow
nloa
ded
by g
uest
on
June
27,
202
1
-
6820 Developmental Biology: Miyazaki et al.
limbs. Four different cDNAs were successfully cloned. Thesewere
homologous to mammalian MMPs and contained severalcommon structural
motifs. In addition, some unique evolu-tionary features were found
in the structure of the newt MMPgenes. Northern blot analysis
suggests that the expression ofMMP genes is dramatically increased
in response to signalsinduced by the amputation.
MATERIALS AND METHODSAnimals and Amputations. Newts, Cynopus
pyrrhogaster,
were obtained from a local animal supplier. Animals, about 10cm
long, were anesthetized in a 0.1% solution of MS222(Sigma), and
their forelimbs were amputated distal to theelbow. Newts were given
no food in the first week after theamputation and were then fed on
earthworms twice a week.They were maintained for appropriate
periods in water at250C.
Isolation of RNA from Regenerating Limbs and Its RT-PCR.
Regenerating limb tissues at the early bud stage were cutat the
elbow and were frozen until use. RNAs were extractedfrom the
tissues by either the acid guanidium phenol chloro-form method (19)
or the guanidine isothiocyanate/cesiumchloride method (20), from
which poly(A)+ RNAs wereseparated using an oligo-dT column. cDNAs
were synthesizedby reverse transcriptase (Boehringer Mannheim) from
thesepoly(A)+ RNAs and used as templates for the PCR. Twodegenerate
PCR primers that correspond to two highly con-served sequences were
synthesized: one in the cysteine switchand the other in the
catalytic domain of mammalian MMPs.These sequences were
5'-TG(T/C)GGTGTICCIGA(T/C)GTand 5'-ICCIGGICC(A/G)TC(A/G)AA and were
used as asense and an antisense primer, respectively. The PCR
wascarried out with 30 cycles of denaturation (940C, 1
min),annealing (55°C, 1 min), and extension (72°C, 2 min),
whichamplified two cDNAs with 262 bp. They were ligated into
thevector pCRII using a TA cloning kit (Invitrogen) and se-quenced
using a DNA sequencer (Applied Biosystems, model373A). The products
were found to be MMP-like fragmentsand named 14-2 and 14-3,
respectively. A PCR fragment (H-1)was amplified from human MMP1
cDNA with the samedegenerate PCR primers and used as one of the
mixed probesfor screening newt cDNA libraries.
Construction and Screening of Newt Limb cDNA Libraries.Poly(A)+
RNAs were isolated from normal limbs, second daypostamputation
regenerating limbs, and early bud-stage re-generating limbs as
described above and were used for con-structing cDNA libraries in
uni-ZaplI (Stratagene) accordingto the manufacture's instructions.
RT-PCR products (14-2 and14-3) and PCR products (H-1) were excised
from the clones ofpCRII, purified by agarose gel electrophoresis,
and labeledwith [32P]dCTP (Amersham) using an oligo labeling kit
(Phar-macia). Approximately 2 x 106 clones of a cDNA library
ofsecond day postamputation regenerating limb tissues werescreened
using the above-mentioned 32P-labeled RT-PCR andPCR products as
probes. Likewise, about 0.8 x 106 clones andabout 1.1 X 106 clones
were screened from cDNA libraries ofearly bud-stage regenerating
and normal limb, respectively. Asa result, three clones (2D-1,
2D-19, and 2D-24) and one clone(EB-1) were obtained from cDNA
libraries of the second daypostamputation limb and the early
bud-stage regeneratinglimb, respectively.Northern Blotting
Analysis. Total RNAs (10 ,tg) that had
been extracted from normal limbs and regenerating limbs at 2,5,
8, 15, 21, and 28-35 days after amputation were denaturedin 10 mM
sodium phosphate buffer (pH 7.0) containing 1 Mglyoxal and 50%
dimethyl sulfoxide. The regenerating limbs at28-35 days had
developed to the palette stage. They wereseparated by
electrophoresis on a 1% agarose gel in 10 mMsodium phosphate buffer
(pH 7.0) and transferred to nylon
membranes (GeneScreen Plus, DuPont) according to
themanufacturer's protocol. The blots were stained with
0.04%methylene blue in 0.5 M sodium acetate (pH 5.2) to
visualizerRNAs. Filters were destained in 10 mM Tris HCl
buffer(pH7.4) containing 5 mM EDTA and 1% SDS, and used
forhybridization. Restriction enzyme fragments of 2D-1, 2D-19,and
2D-24 were obtained as outlined below, purified byagarose gel
electrophoresis, and used as probes. 2D-1 (nucle-otide sequence
1274-2214) was obtained with EcoRI andXbaI; 2D-19 (nucleotides
96-1780) was obtained with PstI andXhoI; and 2D-24 (nucleotides
837-1425) was obtained withSmaI and XbaI. The fragments were
labeled with 32P asdescribed above. Blots were hybridized with
32P-labeled probesat 42°C for 16 hr in 50% formamide, Sx standard
salinephosphate/EDTA, 5x Denhardt's solution, 1% SDS, and
10%dextran sulfate. Filters were washed at 65°C with 0.2x stan-dard
saline citrate and 0.1% SDS. Size of RNAs was deter-mined by
comparing their migration on the gels with themigration of standard
RNAs (GIBCO/BRL).
Construction of the Phylogenetic Trees. The trees
wereconstructed from the deduced amino acid sequence data of
thefour newt MMPs according to UPGMA (Unweighted PairGroup Method
with Arithmetic Mean) which had been oper-ated by a GENEWORKS
software program (IntelliGenetics) (21).
RESULTS AND DISCUSSIONAmplification of Newt cDNA Fragments.
Enzymes of the
MMP superfamily contain several highly conserved domains intheir
molecular structures: a signal peptide, a propeptidecontaining a
cysteine switch, a catalytic domain, a Zn2+_binding domain, and two
Ca2+-binding domains. Of these, thecysteine switch and the
catalytic domain were used to designoligonucleotides as a pair of
degenerate primers for RT-PCR.RT-PCR was performed for poly(A)+ RNA
extracted fromearly bud-stage regenerating newt limbs. A cDNA whose
size,262 bp, was identical to that of clone H-1 (amplified
fromhuman MMP1 cDNAs with the same degenerate primers)
wasamplified. The product was subcloned into pCRII vectors,
20clones of which were sequenced and found to all have ahomology to
known MMP genes; 15 clones (14-2) showed ahomology to MMP9
(gelatinase b) and the other 5 clones(14-3) showed a homology to
MMP13 (collagenase 3).Cloning of MMP cDNA from cDNA Libraries of
Regener-
ating Limbs. cDNA libraries of normal limbs and second
daypostamputation regenerating limbs were screened to isolatenewt
homologues ofMMP genes using three amplified cDNAs(14-2, 14-3, and
H-1) as mixed probes. No meaningful cloneswere obtained from normal
limb libraries. As described below,Northern blot analysis showed
that the amount of MMP genetranscripts was very low in the
unamputated normal limb. Thismight be a part of the reason for the
failure in the cloning fromnormal limbs.The library of second day
postamputation regenerating
limbs yielded three clones of MMP genes named 2D-1, 2D-19,and
2D-24, respectively. 2D-1 was 3.9 kb long, encoded 718amino acids,
and showed 54% homology to human MMP9(22). The other two clones,
2D-19 and 2D-24, were 1.8 kb and1.7 kb long and encoded 484 and 470
amino acids, respectively.As described below, Northern blot
analysis gave the identicalsizes for transcripts of these genes.
The predicted amino acidsequences of 2D-19 and 2D-24 indicated that
they corre-sponded to MMP3 (23), showing 53% and 53%
homology,respectively, or to MMP10 (24), showing 51% and
52%homology, respectively. These two clones encode
differentpolypeptides. The homology between them was 61%.A cDNA
library from early bud-stage regenerating limbs
was screened by mixed probes of 14-3 and H-1. In addition
to2D-19 and 2D-24, a novel clone, EB-1 (4.0 kb long and
Proc. Natl. Acad. Sci. USA 93 (1996)
Dow
nloa
ded
by g
uest
on
June
27,
202
1
-
Developmental Biology: Miyazaki et al.
encoding 471 amino acids), was isolated and showed 68%identity
to human MMP13 (collagenase 3) (25).
Structural Characteristics of Newt MMP Genes. A grosscomparison
of amino acid sequences of MMPs among themammal, chicken (26), and
newt reveals an uneven distribu-tion of conserved sequences in the
molecules. The amino-
_signal -...& PI
Proc. Natl. Acad. Sci. USA 93 (1996) 6821
terminal domains contain high homology regions common toeach
species, whereas the hemopexin domains show highspecies
divergence.The basic motif of the newt MMP genes cloned in this
study
was similar to that of established human MMP genes (Fig. 1).This
motif contains sequences encoding five major domains: a
o-peptideM*P9 K-SLWQPLVL VLLVLGGCFA APRQRQSTLV LFPGDLRTNL
TDRQLAEEYL YRYGTRVAE MRGESKSLCP ALLLLQKQLS LPETGEWSA 'L. 992D-1
MKPQLALLAL GLLALGCRAA PLQSQPQVRV TFPGELVSGI SDDELAESYL ERFGYISKRA
RSSTHVSLSK ALLQMQKKLG LINElELDQS fL 100MMP3 NK--SL-PIL LLLCVAVCSA
YPLDG-AARG E--DT-SMNL VQKYL-ENYY DLKKDVKQFV RRKDSGPVVK KIREMQKFLG
LVWXLDSD U. Pt 922D-19 KK--SL--SL LaLLVVHTYA FPAVP-ATED R--GE-NEQL
AE1YL-KKFY NLNED-GTPI TRKKHSPFSE KLQEHQAFFG LEVIKLDSN U. 902D-24
NK--IL--SL LLLCAAGAYA VQEAP-VHEE D--DT-IRQO VRYLX-KKYY GLNSD-KTPD
LRKAASPLAE KIEMOKFCG LQVTGKVDSN 90MMP13 DIPGVL-A&F LLSWWHCRA
1PLPSGGDED DLSEE-DLQF AERYL-RSYY HPTNL-AGILEB-1 W4PSVLSAAI
FFLSLAFGLP VPVPH-ERDS DVTEQ-ELRL AEKYL-KTFY VASDH-AGIM
catalyvic domainMMP9 !PRE? FEGDLKVHHH NITYWIQNYS EDLPRAVIDD
AFARAFALWS AVTflTFTRV2D-1 GYP GN FDGDLKWDHN DITFRVLNYS PDLDGDVIED
AFRRAFKVWS DVSIPLTFTIMMP3 OHER IFPG IP T HLfYRIVNYT PDLPKDAVDS
AVEKALKVWE EYTPLTFSRL2D-19 IA4VMTSH FGGRPTWRT SLMYILGYT PMAEDVDT
AIRRAFKVWS DVTPLT7SRI2D-24IS PGRPARH ALTYRILNYT PDKRAVDT AIQLAFJVVS
DVTPLTITQIMMP13Jj¶4GYNV FPRTLXWSKM NL!YRIVNYT PDITHSEVEK AFKKAFKVWS
DWTPLNFTRLEB-1 PGNV FPRSKNPRF NLTYRIENYT PDNHAEVDR AIKXAFRVW S
EVTPLHFTRL
3ium-binding tomainMMP9 dADDDL WgS'KG'VVP TRFGNADGAA CHFPFIFEGR
SYSACTTDGR SDGLPWCST2D-1 GUAWDDSF WTLGTGVVVR TEGNANGA CKPPFXFNGN
SYSSCTSEGR TDGLLWCSTTMMP3 (WAEPDDD WTRDT2D-19 GUaNPDZDTAGS-.2D-24
SDAflfl WSXVSMMPi 3 GDAWDD WTSS8.EB- I WDTIWDMT FTSGS-
gelatin-binding_domain
KENAASSMTE RIREI4QSFFG LEVTGKLDDN TL K 96TKKGGNALAS KLREMQSFFD
LEVTGKLDED 1LEVNKQ 96
YSRDADIVIQ FGVAEIGDGY PFDXDGLLA SAFP?GPGIQ 199YSGEADIMIL
FGSDDDGDPY PFDGKDGLL& EAyppGEGvQ 200YEGEADIMIS FAVREUGDFY
WPGNVL& EAYAPGPGIN 192YEGTADIQIS FGAGVHGDFY PPGPHGTL&
KAFAPONSIG 190YYGTWDIQIS FGAREHDFN PFDGPYGTLA UAAPGTGIG
190HDGIADIMIS FGIKEBGDFY PFDGPSGL& SAFPGPNYG 196RSGThDIMIS
GTKEDGDFY PFDGPNGLLA H&FPGQRIG 196
ANYDTDDRFG FCPSERLYTR DGNADGKPCQ FPFIFQSYTDYDKDKKYG FCPSELLYTY
GGNSDGDKCV FPPIFDGDSY
299300
207205205
211211
MMP9 SACTTWGRSD GYRWCATTAN YPRDKLFGFC PTRADSTVMG GNSAGELCVP
PFTFLGKEYS TCTSEGRGDG RLWCATTSNF DSDKKWGF6P DQGYSIIrA 3992D-1
DACTKEQRSD GYRWGGTSDT FDKDKKYGFC PNR-DTAVIG GVSQGDPCVF PFVLKTYH
SCTSDGRGDR KLWCATTSSY DSDRXWPCP DQSY LYG
399MP3--------------------TTNLflVA 216
2D-19 --AGYNLVA 2142D-24 . -.-- - -VTTLA 214M--1--- -K- -NVA
221EB-1
----------------------------------------------------NGYMIUFIA
221
zinc-binding d4in calcjym-binding an hinge domainMMP9 LJL
ZGSVPEALUB YPtYRFTEG -PPLHKDDV NGIRLGPR PEPEPRPPTT TTPQPTAPPT
VCPTGPPTVH PSERPTAGPT GPPSAGPTGP 4962D-1 aInreAIa EHBTVRDALM
YlMfRYIEG- --FQLHQDDI EGVQYLIGSG TGPHPSPPMP T----TKSPD VSGKTTTTV-
TTSPT ---- 476WMP3 £EIMSULS FESANTEAU( YPLYHSLTDL TRFRLSQI
IGIQSLYGPP PD- -SPETPLV PTEPV----- ---------- ---------- 2802D-19
SISi BSGXI "J SYI-DP ARFRLPQV DOIQAILGAS PN-PVPTTPQ ATTPTTTVST
TTTT-2862D-24 LUFUSLO. S3SNIAU FfTSX-DP AFLPKUI ISIQAIlOPS
RK-PSPQTPP PTKPA- 277MMP 1 3 ANhEGUSLS DBKDPGAL FPIfYTY-GK
SHF4LPDDDV QGIQSLPG DE- - ---DPN--- 274EB- 1 AflEKALU. DSRDPGSJK
YPVYSYT-EP SRFLLPDVDV QGIQSLGPG NRD PN- 274
MMP9 PTAGPSTATT VPLSPVDDAC NV-NIFSIA EIGNQLYLFE DGJYWRFSEG
RGSRPQGPFL2D-1- TTTEL VPVDPTTDAC MV-RAFMT SIEGQLHFPK DGKYWMASSA
RPGAIMGPVK IADKNPALPR MLDBVFUEPL SKMLFFSOR QVKVVY?GSV 59 5ISDTWALPA
IIDSAP3DLL TMMIFFPSGR RPIUVYTGTTV 570
MMP3-PPEPGTPANC DPALSFDVS TLRGSILIFX DRHFVRKSLR KLEP--ELHL
ISSFWSLPB GVDA&Y3VTS MDLVFIIKGN QFKAIRGNEV 3682D-19-TSSPINPSIC
DPTLVFDIT TLRGEILFIP DSSFVRRVPT IKEV--YNYP ISTSW SWf GIQAMYINPE
TDQIFLPRGS MYUALQGFDI 374
2D-24 -LQSYC DPAIRWMIT TLRNNILFM GRTFLRSMPH TGRI--ISYT
ISAVWPSLPS GIHAAYRNQQ KDQVLLPRON KYVAGYQM 360MMP13- - PKHCTPIDJC
DPSLSLSIT SLRGTMIPK DRFWLRHPQ QVDA--ELFL TKSFWEWLPN RIDAAYZHPS
HDLIFIFRGItR FVALNGYDI 36tEB-1 - PMHPMTPEMC DPELSLSIT EMRGEMLIVE
DRFFNRQHPQ MTDV--ELVL IRNFWELPS KIDAAYSYPE XDLIYIVRG KMWLNOYDI
362
hemopexin domainMMP9 LG--PRRLDK LOLGADVAQV TGAL-RSGRG XMLIPSGRRL
WRFUVKAQMV DPRSASEVDR MVPGPLDTH DVFQYREMAY FCQDRFYWRV SSRSELNQVD
6922D-1 LG--PLEK LGIGKDVEMI VGSL-QRGRG KVLLFNGDKY WRLWKQVV
DKGYPRDTED A?AGVPINAS DVFLYQENIH PCQFYWRM TPR---RQVD 664MMP3
RAGYPRGIHT LGFPPTVRKI DAAISDKEKN KTYFFVEDKY WRFDEMRNSM EPGFPKQIAE
DFPGIDSKID AVFEEFGFFY PFTGSSQLEF DPNA--KMVT 4662D-19 LPNYPKMIDX
LGPRTVKNI NAAVYLQSTQ ITYFFAGEQY VSYDEARMTM DKESPERIED DVPGIGIKVH
AVFEDNGLLY ?FSGHKQFEF MMS--KKVT 4722D-24 LPWYPQNIYT LOLPRTVTRI
DAAVYHPDTR XTYYVNDRY WSWALQVH DRDSPQQIVT T7PRITMVD AVWYAKGLLY
VFNGQHQFF mI.RL--nvr 458MMP13 LEGYPKISE ROLPKEVKKI SA&VHFEDTG
STPLLFSGNQV WRYDDTNHIM DKDYPRLIEE DFGIGDKVD AVYEKNGYIY FFNGPIQFEY
SIxS--RIV 460EB-1 LAD tIPR IAPSLRTI DAAVYNRA?G XI VGY WNSDEEKQT
ERGYPRFIAD DVPGIETVD AAYQRNGYIY ?FSGSLQFY ST--EKVI 460
MMP9 QVGYVTYDIL QCPED 7072D-1 QVGYVKYDIL NCPENT 680MMP3
HT-LKSNSWL NC 4772D-19 RT-LKNTSWL GC 4832D-24 RV-LKXSSWF SC
469MMP13 RV-MPANSIL WC 471EB- I RV-LKTNMSL wC 471
FIG. 1. Comparison of amino acid sequence between newt and human
MMPs. Amino acid sequences of seven MMPs, human MMP9 (22),
2D-1,human MMP3 (23), 2D-19, 2D-24, human MMP13 (25), and EB-1 were
aligned using UPGMA (operated by a GeneWorks software program)
andare shown in rows 1-7. The residues common to all seven MMPs are
in bold. Shaded boxes indicate homologous regions between MMP9 and
2D-1;among MMP3, 2D-19, and 2D-24; and between MMP13 and EB-1. Each
domain was assigned a signal peptide, propeptide, catalytic
domain,calcium-binding domain, gelatin-binding domain, zinc-binding
domain, hinge domain, and hemopexin domain according to Takino and
coworkers(27). The cysteine switch is boxed and three histidine
residues in the zinc-binding domain are marked by asterisks.
_ _1
Dow
nloa
ded
by g
uest
on
June
27,
202
1
-
6822 Developmental Biology: Miyazaki et al.
putative signal sequence, a propeptide, an active site with
threezinc-binding histidine residues, calcium-binding domains, anda
hemopexin region. In addition, three key cysteine residueswere
similarly found: one in the cysteine switch of propeptidesand the
other two in the hemopexin region that forms thedisulfide bond. The
cysteine switch of the 2D-1 and EB-1 wasidentical to that of known
MMP genes (PRCGVPD), whereasproline at the sixth residue of 2D-19
and 2D-24 was replacedwith alanine and serine, respectively (Fig.
1). These variationsin the sixth amino acid residue were confirmed
by the fact thatRT-PCR with the oligonucleotide corresponding to
CGVPDVas sense primers yielded fragments of 2D-1 and EB-1 but notof
2D-19 and 2D-24. It would be interesting to know whetherthis
replacement of the sixth proline residue with alanine orserine
affects the function of the switch as a suppressor ofenzyme
activity, because the point mutation experiments haveshown that
replacement of the sixth residue with valine orasparagine residues
destroys the function of the switch (28).We speculate at present
that the replacement with alanine orserine should not influence the
function of the cysteine switch.However, we have not attempted to
obtain the expressedenzymes of cDNAs of 2D-19 and 2D-24 and have
not deter-mined whether these expressed proteins show the
actualactivity of the cysteine switch. This type of experiment
isrequired for testing the speculation. The three histidine
resi-dues in the catalytic domain that bind to Zn2+ and interact
withthe cysteine switch (29) are well conserved among the fournewt
MMPs.
The amino acid sequence homology test predicted that the2D-1
clone is a newt homologue of gelatinase b (MMP9). Thisresult was
verified by aligning the predicted amino acidsequence of 2D-1 with
human MMP9 (Fig. 1), which revealedthe presence of a
gelatin-binding domain consisting of fourfibronectin type II
repeats (22).
Site-specific deletion experiments on neutrophil
collagenase(MMP8) have shown the importance of a region
containing16-amino acid residues in the carboxyl-terminal domain
for theinteraction of the enzyme with triple-helical domains of
col-lagen (30) (Fig. 2A). Stromelysins contain an insertion
con-sisting of nine amino acid residues in this region, inhibiting
thebinding of stromelysin to the collagen triple helix (30)
(Fig.2A). EB-1 was identified as MMP13 (collagenase 3) accordingto
its amino acid sequence homology. This identity was sup-ported
further, because EB-1 lacks the nine-amino acid inser-tion (Fig.
2A). Clones of 2D-19 and 2D-24 contained theinsertion (Fig. 2A,
bold), consistent with their high sequencehomology to MMP3
(stromelysin 1) and MMP10 (stromelysin2).However, newt stromelysins
(2D-19 and 2D-24) were found
to be unique in the structure of the insertion. 2D-24 containeda
homologous nine-residue insertion, but the cDNA wasunique in that
it lacked five amino acids adjacent to thecarboxyl terminus of the
insert (Fig. 2A, asterisks). Instead ofa nine-amino acid insert,
the insert (amino acids 269-286) of2D-19 was 18 residues long and
threonine-rich, and lacked thesequence homologous to the
nine-residue insert found in
(A)
mammalian MMP1(H) 258
AIYGRSQNPVQ------------------PIGPQTPKACcollagenase MMP8 (H) 259
AIYGLSSNPIQ------------------PTGPSTPKPC
MMP13(H) 264 SLYGPGDEDPN------------------PKHPKTPDKCnewt
collagenase EB-1 264 SLYGPGNRDPN------------------PKHPKTPEKC
MMP3(H) 261 SLYGPPPDSPETPLVPTEPV--- --PPEPGTPANCmammalian
MMP10(H) 260 SLYGPPPASTEEPLVPTKSV---------PSGSEMPAKCstromelysin
MMP3(R) 259 SLYGPPTESPDVLVVPTKSN---------SLDPETLPMC
MMP10(R) 261 SLYGARP-SSDATVVPVPSV---------SPKPETPVKC
newt 2D-24 258
AIYGPSRKPSPQTPPPTKPA--------------LQSYCstromelysin 2D-19 258
ALYGASPNPVPTTPQATTPTTTVSTTTTTTSSPINPSIC
(B)newt
gelatinase 2D-1 467 TTTTVTTSPTTTT
(C)MMP1(H) 209 NYNLHRVAAHELGHSLGLSHSTDIGA
mammalian MMP8 (H) 208 NYNLFLVAAHEFGHSLGLAHSSDPGAcollagenase
MMP13(H) 213 GYNLFLVAAHEFGHSLGLDHSKDPGA
newt EB-1 213 GYNLFIVAAHEFGHALGLDHSRDPGScollagenase
2D-24 207 GTNLFLVAAHEFGHSLGLSHSNDRNAnewt 2D-19 207
GYNLFLVAAHEFGHLSGLSHSGDRSA
stromelysinMMP3(H) 209 GTNLFLVAAHEIGHSLGLFHSANTEAMMP10(H) 208
GTNLFLVAAHELGHSLGLFHSANTEA
mammalian MMP3(R) 207 GTNLFLVAAHELGHSLGLFHSANAEAstromelysin
MMP1O(R) 209 GTNLFLVAAHELGHDLGLFHSNNKES
FIG. 2. Comparison of subtype-specific amino acid sequences of
newt and mammalian MMPs. Amino acid sequences of known MMPs are
fromFreije and coworkers (25). Arabic numerals at the left side of
each amino acid sequence represent the positions of the left-end
residues of sequences.(A) Amino acid sequences near the
nine-residue insertion found in stromelysins. H, human; R, rat.
Both ends of the 16-amino acid sequence requiredfor the interaction
with type I collagens are marked with arrowheads. The nine-residue
insertions specific for stromelysins in this region and
theinsertions in newt stromelysins (2D-19 and 2D-24) are shown in
bold. The five-residue deletion specific for 2D-24 following the
insertion is markedby asterisks. Dashes are deleted residues when
the sequences were compared with 2D-19. (B) Amino acid sequence of
the threonine cluster foundin 2D-1. The sequence is from Fig. 1.
(C) Amino acid sequences around the zinc-binding domain.
Subtype-specific residues in mammaliancollagenase and stromelysin
and the residues located at the corresponding position in newt
collagenase and stromelysins are shown in bold.
Proc. Natl. Acad. Sci. USA 93 (1996)
Dow
nloa
ded
by g
uest
on
June
27,
202
1
-
Proc. Natl. Acad. Sci. USA 93 (1996) 6823
MMP3 and MMP10. The presence of a threonine-rich regionseemed to
be one of the unique features of newt MMPs,because 2D-1 also
contains the threonine cluster at aminoacids 467-479 (Fig. 2B).
Mammalian MMPs reported hithertodo not contain such a threonine
cluster. Interestingly, thethreonine cluster is found in the sea
urchin hatching enzyme,a homologue of collagenases (31). The
threonine-rich regionin the hatching enzyme has homology with
Drosophila salivaryglue protein sgs-3, which contains the motif
(T)4_5K(A/P)(32). The pattern of this motif in the hatching enzyme
is lessregular and the number of repeats is smaller compared
withthe motif and repeats for sgs-3. Threonine clusters in 2D-1
and2D-19 are shorter and less regular than they are in the
hatchingenzyme. It is intriguing to speculate that some of the
newtMMPs retain traits characteristic of a primitive type of
colla-genase.Three amino acid residues, Tyr-210, Asp-231, and
Gly-233,
in MMP1 are located around the zinc-binding site and are
wellconserved in other collagenases (MMP8 and MMP13) (Fig.2C).
These residues are not found in stromelysins (MMP3 andMMP10).
Instead threonine, asparagine, and glutamic acid arefound at
positions 210, 231, and 233, respectively (Fig. 2C).EB-1 contained
all three residues at the corresponding sites,which is consistent
with the fact that EB-1 is a collagenase.Interestingly, the newt
stromelysins 2D-19 and 2D-24 con-tained two and one of the three
collagenase-specific residuesat the corresponding sites,
respectively: Tyr-208 and Asp-229in 2D-19 and Asp-229 in 2D-24
(Fig. 2C). 2D-24 and 2D-19might be ancestral types of MMPs and show
intermediatecharacteristics of stromelysins and collagenases.The
newt MMPs are located in the phylogenetic tree
constructed from known MMPs as shown in Fig. 3A. Consis-tent
with the structural characteristics of the newt MMPs
A 2D-242D-19human MMP3human MMP1O
EB-1Xenopus MMP13
r Human MMP13Rana MMP1Xenopus MMP11
described above, 2D-1 and EB-1 are closely related to humanMMP9,
and human and Xenopus MMP13, respectively. Sim-ilarly, both 2D-19
and 2D-24 are grouped as members close tohuman MMP3 and MMP10. As
shown in Fig. 3B, 2D-19 and2D-24 represent evolutionary ancestral
molecules of bothmammalian MMP3 and MMP10 in phylogenetic tree.
Appar-ently, newt MMPs are differently located in the tree
ascompared with other known mammalian MMPs. These vari-ations can
be explained by the unique characteristics of newtMMPs described
above. Based upon these data and consid-erations, we propose to
designate 2D-1, 2D-19, 2D-24, andEB-1 as nMMP9 (newt MMP9),
nMMP3/10-a, nMMP3/10-b,and nMMP13, respectively. Considering the
molecular size andsubstrate specificity, nMMP9 seems to be the
90-kDa matrixmetalloproteinase reported by Yang and Bryant
(8).There are some small discrepancies between the phyloge-
netic tree in Fig. 3A and the tree described in Murphy
andcoworkers (37). These discrepancies might be due to
thedifference in constructing phylogenetic tree. We made the
treeusing whole sequence of the genes, whereas the cited
authorsused the sequence of catalytic domains or the sequence
lackingin fibronectin-like domain in case of gelatinases.
This study characterized and identified the nMMPs entirelyby
their sequence homology to known MMPs. As describedabove, we have
not produced the expressed enzymes encodedby genes of nMMPs. It
remains to be tested if the expressedenzymes show the actual
activity.
Expression of nMMP Genes in Regenerating Limb. North-ern blot
analyses of nMMP cDNAs (nMMP9, nMMP3/10-a,and nMMP3/10-b) were
performed to determine the size oftheir transcripts and the change
in their expression levelsduring limb regeneration (Fig. 4). nMMP9
hybridized a tran-script with a size of 4.0 kb. The other two
cDNAs, nMMP3/10-a and nMMP3/10-b, hybridized transcripts with sizes
of 1.8and 1.7 kb, respectively. nMMP3/10-a transcripts were
faint.The size difference between transcripts of nMMP9,
andnMMP3/10-a and nMMP3/10-b is explainable by the fact thatnMMP9
contains extra sequences coding the gelatin-bindingdomain and
longer noncoding regions.Human MMP3 and MMP10 are very close in
sequences
(78%) and hybridize to each other's mRNAs unless thedivergent 3'
untranslated sequences are used as probes (38). Bycontrast,
nMMP3/10-a and nMMP3/10-b can be distinguishedfrom each other in
the Northern blot analysis. This might bedue to less sequence
homology between the probe used (56%between MMP3/10-a probe and the
corresponding sequences
u Human MMP9
nMMP3/10-a
B ~~~~~~~2D-242D-19Human MMP3
-Rabbit MMP3_ | ~Human MMP10
-E--Mouse MMP3Rat MMP3Rat MMP10
FIG. 3. Phylogenetic trees of MMPs. The trees were
constructedusing the whole amino acid sequences predicted from
cDNAs ofnMMPs and those of human MMP3 (23), human MMP9 (22),
humanMMP10 (24), human MMP13 (25), rabbit MMP3 (33), mouse
MMP3(34), rat MMP3 (35), rat MMP10 (36), Rana MMP1 (17),
XenopusMMP11 (18), andXenopus MMP13 (M. E. Fini, S. Scott, Z. Wang,
andD. D. Brown, unpublished; GenBank accession no. L49412). (A)
Atree with newt MMPs, human MMPs, and anuran MMPs. (B) A treewith
2D-19, 2D-24, and known mammalian MMP3s and MMP1Os.
nMMP3/10-b
rRNA--a28S
-.018S
FIG. 4. Northern blot analysis of nMMP transcripts during
limbregeneration. Total RNAs were extracted from normal limbs
(0),regenerating limbs at indicated days after amputation (2, 5, 8,
15, and21), or palette-stage regenerating limbs (P) (28-35 days).
Ten micro-grams of the RNAs was subjected to gel electrophoresis.
The blotswere hybridized with labeled fragments ofnMMP9,
nMMP3/10-a, andnMMP3/10-b. A panel of rRNA indicates the
corresponding blottingfilter stained with methylene blue to show
the amount ofRNA presentin each lane.
Developmental Biology: Miyazaki et al.
0 2 5 81521 PnMMP9
Dow
nloa
ded
by g
uest
on
June
27,
202
1
-
6824 Developmental Biology: Miyazaki et al.
of nMMP3/10-b gene; 64% between nMMP3/10-b probe andthe
corresponding sequences of nMMP3/10-a gene).Although the expression
of nMMP9 was not detected for
unamputated normal limbs as shown in Fig. 4, the signalbecame
faintly positive when more RNAs were charged or theexposure time of
filter to x-ray films was elongated (data notshown). Even in this
case, signals of nMMP3/10-a andnMMP3/10-b were not seen in normal
limbs. Expression of thethree genes was dramatically stimulated in
regenerating limbsas early as 2 days after amputation.
Interestingly, their expres-sion patterns differed among the three
thereafter. nMMP9 wasfully expressed at 5 days, which was sustained
until 15 days, andits expression was drastically declined at 21
days. nMMP3/10-amRNAs continued to be weakly transcribed up to 8
days, andalmost disappeared at 15 days. Regenerating limbs
sustainedthe expression of nMMP3/10-b through the palette
stage.These results suggest that these MMPs might play
differentialroles in the matrix-remodeling of regenerating limbs.
Theexpression of nMMP13 was not detected in either normal
orregenerating limbs on the same filters detected for othernMMPs
(data not shown). This means nMMP13 would betranscribed at a very
low level.
We thank Dr. D. L. Stocum for helpful comments and
carefulreading of the manuscripts, and Ms. Y. Kobayashi for helpful
technicalassistance.
1. Wallace, H. (1981) Vertebrate Limb Regeneration (Wiley,
Toron-to).
2. Stocum, D. L. (1995) Wound Repair, Regeneration and
ArtificialTissues (Springer, New York).
3. Yoshizato, K. (1989) Int. Rev. Cytol. 119, 97-149.4. Gailit,
J. & Clark, R. A. F. (1994) Curr. Opin. Cell Biol. 6,
717-725.5. Hay, E. D. (1993) Curr. Opin. Cell Biol. 5,
1029-1035.6. Talhouk, R. S., Bissel, M. J. & Werb, Z. (1992) J.
Cell Biol. 118,
1271-1282.7. Grillo, H. C., Lapiere, C. M., Dresden, M. H. &
Gross, J. (1968)
Dev. Biol. 17, 571-583.8. Yang, E. V. & Bryant, S. V. (1994)
Dev. Biol. 166, 696-703.9. Mailman, M. L. & Dresden, M.
H.(1976) Dev. Biol. 50, 378-394.
10. Gulati, A. K., Zalewski, A. A. & Reddi, A. H.(1983) Dev.
Biol. 6,355-365.
11. Onda, H., Poulin, M. L., Tassava, R. A. & Chiu, I.-M.
(1991) Dev.Biol. 148, 219-232.
12. Toole, B. P. & Gross, J. (1971) Dev. Biol. 25, 57-77.13.
Birkedal-Hansen, H. (1995) Curr. Opin. Cell Biol. 7, 728-735.14.
Alexander, C. M. & Werb, Z. (1991) in Cell Biology of
Extracel-
lular Matrix, ed. Hay, E. D. (Plenum, New York), 2nd Ed.
15. Liu, X., Wu, H., Byrne, M., Jeffrey, J., Krane, S. &
Jaenisch, R.(1995) J. Cell Biol. 130, 227-237.
16. Stetler-Stevenson, W. G., Aznavoorian, S. & Liotta, L.
A. (1993)Annu. Rev. Cell Biol. 9, 541-573.
17. Oofusa, K, Yomori, S. & Yoshizato, K. (1994) Int. J.
Dev. Biol.38, 345-350.
18. Patterton, D., Hayes, W. P. & Shi, Y.-B. (1995) Dev.
Biol. 167,252-262.
19. Chomczynski, P. & Sacchi, N. (1987) Anal. Biochem. 162,
156-159.
20. Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J. &
Rutter,W. J. (1979) Biochemistry 18, 5294-5299.
21. Nei, M. (1987) in Molecular Evolutionary Genetics
(ColumbiaUniv. Press, New York), pp. 293-298.
22. Wilhelm, S. M., Collier, I. E., Marmer, B. L., Eisen, A. Z.,
Grant,G. A. & Goldberg, G. I. (1989) J. Biol. Chem. 264,
17213-17221.
23. Wilhelm, S. M., Collier, I. E., Kronberger, A., Eisen, A.
Z.,Marmer, B. L., Grant, G. A., Bauer, E. A. & Goldberg, G.
I.(1987) Proc. Natl. Acad. Sci. USA 84, 6725-6729.
24. Muller, D., Quantin, B., Gesnel, M.-C., Millon-Collard,
R.,Abecassis, J. & Breathnach, R. (1988) Biochem. J. 253,
187-192.
25. Freije, J. M. P., Diez-Itza, I., Balbin, M., Sanchez, L. M.,
Blasco,R., Tolivia, J. & L6pez-Otin, C. (1994) J. Biol. Chem.
269,16766-16773.
26. Aimes, R. T., French, D. L. & Quigley, J. P. (1994)
Biochem. J.300, 729-736.
27. Takino, T., Sato, H., Yamamoto, E. & Seiki, M. (1995)
Gene 155,293-298.
28. Sanchez-Lopez, R., Nicholson, R., Gesnel, M.-C. Matrisian,L.
M. & Breathnach, R. (1988) J. Biol. Chem. 263, 11892-11899.
29. Van Wart, H. E. & Birkedal-Hansen, H. (1990) Proc. Natl.
Acad.Sci. USA 87, 5578-5582.
30. Hirose, T., Patterson, C., Pourmotabbed, T., Mainardi, C. L.
&Hasty, K. A. (1993) Proc. Natl. Acad. Sci. USA 90,
2569-2573.
31. Lepage, T. & Gache, C. (1990) EMBO J. 9, 3003-3012.32.
Garfinkel, M. D., Pruitt, R. E. & Meyerowiz, E. M. (1983)J.
Mol.
Biol. 168, 765-789.33. Whitham, S. E., Murphy, G., Angel, P.,
Rahmsdorf, H. J., Smith,
B., Lyons, A., Harris, T. J. R., Herrlich, P. & Docherty A.
J. P.(1986) Biochem. J. 240, 913-916.
34. Hammani, K., Henriet, P. & Eeckhout, Y. (1992) Gene
120,321-322.
35. Matrisian, L. M., Glaichenhaus, N., Gesnel M.-C. &
BreathnachR. (1985) EMBO J. 4, 1435-1440.
36. Breathnach, R., Matrisian, L. M., Gesnel M.-C. Staub A.
&Leroy, P. (1987) Nucleic Acids Res. 15, 1139-1151.
37. Murphy, G. J. P., Murphy, G. & Reynolds, J. J. (1991)
FEBS Lett.289, 4-7.
38. Windsor, L. J., Grenett, H., Birkedal-Hansen, B., Bodden, M.
K.,Engler, J. A. & Birkedal-Hansen, H. (1993) J. Biol. Chem.
268,17341-17347.
Proc. Natl. Acad. Sci. USA 93 (1996)
Dow
nloa
ded
by g
uest
on
June
27,
202
1