Monoclonal antibody MT2 identifies an extracellular matrix glycoprotein that is co-localized with tenascin during adult newt limb regeneration

Differentiation (1992) 50: 133-140

Differentiation Ontogeny, Neoplasia and Differentiation Therapy

Monoclonal antibody MT2 identifies an extracellular matrix glycoprotein that is co-localized with tenascin during adult newt limb regeneration Kenneth P. Klatt’, Eric V. Yang’, and Roy A. Tassava’

*

Department of Biology, Denison University, Granville, Ohio, 43023, USA Department of Molecular Genetics, The Ohio State University, 484 W. 12th Avenue, Columbus, Ohio, 43210, USA

Abstract. Using immunohistochemical techniques and mAb MT2, we describe here a novel extracellular matrix (ECM) molecule that is developmentally regulated during limb regeneration in adult newts. The MT2 antigen appears during preblastema stages, is most abundant during blastema stages, and persists, near undifferentiated cells, until digit stages. The MT2 antigen is located in an acellular layer under the wound epithelium and throughout the ECM of the undifferentiated mesenchyme as a thick, cord-like component. In unamputated limbs mAb MT2 reactivity is restricted to tendons, myotendinous junctions, periosteum and to a layer of material beneath the epidermis. In both unamputated limbs and regenerating limbs, the reactivity to mAb MT2 colocalizes closely with urodele tenascin. Immunoblot analysis of blastema extracts showed that the unreduced form of the MT2 antigen is a large, polydispersed protein of approximately the same size as tenascin. However, based upon (a) molecular weights of reduced subunits, (b) competition experiments on tissue sections, and (c) analysis of molecules immunoprecipitated by mAb MT2, we conclude that the MT2 substance is unrelated biochemically to tenascin. The results from immunoblots, enzyme digestions and DEAE-Sephacell binding studies suggest that the unreduced MT2 antigen is a large protein composed of subunits which are connected by disulfide bonds. Reduction of the MT2 antigen results in three components recognized by mAb MT2. The largest of these reduced components is a chondroitin sulfate-like glycoprotein with a molecular weight (Mr) of 310-325 x lo3. A second component (Mr, 285-300 x lo3) is the core protein of the 31G325 x lo3 glycoprotein. The smallest reduced component, which may be a degradation product of the two larger components, has a Mr of 25G275 x lo3.

*

To whom offprint requests should be sent

A necessary feature of urodele limb regeneration is the formation of a blastema, a mound of embryonic-like cells at the end of the limb stump. Blastema formation involves release of mesodermal cells from specialized tissues, their dedifferentiation and entry into the cell cycle, and their proliferation and accumulation under the newly formed wound epithelium [20, 221. Important to blastema development is remodeling of the extracellular matrix (ECM), a process that involves both degradation of the differentiated stump matrix and synthesis of a matrix unique to the undifferentiated blastema. Several matrix molecules have already been implicated in limb regeneration, including collagen [6], hyaluronic acid [ 11, 231, fibronectin and laminin [7, 151. More recently we showed that tenascin is an abundant matrix protein of both newt and axolotl limb regenerates [ 131. Tenascin is a large, hexameric glycoprotein that has been implicated in various developmental processes [51. With a urodele tenascin-specific monoclonal antibody (mAb MTl), we found that immunoreactivity appears in the area of dedifferentiation already at 5 days after amputation and increases thereafter. Material immunoreactive to mAb MT1 is also present within cells of the wound epithelium in the pre-blastema and blastema stages. During blastema stages, an abundance oftenascin is seen in the blastema and in a thick layer under the wound epithelium. Tenascin persists to late digit stages of regeneration [13]. Based on this temporal and spatial appearance, we postulated an involvement oftenascin in mesodermal cell dedifferentiation, migration, and proliferation, and/or epithelial-mesenchymal interactions [13; see 19, 221. A role for tenascin in newt tail regeneration has also been suggested [2]. During a continuing immunological search for additional macromolecules of developmental significance to regeneration, we obtained another monoclonal antibody (mAb MT2, matrix 2) that is reactive to blastema ECM. With mAb MT2, we initiated investigations to determine

134

the temporal and spatial appearance of the MT2 antigen during regeneration and to biochemically characterize this matrix component. The results show that the MT2 antigen is a novel, large, polydispersed protein with subunits connected by disulfide bonds and which contains covalently attached units of chondroitin sulfate-like polysaccharide. Interestingly, the MT2 antigen exhibits considerable colocalization with tenascin in both the regenerate and the unamputated limb.

Methods General. Adult newts (Notophthalmus viridescens) were collected from ponds in Ohio. Care, feeding, and amputation procedures have been described [9. 121. All surgical procedures were carried out while newts were anesthetized with MS222 (methane sulfonate, Sigma Chemical, St. Louis, Mo. USA). To obtain regenerates, forelimbs were amputated through the mid-radius/ulna and allowed to proceed to pre-blastema, blastema, and differentiation stages of regeneration. Both stylopodium and zeugopodium regions were sampled from unamputated forelimbs. Antibody. mAb MT2 (matrix 2) was obtained by immunizing mice against homogenates of regenerating eye tissues of adult newts. The immunization protocol and survey method for putative mAbs toward regenerating eye and forelimb tissues of the newt were similar to those described for mAb WE3 [21]. The preparation of mAb MT2 from hybridoma cell culture medium was as described for mAb MT1 [13]. mAb MT2 is an IgGl antibody (Calbiochem Isotyping kit, Calbiochem Corporation, San Diego, Calif., USA). The distribution of the MT2 antigen in the normal and regenerating newt eye is being investigated (Yang and Tassava, unpublished). mAb MT1 (anti-urodele tenascin Ab) was obtained as described by Onda et al. [13]. A rabbit polyclonal antibody against chicken tenascin (pAbTN) was a kind gift from Dr. Harold Erickson, Duke University.

Immunohistochemistry. For indirect immunofluorescence, tissues were frozen in OCT compound in a slurry of dry ice and isopropyl alcohol and sectioned at 10 pm with a cryostat as described [21]. lmmunofluorescence methods were identical to those described for the study of mAb MT1 [I31 except that the secondary antibody used to detect the MT2 antigen was rhodamine-conjugated goat anti-mouse IgG (Cappel, Durham, N.C., USA). In some cases adjacent sections were incubated with mAb MTl (anti-urodele tenascin) as a comparison. Pre-blastema, blastema, and differentiating stages of regeneration were sampled for mAb MT2 reactivity. To ascertain the degree of co-localization of the MT2 and MT1 antigens, we employed a double-labeling procedurc as follows: cryostat-cut sections of mid-bud and late-bud newt blastemas were incubated with a 1:10 dilution of mAb MT2 for 45 min and washed in phosphate buffered saline (PBS) with 0.05% Triton X 100 and PBS without Triton X 100 [13]. Sections were then incubated with fluorescein isothiocyanate (F1TC)-labeled goat anti-mouse IgG (Cappel, Durham, N.C., USA) for 45 min and washed as above. To saturate all Ab binding sites on the goat anti-mouse IgG already attached to sections, an incubation in full-strength mAb MT2 was carried out for 45 min. Sections were then incubated with a 1 : 10 dilution of mAb MT1 for 60 min, washed, incubated with rhodamine-labeled goat anti-mouse IgM, washed and mounted in glycerol mounting medium. Sections were viewed with the appropriate green (FITC) and blue (rhodamine) filter combinations. As controls, additional blastema sections were incubated with various combinations of primary and secondary antibodies. The results from these controls showed that the above double-labeling procedure resulted in specific labeling of MT2 (FITC) and MT1 (rhodamine) on the same section.

Because the reactivities of mAbs MT2 and MT1 were co-localized, it was of interest to determine whether the MT2 antigen was a form of tenascin. Competition experiments on tissue sections were designed to test the ability of pAbTN to inhibit binding of mAb MT2 to sections of regeneration limbs. Sections of newt midbud blastemas were incubated with pAbTN either undiluted or diluted 1 :10 in PBS for 1 h at room temperature after which mAb MT2 (diluted 1 :50 or 1 :500) was added to the same sections with no intervening washes. These two dilutions were used to ascertain the degree of competition between mAb MT2 and pAbTN. If mAb MT2 showed strong reactivity even at the highest dilution (1 : 500), then it could be concluded that pAbTN was not competitive. On separate sections mAb MT1 was utilized as a control [13]. After another 1 h incubation and subsequent washes, secondary Abs specific for mAb MT2 (rhodamine-conjugated goat anti-mouse IgG) or mAb MT1 (rhodamine-conjugated goat anti-mouse IgM) were applied and the sections examined for immunofluorescence. Extraction of the M T 2 antigen from regenerating newt limbs. To biochemically characterize this matrix component, the MT2 antigen was extracted from regenerating newt limbs using a modification of the procedure used to extract urodele tenascin [13]. For each extraction, mid-bud bkdstemds from 40 regenerating limbs (approximately 75 mg of tissue) were collected on ice and immediately homogenized at 4" C in a buffer containing 50 mM diethylamine (pH 11.5), 2 p M leupeptin, 2 m M phenylmethylsulphonyl fluoride (PMSF), and 10 mM EDTA [13]. The homogenate was centrifuged at 15000g for 15 min and the supernatant was dialyzed against 2000 volumes of PBS, 10 m M EDTA, overnight at 4' C. This preparation was designated as the crude extract. Immunoprecipitation. To obtain a purified MT2 antigen preparation, irnmunoprecipitations were performed. The crude extract (25 to 100 pl) was incubated with two volumes of a 1 :50 dilution of the MT2 antibody and one volume of a 50% suspension of Protein G/Agarose beads (Gibco BRL, Gaithersburg, Md., USA). After centrifugation, the pellet of beads was washed three times with 1 ml of PBS, 10 mM EDTA. The bound MT2 antigen was then eluted from the beads with 2 x sodium dodecylsulfate (SDS) electrophoresis sample buffer [lo] and examined in Western blots (see below). Enzyme treatments. To obtain information relevant to the biochemical nature of the MT2 antigen, preparations of the crude extract (25-50 pg of protein) were incubated with the following enzymes under the appropriate conditions: (a) chondroitinase ABC (Sigma Chemical, St. Louis, Mo., USA), 0.01 to 0.05 units, at 37O C in 0.05 M TRIS/acetate, pH 7.2, for 4 to 6 h, (b) proteinase K (Boehringer Mannheim, Indianapolis, Ind., USA), 10 pg, at 37" C in 0.05 M TRIS/acetate, pH 7.2, for 1 h, (c) heparinase (Sigma), 0.25 units, at room temperature in 0.05 M TRIS/acetate, pH 7.2, for 7 h, (d) heparitinase (Sigma), 0.29 units, at room temperature in 0.05 M TRIS/acetate, pH 7.2, for 7 h, (e) testicular hyaluronidase (Sigma), 27 units, at room temperature in 0.1 M acetateINaC1, pH 5.0, for 7 h. Absorption of the M T 2 substance onto DEAE-Sephacell. DEAESephacell binding studies were employed to test whether the MT2 antigen is a highly ionic glycoprotein. Andres et al. [l] reported that sulfated glycoproteins bind to DEAE-Sephacell in the presence of 0.3 M NaCI, while other less anionic polymers are eluted from DEAE-Sephacell by this concentration of NaC1. The crude extract (100 to 300 pl) was made 0.15 M in NaCl and then incubated with two volumes of DEAE-Sephacell which had been equilibrated with 0.01 M phosphate buffer containing 0.15 M NaCI. The DEAESephacell was centrifuged and washed with 0.02 M phosphate buffer, 0.3 M NaCI. MT2 antigen that remained bound to the DEAE-Sephacell in the presence of 0.3 M NaCl was eluted by incubating the DEAE-Sephacell with 0.1 M phosphate, 1.5 M NaCI.

135

Irnmunoblotting. Separately, the crude extract, immunoprecipitated material, DEAE-Sephacell eluates, and enzyme treated pieparations were mixed with an equal volume of 2 x SDS electrophoresis sample buffer with or without P-mercaptoethanol [ l o ] , and separated on 5% discontinuous SDS-polyacrylamide gels with either 2.5 or 3% stacking gels [lo]. After electrotransfer of the separated proteins from the gel to 0.45 pm nitrocellulose membranes [24], the membranes were blocked with a 5% solution of dried milk in TRIS-buffered saline (TBS) for 30 min at room temperature. The MT2 antigen and urodele tenascin on the nitrocellulose membranes were detected as follows: (a) The membranes were first incubated in either a 1 :100 dilution of mAb MT2 in dried milk or a 1:500 dilution of pAbTN overnight at room temperature. (b) The membranes were washed three times in TBS plus 0.05% Tween 20 and once in TBS. (c) The membranes were then incubated in a 1 :2000 dilution of goat anti-mouse IgG conjugated to alkaline phosphatase, goat anti-rabbit IgG alkaline phosphatase conjugate, or a 1 :500 dilution of goat anti-rabbit horse-radish peroxidase conjugate. (d) After washing as in step (b), the immunoreactive bands were visualized by incubating the membranes with the alkaline phosphatase substrates, nitro-blue tetrazolium and 5-bromo-4chloro-3-indolylphosphate,or the horse radish peroxidase substrates, 4-chloro-1-naphthol and H,O,.

the wound euithelium and within the ECM of the blastema mesenchime (Fig. 1 a). The MT2 substance appeared in the distal stump during preblastema stages and persisted until late digit stages of regeneration (Table 1). As seen with tenascin [13], the MT2 antigen was also present within cells of the wound epithelium during preblastema stages (Table 1). Of interest was the similar distribution of the MT2 and MTI antigens at the mid-bud stage of blastema development. A high degree of co-localization of these matrix antigens is indicated upon examination of adjacent sections of a blastema reacted to these Abs (Fig. 1). The distribution of the MT2 substance during all stages of regeneration is thus very similar to the distribution of urodele tenasacin (MTI antigen) (Table 1 ; also see P 31). Distribution of the MT2 antigen in unamputated limbs

During the blastema stage of limb regeneration, mAb MTZreacting material was present in a layer beneath

The distribution of the MT2 antigen in unamputated limbs also follows closely that of urodele tenascin. Both mAbs MT2 and MTI reacted to tendons. myotendinous junctions and periosteum (Fig. 2; Table 1). However, tenascin is routinely present as granules within epidermis and is restricted to gland-epidermis junctions [ 131, whereas the MT2 antigen is not within the epidermis but is present in a layer beneath the epidermis (Fig. 2; Table 1).

Fig. 1. Montages of fluorescence micrographs of adjacent sections of an adult newt mid-bud stage blastema reacted to mAbs MT2 (a) and MTI (b). The amputation level is indicated by the distal gland (g). A nerve (n)can be seen entering the blastema. The distributions of the MT2 and MTl antigens match quite closely, indicating a considerable degree of co-localization. Both antigens are pres-

ent in a layer beneath the wound epithelium and within the extracellular matrix of the blastema. The proximal border of mAb MT1 reactivity (b) is somewhat sharper (compare a). The somewhat lower intensity of mAb MT2 reactivity (a) compared to that of mAb MTl (b) is not consistent and is due to Ab concentration differences. Bar, 200 pm

Results Expression of the MT2 antigen during newt limb regeneration

136 Table 1. Distributions and rela.tive amounts' of Ithe MT2 and MTI antigens in unamputa ted itnd I:ege nerating acjult ne:wt linibs -MTI MT2 Antigen Antigen

Unamputated limbs2 Epidermis Epidermis/ dermis border Epidermis/ gland border Myotendinous junctions Tendons Periosteum Amputated limbs Prehlaslema: Wound epithelium Acellular layer under wound epit heli um Mesenchyme ECM Blastema : Wound epithelium Acellular layer under wound epithelium Mesenchyme ECM Palette stage: Mesenchyme ECM Digit Stage: Mesenchyme ECM (a) Proximal (b) Distal

+4-

I

++ ++

++ ++ ++

++ ++ ++ +++ +++ +++

+ ++ ++ + +++ +++ +++

+ ++

+ ++

++

Fig. 2. A fluorescence micrograph of a portion of an adult newt forelimb in cross section, illustrating mAb MT2 reactivity in tendon (smull arrows), periosteum (large arrows) and in a layer underneath the epidermis (e). No reactivity is within the epidermis. m, muscle; g, gland. Bur, 200 pm

+,

+ + +, strong Ab reactivity; weak Ab reactivity. Strength of antibody reactivity was determined subjectively by three unbiased observers Tissues not listed were not reactive to either mAb MTl or MT2

Double-labeling with mAbs MT2 and M T i

It is clear from the examination of mAb MT2 and mAb MT1 reactivities on adjacent sections that the spatial distributions of the respective antigens are similar in both regenerating and unamputated limbs. Double-labeling of the same section with both mAb MT2 and mAb MT1 showed that there was a high degree of colocalization of the respective antigens, particularly in the distal, undifferentiated region of blastemas. It can be seen from Fig. 3a and b that the reactivity patterns of mAbs MT2 and MT1 are essentially identical in the distal-most region of late-bud blastema. In the proximal regions of mid-bud and late-bud blastemas, it appeared that mAb MT1 reactivity was lost prior to the loss of mAb MT2 reactivity (see Fig. 1a, b). Double-labeling often showed mAb MT2 reactivity in the distal stump while mAb MT1 reactivity was nearly or completely absent (not shown). A more extensive study of urodele limb regeneration with double-labeling, using mAbs MT2 and MT1, is in progress.

Fig. 3. Fluorescence micrographs showing double-labeling of the distal portion of a newt forelimb late bud stage blastema with mAbs MT2 and MTI. The reactivity patterns of mAb MT2 (a: FITC) and mAb MT1 (b: rhodamine) are essentially identied (see orrows), illustrating a high degree of co-localization of the MT2 and MT1 antigens. Scale bar, 15 pm

137

Table 2. Competition between rabbit anti-chicken tenascin Ab and mAbs MT2 and MT1 for antigens in newt blastemas

mAb MT2

pAb TN

Amount of fluorescence due to mAb MT2

1 :50 Dilution 1 :50 Dilution 1:50 Dilution 1:500 Dilution 1 : 500 Dilution 1 :50 Dilution

None 1 :10 Dilution Undiluted None 1 : 10 Dilution Undiluted

++++I

++++ ++++ +++ +++ +++

mAb MTI

pAb TN

Amount of fluorescence due to mAb MTl

1:10 Dilution

1:10 Dilution

-

a

b 1

1

2

2

205 116 97

1169 7-

Fig. 5. Immunoblot analysis of blastema crude extract. Crude extracts of newt blastema were mixed with 2 x S D S sample buffer with P-mercaptoethanol (lane I) or without ,&mercaptoethanol (fane 2). After gel electrophoresis, proteins were transferred to nitrocellulose and stained with pAbTN (a) or mAb MT2 (b). Positions of molecular weight standards (in kilodaltons) are shown in the left margins

Strength of antibody reactivity was determined subjectively by two unbiased observers

1

2

20511697-

Fig. 4. Immunoprecipitation of the MT2 antigen from newt blastema extract using mAb MT2. Immunoprecipitated proteins were suspended in 2 x sodium dodecyl sulfate (SDS) sample buffer with 8-mercaptoethanol and exposed to gel electrophoresis. After transfer to nitrocellulose, the proteins were stained with mAb MT2 followed by goat anti-mouse IgG alkaline phosphatase conjugate (lane I) or pAbTN followed by goat anti-rabbit IgG horse radish peroxidase conjugate (fane 2). Positions of molecular weight standards (in kilodaltons) are shown at the left

pAbTN was not reactive to the purified MT2 substance (Fig. 4). In order to determine the molecular nature of the MT2 antigen and compare it to tenascin, an extract that contained both tenascin and MT2 antigen was examined by immunoblot analysis. Non-reduced forms of tenascin and MT2 antigen were similar in relative molecular weights, and both unreduced molecules were composed of disulfide-bound subunits (Fig. 5). However, under reducing conditions, the components of the MT2 antigen and urodele tenascin did not have the same size. The major subunit of tenascin detected by pAbTN had a molecular weight (Mr) of 205 x lo3 (Fig. 5a) whereas mAb MT2 detected three components in Western blots, (a) the largest species with Mr of 31CL325 x lo3, (b) an intermediate size species with Mr 285-300 x lo3, and (c) the smallest component with Mr of 2 5 0 - 2 7 5 ~ 1 0 ~ (Fig. 5b). These lines of evidence show that urodele tenascin and the MT2 substance are immunologically unrelated. Biochemical characterization of the MT2 untigen

Immunological and biochemical relationship of the MT2 antigen with tenascin

The similar tissue distribution suggested that the MT2 antigen is a form of urodele tenascin. To test this possibility, we performed an antibody competition experiment on newt blastema tissue sections between pAbTN and mAb MT2. Undiluted pAbTN did not reduce mAb MT2 immunoreactivity even when the mAb was present in a 1 :500 dilution (Table 2). In the control, pAbTN diluted 1 :10 completely blocked all binding to sections by a 1:10 dilution of anti-urodele tenascin monoclonal antibody (mAb MT1) (Table 2; also see [13]). Furthermore, after the MT2 antigen was purified from crude extracts by immunoprecipitation with mAb MT2,

Periodate does not destroy the antigenicity of MT2 in tissue sections (data not shown), suggesting that the epitope is peptide sequence. But since many of the known macromolecules of the ECM are glycoproteins, we performed a number of enzyme digestions of the crude extracts, with subsequent immunoblot assays, in an attempt to test whether the MT2 substance is a glycoprotein. Digestion of the crude extract with Proteinase K destroyed MT2 antigenicity (Fig. 6), showing that MT2 is a protein. MT2 antigenicity and MT2 size distribution were maintained after treatment of the crude extract with heparinase and heparitinase, suggesting that the MT2 antigen does not contain heparin or heparan sulfate-like polysaccharides (Fig. 6 ) . However, both chondroitinase ABC and testicular hyaluronidase eliminated

138

a

1

2

3

4

1

5

2

3

4

5

2 0511 69 7-

b

1169 7,

1

2

205, 116, 9 7, Fig. 6. lmmunoblot analysis of enzymatic digestions of newt blastema extract. After enzyme treatment, the crude extracts were mixed with 2 x SDS sample buffer with b-mercaptoethanol, and electrophoresed. The proteins were transferred to nitrocellose and reacted with mAb MT2. (a). Carbohydrase digestions. No enzyme digestion (lane I), chondroitinase ABC digestion (lane 2), heparinase digestion (lane 3), testicular hyaluronidase digestion (lane 4), heparitinase digestion (lane 5). The arrow in the right margin marks the MT2 antigen species sensitive to chondroitinase and hyaluronidase. (b) Protease digestion. Proteinase K digestion (lane I) and no enzyme (lane 2). Positions of molecular weight standards (in kilodaltons) are shown in the left margins

the high molecular weight (310-325 x lo3) reduced component of MT2. Since both of these enzymes hydrolyze 8-1,4 glycosidic bonds of N-acetylated amino sugars [16], these results suggest that the largest MT2 component is a chondroitin sulfate- or a hyaluronic acid-like glycoprotein. DEAE-Sephacell binding studies were performed to ascertain the anionic nature of the MT2 antigen. We found that all components of the MT2 substance bound to the anion exchange resin in the presence of 0.15 M NaCl (Fig. 7, lane 2). The less anionic MT2 species, i.e. the two smaller reduced components, were eluted with the buffer containing 0.3 M NaCl (Fig. 7, lane 3). The highly anionic species of MT2, i.e. the largest component, was only eluted from DEAE-Sephacell in the presence of 1.5 M NaCl (Fig. 7, lane 4). Furthermore, chondroitinase digestion of the 1.5 M NaCl eluate eliminated the largest, reduced MT2 component from the Western blot, and increased the amount of the intermediate size, reduced MT2 component (compare lanes 4 and 5 of Fig. 7). Since chondroitin sulfate is highly anionic in nature, these data and the results of the enzyme digestions suggest that the MT2 antigen contains a component, connected to the other subunits via disulfide bonds, of Mr 310-325 x lo3, that is a chondroitin sulfate-like gly-

I

i

Fig. 7. Use of DEAE-Sephacell to determine ionic properties of the MT2 antigen, The following protein preparations were mixed with 2 x SDS sample buffer with 8-mercaptoethanol and, after gel electrophoresis and electrotransfer to nitrocellulose, the proteins were reacted with mAb MT2. Crude extract (lane I), crude extract after treatment with DEAE-Sephacell (lane 2), MT2 antigen eluted from DEAE-Sephacell by 0.3 M NaCl (lane 3), MT2 antigen eluted from DEAE-Sephacell by 1.5 M NaCl (lane 4), and chondroitinase digest of the antigen eluted by 1.5 M NaCl (lane 5). The upper arrow in the right margin marks the highly anionic, chondroitinasesensitive species of the MT2 antigen, and the lower arrow in the right margin marks the major product after chondroitinase treatment. The asterisk in the left margin marks low molecular weight MT2-reactive material. Positions of molecular weight standards are shown in the left margin

coprotein with a core protein of Mr 285-300 x lo3. The chondroitin sulfate-like component gives the MT2 antigen its highly anionic character. Crude blastema extracts, reduced by fl-mercaptoethanol, also contained a third species (Mr, 250-275 x lo3) that reacted with mAb MT2 (Fig. 7, lane 1, band designated by the [*I in the left margin). This material was resistant to digestion with all of the carbohydrases tested (Fig. 6). This smallest MTZreactive material is anionic but not as highly anionic as the putative chondroitin sulfate-like component of the MT2 reduced components.

Discussion The results show that mAb MT2 identifies a molecule that is biochemically and immunologically unrelated to tenascin but which is developmentally regulated during limb regeneration in a fashion similar to that of tenascin. The MT2 antigen is first expressed early in newt limb regeneration concomitant with the start of dedifferentiation and cell cycling [3, 121. The reactivity to mAb MT2 in the newt limb regenerate reaches a peak during blastema growth. Most of the limb blastema extracellular matrix reactivity is present as a fibrous-like material. mAb MT2 also reacts with a thick acellular layer of material under the wound epithelium. During differentiation stages, mAb MT2 reactivity persists in association with the mesenchyme at the tips of the growing digits and to a lesser extent with newly differentiating cartilage. In unamputated limbs, the MT2 antigen co-localizes with tenascin in tendons, myotendinous junctions, and periosteum [13]. Also, mAb MT2 reactivity is generally present in a layer under the epidermis, whereas mAb MTl reactivity is restricted to epidermis-gland junctions ~31.

139

The patterns of reactivity of mAbs MT2 and MTI in the regenerating limbs of Mexican axolotls (Ambystoma mexicanum) are similar to those of newts (Yang and Tassava, unpublished; see also [13]), suggesting relevance to these matrix components across species. mAb MT2 reacts with a large, polydispersed antigen which is composed of subunits connected by disufide bonds. After the MT2 antigen is reduced, mAb MT2 reacts with three different components. The largest of these components (Mr, 310-325 x lo3) has a very high affinity for DEAE-Sephacell, implying that it contains sulfated polysaccharide [I]. This largest reduced component is also digested by testicular hyaluronidase and chondroitinase but not by heparinase or heparatinase. This pattern of carbohydrase reactivities implies that the polysaccharide moiety of the largest MT2 glycoprotein component is chondroitin sulfate [16]. The intermediate size, reduced component is carbohydrase-insensitive and less anionic than the largest component, and can be enhanced by digesting the larger component with chondroitinase (see Fig. 7). This implies that the intermediate component (Mr, 285-300 x lo3) is the core protein of the chondroitin sulfate glycoprotein. The smallest reduced component (Mr, 250-275 x lo3) is also carbohydrase-insensitive and less anionic than the largest component. The smallest component may be a degradation product of the two larger units. Alternatively, it may be a distinct subunit of the polymeric MT2 antigen with little structural relationship to the two larger reduced components. Furthermore, we cannot eliminate the possibility that all three of the MT2 reduced components may be subunits of polymeric proteins that possess other subunits that have not been detected because they do not contain the MT2 epitope. We hope to raise polyclonal antibodies against the unreduced MT2 antigen to further study the subunit structure of this substance. It is clear from examination of mAb MT2 and mAb MT1 reactivities on adjacent sections of both unamputated and regenerating limbs that the MT2 antigen and tenascin (MT1 antigen) are distributed in a similar fashion. The co-distribution is less evident in the proximal regions of the regenerate. Double-labeling with both mAbs on the same section showed that the reactivity of mAb MT2 (FITC) and mAb MT1 (rhodamine) were essentially identical in the distal, undifferentiated regions of the mid-bud and late-bud blastemas, illustrating a high degree of co-localization of the MT2 antigen and tenascin. In terms of proportional amounts of the two antigens in a given region of the blastema, immunofluorescence observations are not conclusive. At this point we can only speculate as to the relevance of this close association of these two matrix molecules, each over one million molecular weight. Proteoglycans can act as cell surface adhesion molecules for ECM macromolecules [171. For example, embryonic chicken muscle has a large chondroitin sulfate proteoglycan that adheres tightly to tenascin [4]. Also, neurons of the embryonic chick brain contain a surface chondroitin sulfate proteoglycan that binds tightly to tenascin [8]. The embryonic neuron ligand is of special interest because the

285 x lo3 size of its core protein [8] closely approximates the size of the MT2 core protein. We hope to raise polyclonal antibodies against the MT2 core protein to determine if there is immunological cross reactivity between the newt MT2 antigen and these two chicken tenascin ligands. If the MT2 substance is a tenascin ligand (or a ligand for any of the other ECM components important in regeneration, for example, fibronectin, laminin, collagen or hyaluronic acid), information concerning these ligand interactions may lead to further knowledge of the role of ECM in this developing system. Our laboratory has completed a series of in situ hybridization studies to locate those cells within the limb regenerate that are transcribing the tenascin mRNA [14]. Tenascin gene expression is seen in cells of the distal stump and subsequently in blastema cells; basal cells of the wound epithelium also express the tenascin gene [14], consistent with mAb MTI-reactivity seen in these cells [13]. Based upon mAb MTZreactivity patterns, we hypothesize that, as in the case of tenascin, both mesenchyma1 and wound epithelial cells synthesize the MT2 antigen. Also, double-labeling experiments suggest that in proximal regions of late-bud blastemas the reactivity of mAb MTI is lost prior to the loss of mAb MT2-reactivity (data not shown). Thus, it can be predicted that in late regenerates transcription of the tenascin gene is terminated while transcription of the MT2 gene persists. We have obtained two cDNAs for MT2, thus we plan to perform in situ hybridization studies to locate the cells that are expressing the MT2 gene. With data on temporal and spatial MT2 gene transcription, we will be better able to understand the temporal and spatial relationships of tenascin and the MT2 antigen in a developmental context. Because the MT2 antigen first appears when distal stump cells enter the cell cycle and then remains abundant in the blastema ECM during periods of active cell cycling, it is conceivable that this putative chondroitin sulfate glycoprotein is a modulator of one or more peptide growth factors involved in limb regeneration. Other chondroitin sulfate containing glycoproteins act as modulators of growth factors [18]. One such chondroitin sulfate proteoglycan, betaglycan, binds transforming growth factor-/? (TGF-/?) [l]. The proposed role of betaglycan is to sequester TGF-/? for delivery to TGF-B signal transducing receptors [l]. TGF-/? also binds to another chondroitin sulfate proteoglycan, decorin [25]. Decorin, which is found in the ECM, inhibits the activities of TGF-8 by competing with the cell surface receptors for TGF-B [25]. By analysis of MT2 cDNAs we can determine if the MT2 protein contains domains in common with the known sequences of betaglycan, decorin, or other known growth factor ligands.

Acknowledgements. Supported by Office of Naval Research Grant NO0014 and NTH grant H022024 to RAT. Most of this research was carried out at The Ohio State University while KPK was on research leave from Denison University.

140

References 1. Andres JL, Stanley K, Cheifetz S, Massague J (1989) Membrane-anchored and soluble forms of betaglycan, a polymorphic proteoglycan that binds transforming growth factor$. J Cell Biol 109:3137-3145 2. Arsanto JP, Diano M, Thouveny Y, Thiery JP, Levi G (1990) Patterns of tenascin expression during tail regeneration of the amphibian urodele Pleurodeles waltl. Development 109: 177188 3. Chalkley DT (1954) A quantitative histological analysis of forelimb regeneration in Triturus viridescens. J Morphol94: 21-70 4. Chiquet M, Fambrough DM (1984) Chick myotendinous antigen: 11. A novel extracellular glycoprotein complex consisting of large disulfide-linked subunits. J Cell Biol98: 1937-1946 5 . Chiquet-Ehrismann R, Mackie EJ, Pearson CA, Sakakura T (1986) Tenascin: an extracellular matrix protein involved in tissue interactions during fetal development and oncogenesis. Cell 47:131-139 6. Grillo HC, LaPiire CM, Dresden MH, Gross J (1968) Collagenolytic activity in regenerating forelimbs of adult newt (Triturus uiridescens).Dev Bid 17:571-583 7. Gulati AK, Zalewski AA, Reddi A (1983) An immunofluorescent study of the distribution of fibronectin and laminin during limb regeneration in the adult newt. Dev Biol96: 355-365 8. Hoffman S, Edelman GM (1987) A proteoglycan with HNK-1 antigenic determinants is a neuron-associated ligand for cytotactin. Proc Natl Acad Sci USA 84:2523-2527 9. Kelly DJ, Tassava RA (1973) Cell division and ribonucleic acid synthesis during the initiation of limb regeneration in larval axolotls (Ambystoma mexicanum). J Exp Zoo1 185:45-54 10. Laemmli UK (1970) Cleavage of structural proteins during the assembly of bacteriophage T4. Nature 227 :680-689 11. Mescher AL, Munaim SI (1986) Changes in the extracellular matrix and glycosaminoglycan synthesis during the initiation of regeneration in adult newt forelimbs. Anat Rec 214:424431 12. Mescher AL, Tassava RA (1975) Denervation effects on DNA replication and mitosis during the initiation of limb regeneration in adult newts. Dev Biol44: 187-197 13. Onda H, Goldhammer DJ, Tassava RA (1990) An extracellular matrix molecule of newt and axolotl regenerating limb blastema

and embryonic limb buds: immunological relationships of MTl antigen with tenascin. Development 108:657-668 14. Onda H, Poulin ML, Tassava RA, Chiu IM (1991) Characterization of a newt tenascin cDNA and localization of tenascin mRNA during newt limb regeneration by in situ hybridization. Dev Biol 148:219-232 15. Repesh LA, Fitzgerald TJ, Furcht LT (1982) Changes in the distribution of fibronectin during limb regeneration of newts using cytohistochemistry. Differentiation 22: 229-240 16. Rodtn L (1980) Structure and metabolism of the connective tissue proteoglycans. In : Lennarz WJ (ed) The biochemistry of glycoproteins and proteoglycans, Plenum Press, New York, pp 267-371 17. Ruoslahti E (1989) Proteoglycans in cell regulation. J Biol Chem 264:13369-13372 18. Ruoslahti E, Yamaguchi Y (1991) Proteoglycans as modulators of growth factor activities. Cell 64:867-869 19. Singer M, Salpeter MM (1961) Regeneration in vertebrates: the role of wound epithelium. In: Zarrow MX (ed) Growth in living systems. Basic Brooks, New York, pp 277-31 1 20. Tassava RA, Olsen CL (1985) Neurotrophic influences on cellular proliferation in urodele limb regeneration: in vivo experiments. In: Sicard RE (ed) Regulation of vertebrate limb regeneration. Oxford University Press, Oxford, pp 81-92 21. Tassava RA, Johnson-Wint B, Gross J (1986) Regenerate epithelium and skin glands of the adult newt react to the same monoclonal antibody. J Exp Zoo1 239 :229-240 22. Thornton CS (1968) Amphibian limb regeneration. Advances in Morphogenesis 7: 205-249 23. Toole BP, Gross J (1971) The extracellular matrix of the regenerating newt limb: synthesis and removal of hyaluronate prior to differentiation. Dev Biol25:57-77 24. Towbin H, Staehelin T, Gordon J (1979) Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some application. Proc Natl Acad Sci USA 76:4350-4354 25. Yamaguchi Y, Mann DM, Ruoslahti E (1990) Negative regulation of transforming growth factor-,9 by the proteoglycan, decorin. Nature 346:281-284