<<

Till:: JOURNAL Ot" I NVf~ST I GAT I VE DEilMATOLOCY Vo l. 59, No. 3 Copy ri ght © 1972 by The \Vill inms & \Vilkins Cu. Printed in U.S ..4..

STUDIES ON THE CHEMISTRY AND FINE STRUCTURE OF ELASTIC FIBERS FROM NORMAL ADULT * DAVID P. VARADI, M.D., F.R. C. P. (C)

ABS TRACT Collagenase-prepared, purified elastic fibers from adult huma n skin consist of at least two morphologically distinct components, the unstained a morphous component, and inter­ nally-located, deeply staining microfibrils. Separation a nd isolation of the amorphous component was achieved by alka li preparation of the elastic fibers. E lectron micrographs of a lkali prepared fibers showed essentia lly a morphous componen t as most of the microfibrils had been solubilized. The a mino acid composition of the amorphous component was that of classical . The microfibrillar component embedded deep wi thin the interstices of the fiber was sepa rated by, first, re movin g the a morphous component with elastase, a nd subse­ quently solubilizing, with dithioerythritol (DTE), the cystine-containing microfibrils con­ tained in the elastase-produced residue. The composition of this DTE-ex­ t racted microfibrilla r material was similar to that of the peripheral microfibrillar compo­ nent enzymatically removed from bovine fetal elastic fibers. Electron microscopi c moni ­ toring showed that only stained microfibrils remained after elastase digestion. A protein portion of the microfibrilla r component was not solubilized by the DTE and was presumed to contain a low concentrat ion of bonds. Studies on purified elastin (amorphous component) revealed that a lanine is concentrated around desmosine crosslinks and that the pyrollidines a re uniformly distributed a long t he elastin molecule thereby precluding a a-helix conformation. Dark fi eld electron mi croscopy suggested that the desmos ine cross­ link region is dumbell -s haped and t hat t he desmosine crosslinks a re not equidistant from each other on the peptide chains.

It has been established by electron microscopy cated microfibrils in large numbers. This investi­ that mature elastic fibers, after staining with cati­ gation reports on a) the separation of microfibril onic lead and uranyl acetate, consist of a centra l from a morphous component, b) the chemical na­ non-staining amorphous core surrounded by ture of adult microfibrils in relation to those iso­ stained tubular appearing microfibrils (1, 2) . In a lated from feta l elastic fibers, c) the distribution recent study on collagenase-prepared elastic fi­ of some of t he amino acids in alkali -prepared bers from fetal bovine li gamentum nuchae, the elastin (amorphous co mponent) and d) the fine central amorphous component was separated structure of acid-solubilized elastin examined by from the periphera l microfibrillar component by da rk fi eld electron microscopy. solubilizing t he cystine-containing microfibrils with the reducing compound, dithioerythri tol (3). MATERIALS AND METHODS Partial characterization of the separated compo­ nents showed that the amorphous component was Isolation and purification of elastic fib ers. A non -hy­ t he desmosine-containing protein, elastin, while drolyti c collage nase meth od of Miller et al. (4) was used the microfibrils consisted of protein(s) with a n to isolate elastic fib ers from norm al adul t human skin. Necropsy specimens of skin we re obtained from the a mino acid composition quite different from that upper, outer thi gh of male ca davers aged forty to sev­ of elastin. enty-two years . The spec imens obta ined from three di f­ In the present study on elastic fibers from ferent cadavers were processed se parately and were not norma l adult human skin, it was noted that elec­ pooled at any stage . Fat and epiderm is were mechani­ tron micrographs of such elastic fibers showed ca lly removed from the skin. The remaining dermis, cut essentially no peripheral but only internally-lo- into small pieces, was fin ely min ced in 3% Na 2 HPO, with a Sorva ll Omni-Mixer. Treatment with 25% KCI , 5 Received Jan uary 7, 1972; accepted for publication M ga unidine HCI and collagenase was ca rri ed out ac­ May 2, 1972. co rding to the method of Miller et a.l. (4). This work was supported by the Medi cal Research A portion of the above prepared elastic fib ers was Coun cil of Canada (MA-3185) and the Womens Auxi l­ fur ther treated with an alkali method in whi ch the fi­ iary of the Welles ley Hospital. bers were suspended in 0.1 N NaO H at 98 ° C for one Part of thi s wo rk was prese nted at the annual hour. Prior to this treatment the fibers had bee n left in meeting of The American Federation for Clini cal Re ­ acetone and then ether, 24 o C, for 24 hours, respec­ search in Atlantic City on May 3, 1970. An Abstract was tively. pub lished in Clini ca l Research, 18: 352, 1970. * From the Department of Medicine, Section of Der­ Electron mi crosco pi c moni torin g of the collage nase­ matology, Wellesley Hospital, University of Toronto, prepared and alkali -treated elastic fibers showed the Toronto, Ontario, Ca nada. former to contain both mi crofibrils and amorphous 238 CHEMISTRY AND FINE STRUCTURE OF ELASTIC FIBERS 239 component while t he latter consisted of essentially stained, unshadowed macromolecules of biological or­ amorpous co mponent devoid of micro!ibrils. igin . For examination, the samples of solubilized elastin Elastase diges tion. Elastin preparations were in cu­ were dissolved in cold 0.01 M phosphate buffer, pH 7.0, bated with electrophoretica lly purified elasta e (Wor­ at conce ntration of 1 ng/ m I. thington Biochemica l Co rp. ) in 0. 1 M glyci ne buffer (pH Chemical analyses. Am in o acid analyses carried out 7.0) at 37° C for 48 hours. One mg of elastase was used by the Beckman-Spinco Auto analyser were performed fo r every 5 mg of elastin (4). on samples of elastin hydrolyzed in 6 N HCI in vacuo Solubilization of microfibrils fr om adult dermal for seventy-two hours at 106° C. Desmosine and isodes­ elastic fibers. The res idue remaining after elastase mosine, respectively, were quantitatively determined on digestion of co ll agenase-prepared dermal elastic fibers the analyser according to a procedure by Anwar (8). was treated with the reducing agent, di thi oerythritol Microfibrillar m aterial was hydrolyzed 24 hours. (DTE), to so lubilize the microfibrils by reduction of di­ Protein concentrations of the oxali c acid solubilized sulfide bonds. The reduction procedure is that of R oss elastin fractions were determined by the method of and Bornstein (3). On dialysis, so me of the solubilized Lowry using solubilized elastin from bovine li gamentum material prec ipi tated out. The still solubilized portion nuchae as the standard. of microfibrillar co mponent was des ignated the 'super­ natant'. RESULTS Preparation of soluble e lastin fra ctions. Soluble elastin was prod uced by partial acid hydrolys is of alkali­ Chemical Studies prepared elastin in boiling 0.25 M oxali c acid solution T he amino acid com posit ion of hot alkali ex­ according to the method of Partridge, Dav id and Adair tracted elastin from the t high skin of a seventy­ (5). After treatin g t he elastin with the oxali c acid solu­ year-old ma n i s s hown in Table I. T h tion in a boiling water bath for a half-hour, t he so lution e composi­ was removed and water washings of the remaining inso l­ t ion is characteristic for elastin. An a lmost iden­ uble elastin added to it. This was des ignated E xtract 1. tical composition was obtained o n elastin isolated Fresh oxali c acid solu tion was added to the res idual in ­ from t he s kin of a forty-three-year-old and s ixty­ soluble elastin and boiling was carried out for one hour. year-old subject, respectively. In the same table Extract 2 was then removed. Subsequently, every hour a re g iven the a mino acid compositions of the 6 for fou r m ore hours the extracts with their washings fractions extracted with oxalic acid solut ion from were removed and fresh solution added. After 5 1!2 hours insolubile elastin. There is a steady increase in the insoluble elastin had dissolved to fo rm 6 extracts the isodesmosine a nd desmosine concent rations numbered 1 to 6. Extracts 4 and 5 which had a hi gh a­ from Extract 1 to 6 while certain amino acids elastin co ntent, were each chromatographed at go C on such as a Sephadex G-100 column equilibrated with 0.05 M , valine, leucine, phenylalanine phosphate buffer containing 8 M urea. The materi al and remain relatively constant. emergin g in the vo id volume, the a-elastin, was selected shows a stead y increase unt il in fraction 6, the for d a rk fi eld electron microscopy. alanine exceeds in concen tration. The The elastin fragments shown in Fi gures 7, 8, 9 were a a nd glycine tended to decrease in gift from Dr. R. A. Anwar. This material, designated amount. The increase in concentration of the Dowex V in a recent publication by Shimada, Bowman, des mosines probably indicates progressive hydro­ Davis and Anwar (6), was obtained by applyin g a n elas­ lytic cleavage of peptide chains with proportion­ tase digest of bov in e li ga mentum nuchae elastin se­ a ll y less extraction of t he desmosine linkage re­ quentially to Sephadex G-25, cellulose phosphate and gions t hereby Dowex 50W co lu mns. The desmosine- ri ch material from leaving an increasin gly more c ross­ one column was fractionated further on the subsequent linked elastin core. Certain amino acids had a column to obtain the final hi gh d es mosine-containing constant concen t ration throughout t he 6 fractions. fraction from the Dowex co lumn designated Dowex V. Thus, proline + hydroxyproline g ive a value of The molecular size of the fragment, determined on a 123 to 127 residues/1000 total residues in each Sepha dex G-50 column, was found to be approx imately fraction thereby s uggesting a uniform distribution 6000. of pyrrolidines in the peptide chains. Ala nine, Light and electron microscopy. Alkali-prepared elas­ from these data see ms to be con cen trated a round tin , of known amino acid co mposition, was fi xed in for­ malin and processed as a histolog ic s pec imen. Staining the desmos ine c rosslinks. The relatively high con­ was w ith Verhoeffs and New Orcein. cen t ration of aspa rt ic acid, , Bright fi eld electron mi croscopy was used to examine and hydroxyproline in Extract 1 (the f raction purified samples of b oth enzy me-prepared and alkali­ with the lowest desmosine content) suggests that prepared elastin. The preparations were fixed for one t here is a small quantity of acidic am ino acid-en­ hour in 2% os mium tetrox ide buffered with s-collidine riched peptides in elastin which may be distant to pH 7.4. They were then dehydrated and embedded in from desmosine crosslinks. These acidic amino epoxy res in . Sections were stained with azure 11-meth­ acids seem unlikely to be derived from contami­ ylene blue for li ght mi croscopy. Thin sections were nating microfibrilla r m aterial as the hydroxypro­ stained first, with ura nly acetate and subsequently, with line is high a nd t he arginine low. No hydroxyly­ lead citrate. The sections were examined in a Philips 300 Electron Mi croscope. s ine was present. Dark field electron mi croscopy, a dark fi eld technique In Table II, t he first column of figures gives t he deve loped b y Ottensmeyer (7) was used to examine sol­ a mino acid composition of the res idue remaining ubilized fragments of elastin. Dark fi eld conditions were after elastase digestion of collagenase-pre pa red achieved by t il tin g the electron beam. This technique elastic fibers from huma n dermis. The residue appears useful for the study of the fine structure of un- constituted 14.7 ± 0.2% of the weight of the 240 THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

TABLE I Amino acid composition of adult dermal elastin and of the oxalic-acid solubilized products obtained from elastin*

Enzyme- Alku li- Oxalic-ncid extracts of alkali -prepltrcd elastin .:f: Amino acid prepared pre pored elastin clos lin 6

Hydroxyproline 7.0 7.2 11.2 7.0 4.9 4.8 5.1 3.3 Aspartic acid 14.4 5.2 12.6 7.2 5.7 4.6 4.2 4.4 T hreonine 16.0 7.0 9.2 7.1 6.9 6.8 7.6 8.2 Serine 14.2 6.3 13 .6 10.3 7.5 7.3 7.7 9.1 Glutamic acid 27.9 18.8 38.1 25 .5 23.2 22.4 22.2 24.0 Proline 103.0 120.0 115.0 117.0 118.0 118.0 120.0 124.0 Glycine 292.0 315.0 306.0 304.0 304.0 293.0 287.0 277.0 Alanine 215.0 249 .0 232.0 252.0 259.0 274.0 277.0 282.0 Cystine/2 5.3 Trace Trace Trace Trace Trace Trace Trace Valine 128.0 120.0 131.0 131.0 131.0 124 119 113 2.4 1.2 0.8 Trace Trace 0.2 0.6 1.2 Isoleucine 28.2 24.7 24.5 23.0 21.4 20.8 20.1 19.0 Leucine 62.8 61.4 50.7 50.9 56.5 55.8 56.0 57.2 23.2 21.6 16.9 22.4 22.4 24.0 27.0 27.3 Phenylalanine 25.8 22.7 21.4 21.5 20.8 20.7 21.0 21.7 Isodesmosine/4 3.7 3.9 1.7 3.8 4.4 5.1 5.7 6. 3 Desmosine/4 5.4 5.7 2.9 5.5 5.8 7.1 8.1 9.1 Lysinonorleucine/2 0.9 0.9 n.c. n.c. n.c. n.c. n.c. n.c. a-Aminoadipic acid§ 2.0 10.4 5.6 7.3 6.8 5.3 5.3 5.5 5.8 1.6 Trace 0.5 0.3 0.2 0.2 0.2 0.5 Arginine 13.2 5.6 5.5 5.3 5.7 5.7 6.3 6.1

* Values are expressed as residues per 1000 residues. No correction was made for losses during hydrolysis. Figures are the average of 2 separate determinations on the sa mple. n.c. indicates ' not calculated'. :j: Note that Extract 1 constituted 4.8% of the total protein extracted; Extract 2, 14.0%; Extract 3, 16.4%; Extract 4, 24.4%; Extract 5, 34.5%; and Extract 6, 5.8%. · §As the a-aminoadipic acid was determined on performic acid-oxidized elastin, the 2 .0 equivalents represent that of a-aminoadipi c-li-semialdehyde. Alkali destroys the semialdehyde and thus cannot be determined on alkali­ prepared elastin.

elastic fibers in the two digestions performed on a cies in amino acid composition between columns given sample. Other samples from subjects aged 2 and 3 are probably due to the different source forty-three and seventy gave similar results. The of the elastic fibers, and to the differences in second column shows the composition of material preparation of the microfibrils. In the case of solubilized by the reducing action of DTE on the adult human skin, elastase had to be used to di­ res idue remaining after elastase digestion of gest away the amorphous component (elastin) as elastic fibers. Approximately 65% of this residue the microfibrils are located mainly within the in­ was solubilized by the DTE. The fact that 35% terstices of the adult e lastic fiber and not periph­ was not solubilized suggested that at least one erall y as in the fetus (see Fig. 2). The non-specific other protein comprised the microfibril and that proteolytic activity of elastase probably partially it probably had a low content of disulfide bonds. digested the microfibrils thereby altering the Both materials have an amino acid composition amino acid composition before extraction with quite different from that of elastin. They contain DTE. In addition, column 3 represents the com­ a much higher concentration of polar (acidic and position of the total microfibrillar component basic) and sulfur containing amino acids, less gly­ (precipitate and supernatant after dialysis) while cine and alanine and no desmosines and hydroxy­ column 2 designates the supernatant fraction proline. The va lues of the second column in Table on ly. Approximately 33% of the DTE solubilized II are quite similar to those of the third column. fraction precipitated during dialysis. The amino The third column represents the composition of acid composition of the precipitate was different the peripheral microfibrils solubilized by chymo­ from that of the supernatant as demonstrated by trypsin from collage nase-prepared elastic fibers of Ross et al. (3) . fetal bovine ligamen tum nuchae. Thus the elas­ tase-produced residue and the DTE extract of it Microscopy probably consist mainly of microfibrils with an Light microscopy. Figure 1 illustrates a stained amino acid composition surprisingly similar to preparation of alkali-prepared, purified dermal that of fetal bovine microfibrils. The discrepan- elastic fibers. Such fibers were the starting mate- CHEMISTRY AND FINE STRUCTURE OF ELASTIC FIBERS 241

TABLE II the fiber. No peripheral microfibrils of the type Amino acid co m position of the microfibrillar component described by Ross and Bornstein are present. from adu.lt dermal elastic fibers * Reduction with DTE does not a lter the e lectron micrographic a ppearance of these elastic fibers. Reduced Residue after , Figure 3 is an dialy Enzymntic electron micrograph of a longitu­ clostnse zed Amino acid digest of digestio n elastase· fetul c dinally oriented, collagenase prepared elastic residue: oif of fibers* microfibrils • supernatant § fiber showing branched, interlacing microfibrils approximately llO J.L in diameter. The beaded Hydroxyproline 0 0 1.7 appearance of the microfibrils is probably due to Aspartic ac id 96.0 86.4 92.5 their wavy cha racter. 68.5 51.0 47.3 Figure 4a shows a collagenase-prepared elastic Serine 78.4 54.4 52.8 fiber with the internally located microfibrils cut Glutamic acid 91.7 93.5 98.3 transversely. Figure 4b demonstrates the appear­ Proline 45.7 91.2 73.5 ance of an prepared with a lkali . Note Glycine 118.0 149.0 142.0 the decrease in the number of interna lly-located Alanine 79.2 113.0 82.6 microfibrils a nd the fuzzy out line a nd reduced Cystine/2 34. 1 34.5 56.3 stainability of the remaining ones. T he sodium Valine 94.9 73.0 69.7 hydroxide t reatment has solubilized most of the Methionine 10.2 6.6 13.0 microfibrillar component. Alkali -preparation of Isoleuci ne 33.6 33.9 43.8 elastic fibers, therefore, y ields essen t ia ll y t he Leucine 78.5 62.6 65.5 amorphous componen t. Tyrosine 42.3 27.9 27.6 Figure 5 is a n electron micrograph of the res­ Phenyl alanine 24.7 26.8 32 .8 idue remaining after elastase treatment of co lla­ Isodesmosine/4 0 0 0 genase-prepared elastic fibers. Only cationic lead­ Desmos ine/4 0 0 0 and urany l acetate-stained microfibrils are Ly sine 25.9 35.2 36.7 present as the amorphous component, the elastin, Histidine 16.5 12.1 11.5 has been solubilized by the e lastase. T he general Arginine 59.3 48.1 42.3 shape of the elastic fiber has been retained but owing to the absence of the e lastin the microfi­ * Values are exp ressed as res idues per 1000 residues. brils are c rowded together. Discrete, round, trans­ Figures are t he ave rage of 2 separate determin ations on versly-cut microfibrils are clearly visible. T he the same sample. a mino acid composition of this substance(s) , the :j: 'Residue' refers to the material remaining after starting material for the DTE extraction, is g iven digestion of collage nase-prepare d, adult dermal elastic in the first column of values in Table II. fibers with pancreatic elastase. Dark field electron microscopy. Extracts 4 a nd § 'S upernatant' refers to mate ri al solubilized by DTE 5 were examined by the dark field technique as treatment o f the residue remaining after elastase diges­ t he high d esmosine a nd relatively low polar tion of co ll agenase-prepared adult dermal elastic fibers. amino acid content indicated a preparation of As some of the DTE-solubilized material prec ipitated solubilized elastin uncontaminated with microfi­ out of solution during dialysis, 'supernatant' designates brillar component. In fact, only t he a-elastin the material remaining in solu tion after dialysis. component of each of these extracts was used for .-These va lues a re from the work of Ross et al . (3). dark field electron microscopy. The molecular The material, soluble peptides, was obtained by diges­ size of the a-elastin, determined on a Sephadex tion with chymotrypsin of coll agenase-prepared elastic G-200 co lumn, was approximately 80,000. fibers from bovine, fetal ligamentum nuchae. Chymo­ Figure 6 s hows a typical fie ld consisting of elec­ trypsin selectively solubilizes the peripheral microfibrils tron dense chains 10 to 25 A thick on which nodes of the elastic fiber. 30 to 40 A in diameter were located. The nodes from which 2 to 3 or occasionally even 4 c hains radiated may represent t he location of desmosine rial for preparation of the soluble e lastin. This crosslinks. The internodal distances varied from photomicrog raph demonstrates that the c hemical 50 to 150 A. Many of these chains a nd nodes were and electron mi croscopic studies tha t followed observed to for m a lattice-like arrangement (Fig. were done o n elastic-staining f ibers corresponding 6). The sizes of the lattice-like structures varied to those seen on histologic examination of the considerably demonstrating t he polydisperse na­ skin. Under the light microscope stained co lla­ ture of a-elastin . The electrophoretically-homo­ genase-prepared elastic f ibers appear similar to geneous e lastase-produced peptide of 6000 molec­ the a lkali-prepared fibers shown in Figure 1. ular size, isolated from b ov ine ligamentum nu­ Bright fi eld electron microscopy. An electron chae, was also examined by this technique. Figure micrograph of co llagenase-prepared purified adult 7 is a photomicrograph of this peptide showing a dermal elastic fibers is illustrated in Figure 2 . dumbell-shaped electron d ense shape. Two elec­ ote the deeply stained microfibrils of transverse, tron dense ci rcles were joined by a short crossbar. longit udinal and oblique c uts lying e mbedded In Figure 8 a nother field , a s imilar structure was within the unstained amorphous component of present. In Figure 9 a t hird field examined, there 242 THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

F'IG . l. (T op) Light micrograph of purified elasti c fib ers isolated from a dult huma n dermis. x 1200 (Verhoeffs Stain). F IG. 2. (Bottom) Elec tron mi crogra ph of coll agenase-prepared , derma l elasti c fib ers. Note the la rge number of dark-s taining mi c rofibrils loca ted interna ll y within the elasti c fib ers but no periperal microfibrils. x 8000. CHEMISTRY AND FINE STRUCTURE OF ELASTIC FIBERS 243

FlG. 3. (Top) Electron microgra ph of a co ll agenase-prepared, dermal elastic fib er l ongitud inall y o ri ented. The beaded. branching nature o f the s tained microfibrils is a ppa rent . x 31,900. FIG. 4. (Bottom) (a) Electron micrograph of a t ransverse section of a co llagenase-prepared elastic fib er. x 27,700. (b) E lectron micrograph of a t ra nsverse section of an alkali -prepa red elastic fib er. Note the marked d ec rease in the num b er and stainability of the microfibrils. x 27.400. 244 THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

FIG. 5. Electron micrograph of the res idue remaining a fter elastase digestion of coll agenase-prepared d ermal elastic fibers. Note the presence of round, t ransversely cut, stain ed microfibrils crowded together and t he absence of unstained amorphous component. x 59,600.

was a g roup of 'dumbell ' structures with no cross­ bars clustered together in the center of the field. Here pairs of electron-dense circles appear tan­ gentia l. In such fragments of elastin, each 6000 molecular weight in size and known to contain desmosine, or isodesmosine there is a high proba­ bility tha t a given fragmen t contains just 1 m ole of a desmos ine. This probability exists because in the source material, insoluble e lastin, there are 10 residues of d esmosine per 1000 amino acid resi­ dues (100,000 molec. wt) . In the fi elds examined, the sizes of these fragments were remarkably uni­ form.

DIS CUSS ION Like fetal bovine elastic fibers, elastic fibers from adult human skin consist of at least two components observable with the electron micro­ FIG. 6. Dark fi eld electron micrograph of solubilized scope. The main component, the unstained amor­ insoluble e lastin (a-elastin) s howing a lattice of elec­ tron -dense chains a nd nodes. The dotted background is phous matrix, was chemically separated from the produced by t he carbon coating on the grid. x 700,000. second component and found to have the amino CHEMISTRY AND FINE STR UCT URE OF ELASTIC FIBERS 245

acid co mposition of classical elastin. T he second co mponent , the stained microfibril, was found to co nstitute, at minimum, 15% of the dry weight of the adult dermal elastic fiber compared with 5- 10% (3) of the fetal bovine elastic fiber. In addi­ tion, the microfibrils in the human adult elastic fiber we re found to be mainly internal within the elastic fiber and not periphera l to it. T he essen­ t ial lack or small quantity of peripheral microfi­ brils probably indicated that there is li ttle de nouo synthesis of elastic fibers in the adult . The internall y- located mi crofibrils are li ke ly, there­ fore, historic markers of previous synthetic ac­ tivity in which the diameter of the elastic fiber was constructed conce nt ri call y much like the Haversia n units in bone. T he diameter of adult elastic fibers is large r than that of feta l fibers and the half- life of elastin in the adult is probably long. It has bee n suggested that the laying down of microfibrils is the initial step in the syn thesis of an elastic fiber (2). T he chemical nature of the microfibrillar compo nent, as determined by the amino acid composition, is similar to that fro m fetal bovine elastic fibers. T his latter finding sug­ gests that the mi crofibril is fo und ge nerally in the co nnective tissues of m ammals. It probably, therefore, rep resents a hi ghly successful evolu ­ tionary feature of ma mmalian co nnective t issue. Ross et al. have suggested that one fun ction of the mi crofibril is to determine the shape of the elastic fiber (3). Alkali -prepared elastic fibers demonstrate a greatly reduce d number of internal microfib rils on electron microscopy (Fig. 4b). T hus, such fi­ bers should be a purer preparation of elastin (amorphous component) than coll agenase-pre­ pared fibers. This finding is supported by t he amino acid composition of alka li prepared elastic fibers in comparison to enzyme prepared ones (Ta­ ble I). T he constancy in a mino acid co mposition of alkali -prepared elastin may be similarly explained. The a lka li -prepa red elastic fibers consisten t ly contain a much lowe r concentration of polar and sulfur-containing amino acids, the ones present in high co nce ntration in the microfibrillar compo­ nent. T hus, there was correlation between the electron microscopic a ppearance of the elastic fibers and their chemical composit ion. It appears now, that this microfibrillar protein (s) , an integral F IG· 7. Dark field electron micrograph of a desmo­ part of the elastic fiber, was the reason fo r the sine-containing 6000 mol. wt. pept ide deri ved from in ­ difficul ty in obtaining purified elastin free of soluble elastin of bov ine origin . Note the electron-dense dumbell shaped structure. x 4,000,000. 'non- contaminant' even with alkali prep­ FIG. 8. Dark field electron micrograph show ing an­ aration. other fi eld on the materi al illustrated in Fig. 7. T wo Treatmen t of collagenase-prepared a dult ring or annular structures tangent ial to each other are dermal elastic fibers with DTE did not alter their demonstrated. x 4,000,000. ap pearance with the electron microscope. T he FIG. 9. Dark field electron microg raph show ing an­ other field of the material in Fig. 7. Note the cluster of in ternall y located microfibrils were as strongly paired ring structures wi th the 2 rings comprisi11g a pair stained and as sha rp as before. In some recent in tangent ial contact with each other. x 4,000,000. unpublished work fro m this laboratory, it was fo und t hat collagenase- prepared elastic fibers fro m human fetal dermis contained a very high 246 THE JOURNAL OF INVESTIGATIVE DERMATOLOGY concentration of polar and sulfur-containing with or without a short crossbar between the amino acids. This suggested a higher percent of rings. This shape was repeatedly observed in the microfibrillar component in human fetal dermal many fields examined. Three such fields (Figs. 7, elastic fibers than in the corresponding adult fi­ 8, 9) are illustrated in the paper. One may specu­ bers. Such a conclusion was supported by the late that the electron dense region between the electron microscopic appearance of 22-week fetal rings represents the desmosine or isodesmosine fibers showing masses of peripheral microfibrils while the rings themselves represent peptide surrounding a small amorphous core that con­ clouds. The width of the circumferential line en­ tained few internal microfibrils. closing a ring correlates we ll with the diameter of Amino acid analyses on the soluble elastin frac­ parallel-oriented, contiguous peptide chains. This tions (solubilized amorphous component) demon­ technique was used recently to verify the helical strated an increasing des mosine content due to structure of DNA filaments and to reveal a U­ the probably artefactual creation of progressively shape for ribonuclease and a clover-leaf form for more highly crosslinked peptides. The greater the transfer RNA (7). enrichment of the peptide with desmosine, the higher was the alanine concentration. Reports by The author wishes to thank Miss Pamela Bourne of Shimuda et al. (6) and Keller et al. (9) have also the Pathology Dept., Princess Margaret Hospital, for her demonstrated alanine enrichment around desmo­ expert assistance with the electron microscopy. This sine crosslinks in mature elastin from bovine liga­ work was carried out on the Canada Life Assurance Company Centennial electron microscope. The author is mentum nuchae. Sandberg et al. (10) have iso­ indebted to Dr. Peter Ottensmeyer, Department of lated relatively small peptides containing 3 and 4 Medical Biophysics, University of Toronto, for the dark moles of alanine, respectively, to 1 mole of lysine, field electron microscopy. from 'tropoelastin' (soluble elastin from copper deficient animals). There is, however, some con­ REFERENCES troversy whether 'tropoelastin' is a precursor of mature elastin. The constant value of t he pyrroli­ 1. Rhodin, J. and Dalhamn, T.: Electron microscopy of dines (proline + hydroxyproline) in the various collagen and elastin in lamina propria of the tra­ cheal mucosa of rat. Exp. Ce ll Res. , 9: 371, 1955. fractions suggests but does not prove a uniform 2. Greenlee, T. K., Jr., Ross, R. and Hartman, J. L.: distribution of this compound throughout the The fine structure of elastic fibers. J . Cell Bioi., elastin protein molecule . Thus approximately 30: 59, 1966. every 8th amino acid is probably a pyrrolidine, a 3. Ross, R. and Bornstein, P.: The elastic fiber. I. The situation which likely precludes a-heli x conforma­ separation and partial characterization of its mac­ romolecular components. J. Cell Bioi., 40: 366. tion. Phenylalanine, lysine and arginine also ap­ 1969. pear to be relatively uniformly distributed along 4. Miller, E. J. and Fullmer, H. M.: Elastin: dimin­ elastin polypeptides. The high concentration of ished reactivity with aldehyde reagents in copper deficiency and lathyrism. J. Exp. Med., 123: 1097, acidic amino acids in very low desmosine cross­ 1966. linked peptides suggests that acid enriched por­ 5. Partridge, S.M., Davis, H. F. and Adair, G. S.: The tions of the polypeptide chains exist in elastin chemistry of connective tissues. II. Soluble pro­ and that they are probably distantly located from teins derived from partial hydrolys is of elastin. Biochem. J ., 61: 11, 1955. the crosslinks. 6. Shimada, W., Bowman, A., Davis, N. R. and Anwar. The fact that a-aminoadipic-o-semialdehyde is R. A.: An approach to the study of the structure present in adult dermal elastic fibers (Table I) of desmosine and isodesmosine containing pep­ suggests that synthesis of elastin is still in prog­ tides isolated from the elastase digest of elastin. Biochem. Biophys. Res. Com ., 37: 191, 1969. ress. It probably indicates de novo synthesis and 7. Ottensmeye r, F. P.: Macromolecular fine structure not merely increased crosslinking of pre-existing by dark fie ld electron microscopy. Biophys. J., 9: elastin. An alternative, however, is that some a­ 1144, 1969. aminoadipic-o-sem ialdehyde residues are not 8. Anwar, R. A.: Comparison of from various sources. Can. J . Biochem. 44: 725, 1966. converted to crosslinks such as the desmosines. 9. Keller, S .. Levi, M. M. and Mandl, 1.: Antigenicity The dark field electron microscopy suggests and chemical composition of an enzymatic digest that the desmosine, if represented by the electron of elastin. Arch. Biochem. Biophys., 132: 565, dense nodes, may not be uniformly distributed 1969. 10. Sandberg, L. B., Weissman, N. and Gray, W. R. : along the peptide chains as the internodal dis­ Structural features of tropoelastin related to the tances vary from 50 to 150 A. The desmosine sites of cross-links in aortic elastin. Biochemistry. linkage region appears to be dumbbell-shaped I 0: 52, 1971. .