Demonstration of a Precursor-Product Relationship between Soluble and Cross-linked Elastin, and the Biosynthesis of the Desmosines in Vitro*

Direct evidence showing that a soluble form of elastin is the precursor of cross-linked elastin was obtained from pulse-chase experiments using chick embryo aortas and by demonstrating the conversion of soluble elastin into cross-linked elastin in a cell-free system. Acetic acid extracts of embryonic chick aortas pulse-labeled with [“Cllysine contain two radioactive proteins of molecular weights 74,000 and 138,000 which have been identified previously as soluble elastin and the pro-o chains of collagen, respectively. In pulse-chase experiments, the radioactivity incorporated into the soluble elastin during the pulse with [“Cllysine disappeared during a 24.hour chase with [%]lysine and 89% of that which disappeared was accounted for in the desmosines of alkali-insoluble elastin. The disappearance of the radioactivity from the soluble fraction and its appearance in the desmosines of elastin were inhibited by fi-aminopropionitrile, a specific inhibitor of the cross-linking enzyme lysyl oxidase. In additional in vitro experiments, it was shown that the radioactivity in the desmosines of elastin can arise only from that present in an acid-soluble precursor protein. This precursor protein is soluble elastin, as demonstrated by the formation of desmosines when

In the experiments illustrated in Fig. 1, of the 1666 cpm present at the end of the pulse period, only 13 and 4% remained after 2 and 6 hours of chase, respectively.
During the same chase period, the 138,000 molecular weight pro-cu collagen chains were converted to N chains of collagen of 100,000 molecular weight within the first 2 hours of chase,3 and 52% of the counts had been lost, presumably by cross-linking, by the end of the 6-hour chase.
Preparation of Alkali-rnsoluble Elastin-After incubation, the aortas were immediately transferred to acetone at 4" and det'atted by successively stirring for 24 hours i n a large excess of acetone and ether at 4'. They were then extracted i n acetic acid as described above.  A, 30-min pulse; B, 2-hour chase; C, 4hour chase; and D, &hour chase. The radioactivity under the soluble elastin peaks was 1666 cpm (loo%), 220 cpm (13%), 70 cpm (4%)), and 62 cpm (4%), respectively, after the pulse and 2, 4, and 6 hours of chase, For the corresponding periods, the collagen (procollagen) peaks had 787 cpm, 706 cpm, 484 cpm, and 379 cpm, respectively. FIG. 2 (center). Sodium dodecyl sulfate-polyacrylamide gel electrophoretic patterns of acid extracts of aortas incubated in the presence of P-aminopropionitrile.
The extracts were treated and equal amounts and if the radioactivity is equally distributed between desmosine and isodesmosine. Therefore, the following transfer experiments were done.
Aortas were pulse-labeled for 15 min followed by cold chase for varying periods of time in the presence or absence of P-aminopropionitrile.
The aortas were defatted and acidextracted as described previously and the soluble elastin separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis.
The residue was extracted with urea, and alkaliinsoluble elastin prepared. The insoluble elastin was hydrolyzed and the radioactivity in the desmosines measured using an amino acid analyzer and liquid scintillation counting. The results of experiments of this type are shown in Table I. After a pulse of 15 min, the soluble elastin contained 6104 counts5 and no radioactivity was present in the desmosines of the crosslinked elastin. When the chase was done in the absence of fl-aminopropionitrile, there was a linear increase with time in the radioactivity of the desmosines and a decrease in that of soluble elastin. After 24 hours of chase, 89% of the radioactivity lost from the soluble elastin fraction could be accounted for as cross-linked elastin (Table I, Experiment 1). The presence of P-aminopropionitrile during the chase period markedly inhibited the formation of radioactive desmosine. After 24 hours, only 31% of the expected radioactive cross-linked elastin was observed, and 40% of the label was still present as soluble SThe acid extracts did not contain radioactive desmosines as revealed by amino acid analysis. The radioactivity under soluble elastin peaks was 1722 cpm (loo%), 1111 cpm (65%), 510 cpm (30%;), and 501 cpm (29%) after the pulse and 2, 4, and 6 hours of chase, respectively.
The radioactrvity under the collagen (or procollagen) peaks for the corresponding time periods was respectively 678 cpm, 1029 cpm, 944 cpm, and 747 cpm. FIG. 3 (right). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the soluble elastin preparation obtained from aortas pulselabeled with [4,5-SH]lysine. The aortas were extracted in acrd, and soluble elastin prepared from the extract by coacervatron followed by organic solvent extractron.
A sample containing 4455 cpm was used. The radioactivity of each 2-mm slice is shown. The migration of authentrc chick soluble elastin and collagen chains is indmated. elastin (Table I, Experiment 2). When &aminopropionitrile was present during the pulse as well as the chase period, no radioactivity appeared in the desmosines (Table I, Experiment 3).
While the above observations indicate rather strongly that a precursor-product relationship exists between soluble elastin and cross-linked elastin, it remains possible that the radioactivity in the desmosines arises from labeled substances present within the acid-insoluble aorta fraction, since large amounts of unidentified radioactive material were always observed in that fraction.' In order to determine whether or not this is the case, the following experiments were done. Acid-soluble and insoluble residue fractions were prepared from aortas pulse-labeled for 15 min without chase and from unlabeled control aortas. Various combinations of the labeled and unlabeled extract and residue fractions were incubated in the presence and absence of P-aminopropionitrile, and the amount of radioactivity appearing in desmosine and isodesmosine measured. The results of these experiments are presented in Table II. Incubation of the labeled residue fraction with unlabeled extract gave rise to only insignificant amounts of radioactivity in desmosines. However, the addition of labeled extract to this mixture resulted in the appearance of 658 counts distributed between desmosine and isodesmosine, and this reaction was inhibited by the presence of /3-aminopropionitrile.
In addition, the incubation of labeled extract and unlabeled residue resulted in the appearance of 479 counts in the desmosines and this reaction too was P-aminopropionitrile-inhibitable. Thus, the radioactivity appearing in the The aortas were pulse-labeled and chased for varying periods as specified. They were then extracted with acetic acid as described under "Methods." The proteins in the extracts were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and the radioactivity in the soluble elastin was determined." Elastin was purified from the residue, hydrolyzed, and the radioactivity of the desmosines measured after separation on amino acid analyzer. desmosines derives from a substance present in the acid-soluble fraction rather than in the insoluble residue fraction. Final proof that the substance in the soluble fraction serving as the precursor of desmosine-containing elastin is indeed soluble elastin was obtained from experiments using purified soluble elastin and lysyl oxidase. Radioactively labeled soluble elastin was obtained from pulse-labeled aortas and purified as described under "Methods" using coacervation and extraction in organic solvents (15,16). The preparation was homogeneous when examined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Fig. 3) and on cationic gels (pH 4.8) containing 6 M urea (not shown). Incubation of this purified material with purified lysyl oxidase and the unlabeled insoluble residue fraction resulted in the biosynthesis of desmosine and isodesmosine (Table III). DISCUSSION During recent years, evidence has been obtained supporting the idea that soluble elastin is the precursor of cross-linked elastin. For example, Miller et al. demonstrated that lathyrogens and copper deficiency interfere with the incorporation of lysine into cross-links of mature elastin (19). Subsequently, Smith et al. (1) and Sandberg et al. (21) showed that an elastin-like soluble protein accumulates in the aorta of copperdeficient pigs. This protein was isolated and partially characterized and, because of its unique amino acid composition and several other elastin-like properties, it was assumed to be the precursor of cross-linked elastin. Observations by others have provided additional support for this assumption. For example, a non-collagenous substance with physical and chemical properties identical with authentic soluble elastin but which, in addition, serves as substrate for lysyl oxidase, has been extracted from embryonic chick aortas pulse-labeled in vitro (13). This substance disappears from the soluble fraction with increasing incubation times (13). Although all of these observations are consistent with a precursor-product relationship between soluble elastin and cross-linked elastin, no compelling experiments have been done previously definitely documenting such a relationship.
The embryonic chick aorta appears to radioactive proteins of molecular weights 74,000 and 138,000 which have been identified as soluble elastin and the pro-a chains of collagen, respectively (13, 14). The soluble elastin fraction rapidly disappears from the extracts during chase with ['2C]lysine, and there is a concurrent loss of lysyl oxidase substrate. Since this loss is prevented in major part by fl-aminopropionitrile, it is likely that cross-linking into an insoluble form, presumably elastin, rather than degradation, is the cause for the disappearance of soluble elastin. If this were true, the lysines of soluble elastin would be converted to desmosines and other elastin cross-links resulting in the transfer of radioactivity from the soluble elastin into the alkali-insoluble elastin; this transfer would be prevented by P-aminopropionitrile.
As seen in Experiment 1 (Table I) cross-linked elastin was not present during the pulse period, but it appeared during the chase; after 24 hours, 89% of the radioactivity lost from the soluble elastin was recovered in the alkali-insoluble fraction. The formation of cross-linked elastin was inhibited by the presence of @-aminopropionitrile during the chase periods (Experiment 2, Table I) and essentially prevented when the inhibitor was present during both the pulse and chase periods (Experiment 3, Table I). Evidence obtained from other experiments demonstrates conclusively that cross-linked elastin derives from soluble elastin. As shown by the mixing experiments described in Table II, the cross-linked radioactive elastin could not arise from the insoluble aorta residue fraction. Nor could it arise as a 'In this experiment the apparent inability of @-aminopropionitrile to completely prevent desmosine formation is probably a result of lysyl oxidase activity during the pulse period when the inhibitor was not present, because when /3-aminopropionitrile was present during the chase as well as the pulse, no desmosines were formed (Experiment 3, Table I).
result of new protein synthesis during the chase period, because cycloheximide, an inhibitor of protein synthesis, was present. Therefore the radioactive elastin must have derived from the proteins of the pulse-labeled aorta extracts of which soluble elastin is the predominant component. Final proof that soluble elastin is indeed the precursor of insoluble elastin was obtained from the experiments described in Table III. The radioactivity present in the lysyl residues of a homogeneous preparation of soluble elastin was converted into desmosines of insoluble elastin by incubation with purified lysyl oxidase and a nonradioactive residue fraction. To our knowledge, these observations comprise the first in vitro demonstration of the biosynthesis of elastin cross-links using a purified protein precursor and lysyl oxidase.
Two other aspects of the conversion of soluble elastin to cross-linked elastin merit comment. While P-aminopropionitrile completely inhibits the cross-linking when it is present during both the pulse and chase periods (Table I), it does not prevent the insolubilization of the soluble elastin (13) (Fig. 2). Thus, reactions other than cross-linking which are not susceptible to fl-aminopropionitrile inhibition appear to be involved in the insolubilization process. The nature of these reactions is not clear, although two possibilities are apparent. Insolubilization may result from the binding of soluble elastin to non-elastin components of the extracellular matrix such as the microfibrillar protein or, the newly synthesized soluble elastin may bind through the lysyl side chains to aldehyde groups produced on the surface of cross-linked elastin prior to the exposure to p-aminopropionitrile.
Another interesting aspect of the conversion of soluble elastin to insoluble elastin is the time course of the reaction. While over 90% of the soluble elastin is lost from the soluble fraction after 3 hours of chase, only 48% of the expected radioactivity appears in the elastin desmosines during this period. The time lag in the formation of desmosines may be a consequence of additional reactions, some of which may involve currently unidentified enzymes, required for the conversion of a-aminoadipic-&semialdehyde to the desmosines and other cross-links.