Chemistry of the collagen cross-links. Origin and partial characterization of a putative mature cross-link of collagen.

The conversion of the reducible divalent cross-links in collagen to non-reducible multivalent cross-links in mature collagen has resulted in the identification of several new amino acids as the putative mature cross-link. None of these compounds has completely satisfied the necessary criteria. We have now isolated an amino acid of high Mr, derived from lysine, that is only present in high-Mr peptides derived from mature collagen. Its increase with age of the tissue correlates with the decrease in the reducible cross-links, and it is present both in mature skin and bone, which are initially cross-linked through the aldimine and oxo-imine divalent cross-link respectively. We propose that this amino acid, as yet incompletely characterized and designated compound M, is a major cross-link of mature collagen.     

Chemistry of the collagen cross-links. Isolation and characterization of two intermediate intermolecular cross-links in collagen

This paper describes the isolation from reduced collagen of two new amino acids believed to be involved, in their non-reduced form, as intermolecular cross-links stabilizing the collagen fibre. The reduction of intact collagen fibrils with tritiated sodium borohydride was found to stabilize the aldehyde-mediated cross-links to acid hydrolysis and thus allowed their location and isolation from acid hydrolysates on an automatic amino acid analyser. Comparison of the radioactive elution patterns from the autoanalyser of collagen treated in various ways before reduction permitted a preliminary classification of the peaks into cross-link precursors, intramolecular and intermolecular cross-links. The techniques employed to isolate the purified components on a large scale and to identify them structurally are described in detail. Two labile intermolecular cross-links were isolated in their reduced forms, one of which was identified by high-resolution mass spectrometry as N-(5-amino-5-carboxypentyl)hydroxylysine. The structure of this compound was confirmed by chemical synthesis. The cross-link precursor α-aminoadipic δ-semialdehyde was isolated in its reduced form, -hydroxynorleucine, together with its acid degradation product -chloronorleucine. A relatively stable intermolecular cross-link was isolated and partially characterized by mass spectrometry as an aldol resulting from the reaction of the δ-semialdehyde derived from lysine and hydroxylysine.     

Rearrangement of interstrand cross-links into intrastrand cross-links in cis-diamminedichloroplatinum(II)-modified DNA.

In the reaction of the anticancer drug cis-diamminedichloroplatinum(II) (cis-DDP) with DNA, bifunctional intrastrand and interstrand cross-links are formed. In this work, we show that at 37 degrees C interstrand cross-links (ICL) are labile and rearrange into intrastrand cross-links. The ICL instability was first studied with a 10 base pairs (bp) double-stranded oligonucleotide containing a unique site-specific ICL resulting from chelation of the N7 position of two guanine residues on the opposite strands of DNA at the d(GC/GC) site by a cis-diammineplatinum(II) residue. The bonds between the platinum and the N7 of guanine residues within the interstrand adduct are cleaved. In 50 mM NaCl or NaClO4, this cleavage results in the formation of monofunctional adducts which subsequently form intrastrand cross-links. One cleavage reaction takes place per cross-linked duplex in either of both DNA strands. Whereas the starting cross-linked 10 bp duplex is hydrogen bonded, the two complementary DNA strands separate after the cleavage of the ICL. Under these conditions, the cleavage reaction is irreversible allowing its rate measurement (t1/2= 29+/-2 h) and closure of monofunctional adducts to intrastrand cross-links occurs within single-stranded DNA. Within a longer cross-linked oligonucleotide (20 bp), ICL are apparently more stable (t1/2= 120+/-12 h) as a consequense of monofunctional adducts closure back to ICL. We propose that the ICL cleavage is reversible in DNA and that these adducts rearrange finally into intrastrand cross-links. Our results could explain an 'ICL unhooking' in previously reported in vivo repair studies [Zhenet al. (1993)Carcinogenesis14, 919-924].     

Collagen cross-linking. Synthesis of collagen cross-links in vitro with highly purified lysyl oxidase.

In this paper, the synthesis of collagen cross-links in vitro was investigated in a defined system consisting of highly purified chick cartilage lysyl oxidase and chick bone collagen fibrils. Cross-link synthesis in vitro was quite similar to the biosynthesis of collagen cross-links in vivo. Enzyme-dependent synthesis of cross-link intermediates and cross-linked collagen derived from lathyritic collagen occurred. The concentration of the two principal reducible cross-links, N6:6'-dehydro-5,5'-dihydroxylysinonorleucine and N6:6'-dehydro-5-hydroxylysinonorleucine, increased to a peak value of approximately two cross-links per molecule and then decreased. Synthesis of histidinohydroxymerodesmosine and a second polyfunctional cross-link of unknown structure began after synthesis of bifunctional cross-links was largely completed and proceeded linearly afterwards. Inhibition of lysyl oxidase after the bulk of bifunctional cross-link synthesis had occurred did not alter the rate of decrease in reducible cross-link concentration but did inhibit further histidinohydroxymerodesmosine synthesis. These results indicate that lysyl oxidase and collagen fibrils are the only macromolecules required for cross-link biosynthesis in vivo. It is likely that the decrease in reducible cross-links observed during fibril maturation results from spontaneous reactions within the collagen fibril rather than additional enzymatic reactions.     

The chemistry of natural enzyme-induced cross-links of proteins

Summary The cross-linking of protein molecules to form stable supramolecular aggregates capable of acting as protective and supporting structures is a common feature of organisms coping with the stresses of life. These new polymeric forms range from thick rigid structures to thin flexible membranes. The formation of such cross-links must be carefully controlled since more or less than optimal cross-linking could lead to malfunction or even death of the organism. The chemistry of the amino acids converted or directly involved in the formation of these cross-links is complex and a range of new amino acids has been identified. Di- and tri-tyrosines are formed by the action of peroxidases, quinones by catechol oxidases, glutamyl lysine iso-peptide bonds by glutamyl transferase and a complex series of lysine- aldehyde derived cross-links induced by lysyl oxidase. These cross-linking mechanisms provide an insight into the complex changes in tissue function during growth of the organism and their effects on the properties of foods.     

Age-related changes in the reducible cross-links on connectin from human skeletal muscle.

After NaB³H₄-reduction of connectin from human skeletal muscle, the changes in the amounts of the reducible cross-links and specific radioactivity of this elastic protein were followed throughout the whole life-span from embryo to old age. The reducible cross-links, aldimine forms of lysinonorleucine and histidino-hydroxymerodesmosine, and unidentified reducible compounds, which were assumed to be cross-linking amino acids, were found to remarkably decrease with age. A progressive decrease in the incorporation of tritium into the reducible compounds was also observed. We conclude that the conversion of the reducible cross-links derived from lysine and hydroxylysine aldehydes to non-reducible compounds is an essential step in the maturation of connectin fibrils, similar to collagen fibrils.     

Structural characterization of pyrrolic cross-links in collagen using a biotinylated Ehrlich's reagent

The structures of pyrrolic forms of cross-links in collagen have been confirmed by reacting collagen peptides with a biotinylated Ehrlich's reagent. This reagent was synthesized by converting the cyano group of N-methyl-N-cyanoethyl-4-aminobenzaldehyde to a carboxylic acid, followed by conjugation with biotin pentyl-amine. Derivatization of peptides from bone collagen both stabilized the pyrroles and facilitated selective isolation of the pyrrole-containing peptides using a monomeric avidin column. Reactivity of the biotinylated reagent with collagen peptides was similar to that of the standard Ehrlich reagent, but heat denaturation of the tissue before enzyme digestion resulted in the loss of about 50% of the pyrrole cross-links. Identification of a series of peptides by mass spectrometry confirmed the presence of derivatized pyrrole structures combined with between 1 and 16 amino acid residues. Almost all of the pyrrole-containing peptides appeared to be derived from N-terminal telopeptide sequences, and the nonhydroxylated (lysine-derived) form predominated over pyrrole cross-links derived from helical hydroxylysine.     

Stable, nonreducible cross-links of mature collagen.

During in vivo maturation, and also during in vitro incubation with physiological buffers, native collagen fibers display a progressive increase in tensile strength and insolubility. Paralleling these physiologically important changes is a progressive loss of the reducible cross-links which initially join the triple-chained subunits of collagen fibers. Although there is evidence suggesting that the reducible cross-links are gradually transformed into more stable, nonreducible cross-links during maturation, the nature of the transformation process and the structure of the stable "mature" cross-links has remained a mystery. In order to test the possibility that cross-link transformation involves addition of a nucleophilic amino acid residue to the reducible cross-links, histidine, arginine, glutamate, aspartate, lysine, and hydroxylysine residues were chemically modified, and the effect of each modification procedure on the in vitro transformation of reducible cross-links was ascertained. The results of these experiments indicated that destruction of histidine, arginine, glutamate, and aspartate residues has no measurable effect on the rate and extent of reducible cross-link transformation in hard tissue collagens. In contrast, modification of lysine and hydrocylysine residues with a wide variety of specific reagents completely blocks the transformation of reducible cross-links. Removal of the reversible blocking groups from lysine and hydroxlylysine residues then allows the transformation to proceed normally. These results indicate that collagen maturation involves nucleophilic addition of lysine and/or hydroxylysine residues to the electrophilic double bond of the reducible cross-links, yielding derivatives which are not only more stable but also capable of cross-linking more collagen molecules than their reducible precursors.     

The nature of the collagen cross-links in bone in the chronic uraemic state.

The cross-links from NaB3H4-reduced bone collagen of chronically uraemic rats and pairfed controls were compared. The ratio of the reduced cross-links deltadelta'-dihydroxylysinonorleucine to delta-hydroxylysinonorleucine was significantly increased in the uraemic animals. The observed increment in the dihydroxylysinonorleucine:hydroxylysinonorleucine ratio was accentuated as the uraemic state advanced. The data indicate that osteodystrophy of chronic renal insufficiency is characterized by an alteration of the quantitative relations between cross-links and aldehydic precursors of bone collagen.     

Diepoxybutane interstrand cross-links induce DNA bending

The bifunctional alkylating agent 1,2,3,4-diepoxybutane (DEB) is thought to be a major contributor to the carcinogenicity of 1,3-butadiene, from which it is derived in vivo. DEB forms DNA interstrand cross-links primarily between distal deoxyguanosine residues at the duplex sequence 5′-GNC. In order for the short butanediol tether to span this distance, distortion of the DNA target has been postulated. We determined that the electrophoretic mobility of ligated DNA oligomers containing DEB cross-links was retarded in comparison with control, uncross-linked DNA. Our data are consistent with DNA bending of ∼34° per lesion towards the major groove.     

Initiation of repair of DNA-polypeptide cross-links by the UvrABC nuclease

Although the biochemical pathways that repair DNA-protein cross-links have not been clearly elucidated, it has been proposed that the partial proteolysis of cross-linked proteins into smaller oligopeptides constitutes an initial step in removal of these lesions by nucleotide excision repair (NER). To test the validity of this repair model, several site-specific DNA-peptide and DNA-protein cross-links were engineered via linkage at (1) an acrolein-derived gamma-hydroxypropanodeoxyguanosine adduct and (2) an apurinic/apyrimidinic site, and the initiation of repair was examined in vitro using recombinant proteins UvrA and UvrB from Bacillus caldotenax and UvrC from Thermotoga maritima. The polypeptides cross-linked to DNA were Lys-Trp-Lys-Lys, Lys-Phe-His-Glu-Lys-His-His-Ser-His-Arg-Gly-Tyr, and the 16 kDa protein, T4 pyrimidine dimer glycosylase/apurinic/apyrimidinic site lyase. For the substrates examined, DNA incision required the coordinated action of all three proteins and occurred at the eighth phosphodiester bond 5' to the lesion. The incision rates for DNA-peptide cross-links were comparable to or greater than that measured on fluorescein-adducted DNA, an excellent substrate for UvrABC. Incision rates were dependent on both the site of covalent attachment on the DNA and the size of the bound peptide. Importantly, incision of a DNA-protein cross-link occurred at a rate approximately 3.5-8-fold slower than the rates observed for DNA-peptide cross-links. Thus, direct evidence has been obtained indicating that (1) DNA-peptide cross-links can be efficiently incised by the NER proteins and (2) DNA-peptide cross-links are preferable substrates for this system relative to DNA-protein cross-links. These data suggest that proteolytic degradation of DNA-protein cross-links may be an important processing step in facilitating NER.     

The effect of cis-diamminedichloroplatinum (II) on the formation of DNA-Pt-DNA cross-links.

The effect of cis-diamminedichloroplatinum (II) (cis-DDP) on the formation of interstrand cross-links in DNA and in DNA and chromatin complex from leukocytes was studied. Following the use of cis-DDP the number of DNA-DNA interstrand cross-links was elevated with the increase of cis-DDP concentration and elongation of reaction time. It was also found that nucleic proteins reduce the quantity of the cis-DDP induced DNA-DNA interstrand cross-links in the DNA in nucleoprotein complex when compared with the links in the isolated DNA.     

Repair of DNA-protein cross-links in mammalian cells

DNA-protein cross-links are generated by both endogenous and exogenous DNA damaging agents, as intermediates during normal DNA metabolism, and during abortive base excision repair. Cross-links are relatively common lesions that are lethal when they block progression of DNA polymerases. DNA-protein cross-links may be broadly categorized into four groups by the DNA and protein chemistries near the cross-link and by the source of the cross-link: DNA-protein cross-links may be found (1) in nicked DNA at the 3' end of one strand (topo I), (2) in nicked DNA at the 5' end of one strand (pol beta), (3) at the 5' ends of both strands adjacent to nicks in close proximity (topo II; Spo 11), and (4) in one strand of duplex DNA (UV irradiation; bifunctional carcinogens and chemotherapeutic agents). Repair mechanisms are reasonably well-defined for groups 1 and 3, and suggested for groups 2 and 4. Our work is focused on the recognition and removal of DNA-protein cross-links in duplex DNA (group 4).     

Processing of targeted psoralen cross-links in Xenopus oocytes.

Psoralen cross-links have been shown to be both mutagenic and recombinagenic in bacterial, yeast, and mammalian cells. Double-strand breaks (DSBs) have been implicated as intermediates in the removal of psoralen cross-links. Recent work has suggested that site-specific mutagenesis and recombination might be achieved through the use of targeted psoralen adducts. The fate of plasmids containing psoralen adducts was evaluated in Xenopus oocytes, an experimental system that has well-characterized recombination capabilities and advantages in the analysis of intermediates in DNA metabolism. Psoralen adducts were delivered to a specific site by a triplex-forming oligonucleotide. These lesions are clearly recognized and processed in oocytes, since mutagenesis was observed at the target site. The spectrum of induced mutations was compared with that found in similar studies in mammalian cells. Plasmids carrying multiple random adducts were preferentially degraded, perhaps due to the introduction of DSBs. However, when DNAs carrying site-specific adducts were examined, no plasmid loss was observed and removal of cross-links was found to be very slow. Sensitive assays for DSB-dependent homologous recombination were performed with substrates with one or two cross-link sites. No adduct-stimulated recombination was observed with a single lesion, and only very low levels were observed with paired lesions, even when a large proportion of the cross-links was removed by the oocytes. We conclude that DSBs or other recombinagenic structures are not efficiently formed at psoralen adducts in Xenopus oocytes. While psoralen is not a promising reagent for stimulating site-specific recombination, it is effective in inducing targeted mutations.     

Amino acid sequences C-terminal to the cross-links in bovine elastin.

All the desmosine-containing elastolytic peptides of bovine ligamentum-nuchae elastin have now been examined for amino acid sequences C-terminal to the cross-links. In addition, amino acid residues C-terminal to lysine residues in bovine tropoelastin were also examined. No tyrosine C-terminal to cross-links in bovine elastin or C-terminal to lysine in tropoelastin was detected. Apparently all the tyrosine residues C-terminal to lysine residues in pig tropoelastin are replaced with phenylalanine in bovine tropoelastin. All the data presented are consistent with the scheme proposed for the formation of desmosine and isodesmosine cross-links of elastin by Gerber & Anwar [(1975) Biochem. J. 149, 685--695].     

Determination of the relationship between collagen cross-links and the bone-tissue stiffness in the porcine mandibular condyle

Although bone-tissue stiffness is closely related to the degree to which bone has been mineralized, other determinants are yet to be identified. We, therefore, examined the extent to which the mineralization degree, collagen, and its cross-links are related to bone-tissue stiffness. A total of 50 cancellous and cortical bone samples were derived from the right mandibular condyles of five young and five adult female pigs. The degree of mineralization of bone (DMB) was assessed using micro-computed tomography. Using high-performance liquid chromatography, we quantified the collagen content and the number of cross-links per collagen molecule of two enzymatic cross-links: hydroxylysylpyridinoline (HP) and lysylpyridinoline (LP), and one non-enzymatic cross-link: pentosidine (Pen). Nanoindentation was used to assess bone-tissue stiffness in three directions, and multiple linear regressions were used to calculate the correlation between collagen properties and bone-tissue stiffness, with the DMB as first predictor. Whereas the bone-tissue stiffness of cancellous bone did not differ between the three directions of nanoindentation, or between the two age groups, cortical bone-tissue stiffness was higher in the adult tissue. After correction for DMB, the cross-links studied did not increase the explained variance. In the young group, however, LP significantly improved the explained variance in bone-tissue stiffness. Approximately half of the variation in bone-tissue stiffness in cancellous and cortical bone was explained by the DMB and the LP cross-links and thus they cannot be considered the sole determinants of the bone-tissue stiffness.     

Simultaneous determination of the reducible and nonreducible cross-links of connective tissue. Analysis of mineralized and nonmineralized bone collagen

Secondary amine cross-links occur in collagen and elastin from a number of tissue sources. Quantification of these cross-links by amino acid analysis is complicated by the problem of separating cross-links, which are often minor components, from the more common amino acids and also because relatively large amounts of a cross-link are required to determine a color factor. A specific radioactive labeling method has been developed and used to quantify cross-links in bone collagen. Primary amines such as lysine and hydroxylysine are first guanidinated with 3,5-dimethylpyrazole-1-carboxamidine nitrate (DMPC). Secondary amines, which are unreactive with DMPC, are then quantitatively cyanoethylated with [14C]acrylonitrile. This procedure can be used to detect any secondary amine cross-link, with higher sensitivity than ninhydrin analysis, in peptide form as well as in acid hydrolysates. It is applied here in conjunction with [3H]NaBH4 reduction to simultaneously quantify Schiff base cross-links and amounts of in vivo reduction of Schiff bases in mineralized versus nonmineralized bovine bone.     

Recombinational and mutagenic repair of psoralen interstrand cross-links in Saccharomyces cerevisiae

Psoralen photoreacts with DNA to form interstrand cross-links, which can be repaired by both nonmutagenic nucleotide excision repair and recombinational repair pathways and by mutagenic pathways. In the yeast Saccharomyces cerevisiae, psoralen cross-links are processed by nucleotide excision repair to form double-strand breaks (DSBs). In yeast, DSBs are repaired primarily by homologous recombination, predicting that cross-link and DSB repair should induce similar recombination end points. We compared psoralen cross-link, psoralen monoadduct, and DSB repair using plasmid substrates with site-specific lesions and measured the patterns of gene conversion, crossing over, and targeted mutation. Psoralen cross-links induced both recombination and mutations, whereas DSBs induced only recombination, and monoadducts were neither recombinogenic nor mutagenic. Although the cross-link- and DSB-induced patterns of plasmid integration and gene conversion were similar in most respects, they showed opposite asymmetries in their unidirectional conversion tracts: primarily upstream from the damage site for cross-links but downstream for DSBs. Cross-links induced targeted mutations in 5% of the repaired plasmids; all were base substitutions, primarily T --> C transitions. The major pathway of psoralen cross-link repair in yeast is error-free and involves the formation of DSB intermediates followed by homologous recombination. A fraction of the cross-links enter an error-prone pathway, resulting in mutations at the damage site.     

The chemistry of the collagen cross-links. A convenient synthesis of the reduction products of several collagen cross-links and cross-link precursors

New synthetic routes to the reduction products of several collagen cross-links and cross-link precursors are described. By the use of these routes hydroxynorleucine, 5,6-dihydroxynorleucine, hydroxylysinonorleucine and hydroxylysinohydroxynorleucine can be synthesized via one common synthetic intermediate. The synthetic routes provide a convenient source of these unusual amino acids, as well as confirming the structure of hydroxylysinohydroxynorleucine and the other lysine-derived residues found in borohydride-reduced collagens.     

Use of oligonucleotides to define the site of interstrand cross-links induced by Adriamycin.

It has been known for several years that Adriamycin forms adducts and interstrand cross-links when reacted for long periods of time with bacterial and mammalian DNA in vitro, with the cross-link being restricted to 2 bp elements containing GpC sequences. The self-complementary 20mer deoxyoligonucleotide TA4T4GCA4T4A has been used in this study as a model of the apparent G-G cross-linking site at GpC sequences. The rate of formation of cross-links, as well as the dependence on both Adriamycin and Fe(III) concentration, were similar with this oligonucleotide as compared with calf thymus DNA. The cross-linking was demonstrated on both denaturing and non-denaturing sequencing gels. The half-life of the G-G cross-link was 40 h, consistent with that implied with high molecular weight, heterogeneous sequence DNA. Exonuclease III digests of adducts formed with 20mer deoxyoligonucleotides containing single, central G-G, G-I and I-I potential cross-links revealed that a guanine residue is required at both ends of the cross-link. No cross-linking was observed with a similar oligonucleotide containing only a single central (G.C) bp.