Mammalian DNA ligase III: molecular cloning, chromosomal localization, and expression in spermatocytes undergoing meiotic recombination.

Three biochemically distinct DNA ligase activities have been identified in mammalian cell extracts. We have recently purified DNA ligase II and DNA ligase III to near homogeneity from bovine liver and testis tissue, respectively. Amino acid sequencing studies indicated that these enzymes are encoded by the same gene. In the present study, human and murine cDNA clones encoding DNA ligase III were isolated with probes based on the peptide sequences. The human DNA ligase III cDNA encodes a polypeptide of 862 amino acids, whose sequence is more closely related to those of the DNA ligases encoded by poxviruses than to replicative DNA ligases, such as human DNA ligase I. In vitro transcription and translation of the cDNA produced a catalytically active DNA ligase similar in size and substrate specificity to the purified bovine enzyme. The DNA ligase III gene was localized to human chromosome 17, which eliminated this gene as a candidate for the cancer-prone disease Bloom syndrome that is associated with DNA joining abnormalities. DNA ligase III is ubiquitously expressed at low levels, except in the testes, in which the steady-state levels of DNA ligase III mRNA are at least 10-fold higher than those detected in other tissues and cells. Since DNA ligase I mRNA is also present at high levels in the testes, we examined the expression of the DNA ligase genes during spermatogenesis. DNA ligase I mRNA expression correlated with the contribution of proliferating spermatogonia cells to the testes, in agreement with the previously defined role of this enzyme in DNA replication. In contrast, elevated levels of DNA ligase III mRNA were observed in primary spermatocytes undergoing recombination prior to the first meiotic division. Therefore, we suggest that DNA ligase III seals DNA strand breaks that arise during the process of meiotic recombination in germ cells and as a consequence of DNA damage in somatic cells.

Coordinated activation of corneal wound response genes in vivo as observed by in situ hybridization.

We used subtractive screening of a cDNA library prepared from corneoscleral rims after cauterizing rat corneas. We identified 76 clones whose corresponding mRNA increased during the wound healing process in an in vivo model of injury which damages the corneal epithelium, stroma, and endothelium. Of these clones, 31 sequences encode known proteins. Another 45 clones are novel sequences based on comparison with the GenBank/EMBL databases. Changes in the level of expression of the novel genes, and a selected number of the known genes, were examined by in situ hybridization 22 and 72 hr after corneal injury. The majority produced a 'wound pattern' of expression such that the mRNAs were highly induced in all cell types adjacent to the wound site at 22 hr post injury. This signal decreased in intensity with distance from the wound site. In a subset of corneoslceral rims examined by in situ hybridization, the mRNAs for these genes were also highly induced in the limbal epithelium, where the progenitor corneal epithelial stem cells reside. By 72 hr, when acute tissue damage had been repaired, the induced mRNA was only faintly present in the thickened epithelium. Our results provide a useful framework for further studies defining the pathophysiological roles of the known and novel proteins encoded by the isolated cDNA clones.

Identification of a role of the vitronectin receptor and protein kinase C in the induction of endothelial cell vascular formation.

When cultured on a basement membrane substratum, endothelial cells undergo a rapid series of morphological and functional changes which result in the formation of histotypic tube-like structures, a process which mimics in vivo angiogenesis. Since this process is probably dependent on several cell adhesion and cell signaling phenomena, we examined the roles of integrins and protein kinase C in endothelial cell cord formation. Polyclonal antisera directed against the entire vitronectin (alpha v beta 3) and fibronectin (alpha 5 beta 1) receptors inhibited cord formation. Subunit-specific monoclonal antibodies to alpha v, beta 3, and beta 1 integrin subunits inhibited cord formation, while monoclonal antibodies to alpha 5 did not, which implicated the vitronectin receptor, and not the fibronectin receptor, in vascular formation. Protein kinase C inhibitors inhibited cord formation, while phorbol 12-myristate 13-acetate (PMA) caused endothelial cells to form longer cords. Since the vitronectin receptor has been shown to be phosphorylated in an in vitro system by protein kinase C, the possible functional link between the vitronectin receptor and protein kinase C during cellular morphogenesis was examined. The vitronectin receptor was more highly phosphorylated in cord-forming endothelial cells on basement membrane than in monolayer cells on vitronectin. Furthermore, this phosphorylation was inhibited by protein kinase C inhibitors, and PMA was required to induce vitronectin receptor phosphorylation in endothelial cells cultured on vitronectin. Colocalization studies were also performed using antisera to the vitronectin receptor and antibodies to protein kinase C. Although no strict colocalization was found, protein kinase C was localized in the cytoskeleton of endothelial cells initially plated on basement membrane or on vitronectin, and it translocated to the plasma membrane of C-shaped cord-forming cells on basement membrane. Thus, both the vitronectin receptor and protein kinase C play a role in in vitro cord formation.