Resolution of Holliday junctions by eukaryotic DNA topoisomerase I.

The Holliday junction, a key intermediate in both homologous and site-specific recombination, is generated by the reciprocal exchange of single strands between two DNA duplexes. Resolution of the junctions can occur in two directions with respect to flanking markers, either restoring the parental DNA configuration or generating a genetic crossover. Recombination can be regulated, in principle, by factors that influence the directionality of the resolution step. We demonstrate that the vaccinia virus DNA topoisomerase, a eukaryotic type I enzyme, catalyzes resolution of synthetic Holliday junctions in vitro. The mechanism entails concerted transesterifications at two recognition sites, 5'-CCCTT decreases, that are opposed within a partially mobile four-way junction. Cruciforms are resolved unidirectionally and with high efficiency into two linear duplexes. These findings suggest a model whereby type I topoisomerases may either promote or suppress genetic recombination in vivo.

The extent of heterocellular communication mediated by gap junctions is predictive of bystander tumor cytotoxicity in vitro.

Herpes simplex virus thymidine kinase (HSV-tk)/ganciclovir (GCV) viral-directed enzyme prodrug gene therapy causes potent, tumor-selective cytotoxicity in animal models in which HSV-tk gene transduction is limited to a minority of tumor cells. The passage of toxic molecules from HSV-tk+ cells to neighboring HSV-tk- cells during GCV therapy is one mechanism that may account for this "bystander" cytotoxicity. To investigate whether gap junction-mediated intercellular coupling could mediate this bystander effect, we used a flow cytometry assay to quantitate the extent of heterocellular coupling between HSV-tk+ murine fibroblasts and both rodent and human tumor cell lines. Bystander tumor cytotoxicity during GCV treatment in a coculture assay was highly correlated (P < 0.001) with the extent of gap junction-mediated coupling. These findings show that gap junction-mediated intercellular coupling contributes to the in vitro bystander effect during HSV-tk/GCV therapy and that retroviral transduction of tumor cells is not required for bystander cytotoxicity.

Rho protein regulates tight junctions and perijunctional actin organization in polarized epithelia.

The rho family of GTP-binding proteins regulates actin filament organization. In unpolarized mammalian cells, rho proteins regulate the assembly of actin-containing stress fibers at the cell-matrix interface. Polarized epithelial cells, in contrast, are tall and cylindrical with well developed intercellular tight junctions that permit them to behave as biologic barriers. We report that rho regulates filamentous actin organization preferentially in the apical pole of polarized intestinal epithelial cells and, in so doing, influences the organization and permeability of the associated apical tight junctions. Thus, barrier function, which is an essential characteristic of columnar epithelia, is regulated by rho.

Leukocytes express connexin 43 after activation with lipopolysaccharide and appear to form gap junctions with endothelial cells after ischemia-reperfusion.

Levels and subcellular distribution of connexin 43 (Cx43), a gap junction protein, were studied in hamster leukocytes before and after activation with endotoxin (lipopolysaccharide, LPS) both in vitro and in vivo. Untreated leukocytes did not express Cx43. However, Cx43 was clearly detectable by indirect immunofluorescence in cells treated in vitro with LPS (1 micrograms/ml, 3 hr). Cx43 was also detected in leukocytes obtained from the peritoneal cavity 5-7 days after LPS-induced inflammation. In some leukocytes that formed clusters Cx43 immunoreactivity was present at appositional membranes, suggesting formation of homotypic gap junctions. In cell homogenates of activated peritoneal macrophages, Cx43, detected by Western blot analysis, was mostly unphosphorylated. A second in vivo inflammatory condition studied was that induced by ischemia-reperfusion of the hamster cheek pouch. In this system, leukocytes that adhered to venular endothelial cells after 1 hr of ischemia, followed by 1 hr of reperfusion, expressed Cx43. Electron microscope observations revealed small close appositions, putative gap junctions, at leukocyte-endothelial cell and leukocyte-leukocyte contacts. These results indicate that the expression of Cx43 can be induced in leukocytes during an inflammatory response which might allow for heterotypic or homotypic intercellular gap junctional communication. Gap junctions may play a role in leukocyte extravasation.

Regulated assembly of tight junctions by protein kinase C.

We have previously shown that protein phosphorylation plays an important role in the sorting and assembly of tight junctions. We have now examined in detail the role of protein kinases in intercellular junction biogenesis by using a combination of highly specific and broad-spectrum inhibitors that act by independent mechanisms. Our data indicate that protein kinase C (PKC) is required for the proper assembly of tight junctions. Low concentrations of the specific inhibitor of PKC, calphostin C, markedly inhibited development of transepithelial electrical resistance, a functional measure of tight-junction biogenesis. The effect of PKC inhibitors on the development of tight junctions, as measured by resistance, was paralleled by a delay in the sorting of the tight-junction protein, zona occludens 1 (ZO-1), to the tight junction. The assembly of desmosomes and the adherens junction were not detectably affected, as determined by immunocytochemical analysis. In addition, ZO-1 was phosphorylated subsequent to the initiation of cell-cell contact, and treatment with calphostin C prevented approximately 85% of the phosphorylation increase. Furthermore, in vitro measurements indicate that ZO-1 may be a direct target of PKC. Moreover, membrane-associated PKC activity more than doubled during junction assembly, and immunocytochemical analysis revealed a pool of PKC zeta that appeared to colocalize with ZO-1 at the tight junction. A preformed complex containing ZO-1, ZO-2, p130, as well as 330- and 65-kDa phosphoproteins was detected by coimmunoprecipitation in both the presence and absence of cell-cell contact. Identity of the 330- and 65-kDa phosphoproteins remains to be determined, but the 65-kDa protein may be occludin. The mass of this complex and the incorporation of ZO-1 into the Triton X-100-insoluble cytoskeleton were not PKC dependent.

RuvC protein resolves Holliday junctions via cleavage of the continuous (noncrossover) strands.

The RuvC protein of Escherichia coli resolves Holliday junctions during genetic recombination and the postreplicational repair of DNA damage. Using synthetic Holliday junctions that are constrained to adopt defined isomeric configurations, we show that resolution occurs by symmetric cleavage of the continuous (noncrossing) pair of DNA strands. This result contrasts with that observed with phage T4 endonuclease VII, which cleaves the pair of crossing strands. In the presence of RuvC, the pair of continuous strands (i.e., the target strands for cleavage) exhibit a hypersensitivity to hydroxyl radicals. These results indicate that the continuous strands are distorted within the RuvC/Holliday junction complex and that RuvC-mediated resolution events require protein-directed structural changes to the four-way junction.

Cultured adherent cells from marrow can serve as long-lasting precursor cells for bone, cartilage, and lung in irradiated mice.

Cells from transgenic mice expressing a human mini-gene for collagen I were used as markers to follow the fate of mesenchymal precursor cells from marrow that were partially enriched by adherence to plastic, expanded in culture, and then injected into irradiated mice. Sensitive PCR assays for the marker collagen I gene indicated that few of the donor cells were present in the recipient mice after 1 week, but 1-5 months later, the donor cells accounted for 1.5-12% of the cells in bone, cartilage, and lung in addition to marrow and spleen. A PCR in situ assay on lung indicated that the donor cells diffusely populated the parenchyma, and reverse transcription-PCR assays indicated that the marker collagen I gene was expressed in a tissue-specific manner. The results, therefore, demonstrated that mesenchymal precursor cells from marrow that are expanded in culture can serve as long-lasting precursors for mesenchymal cells in bone, cartilage, and lung. They suggest that cells may be particularly attractive targets for gene therapy ex vivo.

Abnormal junctions between surface membrane and sarcoplasmic reticulum in skeletal muscle with a mutation targeted to the ryanodine receptor.

Junctions that mediate excitation-contraction (e-c) coupling are formed between the sarcoplasmic reticulum (SR) and either the surface membrane or the transverse (T) tubules in normal skeletal muscle. Two structural components of the junctions, the feet of the SR and the tetrads of T tubules, have been identified respectively as ryanodine receptors (RyRs, or SR calcium-release channels), and as groups of four dihydropyridine receptors (DHPRs, or voltage sensors of e-c coupling). A targeted mutation (skrrm1) of the gene for skeletal muscle RyRs in mice results in the absence of e-c coupling in homozygous offspring of transgenic parents. The mutant gene is expected to produce no functional RyRs, and we have named the mutant mice "dyspedic" because they lack feet--the cytoplasmic domain of RyRs anchored in the SR membrane. We have examined the development of junctions in skeletal muscle fibers from normal and dyspedic embryos. Surprisingly, despite the absence of RyRs, junctions are formed in dyspedic myotubes, but the junctional gap between the SR and T tubule is narrow, presumably because the feet are missing. Tetrads are also absent from these junctions. The results confirm the identity of RyRs and feet and a major role for RyRs and tetrads in e-c coupling. Since junctions form in the absence of feet and tetrads, coupling of SR to surface membrane and T tubules appears to be mediated by additional proteins, distinct from either RyRs or DHPRs.