Efficient repair of hydrogen peroxide-induced DNA damage by Escherichia coli requires SOS induction of RecA and RuvA proteins.

The survival of Escherichia coli following treatment with a low dose (1-3 mM) of hydrogen peroxide (H(2)O(2)) that causes extensive mode-one killing of DNA repair mutants is stimulated by the induction of the SOS regulon. Results for various mutants indicate that induction of recA and RecA protein-mediated recombination are critical factors contributing to the repair of H(2)O(2)-induced oxidative DNA damage. However, because DNA damage activates RecA protein's coprotease activity essential to cleavage of LexA repressor protein and derepression of all SOS genes, it is unclear to what extent induction of RecA protein stimulates this repair. To make this determination, we examined mode-one killing of DeltarecA cells carrying plasmid-borne recA (P(tac)-recA(+)) and constitutively expressing a fully induced level of wild-type RecA protein when SOS genes other than recA are non-inducible in a lexA3 (Ind(-)) genetic background or inducible in a lexA(+) background. At a H(2)O(2) dose resulting in maximal killing, DeltarecA lexA3 (Ind(-)) cells with P(tac)-recA(+) show 40-fold greater survival than lexA3 (Ind(-)) cells with chromosomal recA having a low, non-induced level of RecA protein. However, they still show 10- to 15-fold lower survival than wild-type cells and DeltarecA lexA(+) cells with P(tac)-recA(+). To determine if the inducible RuvA protein stimulates survival, we examined a ruvA60 mutant that is defective for the repair of UV-induced DNA damage. This mutant also shows 10- to 15-fold lower survival than wild-type cells. We conclude that while induction of RecA protein has a pronounced stimulatory effect on the recombinational repair of H(2)O(2)-induced oxidative DNA damage, the induction of other SOS proteins such as RuvA is essential for wild-type repair.

Differential cleavage of LexA and UmuD mediated by recA Pro67 mutants: implications for common LexA and UmuD binding sites on RecA.

In Escherichia coli, RecA-mediated cleavage of LexA repressor is a key regulatory event required for expression of SOS genes involved in the repair of DNA damage. RecA also mediates the cleavage of UmuD protein to UmuD, a form active in SOS mutagenesis. To determine whether LexA and UmuD have common binding determinants on RecA, we have compared the ability of several recA mutants to function in the cleavage of LexA versus UmuD in vivo. The data reveal that while some recA mutations at Pro67 have a similar effect on LexA and UmuD cleavage, others have striking differential effects. For example, a Pro67-->Trp mutation results in a high level of constitutive cleavage of both proteins. However, Pro67-->Asp and Glu mutations promote constitutive cleavage of LexA and reduce induction of UmuD cleavage to just 5 to 10% of wild-type activity. In contrast, Pro67-->Arg prevents LexA cleavage while allowing nearly 50% of wild-type induction of UmuD cleavage. These results are consistent with the idea that Pro67 is located at a site in the nucleoprotein filament where both LexA and UmuD contact RecA.

Mutations at Pro67 in the RecA protein P-loop motif differentially modify coprotease function and separate coprotease from recombination activities.

The functional significance of residues in the RecA protein P-loop motif was assessed by analyzing 100 unique mutants with single amino acid substitutions in this region. Comparison of the effects on the LexA coprotease and recombination activities shows that Pro67 is unique among these residues because only at this position did we find substitutions that caused differential effects on these functions. One mutant, Pro67-->Trp, displays high constitutive coprotease activity and a moderate inhibitory effect on recombination functions. Glu and Asp substitutions result in low level constitutive coprotease activity but dramatically reduce recombination activity. The purified Pro67-->Trp protein shows a completely relaxed specificity for NTP cofactors in LexA cleavage assays and can use shorter length oligonucleotides as cofactors for cleavage of lambda cI repressor than can wild type RecA. Interestingly, both the mutant protein and wild type RecA can use very short oligonucleotides, e.g. (dA)6 and (dT)6, as cofactors for LexA cleavage. We have also found two mutations at position 67, which are completely defective for LexA coprotease activity in vivo but still maintain recombinational DNA repair (Pro67-->Lys) and homologous recombination (Pro67-->Lys and Pro67-->Arg) activities. These findings show that the recombination activities of RecA are mutationally separable from the coprotease function and that Pro67 is located in a functionally important position in the RecA structure.

Functional characterization of residues in the P-loop motif of the RecA protein ATP binding site.

We have introduced a large number of single amino acid substitutions at six positions that lie within the P-loop motif of the ATP binding site in the Escherichia coli RecA protein. The activity of each recA mutant was determined using genetic assays which assess the catalytic proficiency of the RecA protein for both homologous genetic recombination and recombinational repair of damaged DNA. The six residues displayed unique patterns of allowed versus non-allowed substitutions that define the functional and structural constraints at each position. Our results show that while the restricted mutability of Gly66 and Ser70 conform to expectations based on their positions in the RecA crystal structure, strict constraints are in effect at positions 68 and 69 that are not apparent in the RecA structure but would be compatible with the proximity of these side-chains to bound DNA. Thr74 shows a rather unexpected pattern of allowed substitutions that may reflect two distinct structural solutions to optimal stabilization of bound nucleotide. In addition, specific substitutions at Pro67 result in mutant RecA proteins that appear to discriminate between homologous genetic recombination and recombinational DNA repair.

Orientation of the myelin proteolipid protein C-terminus in oligodendroglial membranes.

The topology of the integral membrane proteolipid protein (PLP) has important structural and functional implications for central nervous system myelin. To determine the orientation of the carboxyl-terminal portion of PLP, cultured mouse oligodendrocytes were probed with polyclonal antibodies raised against a synthetic terminal peptide corresponding to PLP residues 264-276 and with ten separate monoclonal antibodies that react with this region. Cells were examined by double-label indirect immunofluorescence for the presence of the PLP C-terminus and either oligodendrocyte-specific surface or intracellular antigens. To detect surface antigens, both living and paraformaldehyde-fixed cells were incubated with primary antibodies and then stained with fluorochrome-conjugated second antibodies. Antigens located within the cytoplasmic space were identified after fixation and permeabilization of cells. Live-labeled oligodendrocytes were stained brightly for myelin-oligodendrocyte glycoprotein, galactocerebroside, and other surface markers but did not stain for the PLP C-terminus or the intracellular proteins myelin basic protein and beta-tubulin. Fixation alone was sufficient for partial permeabilization of oligodendrocytes to antibodies and resulted in limited staining of the PLP C-terminus and intracellular proteins. The permeabilized oligodendrocytes stained intensely for the PLP C-terminus, myelin basic protein, and beta-tubulin. Finally, trypsinization of living oligodendrocytes eliminated surface myelin-oligodendrocyte glycoprotein staining but did not change the immunostaining properties of the PLP C-terminus. These results provide evidence that the carboxyl-terminus of PLP is located at the cytoplasmic face of oligodendroglial membranes.

Monoclonal antibodies against myelin proteolipid protein: identification and characterization of two major determinants.

This report describes the preparation and characterization of a panel of monoclonal antibodies (mAbs) against the myelin proteolipid protein (PLP). A Lewis rat was immunized with bovine proteolipid apoprotein and 27 mAbs were selected based on their reactivity against bovine PLP on enzyme-linked immunosorbent assays. Eleven mAbs recognized the PLP carboxyl-terminal sequence when tested against a panel of synthetic peptides in a solid-phase assay. A carboxyl-terminal pentapeptide (residues 272-276) was sufficient for antibody binding and the terminal phenylalanine residue was found particularly important. Deletion, modification, or replacement of this residue markedly reduced or obliterated antigen-antibody interaction. Nine mAbs reacted with a second antigenic determinant, residues 209-217, but these could be identified only by competitive immunoassays. This peptide was a more effective inhibitor than the longer peptides 202-217 and 205-221, suggesting that flanking residues may interfere with peptide-antibody interaction. Seven antibodies did not react with any of the synthetic peptides tested and their determinants remain unidentified. Immunoblot analysis showed that the mAbs reacted with both the PLP and the DM-20 isoforms. Twenty-three of the mAbs were of the immunoglobulin G2a or b isotype; the remaining antibodies were immunoglobulin M and all of these were specific for residues 209-217. Cultured murine oligodendrocytes were stained by most of the mAbs tested, but the most intense reactivity was observed with the carboxyl-terminus-specific mAbs. The immunocytochemical analyses demonstrate that the mAbs react with the native PLP in situ and show their potential usefulness for studies of the cell biology of myelin and oligodendrocytes.

Expression of plasmolipin in oligodendrocytes.

Plasmolipin is a plasma membrane proteolipid which has recently been described as a component of myelin (Cochary et al.: Journal of Neurochemistry 55:602-610, 1990). The present study reports the expression and localization of plasmolipin in primary glial cultures and secondary oligodendrocyte cultures. Double-label immunofluorescence showed that plasmolipin was expressed by galactocerebroside (GC)-positive oligodendrocytes, but was absent from astrocytes, characterized by their positive staining for glial fibrillary acidic protein (GFAP). At 1 week in culture plasmolipin staining was relatively weak in the cell body of some of the GC-positive cells. During the following 3 weeks in culture plasmolipin staining of oligodendrocytes gradually increased and was present in the cell body, its plasma membrane, and all the processes. However, the plasmolipin antibodies did not stain regions of the flat membrane sheets. Western blot analysis of homogenates from primary glial cultures showed that plasmolipin levels gradually increased during the first 5 weeks in culture. We conclude that the presence of plasmolipin in myelin is a result of its expression by oligodendrocytes.

Somatic cell genetic analysis of the galactocerebrosidase gene: lack of complementation in human Krabbe disease/twitcher mouse cell hybrids.

The inherited deficiency of galactosylceramide beta-galactosidase (E.C. 3.2.1.46: galactocerebrosidase) activity results in globoid cell leukodystrophy in humans (Krabbe disease) and in mice (twitcher mutant). To determine whether Krabbe patients' cells complement twitcher cells to produce, in hybrid combination, greater than deficient levels of galactocerebrosidase activity, five separate crosses were made between an established twitcher mouse cell line and five cell strains from unrelated Krabbe disease patients. A total of 57 twitcher mouse/Krabbe somatic cell hybrid lines developed from all of these crosses were deficient in galactocerebrosidase activity despite the presence of human chromosomes 14 or 17, which have been previously implicated as bearing the galactocerebrosidase gene. A control cross between twitcher mouse/positive control human fibroblasts resulted in 14 of 21 independent hybrid lines that expressed higher than deficient levels of galactocerebrosidase activity. The lack of complementation between Krabbe disease patient and twitcher mutant mouse cells provides further evidence that the twitcher mouse is an authentic murine model for Krabbe disease and supports the hypothesis that the mutations in both species are within the structural gene for the galactocerebrosidase enzyme.

Galactocerebrosidase activity in somatic cell hybrids derived from twitcher mouse/control human fibroblasts is associated with human chromosome 17.

Somatic cell hybrids derived from twitcher mouse cells and from control human fibroblasts were selected by two different methods. One method utilized 6-thioguanine-resistant twitcher cells as a parental line and the other used neomycin-resistant control human fibroblasts as a parental line so that hybrid lines could be selected in either HAT or in G-418 medium, respectively. The hybrid lines were analyzed for galactocerebrosidase activity. Since the twitcher cell lines are deficient in galactocerebrosidase activity, the presence of this activity in these hybrid lines depends upon the presence of human chromosome contents. Both galactocerebrosidase-positive and -deficient hybrid lines were analyzed for their human chromosome contents by the use of isozyme markers. In hybrids derived from both selection methods the expression of galactocerebrosidase activity was associated with the presence of human chromosome 17 marker isozymes. This was confirmed cytogenetically by means of trypsin-banded Giemsa staining of intact human chromosome 17 in three galactocerebrosidase-positive hybrid lines.

Establishment of a galactocerebrosidase-deficient twitcher mouse cell line that expresses galactocerebrosidase activity in hybrids with control human fibroblasts.

Primary cell cultures from twitcher (galactocerebrosidase deficient) mice were made by enzymatic dispersion and explantation of skin obtained from 3-d-old littermates of a twi+/twi X twi+/twi mating. Galactocerebrosidase activity remained deficient for two twitcher cell lines, TM-1 and TM-2, and both lines demonstrated an initial period of growth decline, followed by accelerated growth. The TM-2 line has been subcultured for more than 3.5 yr, has a modal chromosome number of 63, a doubling time of approximately 16 h, and has remained galactocerebrosidase deficient throughout its life span. These data indicate this to be an established twitcher cell line that can be continuously maintained in culture as a transformed galactocerebrosidase-deficient mouse cell line. This established line was rendered 6-thioguanine resistant so that the cells could be fused with control human fibroblasts and selected for hybrid lines in hypoxanthine-aminopterin-thymidine medium. Also, the established twitcher cells were crossed with neomycin-resistant control human fibroblasts and selected in G418 medium. Several of the hybrid lines from both crosses had higher than deficient levels of galactocerebrosidase activity initially, followed by a decrease to twitcher levels during subculture, whereas other lines retained high levels of activity. These results indicate that twitcher-human somatic cell hybrids will express galactocerebrosidase activity and thus may be useful for determining the human chromosome or chromosomes associated with this expression.