RECOMBINATION REPAIR

RR encompasses both homology-dependent recombination and nonhomologous end-joining. The primary focus of the Fishel laboratory is homologous recombination, although several intermediate structures may be common to both. Homology-dependent RR utilizes homologous sequence alignment between participating DNAs to repair lesions (principally DSBs). DSB may result from a variety of DNA insults that include radiation damage, replication fork collapse primarily due to impassable nucleotide lesions, as well as specific endonuclease cleavage events. Sequence alignment requires the biochemical process of homologous pairing and strand exchange (recombinase). The bacterial RecA was the first protein to be found to display recombinase activity. Since this initial observation a number of RecA homologs have been isolated and determined to contain recombinase functions. After 25 years of study the mechanism of RecA-mediated strand exchange remains controversial.

Conservation of DNA Repair genes from bacteria to man

Most of the bacterial DNA repair genes have been conserved through human evolution. In many cases there have been gene duplications leading to multiple related homologs. For example, human cells contain five genes with sequence similarity to the bacterial MMR protein MutS. While bacterial MutS functions as a homodimer, the human homologs function as heterodimers in which there has been a clear separation of function between subunits. This evolutionary process appears to have enhanced the efficiency of MMR and the breadth of MutS signaling processes.

There are at least seven non-redundant human RecA homologs. An apparent transposition of N- and C-terminal function/sequence has resulted in two RecA homolog families: the RecA paralogs and the Rad51 paralogs. Genetic and biochemical studies in a number of organisms have suggested that all members of these families are intimate players in recombination repair (RR). A separation-of-function that appears to have evolved with the RAD51 paralogs provides a superb opportunity to understand the fundamental mechanism of RecA recombinases.

Comparison of RecA and RAD51 reveals multiple biochemical differences: RAD51 is not RecA

Aside form the obvious structural differences (transposition of the N- and C-terminus) the prototypical RecA is capable of catalyzing efficient and complete homologous pairing and strand exchange without additional cofactors. RAD51 has been found to initiate a short stretch of strand exchange but efficient and extensive strand exchange has only been demonstrated with the addition of RPA (the eukaryotic heterotrimeric SSB) and special biochemical conditions (ammonium sulfate salt). The orderly addition of cofactors to the RecA reaction, such as single-stranded binding (SSB) protein or RecF/O/R significantly enhances the efficiency.

RAD51 forms an ATP-bound NPF on DNA that is visibly indistinct from RecA. While not entirely defined in E.coli, nucleation of a RecA NPF likely begins with a monomer and is significantly faster on ssDNA than dsDNA; with the most effective nucleation site for a dsDNA being a ssDNA gap Extension of the NPF appears cooperative with assembly and disassembly taking place unidirectionally (5’ to 3’ on ssDNA) and in an end-dependent manner: protomers are added at one end and subtracted from the other. This “turnover” of RecA appears to require ATP hydrolysis, where the ADP-form is substantially less avid on DNA than the ATP-bound form. If SSB protein is present it replaces the RecA protein as it dissociates. RecA filament extension is faster than disassembly suggesting a mechanism for filling discontinuities following multiple nucleation events. It is generally believed that the 5’ to 3’ directionality and cooperativity of the RecA NPF establishes the foundation for the biochemical directionality of strand exchange as well as the cooperativity in ATPase activity/function. Details of RAD51 filament assembly and disassembly are unknown.

Unlike RecA, RAD51 binds ssDNA and dsDNA equivalently. Both ss- and ds- DNA induce a comparable RAD51 DNA-activated ATPase activity; which appears about ~50-fold lower than the largely ssDNA-dependent RecA ATPase. RAD51-mediated strand exchange can occur in either the 3’ to 5’ or 5’ to 3’ direction and does not require ATP hydrolysis. This is clearly different than RecA where extensive strand exchange is directional (5’ to 3’) and largely requires ATP hydrolysis. ATP hydrolysis by RecA shows cooperativity, while the RAD51 ATPase does not. It has been suggested that cooperative ATP hydrolysis permits RecA to drive strand exchange through large heterologous sequences as well as to perform a “four-strand” exchange reaction between a ss-gapped duplex circular DNA and a linear DNA substrate in which an end overlaps the ss-gap. The RAD51 protein alone is incapable of transiting heterologous DNA sequences during strand exchange and does not appear to catalyze a four-strand reaction.

RAD51 Structure: both answers and questions

The structure of several RAD51 homologs have been solved. These include nucleotide free, ADP-bound and ATP bound forms. The nucleotide-free and ADP bound forms generally recapitulate the structure of RecA determined over a decade ago (no ATP-bound structure for RecA has been solved). The ATP-bound structure of an Archeal RAD51 reveals hitherto unrecognized beauty in the mechanism of RecA/RAD51 homologs while eliciting a host of questions (Figure 2).

Probably the most remarkable is the location of the ATP binding site at the interface between subunits in an NPF (Figure 2). Domains of BOTH protomers coordinate the adenosine nucleotide, important salt ions, a magnesium ion, and water used for the hydrolysis (Figure 3). The addition of salt ions (potassium) also results in a further conformational transition that involves residues on both proteomeric subunits (see dashed circle Figure 3). A second magnesium ion of unknown function(s) appears in the structure near the ATP-bound protomeric interface. In addition, the binding of ATP resolves several disordered domains suggested to be involved in ssDNA and dsDNA binding in the previous RecA/RAD51 structures. One of the major questions that this structure evokes: why is RAD51 not cooperative? It would certainly appear that ATP-dependent functional interactions and conformational ordering occurs in RAD51. We have suggested that one or several of the other RAD51 paralogs may significantly alter RAD51 function(s) in order to further imitate RecA.

Both salt and magnesium play critical roles in RAD51 activities; particularly strand exchange, which only becomes measurable with specific salts. The reordering of the subunit interface with the addition of potassium ions is an enigma.

PROJECTS
  1. Biophysical analysis of hRAD51; hRAD51B; hRAD51C; hRAD51D; hXRCC2; hXRCC3 proteins
  2. Studies of the ADP/ATP processing by the RR proteins – analysis of the Molecular Switch
  3. Structural analysis of the RR proteins
  4. The functional effects of post-translational modification on RR proteins
  5. Proteins that interact with activated RR proteins
  6. The role of RR proteins in therapeutic drug resistance
  7. The interaction of RR proteins and MMR proteins in maintaining genomic stability
Home - Introduction - Switches - Mismatch Repair - Recombinational Repair - Meiosis - Proteome - Publications