The human gene is composed of 10 exons. The study by Kawabata and Saeki (1999) describes alternative splicing of the human gene using a numbering scheme of 12 alternatively spliced exons. The exon alignment is illustrated below.
Further descriptions of mouse alternatively spliced variants are described in Gruver et al., 2009 and Kawabata et al., 2004.

The HsTRAD transcript is the predominant variant. It is the full-length transcript and is made up of 2418 base pairs. This transcript will be used as the reference for the information that follows. There are multiple splice variants for the RAD51L3 gene that translate into one of seven putative protein isoforms.

The Saccharomyces cerevisiae Rad51 protein is homologous to the RecA protein of Escherichia coli. The RecA protein is known to promote repair via ATP-dependent mechanisms and is responsible for pairing and strand transfer between homologous DNA sequences. This is similar to the actions of the RAD51 protein in repair pathways. There are 5 members of the RAD51 family that share similar roles in recombination and DNA repair. RAD51D is one of these RecA-like genes (Pittman et al., 1998; Cartwright et al., 1998).
The RAD51D gene is predicted to encode seven different protein isoforms through alternative splicing. Isoform 1 is the predominant protein and is translated from the HsTRAD transcript mentioned previously (Kawabata and Saeki, 1999). The diagram below is based on this predominant form.

The RAD51D protein contains regions necessary for interactions with other RAD51 paralogs as well as those that are required for proper function of the protein. RAD51D contains an ATP binding domain with highly conserved Walker A and B motifs (Pittman et al., 1998; Cartwright et al., 1998). Mutations targeting the conserved residues of glycine and lysine within the Walker A motif region resulted in a reduction in RAD51C binding ability and were shown to be required for DNA repair (Gruver et al., 2005). The Walker B motif contains a "GGQRE" sequence between residues 219-223 that is also required for DNA repair (Wiese et al., 2006). Furthermore, RAD51D-XRCC2 complex formation is significantly reduced with mutations targeting a highly conserved aspartate residue within the Walker B motif (Wiese et al., 2006).
A carboxyl terminal domain spanning amino acids 77-329 has been identified to be required for RAD51D to interact with RAD51C. In addition, the "linker region" located between residues 54-77 in the amino terminus is required for proper interactions with XRCC2. Together, these interactions aid in the repair of DNA damage (Miller et al., 2004; Gruver et al., 2009).

According to the study by Kawabata and Saeki (1999), RAD51L3 transcripts are expressed to varying degrees in the colon, prostate, spleen, testis, ovaries, thymus, small intestine and leukocytes.

Located in the nucleus. Specifically, RAD51L3 localizes to the telomeres during both mitosis and meiosis (Tarsounas et al., 2004). There is evidence that RAD51L3 is found in the cytoplasm as well (Gruver et al., 2005).

RAD51D is one of five members of the RAD51 gene family that is known to participate in repair of double stranded DNA breaks via homologous recombination. Without repair, the DNA damage can result in cell death or chromosomal aberrations that can ultimately lead to cancer (Thacker, 2005). Knockout studies with mice have shown a dramatic increase in levels of chromosomal aberrations, most notably, chromatid and chromosome breaks that occur through unrepaired replication forks (Smiraldo et al., 2005; Hinz et al., 2006). Proteomic studies have identified an interaction between RAD51D with the SFPQ protein (Rajesh et al., 2011). Exposure of mouse RAD51D-deficient cells to a strong alkylating agent results in G2/M cell cycle arrest and ultimately apoptosis (Rajesh et al., 2010). RAD51D has recently been shown to play a diverse role in cellular processes through its interaction with proteins involved in cell division, embryo development, protein and carbohydrate metabolism, cellular trafficking, protein synthesis, modification or folding, and cellular structure (Rajesh et al., 2009).
RAD51L3 is directly associated with telomeres prevents their dysfunction (Tarsounas et al., 2004). In mouse studies, RAD51L3 foci were present at telomeres in both meiosis and mitosis. Knockout studies showed that "RAD51D-deficient" mice exhibited an increase in end-to-end fusion and telomere attrition (Smiraldo et al., 2005). In addition, human studies using RAD51D-deficient cells have shown repeated shortening of the telomeric DNA, leading to chromosomal instability. This suggests a role for "RAD51D" in telomere capping. Failure to provide this function can lead to chromosomal aberrations (Tarsounas and West, 2005).

Canis lupus familiaris [Dog]	Official gene name: RAD51-like 3 (S. cerevisiae) Genomic location: chromosome 9	Reference material: no primary references found
Pan troglodytes [Chimpanzee]	Official gene name: RAD51-like 3 (S. cerevisiae) Genomic location: chromosome 17	Reference material: no primary references found
Bos taurus [Cow]	Official gene name: RAD51-like 3 (S. cerevisiae) Genomic location: chromosome 19	Reference material: Zimin et al., 2009
Gallus gallus [Chicken]	Official gene name: RAD51-like 3 (S. cerevisiae) Genomic location: chromosome 19	Reference material: no primary references found
Rattus norvegicus [Rat]	Official gene name: RAD51-like 3 (S. cerevisiae) Genomic location: chromosome 10	Reference material: Strausberg et al., 2002
Mus musculus [Mouse]	Official gene name: RAD51-like 3 (S. cerevisiae) Genomic location: chromosome 11	Reference material: Pittman et al., 1998 Cartwright et al., 1998
Arabidopsis thaliana [Thale cress]	Official gene name: RAD51D (ARABIDOPSIS HOMOLOG OF RAD51D) Genomic location: chromosome 1	Reference material: Durrant et al., 2007
Oryza sativa [Rice]	Official gene name: Os09g0104200 Genomic location: chromosome 9 **Hypothetical protein**	Reference material: no primary references found
Danio rerio [Zebrafish]	Official gene name: zgc:77165 Genomic location: chromosome 5	Reference material: no primary references found

** Protein alignments and protein sequences are available at the HomoloGene database.

Single nucleotide polymorphisms have been identified in RAD51L3. However, only a small number of the major mutations occur in coding regions. The majority of the other mutations are present in various locations within the introns. Of the mutations affecting the gene, only one has an observed clinical association. It is observed that a mutation of the mRNA position 954 (SNP ID: rs28363284) results in an allele change to GGG (from the wild type GAG). This point mutation affects the 233^rd amino acid as a glycine residue is observed in this particular mutation rather than the natural glutamic acid. This particular variation in amino acid sequence has been implicated as a precursor to breast cancer (see "Implicated In" section below). Another mutation observed in the coding region is at mRNA position 188 (SNP ID: rs1871892), resulting in a change in the sequence to TCA (from the wild type CCA). This particular substitution results in the insertion of proline at the 36^th protein position rather than a serine. A third mutation observed is noted to occur at mRNA positions 810 (SNP ID: rs4796033). A mutation at this location results in a sequence of CAG (from the natural CGG). The effect of this substitution is the insertion of a glutamine residue at the 185^th amino acid position rather than the arginine observed in the wild type gene. It is noted, that this particular mutation also occurs in 2 additional transcripts of the gene at the mRNA positions 750 and 414 affecting the 165^th and 53^rd amino acid residues respectively. Other mutations in the coding region include E237K (SNP ID: rs115031549), R252Q (SNP ID: rs28363283), A245T (SNP ID: rs28363282), A210T (SNP ID: rs80116829), E177D (SNP ID: rs55942401), and R24S (SNP ID: rs28363257).