PCR-product directed gene disruption with a marker cassette having short homology regions is often used in Candida albicans. However, it is quite inefficient due to the high frequency of non-homologous recombination at non-targeted loci, which necessitates extensive screening to identify the correct disruptants. Thus, many PCR-based methods to introduce long flanking homology regions have been developed to increase the frequency of integration at the targeted loci. However, these methods are not that amenable for use with the widely employed C. albicans marker cassettes having direct repeats, as these repeats tend to recombine during PCR, resulting in shorter amplified products without the selection marker. To circumvent this limitation, we have developed a dinucleotide-sticky-end-ligation strategy to add one flanking homology region to one side of the selection cassette, and the other flanking homology region to the other side of the selection cassette. This method involves release of the selection cassette from the plasmid by digestion with two different restriction enzymes, followed by partial fill-in, to provide a unique two base overhang at each end of the cassette. The flanking homology regions, corresponding to the gene to be disrupted, are individually PCR-amplified, and treated with T4-DNA Polymerase in the presence of appropriate dNTPs to yield two base-5' overhangs. The primers used for the PCR have additional bases at the 5' ends such that after T4 DNA Polymerase treatment, the two flanks will have distinct overhangs compatible with the overhangs of the partially filled-in selection cassette. The selection cassette and the flanks are then ligated together and directly used to transform C. albicans. We have successfully used this method for disruption of several C. albicans genes. We have also used this method to recreate insertion mutations obtained with transposons to reconfirm the mutant phenotypes. This approach can be extended to other organisms like Schizosaccharomyces pombe which also require long flanking regions of homology for targeted gene disruption.