The nuclear encoded catalytic subunit of polymerase gamma (POLG) is the only polymerase in humans that is known to be located in mitochondria and has been proposed as the replicase of mitochondrial DNA. In its coding region lies a short polyglutamine tract, which is absent in POLG of mouse or Drosophila. Microsatellite repeats have been shown to evolve in length during the course of primate evolution and show considerable variation between individuals. Polyglutamine repeat expansions are a frequent cause of disease, in many cases involving also infertility. Mitochondrial involvement in infertility has been proposed, based on the high energy demand of sperm cells for motility, and mtDNA deletions and point-mutations described in sperm cells. POLG harbours a polymerase domain and an exonuclease domain, both needed for faithful replication of mtDNA, and mutations along the gene have been reported in connection with diseases associated with mitochondrial DNA deletions or depletion. Cell lines expressing POLG with an exonuclease domain mutation accumulate high loads of mtDNA point mutations compared to control cell lines. Higher mtDNA point mutation loads have been documented in tissues of ageing subjects than of young individuals. These findings prompted a number of questions: How variable is the POLG microsatellite in different populations? Do our primate relatives have the repeat and does the polyglutamine in humans have a role in common mitochondrial disease? Does POLG sequence variation have an effect on cells that are highly dependent on biological energy during cell maturation, differentiation and function, like sperm? To address these questions I analysed POLG genotype in different human populations, in various types of disease, and in non-human primates. The most common allele of POLG with 10 CAGs was shown to be maintained at a frequency of 0.88 in different human populations, suggesting maintenance by selection. Expansions of the repeat were ruled out as a common cause of mitochondrial disease. A homozygous lack of the common (CAG)10 allele of POLG was seen in 9 % of males in whom no other reason for infertility was known, compared to 1.2 % amongst unselected controls or none of 98 males known to be fertile. The semen of men affected by the mutation showed low number of sperm with poor motility and/or morphological defects, but not azoospermia, severe oligozoospermia or isolated asthenozoospermia. Later studies by others have shown that couples with male POLG CAG mutation can be helped to reproduce with the help of ICSI treatment. I therefore propose that in cases of idiopathic infertility the males should thus be analysed for POLG CAG repeat variation and if found to lack the common allele, the couples should be offered ICSI as the first and effective treatment. Infertile males had more mutations in their mtDNA control region than fertile controls and a positive correlation between the mutation load and the absence of the common POLG length variant was seen. An appreciable proportion of infertile males heterozygous for POLG CAG repeat length harbour also, in addition to the length variant, a second mutation somewhere else in their POLG gene. The length variants of the POLG polyglutamine tract were almost exclusively due to the elongation of the CAG repeat. Nucleotide level variation was seen only in 4 out of 1210 control alleles sequenced, all of whom had the same variant in which one of the CAG codons was changed to a CGG (arginine). The polyglutamine tract appears to have evolved during primate evolution from a short arginine-glutamine-leucine-proline region in monkeys to the ´clean` and extended polyglutamine in humans. Each primate species was shown to have its own specific repeat length with high frequency, similar to humans. Homozygous knock-in mice having a point mutation in the second exonuclease domain of Polg were shown to have a three- to five-fold increase in the amount of mtDNA point mutations, and they also had increased amounts of aberrant mtDNA molecules. The mtDNA point mutations accumulated linearly during lifetime of the mice but affected different mtDNA sequence segments differently. The mice showed a premature ageing phenotype with weight loss, reduced subcutaneous fat, alopecia, osteoporosis, kyphosis, anaemia, heart enlargement and reduced fertility, the males being practically infertile. The ageing related phenotypes start to manifest at ~24 weeks of age and the maximum lifespan of the ´mutator` mice was 61 weeks, which is less than half of the normal life span of laboratory mice. This study combined to the phenotypical and biochemical analysis done by collaborators and the data in the literature underline the critical role of DNA polymerase gamma in maintaining optimal function and healthy status of cells and tissues, and thus whole organisms through their whole lifetime.