Paediatric human immunodeficiency virus (HIV) infection has become a major socio-economic health problem in recent years as the number of HIV-1 infected children steadily increases. The majority of these infants are infected through mother-to-child transmission, with the frequency of vertical transmission varying between 12,9% and 65%. In order to implement appropriate management and possible treatment of these infected neonates, it is essential to have reliable laboratory tests for the early diagnosis of an HIV infection. At the time that this study was initiated, the diagnosis of HIV-1 infection in the Groote Schuur Hospital Virology Laboratory depended almost exclusively on serological assays. Such assays are of limited value for infants under 18 months of age, as maternal lgG antibody to HIV-1 is transferred via the placenta and may persist in the baby for up to 18 months. Available lgG antibody tests do not distinguish reliably between passively acquired maternal antibody and that produced by the infant itself. A valuable method of establishing the presence of true infection is provided by the polymerase chain reaction (PCR) technique which allows the identification, and subsequent exponential amplification of low levels of specific viral nucleic acid using specific oligonucleotide primers. A major aim of this study was to develop and instigate a (PCR) assay for the early diagnosis of HIV infection in infected infants. This was successfully achieved by the adaptation and optimization of an existing standard PCR protocol to suit the specific needs of a routine diagnostic service. Preliminary requirements involved the selection of primers and probes and establishing optimal parameters for: ionic strength, Taq DNA polymerase concentration, primer concentration, deoxynucleotide triphosphate concentration, and hybridization conditions for most efficient functioning of the test. The devised method entailed the extraction of proviral DNA from peripheral blood mononuclear cells, amplification of HIV-1 specific sequences by PCR, and identification by Southern blot hybridization with digoxigenin (DIG)-labelled probes. Thereafter the efficacy of the assay was tested on 45 infants (under 15 months of age) all born to seropositive mothers and therefore at risk for HIV infection. Forty-two of these infants had antibodies to HIV-1 and the remaining 3 were seronegative. The latter 3 also tested negative for HIV proviral DNA when PCR was performed, using at least 2 different HIV-1 primer pairs and their respective DIG-labelled probes. However, 27 (64%) of the 42 seropositive infants were also HIV-PCR positive and the remaining 15 (36%) seropositive infants were negative for HIV proviral DNA. Positive PCR tests correlated well with clinical data indicative of active HIV-1 infection for the majority of infants in the neonatal period, although it could not provide proof of infection in newborn babies (less than 1 week of age). The development of an in-house PCR protocol specific for HIV-1 has not only provided a valuable diagnostic assay for neonatal infection, but has also given insight into the parameters required for high sensitivity and the stringent precautionary measures that need to be applied to avoid contamination problems. The second part of this study was devoted to DNA sequence analysis of cloned HIV isolates from an infected mother and her 3-month-old infant. Nucleotide sequence variation between isolates of HIV-1 has been well documented. Examination of the third variable region (particularly the V3- loop) in the env gene of HIV-1 of our mother-infant pair confirmed this variation and provided the first genetic epidemiological data of this nature in the local community. Proviral DNA from both mother and baby was amplified using V3-specific degenerate primers and cloned. Clones containing the insert DNA were 2 identified by colony-blot hybridization. Their nucleotide and amino acid sequences were analyzed by using various computer programs. The degree of similarity between variants from the mother and infant in this study differed to a large extent from previous studies. The virus population harboured by the mother displayed highly homogeneous V3 sequences (1,04% variation) compared to the isolates from her 3-month-old infant, which showed a higher degree (1,8%) of heterogeneity. Phylogenetic analysis of the different isolates from mother and infant demonstrated that an HIV-1 subtype C virus was the infectious agent. This classification was confirmed by the characteristic amino-acid sequence of the tetrapeptide motif of the V3 loop present in the isolates from both mother and infant as well as the absence of a potential N-linked glycosylation site proximal to the first cysteine of the V3 loop, which is characteristic of subtype C viruses. Based on the amino acids present at positions 306 and 320 of the V3 loop, it could also be concluded that isolates from both the mother and her baby were consistent with the non-syncytium inducing (NSI) phenotype of HIV-1, thus indicating that, contrary to popular belief, NSI variants can be responsible for initiating infection. Data obtained from these genetic investigations of variants involved in vertical transmission of HIV-1 can form a useful basis for future comparative studies.