Correlates of Protective Viruses Damaging to HIV Infection
our report did not justify such additional considerations. Similarly, while more sequence data, from this and other loci, will be required before the full evolutionary history of Y chromosomes and of our species can be deciphered, our report was both an attempt to initiate this evolutionary reconstruction and an example of how the absence of variation represents an evolutionary signal in its own right. Finally, we wish to clarify a point raised by Rogers et al. Although the most parsimonious
... e that can be derived from our data does in fact place the chimpanzeehuman split after the branching off of the gorilla lineage (supported by two characters), we were careful to state, in note 10 of the report, that the next shortest tree describes an unresolved trichotomy. When we calculated an expected mutation rate for this intron (note 11), we assumed such a trichotomy, and used independent estimates of branching times of 5MY for the chimpanzee-human, gorillahuman, and chimpanzee-gorilla splits (14MY for the splitting off of orangutan). We then averaged over all possible pairwise comparisons to obtain a mean mutation rate. Given the small number of changes tak-Correlates of P Damaging to Barton F. Haynes et al. (I) state correctly that concentrations of human immunodeficiency virus (HIV) are low and of cytotoxic T lymphocytes (CTLs) are high in people who are "nonprogressors." Therefore, they argue, our proposals-that HIV is essentially not a lytic virus and that immunosuppression may be caused by virus-specific CD8+ T cell-mediated immunopathology that destroys infected antigen-presenting and T cells-do not apply. This is an incorrect conclusion drawn from our views, because the example of the nonprogressor with a low HIV load and high CTL response does fit into our balance-scheme between the two extremely rare cases that Haynes et al. quote from our proposal (2). If efficient CTL killing (plus neutralising antibody) eliminates HIV completely before it can be integrated into many cells, HIV negativity and immunity will result. If high CTL activity (plus antibody) controls infection early and efficiently, long-term nonprogres-rc sion will result (with potential incubation times of more than 30 years). If the balance is in the middle, the average of 8 to 10 years 1362 ing place along the branches and nodes of this gene tree, our data should not be used in a molecular clock form to estimate the age of the interspecific splits, as was done by Rogers et al. If one considers only the numbers of changes, the observed numbers (5, 10, and 11) for the human-chimpanzee, human-gorilla, and chimpanzee-gorilla comparisons, respectively, are not significantly different from the 8, 8, and 8 expected from a trichotomy (X2 = 2.75).