Some behavioral consequences of H-2:Qa:Tla polymorphism

Abstract

The conjunction of a penetrating idea and a novel observation is an experimentalist’s delight. At about the time that Lewis Thomas was writing on histocompatibility genes and olfaction it was brought to my attention that the social behavior of certain mice of our colonies at Sloan-Kettering Institute suggested that they could sense differences in the H-2 haplotypes of their companions. In short, males and females of dissimilar H-2 type appeared to associate with one another in preference to partners of the same H-2 type. Since the mice in question were congenic, differing only at the H-2 region of chromosome 17, the inference was that H-2 or linked genes are concerned in a sensory communication system among mice. The upshot of succeeding studies was the demonstration of spontaneous mating preferences which, at least among congenic laboratory mice, generally favor H-2 disparity (l-4). Subsequent proof that chemosensory recognition is involved in the identification of H-2 types among mice came from the introduction of a Y-maze testing system which has been exploited with gratifying effect at the Monell Chemical Senses Center in Philadelphia (5). In this method, mice are trained to distinguish the arms of a Y-maze scented by odors of alternative sets of H-2 congenic mice. Training consists in prior withholding of water and provision of a drop of water each time the trainee mouse selects whichever of the two scents is assigned for reinforcement. It was found that urine is a potent source of H-2 distinctive odors (6), and thus that olfaction, which is a function of primary sensory neurons whose bodies are situated in the olfactory mucosa, sufficiently explains the discrimination of H-2 types. We have asked whether the genetic constitution of the odor phenotype is restricted to H-2 as defined by its K and D gene poles. This is not so. Recombinant congenic strains which reduce the genetic difference to the Qa:Tla region adjacent to H-2 can be distinguished in the Y maze, and disparity restricted to the region of H-2K alone is also distinguishable (7). Thus the odor phenotype corresponding to a given H-2