that the MHC may be paramount in the chemosensory ... The rat initiates each trial by interrupting ... Fig. 1). To initiate a trial, the trained rat broke the photobeam.
Proc. Nati. Acad. Sci. USA
Vol. 82, pp. 4186-4188, June 1985 Genetics
Chemosensory recognition of mouse major histocompatibility types by another species (olfactometer/transfer of training/odor phenotypes/olfaction/rats) G. K. BEAUCHAMP*, K. YAMAZAKI*, C. J. WYSOCKI*, B. M. SLOTNICKt, L. THOMASf, AND E. A. BOYSEf *Monell Chemical Senses Center, Philadelphia, PA 19104; tDepartment of Psychology, American University, Washington, DC 20016; and *Memorial Sloan-Kettering Cancer Center, New York, NY 10021
Contributed by Lewis Thomas, February 22, 1985
Mice can recognize one another by individuABSTRACT ally characteristic body scents that reflect their genetic constitution at the extremely polymorphic major histocompatibility complex (MHC) of genes on chromosome 17. Reproductive behavioral manifestations of this sensory communication system include MHC-related mating preferences and neuroendocrine responses that affect preimplantation pregnancy and arise from the MHC-related scent of alien males. We have shown previously that mice can be trained in a Y maze to distinguish the scents of urine of congeneic mice that differ genetically only at the MHC. By means of an automated olfactometer, we now show that rats also can similarly distinguish the urinary scents of MHC congeneic mice. Thus, the mode of individual recognition that depends on scents determined by MHC genes can operate across species barriers.
Olfactory discrimination of major histocompatibility complex (MHC) types among mice is apparent in spontaneous mating preferences favoring one MHC type over another (1, 2) from discrimination of odors of MHC-dissimilar mice or their urine in a Y maze (3, 4) and from differing incidences of pregnancy block in females exposed to the scent of alien males whose MHC type is similar or dissimilar to the mate (5). In all of these studies, the genetic difference was confined to the MHC or to the H-2 or Qa:Tla regions of the extended MHC (6) by the use of congeneic mice. The multiplicity of genes concerned and the high polymorphism of H-2 imply that the MHC may be paramount in the chemosensory marking of individuals according to genotype. The present study was conducted to determine whether MHC-associated odors enable members of one species to identify individuals of another species. A genetic component in the scent of human individuals was inferred by Kalmus in his experiments with identical twins and other subjects tracked by dogs (7), but no particular genetic locus was implicated. As a first step in the study of interspecies chemosensory recognition of individuals by their MHC types, we investigated the ability of trained rats to distinguish the MHC types of MHC congeneic mice by the scent of their urine. This was demonstrated with an automated testing apparatus designed for the rat.
MATERIALS AND METHODS The Apparatus. The odor generator and wind tunnel test chamber were adapted from the design of Slotnick and co-workers (8, 9) (Fig. 1). The odor generator provides two odor streams, which are independently and automatically manifolded into a clean air stream by electrically controlled Teflon body valves according to a set program. In each odor channel, odors were generated by passing air over a Petri dish containing urine collected from one of two panels of H-2
congeneic mice. The test chamber consists of a modified glass funnel fitted with a response bar controlling a waterdelivery mechanism. The wide end of the funnel is connected to an exhaust fan, and the conical end is connected to the odorizing system. The rat initiates each trial by interrupting a photobeam positioned over the neck of the chamber and contacts a bar to receive the drop of water provided if the odor presented is the odor assigned for reinforcement. Subjects and Urine Donors. Five adult male and three adult female inbred W/Fu (Wistar/Furth) rats were trained. All were sexually naive, maintained in separate cages after weaning, and aged from 3 months to 1 year during the studies described. They were given commercial rat food ad lib, but their water intake was restricted to about 15 ml per day. Table 1 gives details of the paired panels of mice used as urine donors. They were maintained under uniform conditions in one animal room and were individually numbered for systematic random sampling. The number of mice per panel ranged from 13 to 38. Urine samples were collected in metabolic cages from four or five mice per cage. The pooled urine was immediately frozen and stored at -20'C until needed. It was then thawed to room temperature, placed in Petri dishes, used within 3 hr, and then discarded. For training, each rat was deprived of water for 23 hr beforehand and received 0.05 ml of water for each correct (S+) response to the reinforced odor in the olfactometer. After each session, 10-12 ml of supplemental water was provided. Training. A discrete-trials "go, no go" successive discrimination training procedure was used (refs. 8 and 9; see Fig. 1). To initiate a trial, the trained rat broke the photobeam at the neck of the chamber and was presented with a brief pulse of one of the two alternative odors. In S+ trials, the correct (rewarded) response was for the animal to touch the bar within 3 sec of termination of the odor pulse, which ended the trial simultaneously with delivery of a 0.05-ml water reward. Touching of the bar during S- trials (nonreinforced odor) neither terminated the trial nor produced a reward during the 3-sec interval. When the rat correctly withheld response during an S - trial, no reward was forthcoming. Positive (S+) and negative (S-) trials were programmed in random order, with the restriction that no more than three of one type occurred in succession and that equal numbers of S+ and S- trials were made during a session. Responding during S- trials and not responding during S+ trials were scored as errors. Each daily session comprised 120-180 trials per rat. Data for S+ and S - trials did not differ significantly and are combined in Fig. 2. Initially all S+ trials were rewarded. To prepare for transfer of training tests (below) once stable responding was evident, the probability of reward on S+ trials was reduced successively from 1.00 to 0.75, then 0.50, then 0.30, and then
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.
Abbreviations: MHC, major histocompatibility complex; strain B6, strain C57BL/6; strain B10, strain C57BL/10.
Proc. Natl. Acad. Sci. USA 82 (1985)
Genetics: Beauchamp et al.
0 manifod 7
flow meter 0
air perlurber ehas
FIG. 1. Olfactometer. Air is pumped into an activated charcoal-containing vessel for cleaning (far left) and then into the initial manifold. Here it is divided into three streams, the top and bottom having access to the odor chambers equipped with removable tops so that urine samples can be easily changed. The center air stream is constantly flowing into the test chamber containing the animal being trained (far right). Flow meters insure that equal amounts of odorized air are drawn from each odor box. Solenoid valves open automatically to allow odorized air to mix (by means of the air perturber) with the center air flowing into the test chamber. At all other times, odorized air is exhausted outside the system. For more details, see text.
0.00 (extinction) for blocks of 12 consecutive trials. For training, four pairs of urine samples were used in each session. Each pair consisted of urine from one metabolic cage collection (four or five mice) of each of the two alternative H-2 types. Four pairs were used so that the rat would learn the class distinction (MHC, H-2) and not sporadic adventitious differences that might distinguish particular groups of mice. When stable responding was achieved, the transfer of training procedure was used with new pairs of urine donor panels of the same inbred and congeneic strains or with H-2-typed homozygous F2 or F3 progeny of the congeneic cross (B6 x B6-H-2k). Transfer of Training. The purpose of transfer of training (4, 6) is to test new panels of donors, representing the same H-2 difference, entirely without reward, to obviate any learning of adventitious odor differences. Typically, each transfer of training session with a given rat began with 60-80 trials with training samples, the last 12 (6 positive and 6 negative) being unrewarded (extinction). Following a series of 12-24 more rewarded trials with the training samples, a series of 12 trials was then made with the new panels (transfer of training, no reward). No more than one series of new samples was
included in each session. At no time was there a reward for any trials with the F2 and F3 panels used to test for segregation of odor phenotype with H-2. The placing of samples was switched every 40-60 trials, particularly to exclude the possibility of interference from residual odors from the training samples preceding the transfer of training samples. In control trials with empty odor boxes, the responses of trained rats were entirely random, showing that responses in these trials were not due to incidental cues such as an auditory signal generated by the apparatus.
RESULTS Disinction of B10 (H-2) vs. B1O.A (H-2) Congeneic Shains. In study 1 of Fig. 2, two rats were trained, a female reinforced for B10 and a male reinforced for B1O.A. In initial training, first with B6 vs. A panels and then with B10 vs. B1O.A congeneic training panels, performance improved to >80%o concordance. The results of extinction (Fig. 2, light stippling; unrewarded trials of the same donor panels interspersed with rewarded trials) indicate that both rats could be trained to distinguish the reinforced H-2 haplotype. The rats also correctly discriminated
Table 1. Age-matched male mouse urine donors
Donor panels Inbred B10 mice* (H-2b) Inbred B1O.A congeneic mice* (H-2a) 2A Inbred B6 micet (H-2b) Inbred B6&H-2k congeneic micet (H-2k haplotype from strain AKR) 2B Inbred B6 mice (H-2h) Inbred B6-H-2k congeneic mice H-2b homozygous segregants (F2bb)t H-2k homozygous segregants (F2kk)t 2C Inbred B6 mice (H-2b) Inbred B&H-2k congeneic mice H-2b homozygous progeny of F2bb segregants (F3bb) H-2k homozygous progeny of F2kk segregants (F3kk) B6, C57BL/6; BiO, C57BL/10. *From The Jackson Laboratory. tFrom Sloan-Kettering Institute. tOf the cross (B6 x B6-H-2k) F2, selected by cytotoxicity typing of lymph node lymphocytes with H-2b anti-H-2" and H-2k anti-H-2b antisera. 1
Beauchamp et al.
Proc. Natl. Acad. Sci. USA 82 (1985) Extinction
a Transfer of training 95% confidence interval
I i~ I CD
0. 0 r_ 0 Q 01
v a) c: o
Number of trials Rat number FIG. 2. Distinction of the scents of H-2 congeneic mice by trained rats. In study 1, the first bar (extinction) of each pair represents unrewarded trials of urine samples from the B10 vs. B1O.A mouse donor panels used in training; the second bar (transfer of training) represents unrewarded trials of samples from new B10 vs. B1O.A panels. In study 2, as for study 1, with B6 vs. B6-H-2k panels, but the second for studies 2B and 2C represent transfer of training from B6 vs. B6-H-2k to F2bb vs. F2kk segregant panels (study 2B) and to F3bb vs. F3kk panels (study 2C) (see Table 1).
between new samples of urine, without reward, from new panels of B10 and B10.A mice not previously encountered [Fig. 2, dark stippling (transfer of training)]. Distinction of H-2b and H-2k Congeneic Strains. In study 2 of Fig. 2, six rats were trained, two males and a female reinforced for B6 and two males and a female reinforced for B6-H-2k. In initial training, first with B6 vs. AKR panels (AKR being the H-2k haplotype donor for the B6-H-2k congeneic strain) and then with B6 vs. B6H-22k congeneic training panels, performance increased to >80%o. Successful transfer of training without reward is shown in Fig. 2 to new B6 and B6-H-2k congeneic donor panels (study 2A), to panels of typed homozygous H_2b and H-2k (B6 x B6-H-2k) F2 segregants (study 2B) and to panels of H_2b and H_2k F3 progeny of typed homozygous F2 segregants (study 2C).
DISCUSSION The nature of the odorants governed by the extended MHC is unknown. Although at least one class I H-2 gene is involved (10), this need not mean that the odorants are structural derivatives of known or unknown MHC products, nor do the differences in odor necessarily require structurally different odorants because distinctive compound odors can be constituted by a set of odorants combined in different proportions. MHC polymorphism is associated with many normal biological variations (see ref. 11), which could be the basis of constitutive quantitative metabolic variations giving rise to compound odors distinctive of genotype (11).
We thank D. Kupniewski, Y. Okada, H. Okada, and K. Suzuki for technical assistance. This work was supported in part by National Institutes of Health Grant GMCA-32096 and National Science Foundation Grant BNS 8201759. E.A.B. is American Cancer Society Research Professor of Cell Surface Immunogenetics. 1. Yamazaki, K., Boyse, E. A., Mike, V., Thaler, H. T., Mathieson, B. J., Abbott, J., Boyse, J., Zayas, Z. A. & Thomas, L. (1976) J. Exp. Med. 144, 1324-1335. 2. Yamazaki, K., Yamaguchi, M., Andrews, P. W., Peake, B. & Boyse, E. A. (1978) Immunogenetics 6, 253-259. 3. Yamazaki, K., Yamaguchi, M., Baranoski, L., Bard, J., Boyse, E. A. & Thomas, L. (1979) J. Exp. Med. 150, 755-760. 4. Yamaguchi, M., Yamazaki, K., Beauchamp, G. K., Bard, J., Thomas, L. & Boyse, E. A. (1981) Proc. Nadl. Acad. Sci. USA 78, 5817-5820. 5. Yamazaki, K., Beauchamp, G. K., Wysocki, C. J., Bard, J., Thomas, L. & Boyse, E. A. (1983) Science 221, 186-188. 6. Yamazaki, K., Beauchamp, G. K., Bard, J., Thomas, L. & Boyse, E. A. (1982) Proc. Nati. Acad. Sci. USA 79,
7828-7831. 7. Kalmus, H. (1955) Anim. Behav. 3, 25-31. 8. Slotnick, B. M. & Ptak, J. E. (1977) Physiol. Behav. 19,
795-802. 9. Slotnick, B. M. & Born, W. S. (1979) Physiol. Behav. 23, 589-591. 10. Yamazaki, K., Beauchamp, G. K., Egorov, I. K., Bard, J., Thomas, L. & Boyse, E. A. (1983) Proc. Nati. Acad. Sci. USA 80, 5685-5688. 11. Boyse, E. A., Beauchamp, G. K. & Yamazaki, K. (1983) Hum. Immunol. 6, 177-183.