Chemical Senses Advance Access published online on July 5, 2008
Chemical Senses, doi:10.1093/chemse/bjn035
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Mice Recognize Recent Urine Scent Marks by the Molecular Composition
Department of Human Anatomy and Physiology, University of Padua, Padua, Italy
Correspondence to be sent to: A. Cavaggioni, Dipartimento di Anatomia e Fisiologia Umana, Università di Padova, Via Marzolo 3, 35100 Padova, Italia. e-mail: andrea.cavaggioni{at}unipd.it
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Male mice mark the territory with urine scent marks that are frequently renewed to maintain the territory ownership. We measured the response of male mice to small spots of urine deposed either 0, 5, 11, 22, 45, 90 min, or 24 h before testing and show that mice loose interest in sniffing scent marks as they become older and older. We asked what scent features tell a mouse how recent a scent mark is, and therefore, we studied the molecule-to-behavior relationship by correlating 6 behavioral variables—the number of sniffing acts, the latency to the first sniff, the number of urine marks, the latency to the first mark, the area of the marks, and the number of fecal pellets—to 2,4-dehydro-exo-brevicomin, linalool, 2-sec-butyl-4,5-dihydrothiazole, 2,4-dimethylphenol, 4-ethylphenol, and 6,10-dimethyl-5,9-undecadien-2-one released from urine spots over the time, identified, and quantified by gas chromatography and mass spectrometry. Canonical correlation between the molecular and the behavioral principal components was strong (R(1)= 0.96, P = 0.026). The principal component based on 2,4-dehydro-exo-brevicomin, linalool, and 2-sec-butyl-4,5-dihydrothiazole correlated negatively with countermarking and positively with the sniffing behavior, suggesting a semantic feature of fresh male mouse urine.
Key words: behavior, mouse, olfaction, scent, time
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Most mammals have a highly developed olfactory sense which they employ in social communication using chemical signals originating in urine, faeces, or skin scent glands (Bossert and Wilson 1963
; Thiessen and Rice 1976; Burger 2005). A large share of the mammalian chemical advertising behavior centers around the active deposition of organic chemicals; this implies specialized motor patterns and delivery systems to deposit the chemical signals on environmental objects or other animals of the same species (Ralls 1971; Mucignat-Caretta et al. 2004).
Nocturnal habits and dark living environments have led to the evolution of olfaction as a major method of communication among mice. Their olfactory signals originate mainly from urine. Male mice very frequently scent mark the environment with spots of urine to prove their ability to maintain ownership over an area; they also countermark spots released by other mice invading their territory, to make sure that their scent marks are the most recent in that area. Males which fail to countermark any competing scent marks can no longer ensure that their own marks are the most recent and, thus, indicate their inability to prevent or overcome challenges to their dominance (Hurst 1990; Rich and Hurst 1998). Hence, the olfactory communication by means of recent scent marks is supposed to give an advantage and enforce the position of a male mouse in a "scent war" (Hurst and Beynon 2004). Recent studies have focussed on the role of mammalian scent molecules on social, sexual, maternal behaviors as to the relevance of olfactory communication (Burger 2005). To a large extent, the studies did not take into consideration the "ageing" process of the scent marks brought about by evaporation, diffusion, and possibly chemical changes of the urine molecular components, when the marks are left in the field exposed to the air.
The volatile molecules of urine confer a characteristic scent. Several chemical scent molecules have been identified in the urine of adult male mice (Holland et al. 1984; Novotny et al. 1985; Ma et al. 1999). Moreover, a protein family abundant in male mouse urine, the major urinary protein complex (MUP), binds and hence concentrates scent molecules in urine and slowly releases them into the air, thus prolonging the lifetime of the scent (Bacchini et al. 1991; Böcskei et al. 1992; Robertson et al. 1996; Jemiolo et al. 2002; Sharrow et al. 2002; Armstrong et al. 2005). The scent molecules dissociate from MUP and are released in the air with a different time course (Cavaggioni et al. 2006).
The scent molecules released from urine induce a sequence of behavioral responses in mice consisting in an early exploration by repetitively sniffing and contacting the urine followed by countermarking (Novotny et al. 1990; Jemiolo et al. 1992; Mucignat-Caretta and Baldini 1998; Ma et al. 1999; Mucignat-Caretta 2002). The scent molecules bound and released by MUP are a stimulus for behavioral responses to explore and countermark (Hurst et al. 1998; Mucignat-Caretta and Caretta 1999a, 1999b; Mucignat-Caretta 2002; Cavaggioni et al. 2003). Interestingly, mice prefer to explore and countermark recent scent marks (Hurst 1990; Cavaggioni et al. 2003). Thus, the timely release of scent molecules from MUP could work as a clock for mice, to "tell the time" of the spot release. Also, golden hamsters (Wilcox and Johnston 1995) are able to time scent marks and this ability may be more widespread among mammals than presently appreciated.
The study of the diffusion rates of scent molecules in the air and the extent to which these differences influence the outcome of behavioral tests in mammals is a neglected area due to the difficulty of analyzing the diffusion rate of scent molecules. Solid phase microextraction (SPME), however, is an extraction method that can be used with confidence to study the diffusion of molecules in the air (Arthur and Pawliszyn 1990). SPME from the headspace of scent marks was first used by Burger et al. (1997) to analyze the diffusion rate of volatile pheromone molecules from the preorbital secretion of the South African antelope, Oreotragus oreotragus. We recently adopted SPME to identify 9 scent molecules and analyze their diffusion from spots of mouse urine (Cavaggioni et al. 2006).
In addition, a number of molecules diffusing from scent marks in the air may be perceived holistically as a complex unity that influences a number of behaviors. We here use the release in the air of 6 molecules from urine spots of different "age" and canonical correlation to describe a multivariate relationship with the mouse behavior (Dillon and Goldstein 1984).
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Subjects
All procedures were authorized by a decree of the Italian Ministry of Health (154/2003-B) and were performed according to the European laws on animal experimentation and handling. Twelve adult male Swiss mice, Mus musculus, between 24 and 28 weeks old (body weight 38.25–47.00 g) were born and reared with both parents and littermates in the animal facility of the department in standard polycarbonate cages (42 x 26 x 15 cm) with water and food pellets (Altromin, Riper Co., Bolzano Italy) ad libitum. The mice were weaned at 21 days of age and housed 6 per cage with same sex and age cage mates up to beginning of the experiment. Cage bedding was made of wood shavings changed twice a week according to standard animals’ room procedure. The animals’ room temperature was 25 ± 1 °C, with relative humidity 70% and 6 air renewals per hour. The light cycle, with 12-h light on, started at 0500 hours. Each mouse was housed isolated in the polycarbonate cages for at least 30 days before the experiments. Mice were isolated in order to increase the excretion of pheromones under androgen control (Lombardi et al. 1976) and hence to reduce the number of mice employed in the tests. Isolated mice are not more stressed than group-housed mice (Hunt and Hambly 2006). On completion of the experiment, the mice were sacrificed with excess anesthesia, 3 times the surgical dose (ketamine 225 mg/kg body weight and xylazine 60 mg/kg body weight) given with an intraperitoneal injection according to the suggestions of the ethical committee and legal requirements.
Urine collection and spot deposition time
Urine was collected daily from the 12 mice for 14 days. A mouse was placed for 2 h in a cage with a mesh grid fixed 1 inch (2.54 cm) above a clean polycarbonate floor, and the urine was collected with a plastic pipette as soon as it was spontaneously voided on the floor. The urine was kept over ice during the collection and was stored at –20 °C. Mice were randomly assigned to 2 groups, A and B, balanced for litter and body weight. The urine samples of group-A mice were then pooled and so were the urine samples of group-B mice in order to minimize individual variability, up to a volume of 7.5 ml from group-A mice and 6.0 ml from group-B mice, a volume sufficient for the chemical analysis of the molecules and the behavioral tests. The 2 pools were spun down with a centrifuge (1130 x g, 10 min, 25 °C), split into 0.3-ml aliquots, and stored at –20 °C. The behavioral tests and the chemical analysis were then carried out within 6 months. In order to avoid self-odors, in the behavioral tests, group-A mice were exposed to the group-B urine and group-B mice were exposed to the group-A urine.
A glass tile (10 x 10 x 1 cm) covered with filter paper was placed 1 cm from the short walls of a polycarbonate cage (42 x 26 x 15 cm), and a 15 µl spot of urine (15.2 ± 0.5 µl) was laid down in the center of the tile in a dedicated room with no air drifts (temperature 20 ± 2 °C and relative humidity 72%). It was then left exposed to the air for 0, 5, 11, 22, 45, or 90 min or, as an end point, for 24 h, to obtain a 0-min-old spot, a 5-min-old spot, and so on.
Behavioral tests
The urine was presented to mice to mimic progressively older spots. A repeated measures experimental design was used. The mice were taken from their home cage and tested in a transparent polycarbonate cage (42 x 26 x 15 cm) whose floor was lined with Bench Guard (Bibby Sterilin, Staffordshire, UK) and equipped with 2 glass tiles (10 x 10 x 1 cm) covered with filter paper placed 1 cm from the short walls of the cage. The mice were tested with urine aged for a different time (0-, 5-, 11-, 22-, 45-, 90-min-, and 24-h-old spots) in random order, different for each mouse, once a day and between 0800 hours and 1100 hours. Two tests were made. In double choice tests, a tile was spotted with 15 µl of male urine and the other tile with water or both the tiles were spotted with the same urine. In a blank experiment, the 12 mice were tested with both tiles spotted with 15 µl of water in order to exclude a side preference. The 180 tests (2 tests x 12 mice x 7 urine spots, a blank test x 12 mice) were done in a testing room under dim light. The mice were observed for 10 min by the same observer. During the test, the observer recorded 5 variables for each tile: The latency to the first sniffing act (sniffing) within 1 cm of the spot and lasting at least 2 s and the number of the sniffings as indexes of exploratory behavior; the latency to the first urine voiding and the number of the voidings on the tile (1 drop or streak of urine) as indexes of countermarking behavior, and the number of fecal pellets on the cage floor as an index of autonomic activation. Subsequently, the area of the spots was measured: a digital record of the fluorescence of the spots under UV light was taken and the area was determined on a personal computer using the software Corel Photo Paint 12 (www.corel.com) and Scion Image 4.02 software (www.scioncorp.com). In the end, 6 behavioral variables were evaluated.
Data analysis
Analysis of behavior
Each behavioral variable was separately analyzed for each time of exposure of urine to the air with the Wilcoxon test. The trend of responses in the spot-age groups was analyzed with regression analysis. The value for statistical significance was P 
0.05.
The data sets
The 6 behavioral variables observed on the K = 7 experimental groups (spot age), with each observation being repeated on 12 mice, formed the behavioral data set. The K experimental groups were defined by the postdeposition time (0, 5, 11, 22, 45, 90 min, or 24 h) of the test urine, K = K(t). The observations were the number of sniffing acts (yi), the latency to the first sniff (yii), the number of urine marks (yiii), the latency (yiv), the area (yv) of the marks, and the number of fecal pellets (yvi). The data set of the chemical variables was the measure of the release xI of 2,4-dehydro-exo-brevicomin (I), xIII of linalool (III), xIV of 2-sec-butyl-4,5-dihydrothiazole (IV), xV of 2,4-dimethylphenol (V), xVI of 4-ethylphenol (VI), and xXI of 6,10-dimethyl-5,9-undecadien-2-one (XI), from K = 7 samples of the test urine defined by the postdeposition time, with each measure being made on the urine pool A and on the urine pool B. The measurements were made by gas chromatography and mass spectrometry of the headspace of 15 µl spots of urine and the compounds are numbered according the gas chromatographic retention time as described in Cavaggioni et al. (2006).
Factor analysis
Canonical correlation analysis proceeded in sequential stages. A first step reduced the dimensions of the behavioral and the chemical data sets by extracting the principal components. A second step analyzed the canonical correlation between the behavioral and the chemical principal components. The statistical approach is outlined by the boxes and arrows of Figure 3 of the Results.
|
Principal components
Principal component (PC) analysis is based on the equation
 |
(i) are the roots of R from the greatest to the smallest, I is the identity matrix, and 
(i) are the characteristic vectors expressing the coefficients of the variables in the PC (Dillon and Goldstein 1984). Hence, the first PC of the vector of the behavioral data Y = Y(t) is b-PC(1) = (1) Y, the second is b-PC(2) = (2) Y, and so on, where (i) are the characteristic vectors. A similar analysis with the vector X = X(t) of the data of the chemical variables gives a first chemical c-PC(1) = 
(1) X, a second c-PC(2) = (2) X, and so on, where (i) are the characteristic vectors.
Canonical correlation analysis
Canonical correlation here analyses the relationship between the b-PCs and the c-PCs by finding 2 linear combinations (canonical variates or CVs), one for the c-PCs with coefficients (weights) b(i) and one for the b-PCs with coefficients a(i) (weights), such that their ordinary correlation is as large as possible. The first canonical root is
 |
The programs Statistica 5.1 (StatSoft, Tulsa, OK) and SPSS 12.0 (SPSS, Chicago, IL) were used and the varimax orthogonal rotation was applied to the principal components.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Behavioral data
The mice did not prefer a particular tile when both tiles were scented with urine or both were blank tiles. During the test, the mice sniffed the tile scented with a spot of urine more than the control tile spotted with water (Wilcoxon test, P < 0.05), Figure 1 (sniffing number and time to the first sniff, means, and standard error of the mean [SEM]). The number of sniffing decreased monotonically with the spot age, 0, 5, 11, 22, 45, 90 min, and 24 h. The t(1/2) of the sniffing rate was approximately 2 h. The Pearson's product-moment correlation of the sniffs with the trace age was r = –0.27 (95% confidence interval –0.47 to –0.039) and the correlation was significant (P < 0.05).
|
The outcomes of the countermarking are shown in Figure 1 (urine deposition number, time to the first deposition, area of the urine trace, means, and SEM). Analysis of the differences in response within the groups showed that, relative to water, the mice countermarked significantly faster and with a larger area the 22-min-old spots; in addition they released more fecal pellets (Wilcoxon test, P < 0.05), Figure 1.
In conclusion, exploration depends on the age of the urine declining over time as urine becomes older.
Pearson's correlation
The release in the air of molecules I, III, IV, V, VI, and XI measured as a function of the age of the urine is shown in the Figure 2. The figure shows that I and III were released from urine in the air only for a short period after spot deposition (early molecules), whereas the other molecules were released for a longer time. The ordinary product-moment correlation coefficient r of the data shown in the Figures 1 and 2 described the one-to-one relationship between molecular and behavioral variables. The Pearson's r between the number of sniffs and V was 0.84, between the number of sniffs and XI was 0.81, and 0.75 between the number of pellets and IV (P < 0.05). The correlations prompted an analysis of the relationship between sets of molecules and types of behaviors.
|
Principal components of the chemical data set
We studied the relationship between I, III, IV, V, VI, and XI and the behavioral variables by analyzing sets of variables rather than the single variables. Given the great number of variables, a reduction of the dimensions was made by principal component analysis while retaining as much as possible of the original information. The analysis extracted a first principal component c-PC(1) based on xI – xIII – xIV. The Figure 2 shows that I and III were characterized by an early release. The analysis extracted also a second c-PC(2) based on xV – xVI – xXI. The Figure 2 shows that V, VI, and XI had a longer lasting release. The respective coefficients and loadings (italics) in the c-PCs are shown in the caption of Figure 3. The positive signs on top of the arrows indicated that the variables contributed positively to the size of the PCs. The loadings give the ordinary correlation of the original variable and the respective PC. The negative loading of xVI indicates an inverse relationship. The loadings showed that the molecules had correlations greater than 0.7 (absolute value). The c-PC(1) explained 42% of the original variability, the c-PC(2) about 41%, and together they explained 83% of the original variability.
To extract the behavioral PCs (b-PCs), we used the observations on the test tile from yi to yvi. The analysis extracted a first b-PC(1) based on the set yiii – yiv – yv – yvi that describes the countermarking behavior and a second b-PC(2) based on yi – yii that describes the sniffing behavior. The respective coefficients and the loadings (italics) in the b-PCs are shown near the lower arrows in the scheme of Figure 3. The positive signs on top of the arrows indicate that the variables contributed positively to the size of the PCs. The loading of the latencies is negative indicating that an increase of the b-PC corresponds to a decrease of the latencies. The loadings showed that the behavioral variables had correlations greater than 0.7 (absolute value). The b-PC(1) explained about 48%, the b-PC(2) about 36%, and together they explained 84% of the original variability.
The data of the Figures 1 and 2 are now represented by the PCs with a contraction of the dimensions from 6 + 6 to 2 + 2 and with least information loss. The scores of the PCs in the K = 7 groups that describe the unfolding in time of the PCs are reported in Table 1. In order to detect a relationship between the scores, the correlation between c-PCs and b-PCs was analyzed by canonical correlation.
|
Canonical correlation
Canonical correlation between the c-PCs and the b-PCs extracted a first canonical root with correlation R(1) = 0.959 between the chemical canonical variate c-CV(1) and the behavioral canonical variate b-CV(1). The coefficients of the CVs are reported near the arrows in the upper part of the scheme of Figure 3. The CVs values of the 0-min-old and 5-min-old spots clustered far apart from the older spots, Figure 4. The index of goodness of fit was P = 0.026. The first root extracted 46% of the variance shared by the c-PCs and the b-PCs. The canonical weight of the PCs, expressing the importance of a PC for obtaining a maximum correlation between CVs, is reported near the arrow of Figure 3. The canonical weight of c-PC(2) in c-CV(1) was 1.000 and that of c-PC(1) was –0.017. Thus, the canonical weight of c-PC(2) gave a great contribution increasing the c-CVs(1), whereas the canonical weight of c-PC(1) gave a very small contribution decreasing the c-CVs(1). The canonical correlation between the CVs showed that an increase of the CVs corresponded to decreasing the b-PC(1) of the countermarking behavior (canonical weight –0.871) and to increasing the b-PC(2) of the sniffing behavior (canonical weight 0.489). The coefficients and the loadings (italics) of xI – xIII – xIV in c-PC(2), reported in the lower part of the scheme of Figure 3, did not differ substantially suggesting that each molecule was similarly important.
|
The second canonical root extracted 23.5% of the residual variance shared by the c-PCs and the b-PCs and had a correlation R(2) = 0.684. The index of goodness of fit P = 0.137 indicated a poor fit. The smaller fraction of variance shared by the CVs(2), the lower value of R(2) and the worse fit, showed that the second canonical root was less relevant than the first one.
In order to test the specificity of the molecules used in the analysis, we repeated the analysis with the set of molecules IV, V, VI, VIII, X, and XI, thus excluding I and III and including VIII and X. The Figure 2 shows that the new molecules VIII (indole) and X (2-butyl-1-octanol) are characterized by a longer lasting release than I and III. The canonical correlation was lower and the fit was worse than with the original set (R(1) = 0.886, P = 0.139).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
The aim of the present study is to determine whether the scent molecules of the urine help a mouse in timing scent marks. The present relationship between the scent molecules and the mouse behavior lends credit to this possibility.
The behavior of the male mice observed in the presence of scent marks confirms that male mice strongly respond to scent marks with exploratory behavior and by countermarking (Hurst 1990). The rate of countermarking is less than reported by other authors (Maruniak et al. 1974) although it is within the range of previous experiments in our laboratory. The reason of the difference may be that we scored the scent marks on the surface of the tiles excluding those on the cage floor outside of the tiles. The effect of the age of scent marks on the behavior of mice is not new. Cavaggioni et al. (2003) compared countermarking of recent and old scent marks. At variance with the previous studies, the present data focus on the behavior elicited by traces in the minutes after urine deposition.
Firstly, we show that the sniffing rate decreases with the age of the scent mark, up to 90 min old.
Secondly, we here show the effect on the mouse behavior of a set of 6 scent molecules that comprises 2 early molecules that diffuse rapidly and for a short time from urine spots, 2,4-dehydro-exo-brevicomin (I) and linalool (III), and 4 molecules with a slower diffusion, 2-sec-butyl-4,5-dihydrothiazole (IV), 2,4-dimethylphenol (V), 4-ethylphenol (VI), and 6,10-dimethyl-5,9-undecadien-2-one (XI). The aggressive behavior in response to I and IV (Novotny et al. 1985, 1990) and the countermarking behavior in response to IV stereoisomers (Cavaggioni et al. 2003) were investigated in earlier studies. The present investigation confirms the activity of these molecules and considers also III, V, and XI. Ordinary correlation analysis between the behavioral variables and the molecules, however, gives low correlation indexes and suggests that a statistical analysis that treats the variables simultaneously rather than one by one should be used to unravel a possible relationship. Hence, we first extracted the molecular and behavioral principal components that describe features and maintain as much of the original information as possible while reducing the number of variables. We then found by canonical correlation analysis, the linear composites of the principal components that maximally correlate between them. The canonical correlation obtained is higher than the ordinary correlations determined between the variables. Canonical correlation associates the principal component based on I, III, and IV to reducing the countermarking component and increasing the sniffing component. None of the molecules is more important than the others in the principal component. It is very interesting that both I and III are chemically related to insect pheromones. I is a dehydrogenated form of (+)-endo-brevicomin that is an aggregating pheromone of the southern pine bark beetle Dendroctonus frontalis (Vité et al. 1985; Sullivan et al. 2007) and S-(+)-linalool, a mate attractant pheromone of the solitary bee Colletes cunicularius (Borg-Karlson et al. 2003), is an isomer of III. The principal component based on V, VI, and XI seems less relevant for inducing behavioral modifications in the recipient mouse. The substantial amount of variability, that is, information, shared by the behavioral and scent principal components and accounted for by the behavioral and scent variables, shows a fair description. Thus, the canonical correlation describes a relationship between the urine scent and the mouse behavior giving insights into the mechanisms that mice use to recognize recent scent marks.
Canonical correlation analysis as a way of guiding feature selection toward the underlying semantics was never applied to olfaction before. The present study, however, has a number of limitations. The behavioral variables did not meet the requirements of multivariate normality and homogeneity of the variance for testing the significance of the relationships. Hence, the canonical correlation is here used for descriptive purposes only. Moreover, the number of the groups investigated reduced the number of molecules that could be analyzed with statistical consistency to 6. The 6 molecules considered are a small subset of the urinary molecules that a mouse may conceivably perceive. The SPME extracted some molecules more efficiently than others from the urine headspace depending on the polarity of the molecules. As an example, farnesene isomers were below our detection threshold although their wide-ranging role in the mouse behavior is well demonstrated (Ma et al. 1999). Finally, mice approaching a scent mark often contacted the urine to investigate nonvolatile molecules (Hurst 1990; Nevison et al. 2000); thus, we do not know whether the countermarking behavior is a response to molecules of different volatility.
We here combine behavior, chemistry, and canonical correlation analysis to provide some insight into the specialized perceptual and motivational mechanisms that mice use to respond to the molecules that unfold from scent marks over time—the scent of time.
|
Funding |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
This work was supported by EU contract N. 507610 GOSPEL and by the Padua University
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
We thank Dr Michela Bondì for technical collaboration.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Armstrong SD, Robertson DH, Cheetham SA, Hurst JL, Beynon RJ. Structural and functional differences in isoforms of mouse major urinary proteins: a male-specific protein that preferentially binds a male pheromone. Biochem J (2005) 391:343–350.[CrossRef][Web of Science][Medline]
Arthur C, Pawliszyn J. Solid phase microextraction with thermal desorption fused silica optical fibres. Anal Chem (1990) 62:2145–2148.
Bacchini A, Gaetani E, Cavaggioni A. Pheromone binding proteins of the mouse, Mus musculus. Experientia (1991) 48:419–421.[CrossRef][Web of Science]
Böcskei Z, Groom CR, Flower DR, Wright CE, Phillips SE, Cavaggioni A, Findlay JB, North AC. Pheromone binding to two rodent urinary proteins revealed by X-ray crystallography. Nature (1992) 360:186–188.[CrossRef][Medline]
Borg-Karlson AK, Tengö J, Valterová I, Unelius CR, Taghizadeh T, Tolasch T, Francke W. (S)-(+)-linalool, a mate attractant pheromone component in the bee Colletes cunicularius. J Chem Ecol (2003) 29:1–14.[CrossRef][Web of Science][Medline]
Bossert W, Wilson EO. Chemical communication between animals. Recent Prog Horm Res (1963) 19:673–716.[Web of Science][Medline]
Burger BV. Mammalian semiochemicals. Topics Curr Chem (2005) 249:231–278.
Burger BV, Yang T-P, le Roux M, Brandt WF, Cox AJ, Hart PF. Mammalian exocrine secretions XI. Constituents of the preorbital secretion of klipspringer, Oreotragus oreotragus. J Chem Ecol (1997) 23:2383–2400.[CrossRef][Web of Science]
Cavaggioni A, Mucignat-Caretta C, Redaelli M, Zagotto G. The scent of urine spots of male mice, Mus musculus: changes of the chemical composition over time. Rapid Commun Mass Spectrom (2006) 20:3741–3746.[CrossRef][Web of Science][Medline]
Cavaggioni A, Mucignat-Caretta C, Zagotto G. Absolute configuration of 2-sec-butyl-4,5-dihydrothiazole in male mouse urine. Chem Senses (2003) 28:791–797.
Dillon WR, Goldstein M. Multivariate Analysis (1984) New York: John Wiley & Sons.
Holland ML, Rhodes GR, Wiesler D, Novotny M. Chromatographic profiling of urinary volatile and organic acid metabolites of normal and diabetic C57BL/Ks mice. J Chromatogr (1984) 306:23–37.[CrossRef][Web of Science][Medline]
Hunt C, Hambly C. Faecal corticosterone concentrations indicate that separately housed male mice are not more stressed than group housed males. Physiol Behav (2006) 87:519–526.[CrossRef][Medline]
Hurst JL. Urine marking in populations of wild mice Mus domesticus Rutty. I. Communication between males. Anim Behav (1990) 40:209–222.[CrossRef][Web of Science]
Hurst JL, Beynon RJ. Scent wars: the chemobiology of competitive signalling in mice. Bioessay (2004) 26:1288–1298.[CrossRef][Web of Science][Medline]
Hurst JL, Robertson DHL, Tolladay U, Beynon RJ. Proteins in urine scent marks of male house mice extend the longevity of olfactory signals. Anim Behav (1998) 55:1289–1297.[CrossRef][Web of Science][Medline]
Jemiolo B, Xie TM, Novotny M. Urine marking in male mice: responses to natural and synthetic chemosignals. Physiol Behav (1992) 52:521–526.[CrossRef][Medline]
Jemiolo B, Xie TM, Novotny M. Pheromone binding by polymorphic mouse major urinary proteins. Protein Sci (2002) 11:2247–2256.[CrossRef][Web of Science][Medline]
Lombardi JR, Vandenbergh JG, Whitsett JM. Androgen control of the sexual maturation pheromone in house mouse urine. Biol Reprod (1976) 15:179–186.[Abstract]
Ma W, Miao Z, Novotny V. Induction of estrus in grouped female mice (Mus domesticus) by synthetic analogues of preputial gland constituents. Chem Senses (1999) 24:289–293.
Maruniak JA, Owen K, Bronson FK, Desjardins C. Urinary marking in male house mice: responses to novel environmental and social stimuli. Physiol Behav (1974) 12:1035–1039.[CrossRef][Medline]
Mucignat-Caretta C. Modulation of exploratory behavior in female mice by protein-borne male urinary molecules. J Chem Ecol (2002) 28:1853–1863.[CrossRef][Web of Science][Medline]
Mucignat-Caretta C, Baldini E. Protein-bound male urinary pheromones: differential responses according to age and gender. Chem Senses (1998) 23:67–70.
Mucignat-Caretta C, Bondì M, Caretta A. Endocrine status affects bladder size and postvoid residual urinary volume in mice. Horm Behav (2004) 46:11–18.[CrossRef][Medline]
Mucignat-Caretta C, Caretta A. Chemical signals in male house mice urine: protein bound molecules modulate interactions between sexes. Behaviour (1999a) 136:331–343.[CrossRef]
Mucignat-Caretta C, Caretta A. Urinary chemical cues affect light avoidance behaviour in male laboratory mice, Mus musculus. Anim Behav (1999b) 57:765–769.[CrossRef][Web of Science][Medline]
Nevison CM, Barnard CJ, Beynon RJ, Hurst JL. The consequences of inbreeding for recognizing competitors. Proc R Soc Biol Sci (2000) 267:687–694.[Medline]
Novotny M, Harvey S, Jemiolo B. Chemistry of male dominance in the house mouse, Mus domesticus. Experientia (1990) 46:109–113.[CrossRef][Web of Science][Medline]
Novotny M, Harvey S, Jemiolo B, Alberts J. Synthetic pheromones that promote inter-male aggression in mice. Proc Natl Acad Sci USA (1985) 82:2059–2061.
Ralls K. Mammalian scent marking. Science (1971) 171:443–449.
Rich TJ, Hurst JL. Scent marks as reliable signs of the competitive ability of mates. Anim Behav (1998) 56:727–735.[CrossRef][Web of Science][Medline]
Robertson DH, Cox KA, Gaskell SJ, Evershed RP, Beynon RJ. Molecular heterogeneity in the major urinary proteins of the house mouse Mus musculus. Biochem J (1996) 316:265–272.[Web of Science][Medline]
Sharrow SD, Vaughn JL, Zidek L, Novotny MV, Stone MJ. Pheromone binding by polymorphic mouse major urinary proteins. Protein Sci (2002) 11:2247–2256.[CrossRef][Web of Science][Medline]
Sullivan BT, Shepherd WP, Pureswaran DS, Tashiro T, Mori K. Evidence that (+)-endo-brevicomin is a male-produced component of the Southern pine beetle aggregation pheromone. J Chem Ecol (2007) 33:1510–1527.[CrossRef][Web of Science][Medline]
Thiessen D, Rice M. Mammalian scent gland marking and social behavior. Psychol Bull (1976) 83:505–539.[CrossRef][Web of Science][Medline]
Vité JP, Billings RF, Ware CV. Southern pine beetle: enhancement or inhibition of aggregating response mediated by enantiomers of endo-brevicomin. Naturwissenschaften (1985) 72:99–100.[CrossRef][Web of Science]
Wilcox RM, Johnston RE. Scent counter-marks: specialized mechanisms of perception and response to individual odors in golden hamsters (Mesocricetus auratus). J Comp Psychol (1995) 109:349–356.[CrossRef][Web of Science][Medline]
Accepted 5 June 2008
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||