Introduction

Urinary scent marking is an important means of chemical communication in mammals. Examples include rodents (Ralls, 1971; Novotny et al., 1990), canids (Jorgenson et al., 1978; Raymer et al., 1984), and red deer (Cervus elaphus) (Bakke and Figenschou, 1990), among others. Urinary scent marks may convey multiple messages, ranging from the species chemosignals and individual-recognition compound patterns to aggression and dominance signaling to various primer pheromone activities. Chemical identification of various scent constituents has become essential in elucidating the physiological and behavioral responses pertaining to olfaction. Such activities have thus far been most successful with investigations in the house mouse (Mus domesticus) where distinct chemical compounds have been linked to endocrinology and behavior (for review, see Schwende et al., 1986; Novotny et al., 1990), whereas their direct action on the vomeronasal neurons has also been verified (Leinders-Zufall et al., 2000; Sam et al., 2001).

Hamsters have been yet another frequently studied rodent in terms of olfaction. Here most studies have centered on investigations of flank gland markings (in both genders) and female vaginal secretions. Gland marking behavior of the Syrian golden hamster (Mesocricetus auratus) has been well established (Johnston, 1975, 1985), whereas the dense and viscous urine of this desert animal is generally viewed as less important in terms of chemical communication. Vaginal secretions of the Syrian golden hamster are also involved in a distinct communication function (Johnston, 1974; Singer et al., 1976) containing aphrodisin (a protein of the lipocalin family). However, it is improbable that aphrodisin serves a direct pheromonal function (Singer et al., 1987; Singer and Macrides, 1990; Vincent et al., 2001); it is more likely that it functions as a pheromone carrier (Briand et al., 2000).

Among the additional members of the hamster family, such as Phodopus sungorus (P. sungorus), P. campbelli, and P. roborowsky, fewer studies have accumulated to date. Secretions of the ventral gland and supplementary sacculi have recently been characterized chemically (Burger et al., 2001a,b). Additional glands and secretions may also be used in chemical communication. Unlike with the Syrian hamster, the Phodopus hamsters seem to involve urine odors (Lai et al., 1996). In a recent study of the effects of cross-fostering between P. campbelli andP. sungorus (Vasilieva et al., 2001), the young hamster appeared to learn quickly the olfactory cues from family members. Whereas we do not know currently the origin of the scent cues used in the adaptation process, the urinaryodor (if composed of chemically distinct molecules) would be suggested. Earlier studies (Laska and Hudson, 1995) demonstrated that—at least in the squirrel monkey (Saimiri sciureus)—urine contains a considerable amount of information of a potential signal value. Consequently, the purpose of this study was tocharacterize the volatile urinary constituents of P. campbelli and P. sungorus with respect to gender and age. In addition, due to the availability ofafew P. roborowsky urinary samples, a comparison was also made with this species.

Whereas species and individual recognition may involve quantitative arrays of different chemosignals, it was deemed necessary to employ a highly quantitative technique for the extractions of urinary organic constituents. Based on the previously described stir bar extraction procedure for aqueous media (Baltussen et al., 1999, 2002), we have recently adapted this methodology to quantitative profiling of volatile and semivolatile compounds in biological media (Soini et al., 2005). Its excellent reproducibility has permitted reliable determinations of differences in volatile urinary profiles of different hamster species in this study. Numerous chemically distinct profile constituents were subsequently identified through a combined capillary gas chromatography–mass spectrometry and quantified with an element-specific detector.

Methods and Materials

Animals and Age Groups.

All three species of the genus Phodopus (P. campbelli, P. roborowsky, and P. sungorus) were available; gender, ages, and number of animals are listed in Table 1. Animals were born and raised in captivityand kept in indoor rooms in solid-bottom, polycarbonate cages (30 × 15 × 15cm) with wood-chip bedding material. For all animals, the bedding material wasidentical (autoclaved, natural wood chips from pinewood). Bedding material was not subjected to chemical analyses. Since Phodopus hamsters are social animals, they were kept in groups (three to five animals) of the same sex or in family units (a pair and its litter) for several weeks. The colony was maintained on a 14-hr light/10-hr dark light cycle with lights off at 10:00 hr; temperature was 21 ± 1°C. All animals had free access to hamster chew and water. Sunflower seeds, fruits, and lettuce were occasionally provided as a dietary supplement.

Table 1 Codes and Ages for the Phodopus Hamster Groups
Full size table

In P. campbelli and P. sungorus, the males and females represented different age groups. In Phodopus species, sexual maturation is reached between postnatal days 30 and 45. During the first 4 mo of life, animals may be regarded as young adults; animals between 9 and 14 mo-old are in their prime. Life span lasts for up to 2 yr.

Sample Collection.

Urine was collected from all three hamster species (University of Tübingen, Germany). To collect urine, animals were removed from colonies and kept individually in small metabolic cages until they produced about 1 ml of urine, or up to 4 hr. If an animal did not produce enough urine in one sampling session, the procedure was repeated the following day. Estrous female hamsters were not subjected to urine collection (estrous state was checked regularly). Samples were kept frozen until analyzed.

Sample Preparation.

All glassware was washed with distilled water and acetone and dried at 80°C in the oven. Volatile and semivolatile compounds were extracted from 1.0 ml of undiluted urine by sorptive extraction with a Twister PDMS polymer-coated stir bar (10 mm, 0.5-mm film thickness, 24 μl PDMS volume, Gerstel GmbH, Mülheim an der Ruhr, Germany) for 60 min. Stirring speed was 800+ rpm on the Variomag Multipoint HP 15 stirplate (H+P Labortechnic, Oberschleissheim, Germany). After 60-min extraction time, a stir bar was rinsed with a small amount of distilled water, dried gently on the paper tissue, and was placed in the glass injector liner for mass spectrometry (MS) identification or in the TDSA autosampler tube for a gas-chromatographic (GC) quantification.

Mass Spectrometry.

A Finnigan MAT Magnum ion trap gas chromatograph–mass spectrometer (GC-MS) system was used for the compound identification (Finnigan MAT, San Jose, CA, USA). The system was provided with a DB-5 capillary column (30 m × 0.25 mm, i.d., 0.25-μm film thickness, J&W Scientific, Folsom, CA, USA). Helium carrier gas head pressure was 12 psi. At the beginning of the column, a loop of uncoated deactivated silica tubing (30 cm × 0.25 mm, i.d.) was attached by using a universal Press-Tight Connector (Restek Corporation, Bellefonte, PA, USA) as described earlier (Ma et al., 1999). The loop was cooled with liquid nitrogen, while the Twister stir bar was held in the injector liner for 15 min at 250°C for the thermal desorption of the analytes. Subsequently, the desorbed compounds were cryotrapped into the liquid nitrogen cooled loop. After removing liquid nitrogen cooling, the GC temperature was held at 40°C for 5 min and increased to 200°C at the rate of 2°C/min. The final temperature was held for 10 min. The manifold and transfer line temperatures were 220 and 300°C, respectively. The ion trap was operating in the positive electron ionization mode. Spectra were scanned from 40 to 350 msu (1 scan/sec).

Gas Chromatography.

Gas-chromatographic (GC) equipment for the quantitative analysis consisted of an Agilent GC Model 6890 with an Atomic EmissionDetector (AED) Model G2350A (Agilent Technologies Inc., Wilmington, DE, USA) and a Thermal Desorption Autosampler (TDSA, Gerstel GmbH). Theseparation capillary was DB-5 (30 m × 0.25 mm, i.d., 0.25-μm film thicknessfrom J&W Scientific Folsom, CA, USA). Samples were thermally desorbed in a TDSA automated system, followed by injection into the column with a cooled injection system CIS-4. TDSA operated in a splitless mode. Temperature program for desorption was 20°C (0.5 min), then 60°C/min to 280°C(10 min). Temperature of the transfer line was set at 280°C. CIS wascooled with liquid nitrogen to −60°C. After desorption and cryotrapping, CIS was heated at 12°C/sec to 280°C with the holding time of 10 min. Temperature program in the GC was 40°C for 5 min, then increasing to 200°C at the rate of 2°C/min. The final temperature was held for 10 min. Carrier gasheadpressure was 14 psi (flow rate 1.2 ml/min). The GC unit was operated inthe constant flow mode. The emission lines for carbon (193-nm), sulfur (181-nm), and nitrogen (174-nm) were monitored in the atomic plasma emission detection.

Statistical Analysis.

Pirouette Lite (Infometrix, Inc., Woodinville, WA, USA) was used for exploratory multivariate analysis to obtain hierarchical cluster patterns for classification of analysis data and for establishment of chemical relations within the subject groups. When appropriate, group comparisons were calculated using the nonparametric Mann–Whitney U test.

Results and Discussion

According to the literature (Sokolov and Vasilieva, 1993; Vasilieva et al., 1990) and the results presented here, P. campbelli and P. sungorus seem closely related. Yet, they are widely considered as two distinct species. Therefore these two species were compared against each other. P. roborowsky will be discussed separately.

I. Campbelli and Sungorus Groups.

Campbelli and sungorus hamster groups each consisted of 10 males and females (Table 1). Preliminary screening showed that both groups exhibited relatively similar chemical constituents in their profiles. Differences were seen mainly in the levels of compounds within the same age and gender groups. Gender and age inflicted certain profile differences, which will be discussed in more detail below.

The compounds identified by GC-MS in the adult animals are shown in Table 2A. A short list of compounds (numbers 1–17) in Table 2B indicates substances that have been used for quantitative measurements by TDSA-GC-AED. A typical MS total ion chromatogram (TIC) of the male hamster urine is shown in Figure 1. Peak numbers refer to Table 2B.

Table 2A Identified and Partially Identified Compounds in Urine of Male P. sungorus and P. campbelli Hamsters (Age 10–11 mo)
Full size table
Table 2B Compounds Quantified by GC-AED
Full size table
Fig. 1
figure 1figure 1

A GC-MS total ion chromatogram (TIC) of the SBSE-extracted urine of a male P. sungorus hamster. Analytical conditions are described in the text. Numbers refer to identified compounds in Table 2B used for quantification by GC-AED.

Full size image

Compound identification was made by GC-MS based on retention times, spectra, and known standard compounds. Based on the total ion chromatogram (TIC) profiles (as shown in Figure 1), the compound profiles were quantitatively compared with the GC-AED profiles, whereas the peak identities were assigned for these measurements. Figure 2 shows a typical GC-AED compound profile (carbon 193-nm line) for a male campbelli hamster.

Fig. 2
figure 2figure 2

A urinary volatile profile of a male P. campbelli hamster (cm-9) by GC-AED, carbon line 193-nm. Separation conditions are described in the text.

Full size image

Within the complex urinary compound profiles, average levels were calculated for 17 identified compounds (Table 2B). Averages were grouped pertaining to the different species/gender/age groups. Calculations were based on the corresponding integrated peak areas for each compound. Logarithmic (log10) transformations log(peak area + 1) were used to normalize the graphs due to large numerical values of the integrated peak areas (Zar, 1999).

In general, male hamster urine contained higher levels of all compounds in both campbelli and sungorus species when compared with females. Typically, urinary profiles in campbelli and sungorus males were closely related. Figure 3 shows average levels of 17 compounds (from Table 2B) for campbelli females (9 mo) and males (9 mo) and sungorus females (10–12 mo) and males (10–11 mo).

Fig. 3
figure 3figure 3

Averages of levels for compound numbers 1–17 for mature female and male campbelli and sungorus hamsters. Logarithmic values for peak areas were used [log(peak area + 1)].

Full size image

Several quantified compounds were either gender- or age-specific in campbelli and sungorus. Also, some of the compounds were typical for the particular hamster species. Hamster groups campbelli females (cf), campbelli males (cm), sungorus females (sf), and sungorus males (sm) were further divided into subgroups based on their ages. The group codes and number of individuals in each group are shown in Table 1. Summary of the identified species/gender/age-specific compounds is shown in Table 3.

Table 3 Summary of Gender- and Age-Specific Compounds and Their Averagea Values for P. campbelli and P. sungorus Groups Quantified by GC-AED
Full size table

Multivariate hierarchical cluster analysis (HCA) in Figure 4 shows “similarity degrees” of different hamster groups based on the averages of 17 quantified compounds. Clusters connected closest to 1.0 on the X-axis mark the most similar group properties. As shown in Figure 4, volatile profiles of male campbelli and sungorus hamsters for 17 identified compounds relate closely to each other. Within the corresponding female profiles, relation is not so clear.

Fig. 4
figure 4figure 4

A hierarchical cluster analysis (HCA) graph for all hamster groups using levels of compounds 1–17 as variables (see Table 2B).

Full size image

Additional comparative qualitative information was obtained from compound profiles by GC-AED sulfur 181-nm line chromatograms (Figure 5). GC-AED facilitated detection of sulfur compound profiles at trace picogram levels (Soini et al., 2005). Characterization by GC-MS was not successful for these compounds due to low levels. As noted for the GC-AED carbon 193-nm line chromatograms (average results in Figure 3), also sulfur-containing, yet unidentified compounds appeared at higher levels in male hamster urine than in female urine.

Fig. 5
figure 5figure 5

Comparative urinary volatile profiles for pooled (A) male sungorus and (B) female sungorus urine by GC-AED sulfur line 181-nm. Compounds were not identified.

Full size image

Method Precision.

Reproducibility of extraction and the GC analysis was investigated with the pooled male campbelli hamster urine. Four parallel extractions were performed, and the relative standard deviations of peak areas were calculated (RSD). Typical variability for the analytes was between 0.1 and 2.0% (RSD, N = 4) as shown in Table 4. A uniform sorptive polymer coating, automated thermal desorption sample introduction, and a constant flow control in the GC-AED system have all contributed to the acceptable reproducibility of the results when no internal standard was deemed necessary.

Table 4 Reproducibility of 17 Quantified Compounds by GC-AED Exemplified by Determination of Samples of Pooled Male P. campbelli Urine (N = 4)
Full size table

Role of Pyrazines.

In all male campbelli and sungorus urinary profiles, the presence of pyrazines was a dominating factor, as seen in Table 2A. Compounds with numbers 3, 4, 7, 8, 12, 15, and 16 (Table 2B) appeared to be related to gender and age in both hamster groups. Campbelli males showed more male-specific pyrazines (2,5-dimethylpyrazine, a propylpyrazine, a methylpropylpyrazine, a C-6 alkylpyrazine, and alkenyl C-4 and C-6 pyrazines). In the sungorus group, female urine also contained 2,5-dimethylpyrazine and a C-4 alkylpyrazine, however, with the latter appearing at higher levels in sungorus males than in females (P < 0.02). Ethylpropylpyrazine (compound 12) levels were significantly higher in both campbelli (P < 0.002) and sungorus males (P < 0.02) compared with females. Ethylpropylpyrazine levels were also significantly higher in mature sungorus males (10–14 mo) compared with juvenile (1 mo) animals (P < 0.002). In all campbelli and sungorus females, the less volatile alkylated C-6 pyrazines (compounds 15 and 16) were absent (below detection limit). Urinary alkenyl C-4 and C-6 pyrazines (compounds 8 and 15) appeared higher in old male hamsters as well as in younger individuals, although the differences were not statistically significant. Figure 6 illustrates average levels of unsaturated alkenyl pyrazines in the campbelli male urinary profiles in the different age groups. Large differences in pyrazine levels within gender and age suggest that pyrazines may be under endocrinological control and may thus be an important means of chemical communication for both campbelli and sungorus hamsters.

Fig. 6
figure 6figure 6

Averages of alkenyl pyrazine compounds 8 and 15 (see Table 2B) in different male campbelli age groups (ages 4, 9, and 14 mo).

Full size image

Ketones.

Relatively small differences of the levels of 4- and 2-heptanone over the age, gender, and species groups were observed. Figure 7 shows that 2-heptanone levels, as an example (compound 2), in campbelli (9 mo) and sungorus (10–11 mo) males were relatively close to each other. The same applied to campbelli (9 mo) and sungorus (10–12 mo) females (data not shown). Statistically significant higher levels of 4-nonanone were found in male urine compared with females in both campbelli (P < 0.02) and sungorus groups (P < 0.002). These findings suggest that urinary ketones may not carry specific signaling properties as pyrazines do for campbelli and sungorus hamsters, although they may be involved in the creation of the baseline scent for the species.

Fig. 7
figure 7figure 7

Average levels of 2-heptanone (compound 2) in different hamster groups.

Full size image

Alcohols.

Among the identified alcohols, branched 2-undecanol (compound 17) appeared male-specific in both campbelli and sungorus species (Figure 3). Surprisingly, branched 2-undecanol levels were already high in the urinary profiles of young sungorus males. Levels were declining with age. Geraniol (a terpene alcohol, compound 14) levels appeared somewhat higher in male and female campbelli mature hamsters (not a statistically significant difference). Only in the campbelli group, were geraniol levels higher in males than in females (P < 0.02). Young females in both campbelli and sungorus groups lacked geraniol in their profiles. A branched alcohol, 2-nonanol (compound 9), was found in all age, gender, and species groups. Sungorus females showed the lowest levels of 2-nonanol (data not shown) from all groups.

Selected Nitrogen Compounds.

Methylaniline (compound 5), phenylacetonitrile (compound 10), and formanilide (compound 13) were independent variables related to each other based on a hierarchical cluster analysis (HCA, data not shown). Within the male profiles, methylaniline was specific for the sungorus males only (Figure 3). In the male profiles, methylaniline was shown only in the age group of 9–11 mo (not seen in juvenile and 14-mo-old males). Phenylacetonitrile was not seen in the urine of immature female or male sungorus hamsters (1 mo). Among mature sungorus males and females, there were no statistically significant differences, whereas the levels in male mature campbelli were higher than in females (P < 0.02).

Figure 8 shows comparisons of the averages of selected compounds in campbelli (9 mo) and sungorus (10 mo) male urinary profiles. Campbelli average levels tended to be slightly higher than sungorus levels. However, the individual variation in their concentrations was clearly larger in the sungorus group.

Fig. 8
figure 8figure 8

Comparison of averages and standard deviations of selected compound levels in campbelli (cm) and sungorus (sm) urinary volatile profiles.

Full size image

II. P. roborowsky.

Within the roborowsky group (four female subjects), a completely different urinary volatile pattern was seen. A list of identified or tentatively identified roborowsky female urinary compounds is shown in Table 5. Screening of urinary profiles of P. roborowsky revealed that few compounds were in common with campbelli and sungorus (Table 2A). A dominant array of alkyl- and alkenylpyrazines was not present in these samples. Instead, the roborowsky urine featured a group of higher-boiling alkyl quinoxalines. The presence of ketones, alcohols, and esters was characteristic for the group of 60 identified compounds. The origin of ketones is likely to be β-and ω-oxidation of fatty acids. This suggests that metabolic pathways in P. roborowsky hamsters differ markedly from P. sungorus and P. campbelli hamsters. Also, based on the large difference on the chemical constituent types in urine, the baseline scent properties of the roborowsky female hamster are expected to differentiate from those of sungorus and campbelli females.

Table 5 Identified and Partially Identified Compounds in the Urine of Female P. roborowsky (Age 12 mo)
Full size table

In summary, quantitative data were proven highly reproducible using the stir bar sorptive extraction (SBSE) methodology. Typically, relative standard deviations (RSD, N = 4) were 0.1–2.0% for normalized peak areas. This analytical feature provided reliability for the urinary profile comparisons. Ultrahigh sensitivity of the atomic emission detection for sulfur-containing compounds provided extra information about comparative urinary volatile profiles.

The chemical characterization data on P. campbelli, P. sungorus, and P. roborowsky verified relatively similar compound profiles in campbelli and sungorus (males and females), which differed substantially from P. roborowsky (females only, males were not available in the screening study). This suggests that metabolic pathways in campbelli and sungorus hamsters resemble each other but differ substantially from the roborowsky species. In campbelli and sungorus, different substituted pyrazines dominated the urinary profiles. Several pyrazines and branched 2-undecanol were male-specific in both species. Methylaniline and phenylacetonitrile were age-specific in both male hamster species. The individual compound level variability within the P. sungorus was clearly larger than in the P. campbelli species.

The urinary profiles of roborowsky female hamsters were dominated by ketones, alcohols, and esters. Similar pyrazine arrays, as seen in campbelli and sungorus, were clearly not observed. Instead, low-volatility alkyl quinoxalines were detected.

One could hypothesize that urinary compound classes such as pyrazines, within a certain volatility range (early eluting pyrazines vs. later eluting pyrazines), may classify the overall perception of the urine odor so that a closely related species may learn the scent codes, as reported (Vasilieva et al., 2001) under cross-fostering conditions. The question remains which urinary compounds carry most crucial information and whether concentration level differences play a role in scent recognition between campbelli and sungorus. At this time, it is not known whether there is any effect of a seasonal variation on the urinary profiles.