Skip to main content
Open Access

Intra- and Interspecific Variation in Production of Bile Acids That Act as Sex Pheromones in Lampreys

Abstract

Pheromones are important sexual signals in most animals, but research into their evolution is largely biased toward insects. Lampreys are a jawless fish with a relatively well-understood pheromone communication system, and they offer a useful opportunity to study pheromone evolution in a vertebrate. Once sexually mature, male sea lamprey (Petromyzon marinus) and likely other lampreys produce and release bile acids that act as sex pheromones. Spawning males do not feed and therefore produce bile acids primarily for sexual communication, whereas larvae produce the same bile acids but for digestion, offering an opportunity to compare the evolution of bile acids produced for sexual versus nonsexual functions. We profiled eight pheromone-related bile acids in livers from larvae and males and determined the effect of life stage on intra- and interspecific variation in bile acid production. Our results indicate less variation among males than larvae within P. marinus but more variation among species for males than larvae. We postulate that bile acid production in males is shaped by directional or stabilizing selection that reduces variance within P. marinus and directional or disruptive selection that promotes diversification across species. Although our results offer support for the role of sexual selection in the evolution of lamprey pheromones, they do not eliminate possible roles of other aspects of lamprey ecology.

Introduction

Animals use a diverse collection of chemical structures as pheromones during mate search and assessment (Wyatt 2014). Pheromone diversity is presumably the result of selection related to species recognition, mating success (Steiger and Stökl 2014), or other aspects of animals’ ecology (Baeckens et al. 2018) as well as neutral processes (e.g., genetic drift). For many animals, selection on visual (Seddon et al. 2013) and acoustic (Wilkins et al. 2013) sexual signaling traits leads to rapid divergence compared with nonsexual traits. However, parallel research on the evolution of pheromones lags behind that on visual or auditory signals (Andersson 1994; Coleman 2009), especially for vertebrates (Symonds and Elgar 2008).

Lampreys offer a useful opportunity to explore the evolution of pheromones (Buchinger et al. 2015). Larval lamprey live in streams and filter feed for several years, after which they metamorphose into spawning adults (nonparasitic species) or into parasitic juveniles that feed on fish in rivers, lakes, or oceans for approximately 1.5 yr before spawning. Prespawning sea lamprey (Petromyzon marinus) migrate into streams following bile acids and other molecules released as metabolic waste by stream-resident larvae (Teeter 1980; Sorensen et al. 2005; Li et al. 2018a). Once sexually mature, males release several bile acids that act as sex pheromones (Li et al. 2002, 2013, 2017a, 2017b, 2018c). Other lamprey species produce similar odors (including bile acids) as larvae (Fine et al. 2004; Robinson et al. 2009; Yun et al. 2011; Stewart and Baker 2012; Buchinger et al. 2013) but use sex pheromones that are partially distinct from those used by P. marinus (Buchinger et al. 2017b, 2017c). Notably, P. marinus larvae and sexually mature males synthesize many of the same bile acids (Li et al. 1995; Sorensen et al. 2005; Brant et al. 2013, 2016; Buchinger et al. 2013). However, bile acids produced by larvae are nonsexual traits, whereas bile acids produced by sexually mature males (which do not feed and have a degenerate digestive system) are sexual traits that affect access to mates (Buchinger et al. 2017a).

Here, we report intra- and interspecific comparisons of variation in production of bile acids with sexual versus nonsexual functions. Specifically, we profiled eight known bile acids in livers of larvae (nonsexual traits) and sexually mature males (sexual traits) across several species of lamprey. We then determined the effect of life stage on (1) among-individual variation within P. marinus and (2) among-species variation. Sexual signaling traits are generally hypothesized to be under (1) directional or stabilizing selection that reduces variance within species and (2) directional or disruptive selection that promotes diversification across species (West-Eberhard 1983; Löfstedt 1993). Therefore, we predicted that sexually mature male bile acid profiles would be more similar than larval bile acid profiles within P. marinus but less similar than larval bile acid profiles across species. Although liver bile acids are not direct measures of pheromone signals, they allow useful comparisons despite the lack of specific information on pheromone identity in most lampreys. Furthermore, we reasoned that the bile acids produced in the liver are important sexual traits because (1) the liver (which synthesizes bile acids) is essentially the pheromone gland of lamprey and a major determinant of pheromone release (Buchinger et al. 2017a) and (2) known liver bile acids are either sex pheromones or pheromone precursors for P. marinus (Li et al. 2018b).

Methods

General Approach

We quantified liver concentrations of eight pheromone-related bile acids in larvae and sexually mature males. Michigan State University’s Institutional Animal Use and Care Committee approved all sampling (approval 4/10-043-00, 02-13-040-00). Larvae were captured via backpack electroshocking, and males were captured using traps, backpack electroshocking, fyke nets, or by hand. All males were sexually mature, as determined by expression of milt upon gentle pressure to the abdomen. The sex of larvae was not assessed but may remain labile up until metamorphosis into the parasitic phase (Lowartz and Beamish 2000; Beamish and Barker 2002). Larvae were fed baker’s yeast; males were not fed because they cease feeding several months before spawning. Notably, bile acid synthesis is not expected to be affected by diet (Hofmann et al. 2010). Lamprey were euthanized with an overdose of 3-aminobenzoic acid ethyl ester (MS222; Sigma, St. Louis, MO) and liver tissues frozen at −80°C. Later, liver tissues were weighed and, if needed, a subsample of approximately 100 mg collected as a cross section of the middle-posterior region. Whole livers were used for nonparasitic species and larvae because they are approximately 100 mg or less.

Bile acids were then extracted using established methods (Brant et al. 2013). Briefly, samples were spiked with 5 ng [2H5] 3kPZS as an internal standard, homogenized in 1 mL of 75% ethanol∶water (v:v), incubated and shaken at room temperature and 70 rpm overnight, and centrifuged at 15,800 g and 4°C for 20 min. The supernatant was then freeze-dried, reconstituted in 50% methanol∶water (v:v), and bile acids quantified using ultrahigh performance liquid chromatography with tandem mass spectrometry (Li et al. 2011; Brant et al. 2013). The compounds 3-keto petromyzonol sulfate (3kPZS), petromyzonol sulfate (PZS), petromyzonamine monosulfate (PAMS), petromyzonamine disulfate (PADS), petromyzosterol disulfate (PSDS), 3-keto petromyzonol (3kPZ), petromyzonol (PZ), and 3-keto allocholic acid (3kACA) were quantified. All of these compounds are either known pheromones or suspected pheromone precursors in P. marinus (Li et al. 2018b). Concentration data were standardized by the subsample mass. All analyses were conducted using R programming language version 3.4.3 (R Development Core Team 2017) and two-tailed tests with α=0.05.

Variation in Bile Acid Profiles within Petromyzon marinus

To determine the effect of life stage on among-individual variation within species, we profiled liver bile acids in larval and sexually mature male P. marinus. Livers from 100 larvae (mean ± SE; body mass=1.3±0.05 g, total length=100.01±1.08 mm) and 100 males (body mass=202.05±6.32 g, total length=456.51±4.5 mm) were sampled as described above. Larvae for this experiment were collected in late May and early June 2016 and held in 1,000-L tanks supplied with Lake Huron water at Hammond Bay Biological Station until they were dissected on September 28, 2016. Males were dissected on July 13–21, 2016. Relative among-individual variation for each compound was compared between larvae and males using asymptotic tests for the equality of coefficients of variation (cvequality asymptotic_test function in R; Feltz and Miller 1996; Marwick and Krishnamoorthy 2016). Multivariate among-individual variation was compared between larvae and males using a permutational ANOVA (PERMANOVA) on Bray-Curtis dissimilarity indices (Vegan adonis function; permutations = 1,000). We also tested the dissimilarity indices for homogeny of variances (Vegan betadisper function), although heterogeneous variances have little effect on PERMANOVA results for balanced designs (Anderson and Walsh 2013).

Variation in Bile Acid Profiles across Lamprey Species

In a second experiment, we determined the effect of life stage on similarity of bile acid profiles across species. On the basis of availability, larvae were sampled in six species and sexually mature males in 10 species (table 1). First, ANOVA and post hoc Tukey’s tests were used to compare log transformed concentrations of different bile acids within each species and life stage. Second, bile acid compositions were compared between species within life stage using multivariate factor analyses. Bile acid composition data were arcsine square root transformed and reduced to factors using the factanal function in R. Scree plots were used to determine the number of factors to extract. MANOVA tests followed by post hoc t-tests with Benjamini and Hochberg (1995) adjustments were used to evaluate differences in factor scores. Third, interspecies variation in the composition (C1/sum of C1–C8) of each compound was compared between larvae and males using asymptotic tests for the equality of coefficients of variation (cvequality asymptotic_test function in R; Feltz and Miller 1996; Marwick and Krishnamoorthy 2016). Fourth, multivariate interspecies variation was compared between larvae and males using a PERMANOVA on Bray-Curtis dissimilarity indices. The focus of this analysis was among-species variation, not among-individual variation; hence, we used bile acid concentrations averaged across individuals (larva n = 3–11, male n = 4–11, per species), with the replication of interest being at the species level (species of larva n=6, species of male n=10). As heterogeneous variances and unbalanced designs can affect the results of PERMANOVAs, we tested for homogeny of variances and analyzed the data first with data from all species and second with only the species for which we sampled both larvae and males (Anderson and Walsh 2013).

Table 1. 

Liver bile acid sampling from across life stage and species of lampreys

Family, genus, and speciesStageLocationNo.Weight (SE; g)Length (SE; mm)
Geotridae:     
Geotria australisMaleCanterbury, New Zealand4164.36 (22.21)488.25 (26.21)
 LarvaWaikato, New Zealand11.84 (.05)75.91 (2.14)
Petromyzontidae:     
Ichthyomyzon unicuspisMaleMichigan1056.13 (6.8)280.9 (10.27)
Ichthyomyzon fossorMaleMichigan93.25 (.3)113.71 (1.48)
 Larva 102.55 (.22)114.7 (2.88)
Ichthyomyzon castaneusMaleMichigan734.43 (2.03)250.57 (6.52)
Petromyzon marinusMaleMichigan10204.6 (18.26)461.3 (10.74)
 Larva 101.49 (.2)101.1 (4.33)
Lethenteron appendixMaleMichigan103.72 (.31)130.7 (2.18)
 Larva 103.32 (.18)130.9 (2.05)
Lethenteron camtschaticumMaleJilin, China897.05 (9.57)384.13 (10.26)
Lethenteron reisneriMaleLiaoning, China86.88 (.63)165.0 (5.67)
 Larva 32.72 (.25)109.67 (5.55)
Lethenteron moriiaMaleLiaoning, China1127.27 (1.1)255.46 (4.64)
 Larva 107.6 (.44)177.9 (5.09)
Lampetra aerypteraMaleIndiana86.98 (.62)153.25 (3.98)

a Note that the classification of Lethenteron morii remains disputed, and Potter et al. (2015) considers the species to be Eudontomyzon.

View Table Image

Results

Variation in Bile Acid Profiles within Petromyzon marinus

Within Petromyzon marinus, among-individual variation in bile acids was generally lower for males than for larvae. Coefficients of variation were significantly lower in males than in larvae for 3kPZS (D’AD=5.99, χ2 df=1, P=0.014), PAMS (D’AD=18.37, χ2 df=1, P<0.001), PADS (D’AD=6.03, χ2 df=1, P=0.014), and PZ (D’AD=8.79, χ2 df=1, P=0.003); not significantly different for PZS (D’AD=0.004, χ2 df=1, P=0.952) and 3kPZ (D’AD=1.12, χ2 df=1, P=0.291); and significantly higher in males for 3kACA (D’AD=7.20, χ2 df=1, P=0.007; fig. 1). PSDS was not detected in samples from males. The Bray-Curtis dissimilarity indices were slightly but significantly higher for males (median=0.282) than for larvae (median=0.278; PERMANOVA F1,198=290.85, P<0.001) and had homogenous variances (F1,198=0.0026, P=0.959). However, box plots revealed many outliers, and the difference in medians was trivial (male=0.282, larva=0.278); hence, we serially reran the model after removing data for each compound. Only PZS resulted in the outliers and had a major influence on the results. When the data for PZS were excluded, the Bray-Curtis dissimilarity indices were significantly higher for larvae (median=0.677) than for males (median=0.413; PERMANOVA F1,198=132.71, P<0.001; fig. 2) and had heterogeneous variances (F1,198=78.284, P<0.001).

Figure 1. 
Figure 1. 

Lower coefficients of variation in liver bile acids among males than larvae in Petromyzon marinus. The bile acids 3-keto petromyzonol sulfate (3kPZS), petromyzonol sulfate (PZS), petromyzonamine monosulfate (PAMS), petromyzonamine disulfate (PADS), petromyzosterol disulfate (PSDS), 3-keto petromyzonol (3kPZ), petromyzonol (PZ), and 3-keto allocholic acid (3kACA) were quantified in liver tissues from 100 larva (filled bars) and 100 sexually mature males (open bars). Asymptotic test for the equality of coefficients of variation was used to evaluate differences in the relative variance between larva and males for each compound. Asterisks indicate P<0.05.

Figure 2. 
Figure 2. 

Lower dissimilarity in liver bile acids among males than larvae in Petromyzon marinus. The bile acids 3-keto petromyzonol sulfate, petromyzonamine monosulfate, petromyzonamine disulfate, petromyzosterol disulfate, 3-keto petromyzonol, petromyzonol, and 3-keto allocholic acid were quantified in liver tissues from 100 larva and 100 sexually mature males. Differences in Bray-Curtis dissimilarity indices between larva and males were evaluated using a permutational analysis of variance (PERMANOVA; Vegan adonis function in R). Data for petromyzonol sulfate are not included because they resulted in many outlier dissimilarity indices. Asterisk indicates P<0.05. Bar represents the median, box the interquartile range, whiskers the range between each quartile and the maximum or minimum data point, and circles the individual indices.

Variation in Bile Acid Profiles across Species

Bile acid profiles were less similar across species for males than for larvae. The most concentrated bile acid produced by larvae in all species was PZS, whereas the relative concentrations of all bile acids produced by males differed across species (fig. 3). Among-species factor analyses indicated more overlap of larval bile acid profiles than male bile acid profiles (table 2). With all species included, coefficients of variation were significantly higher in males than in larvae for PZS (D’AD=8.35, χ2 df=1, P=0.004) but no other compound (3kPZS D’AD=0.82, χ2 df=1, P=0.366; PAMS D’AD=0.002, χ2 df=1, P=0.966; PADS D’AD=0.045, χ2 df=1, P=0.831; PSDS D’AD=0.906, χ2 df=1, P=0.341; PZ D’AD=0.028, χ2 df=1, P=0.868; 3kPZ D’AD=0.386, χ2 df=1, P=0.534; 3kACA D’AD=0.015, χ2 df=1, P=0.904; fig. 4). With all species included, Bray-Curtis dissimilarity indices were significantly higher for males (median=0.849) than for larvae (median=0.53; PERMANOVA F1,14=3.22, P=0.02; fig. 5) and had heterogeneous variances (F1,14=5.74, P=0.031). The results were similar when we included only species for which we sampled both life stages; male dissimilarity was higher (median=0.934) than larvae (median=0.53; PERMANOVA F1,10=3.57, P=0.006), although the variances were not significantly different (F1,10=2.37, P=0.154).

Figure 3. 
Figure 3. 

Bile acid concentration in liver tissues from sexually mature males in 10 species of lamprey and from larvae in six species of lamprey. The bile acids 3-keto petromyzonol sulfate (3kPZS), petromyzonamine monosulfate (PAMS), petromyzonamine disulfate (PADS), petromyzosterol disulfate (PSDS), 3-keto petromyzonol (3kPZ), petromyzonol (PZ), and 3-keto allocholic acid (3kACA) were quantified, and the log transformed concentrations compared within species and life stage using ANOVA and post hoc Tukey’s tests (α=0.05). Error bars signify standard error of the mean.

Table 2. 

Results from factor analyses on effect of species on bile acid profiles in larva and male lampreys

Stage and factorVarianceSpecies effectSpecies overlap
FdfPGroup aGroup bGroup cGroup dGroup e
Larva:         
1.2561.855, 48.121Ga, If, Pm, Lap, Lr, Lm    
2.17521.985, 48<.001If, Pm, Lap, Lr, LmGa   
3.152.005, 48.1Ga, If, Pm, Lap, Lr, Lm    
Male:         
1.24328.949, 75<.001Ga, Lc, Lr, LmIu, If, Ic, Lap, LaeIu, If, Pm, Lap, Lae  
2.2247.299, 75<.001Lap, Lc, Lr, LmLc, LaeGa, Ic, Lap, Lr, LmGa, If, Ic, Pm, Lr,Ga, If, Ic, Pm
3.1017.039, 75<.001Ga, If, Ic, LcGa, Iu, Lap, Lc, LrIu, Pm, Lap, Lr, Lm, Lae  

Note. Variance is the proportion of variance explained by each factor. Species effect is the significance of species effects on factor scores, as determined using a MANOVA. Species overlap is the grouping of species based on factor scores as determined using pairwise t-tests followed by Benjamini and Hochberg adjustments for multiple comparisons. Species abbreviations: Ga = Geotria australis; Iu = Ichthyomyzon unicuspis; If = Ichthyomyzon fossor; Ic = Ichthyomyzon castaneus; Pm = Petromyzon marinus; Lap = Lethenteron appendix; Lc = Lethenteron camtschaticum; Lr = Lethenteron resneri; Lm = Lethenteron morii; Lae = Lampetra aeryptera.

View Table Image
Figure 4. 
Figure 4. 

Among-species variation in proportional concentrations of eight pheromone metabolites for males and larvae. The bile acids 3-keto petromyzonol sulfate (3kPZS), petromyzonol sulfate (PZS), petromyzonamine monosulfate (PAMS), petromyzonamine disulfate (PADS), petromyzosterol disulfate (PSDS), 3-keto petromyzonol (3kPZ), petromyzonol (PZ), and 3-keto allocholic acid (3kACA) were quantified in liver tissues in larvae (filled bars) of six species and sexually mature males (open bars) of 10 species. Data are the percent concentrations of each compound. Asymptotic test for the equality of coefficients of variation were used to evaluate differences in the relative variance between larva and males for each compound. Asterisks indicate P<0.05.

Figure 5. 
Figure 5. 

Higher dissimilarity in liver bile acids across species for males than for larvae. The bile acids 3-keto petromyzonol sulfate, petromyzonol sulfate, petromyzonamine monosulfate, petromyzonamine disulfate, petromyzosterol disulfate, 3-keto petromyzonol, petromyzonol, and 3-keto allocholic acid were quantified in liver tissues in larvae of six species and sexually mature males of 10 species. Differences in Bray-Curtis dissimilarity indices between larva and males were evaluated using a permutational analysis of variance (PERMANOVA; Vegan adonis function in R). Asterisk indicates P<0.05. Bar represents the median, box the interquartile range, whiskers the range between each quartile and the maximum or minimum data point, and circles the individual indices.

Discussion

Our results indicate distinct patterns of variation in bile acids produced by sexually mature males compared with those produced by larvae for nonsexual functions. Specifically, liver bile acids varied less among males than larvae within Petromyzon marinus but more for males than larvae across species. Lampreys do not feed for several months before spawning, and high rates of bile acid synthesis are related to pheromone signaling in sexually mature male P. marinus. Therefore, we postulate that the observed variation in production of bile acids is directly related to variation in the pheromones released. Alternatively, different bile acid profiles may be evidence for parallel biosynthetic pathways that result in release of the same mixture of bile acids. However, intra- and interspecific variation in male odors supports our position that variation in liver bile acids correlates to variation in male odors (Buchinger et al. 2017a, 2017c). Notably, the intra- and interspecific patterns differed for individual compounds, supporting previous observations that male odor is composed of multiple components with different functions (Johnson et al. 2012). Although most of the measured bile acids are either pheromones or precursors for P. marinus (Brant et al. 2013; Li et al. 2018b), our results should be interpreted bearing in mind uncertainties regarding their specific functions across species. Regardless, our results indicate distinct evolutionary patterns for male versus larval bile acids. We consider the potential role of sexual selection as well as some alternatives.

Sexual selection can drive diversification of sexual signaling traits (Panhuis et al. 2001; Ritchie 2007; Kraaijeveld et al. 2011; Maan and Seehausen 2011; Scordato et al. 2014; Servedio and Boughman 2017). One approach to parse the effects of sexual selection and other processes on diversification is to compare divergence rates of traits used for sexual communication and traits used for nonsexual functions. For example, Seddon et al. (2013) report an association between the strength of sexual selection and divergence rate of sexual plumage traits (but not nonsexual traits) in passerine birds. Similar studies on insects report greater divergence in sex pheromones or associated glands compared with nonsignaling traits (Weber et al. 2016; Weiss et al. 2017). Our interspecific results indicate greater divergence in bile acids produced by males compared with larvae and therefore support a role of sexual selection in shaping sex pheromone diversification.

Alternatively, interspecific variation in male bile acid production might be related to differences in feeding or migratory ecology. The ecology of larval lampreys is highly conserved across species; larvae of all species burrow in stream sediment and filter feed on detritus and microorganisms for several years before metamorphoses (Dawson et al. 2015). In contrast, postmetamorphic lampreys exhibit considerable interspecific variation in feeding and migratory ecology; some lampreys migrate to lakes or oceans to parasitize on fish and marine mammals, some parasitize on fish in rivers, and others skip the parasitic phase entirely and spawn after the larval phase (Potter et al. 2015). Our study includes three anadromous parasites (P. marinus, Geotria australis, and Lethenteron camtschaticum), one lacustrine parasite (Ichthyomyzon unicuspis), two riverine parasites (Ichthyomyzon castaneus and Lethenteron morii), and four nonparasitic species (Ichthyomyzon fossor, Lethenteron appendix, Lethenteron reisneri, and Lethenteron aeryptera). Therefore, although all lampreys spawn in similar habitats (Johnson et al. 2015), the greater among-species variation for males than larvae may be residual from differences in feeding or migratory ecology. However, we suggest that sexual selection is a more likely driver of diverged bile acid profiles in sexually mature males because (1) more of the bile acid production by parasitic phase lamprey appears to occur in the intestine than in the liver (Yeh et al. 2012), (2) lampreys cease feeding for several months before spawning, and (3) the synthesis of primary bile acids is unlikely influenced by diet (Hofmann et al. 2010). Our data do not appear to group by feeding or migratory ecology but—with only one to four representatives from each strategy—is insufficient for a formal statistical test. A broader set of lampreys with different ecologies would allow more direct tests of how sexual selection and other aspects of ecology shape pheromone evolution.

Sexual selection and ecology often shape the evolution of sexual traits interactively (Servedio and Boughman 2017). Environmental and social conditions affect signal transmission and perception, the benefits accrued by choosing particular mates, and the attributes that influence male mating success (Endler 1992, 1993). For example, the environments inhabited by lacertid lizards appear to influence the design of their chemical signals (Baeckens et al. 2018), whereas sexual selection on its own may not (Baeckens et al. 2017). In lampreys, different mating systems (polygynous, polygynandry, monogamy) could shift the strength and direction of sexual selection (Johnson et al. 2015; Baker et al. 2017). For example, communal spawning in most lampreys may exert selective pressure on sperm production, whereas lek-like spawning in P. marinus may shift selective pressure onto pheromone production. Divergence of bile acid profiles might also be the result of selection for pheromones that better distinguish suitable mates from heterospecific males (reproductive character displacement; Butlin 1987). Indeed, heterospecific interactions appear to shape the evolution of sex pheromones in butterflies (Bacquet et al. 2015). Most lamprey spawn in sympatry with at least one other lamprey species (Johnson et al. 2015) and do not produce viable offspring when mated with certain heterospecifics (Piavis et al. 1970; but see Hume et al. 2013 for an example of viable hybridization). However, widespread observations of spawning nests shared by multiple lamprey species (Johnson et al. 2015) indicates that pheromones either are not fully species specific (Buchinger et al. 2017c) or are not effective barriers among species. Although we hypothesize that female mate choice and male competition have some role in the diversification of lamprey pheromones, their effects are almost certainly intertwined with other aspects of lamprey ecology.

We hypothesize that stabilizing and positive sexual selection shapes bile acid production in male P. marinus. Animals that rely on species-specific pheromones for mate recognition (such as moths) often use mixtures of several compounds and face strong selective pressure against releasing blends that deviate from an optimal mixture and are therefore less attractive to narrowly tuned mates (Cardé and Baker 1984; Groot et al. 2016). Therefore, our observation of lower among-individual variation for males compared with larvae within P. marinus may be the result of selection against males with extreme pheromone blends. Whether lampreys are narrowly tuned to a species-specific pheromone blend remains unknown, although female P. marinus respond to pheromones released by other species (Buchinger et al. 2017c). Reduced variation may also be the result of linear preferences that impose directional selection (Löfstedt 1993). For example, male Drosphila serrata face directional selection on pheromone signaling because of linear female preferences (Chenoweth and Blows 2004). Similarly, female P. marinus prefer the higher of two adjacent pheromone concentrations (Johnson et al. 2009; Buchinger et al. 2017a) and conceivably exert positive selection on bile acid production in males. Interestingly, the relative pheromone compositions might be under stabilizing selection for species specificity and the absolute production under directional selection for higher release rates. Stabilizing and directional selection act on different levels of pheromone communication in some animals (i.e., Watts et al. 2004), but mate choice assays are needed to determine whether similar selective regimes shape pheromone evolution in sea lamprey.

In conclusion, we report possible signatures of sexual selection in production of bile acids that act as sex pheromones in lampreys. However, further investigation of the role of pheromones in mate choice for lampreys and a larger set of species that allows specific tests of the role of phylogenetic and life history differences is needed. Regardless, our results help address biases in the animal communication literature, in which research on the evolution of vertebrate pheromones remains relatively scarce (Symonds and Elgar 2008; Coleman 2009).

The US Fish and Wildlife Service Marquette Biological Station and Don Jellyman assisted with lamprey collection. Elizabeth Buchinger, Michael Connolly, Scott Couture, Skye Fissette, Brooklyn Idalski, and Elizabeth Stieber assisted with sampling and extractions. Two anonymous reviewers provided useful suggestions on an early version of the manuscript. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the US Government. This work was supported by the Great Lakes Fishery Commission, the New Zealand Ministry of Business Innovation and Employment (contract CO1X1002), and the Shanghai Ocean University and Michigan State University Joint Research Center Program (A1-0209-13-0805).

Literature Cited