Spermine in semen of male sea lamprey acts as a sex pheromone

Abstract

Semen is fundamental for sexual reproduction. The non-sperm part of ejaculated semen, or seminal plasma, facilitates the delivery of sperm to the eggs. The seminal plasma of some species with internal fertilization contains anti-aphrodisiac molecules that deter promiscuity in post-copulatory females, conferring fitness benefits to the ejaculating male. By contrast, in some taxa with external fertilization such as fish, exposure to semen promotes spawning behaviors. However, no specific compounds in semen have been identified as aphrodisiac pheromones. We sought to identify a pheromone from the milt (fish semen) of sea lamprey (Petromyzon marinus), a jawless fish that spawns in lek-like aggregations in which each spermiating male defends a nest, and ovulatory females move from nest to nest to mate. We postulated that milt compounds signal to ovulatory females the presence of spawning spermiating males. We determined that spermine, an odorous polyamine initially identified from human semen, is indeed a milt pheromone. At concentrations as low as 10-14 molar, spermine stimulated the lamprey olfactory system and attracted ovulatory females but did not attract males or pre-ovulatory females. We found spermine activated a trace amine-associated receptor (TAAR)-like receptor in the lamprey olfactory epithelium. A novel antagonist to that receptor nullified the attraction of ovulatory females to spermine. Our results elucidate a mechanism whereby a seminal plasma pheromone attracts ready-to-mate females and implicates a possible conservation of the olfactory detection of semen from jawless vertebrates to humans. Milt pheromones may also have management implications for sea lamprey populations.

Fig 1. The seminal content and olfactory potency of spermine.

(A) The concentration of spermine in sea lamprey milt and ovarian fluid with eggs (mean ± SEM, n = 6), measured with UHPLC–MS/MS. (Inset) Structure of spermine. (B) Percentage of spermine by weight in the milt components, seminal plasma and sperm, determined with UHPLC–MS/MS (mean ± SEM, n = 6). (C) Semi-logarithmic plot of normalized EOG amplitude (mean ± SEM) elicited by spermine (10⁻¹⁵ M to 10⁻⁶ M) in pre-ovulatory female (n = 10) and pre-spermiating male (n = 11) sea lampreys. The EOG response to spermine at each concentration was blank-subtracted and normalized to the response of 10⁻⁵ M _L-ARG (standard) for each fish. (D) Representative EOG traces of pre-ovulatory female olfactory epithelia exposed to spermine at concentrations between 10⁻¹⁵ and 10⁻⁶ M. The number above each trace is the logarithmic value of the molar concentration of each stimulant. The bar above the _L-ARG trace on the left represents the duration of odorant treatment. (E) EOG traces of pre-spermiating male olfactory epithelia exposed to spermine. Underlying data are available in S1 Data. blank, vehicle solution; EOG, electro-olfactogram; _L-ARG, _L-arginine; N.D., not detectable; UHPLC–MS/MS, ultrahigh performance liquid chromatography-tandem mass spectrometry.

Fig 2. Adult sea lamprey behavioral responses to milt and spermine in a two-choice maze.

The index of preference (mean ± SEM) was calculated as the value of [Ae ÷ (Ae + Be)–Ac ÷ (Ac + Bc)], where Bc is the cumulative amount of time spent in the control channel before odorant application, Be is the cumulative amount of time spent in the experimental channel before odorant application, Ac is the cumulative amount of time spent in the control channel after odorant application, and Ae is the cumulative amount of time spent in the experimental channel after odorant application in a two-choice maze (S2 Fig; see Materials and methods). A positive index value indicates attraction, and a negative index value indicates repulsion. Significance was evaluated using a Wilcoxon signed-rank test. *p < 0.05. **p < 0.01. The number in the parentheses indicates the number of test subjects spending more time in the experimental channel out of the total sample size. Milt was applied to produce a final spermine concentration of (1) 2.2 × 10⁻¹⁴ M or (2) 3.8 × 10⁻¹⁴ M in the maze. Underlying data are available in S1 Data.

Fig 3. Spermine activated sea lamprey TAAR348.

(A) HEK293T cells were incubated with TAAR plasmids or mock (empty vector) along with a CRE-luciferase reporter vector for 48 hours and subsequently stimulated with 10 or 100 μM spermine for 4 hours. Luciferase activity was indicated by luminescence value and was normalized to the response from the control stimuli DMSO (mean ± SEM, n = 2). (B) As an inset to (A), (B) shows spermine-induced cAMP production in HEK293T cells expressing TAAR348. The cAMP levels were measured with a TR-FRET assay and normalized to the cAMP level in buffer-treated cells (mean ± SEM, n = 5). Fold increase over basal was calculated as [(measured cAMP − basal cAMP) ÷ basal cAMP] − 1. (C) TAAR348 targeted to HEK293T cell membrane as shown by the immunostained Rho-tag antibody (red) for total (bottom panel, with permeation using Triton X-100) or membrane-bound (top panel, without permeation) expression. Receptors located on the cell surface appeared as red rings around the nucleus (denoted with arrow). The nucleus was counterstained with DAPI (blue). EGFP was used as the negative control to evaluate transfection efficiency (green). Scale bar: 50 μm. (D) Representative olfactory receptor neurons expressing taar348 (purple), labeled with a DIG-labeled antisense RNA probe in a cross-sectional view of the main olfactory epithelium of an adult female sea lamprey. The section was counterstained with Nuclear Fast Red. Scale bar: 50 μm. Underlying data are available in S1 Data. cAMP, cyclic-adenosine monophosphate; CRE, cyclic-adenosine monophosphate response element; DIG, digoxigenin; EGFP, enhanced green fluorescent protein; HEK293T, human embryonic kidney 293T; LP, lamina propria; LU, lumen; OE, olfactory epithelium; TAAR, trace amine-associated receptor; TR-FRET, time-resolved fluorescence energy transfer.

Fig 4. Cyclen antagonized spermine-induced cAMP responses to receptor TAAR348.

(A) Cyclen treatment inhibited cAMP production induced by 0.1 mM spermine in TAAR348-expressing HEK293T cells as measured with a TR-FRET assay and normalized to the cAMP level in buffer-treated cells (mean ± SEM, n = 5). The filled triangle represents cAMP production induced by 0.1 mM spermine in the cells without exposure to cyclen. Fold increase over basal was calculated as [(measured cAMP − basal cAMP) ÷ basal cAMP] − 1. (B) Cyclen did not induce cAMP production in HEK293T with TAAR348 (mean ± SEM; reads from triplicate wells). (Inset) Structure of cyclen (1,4,7,10-tetraazacyclododecane). (C) Cadaverine induced cAMP production in a concentration-dependent manner in HEK293T cells expressing another sea lamprey TAAR-like gene—TAAR346a (mean ± SEM, n = 5). (D) Cyclen treatment did not inhibit cAMP production induced by 1 mM of cadaverine in HEK293T cells expressing TAAR346a. The filled triangle represents the cAMP production induced by 1 mM of cadaverine in the cells without exposure to cyclen (mean ± SEM, n = 5). Underlying data are available in S1 Data. cAMP, cyclic-adenosine monophosphate; HEK293T, human embryonic kidney 293T; TAAR, trace amine-associated receptor; TR-FRET, time-resolved fluorescence energy transfer.

Fig 5. Cyclen impeded spermine-induced olfactory and behavioral responses.

(A) Cyclen treatment (10⁻⁵ M) reduced the EOG response magnitude to spermine (10⁻⁵ M) but not to 3kPZS (10⁻⁷ M), spermidine (10⁻⁵ M), or _L-arginine (10⁻⁵ M). EOG responses were blank-subtracted and normalized to the amplitude of the responses to 10⁻⁵ M _L-arginine (mean ± SEM, n = 5). The difference in responses for each stimulus before (gray) and during (white) exposure of the naris to cyclen was evaluated with a paired t test. (B) Ovulatory female sea lamprey behavioral responses to spermine and 3kPZS (positive control) in the absence and presence of 10⁻¹² M cyclen (cyclen background) in a two-choice maze assay (S2 Fig). The effect of cyclen was reversible because the attraction to spermine was rescued when cyclen application ceased. For the definition of the index of preference, see Materials and methods. A positive index value indicates attraction, and a negative index value indicates repulsion. Significance was evaluated using a Wilcoxon signed-rank test. The number in the parentheses indicates the number of test subjects spending more time in the experimental channel out of the total sample size. *p < 0.05, **p < 0.01. Underlying data are available in S1 Data. EOG, electro-olfactogram; n.s., not significant; 3kPZS, 3-keto petromyzonol sulfate.

Fig 6. Nap-spermine, an agonist of spermine receptor TAAR348, induced responses virtually identical to those induced by spermine.

(A) Nap-spermine induced cAMP production in a concentration-dependent manner in TAAR348-expressing HEK293T cells (mean ± SEM, n = 5). The cAMP assay was done as in Fig 3B. (Inset) Structure of nap-spermine. (B) EOG responses to 10⁻⁷ M spermine and 10⁻⁷ M nap-spermine (mean ± SEM, n = 12 and n = 6, respectively). Normalized EOG responses were blank-subtracted and normalized to the amplitude of the responses to 10⁻⁵ M _L-arginine. (C) Cyclen inhibited cAMP production induced by 10⁻⁵ M nap-spermine in HEK293T cells with TAAR348. The filled triangle represents the cAMP production induced by TAAR348-expressing HEK293T cells exposed to only 10⁻⁵ M nap-spermine (mean ± SEM, n = 5). (D) Cyclen treatment (10⁻⁵ M) reduced the EOG response magnitude to nap-spermine (10⁻⁶ M). Cyclen exposure conditions are the same as in Fig 5A. EOG responses were blank-subtracted and normalized to the amplitude of the responses to 10⁻⁵ M _L-arginine (mean ± SEM, n = 5). (E) Adult sea lamprey behavioral responses to spermine and nap-spermine in a two-choice maze (S2 Fig). For definition of the index of preference, see Materials and methods. A positive index value indicates attraction, and a negative index value indicates repulsion. Significance was evaluated using a Wilcoxon signed-rank test. The number in the parentheses indicates the number of test subjects spending more time in the experimental channel out of the total sample size. *p < 0.05, **p < 0.01. Underlying data are available in S1 Data. cAMP, cyclic-adenosine monophosphate; EOG, electro-olfactogram; HEK293T, human embryonic kidney 293T; nap-spermine, 1-naphthylacetyl spermine; TAAR, trace amine-associated receptor.