Introduction

Reproductive behaviors are to a large degree hard-wired in the brain to guarantee reproductive success making the underlying neuronal circuits amenable to genetic analysis (Anderson, 2016; Dulac and Kimchi, 2007; Rings and Goodwin, 2019; Yamamoto and Koganezawa, 2013).

During development, sex-specific circuits are built into the brain under the control of the sex determination genes doublesex (dsx) and fruitless (fru) in Drosophila (Billeter et al., 2006; Schutt and Nothiger, 2000). They encode transcription factors that are alternatively spliced in a male or female specific mode (Schutt and Nothiger, 2000). By default, the dsx gene generates the male-specific isoform DsxM, while a female-specific isoform DsxF is generated by alternative splicing and expressed in about ∼700 distinct neurons in the brain important for female reproductive behaviors directing readiness to mate and egg laying (Rezaval et al., 2012; Rideout et al., 2010). FruM is expressed in about ∼1000 neurons in males and implements development of neuronal circuitry key to display male courtship behavior, but is switched off in females through alternative splicing by incorporation of a premature stop codon (Demir and Dickson, 2005; Manoli et al., 2005; Stockinger et al., 2005).

The circuitry of female specific behaviors including receptivity to courting males for mating and egg laying have been mapped using intersectional gene expression via the split-GAL4 system to restrict expression of activators or inhibitors of neuronal activity to very few neurons (Aranha and Vasconcelos, 2018; Cury and Axel, 2023; Wang et al., 2020a; Wang et al., 2020b; Wang et al., 2021). Through this approach, sensory neurons in the genital tract have been identified as key signal transducers for the readiness to mate and the inhibition of egg laying connecting to central parts of the brain via projection to abdominal ganglion neurons (Feng et al., 2014; Hasemeyer et al., 2009; Rezaval et al., 2012; Yang et al., 2009). This circuit then projects onto centrally localized pattern generators in the brain to direct a behavioral response via efferent neurons (Wang et al., 2020a; Wang et al., 2020b; Wang et al., 2021).

Once females have mated, they will reject courting males and lay eggs (Manning, 1967). These post-mating responses (PMRs) are induced by male derived sex-peptide (SP) transferred during mating (Avila et al., 2011; Chen et al., 1988; Hopkins and Perry, 2022). In addition, SP will induce a number of other behavioral and physiological changes including increased egg production, feeding, a change in food choice, sleep, memory, constipation, midgut morphology, stimulation of the immune system, and sperm storage and release (Carvalho et al., 2006; Domanitskaya et al., 2007; Kim et al., 2010; Peng et al., 2005; Ribeiro and Dickson, 2010; Scheunemann et al., 2019; Soller et al., 1999) (Cognigni et al., 2011) (Avila et al., 2010; Isaac et al., 2010; Wainwright et al., 2021; White et al., 2021). SP binds to broadly expressed Sex Peptide Receptor (SPR), an ancestral receptor for myoinhibitory peptides (MIPs) (Jang et al., 2017; Kim et al., 2010; Yapici et al., 2008). Although MIPs seem not to induce PMRs, excitatory activity of MIP expressing neurons underlies re-mating (Jang et al., 2017; Kim et al., 2010; Yapici et al., 2008). Expression of membrane-tethered SP (mSP) induces PMRs in an autocrine fashion when expressed in neurons, but not glia (Haussmann et al., 2013; Nakayama et al., 1997).

First attempts to identify SP target neurons by enhancer GAL4 induced expression of UASmSP only identified lines with broad expression in the nervous system (Nakayama et al., 1997). Later, drivers with more restricted expression including dsx, fru and pickpocket (ppk) genes were identified, but they are expressed in all parts of the nervous system throughout the body eluding to reveal the location of SP target sites unambiguously (Hasemeyer et al., 2009; Haussmann et al., 2013; Rezaval et al., 2012; Yang et al., 2009; Yapici et al., 2008). To delineate where in the Drosophila SP target neurons are located which induce the main PMRs, refusal to mate and egg laying, we expressed mSP specifically only in the head or trunk. These experiments separate reduction of receptivity induced in the head from trunk induction of egg laying. To further restrict our search for SP target neurons, we focused on three genes, SPR, dsx and fru, because SPR is broadly expressed but anticipated to induce PMRs only from few neurons, and because GAL4 inserted in the endogenous dsx and fru loci induces PMRs from mSP expression. Using GAL4 tiling lines with fragments encompassing the regulatory regions of complex SPR, fru and dsx genes (Jenett et al., 2012; Kvon et al., 2014; Pfeiffer et al., 2008), we identified one regulatory region in each gene reducing receptivity and inducing egg laying upon mSP expression, and one additional region in SPR only inducing egg laying. To further refine this analysis, we used intersectional gene expression using split-GAL4 and flipase (flp) mediated excision of stop cassettes in UAS reporters (Luan et al., 2006; Struhl and Basler, 1993). Consistent with previous results that the SP response can be induced via multiple pathways (Haussmann et al., 2013), we found distinct sets of neurons in the central brain and the abdominal ganglion that can induce PMRs via expression of mSP. Mapping the pre- and post-synaptic connections of the distinct SP target neurons by retro- and trans-Tango (Sorkaç et al., 2023; Talay et al., 2017) revealed that SP target neurons direct higher order sensory processing in the central brain. These neurons feed into two common post-synaptic neuronal subtypes indicating that SP interferes with the integration of diverse sensory inputs to build a stereotyped output either reducing receptivity and/or increasing egg laying.

Results

Reduction of receptivity and induction of egg laying are separable by head and trunk expression of membrane-tethered SP

Due to the complex behavioral and physiological changes induced by SP, neurons in the central nervous system have been suspected as main targets for SP (Kubli, 1992). To express mSP only in the head we used an elav FRTstopFRT GAL4 in combination with otdflp, that expresses in the head to drive recombination and head-specific expression of mSP from UAS (Figure 1A) (Asahina et al., 2014; Haussmann et al., 2008; Nallasivan et al., 2021; Zaharieva et al., 2015). To express mSP only in the trunk we used tshGAL4 (Figure 1B) (Soller et al., 2006).

The main PMRs in females can be separated.

A, B) Schematic depiction of head and trunk expression in Drosophila elav FRTstopFRT GAL4; otdflp (A) and in tshGAL4 (B) visualized by UAS GFP (green).

C, D) Receptivity (C) and oviposition (D) of wild type control virgin (red) and mated (orange) females, and virgin females expressing UAS mSP (green) pan-neuronally with nsybGAL4 or in head and trunk patterns shown as means with standard error from three repeats for receptivity (21 females per repeat) by counting the number of females mating within a 1 h period or for oviposition by counting the eggs laid within 18 hours from 30 females. Statistically significant differences from ANOVA post-hoc comparison are indicated by different letters (p<0.0001).

E-H) Representative adult female genital tract showing tshGAL4 UAS H2BYFP (green) and elavLexA LexAop NLStomato (red) nuclear expression. The magnification (F-H) shows sensory genital tract neurons. Scale bar shown in E and H are 100 μm and 20 μm, respectively.

When we expressed mSP in the head, females reduced receptivity indistinguishable from mated females, but did not lay eggs, thereby again demonstrating that the two main PMRs can be separated (Figures 1C and 1D) (Haussmann et al., 2013). In contrast, when we expressed mSP in the trunk, females remained receptive, but laid eggs in numbers indistinguishable from mated females (Figures 1C and 1D).

Moreover, tshGAL4 is expressed in fru, dsx, ppk genital tract sensory neurons (Figures 1E-1H). Since mSP expression with tshGAL4 does not affect receptivity, these genital tract neurons unlikely are direct targets for SP (Haussmann et al., 2013). Taken together, these results indicate presence of SP target neurons in the brain and ventral nerve cord (VNC) for the reduction of receptivity and induction of egg laying, respectively.

Few restricted regulatory regions in large SPR, fru and dsx genes can induce the SP response

Expression of mSP from UAS via GAL4 inserts in fru and dsx genes induces a robust reduction in receptivity and increase in egg laying (Haussmann et al., 2013; Rezaval et al., 2012). To identify SP target neurons, we thought to dissect the broad expression pattern of complex SPR, fru and dsx genes spanning 50-80 kb by identifying regulatory DNA fragments in the enhancer regions that drive UAS mSP in a subset of neurons. For these experiments, we analysed 22, 27 and 25 GAL4 lines from the VDRC and Janelia tiling GAL4 projects (Jenett et al., 2012; Kvon et al., 2014; Pfeiffer et al., 2008) (Figures 2A-2C).

Distinct regulatory regions in SPR, fru and dsx genes induce PMRs from mSP expression.

A-C) Schematic representation of SPR, fru, and dsx chromosomal regions depicting coding and non-coding exons as black or white boxes, respectively, and splicing patterns in solid lines. Vertical lines below the gene model depict enhancer GAL4 lines with names and those in red showed PMRs by expression of mSP.

D, E) Receptivity (D) and oviposition (E) of wild type control virgin (red) and mated (orange) females, and virgin females expressing UAS mSP (green) under the control of GAL4 pan-neuronally in nsyb or in SPR8, SPR12, fru11, fru12, and dsx24 patterns shown as means with standard error from three repeats for receptivity (21 females per repeat) by counting the number of females mating within a 1 h period or for oviposition by counting the eggs laid within 18 hours from 30 females. Statistically significant differences from ANOVA post-hoc comparison are indicated by different letters (p≤0.0001).

F-O) Representative adult female brains (F-J) and ventral nerve cords (VNC, K-O) expressing UAS CD8GFP under the control of SPR8, SPR12, fru11, fru12 and dsx24 GAL4. Scale bars shown in J and O are 50 µm and 100 µm, respectively.

Strikingly, in SPR, fru and dsx genes we identified only one regulatory region in each gene (SPR8, fru11/12 and dsx24) that reduced receptivity and induced egg laying through GAL4 UAS expression of mSP (Figures 2D and 2E). In addition, we identified one line (SPR12) in the SPR gene, that induced egg laying, but did not reduce receptivity consistent with previous results that SP regulation of receptivity and egg laying can be split (Haussmann et al., 2013). All of these lines expressed in subsets of neurons in the central brain and the ventral nerve cord in distinct, but reduced patterns compared to the expression of the SPR, fru and dsx genes (Rideout et al., 2010; Yapici et al., 2008; Zhou et al.) (Figures 2F-2O). Moreover, these lines showed prominent labelling of abdominal ganglion neurons in the VNC (Figures 2K-2O). In addition, all of these lines except SPR12 are also expressed in genital tract sensory neurons (Figure S1).

From all the 74 lines that we have analyzed for PMRs from SPR, fru and dsx genes, we also analysed expression in genital tract sensory neurons as they had been postulated to be the primary targets of SP (Hasemeyer et al., 2009; Rezaval et al., 2012; Yang et al., 2009; Yapici et al., 2008). Apart from PMR inducing lines SPR8, fru11, fru12 and dsx24, that showed expression in genital tract sensory neurons, we identified three lines (SPR3, SPR 21 and fru9), which also robustly expressed in genital tract sensory neurons but did not induce PMRs from expression of mSP (Figure S2).

Secondary ascending abdominal ganglion neurons can induce the PMRs from mSP expression

A set of neurons has been identified in the abdominal ganglion important for egg laying by silencing neuronal activity from a screen expressing the rectifying potassium channel Kir2.1 that identified six driver lines (FD1-6) (Feng et al., 2014). These neurons have been termed SAG (secondary ascending abdominal ganglion neurons) neurons, that are also interconnected with myoinhibitory peptide sensing neurons (Jang et al., 2017). Since enhancer lines identified in SPR, fru and dsx genes are prominently expressed in the abdominal ganglion, we tested whether mSP expression from these SAG neuron expressing lines induced PMRs.

From these six lines, one robustly suppressed receptivity and induced egg laying (FD6/VT3280), while two lines only induced egg laying (FD3/VT4515 and FD4/V000454) similar to controls from mSP expression (Figure 3). Again, all three lines also expressed in subsets of neurons in the central brain and VNC, particularly in the abdominal ganglion (Zhou et al.). In addition, FD3 and FD4 did not express in genital tract sensory neurons, in contrast to FD6 (Feng et al., 2014).

Expression of mSP in secondary ascending abdominal ganglion neurons induces PMRs.

A, B) Receptivity (A) and oviposition (B) of wild type control virgin (red) and mated (orange) females, and virgin females expressing UAS mSP (green) under the control of GAL4 pan-neuronally in nsyb or in FD1, FD2, FD3, FD4, FD5, and FD6 patterns shown as means with standard error from three repeats for receptivity (21 females per repeat) by counting the number of females mating within a 1 h period or for oviposition by counting the eggs laid within 18 hours from 30 females. Statistically significant differences from ANOVA post-hoc comparison are indicated by different letters (p<0.0001).

Intersectional expression reveals distinct mSP responsive neurons in the central brain and abdominal ganglion

To further restrict the expression to fewer neurons, we intersected the expression patterns of those lines that induced robust reduction of receptivity and increase of egg laying using split-GAL4 (SPR8, fru11/12, dsx and FD6, for further experiments we used dsxGAL4-DBD, because dsx24 is less robust and fru11 and fru12 were made into one fragment) that activates the UAS reporter when GAL4 is reconstituted via dimerization of activation (AD-GAL4) and DNA binding (GAL4-DBD) domains (Luan et al., 2006) (Figure 4A).

Distinct circuits from intersection of SPR, fru, dsx and FD6 patterns in the brain and VNC induce PMRs from mSP expression.

A) Schematic showing the intersectional gene expression approach: GAL4 activation (AD, orange) and DNA binding domains (DBD, blue) are expressed in different, but overlapping patterns. Leucine zipper dimerization reconstitutes a functional split-GAL4 in the intersection (pink) to express UAS reporters.

B, C) Receptivity (B) and oviposition (C) of wild type control virgin (red) and mated (orange) females, and virgin females expressing UAS mSP (green) under the control of split-GAL4 intersecting SPR8 ∩ fru11/12, SPR8 ∩ dsx, SPR8 ∩ FD6, fru11/12 ∩ dsx and fru11/12 ∩ FD6 patterns shown as means with standard error from three repeats for receptivity (21 females per repeat) by counting the number of females mating within a 1 h period or for oviposition by counting the eggs laid within 18 hours from 30 females. Statistically significant differences from ANOVA post-hoc comparison are indicated by different letters (p<0.0001).

D-M) Representative adult female brains and ventral nerve cords (VNC) expressing UAS CD8GFP under the control of SPR8 ∩ fru11/12, SPR8 ∩ dsx, SPR8 ∩ FD6, fru11/12 ∩ dsx and fru11/12 ∩ FD6. Scale bars shown in H and M are 50 µm and 100 µm, respectively.

Again, intersection of SPR8 with fru11/12, dsx or FD6, and fru11/12 with dsx or FD6 expression robustly reduced receptivity and increased egg laying upon expression of mSP (Figures 4B and 4C).

When we further analyzed the expression of these split-GAL4 intersections in the brain, we found that each combination first showed very restricted expression, but second, that none of these combinations labeled the same neurons (Figures 4D-4H). For dsx neurons, split-GAL4 intersections correspond to a subset of dPC2l (SPR8 ∩ dsx) and dPCd-2 (fru11/12 ∩ dsx) neurons (Deutsch et al., 2020; Nojima et al., 2021; Schretter et al., 2020). These results suggest the SP targets interneurons in the brain that feed into higher processing centers from different entry points likely representing different sensory input.

In the ventral nerve cord, we found expression in the abdominal ganglion with all split-GAL4 combinations (Figures 4I-4M). In particular, intersection of dsx with SPR8 or fru11/12 showed exclusive expression in the abdominal ganglion, while the other combinations also expressed in other cells of the VNC. All together, these data suggest that the abdominal ganglion harbors several distinct type of neurons involved in directing PMRs (Oliveira-Ferreira et al., 2023).

In the female genital tract, these split-Gal4 combinations show expression in genital tract neurons with innervations running along oviduct and uterine walls (Figures S3A-S3E). In addition, SPR8 ∩ fru11/12 and SPR8 ∩ dsx were also expressed in the spermathecae (Figures S3A-S3B).

Both inhibitory and activating neurons have been attributed to impact on PMRs (Kvitsiani and Dickson, 2006; Rezaval et al., 2012; Yapici et al., 2008). These neurons seem to be part of intersecting circuitry as general inhibition of ppk neurons by tetanus toxin (TNT) only partially blocks the SP response in contrast to inhibition of ppk neurons in the brain alone (Nallasivan et al., 2021).

When we inhibited neuronal activity by expression of TNT (Sweeney et al., 1995), we observed a significant reduction of receptivity for all split-Gal4 combinations, though only partially for inhibition in fru11/12 ∩ FD6 neurons. Likewise, all split-Gal4 combinations induced a significant increase in egg laying (Figures S4A and S4B). Ablation of these neurons by expression of apoptosis inducing reaper and hid genes essentially replicated the results from neuronal inhibition indicating that SPR target neurons are modulatory and are not part of motor circuits because females laid eggs and performed normally in receptivity assays (Figures S4C and S4D).

To evaluate the composition of the intersected expression patters into inhibitory and activating neurons we also expressed the Bacillus halodurans sodium channel (NaChBac) (Feng et al., 2014) to activate all of the intersected neurons. Here, we found a significant reduction of receptivity for four of the five split-GAL4 combinations, though only partially for activation of SPR8 ∩ dsx neurons (Figure S4E). Activating fru11/12 ∩ FD6 neurons did not reduce receptivity (Figure S4E). Likewise, we found the same pattern for the induction of egg laying (Figure S4F). Four of the five split-GAL4 combinations induced a significant increase which was only partial in SPR8 ∩ dsx neurons and no egg laying was induced by activating fru11/12 ∩ FD6 neurons.

Essentially, these results are consistent with previous findings that inhibitory neurons prevail (Nallasivan et al., 2021), possibly as input from trunk neurons as found for ppk expressing neurons (see below).

ppk neurons do not intersect with SPR, fru, dsx and FD6 neurons in inducing PMRs by mSP

Expression of mSP in ppk neurons can also induce PMRs (Figures S5A and S5B) (Hasemeyer et al., 2009; Yang et al., 2009). The complement of ppk neurons labeled with ppkGAL4 consists of at least two populations including sensory neurons and eight interneurons in the central brain (Nallasivan et al., 2021). These brain neurons show severe developmental defects in SP-insensitive Nup54 mutant alleles, but they receive inhibitory input from sensory neurons (Nallasivan et al., 2021).

To evaluate whether ppk neurons are part of the previously identified expression patterns, we intersected them by crossing GAL4-AD lines SPR8, SPR12 and fru11/12 with a ppk GAL4-DBD line containing the previously used 3 kb promoter fragment (Grueber et al., 2003; Seidner et al., 2015). Surprisingly, none of these split-GAL4 combinations reduced female receptivity or increased egg laying (Figures S5A and S5B). Likewise, no GFP expressing neurons were detected in the brain, abdominal ganglion or the genital tract (Figures S5C-S5E, S5F-S5H and S5I-S5K). Also, inhibiting or activating neurons with these split-Gal4 combinations did not reduce receptivity or induce egg laying (Figures S5L and S5O). How exactly ppk neurons impact on PMRs, however, needs to be further evaluated in follow-up studies.

mSP responsive neurons rely on SPR and are required for PMRs induced by SP delivered through mating

Next, we tested whether PMRs induced by mSP expression in the SPR8 ∩ dsx, fru11/12 ∩ dsx or SPR8 ∩ fru11/12 rely on SPR. Expression of mSP in dsx ∩ SPR8 and dsx ∩ fru11/12 neurons in SPR mutant females did not reduce receptivity or induce egg laying (Figures 5A and B, see also Figures 4A and B), while a partial response was observed for SPR8 ∩ fru 11/12 induced mSP expression in SPR mutant females, which is consistent with presence of additional receptors for SP (Haussmann et al., 2013).

Distinct neuronal circuitries from intersection of SPR, fru and dsx sense SP after mating to induce PMRs.

A, B) Receptivity (A) and oviposition (B) of wild type control virgin (red) and mated (orange) females, and virgin females expressing UAS mSP (green) under the control of split-Gal4 intersecting SPR8 ∩ dsx, fru11/12 ∩ dsx, and SPR8 ∩ fru11/12 patterns in SPR/Df mutant females or SPR RNAi knock-down shown as means with standard error from three repeats for receptivity (21 females per repeat) by counting the number of females mating within a 1 h period or for oviposition by counting the eggs laid within 18 hours from 30 females. Statistically significant differences from ANOVA post-hoc comparison are indicated by different letters (p<0.0001 except p=0.002 and p=0.006 for c and d in A, and p=0.004 for c in B).

Since SP is transferred during mating to females and enters the hemolymph (Haussmann et al., 2013), we wanted to test whether SPR is required in these neurons for inducing PMRs after mating. For SPR RNAi in dsx ∩ fru11/12 and SPR8 ∩ fru 11/12 neurons, no reduction, or a partial reduction, of receptivity was observed, respectively, while SPR RNAi in dsx ∩ SPR8 neurons turned virgin females unreceptive (Figure 5A). Expression of mSP in dsx ∩ fru11/12 neurons in the context of SPR RNAi partially reduced receptivity again suggesting additional receptors for SP (Haussmann et al., 2013).

Strikingly, however, SPR RNAi in these neurons prevented egg laying independent of whether SP was delivered by mating or when tethered to the membrane of these neurons (Figure 5B).

These results demonstrate that neurons identified by split-GAL4 intersected expression of SPR8 with dsx or fru11/12, or fru11/12 with dsx are genuine SP targets as they rely on SPR and PMRs are induced by SP delivered through mating.

Expression of mSP in distinct neurons in the brain induces PMRs

The analysis of ppk neurons in SP-insensitive Nup54 alleles revealed a hierarchy of trunk neurons that dominate over central brain neurons (Nallasivan et al., 2021). To focus on the role of central brain neurons, we generated a UAS mSP line with a stop cassette (UAS FRTstopFRT mSP) that allows to restrict expression of mSP to the head in the presence of otdflp, which only expresses in the head (Figure 6A), but not in the trunk (Asahina et al., 2014; Nallasivan et al., 2021).

Distinct neuronal circuitries in the brain senses SP to induce PMRs.

A, B) Schematic depiction of UAS GFP (green) expression in the head of Drosophila (A) combining split-GAL4 intersectional expression (AD-GAL4 and GAL4-DBD) with brain-expressed otdflp mediated recombination of UAS FRTGFPstopFRTmSP (B).

C, D) Receptivity (C) and oviposition (D) of wild type control virgin (red) and mated (orange) females, and virgin females expressing UAS FRTGFPstopFRTmSP (grey), UAS FRTGFPstopFRTTrpA1 (purple) and UAS FRTGFPstopFRTTNT (pink) under the control of split-GAL4 intersecting SPR8 ∩ dsx, fru11/12 ∩ dsx and SPR8 ∩ fru11/12 patterns with brain-specific FRT-mediated recombination by otdflp shown as means with standard error from three repeats for receptivity (21 females per repeat) by counting the number of females mating within a 1 h period or for oviposition by counting the eggs laid within 18 hours from 30 females. Statistically significant differences from ANOVA post-hoc comparison are indicated by different letters (p<0.0001 except p<0.0004 for c in C, p<0.007 for c in D).

In combination with the intersectional approach, we now can restrict mSP expression to few central brain neurons, or alternatively activate or silence these neurons (Figure 6B). Expression of mSP in SPR8 ∩ dsx, fru11/12 ∩ dsx or SPR8 ∩ fru11/12 neurons mSP in the central brain significantly reduced receptivity, but oviposition was only substantially induced in SPR8 ∩ dsx brain neurons (Figures 6C and 6D). These results clearly demonstrate a role for brain neurons in the SP response. However, we noticed that the flipase approach can result in false negatives as fruflp inserted in the same position in the endogenous locus as fruGAL4 does not induce a response with UAS FRTstopFRT mSP in contrast to fruGAL4 induced expression of mSP. In contrast, the same experiment with dsxGAL4 and dsxflp results in a positive SP response indistinguishable from mated females (Haussmann et al., 2013).

Next, we tested whether neuronal activation or inhibition would induce a post-mating response. Strikingly, conditional activation of SPR8 ∩ dsx, fru11/12 ∩ dsx or SPR8 ∩ fru11/12 brain neurons with TrpA1in adult females completely inhibited receptivity and induced egg laying comparable to mated females (Figures 6C and 6D). In contrast, inhibition of these neurons with tetanus toxin (TNT) did not alter the virgin state, e.g. receptivity was not reduced and egg laying was not induced (Figures 6C and 6D).

mSP responsive neurons operate in higher order sensory processing in the brain

With the split-GAL4 approach we identified five distinct neuronal sub-types that can induce PMRs. To find out whether these neurons receive input from distinct entry points in the brain and to identify the target neurons of these mSP responsive neurons, we used the retro- and trans-Tango technique to specifically activate reporter gene expression in up- and down-stream neurons (Sorkaç et al., 2023; Talay et al., 2017)(Figures 7A-O).

retro- and trans-Tango identification of pre- and post-synaptic neurons of SP target neurons reveals higher order neuronal input canalized into shared output circuitries.

A-O) Representative adult female brains expressing QUAST tomato3xHA retro-Tango (left, A-E), UAS myrGFP (middle, F-J) and QUAST tomato3xHA trans-Tango (right, K-O) in SPR8 ∩ dsx, fru11/12 ∩ dsx, SPR8 ∩ fru11/12, SPR8 ∩ FD6 and fru11/12 ∩ FD6 split-GAL4s. The presynaptic (A-E, left), split-GAL4 (F-J, middle) and postsynaptic (K-O, right) neuronal circuitries are shown in an inverted grey background. Arrows (magenta) indicate neurons and their corresponding projections in different regions in female brain. The scale bar shown in O is 50 μm.

P) Model for the SP induced post-mating response. SP interferes with interpretation of sensory cues, e.g. vison, hearing, smell, taste, and touch at distinct sites in the brain indicated by higher order projections revealed by intersectional expression in the following patterns: SPR8 ∩ dsx (blue), fru11/12 ∩ dsx (black), SPR8 ∩ fru11/12 (yellow), SPR8 ∩ FD6 (pink) and fru11/12 ∩ FD6 (olive). and VNC (fru11/12 ∩ dsx) during higher order neuronal processing.

In the brain, the retro-Tango analysis did not identify primary sensory neurons, but higher order neurons in the central brain in all five split-GAL4 combinations (Fig 7A-E). In addition, neurons in the suboesophagal ganglion were marked from SPR8 intersections with dsx and FD6, and in dsx ∩ fru11/12. In dsx ∩ fru11/12, neurons in the optic lobe (medulla) were marked. In addition, a strong signal was observed in all five split-GAL4 combinations in the mushroom bodies (Figs 7A-E). Although mushroom bodies are dispensable for PMRs (Fleischmann et al., 2001) their connection to SP target neurons indicates an experience dependent component of PMRs.

The trans-Tango analysis identified a subset of neurons with cell bodies in the suboesophageal ganglion with projections to the pars intercerebralis for SPR8 ∩ dsx and fru11/12 ∩ dsx neurons (Figures 7K and 7L). For SPR8 ∩ fru11/12 and SPR8 ∩ FD6 neurons common target neurons were found in the antennal mechanosensory and motor centre (AMMC) region with a single neuron identified near the mushroom body region (Figures 7M and 7N) (Ishimoto and Kamikouchi, 2021). For fru11/12 ∩ FD6 no obvious targets were identified in the central brain (Figure 7O).

In the VNC, the trans-Tango analysis showed post-synaptic targets within the abdominal ganglion with all five split-GAL4 combinations indicating an interconnected neuronal network (Figure S6A-S6O), which needs to be elaborated in detail. In the genital tract, no post-synaptic targets were detected indicating that these are afferent neurons integrating sensory input (Figure S6P-S6AD).

Taken together, circuitries identified via retro- and trans-Tango place SP target neurons at the interface of sensory processing interneurons connecting to two commonly shared post-synaptic processing neuronal populations in the brain. Hence, our data indicate that SP interferes with sensory input processing from multiple modalities that are canalized to higher order processing centres to generate a behavioural output.

Discussion

Much has been learned about the neuronal circuitry governing reproductive behaviors in Drosophila from interfering with neuronal activity in few neurons selected by intersectional expression using split-GAL4 (Wang et al., 2020a; Wang et al., 2020b; Wang et al., 2021). However, how sex-peptide signaling as main inducer of the post-mating response, prominently consisting of refractoriness to re-mate and induction of egg laying, is integrated in this circuitry is not completely understood (Haussmann et al., 2013). Here, we addressed this gap by identifying regulatory regions in SPR, fru and dsx genes driving membrane-tethered expression of SP in subsets of neurons to delineate SP targets to very few neurons in the central brain and the ventral nerve cord by intersectional expression. Consistent with previous analysis describing multiple pathways for the SP response (Haussmann et al., 2013), we find five distinct populations of interneurons in the central brain directing PMRs. In SP target neurons in the central brain, SPR is essential to induce PMRs when receiving SP from males through mating. From mapping post-synaptic targets by trans-Tango, we identified two populations of interneurons. The architecture of this circuitry is reminiscent for processing of sensory input transmitted to central brain pattern generators for behavioral output. Hence, SP interferes at several levels for coordinating PMRs, but also leaves the female the opportunity to interfere under unfavorable conditions with specific elements of PMRs, e.g., if there is no egg laying substrate, females will still not remate (Haussmann et al., 2013). Likewise, mated females will not lay eggs despite suitable egg laying substrates if parasitoid wasp are present (Kacsoh et al., 2015). Thus, the architecture of female PMRs contrasts with male-courtship behavior consisting of a sequel of behavioral elements that once initiated will always follow stereotypically to the end culminating in mating, or start from the beginning when interrupted (Greenspan and Ferveur, 2000; Hall, 1994).

SP induces PMRs via entering the hemolymph to target neurons in the central brain and ventral nerve cord

Early characterization of the SP signaling cascade demonstrated induction of PMRs from various other sources than mating including transgenic secretion from the fat body, expression as membrane-tethered form on neurons or injection of synthetic peptide into the hemolymph (Aigaki et al., 1991; Chen et al., 1988; Nakayama et al., 1997; Schmidt et al., 1993). Likewise, SP is detected in the hemolymph after mating at a PMR inducing concentration (Haussmann et al., 2013). Moreover, PMRs are induced faster, when SP is injected compared to induction by mating (Haussmann et al., 2013). This delay, however, is not attributed to sperm binding of SP as it is unchanged after mating with spermless males. These results suggest that SP reaches its targets through entering the circulatory system to target neurons and contrasts a previously proposed model favoring genital tract neurons as SP sensors from the lumen of the genital tract (Hasemeyer et al., 2009; Rezaval et al., 2012; Yang et al., 2009).

In further support of the internalization model, we identified GAL4 drivers that express mSP in genital tract neurons, but do not induce PMRs. Also, SPR12 does not express in genital tract neurons, but induces egg laying by expression of mSP. Moreover, expression of mSP in the trunk (including all genital tract sensory neurons), only induces egg laying, but does not change receptivity. Likewise, expression of mSP specifically in the brain (SPR8 ∩ dsx) can reduce receptivity and induce egg laying indistinguishable from mated females.

These results are in strong favor for SP entering the hemolymph to target neurons in the ventral nerve cord for inducing egg laying, and in the central brain for reducing receptivity and inducing egg laying.

Integration of SP signaling into the circuitry directing reproductive behaviours

Reduction of receptivity and induction of egg laying are both induced by the same critical concentration of injected SP (Haussmann et al., 2013; Schmidt et al., 1993) initially suggesting a simple on/off system for PMRs likely initiated from a small population of neurons. However, such model would not allow to split the SP response into individual PMR components by expression of mSP.

Here, we identified several GAL4 drivers, that can induce only egg laying (SPR12, FD3, FD4 and tsh GAL4), but do not reduce receptivity. Most striking is tshGAL4, that expresses only in the trunk. All of these lines express in the abdominal ganglion and dsx neurons in the abdominal ganglion have been identified to induce egg laying (Rezaval et al., 2012; Zhou et al.). Hence, this neuronal structure has a key role in regulating egg laying. Since more than a single neuronal population seems to direct egg laying, further high-resolution mapping is required to identify individual neuronal population within the abdominal ganglion (Jang et al., 2017; Oliveira-Ferreira et al., 2023).

Since tshGAL4 only induces egg laying, neurons in the brain must direct reduction of receptivity. Through intersectional expression in combination with head-specific expression of otdflp, we could express mSP only in the brain by FLP mediated brain-specific excision of a stop cassette. We observed a significant reduction in receptivity for all five intersections tested, but for four the response is only partial likely due to the inefficiency of FLP mediated recombination.

Moreover, brain neurons can also induce egg laying when SPR8 is intersected with dsx, and to some extent also from SPR8 intersection with fru11/12. Due to the inefficiency of FLP mediated recombination, however, this is likely an underestimate and solving this issue requires development of more robust tools.

In any case, however, our results show that PMRs can be induced from mSP expression from several sites suggesting interference with processing of sensory information at the level of interneurons. In particular, SPR8 ∩ fru11/12 neurons resemble auditory AMMC-B2 neurons involved in processing of information of the male love song (Yamada et al., 2018). Likewise, SPR8 ∩ dsx neurons seem to overlap with dimorphic dsx pCL2 interneurons that are part of the 26 neurons constituting the pC2 neuronal population involved in courtship song sensing, mating acceptance and ovipositor extrusion for rejection of courting males (Deutsch et al., 2019; Kimura et al., 2015; Wang et al., 2020a). The SPR8 ∩ FD6 neurons resemble dopaminergic fru P1 neurons involved in courtship and the fru11/12 ∩ dsx neurons seem to overlap with dsx pCd and neuropeptide F neurons involved in courtship (Zhang et al., 2021). In females, pC1d neurons have been linked to aggression (Deutsch et al., 2020; Schretter et al., 2020). The fru11/12 ∩ FD6 neurons resemble a class of gustatory pheromone sensing neurons (Sakurai et al., 2013). Although we likely have not identified all SP sensing neurons, our resources will provide a handle to future exploration of the details of this neuronal circuitry incorporating SP signaling for inducing PMRs.

Conclusions

We have identified distinct SP sensing neurons in the central brain and the ventral nerve cord. Since these five different SP sensing neuronal populations in the central brain converge into two target sites, our data suggest a model (Figure 7P), whereby SP signaling interferes with integration of sensory input. Independent interference with different sensory modalities opts for the female to counteract male manipulation at the level of perception of individual sensory cues to adapt to varying physiological and environmental conditions to maximize reproductive success.

Acknowledgements

We thank T. Aigaki, G. Barnea, P. Soba, W.J. Joiner, B. Dickson, S. Goodwin, C. Rezaval, D. Anderson, J.J. Hodge, A. Hidalgo, S. Collier, the Bloomington stock center, the Vienna Drosophila RNAi Center for flies, T. Aigaki and W.J. Joiner for plasmids, the University of Cambridge Department of Genetics Fly Facility and FlyORF for injections, D. Scocchia for help with PCR and, I.U. Haussmann, Y.J. Kim, J.C. Billeter and J-R Martin for comments on the manuscript. We acknowledge funding by the Biotechnology and Biological Science Research Council to MS.

Author contributions

MS conceived and directed the project. MPN performed genetic experiments and imaging. MS and SS performed genetic and DS imaging experiments. MPN and MS analyzed data. MS wrote the manuscript with support from MPN. All authors read and approved the final manuscript.

Data availability

Data is included in the paper and supplementary files. Source data is provided.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Materials and methods

Key resources table

Fly strains and husbandry

Flies were kept on standard cornmeal-agar food (1%industrial-grade agar, 2.1% dried yeast, 8.6% dextrose, 9.7% cornmeal and 0.25% Nipagin, all in (w/v)) in a 12 h light: 12 h dark cycle. Propionic acid was omitted from fly food as acidity affects egg laying (Gou et al., 2014). Genetic crosses were done in vials and kept at low density to ensure larvae were not competing for food and if necessary, additional live yeast was added. For all behavioral assays, virgin and mated Canton-S were used as controls. Virgin females, e.g. from crosses of GAL4 with UASmSP, were collected after emergence within a 5 h window and well-fed with live yeast sprinkled on food for maximum egg production and allowed to sexually mature (3– 5 days). To recombine 2nd chromosome inserts for splitGAL4AD (attP40) and 3rd chromosome splitGAL4DBD (attP2), standard genetic crossing schemes were used and final stocks were balanced with CyO and TM3 Sb (combined from from ST and CT stock, see key resource list). SplitGal4AD and DBD combination lines were then crossed to UASmSP. For meiotic recombination, final stocks were validated by behavioral analysis for UAS mSP, for flp with eFeG UASCD8GFP to monitor GFP expression and for otdflp UASstopTrpA and otdflp UASstopTNT by crossing to elavGAL4 and monitored by lethality.

For enhanced recombination with flp, virgin females were transferred to 30° C after eclosion and kept for 5 d at this temperature before performing the behavioral assays. For induction of neuronal activity by temperature sensitive TrpA1, females were kept at 30° C.

To make UAS FRTstopFRT mSP, a gBlock (IDT) stop cassette with the FRT sequences used in the eFeG plasmid (Haussmann et al., 2008) was inserted into NotI cut pUAST-GGTmSP (gift from T. Aigaki) by Gibson assembly. In the stop-cassette, the FRT sequence is followed by a GFP with a 3’UTR from ewg containing polyA site 1 from intron 6 (Haussmann et al., 2011). Flies were transformed by P-element mediated transgenesis and a inserts on each chromosome were established that show a robust post-mating response with dsxflp indistinguishable from mated females.

Behavioral analysis

Females were examined for the main post-mating behaviors receptivity and oviposition as described previously and as follows (Soller et al., 1999; Soller et al., 2006). To generate mated females, one female and three males were added to fly vials and observed until mating and males were removed after mating. For receptivity tests, mature virgin or mated females were added to fly vials (95 mm length and 24 mm diameter) containing Canton S males with an aspirator and observed for 1 h, generally 3 females and 7 males. For these experiments, males were separated from females at least one day before the experiment. Receptivity tests were done in the afternoon with virgins, or 5-24 h after mating for controls. For oviposition, females were placed individually in fly vials in the afternoon and the number of eggs laid was counted the next day.

Statistical analysis

Sample size was based on previous studies, non-blinded and not pre-determined by statistical methods (Haussmann et al., 2013; Nallasivan et al., 2021; Soller et al., 1997). Behavioral data are representatives of at least three replicates that were performed on three different days. Statistical analysis of behavioral experiments were performed using GraphPadPrism 9 (GraphPad by Dotmatics) using one way Anova followed by pairwise comparisons with Tukey’s test.

Immunohistochemistry and imaging

For the analysis of adult neuronal projection from UAS CD8GFP, UAS H2BYFP,UASmyrGFP, lexAopNLStomato or QUAS mtdtomato3xHA expressing brains, ventral nerve cords or genital tracts, tissues were dissected in PBS (137 mM NaCl, 10 mM phosphate, 2.7 mM KCl, pH7.4), fixed in 4% (w/v in PBS) paraformaldehyde for 15 minutes, washed three times in PBST (PBS with 1% BSA and 0.3% Triton-X100), then once in PBS for 10 mins, mounted in Vectashield (Vector Labs) and visualized with confocal microscopy using a Leica TCS SP8. If signals were weak, antibody in-situ stainings were done as described previously (Haussmann et al., 2008) for validation using rat anti-HA (MAb 3F10, 1:20; Roche), rabbit anti-GFP (Molecular Probes, 1:100) and visualized with Alexa Fluor 488 (1:250; Molecular Probes or Invitrogen), Alexa Fluor 546 (1:250; Molecular Probes or Invitrogen) or Alexa Fluor 647 (1:250; Molecular Probes or Invitrogen). For imaging, tissues were mounted in Vectashield (Vector Labs).

Confocal microscopy and image processing

Adult tissues were scanned using a Leica SP8 confocal microscope equipped with a set of fluorescent filters and hybrid detector (HyD). Adult brains were scanned using a 40x HC PL APO 40x/1.30 lens with oil, 1024 x 1024 resolution and 0.96 µm Z-step. VNC and genital tracts were scanned using a HC PL APO CS2 20X/0.75 with oil, 1024 x 1024 resolution and 0.96 µm Z-step. Images were obtained using Leica Application Suite X (LAS X) imaging acquisition software. Raw data files were in LIF format and were processed using FIJI. For high resolution mapping neurons were identified in the virtual fly brain based on registered GAL4 expression and traces retrieved for modelling (Galili et al., 2022; Phelps et al., 2021; Scheffer et al., 2020).

Supplementary information

Supplementary Figure S1: Expression analysis of PMR-inducing GAL4 in the genital tract.

A-E) Representative adult female genital tracts expressing UAS CD8GFP under the control of SPR8, SPR12, fru11, fru12 and dsx24 GAL4, and LexAop NLStomato under the control of elavLexA. Arrows indicate genital tract sensory neurons. The insert shows expression of GFP in the genital tract sensory neurons. Scale bars shown in A and insets are 100 µm and 20 µm, respectively.

Supplementary Figure S2: Expression analysis of non-PMR-inducing GAL4 in the genital tract.

A-C) Representative adult female genital tracts expressing UAS CD8GFP under the control of SPR3, SPR21 and fru9 GAL4. Arrows indicate genital tract sensory neurons. The insert shows expression of GFP in the genital tract sensory neurons. Scale bars shown in C and insets are 100 µm and 20 µm, respectively.

Supplementary Figure S3: Expression analysis of split-GAL4 in the genital tract.

A-E) Representative adult female genital tracts expressing UAS CD8GFP under the control of SPR8 ∩ fru11/12, SPR8 ∩ dsx, SPR8 ∩ FD6, fru11/12 ∩ dsx and fru11/12 ∩ FD6 split-GAL4 intersectional patterns. The scale bar shown in E is 100 µm.

Supplementary Figure S4: PMRs after neuronal inhibition, ablation or activation of distinct circuits from intersection of SPR, fru, dsx and FD6 patterns in the brain and VNC.

A-F) Receptivity (A, C and E) and oviposition (B, D and F) of wild type control virgin (red) and mated (orange) females, and virgin females expressing either UAS TNT (azure, A and B) or UAS reaper hid to inhibit or ablate neurons (yellow, C and D), respectively, or UAS NaChBac (brown, E and F) to activate neurons in SPR8 ∩ fru11/12, SPR8 ∩ dsx, SPR8 ∩ FD6, fru11/12 ∩ dsx and fru11/12 ∩ FD6 split-Gal4 patterns shown as means with standard error from three repeats for receptivity (21 females per repeat) by counting the number of females mating within a 1 h period or for oviposition by counting the eggs laid within 18 hours from 30 females. Statistically significant differences from ANOVA post-hoc comparison are indicated by letters (p≤0.0095 in A and B, p<0.0001 in C and D except p=0.016 for c in D, p<0.0001 in E and p<0.0002 in F).

Supplementary Figure S5: ppk is not part of the SPR8, SPR12 and fru11/12 PMR-inducing neuronal circuitry

A, B) Receptivity (A) and oviposition (B) of wild type control virgin (red) and mated (orange) females, and virgin females expressing UAS mSP (green) under the control of GAL4 in ppk or in SPR8 ∩ ppk, SPR12 ∩ ppk, and fru11/12 ∩ ppk patterns shown as means with standard error from three repeats for receptivity (21 females per repeat) by counting the number of females mating within a 1 h period or for oviposition by counting the eggs laid within 18 hours from 30 females. Statistically significant differences from ANOVA post-hoc comparison are indicated by different letters (p<0.0001).

C-K) Representative adult female brains, ventral nerve cords (VNC) and genital tracts expressing UAS CD8GFP under the control of UAS by SPR8 ∩ ppk, SPR12 ∩ ppk, and fru11/12 ∩ ppk. Scale bars shown in E are 50 µm and in H and K are 100 µm, respectively. L-O) Receptivity (L, N) and oviposition (M, O) of wild type control virgin (red) and mated (orange) females, and virgin females expressing either UAS TNT (azure) or UAS NaChBac (brown) to inhibit or activate neurons in SPR8 ∩ ppk, SPR12 ∩ ppk, and fru11/12 ∩ ppk patterns shown as means with standard error from three repeats for receptivity (21 females per repeat) by counting the number of females mating within a 1 h period or for oviposition by counting the eggs laid within 18 hours from 30 females. Statistically significant differences from ANOVA post-hoc comparison are indicated by different letters (p<0.001 for b, and p<0.01 for c in L and N).

Supplementary Figure S6: trans-Tango identifies post-synaptic proceeding neurons of SP targets in the VNC, but not the genital tract

A-O) Representative adult female ventral nerve cords (VNC, A-O) and genital tracts (P-AD) expressing UAS myrGFP; QUAST tomato3xHA trans-Tango in SPR8 ∩ fru11/12, SPR8 ∩ dsx, SPR8 ∩ FD6, fru11/12 ∩ dsx and fru11/12 ∩ FD6 split-GAL4s. The presynaptic (A-E and P-T) and postsynaptic (F-J and U-Y) neuronal circuitries are shown in an inverted grey background and the merge is shown in colour. In the merged picture (K-O and Z-AD), the pre-synaptic and post synaptic neuronal circuitry is shown in green and magenta, respectively. Scale bars shown in O and AD are 100 μm.