Correlated evolution of elaborate intromission mechanics during copulation between the sexes in leaf beetles

Abstract

Diverse female and male genitalia have evolved. Although the evolutionary mechanisms behind this phenomenon have been studied extensively, the mechanical interactions between female and male genitalia are much less understood. Here, we provide an in-depth biomechanical study on the intromission of elongated tube-like female and male genitalia, termed the spermathecal duct and flagellum, in tortoise leaf beetles. Our findings reveal that this seemingly straightforward penetration mechanism is underpinned by structural and material specializations of the genitalia. We employed synchrotron-based micro-computed tomography and confocal laser scanning microscopy (CLSM) to visualize the precise mechanical interactions between female and male genitalia. Mechanical tests revealed a stiffness gradient within the flagellum and correlated stiffness variations between the sexes. Combining CLSM and cryo-microtome techniques, we characterized a heterogeneous material distribution at the flagellum tip; this specialization was more remarkable in the species with a more sclerotized spermathecal duct. Finite element analyses incorporating the observed material properties demonstrated that the observed material distribution reduced the penetration force required tremendously and dispersed the stress at the flagellum tip. This study unveils that structural and material adaptations in female and male genitalia have probably coevolved and highlights the importance of biomechanics in genital studies.

1. Introduction

Animal genitalia have evolved divergently and rapidly [1]. However, in earlier studies, female genitalia were largely neglected, as they were considered less variable compared with male genitalia [2]. Recent studies focusing on female genital variations have revealed a previously hidden diversity in their morphology [3,4]. The coevolution of female and male reproductive organs is now recognized as a widespread phenomenon (e.g. [5–13]). Moreover, recent studies have shown that at least in some cases, female genitalia evolved even more rapidly than male genitalia [14–16]. This rapid evolution is thought to be driven not only by sexual selection but also by natural selection, which partially shapes female genital diversity, as female genitalia serve multiple functions, including intromission and oviposition/parturition [17–19].

These previous studies highlight the importance of gaining a deeper insight into female genital diversity, genital coupling and evolutionary mechanisms, not only sexual selection but also natural selection and pleiotropy [20]. Genital coupling, in particular, has been extensively studied in arthropods using traditional techniques, such as a combination of instant fixation and mainly dissection/histology (e.g. [21–24]). One of the modern, non-destructive three-dimensional imaging techniques, micro-computed tomography (µCT), has promoted studies on genital coupling (e.g. [25–29]). Understanding genital coupling provides detailed insights into the function of genital elements, and this information helps us understand what mechanism of sexual selection shapes their variations [20,30]. Techniques such as in vivo X-ray cineradiography employed on genital coupling of bush crickets [29], magnetic resonance imaging of human coitus [31,32] and video recording of penile intromission into glass tubes using trained ducks [33] provided a clearer view of genital interactions, highlighting the remarkable shape deformation that genitalia undergo during copulation as suggested at least partially by flush fixation studies mentioned above.

The research field of genital evolution heads towards a comprehensive understanding of genital evolution; nevertheless, the mechanical behaviour of genitalia during copulation has largely been neglected [30,34]. The female and male genitalia are not static structures but rather undergo significant shape alterations during copulation, and the interaction of the genitalia is ‘much more complicated than sliding Tab A into Slot B’ [34]. Given that biomechanics builds a bridge between morphology and functionality/functions and selection pressure does not act directly on morphological traits but on their functionality and functions, biomechanical insights are crucial to address how genital morphology responds to selection pressure and how functional constraints affect genital diversification. To the best of our knowledge, only a few series of studies have examined the biomechanics of genitalia, with focuses on vertebrate penile erection (reviewed in Kelly [34]), harvestmen [35,36], traumatic insemination in bed bugs [37,38] and our series of studies on penetration mechanics of tortoise beetles [39–42] in addition to a few individual biomechanical studies (see examples reviewed in [18] for vertebrate genital studies). However, none of these studies has yet fully focused on the mechanics of genitalia interactions, rather on either female or male structures.

Our study species, the tortoise beetles Cassida rubiginosa Müller, 1776 and C. vibex Linnaeus, 1767 (Coleoptera, Chrysomelidae and Cassidinae), exhibit distinct thicknesses of the spermathecal duct and flagellum despite their comparable lengths (figure 1a,b, [39,43]). The flagella are 10.2 and 10.9 mm, and the spermathecal ducts are much shorter, 3.1 and 2.9 mm in C. rubiginosa and C. vibex, respectively [39,43,44]. These species sometimes occur sympatrically, feed on thistle plants (Cirsium spp.) and engage in mounting behaviour even in heterospecific pairings [45,46], suggesting that they are closely related. The female spermathecal duct is regularly coiled with 29 and 7 coils, on average, in C. rubiginosa and C. vibex, respectively [39,43], as seen in many other tortoise beetles [47–49]. The cassidine male has the flagellum as a sperm-transfer organ in addition to a usual intromittent organ called aedeagus, which is inserted into the vagina. The male flagellum is a tube-like cuticular structure situated in the ejaculatory duct at rest [43]. The ejaculatory duct is surrounded by circular and longitudinal muscles, and the contraction of the longitudinal ones generates the penetration force, meaning that the flagellum is simply pushed at the base towards the spermathecal duct [43]. Despite this simple penetration mechanism, the flagellum is deeply and quickly inserted into the coiled spermathecal duct [41].

Figure 1.

Spermathecae and flagella of C. vibex (a,c) and C. rubiginosa (b,d). Samples from mature and teneral individuals are shown in (a,b) and (c,d), respectively. Asterisks * and ** denote the base and apex of the spermatheca, respectively, and arrowheads point to the very tip of the flagellum in C. vibex and the coiled tip region in C. rubiginosa.

Focusing on Cassida beetles, the present study aimed to address the following questions: (i) How do females and males manage penetration of the fine elongated flagellum within the coiled spermathecal duct? (ii) What role do structural and material specializations at the flagellum tip play in genital interaction? For this purpose, we combined various techniques and documented (a) precise shape variations of the genitalia (curvature and geometry), (b) flexural rigidity measured with a three-point bending test, (c) behaviour of the flagellum tip, (d) genital coupling with a combination of synchrotron µCT (SR-µCT) and confocal laser scanning microscopy (CLSM), (e) material composition and properties using CLSM, and (f) performance of the flagellum tip during penetration by employing modelling with finite element analyses (FEA). We also discuss the broader implications of mechanical insights for the study of genital evolution.

2. Material and methods

2.1. Animals

Adult and larval beetles C. rubiginosa and C. vibex (6.5 and 5.9 mm in body size [43,44]) were collected in Kiel and Jena, Germany, in 2017−2019. Cassida rubiginosa is abundant at both sites and copulates much more often in captivity, while C. vibex is much less abundant and their genitalia are much easier to handle due to their larger size. Therefore, we used both of them in a compensational way for each experiment.

All samples were prepared in 0.1 M phosphate-buffered saline (PBS, pH = 7.4, Carl Roth GmbH & Co. KG, Karlsruhe, Germany) under an Olympus SZX12 stereomicroscope (Olympus Corporation, Tokyo, Japan) with the exception of copulating pairs fixed with cooled 99.5% ethanol (approx. −20°C). If necessary, we photographed specimens using a Leica M205A stereomicroscope equipped with a Leica DFC420 camera, and the photographs were stacked with the aid of the software LAS V3.8 (Leica Microscopy GmbH, Wetzlar, Germany).

2.2. Curvature measurement

The curvature of the spermathecal duct was quantified for each coil based on published micrographs (n = 10 for C. rubiginosa, n = 9 for C. vibex [43]). For the male of C. vibex, five micrographs taken for the present study were utilized for the curvature measurement as well. Six flagella of C. rubiginosa were submerged in either distilled water, 0.1 M PBS, or 1 M PBS and immediately photographed under the stereomicroscope Leica M205 A. Those flagella were visualized after 20 h, 5 days and 8 days again to observe the effect of osmotic pressure on the curvature. Using the software Fiji [50], curvature (radius⁻¹) was calculated.

2.3. Stiffness characterization

To test the generality of our previous finding on C. rubiginosa that the flexural rigidity of the flagellum is not constant along the structure with the more flexible flagellum apical region [41], we applied the previously used experimental set-up of the three-point bending test. Based on the micrograph series capturing flagellum deformation during the experiments (figure 2a), we calculated the flexural rigidity and elastic modulus as previously described in detail [41]. Briefly, we conducted the three-point bending test under an inverted microscope Axio Observer A1 (Carl Zeiss Microscopy GmbH). Three flagella were used. Each of them was submerged in a PBS droplet on a cover glass, and the very tip and very base were fixed with fine insect pins to gently straighten the entire structure. Thus, the very tip and very basal regions of the flagellum were not analysed.

Figure 2.

Bending experiment (a), flexural rigidity (b) and elastic modulus (c) of the flagellum of C. vibex. Each individual (n = 3) is shown as a different symbol.

We used a handmade hook with the arm distance being adjusted to approximately 300 µm and set two supporting points of the three-point bending test. A load was applied approximately at the middle between the supporting points using a stiffer cantilever prepared from a glass capillary since the flagellum of C. vibex was markedly less pliable than that of C. rubiginosa. The spring constant of the cantilever was 1.35 N m⁻¹, which was estimated using an ultramicrobalance (UMX2, Mettler-Toledo Inc., Columbus, OH, USA). The displacement of the glass capillary was controlled with a three-axial nano- and micro-positioning system (F-130, Physik Instrumente GmbH & Co. KG, Karlsruhe, Germany), with a step size of 20 µm and a velocity of 100 µm s⁻¹. The flagellum deformation was recorded using a camera Basler piA1900-32g (Basler Vision Technologies, Winterbach, Germany) and the software StreamPix 5 (NorPix Inc., Montreal, Canada). For the calculation of elastic modulus, we assumed that the flagellum is a hollow cylinder with an external radius of 9.00 µm and a wall thickness of 3.97 µm, which were estimated based on measurements from our previous study [44].

2.4. Genital coupling

Copulating pairs were fixed with cooled 99.5% ethanol (approx. −20°C), and their genital coupling was observed with two methods.

2.4.1. Synchrotron micro-computed tomography/post-processing of tomographic data

SR-µCT of a copulating pair of C. rubiginosa within 99.5% ethanol was performed at the Imaging Cluster of the KIT Light Source. A polychromatic X-ray beam produced by a 1.5 T bending magnet with a spectrum peak of approximately 15 keV was filtered with 0.2 mm aluminium and 0.2 mm beryllium. We employed a fast indirect detector system consisting of a LSO:Tb scintillator [51] and a diffraction-limited optical microscope (Optique Peter, Lentilly, France; [52]) coupled to a 12-bit pco.dimax high-speed camera with 2016 × 2016 pixels [53]. To cover the whole region of interest, the couple was scanned in two height steps. For both scans, we acquired 3000 projections at 70 frames per second at 10× optical magnification, resulting in an effective pixel of 1.22 µm. The control system concert [54] was used for automated data acquisition. Final three-dimensional tomographic reconstructions were performed by tofu [55] and additionally included phase recovery [56], ring removal, 8-bit conversion and mixing of phase and absorption three-dimensional reconstructions to increase the contrast between the background and homogeneous regions while highlighting the edges. Afterwards, both tomograms were merged into a single volumetric image. Segmentation and photographing were performed with Amira 5.4 software (Visualization Sciences Group, Mérignac, France).

2.4.2. Confocal laser scanning microscopy

The coupling between the spermatheca and flagellum of the rest of the pairs was visualized with a Zeiss LSM 700 CLSM (Carl Zeiss Microscopy GmbH, Jena, Germany). Four pairs of C. rubiginosa and one pair of C. vibex were visualized. The sample of C. vibex was re-used from our previous study [43]. The pair had been kept in benzyl alcohol/benzyl benzoate for years till the reuse, and their cuticula was much less melanized. All samples were embedded in glycerine (greater than or equal to 99.5%, free of water, Carl Roth GmbH & Co. KG, Karlsruhe, Germany). The samples were scanned with a wavelength of 555 nm for excitation and with a long-pass emission filter transmitting light greater than or equal to 580 nm. An objective lens with 20× magnification (Zeiss Plan-Apochromat, air immersion, numerical aperture (NA = 0.8)) was applied. Based on the resultant CLSM micrographs, three-dimensional models were reconstructed with Amira.

2.5. Autofluorescence composition

We applied the method established by Michels & Gorb [57] to the spermathecal duct and flagellum to estimate the distribution of the material compositions using the Zeiss LSM 700 CLSM. All samples were embedded in glycerine. For excitation wavelengths, we used 405, 488, 555 and 639 nm laser lines and applied a band-pass emission filter of 420 to 480 nm and long-pass emission filters transmitting light with wavelengths greater than or equal to 490, 560 and 640 nm, respectively. Subsequently, we assigned artificial colours, blue, green and red (50% intensity), respectively, and stacked them into two-dimensional images with maximum intensity projection. We applied an objective lens with either 10× or 20× magnification (both Zeiss Plan-Apochromat, air immersion, NA = 0.8).

For the analyses of the entire spermatheca, flagellum and its tip region, two spermathecae and one flagellum of C. rubiginosa and two spermathecae and three flagella of C. vibex were examined. In addition, three flagella from immature C. vibex were analysed.

We conducted additional analyses of material compositions within cross sections. To prevent any alteration of fluorescent substances of cuticular structures through chemical processes, we got cross sections of cryo-preserved spermathecae and flagella using a Leica EM UC7 ultramicrotome equipped with an EM FC7 cryo-chamber and an EM CRION ionizer (Leica Microscopy GmbH, Germany). Dissected samples were immediately transferred to commercially available gelatin. We cut them into small dice (approx. 1 × 1 mm) and incubated them in 2.3 M sucrose solution at 4°C to prevent ice crystal formation. Then each dice was meticulously placed onto a sample pin and preserved in liquid nitrogen till further experiment. Each sample pin was transferred to the cryo-chamber cooled down with liquid nitrogen. We cut semi-thin sections, approximately 700−900 nm in thickness, with a cryo immuno diamond knife (Diatome Ltd, Switzerland). Subsequently, material compositions of every cross section were studied using the CLSM with the above-mentioned method. The objective lens with 20× magnification was applied. One spermatheca and three flagella from C. rubiginosa and one spermatheca and six flagella from C. vibex were used.

2.6. Behavioural experiment

To visualize the movement of the flagellum tip during the penetration process, we used three flagella of C. vibex. Two of these were re-used after the CLSM observation described above. We applied our previous set-up of breaking stress measurement for a wet mode [37]. Briefly, the flagellum tips were glued to insect pins attached to a force transducer (25 g capacity, World Precision Instruments, Sarasota, FL). The movement of the force transducer was controlled with a PI-M111.12S micromanipulator (Physik Instrumente GmbH & Co. KG), and the flagellum tip was pressed onto sandpaper at a speed of 0.05 µm s⁻¹ with a step size of 0.01 µm. The force generated during this process was recorded with the software AcqKnowledge 3.7.3 (Biopac System Inc., Goleta, CA, USA).

2.7. Geometry of the flagellum tip

To examine the geometry of the flagellum tip, CLSM was employed. We scanned one flagellum tip for each species with a wavelength of 488 nm for excitation and with a long-pass emission filter transmitting light greater than or equal to 490 nm. Objective lenses with 25× (Plan-Apochromat, NA = 0.8) and 40× (Plan-Apochromat, NA = 1.3) magnification were employed with immersion oil for C. vibex and C. rubiginosa samples, respectively. Based on micrographs, we segmented the flagellum tip with Amira, smoothed the models and photographed their cross sections with the software Blender 2.82 a (Blender Foundation, Amsterdam, The Netherlands). Using the cross-sectional images, we calculated the second moment of area with Fiji and the plug-in BoneJ [58].

2.8. Finite element analyses

We focused on C. vibex because both species exhibited fundamentally the same characteristics during the aforementioned inspection. Additionally, the behaviour of the flagellum tip could only be observed in C. vibex due to its larger size. The three-dimensional model of the male C. vibex, developed for the geometric study, was used as the realistic model. For the female, a simplified geometrical spiral model was created.

Based on CLSM images of the flagellum, the flagellum tip was subdivided into three regions, each assigned specific material properties: an elastic modulus of 1 GPa for the thin outer layer (green), 3 GPa for the middle stripe (dotted red) and 100 MPa for the thick inner layer (striped blue). These layers were set with thicknesses of 2, 3 and 4 µm, respectively, based on microscopic measurements. A Poisson’s ratio of 0.45 was assumed for all layers. The density of the outer and middle layers was set to 1200 kg m⁻³, while the inner layer was assigned a density of 1000 kg m⁻³. These properties correspond to the highly sclerotized, less-sclerotized and non-sclerotized cuticles, as measured in experimental studies and used in previous investigations of insect cuticular structures [59–63].

For comparison, six additional male models were developed:

• (1) Simplified model. The green outer layer was omitted, leaving only two regions with differing material properties.

• (2) Reversed model. Similar to the simplified model but with reversed material property assignments.

• (3) Bent model. Identical material properties to the realistic model, but without the swollen part.

• (4) Homogeneous model. A uniform elastic modulus of 1 GPa across the entire model.

• (5) Straight model. A straight tubular model with material properties matching those of the realistic model.

• (6) Straight, swollen model. Similar to the straight model but including the swollen part.

The force required for the male’s flagellum to penetrate the spermathecal duct was evaluated using the Abaqus FEA software package (Simulia, Providence, RI, USA). A force was applied to the base of the flagellum model and incremented in 1 µN intervals. The penetration pressure was then calculated by dividing the force by the cross-sectional area at the flagellum base. Stress distribution within the flagellum was also analysed during the entry and exit phases into the spiral model. For easier comparison, the results were normalized using the min-max normalization method, which rescales the dataset to a specified range of [0, 1].

In all simulations, the flagellum was modelled using S4R elements (four-node curved shell elements) with reduced integration and hourglass control. Given the significantly higher stiffness of the spiral structure, the female spiral was modelled using rigid elements. Friction between the female and male structures was assumed negligible (friction coefficient = 0.1). Mesh convergence analyses were performed for all simulations to ensure results were independent of mesh size.

2.9. Statistics

All data analyses were conducted with R (v. 4.4.1) in RStudio [64], and the results were plotted using the ‘ggplot2’ package [65]. To estimate 95% confidence intervals for the curvature data, we fitted the models. The effects of explanatory factors (fixed effects) on curvature data were analysed by fitting generalized linear mixed effects models using the ‘glmer()’ function in the ‘lme4’ package [66], considering relative positions as explanatory factors. A gamma error structure and a log link function were used. Individuals were included with a random factor for intercepts.

For the same purpose, but for the results of the bending test, the effects of explanatory factors (fixed effects) on flexural rigidity were analysed by fitting generalized linear models using the ‘glm()’ function, considering relative positions as explanatory factors. A gamma error structure and a log link function were used. Due to the small sample size, a random factor was not considered.

3. Results

3.1. Curvature measurement

The recipient structure, the spermathecal duct, is highly coiled (figure 1a,b), and the curvature varies along the duct in both species (figure 3). The calculated curvature values of C. rubiginosa get bigger from the base to the apex, meaning that the radius of curvature becomes smaller towards the apex. In contrast, that of C. vibex shows the opposite trend getting smaller in curvature. The flagellum was naturally coiled, especially at the apical region in both species (figure 1a,b). This nature was not influenced by the osmotic pressure since no particular difference was observed in the flagella submerged in either distilled water or concentrated PBS (electronic supplementary material, figure S1). The measured curvature of the flagellum was smaller (C. rubiginosa: 0.030 ± 0.0058 µm⁻¹ (mean ± s.d.), n = 6; C. vibex: 0.0082 ± 0.00099 µm⁻¹; n = 5) than the smallest values of the spermathecal duct (figure 3). Flagella of teneral beetles of both species are not coiled (figure 1c,d, electronic supplementary material, figure S2).

Figure 3.

Curvature measurements of spermathecal ducts in C. rubiginosa (a) and C. vibex (b). Individuals are shown as different colours and symbols for each species.

3.2. Bending test

The measured flexural rigidity of the C. vibex flagellum showed the same tendency as that of the C. rubiginosa flagellum previously reported [41], meaning that the flagellum shows a bending stiffness gradient along the structure with a stiff base and a flexible apex (figure 2b). Unlike the calculated elastic modulus of C. rubiginosa, which was constant along the flagellum [41], the calculated elastic modulus of C. vibex showed a material stiffness gradient with a hard base and a soft tip (figure 2c).

3.3. Genital coupling

Our SR-µCT observation provided us with more detailed information on the genital interaction of C. rubiginosa than a previously used desktop µCT [43] (figure 4; electronic supplementary material, figure S3). In the observed pair, the membranous and everted endophallus, which is a part of the aedeagus and usually stored in it, was fully inserted into the vagina, and the tip of the endophallic sclerite was placed in the opening of the spermathecal duct (figure 4a–c). The flagellum of C. rubiginosa is too thin to be observed with the scans of the whole pairs; therefore, we used an invasive method, i.e. the dissected genitalia in pairs were visualized using CLSM (figure 4d–f). The flagellum tip located in the endophallic sclerite was visualized in one pair, and the tip region of the flagellum was straightened (figure 4d). In the other samples, the flagellum was placed in the spermathecal duct and the genital interaction was reconstructed for C. rubiginosa (figure 4e) and C. vibex (figure 4f). The data, quantified where the flagellum has contact with the spermathecal duct, show that two of four samples observed in C. rubiginosa have contact only with the outer curves, while one pair shows the opposite trend, mainly contacting the inner curves (table 1). The pair of C. vibex shows the trend that the flagellum has contact with the outer curve of the spermathecal duct more often (table 1).

Figure 4.

Genital coupling of C. rubiginosa (a–e) and C. vibex (f). Arrowhead in (f) denotes the flagellum tip. Abbreviations: ed, ejaculatory duct; end, endophallus; fl, flagellum; ml, median lobe; od, oviduct; scl, sclerite; spd, spermathecal duct, va, vagina including surrounding muscles.

Table 1.

The contact positions of the flagellum in the spermathecal duct at coils. The number of observed conditions is listed for each individual.

	inner curve	outer curve	middle	not visible
C. rubiginosa P3	0	21	0	1
C. rubiginosa P9	0	30	0	2
C. rubiginosa P10	26	10	0	0
C. rubiginosa P11	1	4	5	0
C. vibex	2	9	0	1

3.4. Behavioural experiment

We recorded the behaviour of three wet flagellum tips in C. vibex while they were pressed onto sandpaper. The used flagellum tips of C. vibex had variations in shape and orientation to a sample holder (figure 5 and electronic supplementary material, videos SA1−SA3). In one sample, the swollen part was pressed perpendicular to the sandpaper, and buckling failure occurred just in front of the swollen part (figure 5c). In the other samples, flagella were slightly inclined towards the direction of the movement, and the local bending occurred at the same point, and then the entire tip region speeded up and slipped forward (figure 5a,b). The highest force recorded during this observation was 23.9 mN for the sample buckled. The forces of the bent samples were smaller, i.e. 6.9 and 5.6 mN.

Figure 5.

Behaviour of the flagellum tip while the tip region was pressed onto the substrate. Three samples (a–c) of C. vibex were observed, and their sequential deformation during the test is shown. The short arrows in (a'') and (b') denote local bending, and the long arrows in (a''') and (b'') denote the slip of the tip region. The broken circle in (c''') indicates buckling failure.

3.5. Geometry of the flagellum tip

The geometry of cross sections of the spermathecal duct and flagellum except for the very tip in C. rubiginosa and C. vibex is more or less circular as partially reported (figure 6b,d,e, [41,42,44]). In contrast, the geometry of the flagellum tip is not just cylindrical with a circular cross section, as the second moment of area calculated based on the cross sections shows the remarkable alteration along the flagellum tip in both species (figure 7). The flagellum tip is curved, and the inner curve is partially swollen (arrowheads in figure 7, especially well developed in C. vibex). The highest second moment of area was observed at the swollen part. The discrepancy between the values around the minor and major axes is remarkable in the swollen part and the very tip.

Figure 6.

Confocal laser scanning micrographs of C. vibex. (a) Spermatheca, (b) cross section of spermathecal duct, (c) flagellum, (d) cross section of the basal part of the flagellum, (e) cross section of the middle part of the flagellum, (f,g) cross section of the apical part of the flagellum, (h,i) flagella of teneral males.

Figure 7.

Geometry of the flagellum tip in C. rubiginosa (left) and C. vibex (right). Arrowheads denote the swollen part, and asterisks indicate the very tip. Light blue lines: second moment of area around the minor axis (the larger I value); pale red lines: second moment of area around the major axis (the smaller I value).

3.6. Autofluorescence composition

The spermathecal duct was imaged in green and red for C. rubiginosa and C. vibex, respectively, and no remarkable heterogeneity in autofluorescence composition was observed (figure 6a, electronic supplementary material, figures S4, S5) as previously reported for C. rubiginosa [40] and C. vibex [67], meaning that the spermathecal duct is homogeneously well-sclerotized. The autofluorescence composition of cross sections also showed a homogeneous pattern outwards from the centre (figure 6b).

The flagellum of C. vibex mainly appeared in red, with an exception at the apical region, which appeared in greenish red (figure 6c). The tip region exhibits a rather complicated pattern: the outer curve is imaged in red to green, while the inner curve is blue, especially at the swollen part. The same pattern was observed in that of C. rubiginosa (electronic supplementary material, figure S4) for the longitudinal gradient [40] and the tip characteristics [41]. This pattern indicates that the flagellum is homologously well-sclerotized but the apical region is less sclerotized compared with the base and middle of the flagellum. The outer curve of the tip region is well-sclerotized while the inner curve is enriched with the rubber-like protein, resilin and elastic.

Since all CLSM micrographs of cross sections were taken with separately adjusted laser settings, we cannot compare the colour between micrographs, but the colour pattern. Cross sections cut from the basal and middle regions appeared mainly in red, but in some cases more greenish, indicating relatively well-sclerotized, and in any case, the colour distributes homogeneously from the centre to the circumference (figure 6d,e, electronic supplementary material, figures S4, S5). In contrast, the apical region exhibits other patterns. The main colour is blue, but with reddish patches on the upper side of the shown images (figure 6f,g, electronic supplementary material, figures S4, S5). In the case of the proximal part of the tip region in C. vibex, reddish patches are situated closer to the centre not to the circumference and surrounded by the bluish part, and the reddish well-sclerotized area is heterogeneously distributed (figure 6f).

3.7. Finite element analyses

The calculated pressure needed for the penetration and maximum stress detected during the penetration process are plotted in figure 8 (electronic supplementary material, videos SB1−SB14; for the female reaction forces recorded for calculating the pressure, see electronic supplementary material, figure S6). The realistic model shows the best performance among all simulated models with the second lowest penetration pressure and the lowest stress. The lowest pressure is observed in the reversed model, while the stress is much higher in comparison with the realistic model. The bent (the swollen part is missing) and simplified (the physical property, green, is missing) models show similar performance to the realistic one but their required penetration pressures are higher than those in the realistic model. The omogeneous model requires even higher pressure and experiences higher maximum stress. The straight model shows similar performance to the homogeneous one, while the straight and swollen model gets remarkably higher maximum stress.

Figure 8.

Finite element simulation of the mechanical performance of the flagellum tip.

4. Discussion

This study provides detailed information on the geometry of genitalia including curvature, physical properties, genital coupling, tip behaviour and stress and pressure involved in flagellum intromission. Below, we discuss the implications of these findings on the mechanics of genital intromission and genitalia evolution.

4.1. Intromission mechanics: female–male interaction

In both species studied, the flagellum penetration mechanism is straightforward [43]. The flagellum is housed in the ejaculatory duct surrounded by longitudinal muscles, whose contraction causes its penetration. Thus, the penetration force is applied at the base of the flagellum and pushed into the tightly spiralled spermathecal duct. Our findings, alongside previous work [41,42], revealed the underlying mechanics behind this simple penetration mechanism, i.e. the flagellum is much more flexible than the spermathecal duct. The flagellum of C. rubiginosa shows a stiffness gradient with a flexible apical part and stiffer rest [41]. The spermathecal duct of C. rubiginosa is approximately five times less pliable than the apical one-third of the flagellum inserted into the female duct [42]. The flagellum of C. vibex shows the same structural stiffness gradient as that of C. rubiginosa, but the structural stiffness values are tenfold higher (1−4 × 10^–12 N m² in C. vibex reported here in comparison with 0.1−0.4 × 10^–12 N m² in C. rubiginosa [41]) (figure 2). The well-sclerotized spermathecal duct of C. vibex shows no remarkable deformation, even when manipulated by fine tweezers under a stereomicroscope and is much harder than the flagellum of the conspecific species, implying that the spermathecal duct resists deformation, and only the flagellum flexes during intromission. Given that the flagellum has contact mainly at the outer curves of the coiled spermathecal duct (figure 4, table 1), the spermathecal duct functions as a mould, and the sole actuator of the flagellum is located at its proximal end.

The newly observed nature of the flagellum in both species, i.e. coiling of the apical region, also appears to play an important role. The coiling state of the flagellum must reduce the force required for intromission. However, the measured curvature of the coiling region in the natural condition is lower than that of the spermathecal duct in conspecific females, and we intuitively expected the curvature to change in response to osmotic pressure, possibly caused by a chemical component difference between the ejaculatory duct and spermathecal duct. Contrary to this prediction, our results showed no remarkable alteration in curvature under varied osmotic conditions. The coiling state was less conspicuous in immature samples, implying that sclerotization causes the coiling. Our detailed CLSM analysis of the material distribution, including cross sections (figure 6), shows the concentrically heterogeneous material distribution, which presumably contributes to the coiling state of the apical region of the flagellum.

The ejaculatory duct housing the flagellum is convoluted in the male abdomen [43], and the endophallic sclerite is inserted into the opening of the spermathecal duct, in which the flagellum is completely straightened (figure 4). The male inflated genitalia called endophallus and the female vagina called bursa copulatrix fit tightly with each other (figure 4a–d), assuring the insertion of the endophallic sclerite and then the coiled flagellum. Given that the flagellum goes through dynamic alteration of the shape (convoluted and straightened) before penetration into the tightly coiled spermathecal duct, the smaller curvature of the flagellum relative to the spermathecal duct may reflect adaptive compensation for these varying shapes.

4.2. Intromission mechanics: male tip and female interaction

Our study also unveils the material and structural specialization of the flagellum tip in both species. Similar to C. rubiginosa [41], the flagellum tip of C. vibex is curled with prominent swelling and shows a material gradient with a well-sclerotized outer curve and a resilin-enriched inner curve (figure 6), while the spermathecal duct is uniformly sclerotized (figure 6). Unlike the circular cross sections of the spermathecal duct and most of the flagellum (figure 4e,f, [41,42,44]), the flagellum tip exhibits complex cross-sectional shapes (figure 7). In particular, structural stiffness increases abruptly in the swollen region. This fact explains the tip behaviour observed while being pressed onto a substrate (figure 5), meaning that the vicinity of the swollen region bends much and the entire tip enriched with resilin moves forward suddenly by releasing the accumulated energy. Our previous study, which implemented a numerical simulation of penetration in C. rubiginosa, predicted, for example, an increase in propulsion velocity and associated energy cost at reversal turns of the spermathecal duct [39]. The observed spring-like behaviour of the flagellum tip might occur at the reversal turns and facilitate the penetration process.

Matsumura et al. [41] hypothesized that the material gradient at the flagellum tip in C. rubiginosa enhances the toughness of the tip since the elongated male structures face the risk of breakage that occurs once in a while [68–70]. Our FE simulations demonstrate that the realistic model shows the best performance with the lowest stress accumulation at the tip and the second lowest pressure needed for penetration (figure 8), supporting the ‘toughness-increase’ hypothesis. Additionally, the observed material and structural specializations reduce the necessary pressure for penetration. The reversed model shows even lesser pressure, although the maximum stress is much higher, meaning that the ‘material gradient’ itself is a key for the lesser pressure. Considering that both the homogeneous model (the tip shape is realistic) and the straight and swollen model (the material gradient is realistic) showed tremendously poor performance, the shape and material gradient both contribute to the better performance observed in the realistic model. The minor alteration, i.e. the reduction of the material stiffness gradient (the green part is removed) and the omission of the swollen part, slightly increased the necessary pressure in comparison with the realistic model. The swollen part and material gradient of C. rubiginosa were less conspicuous than those of C. vibex (figure 6, [41]), and the bent and simplified models would represent the performance of C. rubiginosa. Although the flagellum of C. rubiginosa is as long as that of C. vibex, the thickness is much smaller (figure 1, [44]), requiring less penetration force than that of C. vibex. The remarkable swelling and material gradient in C. vibex presumably evolved alongside an enlargement of the wall thickness of the flagellum.

We are aware that the above-mentioned complicated mechanics still lacks a comprehensive understanding of penetration, as the presence of sperm has not been taken into account. The flagellum penetration probably takes place when the spermathecal duct and flagellum are filled with sperm, which may affect flagellum performance. Moreover, females are known to control sperm distribution within the spermatheca using the spermathecal muscles [71,72], implying that females may actively facilitate and/or hinder this penetration process, including sperm transfer. Given that females are capable of ejecting sperm from males with a short flagellum [13], indicating that females can sense the presence of the flagellum, this hypothesis of female contribution warrants further investigation. To achieve a comprehensive understanding, biomechanical studies including these various aspects have to be further accumulated.

4.3. Evolutionary implications of the biomechanical study

Our in-depth biomechanical study on the genital interaction between tube-like female and male structures revealed unexpectedly complicated intromission mechanics as discussed above, confirming Kelly’s statement that the interaction of the genitalia is ‘much more complicated than sliding Tab A into Slot B’ [34] in invertebrates.

The length of the spermathecal duct in tortoise beetles probably coevolved with that of the flagellum [13]. The evolution of at least the flagellum length has been driven by the cryptic female choice [13]. Although several studies highlight coevolutionary patterns between tube-like female and male structures in insects, all of them focus only on length (e.g. [11,13]). Given that our study species are probably closely related, the distinct shape and physical property variations of the tube-like structures between the species are noteworthy, and the structural and material adaptations detected in the present study suggest the coevolution of those properties between the sexes.

In studies on animal genitalia, the morphological diversity of female genitalia and their role in genital evolution had long been neglected [2], and many recent contributions started unveiling them. Our present study highlights that not only morphological variations but also material properties and surface interactions should be taken into account for a comprehensive understanding of genital evolution. Given the unexpectedly complicated mechanics of one of the possibly simplest genital interactions in internal fertilizing animals, we need to elucidate not only evolutionary driving forces but also the roles of functional constraints in genital diversification. Despite the argument emphasizing the importance of biomechanical aspects [30,34], biomechanical studies are very limited, especially for invertebrates. Comprehensive studies are desired to obtain various facets of evidence elucidating the evolutionary mechanisms of genital evolution. The present study clearly shows the potential significance of the biomechanics of genital interactions, and we hope that our study may convince many researchers in the field to pay more attention to the biomechanical aspects of genital evolution in the future.

Ethics

We minimized sample sizes used in the study, and euthanasia of beetles was carried out using appropriate methods.

Data accessibility

SR-µCT data analysed in this study is accessible through the RADAR4KIT repository at [73]. The other data are provided in the electronic supplementary material [74].

Declaration of AI use

We used an AI-assisted technique solely for linguistic improvement of the first draft. We wrote all the texts by ourselves and incorporated some of the linguistic suggestions generated by an AI tool into the draft after a critical assessment of the suggestions. Not a single sentence was generated by the AI tool.

Authors’ contributions

Y.M.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, validation, visualization, writing—original draft; S.R.: formal analysis, investigation, validation, writing—review and editing; T.v.d.K.: data curation, investigation, methodology, validation, writing—original draft; S.G.: methodology, resources, validation, writing—review and editing; A.K.: formal analysis, methodology, validation, writing—review and editing; H.R.: investigation, methodology, supervision, visualization, writing—original draft.

All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Conflict of interest declaration

We declare we have no competing interests.

Funding

This study was partially supported by the German Research Foundation (DFG) funding (MA7400/1-1 and 7400/1-2) and the Japan Society for the Promotion of Science (23H02534) to Y.M.