Sperm replacement in sperm-storage tubules causes last-male sperm precedence in chickens

ABSTRACT 1. This study elucidated the last-male sperm precedence (LMSP) mechanism in chickens by examining replacement in storage tubules (SSTs) after multiple artificial inseminations (AI) and the effects of seminal plasma (SP) and male breed on sperm replacement in SSTs. 2. Hens were artificially inseminated with fluorescent dye-labelled spermatozoa from White Leghorn (WL) chickens. Secondary AI was conducted 3 d later with sperm labelled with different nuclear fluorescent dye. Percentage of first and second inseminated sperm in SSTs and their logarithmic odds were calculated. The effect of SP on LMSP was examined using (1) Lake’s solution-washed sperm before second insemination, and (2) SP injected continuously after first insemination. Effect of breed difference on sperm replacement was investigated using Barred Plymouth Rock (BP) sperm. 3. Successive WL-sperm inseminations at three-day intervals caused > 70% stored sperm replacement in SSTs. Although SP removal from sperm from second insemination significantly decreased replacement, its intra-vaginal injection did not affect release. Secondary insemination using BP sperm significantly increased replacement. 4. Sperm replacement is a major factor favouring LMSP in domestic chickens. Two fluorescent staining of sperm, and intra-vaginal multiple AI technique have enabled visualisation, differentiation, and quantification of multiple inseminated sperm stored in the SSTs.


Introduction
Although more than 90% of birds have been reported to be monogamous, extra-pair copulations are frequently observed in male and female individuals from many species (Brouwer and Griffith 2019).Post-copulatory sexual selection, comprising sperm competition and cryptic female choice, occurs in the female reproductive tract after mating in more promiscuous animals (Birkhead and Pizzari 2002;Orr and Brennan 2015).Competition is a process by which sperm from multiple males compete to fertilise the same ova (Karr and Pitnick 1999;Parker 1970).On the contrary, cryptic female choice is the biased selection of sperm by females to produce progeny from preferred males (Birkhead 1987;Olsson et al. 1996;Thornhill 1983).Although the mechanism of post-copulation selection has been studied over several decades in avian models (Birkhead and Montgomerie 2020;Firman et al. 2017), the understanding of this phenomenon remains limited.
Sperm storage, wherein the ejaculated sperm are stored in the female reproductive tract, has been observed in most internal fertilisation (Birkhead and Møller 1992;Holt 2011;Orr and Brennan 2015).In birds, the sperm storage tubules (SSTs) embedded in the utero-vaginal junction (UVJ) of the oviduct are the primary sites of sperm storage.Only spermatozoa residing in the SSTs are thought to contribute to fertilisation for the following reasons.(1) after mating, less than 1% of the ejaculated sperm enter the SST and the remaining are ejected by vaginal contraction (Allen and Grigg 1957); (2) stored sperm has been reported to release in response to progesterone to achieve fertilisation, and this phenomenon helps the sperm avoid the descending egg in the oviduct (Ito et al. 2011;Sasanami et al. 2013); (3) SSTs act as the shelter for protecting the sperm from immune response of the oviduct (Das et al. 2008).It is reasonable to suggest that sperm-storage in the SSTs is the key event where post-copulatory selection occurs, because sperm stored in the SSTs may contribute to maintaining long-term fertility in birds.
The LMSP phenomenon, where sperm of the last mated male sires more offspring, has been observed in domestic chickens using multiple AI (Bakst et al. 1994;Bui et al. 2018;Compton et al. 1978).A recent study suggested that the paternity of the offsprings were mainly switched to the last male when multiple AI was performed (Bui et al. 2018).Such LMSP has been speculated to be a consequence of postcopulatory sexual selection.Several models have been suggested to explain the mechanism of LMSP, including sperm stratification, passive sperm loss and sperm replacement (Lessells and Birkhead 1990).
Previous studies have statistically analysed paternal odds of offspring under these model hypotheses and concluded that passive sperm loss was most plausible (Birkhead and Biggins 1997;Lessells and Birkhead 1990).In each of these models, offspring paternity bias is thought to be attributed to uneven sperm storage or utilisation in the SSTs; however, these hypotheses lacked direct observational evidence because it is difficult to individually distinguish the sperm from different males within the SSTs.
A recent study successfully differentiated sperm populations from different males within the SSTs using two fluorescent dye labels (blue vs. red), namely Hoechst 33342 and pHrodo-Red, prior to artificial insemination (Matsuzaki et al. 2021).The following study examined the mechanism of LMSP by investigating the quantity and distribution of stored sperm within the SSTs following the AI of fluorescently labelled sperm.

Ethical statement
All experimental procedures involving the care and use of animals were carried out in accordance with the approved guidelines of the Animal Experiment Committee of Hiroshima University and the Animal Care Committees of Shizuoka University (approval number: 2018A-5).

Animals
One-to two-year-old White Leghorn (WL) and Barred Plymouth Rock (BP) males, and six-to twelve-month-old WL females were maintained under a photoperiod of 14 h light:10 h dark with water and a commercial diet ad libitum.Although sperm motility tests were not performed, all males used in the experiments were confirmed to be fertile.

Seminal plasma preparation
Semen was collected from WL and BP males using the abdominal massage method (Burrows and Quinn 1937), and pooled by breed.The WL and BP semen samples were centrifuged individually at 1,000 × g for 10 min, and the seminal plasma (SP) of both was collected as the supernatant (WLSP and BPSP, respectively).The WLSP and BPSP samples were stored at − 20°C until use.

Artificial insemination
Semen was collected from four roosters by the method described above (Burrows and Quinn 1937) and pooled for AI.The sperm concentration was then determined using a haemocytometer, and the progressive motility of the sperm was observed using a stereo microscope according to the method described by Matsuzaki et al. (2017).Spermatozoa were stained with fluorescent dyes, 10 μM pHrodo-Red AM (red) or 10 μM Hoechst 33342 (blue), prior to AI as previously described (Matsuzaki et al. 2021).Briefly, semen was added 1/1000 volume of 10 mM fluorescent dye solution, mixed gently, loaded at 37°C for 5 min, and used directly for AI without washing.Hens were artificially inseminated with 100 μl/bird of fluorescently labelled sperm, according to the method described by Quinn and Burrows (1936).A portion of the stained sperm was observed under a microscope and confirmed no detectable adverse effects of fluorescent dyes on sperm motility in each trial.Eggs laid by the hens the day after AI were collected and incubated for two days at 37.5°C, and then cracked to confirm fertilisation.Eggs showing a developing embryo in the germinal disc were classified as fertilised, whereas undeveloped eggs with vacuoles in the germinal disc were classified as unfertilised.

Tissue preparation and observation
The UVJ were collected according to the method described by Matsuzaki et al. (2021a).The UVJ tissue was dissected and placed in phosphate-buffered saline (PBS).The mucosa containing the SSTs was excised, and the smooth muscle layer was carefully removed using forceps and scissors under a stereomicroscope.After washing with PBS, the UVJ mucosa was mounted in 95% glycerol and observed under a fluorescence microscope (BX51, Olympus Optics) with a 20× objective lens (UplanApo20X, NA0.70; BX51, Olympus).The pHrodo-Red and Hoechst 33342 signals were detected using fluorescence filter cubes U-WMG2 (Olympus Optics) and U-MWU2 (Olympus Optics), respectively.A previous study found that dye transfer to the adjacent sperm did not occur when the stained sperm were mixed (Matsuzaki et al. 2021).The images were recorded using a digital camera (DP-70; Olympus Optics).In all experiments, a minimum of five images were captured per specimen.

Time-dependent changes in the quantity of resident sperm in the SSTs
The study included time-dependent changes in sperm occupancy in the SSTs after AI to observe passive sperm loss from the SSTs.Semen collected from the WL males was pHrodolabelled and inseminated into the WL females at 15:00.Three to four females were sacrificed every subsequent day at 9:00 and UVJs were collected for five days (a total of 19 hens used).Fluoromicroscopic images of the SSTs were used to count the fluorescent area of the stored sperm within the SSTs per 1 mm 2 of UVJ (ImageJ ver.1.53; https://imagej.nih.gov/ij/).The area occupied by sperm was semi-automatically quantified by manually setting a fluorescence intensity threshold to ensure that only resident sperm in the SSTs were selected as the region of interest, and automatically calculating the sum of these areas.These measurements were repeated three times to confirm that there was little variation in the calculated values.

Effect of SP injection on sperm release from the SSTs
The trial investigated the effect of SP injection on sperm release from the SSTs.After one day of insemination with pHrodo-labelled WL semen, 100 μl of Lake's solution (Lake 1960), WLSP or BPSP were injected intravaginally at 15:00 each subsequent day.After three and five days of AI, the UVJs were collected at 9:00.The fluorescent area indicating sperm on these days was determined as described previously.

Replacement of stored sperm in the SSTs by additional insemination
Successive AIs were performed at three-day intervals to investigate sperm replacement in the SSTs.The experimental design was according to our previous report (Bui et al. 2018).The semen from WL was stained with 10 μM pHrodo-Red AM at 37°C for 5 min and artificially inseminated into WL hens (100 μl/hen) at 15:00.Three days after the first AI, WL or BP semen was stained with 10 μM Hoechst 33342 at 37°C for 5 min and artificially inseminated into the WL hens (100 μl/hen) at 15:00.The effects of SP removal from semen of the second male were also tested.At the second AI, sperm isolated from each WL and BP were centrifuged at 1,000 × g and the pellet was washed with 2.0 ml Lake's solution before staining.After resuspension with equal volume of Lake's solution, the washed sperm was stained with fluorescent dye and used for AI, as described above.A day after the second AI, UVJs were collected at 9:00 and SST images were captured using a fluorescent microscope.Fluorescence areas of pHrodo-Red (first insemination) and Hoechst 33342 (second insemination) were measured using ImageJ, and their percentage to the total stored sperm in the SSTs was calculated.The logarithmic odds of the sperm from second insemination (1−p) to the sperm from first insemination (p) was calculated using the following formula:

Statistical analysis
All data were statistically analysed using the statistical package R version 3.4.2(https://cran.r-project.org/bin/macosx/).To determine whether parametric or non-parametric tests should be selected, normality and homoscedasticity of the data were confirmed using the Shapiro-Wilk test and Bartlett test, respectively.The time-dependency data of sperm release from SST were subjected to regression analysis, and the significance of Pearson's correlation coefficient was tested.The effect of SP injection on sperm release from the SSTs was assessed using one-way ANOVA and Dunnett's test with one factor (Lake's solution, WLSP, BPSP injection, or no injection (control) groups).Two-way ANOVA was performed to evaluate the effects of male breed (WL or BP) and semen treatment (intact or washed with Lake's solution) and their interaction.All data are shown as mean ±S.E.

Time-dependent changes in the quantity of resident sperm in the SSTs
The protocol measured the quantity of the resident sperm in the SSTs by determining the fluorescence area occupied by the sperm using a fluorescence microscope (Figure 1).
The pHrodo Red-labelled spermatozoa were observed in the lumen of SSTs (Figure 1b-d), and the area occupied by the sperm in the SSTs decreased in a time-dependent manner over a period of five days after insemination (Figure 1e; P = 2.27 × 10 −5 , n = 19).The percentage of fertilised eggs obtained during the experiment was 97.5% (39/40).

Effect of SP injection on sperm release from the SSTs
Intravaginal injections of SP after AI were examined to determine if this affected the quantity of sperm stored within the SSTs (Figure 2a).Two days after the P injection (at d 3 after the AI), there were no significant differences between the birds injected with Lake's solution, WL-or BP-derived SP, and the noninjected control group (Figure 2b; n = 4 (control, Lake's solution and WLSP) or 3 BPSP).Likewise, 4 d after the injection (d 5 after the AI), the fluorescent area indicating the sperm within the SSTs showed no variation in all experimental groups, including the control group (Figure 2c; n = 4 all respective groups).The percentage of fertilised eggs obtained during the two experiments were 97.1% (33/34) and 94.5% (52/55).

Replacement of stored sperm in the SSTs by additional insemination
AI was performed twice at defined intervals and the occupancy of the first and second inseminated sperm in the SSTs were observed to investigate if the resident sperm in the SSTs were replaced by those that were introduced during the subsequent insemination (Figure 3).
Four trials were conducted, and the numbers of inseminated sperm (×10 8 sperm/hen) in each trial at first and second AI were as follows: trial #1, 2.67 and 2.50; trial #2, 2.12 and 2.06; trial #3, 2.78 and 2. 86; trial #4, 2.10 and 1.76.The percentage of fertilised eggs obtained during the experiment was 100% (24/24).As shown in Figure 3(b-g), sperm labelled with pHrodo Red (red fluorescence) and those with Hoechst 33342 (blue fluorescence) were distinguished without interference.When inseminated the WL sperm twice with a three-day interval, it was found that the percentage of the second-AI sperm (red) was 74%, which decreased to 38% upon removal of SP (Figure 3h,i); n = 4 (second AI: intact) or 3 (second AI: washed)).
The trial examined if sperm replacement was affected when sperm isolated from different breeds were used for the second AI (Figure 4a).
The numbers of inseminated sperm in all three trials were 3.02 (first AI) and 3.08 (second AI, ×10 8 sperm/hen).The percentage of fertilised eggs obtained during the experiment was 96.3% (26/27).When BP semen was inseminated into WL hens previously inseminated with the same semen, fewer sperm were found in the SSTs (Figures 4b-d).The percentage of second (BP) sperm to the total sperm in the SST was 93%, which reduced to 64% upon the removal of SP at second AI (Figure 4h,i; n = 3 both groups).In all trials, no obvious adverse effects of the fluorescent dyes on sperm motility and viability were observed.
Finally, this was conducted from statistical analysis on the effects of SP removal and sperm origin on replacement.The logarithmic odds of stored sperm within the SSTs were higher when the second AI was performed with different breeds (extra-breed) than with the same breed (withinbreed).
Removal of SP caused a reduction in the log odds in both within-and extra-breed inseminations (Figure 5).
The effects of semen treatment and breed on sperm replacement were assessed using two-way ANOVA, and both factors were found to be statistically significant (semen treatment: P = 0.00585; breed: P = 0.01464); however, the interaction between them was found to be nonsignificant (P = 0.3125).

Discussion
The results showed that the resident sperm in the SSTs were replaced by another sperm population introduced in   a subsequent insemination.This indicated that LMSP occurs in birds, at least partially, during the sperm storage phase in SSTs.Although the mechanism is unknown, it was found that sperm replacement occurred more efficiently when subsequent insemination contained sperm obtained from different male breeds.As it has been speculated that several organisms avoid inbreeding, this phenomenon has the potential to be a good model to investigate the mechanism of female choice.Lessells and Birkhead (1990) proposed the following three models to elucidate the mechanism of LMSP in birds: (1) sperm are stored as stratified layers in the SSTs, with those arriving later present in the upper layers near the entrance of the SSTs and being preferentially used for fertilisation; (2) incoming sperm displace or destroy the stored sperm; or (3) passive sperm loss occurs, where older sperm disappear from the SSTs during storage.Based on this analysis, it was concluded that the most likely mechanism of LMSP involves the replacement of resident sperm by the sperm entering later.This provided further validation through experimental data, which suggested that passive sperm loss may be one of the other major factors in LMSP (Birkhead et al. 1995).In the current study, spermatozoa were not stored as stratified layers in the SSTs (Figure 3d,g); (Figure 4d,g), and the reduction in the amount of resident sperm stored in the SSTs varied depending on the treatment of the second inseminated sperm population (Figures 3h and 4h).Santiago-Moreno et al. (2014) reported that the effect of LMSP in sequential insemination of two different varieties decreased with increased motility of the sperm inseminated first.Their observation was explained by both sperm passive loss and sperm replacement models because the result indicated that a higher storage rate of the first sperm or relatively lower activity of the second sperm may decrease LMSP effects.Although the current results obtained using dualfluorescent dyes were explained by the sperm replacement model, it was reasonable to suggest that the process of sperm passive loss in the SSTs was affected by the trait of subsequently inseminated sperm.Since the experimental techniques used in this study alone provided only limited information, further studies such as direct observation of the process of sperm replacement after subsequent insemination could be a key to uncovering the mechanism of LMSP.These findings showed that the removal of SP from the second spermatozoa increased the relative quantity of the firstinseminated sperm in the SSTs and reduced the apparent LMSP effect (Figures 3h and 4h), which indicated that the degree of first sperm exclusion probably depended on the condition of the second sperm.This result was reasonable because the studies on SP in many species highlights its pivotal roles in successful fertilisation, including augmentation of sperm motility, decreasing immune responses against allogenic spermatozoa and enhancing sperm transport within the oviduct (Juyena and Stelletta 2012;Robertoson 2005;Rodríguez-Martínez et al. 2011).Although direct evidence is not available, some mechanisms by which females actively exclude resident sperm in the SSTs in response to the condition of incoming sperm have been predicted.Sperm that fail to reside in the SST may be removed by an immune response in the oviduct (Higaki et al. 1995), and it is possible that the female uses the SST to sort out sperm that should be protected from the immune response.This study provided a direction to identify such unknown mechanisms that may occur in the female reproductive tract.
The major histocompatibility complexes (MHC), expressed on the sperm surface, is a factor involved in selection within the female reproductive tract (Martin-Villa et al. 1999;Ziegler et al. 2002).The MHC encodes numerous genes necessary for immune responses and recognise self and nonself by antigen display in vivo.In the red junglefowl, higher numbers of sperm reach the egg after mating between males and females with dissimilar MHC class I minor loci (Løvlie et al. 2013).However, this bias was not detectable after AI, which suggested the possibility that females may need unknown male phenotypic cues to select sperm (Løvlie et al. 2013).Moreover, in the present study, sperm selection was observed after AI (Figures 3h and 4h).Although the relevant experiments were not performed to elucidate the reasons for the lack of changes after AI, this could be attributed to species divergence, experimental conditions or differences in methods used.The MHC was not expected to be the primary cause of bias in sperm storage in SSTs.
Oligosaccharides are another candidate molecules that contribute to sperm selection.In chickens, the removal of sialic acid from the sperm surface prevented their passage through the vaginal segment (Steel and Wishart 1996), and, in quail, ABA/ConA-binding carbohydrates on the sperm surface play an important role in their entry into the SSTs (Matsuzaki et al. 2021).Glycoproteins modified by these glycans can regulate the sperm selection in birds.Although this mechanism is uncertain, the present study demonstrated the possibility that sperm selection occurs even after their entry into the SSTs, because the degree of sperm replacement varied depending on the strain used for the second insemination.
The current results showed that single AI produced fertilised eggs for a minimum of five days (Figure 1e), but the storage duration of these sperm decreased after the second AI was performed (Figures 3h and 4h).The injection of SP alone did not reduce the sperm storage rate (Figure 2b,c), hence, these findings indicated that the interaction between sperm, rather than the SP constituents, and the female reproductive tract may respond to the inseminated sperm and exclude the resident sperm in the SSTs.The sperm stored in the SSTs were more efficiently replaced when subsequent AI were performed using different breeds; hence, it is reasonable to speculate that females may sense the genetic distance of sperm and selectively exclude them from more closely related males.
In conclusion, the results indicated that sequential insemination decreased the storage of the initially stored sperm, and this reduction was affected by certain conditions (presence or absence of seminal plasma) and sperm origin (whether from the same or different breeds) of the subsequent AI.Although an alternative explanation was not completely excluded, these findings supported the hypothesis that the sperm replacement model was the most likely mechanism of LMSP in chickens.

Figure 1 .
Figure 1.Temporal changes in the quantity of stored sperm within the SSTs after AI.(a) Experimental outline.Nineteen WL hens were artificially inseminated with WL semen stained with pHrodo Red, and 3-4 hens were sacrificed per day following the day of AI.The UVJ mocosa was isolated and the pHrodo-positive area was measured using a fluorescent microscope.White arrow indicates the time of AI.The endpoint of the black arrow indicates the time of UVJ collection.(b-d) Observation of the UVJ whole-mount specimens three days after AI, showing (b) bright field, (c) Ex = 530-550 nm and Em = 575 nm, and (d) composite image of (b) and (c) (bar = 200 μm).(e) Changes in the pHrodo-positive area (μm 2 /mm 2 UVJ) five days after AI (mean ± SE).The black line represents the regression curve and the grey region indicates 95% confidence interval.

Figure 2 .
Figure 2. Effects of SP injection on the sperm release from the SSTs after AI.(a) Experimental outline.A total of 32 hens were artificially inseminated with WL semen labelled with pHrodo Red, and intravaginally injected with Lake's solution, WLSP, or BPSP from the following day.The hens were sacrificed on day 3 (top) or 5 (bottom) after AI. White arrow indicates the time of AI.Black arrowhead indicates the time of intravaginal injection of test solution; the endpoint indicates the time of the UVJ collection.The area of pHrodo-positive sperm on (b) day 3 and (c) day 5 is shown (mean ± SE).No significant differences were detected amongst the experimental groups (P > 0.05).

Figure 4 .
Figure 4. Replacement of stored sperm in the SSTs by subsequent insemination of semen collected from a different breed.(a) Experimental outline.WL hens were artificially inseminated with WL semen labelled with pHrodo Red, and subsequently inseminated with Hoechst-stained BP semen three days later.The sperm were washed with PBS, stained with Hoechst 33342, and resuspended in Lake's solution to test the effects of SP removal.White and grey arrows indicate the time of AI with WL and BP semen, respectively.The endpoint of the black arrow indicates the time of UVJ collection.(b-g) UVJ whole-mount specimens collected from females inseminated with intact (b-d) or Lake's-resuspended BP semen (e-g) at the second AI (bar = 200 μm).(b,e) Hoechst detection at Ex = 340 nm and Em = 420 nm.(c,f) pHrodo detection at Ex = 530-550 nm and Em = 575 nm.(d,g) Merged images of (b,c) and (d,g), respectively.(h) Total fluorescent area of first (red columns) and second (blue columns) inseminated sperm per 1 mm 2 of UVJ mucosa (mean ± SE of three repeated experiments).(i) The percentage of total fluorescent area of the first (red columns) and second (blue columns) inseminated sperm (mean ± SE of three independent experiments).

Figure 3 .
Figure 3. Replacement of the stored sperm in the SSTs by subsequent insemination.(a) Experimental outline.WL hens were artificially inseminated with WL semen labelled with pHrodo Red.Second AI was performed with Hoechst-stained WL semen three days after the first AI.The sperm were washed with PBS, stained with Hoechst 33342, and resuspended in Lake's solution to determine the effects of SP removal.White arrow indicates the time of AI with WL semen.The endpoint of the black arrow indicates the time of UVJ collection.(b-g) UVJ whole-mount specimens collected from the females inseminated with intact (b-d) or Lake's-resuspended WL semen (e-g) (bar = 200 μm).(b,e) Detected at Ex = 340 nm and Em = 420 nm.(c,f) Detected at Ex = 530-550 nm and Em = 575 nm.(d, g) Merged images of (b,c) and (d,g), respectively.(h) Total fluorescent area of first (red columns) and second (blue columns) inseminated sperm per 1 mm 2 of UVJ mucosa (mean ± SE of three repeated experiments).(i) Percentage of total fluorescent area of first (red columns) and second (blue columns) inseminated sperm (mean ± SE of three repeated experiments).

Figure 5 .
Figure 5. Sperm replacement is more likely to occur when the sperm of different breeds are used for insemination.Sperm replacement was expressed as the logarithmic odds of the area occupied by the second sperm to first sperm (median and interquartile range ± SE). White boxes refer to AI pairs within the same breed (first AI: WL, second AI: WL) and while grey boxes refer to those from a different breed (first AI: WL, second AI: BP).Data were expressed as boxplot (N = 3).