Evaluating the efficacy of the use of trojan male brook trout with two Y chromosomes (M_YY) requires a better understanding of reproductive performance. We measured the reproductive performance of hatchery age-0 and age-1 M_YY brook trout compared to hatchery XY males using laboratory crosses. Offspring of XY males had higher survival than offspring of age-1 M_YY one day post-fertilization but not offspring of age-0 M_YY. We found no detectable differences in survival from eyed-egg to the juvenile-fry stage. However, size-at-age differed, where offspring of age-0 M_YY were 3.6% smaller in length and 25.2% smaller in weight than those of XY males. For crosses fertilized by both M_YY and XY males, we found that a significantly higher proportion of offspring within families were sired by M_YY versus XY males. These results show, under controlled conditions, evidence for possible fitness advantage for M_YY under sperm competition, but a possible fitness disadvantage associated with early growth of their offspring. Overall, our results hold promise for the use of M_YY brook trout to serve as an effective eradication tool. 

M_YY brook trout were obtained as eyed-eggs from Hayspur Hatchery (Idaho) and reared at Abernathy Fish Technology Center. Females were assessed for ripeness weekly to determine when to begin spawning. Females that were considered ripe (eggs readily released when light pressure was applied to the abdomen) were then placed into a separate holding area to prepare for spawning. Before spawning, all three male fish (XY male, age-0 M_YY, and age-1 M_YY) to be used for each cross were anesthetized using MS-222, then weighed (g) and measured for fork length (mm). The female was euthanized using MS-222, wiped dry, and a ventral incision was made and all eggs were manually expelled into a colander to remove excess fluid. The overall weight (g) of the eggs was determined, and eggs were divided into four evenly weighted groups and placed into four large Petri dishes. Each group was then fertilized using milt from one male type: XY male (cross-type 1), age-0 M_YY (cross-type 2), age-1 M_YY (cross-type 3), or a mixture of milt from the XY and age-0 M_YY (cross-type 4). For each female, we randomized the order of fertilization across each cross-type. The collection of milt and fertilization of eggs for each cross was as follows: One male was randomly chosen, anesthetized, and wiped dry. Milt was then expressed into a Petri dish, collected using a 200 to 1000µL pipette, and deposited directly into the Petri dish containing the eggs for that cross. We standardized milt volume, but not sperm concentration, for each cross-type for each female by determining the male of the three male types with the smallest total volume of milt and halving it. For cross-type 4, the same volume of milt from the XY and age-0 M_YY used for each of the other crosses for that female was combined within a separate 0.5 mL microcentrifuge tube and gently mixed using the pipette. Half of that mixture was then used to fertilize the eggs so that the same volume of milt was used to fertilize each egg group for each female. For this cross-type, we chose to combine milt between the two male types before fertilization to increase the effects of sperm competition (Compton 2004). Once milt was added to all groups, water was poured onto each cross-type, then gently mixed and allowed to fertilize for 10 minutes. After fertilization, all males were euthanized. Tissue samples were collected from all adults, including females, and stored in 95% ethanol for subsequent parentage analysis.  Fertilized eggs were placed in four circular screened rearing vessels (one for each cross-type) within Heath trays, such that each tray contained the four crosses for a single female.

This procedure was replicated across all 43 females and 43 males that were spawned so that a separate male of each type was used to fertilize one cross-type for each female. Each Heath stack was supplied with ~12^o C well water at a flow rate of 11.4 L per minute. The placement of individual rearing vessels within each Heath tray was randomized by cross-type to account for potential variability in flow from front to back. Twenty-four hours after fertilization, unfertilized eggs were removed and enumerated. 

Early Rearing

Developing embryos were examined daily, and mortalities were removed and enumerated. At the eyed-egg stage, eggs were physically shocked by dropping them into a 600 mL glass beaker, and non-viable eggs were removed twenty-four hours later. Surviving eggs from cross-type 4 were incubated until the eyed-egg stage (Figure 1), whereupon eggs were removed and preserved in 95% ethanol for subsequent dissection and parentage analysis. Surviving eggs from cross-type 3 (age-1 M_YY sire) were incubated until hatching (Figure 1) and were then removed and enumerated. Eggs from cross-types 1 (XY) and 2 (age-0 M_YY sire) were incubated until the swim-up fry stage with continued daily checks for non-viable eggs. Once the swim-up fry stage was reached (after yolk-sac depletion), individuals from crosses 1 and 2 were transferred into 57 L fiberglass tanks supplied with 12°C well water at 1.9 L per minute. The fry was then fed a daily ration of 3.5% of body weight (biomass)/day. Tanks were monitored and cleaned daily, and all mortalities were removed and enumerated. After approximately one month (post swim-up fry stage), we standardized fish density by selecting a random sample of 80 individuals from each cross (i.e., tank) for continued rearing. The remaining fish within each cross (in excess of 80) were removed, euthanized with an overdose of MS-222, and enumerated. The 80 individuals remaining in each tank were reared to the juvenile-fry stage (Figure 1) with continued daily tank cleaning and mortality checks. At the juvenile-fry stage, all fish were removed from tanks and euthanized with an overdose of MS-222. All crosses were then photographed for subsequent measurement of length using the ImageJ software package (Schneider et al. 2012). We also weighed (g) each offspring for a subset of six of the crosses: cross-types 1 and 2 from three randomly chosen females.

Survival and Growth Analysis

Survival was calculated for six stages of development: 1-day post-fertilization (1 dpf; the measure of fertilization success), eyed-egg (30 dpf), hatch (40 dpf), swim-up fry (65 dpf), post swim-up fry (105 dpf), and juvenile-fry (150 dpf). For survival 1-day post-fertilization, we were unable to differentiate between mortalities that occurred post-fertilization and eggs that were unfertilized. Therefore, we classified all dead eggs as eggs that were unfertilized and then used the surviving eggs as a measure of fertilization success. For cross-types 1 and 2, survival to each stage was calculated. For cross-type 3, only survival to 1 day post-fertilization, the eyed-egg stage, and hatch stage were calculated. This was due to limited space for post-hatching rearing vessels within the laboratory, and individual families for cross-types 1 and 2 needed to be reared separately after hatching. Survival to the first five stages of development was calculated by determining the number of individuals per cross at fertilization, then subtracting the number of mortalities observed to the designated developmental stage (Figure 2). We then calculated the proportion survival of each cross to each stage by taking the survival of each cross at each life stage and dividing by the total number of individuals per cross at fertilization. Survival to juvenile-fry stage for cross-types 1 and 2 was determined by subtracting the number of individuals that survived to 150 days post-fertilization from the number of individuals at post-swim-up fry stage (determined upon subsampling to 80 fish per cross). We then calculated the proportion survival of each cross to the juvenile-fry stage by taking the total number survived for each cross at the juvenile-fry stage and dividing it by the total number of individuals per cross at the post-swim-up fry stage.

Differences in the proportion of offspring survival to each developmental stage for cross-types 1-3 were analyzed using a Bayesian model with a beta-binomial distribution in Stan (Stan Development Team 2023) using the package "brms" in R (Bürkner 2018) with the female as a random intercept. Four parallel chains were run. The number of MCMC iterations was set at 5,000 and burn-in was set at 1,000. All priors were drawn from uniform distributions. Convergence between chains was assessed using the statistic Ȓ (Brooks and Gelman 1998) where Ȓ values of less than < 1.01 suggest appropriate convergence of the model (Gelman et al. 2014).

Growth differences at the juvenile-fry stage were examined by comparing length (mm) between all offspring of XY and all offspring of age-0 M_YY males and comparing weight (g) for the subset of six crosses using a glmm. For the growth analyses, the total number of individual juvenile-fry within each cross was used as a random intercept to account for the density of tanks with either length or weight as the dependent variable and cross-type as a covariate. All glmm’s were fit using the R package glmmTMB (R v 4.1.1). 

Parentage of Milt-Mixture Crosses (Cross-Type 4)

The parentage of cross-type 4 offspring was determined through genetic analysis of eyed-eggs. We allowed offspring from cross-type 4 to develop to the eyed-egg stage to ensure we would be able to successfully extract enough embryonic DNA to determine parentage. DNA was extracted from 100 randomly chosen eggs from each cross using a modified version of the extraction method found in Ali et al. (2016). Extracted DNA was then sent to the Idaho Fish and Game (IDFG) Eagle Fish Genetics Lab for library preparation and sequencing analysis. Library preparation was completed following the GTSeq approach developed by Campbell et al. (2015). GTSeq was performed for each individual using an established 240 GTSeq SNP panel. This SNP panel has been used to determine the sire type (either XY or M_YY) of offspring sampled from experimental systems into which M_YY brook trout have been released (Kennedy et al. 2018).

Screening of SNPs

We conducted a locus-specific analysis of conformation to Hardy–Weinberg proportions in a companion brook trout population genomic analysis (Davenport et al. in prep). Loci that showed significant deviation (P < 0.05) from HW proportions across large numbers of population samples (≥ 7) in this analysis were removed (Waples 2015). We did not screen for linkage disequilibrium among loci as that is not an assumption of exclusion-based parentage assignments.

Parentage Estimation for Milt-Mixture Crosses

The parentage of cross-type 4 offspring at the eyed-egg stage was estimated from genotypes based on the exclusion approach implemented by the R package Hiphop (Cockburn et al. 2020). Adults were excluded from possible parentage if the offspring possessed two identical copies of alleles at a given locus and adults possessed two identical copies of the alternate allele at that same locus (i.e. they were alternate homozygotes) or if both possible parents were homozygous for the same allele while the putative offspring was heterozygous at that locus. Each cross was examined separately, and the putative parent with the fewest number of mismatches was the assigned parent. We expected high accuracy of parent assignments for each cross because the female parent was known and there were only two possible male parents (the XY and M_YY used for that cross). Differences in the proportion of offspring that were assigned to an XY compared to an M_YY were tested using a beta-binomial glmm (Gelman and Hill 2007) in R with sire type as the fixed effect and female as a random intercept.