Evidence of presence of Mediterranean origin fish in the Gulf of Mexico could suggest recent changes in Atlantic bluefin tuna connectivity patterns
Overall, our results support a historical split between the western Atlantic and Mediterranean spawning grounds followed by a subsequent split between the Gulf of Mexico and Slope Sea and trans-Atlantic unidirectional gene flow from the Mediterranean into western Atlantic spawning grounds. While admixture in the Slope Sea is reflected in the larval and juvenile individual genetic profiles, larvae captured in the Gulf of Mexico were pure GOM-like. Previous work suggested weak input of Mediterranean alleles in the larvae collected from the western Gulf of Mexico in the year 2014 (Johnstone et al. 2021). However, we have not detected evidence of such genetic connectivity in larval samples from the western and eastern sides of the Gulf of Mexico collected before this date (from years 2007 to 2010) despite using thousands of SNP genetic markers. Nevertheless, the number of larvae collected in the Gulf of Mexico with individual genetic profile available for this study remains limited (n=27) and the presence of MED-like spawning adults suggests potential genetic connectivity between the Mediterranean Sea and the Gulf of Mexico. Given that Slope Sea individuals’ ancestry proportions cover the range of MED-like individual genetic profiles, it would also be possible that these MED-like individuals have their origin in the Slope Sea. The detection of MED-like individuals in the Gulf of Mexico originating from the Slope Sea is made likely due to its proximity.
Due to the heterogeneous profile of the Slope Sea, the observed proportion of MED-like individuals in the Gulf of Mexico originated in the Slope Sea could only be explained by a high number of Slope Sea MED-like individuals entering the Gulf of Mexico, unless MED-like individuals originated in the Slope Sea under a scenario of even higher inflow from the eastern Atlantic. Otolith microchemistry analyses revealed that genetically MED-like individuals captured in the Gulf of Mexico showed an otolith isotopic composition of oxygen (δ18O) intermediate between the Gulf of Mexico and the Mediterranean spawning areas, suggesting that these were probably not born in the Gulf of Mexico. Interestingly, these intermediate values are consistent with the proposed signature range of a potential third contingent, compatible with a Slope Sea or Mediterranean origin of individuals showing early and/or more intense migratory behavior (Brophy et al. 2020). These observations allow for different possible origins of the MED-like individuals captured in the Gulf of Mexico. Additional observations of the genetic composition of adult ABFT spawning in the Slope Sea coupled with further knowledge about the migratory behavior of Mediterranean ABFT would be needed to assign the origin of these MED-like individuals more accurately.
Regardless of the origin of the MED-like individuals in the Gulf of Mexico, a few dozen migrants exchanged per generation is theoretically sufficient to erase genetic differentiation between populations (Waples 1998, Lowe and Allendorf 2010, Gagnaire et al. 2015). The low FST values reported in this study are common among marine fishes with large population sizes, high dispersal rates and wide-ranging distributions (Hauser and Carvalho 2008, da Fonseca et al. 2022, Fuentes-Pardo et al. 2022). While genetic differentiation of ABFT between the Mediterranean Sea and the Gulf of Mexico could persist despite admixed individuals in the Slope Sea, the number of migrants detected in the Gulf of Mexico is theoretically expected to lead to genetic homogeneity between eastern and western born ABFT and is thus not easily compatible with significant FST values. Histological inspection confirmed the presence of at least one MED-like female which had spawned less than 48 hours before capture in the Gulf of Mexico, suggesting that MED-like individuals spawn in the Gulf of Mexico. One possible explanation for the observed levels of genetic differentiation would be negative selection against Mediterranean genes preventing successful gene flow. Hence, we have explored different sources of genetic variation, which could help to explain the maintenance of genetic differentiation between highly demographically connected populations. More specifically, we have explored the effect of locally adaptive alleles, which could maintain genetic differentiation in the presence of gene flow (Tigano and Friesen 2016) and inter-specific introgression, which can trigger different evolutionary processes, such as the input of adaptive alleles (Huerta-Sánchez et al. 2014) or the contribution to reproductive isolation (Duranton et al. 2020). However, we did not find evidence to confirm the maintenance of genetic differentiation through either local adaptation or reproductive isolation. These would lead to much higher levels of FST at loci involved in incompatibilities and/or selection than the ones observed in this work. Besides, genome-wide and homogeneously distributed genetic differentiation between Mediterranean Sea and Gulf of Mexico reference individuals at neutral alleles reflects that differentiation is not primarily driven by introgression or adaptation, but by the effect of historical long-term isolation between Mediterranean and Atlantic populations, as indicated by their inferred demographic history. Moreover, interbreeding in the Slope Sea implies genetic compatibility between GOM-like and MED-like individuals, which makes the barrier to gene flow hypothesis unlikely to explain the maintenance of genetic differentiation. Another possibility is that selection undetected in this study impedes incorporation of MED-like alleles in the Gulf of Mexico. On one hand, widespread polygenic selection remains difficult to reject based on outlier detection tests as the ones used here, as these are underpowered to detect weakly selected loci, although genetic differences are unlikely maintained by weak polygenic selection in the presence of gene flow. On the other hand, the use of reduced representation sequencing could lead to missed localized selective sweeps. Analyses based on whole genome sequencing data would allow one to detect adaptation signals missed in our study. Alternatively, we propose that the currently observed ancestry patterns could be explained by recent secondary contact following genetic divergence of both ancestral populations after long-term isolation with reduced or no migration, and that the high observed migration rates, in the absence of barriers to gene flow and if sustained over time, could ultimately lead to genetic homogenization and the consequently the loss of genetic differentiation.